vecId
stringlengths
12
23
id
stringlengths
2
13
conference
stringclasses
11 values
year
float64
2.02k
2.03k
title
stringlengths
6
189
abstract
stringlengths
10
4.74k
author
stringlengths
0
7.45k
aff
stringlengths
0
7.16k
status
stringclasses
11 values
track
stringclasses
4 values
keywords
stringlengths
0
804
github
stringlengths
0
141
site
stringlengths
0
193
gsCitation
float64
-1
11.1k
arxiv
stringlengths
0
12
text
stringlengths
58
4.82k
vector
list
wacv_2025_7e9da3e72f
7e9da3e72f
wacv
2,025
Planar Gaussian Splatting
This paper presents Planar Gaussian Splatting (PGS) a novel neural rendering approach to learn the 3D geometry and parse the 3D planes of a scene directly from multiple RGB images. The PGS leverages Gaussian primitives to model the scene and employ a hierarchical Gaussian mixture approach to group them. Similar Gaussia...
Farhad G. Zanjani; Hong Cai; Hanno Ackermann; Leila Mirvakhabova; Fatih Porikli
Qualcomm AI Research*; Qualcomm AI Research*; Qualcomm AI Research*; Qualcomm AI Research*; Qualcomm AI Research*
Poster
main
https://openaccess.thecvf.com/content/WACV2025/html/Zanjani_Planar_Gaussian_Splatting_WACV_2025_paper.html
1
2412.01931
Planar Gaussian Splatting This paper presents Planar Gaussian Splatting (PGS) a novel neural rendering approach to learn the 3D geometry and parse the 3D planes of a scene directly from multiple RGB images. The PGS leverages Gaussian primitives to model the scene and employ a hierarchical Gaussian mixture approach to g...
[ -0.049824513494968414, -0.01789354719221592, -0.027409030124545097, -0.046310532838106155, -0.011568386107683182, -0.01399118173867464, 0.033364299684762955, -0.013353116810321808, 0.011309461668133736, 0.05696343630552292, -0.041908811777830124, 0.014767955988645554, 0.02761247009038925, ...
wacv_2025_60fafbd26a
60fafbd26a
wacv
2,025
PocoLoco: A Point Cloud Diffusion Model of Human Shape in Loose Clothing
Modeling a human avatar that can plausibly deform to articulations is an active area of research. We present PocoLoco - the first template-free point-based pose-conditioned generative model for 3D humans in loose clothing. We motivate our work by noting that most methods require a parametric model of the human body to ...
Siddharth Seth; Rishabh Dabral; Diogo C Luvizon; Marc Habermann; Ming-Hsuan Yang; Christian Theobalt; Adam Kortylewski
UC Merced; Max Planck Institute for Informatics; Max Planck Institute for Informatics; Max Planck Institute for Informatics; UC Merced; Max Planck Institute for Informatics; University of Freiburg+Max Planck Institute for Informatics
Poster
main
https://github.com/sidsunny/pocoloco
https://openaccess.thecvf.com/content/WACV2025/html/Seth_PocoLoco_A_Point_Cloud_Diffusion_Model_of_Human_Shape_in_WACV_2025_paper.html
0
2411.04249
PocoLoco: A Point Cloud Diffusion Model of Human Shape in Loose Clothing Modeling a human avatar that can plausibly deform to articulations is an active area of research. We present PocoLoco - the first template-free point-based pose-conditioned generative model for 3D humans in loose clothing. We motivate our work by ...
[ 0.003631026018410921, -0.028957773000001907, -0.03986441716551781, 0.016115786507725716, 0.00033320204238407314, -0.01598917506635189, 0.030097272247076035, -0.007718756794929504, -0.00018638400069903582, -0.003334846580401063, -0.003418500302359462, -0.04113052785396576, -0.0194076765328645...
wacv_2025_91fc93cd19
91fc93cd19
wacv
2,025
Point Cloud Color Upsampling with Attention-Based Coarse Colorization and Refinement
Point cloud color upsampling is an important and less explored research topic. State-of-the-art methods colorize points based on the colors of neighboring points and geometric distances. However these methods often suffer from blurring and noise at color boundaries since object textures can have large color variations ...
Kohei Matsuzaki; Keisuke Nonaka
KDDI Research, Inc.; KDDI Research, Inc.
Poster
main
https://openaccess.thecvf.com/content/WACV2025/html/Matsuzaki_Point_Cloud_Color_Upsampling_with_Attention-Based_Coarse_Colorization_and_Refinement_WACV_2025_paper.html
0
Point Cloud Color Upsampling with Attention-Based Coarse Colorization and Refinement Point cloud color upsampling is an important and less explored research topic. State-of-the-art methods colorize points based on the colors of neighboring points and geometric distances. However these methods often suffer from blurring...
[ -0.025320760905742645, -0.0016287871403619647, -0.04683138430118561, -0.019531574100255966, -0.010810799896717072, -0.00959007628262043, 0.009136929176747799, 0.022102490067481995, -0.01755252107977867, 0.004915261175483465, -0.013095032423734665, -0.014075309969484806, 0.008179770782589912,...
wacv_2025_ac5ac3e8a6
ac5ac3e8a6
wacv
2,025
Point-GN: A Non-Parametric Network using Gaussian Positional Encoding for Point Cloud Classification
This paper introduces Point-GN a novel non-parametric network for efficient and accurate 3D point cloud classification. Unlike conventional deep learning models that rely on a large number of trainable parameters Point-GN leverages non-learnable components-specifically Farthest Point Sampling (FPS) k-Nearest Neighbors ...
Marzieh Mohammadi; Amir Salarpour
Sirjan University of Technology, Iran; Sirjan University of Technology, Iran
Poster
main
https://github.com/asalarpour/Point_GN
https://openaccess.thecvf.com/content/WACV2025/html/Mohammadi_Point-GN_A_Non-Parametric_Network_using_Gaussian_Positional_Encoding_for_Point_WACV_2025_paper.html
2
Point-GN: A Non-Parametric Network using Gaussian Positional Encoding for Point Cloud Classification This paper introduces Point-GN a novel non-parametric network for efficient and accurate 3D point cloud classification. Unlike conventional deep learning models that rely on a large number of trainable parameters Point-...
[ -0.0069127157330513, -0.022059760987758636, -0.03252515569329262, -0.05576765164732933, -0.007513042073696852, -0.03788329288363457, 0.0207515861839056, 0.003124834271147847, 0.02702365629374981, 0.02671901322901249, -0.03397669270634651, -0.0075623225420713425, 0.012194693088531494, 0.033...
wacv_2025_5d4fecb1fe
5d4fecb1fe
wacv
2,025
Point-JEPA: A Joint Embedding Predictive Architecture for Self-Supervised Learning on Point Cloud
Recent advancements in self-supervised learning in the point cloud domain have demonstrated significant potential. However these methods often suffer from drawbacks such as lengthy pre-training time the necessity of reconstruction in the input space and the necessity of additional modalities. In order to address these ...
Ayumu Saito; Prachi Kudeshia; Jiju Poovvancheri
Graphics and Spatial Computing Lab, Saint Mary’s University; Graphics and Spatial Computing Lab, Saint Mary’s University; Graphics and Spatial Computing Lab, Saint Mary’s University
Poster
main
https://github.com/Ayumu-J-S/Point-JEPA
https://openaccess.thecvf.com/content/WACV2025/html/Saito_Point-JEPA_A_Joint_Embedding_Predictive_Architecture_for_Self-Supervised_Learning_on_WACV_2025_paper.html
6
Point-JEPA: A Joint Embedding Predictive Architecture for Self-Supervised Learning on Point Cloud Recent advancements in self-supervised learning in the point cloud domain have demonstrated significant potential. However these methods often suffer from drawbacks such as lengthy pre-training time the necessity of recons...
[ -0.0041976263746619225, 0.01171770691871643, -0.03385521471500397, 0.0032104873098433018, -0.001596103422343731, -0.011927931569516659, -0.007997655309736729, 0.030784117057919502, 0.012805388309061527, 0.023746179416775703, -0.005831432994455099, -0.017101271077990532, -0.009770849719643593...
wacv_2025_81a999a413
81a999a413
wacv
2,025
Polarization as Texture: Microscale 3D Shape from Polarized Light Focus
Defocus is a crucial cue for image-based microscale depth estimation yet its measurement depends on spatial appearance changes such as texture. We show that passively observed polarization is responsive to small irregularities of the surface visible in the microscopic world and can be leveraged for focus measure as a s...
Ren Matsumoto; Takahiro Okabe; Ryo Kawahara
Kyushu Institute of Technology*; Okayama University†; Kyoto University‡
Poster
main
https://openaccess.thecvf.com/content/WACV2025/html/Matsumoto_Polarization_as_Texture_Microscale_3D_Shape_from_Polarized_Light_Focus_WACV_2025_paper.html
0
Polarization as Texture: Microscale 3D Shape from Polarized Light Focus Defocus is a crucial cue for image-based microscale depth estimation yet its measurement depends on spatial appearance changes such as texture. We show that passively observed polarization is responsive to small irregularities of the surface visibl...
[ -0.023703349754214287, -0.0076413326896727085, -0.019405387341976166, -0.03395942971110344, -0.01063422579318285, 0.01058811042457819, 0.019663633778691292, -0.010652672499418259, 0.003202719148248434, 0.010173071175813675, -0.059544287621974945, 0.010855580680072308, 0.02980903722345829, ...
wacv_2025_54a0079509
54a0079509
wacv
2,025
PoolAtnRes: Towards Generalisable Differential Morphing Attack Detection
Morphing attacks can successfully deceive face recognition systems resulting in unreliable access control especially in the border control scenario. Consequently the development of Morphing Attack Detection (MAD) algorithms is crucial for detecting morphing attacks based on either a single facial image (S-MAD) or two f...
Raghavendra Ramachandra; Sushma Krupa Venkatesh; Guoqiang Li
Norwegian University of Science and Technology (NTNU), Norway; MOBAI AS, Norway; MOBAI AS, Norway
Poster
main
https://openaccess.thecvf.com/content/WACV2025/html/Ramachandra_PoolAtnRes_Towards_Generalisable_Differential_Morphing_Attack_Detection_WACV_2025_paper.html
0
PoolAtnRes: Towards Generalisable Differential Morphing Attack Detection Morphing attacks can successfully deceive face recognition systems resulting in unreliable access control especially in the border control scenario. Consequently the development of Morphing Attack Detection (MAD) algorithms is crucial for detectin...
[ 0.016952726989984512, 0.012876452878117561, -0.03188636526465416, 0.00638584466651082, -0.037219807505607605, -0.009266856126487255, -0.017533691599965096, 0.024343354627490044, 0.000880970386788249, 0.051505811512470245, -0.044534243643283844, -0.025524331256747246, -0.0002026827132795006, ...
wacv_2025_49390744bc
49390744bc
wacv
2,025
PositiveCoOp: Rethinking Prompting Strategies for Multi-Label Recognition with Partial Annotations
Vision-language models (VLMs) like CLIP have been adapted for Multi-Label Recognition (MLR) with partial annotations by leveraging prompt-learning where positive and negative prompts are learned for each class to associate their embeddings with class presence or absence in the shared vision-text feature space. While th...
Samyak Rawlekar; Shubhang Bhatnagar; Narendra Ahuja
University of Illinois Urbana-Champaign; University of Illinois Urbana-Champaign; University of Illinois Urbana-Champaign
Poster
main
https://github.com/SamyakR99/PositiveCoOp
https://openaccess.thecvf.com/content/WACV2025/html/Rawlekar_PositiveCoOp_Rethinking_Prompting_Strategies_for_Multi-Label_Recognition_with_Partial_Annotations_WACV_2025_paper.html
0
PositiveCoOp: Rethinking Prompting Strategies for Multi-Label Recognition with Partial Annotations Vision-language models (VLMs) like CLIP have been adapted for Multi-Label Recognition (MLR) with partial annotations by leveraging prompt-learning where positive and negative prompts are learned for each class to associat...
[ -0.10603615641593933, -0.03855202719569206, 0.0051896958611905575, 0.010786249302327633, -0.017377344891428947, -0.014909074641764164, -0.010524052195250988, -0.00659109465777874, 0.00913169514387846, 0.013164106756448746, -0.029854314401745796, -0.05515182018280029, -0.010288978926837444, ...
wacv_2025_c3103e6061
c3103e6061
wacv
2,025
PostoMETRO: Pose Token Enhanced Mesh Transformer for Robust 3D Human Mesh Recovery
With the recent advancements in single-image-based 3D human pose and shape estimation (3DHPSE) there is a growing amount of works that can achieve good results on standard benchmarks but struggle to yield accurate human mesh in extreme scenarios like occlusion. Previous works propose to leverage 2D poses to help 3D HPS...
Wendi Yang; Zi-Hang Jiang; Shang Zhao; S. Kevin Zhou
;;;
Poster
main
https://github.com/PostoMETRO/PostoMETRO-Paper
https://openaccess.thecvf.com/content/WACV2025/html/Yang_PostoMETRO_Pose_Token_Enhanced_Mesh_Transformer_for_Robust_3D_Human_WACV_2025_paper.html
2
2403.12473
PostoMETRO: Pose Token Enhanced Mesh Transformer for Robust 3D Human Mesh Recovery With the recent advancements in single-image-based 3D human pose and shape estimation (3DHPSE) there is a growing amount of works that can achieve good results on standard benchmarks but struggle to yield accurate human mesh in extreme s...
[ -0.0015416829846799374, -0.030726775527000427, -0.032213855534791946, 0.015521390363574028, -0.012714527547359467, 0.00874123815447092, 0.018133072182536125, 0.0342399999499321, 0.020688990131020546, -0.0034876656718552113, -0.017380239441990852, -0.0003799022815655917, 0.03180490806698799, ...
wacv_2025_6aafbd33bc
6aafbd33bc
wacv
2,025
Pre-Capture Privacy via Adaptive Single-Pixel Imaging
As cameras become ubiquitous in our living environment invasion of privacy is becoming a significant concern. A common approach to privacy preservation is to remove personally identifiable information from a captured image but there is a risk of the original image being leaked. In this paper we propose a pre-capture pr...
Yoko Sogabe; Shiori Sugimoto; Ayumi Matsumoto; Masaki Kitahara
NTT Corporation, Japan; NTT Corporation, Japan; NTT Corporation, Japan; NTT Corporation, Japan
Poster
main
https://openaccess.thecvf.com/content/WACV2025/html/Sogabe_Pre-Capture_Privacy_via_Adaptive_Single-Pixel_Imaging_WACV_2025_paper.html
0
2407.00991
Pre-Capture Privacy via Adaptive Single-Pixel Imaging As cameras become ubiquitous in our living environment invasion of privacy is becoming a significant concern. A common approach to privacy preservation is to remove personally identifiable information from a captured image but there is a risk of the original image b...
[ -0.03949077054858208, -0.03172511234879494, -0.027566252276301384, 0.01294583361595869, -0.010608774609863758, -0.06576888263225555, 0.0652904286980629, 0.008570749312639236, -0.003691908670589328, 0.0013306973269209266, -0.043023962527513504, -0.016911473125219345, -0.013534698635339737, ...
wacv_2025_71ffd4e444
71ffd4e444
wacv
2,025
Pre-Trained Multiple Latent Variable Generative Models are Good Defenders Against Adversarial Attacks
Attackers can deliberately perturb classifiers' input with subtle noise altering final predictions. Among proposed countermeasures adversarial purification employs generative networks to preprocess input images filtering out adversarial noise. In this study we propose specific generators defined Multiple Latent Variabl...
Dario Serez; Marco Cristani; Alessio Del Bue; Vittorio Murino; Pietro Morerio
Istituto Italiano di Tecnologia, Italy+University of Genoa, Italy; University of Verona, Italy; Istituto Italiano di Tecnologia, Italy; Istituto Italiano di Tecnologia, Italy+University of Verona, Italy+University of Genoa, Italy; Istituto Italiano di Tecnologia, Italy
Poster
main
https://github.com/SerezD/gen_adversarial
https://openaccess.thecvf.com/content/WACV2025/html/Serez_Pre-Trained_Multiple_Latent_Variable_Generative_Models_are_Good_Defenders_Against_WACV_2025_paper.html
0
2412.03453
Pre-Trained Multiple Latent Variable Generative Models are Good Defenders Against Adversarial Attacks Attackers can deliberately perturb classifiers' input with subtle noise altering final predictions. Among proposed countermeasures adversarial purification employs generative networks to preprocess input images filteri...
[ -0.046196166425943375, -0.020034827291965485, -0.02247020974755287, 0.007929263636469841, -0.0032511407043784857, -0.03578871116042137, -0.010702365078032017, 0.004076411947607994, -0.0019181016832590103, 0.010845063254237175, -0.026560893282294273, -0.04048823565244675, -0.05742175877094269...
wacv_2025_4a70621c97
4a70621c97
wacv
2,025
Precise Integral in NeRFs: Overcoming the Approximation Errors of Numerical Quadrature
Neural Radiance Fields (NeRFs) use neural networks to translate spatial coordinates to corresponding volume density and directional radiance enabling realistic novel view synthesis through volume rendering. Rendering new viewpoints involves computing volume rendering integrals along rays usually approximated by numeric...
Boyuan Zhang; Zhenliang He; Meina Kan; Shiguang Shan
Key Lab of AI Safety, Institute of Computing Technology, CAS, China; Key Lab of AI Safety, Institute of Computing Technology, CAS, China; Key Lab of AI Safety, Institute of Computing Technology, CAS, China+University of Chinese Academy of Sciences, China; Key Lab of AI Safety, Institute of Computing Technology, CAS, Ch...
Poster
main
https://openaccess.thecvf.com/content/WACV2025/html/Zhang_Precise_Integral_in_NeRFs_Overcoming_the_Approximation_Errors_of_Numerical_WACV_2025_paper.html
0
Precise Integral in NeRFs: Overcoming the Approximation Errors of Numerical Quadrature Neural Radiance Fields (NeRFs) use neural networks to translate spatial coordinates to corresponding volume density and directional radiance enabling realistic novel view synthesis through volume rendering. Rendering new viewpoints i...
[ -0.019795304164290428, -0.009704122319817543, -0.003990394063293934, -0.014689810574054718, 0.0041355411522090435, -0.019445106387138367, 0.024458443745970726, -0.022854913026094437, 0.0232972651720047, 0.022283539175987244, -0.023997658863663673, -0.0016357850981876254, -0.01266235858201980...
wacv_2025_3cee677dff
3cee677dff
wacv
2,025
Predicting Event Memorability using Personalized Federated Learning
Lifelog images are very useful as memory cues for recalling past events. Estimating the level of event memory recall induced by a given lifelog image (event memorability) is useful for selecting images for cognitive interventions. Previous works for predicting event memorability follow a centralized model training para...
Sourasekhar Banerjee; Debaditya Roy; Vigneshwaran Subbaraju; Monowar Bhuyan
Uppsala University1+Ume ˚a University3; IHPC, A*STAR Singapore2; IHPC, A*STAR Singapore2; Ume ˚a University3
Poster
main
https://openaccess.thecvf.com/content/WACV2025/html/Banerjee_Predicting_Event_Memorability_using_Personalized_Federated_Learning_WACV_2025_paper.html
1
Predicting Event Memorability using Personalized Federated Learning Lifelog images are very useful as memory cues for recalling past events. Estimating the level of event memory recall induced by a given lifelog image (event memorability) is useful for selecting images for cognitive interventions. Previous works for pr...
[ -0.021600820124149323, -0.03060739114880562, -0.04764886945486069, -0.015761498361825943, -0.028570635244250298, -0.02059178613126278, 0.023450717329978943, 0.012444763444364071, -0.02432895079255104, -0.0006516683497466147, -0.027038397267460823, -0.005199332255870104, -0.04058562219142914,...
wacv_2025_b78bf2fd84
b78bf2fd84
wacv
2,025
PrevPredMap: Exploring Temporal Modeling with Previous Predictions for Online Vectorized HD Map Construction
Temporal information is crucial for detecting occluded instances. Existing temporal representations have progressed from BEV or PV (perspective view) features to more compact query features. Compared to these aforementioned features predictions offer the highest level of abstraction providing explicit information. In t...
Nan Peng; Xun Zhou; Mingming Wang; Xiaojun Yang; Songming Chen; Guisong Chen
Ruqi Mobility; Ruqi Mobility; GAC R&D Center; Ruqi Mobility; GAC R&D Center; Ruqi Mobility
Poster
main
https://github.com/pnnnnnnn/PrevPredMap
https://openaccess.thecvf.com/content/WACV2025/html/Peng_PrevPredMap_Exploring_Temporal_Modeling_with_Previous_Predictions_for_Online_Vectorized_WACV_2025_paper.html
5
2407.17378
PrevPredMap: Exploring Temporal Modeling with Previous Predictions for Online Vectorized HD Map Construction Temporal information is crucial for detecting occluded instances. Existing temporal representations have progressed from BEV or PV (perspective view) features to more compact query features. Compared to these af...
[ -0.05194185674190521, -0.013573207892477512, -0.0212322436273098, 0.03541156277060509, -0.0057350932620465755, -0.037762537598609924, 0.00939471647143364, 0.027936195954680443, 0.003962678834795952, 0.058994781225919724, -0.03089328110218048, -0.03249121084809303, -0.005312652792781591, 0....
wacv_2025_eb0064f706
eb0064f706
wacv
2,025
Prior2Posterior: Model Prior Correction for Long-Tailed Learning
Learning-based solutions for long-tailed recognition face difficulties in generalizing on balanced test datasets. Due to imbalanced data prior the learned a posteriori distribution is biased toward the most frequent (head) classes leading to an inferior performance on the least frequent (tail) classes. In general the p...
S Divakar Bhat; Amit More; Mudit Soni; Surbhi Agrawal
;;;
Poster
main
https://openaccess.thecvf.com/content/WACV2025/html/Bhat_Prior2Posterior_Model_Prior_Correction_for_Long-Tailed_Learning_WACV_2025_paper.html
1
2412.16540
Prior2Posterior: Model Prior Correction for Long-Tailed Learning Learning-based solutions for long-tailed recognition face difficulties in generalizing on balanced test datasets. Due to imbalanced data prior the learned a posteriori distribution is biased toward the most frequent (head) classes leading to an inferior p...
[ -0.04662872105836868, -0.040953148156404495, -0.0223498847335577, 0.019864503294229507, -0.01891857385635376, -0.025781193748116493, 0.005156239029020071, 0.02837786078453064, -0.026875503361225128, 0.005731215234845877, -0.04740772023797035, 0.0006561221089214087, -0.010386668145656586, 0...
wacv_2025_b4b6cdbdac
b4b6cdbdac
wacv
2,025
PrivateEye: In-Sensor Privacy Preservation Through Optical Feature Separation
We address privacy issues in applications where images captured by an edge device (camera) are sent to the cloud for inference on utility tasks such as classification. Sending raw images to the cloud exposes them to data sniffing attacks and misuse by untrusted third-party service providers beyond the user's intended t...
Adith Boloor; Weikai Lin; Tianrui Ma; Yu Feng; Yuhao Zhu; Xuan Zhang
Washington University in St. Louis; University of Rochester; Washington University in St. Louis; Shanghai Jiao Tong University; University of Rochester; Northeastern University
Poster
main
https://openaccess.thecvf.com/content/WACV2025/html/Boloor_PrivateEye_In-Sensor_Privacy_Preservation_Through_Optical_Feature_Separation_WACV_2025_paper.html
0
PrivateEye: In-Sensor Privacy Preservation Through Optical Feature Separation We address privacy issues in applications where images captured by an edge device (camera) are sent to the cloud for inference on utility tasks such as classification. Sending raw images to the cloud exposes them to data sniffing attacks and ...
[ -0.08582331240177155, 0.01967564970254898, -0.03603454679250717, -0.004031165968626738, 0.011271336115896702, 0.001291800057515502, 0.017970426008105278, 0.030337978154420853, 0.013688630424439907, 0.03329869359731674, -0.03534121438860893, -0.0006453144596889615, -0.01717403158545494, -0....
wacv_2025_2a36e8446d
2a36e8446d
wacv
2,025
Pruning One More Token is Enough: Leveraging Latency-Workload Non-Linearities for Vision Transformers on the Edge
This paper investigates how to efficiently deploy vision transformers on edge devices for small workloads. Recent methods reduce the latency of transformer neural networks by removing or merging tokens with small accuracy degradation. However these methods are not designed with edge device deployment in mind: they do n...
Nicholas John Eliopoulos; Purvish Jajal; James C. Davis; Gaowen Liu; George K. Thiruvathukal; Yung-Hsiang Lu
;;;;;
Poster
main
https://openaccess.thecvf.com/content/WACV2025/html/Eliopoulos_Pruning_One_More_Token_is_Enough_Leveraging_Latency-Workload_Non-Linearities_for_WACV_2025_paper.html
1
2407.05941
Pruning One More Token is Enough: Leveraging Latency-Workload Non-Linearities for Vision Transformers on the Edge This paper investigates how to efficiently deploy vision transformers on edge devices for small workloads. Recent methods reduce the latency of transformer neural networks by removing or merging tokens with...
[ -0.02187177911400795, -0.0033344454132020473, -0.029186632484197617, -0.00946200080215931, -0.009835021570324898, 0.031497545540332794, 0.011281616054475307, 0.03053314797580242, -0.001131573342718184, 0.029204828664660454, -0.011290714144706726, -0.015312064439058304, -0.007237520534545183,...
wacv_2025_51caaf3d77
51caaf3d77
wacv
2,025
Psych-Occlusion: using Visual Psychophysics for Aerial Detection of Occluded Persons during Search and Rescue
The success of Emergency Response (ER) scenarios such as search and rescue is often dependent upon the prompt location of a lost or injured person. With the increasing use of small Unmanned Aerial Systems (sUAS) as "eyes in the sky" during ER scenarios efficient detection of persons from aerial views plays a crucial ro...
Arturo Miguel Russell Bernal; Jane Cleland-Huang; Walter Scheirer
Computer Science and Engineering Department, University of Notre Dame, Indiana, USA; Computer Science and Engineering Department, University of Notre Dame, Indiana, USA; Computer Science and Engineering Department, University of Notre Dame, Indiana, USA
Poster
main
https://github.com/ArtRuss/NOMAD
https://openaccess.thecvf.com/content/WACV2025/html/Bernal_Psych-Occlusion_using_Visual_Psychophysics_for_Aerial_Detection_of_Occluded_Persons_WACV_2025_paper.html
0
Psych-Occlusion: using Visual Psychophysics for Aerial Detection of Occluded Persons during Search and Rescue The success of Emergency Response (ER) scenarios such as search and rescue is often dependent upon the prompt location of a lost or injured person. With the increasing use of small Unmanned Aerial Systems (sUAS...
[ -0.040539320558309555, -0.02089131437242031, -0.059565674513578415, -0.01633252389729023, -0.024790775030851364, 0.008312256075441837, 0.043063610792160034, 0.0053876605816185474, 0.03982347622513771, 0.03274039924144745, 0.012715633027255535, -0.017396869137883186, -0.007054821588099003, ...
wacv_2025_b5242d2395
b5242d2395
wacv
2,025
PureForest: A Large-Scale Aerial Lidar and Aerial Imagery Dataset for Tree Species Classification in Monospecific Forests
Sustainable forest management is a cornerstone of climate and environmental action. Responsible management relies on forest models such as biomass or fire vulnerability estimates which depend on mapping the spatial distribution of tree species. Because forest mapping relies heavily on visual identification it is a time...
Charles Gaydon; Floryne Roche
;
Poster
main
https://openaccess.thecvf.com/content/WACV2025/html/Gaydon_PureForest_A_Large-Scale_Aerial_Lidar_and_Aerial_Imagery_Dataset_for_WACV_2025_paper.html
3
2404.12064
PureForest: A Large-Scale Aerial Lidar and Aerial Imagery Dataset for Tree Species Classification in Monospecific Forests Sustainable forest management is a cornerstone of climate and environmental action. Responsible management relies on forest models such as biomass or fire vulnerability estimates which depend on map...
[ -0.06832289695739746, -0.033709339797496796, -0.07367588579654694, -0.049334269016981125, -0.014250514097511768, -0.011031489819288254, -0.005452449433505535, 0.023672491312026978, 0.001797138131223619, 0.016619572415947914, -0.025444762781262398, -0.015643013641238213, 0.017433369532227516,...
wacv_2025_a8c517f346
a8c517f346
wacv
2,025
Q-TempFusion: Quantization-Aware Temporal Multi-Sensor Fusion on Bird's-Eye View Representation
Recent advancements in bird's-eye view (BEV) perception models have highlighted the superior performance of LiDAR-camera fusion systems over single-modality approaches garnering considerable interest in the field. Despite the progress the integration of temporal information a technique that has considerably benefitted ...
Pinrui Yu; Zhenglun Kong; Pu Zhao; Peiyan Dong; Hao Tang; Fei Sun; Xue Lin; Yanzhi Wang
Northeastern University; Northeastern University; Northeastern University; Northeastern University; CVL, ETH Zurich; Meta Inc.; Northeastern University; Northeastern University
Poster
main
https://openaccess.thecvf.com/content/WACV2025/html/Yu_Q-TempFusion_Quantization-Aware_Temporal_Multi-Sensor_Fusion_on_Birds-Eye_View_Representation_WACV_2025_paper.html
0
Q-TempFusion: Quantization-Aware Temporal Multi-Sensor Fusion on Bird's-Eye View Representation Recent advancements in bird's-eye view (BEV) perception models have highlighted the superior performance of LiDAR-camera fusion systems over single-modality approaches garnering considerable interest in the field. Despite th...
[ -0.05679621547460556, -0.021943992003798485, -0.028125397861003876, 0.02445291541516781, -0.02081679366528988, 0.026216434314846992, -0.008344898000359535, 0.036942992359399796, 0.03408863767981529, 0.04636054486036301, -0.016162559390068054, -0.037088435143232346, 0.0078040254302322865, 0...
wacv_2025_ef0e8583e3
ef0e8583e3
wacv
2,025
QuantAttack: Exploiting Quantization Techniques to Attack Vision Transformers
In recent years there has been a significant trend in deep neural networks (DNNs) particularly transformer-based models of developing ever-larger and more capable models. While they demonstrate state-of-the-art performance their growing scale requires increased computational resources (e.g. GPUs with greater memory cap...
Amit Baras; Alon Zolfi; Yuval Elovici; Asaf Shabtai
Ben-Gurion University of the Negev, Israel; Ben-Gurion University of the Negev, Israel; Ben-Gurion University of the Negev, Israel; Ben-Gurion University of the Negev, Israel
Poster
main
https://github.com/barasamit/QuantAttack
https://openaccess.thecvf.com/content/WACV2025/html/Baras_QuantAttack_Exploiting_Quantization_Techniques_to_Attack_Vision_Transformers_WACV_2025_paper.html
0
QuantAttack: Exploiting Quantization Techniques to Attack Vision Transformers In recent years there has been a significant trend in deep neural networks (DNNs) particularly transformer-based models of developing ever-larger and more capable models. While they demonstrate state-of-the-art performance their growing scale...
[ -0.024216918274760246, -0.005848366767168045, -0.015402259305119514, 0.007495267316699028, -0.03338715806603432, -0.0018913618987426162, 0.02822187915444374, 0.0362505167722702, -0.002269166288897395, 0.015477118082344532, -0.006147803273051977, -0.02225186675786972, -0.04053620249032974, ...
wacv_2025_1a5de5c96f
1a5de5c96f
wacv
2,025
RAW-Diffusion: RGB-Guided Diffusion Models for High-Fidelity RAW Image Generation
Current deep learning approaches in computer vision primarily focus on RGB data sacrificing information. In contrast RAW images offer richer representation which is crucial for precise recognition particularly in challenging conditions like low-light environments. The resultant demand for comprehensive RAW image datase...
Christoph Reinders; Radu Berdan; Beril Besbinar; Junji Otsuka; Daisuke Iso
Leibniz University Hannover; Sony AI; Sony AI; Sony Group Corporation; Sony AI
Poster
main
https://github.com/SonyResearch/RAW-Diffusion
https://openaccess.thecvf.com/content/WACV2025/html/Reinders_RAW-Diffusion_RGB-Guided_Diffusion_Models_for_High-Fidelity_RAW_Image_Generation_WACV_2025_paper.html
3
RAW-Diffusion: RGB-Guided Diffusion Models for High-Fidelity RAW Image Generation Current deep learning approaches in computer vision primarily focus on RGB data sacrificing information. In contrast RAW images offer richer representation which is crucial for precise recognition particularly in challenging conditions li...
[ -0.06605646759271622, -0.03193095326423645, -0.011905526742339134, 0.00767641281709075, -0.029480351135134697, -0.0020756947342306376, 0.02503635361790657, 0.04111155867576599, 0.016066059470176697, 0.02329898811876774, -0.04619563743472099, -0.013670324347913265, -0.018370356410741806, 0....
wacv_2025_3c0748a8db
3c0748a8db
wacv
2,025
RD-DPP: Rate-Distortion Theory Meets Determinantal Point Process to Diversify Learning Data Samples
Selecting representative samples plays an indispensable role in many machine learning and computer vision applications under limited resources (e.g. limited communication bandwidth and computational power). Determinantal Point Process (DPP) is a widely used method for selecting the most diverse representative samples t...
Xiwen Chen; Huayu Li; Peijie Qiu; Wenhui Zhu; Rahul Amin; Abolfazl Razi
Clemson University; University of Arizona; Washington University in St. Louis; Arizona State University; MIT Lincoln Laboratory; Clemson University
Poster
main
https://github.com/xiwenc1/RD-DPP1
https://openaccess.thecvf.com/content/WACV2025/html/Chen_RD-DPP_Rate-Distortion_Theory_Meets_Determinantal_Point_Process_to_Diversify_Learning_WACV_2025_paper.html
5
RD-DPP: Rate-Distortion Theory Meets Determinantal Point Process to Diversify Learning Data Samples Selecting representative samples plays an indispensable role in many machine learning and computer vision applications under limited resources (e.g. limited communication bandwidth and computational power). Determinantal...
[ -0.019835803657770157, -0.016442837193608284, -0.03315567225217819, 0.03391166403889656, -0.038591619580984116, -0.005498945713043213, 0.014471856877207756, -0.010250899009406567, 0.021095791831612587, 0.040823597460985184, -0.04031960293650627, -0.022193780168890953, -0.0037619629874825478,...
wacv_2025_e50bce3538
e50bce3538
wacv
2,025
RGB-D Video Mirror Detection
Mirror detection aims to identify mirror areas in a scene with recent methods either integrating depth information (RGB-D) or making use of temporal information (video). However utilizing both data is still under-explored due to the lack of a high-quality dataset and an effective method for the RGB-D Video Mirror Detec...
Mingchen Xu; Peter Herbert; Yu-Kun Lai; Ze Ji; Jing Wu
School of Computer Science and Informatics, Cardiff University; School of Computer Science and Informatics, Cardiff University; School of Computer Science and Informatics, Cardiff University; School of Engineering, Cardiff University; School of Computer Science and Informatics, Cardiff University
Poster
main
https://github.com/UpChen/2025_DVMDNet
https://openaccess.thecvf.com/content/WACV2025/html/Xu_RGB-D_Video_Mirror_Detection_WACV_2025_paper.html
0
RGB-D Video Mirror Detection Mirror detection aims to identify mirror areas in a scene with recent methods either integrating depth information (RGB-D) or making use of temporal information (video). However utilizing both data is still under-explored due to the lack of a high-quality dataset and an effective method for...
[ -0.07490944117307663, -0.030219784006476402, -0.031388502568006516, -0.0020707694347947836, -0.0465262234210968, 0.02931077778339386, 0.010462834499776363, 0.00004304440881242044, -0.005212866235524416, 0.017874009907245636, -0.0019339547725394368, -0.032557226717472076, -0.00909932702779769...
wacv_2025_fb8d7b53ef
fb8d7b53ef
wacv
2,025
RGB2Point: 3D Point Cloud Generation from Single RGB Images
We introduce RGB2Point an unposed single-view RGB image to a 3D point cloud generation based on Transformer. RGB2Point takes an input image of an object and generates a dense 3D point cloud. Contrary to prior works based on CNN layers and diffusion-denoising approaches we use pre-trained Transformer layers that are fas...
Jae Joong Lee; Bedrich Benes
Department of Computer Science, Purdue University, West Lafayette, USA; Department of Computer Science, Purdue University, West Lafayette, USA
Poster
main
https://openaccess.thecvf.com/content/WACV2025/html/Lee_RGB2Point_3D_Point_Cloud_Generation_from_Single_RGB_Images_WACV_2025_paper.html
1
2407.14979
RGB2Point: 3D Point Cloud Generation from Single RGB Images We introduce RGB2Point an unposed single-view RGB image to a 3D point cloud generation based on Transformer. RGB2Point takes an input image of an object and generates a dense 3D point cloud. Contrary to prior works based on CNN layers and diffusion-denoising a...
[ -0.038382939994335175, -0.00841458048671484, -0.015556445345282555, -0.04534168168902397, -0.02281733974814415, 0.0035137056838721037, -0.026608021929860115, 0.019997218623757362, 0.006761880125850439, 0.004701727069914341, -0.03715600445866585, -0.006368162110447884, 0.0311312023550272, 0...
wacv_2025_455e08731e
455e08731e
wacv
2,025
ROADS: Robust Prompt-Driven Multi-Class Anomaly Detection under Domain Shift
Recent advancements in anomaly detection have shifted focus towards Multi-class Unified Anomaly Detection (MUAD) offering more scalable and practical alternatives compared to traditional one-class-one-model approaches. However existing MUAD methods often suffer from inter-class interference and are highly susceptible t...
Hossein Kashiani; Niloufar Alipour Talemi; Fatemeh Afghah
Clemson University; Clemson University; Clemson University
Poster
main
https://openaccess.thecvf.com/content/WACV2025/html/Kashiani_ROADS_Robust_Prompt-Driven_Multi-Class_Anomaly_Detection_under_Domain_Shift_WACV_2025_paper.html
2
2411.16049
ROADS: Robust Prompt-Driven Multi-Class Anomaly Detection under Domain Shift Recent advancements in anomaly detection have shifted focus towards Multi-class Unified Anomaly Detection (MUAD) offering more scalable and practical alternatives compared to traditional one-class-one-model approaches. However existing MUAD me...
[ -0.08612304925918579, -0.0228614192456007, -0.016864223405718803, -0.00291081122122705, -0.021881908178329468, 0.00035865354584529996, 0.006842716597020626, 0.025485770776867867, 0.03895866870880127, 0.027241498231887817, -0.0415460541844368, -0.01446164958178997, 0.013861006125807762, -0....
wacv_2025_d272e45fa2
d272e45fa2
wacv
2,025
ROSA: Reconstructing Object Shape and Appearance Textures by Adaptive Detail Transfer
Reconstructing an object's shape and appearance in terms of a mesh textured by a spatially-varying bidirectional reflectance distribution function (SVBRDF) from a limited set of images captured under collocated light is an ill-posed problem. Previous state-of-the-art approaches either aim to reconstruct the appearance ...
Julian Kaltheuner; Patrick Stotko; Reinhard Klein
University of Bonn; University of Bonn; University of Bonn
Poster
main
https://openaccess.thecvf.com/content/WACV2025/html/Kaltheuner_ROSA_Reconstructing_Object_Shape_and_Appearance_Textures_by_Adaptive_Detail_WACV_2025_paper.html
0
2501.18595
ROSA: Reconstructing Object Shape and Appearance Textures by Adaptive Detail Transfer Reconstructing an object's shape and appearance in terms of a mesh textured by a spatially-varying bidirectional reflectance distribution function (SVBRDF) from a limited set of images captured under collocated light is an ill-posed p...
[ -0.021381527185440063, 0.007841135375201702, -0.0076222908683121204, 0.01215282827615738, -0.02344890497624874, 0.029315786436200142, 0.02194027788937092, 0.029539287090301514, 0.03993204981088638, 0.024901656433939934, -0.034419044852256775, 0.02682003378868103, 0.034270044416189194, 0.04...
wacv_2025_0d0e613c4e
0d0e613c4e
wacv
2,025
RT-DETRv3: Real-Time End-to-End Object Detection with Hierarchical Dense Positive Supervision
RT-DETR is the first real-time end-to-end transformer-based object detector. Its efficiency comes from the framework design and the Hungarian matching. However compared to dense supervision detectors like the YOLO series the Hungarian matching provides much sparser supervision leading to insufficient model training and...
Shuo Wang; Chunlong Xia; Feng Lv; Yifeng Shi
Baidu Inc, China; Baidu Inc, China; Baidu Inc, China; Baidu Inc, China
Poster
main
https://github.com/clxia12/RT-DETRv3
https://openaccess.thecvf.com/content/WACV2025/html/Wang_RT-DETRv3_Real-Time_End-to-End_Object_Detection_with_Hierarchical_Dense_Positive_Supervision_WACV_2025_paper.html
13
RT-DETRv3: Real-Time End-to-End Object Detection with Hierarchical Dense Positive Supervision RT-DETR is the first real-time end-to-end transformer-based object detector. Its efficiency comes from the framework design and the Hungarian matching. However compared to dense supervision detectors like the YOLO series the H...
[ -0.02104336954653263, -0.04126589745283127, -0.00032472176826559007, 0.0009805781301110983, -0.016780594363808632, 0.006063027773052454, 0.008138448931276798, -0.004216136410832405, 0.022498495876789093, 0.015987737104296684, -0.04656405374407768, 0.008492902852594852, 0.0004063394444528967,...
wacv_2025_e13a118580
e13a118580
wacv
2,025
Radiance Field-Based Pose Estimation via Decoupled Optimization Under Challenging Initial Conditions
Estimating six-degree-of-freedom poses is essential but difficult especially under challenging initial conditions as existing methods relying on photometric loss often fail catastrophically. To address this we propose a novel radiance field-based pose estimation framework that combines Monte Carlo initialization and de...
Si-Yu Lu; Yung-Yao Chen; Yi-Tong Wu; Hsin-Chun Lin; Sin-Ye Jhong; Wen-Huang Cheng
National Taiwan University; National Taiwan University of Science and Technology; National Taiwan University of Science and Technology; National Taiwan University of Science and Technology; National Taiwan University of Science and Technology; National Taiwan University
Poster
main
https://openaccess.thecvf.com/content/WACV2025/html/Lu_Radiance_Field-Based_Pose_Estimation_via_Decoupled_Optimization_Under_Challenging_Initial_WACV_2025_paper.html
0
Radiance Field-Based Pose Estimation via Decoupled Optimization Under Challenging Initial Conditions Estimating six-degree-of-freedom poses is essential but difficult especially under challenging initial conditions as existing methods relying on photometric loss often fail catastrophically. To address this we propose a...
[ -0.017000624909996986, 0.0067900256253778934, -0.049970127642154694, 0.0043802871368825436, -0.012074255384504795, 0.02215937152504921, 0.029483862221240997, 0.06755634397268295, 0.049858588725328445, 0.025170965120196342, -0.04662391170859337, -0.028796030208468437, 0.0006105678621679544, ...
wacv_2025_147053ed58
147053ed58
wacv
2,025
RapidNet: Multi-Level Dilated Convolution Based Mobile Backbone
Vision transformers (ViTs) have dominated computer vision in recent years. However ViTs are computationally expensive and not well suited for mobile devices; this led to the prevalence of convolutional neural network (CNN) and ViT-based hybrid models for mobile vision applications. Recently Vision GNN (ViG) and CNN hyb...
Mustafa Munir; Md Mostafijur Rahman; Radu Marculescu
The University of Texas at Austin; The University of Texas at Austin; The University of Texas at Austin
Poster
main
https://github.com/mmunir127/RapidNet-Official
https://openaccess.thecvf.com/content/WACV2025/html/Munir_RapidNet_Multi-Level_Dilated_Convolution_Based_Mobile_Backbone_WACV_2025_paper.html
0
2412.10995
RapidNet: Multi-Level Dilated Convolution Based Mobile Backbone Vision transformers (ViTs) have dominated computer vision in recent years. However ViTs are computationally expensive and not well suited for mobile devices; this led to the prevalence of convolutional neural network (CNN) and ViT-based hybrid models for m...
[ -0.03837442025542259, -0.014609930105507374, -0.03375977277755737, 0.00451189000159502, 0.0019873802084475756, 0.026398761197924614, -0.01125637162476778, 0.028323085978627205, 0.023652393370866776, 0.02703397534787655, -0.018411874771118164, -0.0035917633213102818, -0.023316102102398872, ...
wacv_2025_e983ae53b9
e983ae53b9
wacv
2,025
RayGauss: Volumetric Gaussian-Based Ray Casting for Photorealistic Novel View Synthesis
Differentiable volumetric rendering-based methods made significant progress in novel view synthesis. On one hand innovative methods have replaced the Neural Radiance Fields (NeRF) network with locally parameterized structures enabling high-quality renderings in a reasonable time. On the other hand approaches have used ...
Hugo Blanc; Jean-Emmanuel Deschaud; Alexis Paljic
Centre for Robotics, Mines Paris - PSL, PSL University, 75006 Paris, France; Centre for Robotics, Mines Paris - PSL, PSL University, 75006 Paris, France; Centre for Robotics, Mines Paris - PSL, PSL University, 75006 Paris, France
Poster
main
https://openaccess.thecvf.com/content/WACV2025/html/Blanc_RayGauss_Volumetric_Gaussian-Based_Ray_Casting_for_Photorealistic_Novel_View_Synthesis_WACV_2025_paper.html
4
2408.03356
RayGauss: Volumetric Gaussian-Based Ray Casting for Photorealistic Novel View Synthesis Differentiable volumetric rendering-based methods made significant progress in novel view synthesis. On one hand innovative methods have replaced the Neural Radiance Fields (NeRF) network with locally parameterized structures enabli...
[ 0.007569345645606518, 0.006826342549175024, 0.003698762971907854, 0.010680671781301498, -0.019429534673690796, -0.03356517106294632, 0.000613558164332062, 0.003410849254578352, 0.021287042647600174, 0.05999751389026642, -0.029144302010536194, -0.016568971797823906, -0.003529265522956848, 0...
wacv_2025_c1df12f77b
c1df12f77b
wacv
2,025
Re-Evaluating Group Robustness via Adaptive Class-Specific Scaling
Group distributionally robust optimization which aims to improve robust accuracies--worst-group and unbiased accuracies--is a prominent algorithm used to mitigate spurious correlations and address dataset bias. Although existing approaches have reported improvements in robust accuracies these gains often come at the co...
Seonguk Seo; Bohyung Han
ECE; ECE + IPAI
Poster
main
https://openaccess.thecvf.com/content/WACV2025/html/Seo_Re-Evaluating_Group_Robustness_via_Adaptive_Class-Specific_Scaling_WACV_2025_paper.html
0
2412.15311
Re-Evaluating Group Robustness via Adaptive Class-Specific Scaling Group distributionally robust optimization which aims to improve robust accuracies--worst-group and unbiased accuracies--is a prominent algorithm used to mitigate spurious correlations and address dataset bias. Although existing approaches have reported...
[ -0.035777006298303604, -0.016863591969013214, -0.0011921714758500457, 0.007345200050622225, -0.006173194386065006, -0.01647450588643551, 0.03492291271686554, 0.010799531824886799, 0.014652440324425697, 0.03644130378961563, -0.05166313797235489, 0.0024958495050668716, -0.008669043891131878, ...
wacv_2025_2032f78b5e
2032f78b5e
wacv
2,025
Re-Identifying People in Video via Learned Temporal Attention and Multi-Modal Foundation Models
Biometric recognition from security camera video is a challenging problem when the individuals change clothes or when they are partly occluded. Others have recently demonstrated that CLIP's visual encoder performs well in this domain but existing methods fail to make use of the model's text encoder or temporal informat...
Cole Hill; Florence Yellin; Krishna Regmi; Dawei Du; Scott McCloskey
Kitware Inc., USA+University of South Florida, USA; Kitware Inc., USA; Kitware Inc., USA; Kitware Inc., USA; Kitware Inc., USA
Poster
main
https://openaccess.thecvf.com/content/WACV2025/html/Hill_Re-Identifying_People_in_Video_via_Learned_Temporal_Attention_and_Multi-Modal_WACV_2025_paper.html
0
Re-Identifying People in Video via Learned Temporal Attention and Multi-Modal Foundation Models Biometric recognition from security camera video is a challenging problem when the individuals change clothes or when they are partly occluded. Others have recently demonstrated that CLIP's visual encoder performs well in th...
[ -0.05353478714823723, -0.03224000334739685, -0.026525406166911125, 0.026934921741485596, -0.03235168755054474, 0.004041634500026703, 0.009716675616800785, -0.001762543455697596, 0.009204781614243984, 0.022616399452090263, -0.03341270610690117, -0.05774162337183952, -0.004239411558955908, 0...
wacv_2025_594ff86fde
594ff86fde
wacv
2,025
ReBotNet: Fast Real-Time Video Enhancement
Most video restoration networks are slow have high computational load and can't be used for real-time video enhancement. In this work we design an efficient and fast framework to perform real-time video enhancement for practical use-cases like live video calls and video streams. Our proposed method called Recurrent Bot...
Jeya Maria Jose Valanarasu; Rahul Garg; Andeep Toor; Xin Tong; Weijuan Xi; Andreas Lugmayr; Vishal M. Patel; Anne Menini
Google; Google; Google; Google; Google; ETH Zurich; Johns Hopkins University; Google
Poster
main
https://github.com/jeya-maria-jose/rebot-net
https://openaccess.thecvf.com/content/WACV2025/html/Valanarasu_ReBotNet_Fast_Real-Time_Video_Enhancement_WACV_2025_paper.html
1
2303.13504
ReBotNet: Fast Real-Time Video Enhancement Most video restoration networks are slow have high computational load and can't be used for real-time video enhancement. In this work we design an efficient and fast framework to perform real-time video enhancement for practical use-cases like live video calls and video stream...
[ -0.004245179705321789, 0.006120000965893269, -0.007280061021447182, 0.0010550151346251369, -0.021849317476153374, 0.012952479533851147, 0.03935069218277931, 0.014505315572023392, 0.003975716885179281, 0.03354125842452049, -0.020625317469239235, -0.015318271704018116, -0.027932781726121902, ...
wacv_2025_58f5a08c9c
58f5a08c9c
wacv
2,025
ReC-TTT: Contrastive Feature Reconstruction for Test-Time Training
The remarkable progress in deep learning (DL) showcases outstanding results in various computer vision tasks. However adaptation to real-time variations in data distributions remains an important challenge. Test-Time Training (TTT) was proposed as an effective solution to this issue which increases the generalization a...
Marco Colussi; Sergio Mascetti; Jose Dolz; Christian Desrosiers
Universit `a degli studi di Milano; Universit `a degli studi di Milano; ´ETS Montr ´eal; ´ETS Montr ´eal
Poster
main
https://github.com/warpcut/ReC-TTT
https://openaccess.thecvf.com/content/WACV2025/html/Colussi_ReC-TTT_Contrastive_Feature_Reconstruction_for_Test-Time_Training_WACV_2025_paper.html
0
ReC-TTT: Contrastive Feature Reconstruction for Test-Time Training The remarkable progress in deep learning (DL) showcases outstanding results in various computer vision tasks. However adaptation to real-time variations in data distributions remains an important challenge. Test-Time Training (TTT) was proposed as an ef...
[ -0.028928283601999283, -0.026192067191004753, -0.030134379863739014, 0.033464640378952026, 0.015220201574265957, 0.014707161113619804, 0.018451455980539322, 0.012294970452785492, 0.0027789694722741842, 0.04741574451327324, -0.0214396920055151, -0.05864863097667694, -0.022969814017415047, 0...
wacv_2025_cdc6889db3
cdc6889db3
wacv
2,025
ReEdit: Multimodal Exemplar-Based Image Editing
Modern Text-to-Image (T2I) Diffusion models have revolutionized image editing by enabling the generation of high-quality photorealistic images. While the de-facto method for performing edits with T2I models is through text instructions this approach is non-trivial due to the complex many-to-many mapping between natural...
Ashutosh Srivastava; Tarun Ram Menta; Abhinav Java; Avadhoot Gorakh Jadhav; Silky Singh; Surgan Jandial; Balaji Krishnamurthy
Indian Institute of Technology, Roorkee; Adobe MDSR; Microsoft Research; Indian Institute of Technology, Bombay; Stanford University; Carnegie Mellon University; Adobe MDSR
Poster
main
https://reedit-diffusion.github.io/
https://openaccess.thecvf.com/content/WACV2025/html/Srivastava_ReEdit_Multimodal_Exemplar-Based_Image_Editing_WACV_2025_paper.html
0
2411.03982
ReEdit: Multimodal Exemplar-Based Image Editing Modern Text-to-Image (T2I) Diffusion models have revolutionized image editing by enabling the generation of high-quality photorealistic images. While the de-facto method for performing edits with T2I models is through text instructions this approach is non-trivial due to ...
[ -0.04068584740161896, 0.013687408529222012, -0.00782596692442894, 0.01463295053690672, -0.01078652311116457, -0.007936126552522182, 0.0057007921859622, 0.006508633494377136, 0.03811544179916382, -0.006660103797912598, -0.04777282103896141, -0.018268238753080368, -0.029063941910862923, 0.03...
wacv_2025_d32b6678b4
d32b6678b4
wacv
2,025
ReFu: Recursive Fusion for Exemplar-Free 3D Class-Incremental Learning
We introduce a novel Recursive Fusion model dubbed ReFu designed to integrate point clouds and meshes for exemplar-free 3D Class-Incremental Learning where the model learns new 3D classes while retaining knowledge of previously learned ones. Unlike existing methods that either rely on storing historical data to mitigat...
Yi Yang; Lei Zhong; Huiping Zhuang
The University of Edinburgh; The University of Edinburgh; South China University of Technology
Poster
main
https://openaccess.thecvf.com/content/WACV2025/html/Yang_ReFu_Recursive_Fusion_for_Exemplar-Free_3D_Class-Incremental_Learning_WACV_2025_paper.html
0
2409.12326
ReFu: Recursive Fusion for Exemplar-Free 3D Class-Incremental Learning We introduce a novel Recursive Fusion model dubbed ReFu designed to integrate point clouds and meshes for exemplar-free 3D Class-Incremental Learning where the model learns new 3D classes while retaining knowledge of previously learned ones. Unlike ...
[ 0.009785287082195282, -0.017944209277629852, -0.014790872111916542, -0.03285254165530205, -0.009523261338472366, -0.002785147400572896, 0.042357731610536575, 0.03830989450216293, 0.012513963505625725, 0.000039282607758650556, -0.029961228370666504, -0.026979561895132065, -0.00243954523466527...
wacv_2025_06ecd55a0e
06ecd55a0e
wacv
2,025
ReMP: Reusable Motion Prior for Multi-Domain 3D Human Pose Estimation and Motion Inbetweening
We present Reusable Motion prior (ReMP) an effective motion prior that can accurately track the temporal evolution of motion in various downstream tasks. Inspired by the success of foundation models we argue that a robust spatio-temporal motion prior can encapsulate underlying 3D dynamics applicable to various sensor m...
Hojun Jang; Young Min Kim
Dept. of Electrical and Computer Engineering, Seoul National University + Interdisciplinary Program in Artificial Intelligence and INMC, Seoul National University; Dept. of Electrical and Computer Engineering, Seoul National University + Interdisciplinary Program in Artificial Intelligence and INMC, Seoul National Univ...
Poster
main
https://openaccess.thecvf.com/content/WACV2025/html/Jang_ReMP_Reusable_Motion_Prior_for_Multi-Domain_3D_Human_Pose_Estimation_WACV_2025_paper.html
0
2411.09435
ReMP: Reusable Motion Prior for Multi-Domain 3D Human Pose Estimation and Motion Inbetweening We present Reusable Motion prior (ReMP) an effective motion prior that can accurately track the temporal evolution of motion in various downstream tasks. Inspired by the success of foundation models we argue that a robust spat...
[ -0.06298260390758514, 0.002029244787991047, -0.04717245697975159, 0.0017309619579464197, -0.03475283458828926, 0.004196690395474434, -0.02001143991947174, -0.006366439163684845, -0.016860468313097954, -0.005629369523376226, -0.055059101432561874, -0.0011827663984149694, -0.007665524259209633...
wacv_2025_cb7eabf460
cb7eabf460
wacv
2,025
ReMix: Training Generalized Person Re-Identification on a Mixture of Data
Modern person re-identification (Re-ID) methods have a weak generalization ability and experience a major accuracy drop when capturing environments change. This is because existing multi-camera Re-ID datasets are limited in size and diversity since such data is difficult to obtain. At the same time enormous volumes of ...
Timur Mamedov; Anton Konushin; Vadim Konushin
Tevian, Moscow, Russia+Lomonosov Moscow State University; AIRI, Moscow, Russia+Lomonosov Moscow State University; Tevian, Moscow, Russia
Poster
main
https://openaccess.thecvf.com/content/WACV2025/html/Mamedov_ReMix_Training_Generalized_Person_Re-Identification_on_a_Mixture_of_Data_WACV_2025_paper.html
3
2410.21938
ReMix: Training Generalized Person Re-Identification on a Mixture of Data Modern person re-identification (Re-ID) methods have a weak generalization ability and experience a major accuracy drop when capturing environments change. This is because existing multi-camera Re-ID datasets are limited in size and diversity sin...
[ -0.05387163162231445, -0.001889747567474842, -0.042219724506139755, -0.015293127857148647, -0.006369831971824169, 0.01230640523135662, 0.006337568163871765, 0.04174037650227547, -0.014657066203653812, 0.018215322867035866, -0.04977871850132942, -0.05719021335244179, -0.03314893692731857, 0...
wacv_2025_af25ee3841
af25ee3841
wacv
2,025
Realistic and Efficient Face Swapping: A Unified Approach with Diffusion Models
Despite promising progress in face swapping task realistic swapped images remain elusive often marred by artifacts particularly in scenarios involving high pose variation color differences and occlusion. To address these issues we propose a novel approach that better harnesses diffusion models for face-swapping by maki...
Sanoojan Baliah; Qinliang Lin; Shengcai Liao; Xiaodan Liang; Muhammad Haris Khan
MBZUAI, UAE; Shenzhen University, China; United Arab Emirates University, UAE; MBZUAI, UAE; MBZUAI, UAE
Poster
main
https://openaccess.thecvf.com/content/WACV2025/html/Baliah_Realistic_and_Efficient_Face_Swapping_A_Unified_Approach_with_Diffusion_WACV_2025_paper.html
5
2409.07269
Realistic and Efficient Face Swapping: A Unified Approach with Diffusion Models Despite promising progress in face swapping task realistic swapped images remain elusive often marred by artifacts particularly in scenarios involving high pose variation color differences and occlusion. To address these issues we propose a...
[ -0.05420362576842308, -0.018835581839084625, -0.04334861785173416, 0.007172696758061647, -0.024655865505337715, -0.018174996599555016, -0.009703448973596096, 0.032582879066467285, 0.0181660708039999, 0.021388651803135872, -0.03540375083684921, -0.014693539589643478, 0.011961933225393295, 0...
wacv_2025_9720c5b894
9720c5b894
wacv
2,025
Recognizing Unseen States of Unknown Objects by Leveraging Knowledge Graphs
We investigate the problem of Object State Classification (OSC) in the context of zero-shot learning. Specifically we propose the first method for Zero-shot Object-agnostic State Classification (OaSC) that given an image infers the state of a single object without relying on the knowledge or the estimation of the objec...
Filippos Gouidis; Konstantinos Papoutsakis; Theodore Patkos; Antonis Argyros; Dimitris Plexousakis
Foundation for Research and Technology-Hellas, Greece + University of Crete, Greece; Hellenic Mediterranean University, Greece; Foundation for Research and Technology-Hellas, Greece; Foundation for Research and Technology-Hellas, Greece + University of Crete, Greece; Foundation for Research and Technology-Hellas, Greec...
Poster
main
https://github.com/philipposg/OaSC.git
https://openaccess.thecvf.com/content/WACV2025/html/Gouidis_Recognizing_Unseen_States_of_Unknown_Objects_by_Leveraging_Knowledge_Graphs_WACV_2025_paper.html
0
Recognizing Unseen States of Unknown Objects by Leveraging Knowledge Graphs We investigate the problem of Object State Classification (OSC) in the context of zero-shot learning. Specifically we propose the first method for Zero-shot Object-agnostic State Classification (OaSC) that given an image infers the state of a s...
[ -0.07519716024398804, -0.001466084853745997, -0.029117558151483536, 0.005289040040224791, 0.01203488651663065, -0.04123595356941223, 0.04431658983230591, 0.010188361629843712, 0.010707986541092396, 0.029896995052695274, -0.03307041898369789, -0.03310753405094147, 0.01300918310880661, 0.007...
wacv_2025_10106a0e44
10106a0e44
wacv
2,025
Recoverable Anonymization for Pose Estimation: A Privacy-Enhancing Approach
Human pose estimation (HPE) is crucial for various applications. However deploying HPE algorithms in surveillance contexts raises significant privacy concerns due to the potential leakage of sensitive personal information (SPI) such as facial features and ethnicity. Existing privacy-enhancing methods often compromise e...
Wenjun Huang; Yang Ni; Arghavan Rezvani Dehaghani; SungHeon Evan Jeong; Hanning Chen; Yezi Liu; Fei Wen; Mohsen Imani
University of California, Irvine, USA; University of California, Irvine, USA; University of California, Irvine, USA; University of California, Irvine, USA; University of California, Irvine, USA; University of California, Irvine, USA; Texas A&M University, USA; University of California, Irvine, USA
Poster
main
https://openaccess.thecvf.com/content/WACV2025/html/Huang_Recoverable_Anonymization_for_Pose_Estimation_A_Privacy-Enhancing_Approach_WACV_2025_paper.html
3
Recoverable Anonymization for Pose Estimation: A Privacy-Enhancing Approach Human pose estimation (HPE) is crucial for various applications. However deploying HPE algorithms in surveillance contexts raises significant privacy concerns due to the potential leakage of sensitive personal information (SPI) such as facial f...
[ -0.04668357968330383, 0.0021558604203164577, -0.026059022173285484, 0.027644075453281403, -0.032455842941999435, -0.020945340394973755, 0.04872150346636772, 0.03394654765725136, 0.03607882186770439, 0.025341974571347237, -0.05249543860554695, 0.01639774814248085, -0.008453615009784698, -0....
wacv_2025_35fb5d7181
35fb5d7181
wacv
2,025
Recurrence-Based Vanishing Point Detection
Classical approaches to Vanishing Point Detection (VPD) rely solely on the presence of explicit straight lines in images. Recent supervised deep learning methods rely on learned filters from labeled datasets. We propose an alternative unsupervised approach: Recurrence-based Vanishing Point Detection (R-VPD) that uses i...
Skanda Bharadwaj; Robert T. Collins; Yanxi Liu
School of EECS, The Pennsylvania State University, University Park, PA; School of EECS, The Pennsylvania State University, University Park, PA; School of EECS, The Pennsylvania State University, University Park, PA
Poster
main
https://openaccess.thecvf.com/content/WACV2025/html/Bharadwaj_Recurrence-Based_Vanishing_Point_Detection_WACV_2025_paper.html
0
2412.20666
Recurrence-Based Vanishing Point Detection Classical approaches to Vanishing Point Detection (VPD) rely solely on the presence of explicit straight lines in images. Recent supervised deep learning methods rely on learned filters from labeled datasets. We propose an alternative unsupervised approach: Recurrence-based Va...
[ -0.06291091442108154, -0.04379203915596008, -0.03198443725705147, 0.021744882687926292, -0.005743218585848808, 0.011127485893666744, 0.07776015251874924, 0.043451979756355286, 0.024540923535823822, 0.032513417303562164, -0.04583239555358887, -0.0552784763276577, 0.013328422792255878, 0.046...
wacv_2025_d2466691c2
d2466691c2
wacv
2,025
Reducing the Content Bias for AI-Generated Image Detection
Identifying AI-generated content is critical for the safe and ethical use of generative AI. Recent research has focused on developing detectors that generalize to unknown generators with popular methods relying either on high-level features or low-level fingerprints. However these methods have clear limitations: biased...
Seoyeon Gye; Junwon Ko; Hyounguk Shon; Minchan Kwon; Junmo Kim
School of Electrical Engineering, KAIST, South Korea; School of Electrical Engineering, KAIST, South Korea; School of Electrical Engineering, KAIST, South Korea; School of Electrical Engineering, KAIST, South Korea; School of Electrical Engineering, KAIST, South Korea
Poster
main
https://openaccess.thecvf.com/content/WACV2025/html/Gye_Reducing_the_Content_Bias_for_AI-Generated_Image_Detection_WACV_2025_paper.html
0
Reducing the Content Bias for AI-Generated Image Detection Identifying AI-generated content is critical for the safe and ethical use of generative AI. Recent research has focused on developing detectors that generalize to unknown generators with popular methods relying either on high-level features or low-level fingerp...
[ -0.0664151981472969, -0.052968718111515045, -0.0326504223048687, 0.0019129656720906496, 0.00038682998274452984, -0.03818502277135849, -0.015052996575832367, 0.03881648927927017, 0.0029205228202044964, 0.040748026221990585, -0.04085946083068848, 0.004995997529476881, -0.045948319137096405, ...
wacv_2025_857efb304c
857efb304c
wacv
2,025
RefVSR++: Exploiting Reference Inputs for Reference-Based Video Super-Resolution
Smartphones with multi-camera systems featuring cameras with varying field-of-views (FoVs) are increasingly common. This variation in FoVs results in content differences across videos paving the way for an innovative approach to video super-resolution (VSR). This method enhances the VSR performance of lower resolution ...
Han Zou; Masanori Suganuma; Takayuki Okatani
Graduate School of Information Sciences, Tohoku University+RIKEN Center for AIP; Graduate School of Information Sciences, Tohoku University+RIKEN Center for AIP; Graduate School of Information Sciences, Tohoku University+RIKEN Center for AIP
Poster
main
https://openaccess.thecvf.com/content/WACV2025/html/Zou_RefVSR_Exploiting_Reference_Inputs_for_Reference-Based_Video_Super-Resolution_WACV_2025_paper.html
0
RefVSR++: Exploiting Reference Inputs for Reference-Based Video Super-Resolution Smartphones with multi-camera systems featuring cameras with varying field-of-views (FoVs) are increasingly common. This variation in FoVs results in content differences across videos paving the way for an innovative approach to video supe...
[ -0.06873297691345215, -0.01534892525523901, -0.05046172812581062, 0.05513744056224823, -0.02404395304620266, 0.06053249537944794, 0.019727909937500954, 0.031363241374492645, -0.009441342204809189, 0.0077463965862989426, -0.033719081431627274, 0.015465818345546722, -0.02285704016685486, 0.0...
wacv_2025_7061d1afa5
7061d1afa5
wacv
2,025
Refining Text-to-Image Generation: Towards Accurate Training-Free Glyph-Enhanced Image Generation
Over the past few years Text-to-Image (T2I) generation approaches based on diffusion models have gained significant attention. However vanilla diffusion models often suffer from spelling inaccuracies in the text displayed within the generated images. The capability to generate visual text is crucial offering both acade...
Sanyam Lakhanpal; Shivang Chopra; Vinija Jain; Aman Chadha; Man Luo
Arizona State University; Georgia Institute of Technology; Meta AI; Amazon GenAI; Intel Lab
Poster
main
https://github.com/SanyamLakhanpal/SAOcr/tree/main
https://openaccess.thecvf.com/content/WACV2025/html/Lakhanpal_Refining_Text-to-Image_Generation_Towards_Accurate_Training-Free_Glyph-Enhanced_Image_Generation_WACV_2025_paper.html
9
2403.16422
Refining Text-to-Image Generation: Towards Accurate Training-Free Glyph-Enhanced Image Generation Over the past few years Text-to-Image (T2I) generation approaches based on diffusion models have gained significant attention. However vanilla diffusion models often suffer from spelling inaccuracies in the text displayed ...
[ -0.09155071526765823, -0.0035478624049574137, 0.0182085819542408, -0.011987920850515366, -0.010319405235350132, -0.03132093697786331, -0.023576848208904266, 0.04954765364527702, -0.006692197639495134, 0.00634761294350028, -0.035056959837675095, -0.019750144332647324, -0.03217333182692528, ...
wacv_2025_9ef5bc9f8d
9ef5bc9f8d
wacv
2,025
Reflective Teacher: Semi-Supervised Multimodal 3D Object Detection in Bird's-Eye-View via Uncertainty Measure
Applying pseudo labeling techniques has been found to be advantageous in semi-supervised 3D object detection (SSOD) in Bird's-Eye-View (BEV) for autonomous driving particularly where labeled data is limited. In the literature Exponential Moving Average (EMA) has been used for adjustments of the weights of teacher netwo...
Saheli Hazra; Sudip Das; Rohit Choudhary; Arindam Das; Ganesh Sistu; Ciarán Eising; Ujjwal Bhattacharya
;;;;;;
Poster
main
https://openaccess.thecvf.com/content/WACV2025/html/Hazra_Reflective_Teacher_Semi-Supervised_Multimodal_3D_Object_Detection_in_Birds-Eye-View_via_WACV_2025_paper.html
0
Reflective Teacher: Semi-Supervised Multimodal 3D Object Detection in Bird's-Eye-View via Uncertainty Measure Applying pseudo labeling techniques has been found to be advantageous in semi-supervised 3D object detection (SSOD) in Bird's-Eye-View (BEV) for autonomous driving particularly where labeled data is limited. In...
[ -0.0944538339972496, -0.033762939274311066, -0.00664304057136178, 0.004915568977594376, 0.013641874305903912, -0.015551924705505371, 0.020205330103635788, 0.047077108174562454, 0.009924768470227718, 0.02473701536655426, -0.024418674409389496, 0.012134433723986149, -0.021347614005208015, 0....
wacv_2025_26ba90ced2
26ba90ced2
wacv
2,025
Reframing Image Difference Captioning with BLIP2IDC and Synthetic Augmentation
The rise of the generative models quality during the past years enabled the generation of edited variations of images at an important scale. To counter the harmful effects of such technology the Image Difference Captioning (IDC) task aims to describe the differences between two images. While this task is successfully h...
Gautier Evennou; Antoine Chaffin; Vivien Chappelier; Ewa Kijak
IMATAG, France+IRISA, CNRS, France; LightOn, France+IMATAG, France+IRISA, CNRS, France; IMATAG, France; IRISA, CNRS, France
Poster
main
https://github.com/gautierevn/BLIP2IDC
https://openaccess.thecvf.com/content/WACV2025/html/Evennou_Reframing_Image_Difference_Captioning_with_BLIP2IDC_and_Synthetic_Augmentation_WACV_2025_paper.html
0
2412.15939
Reframing Image Difference Captioning with BLIP2IDC and Synthetic Augmentation The rise of the generative models quality during the past years enabled the generation of edited variations of images at an important scale. To counter the harmful effects of such technology the Image Difference Captioning (IDC) task aims to...
[ -0.06638073921203613, -0.0016063584480434656, -0.004696575924754143, -0.013571995310485363, -0.03642619773745537, -0.04526463896036148, 0.017112918198108673, 0.048444997519254684, -0.011852382682263851, 0.009180511347949505, -0.05188422277569771, -0.03563110902905464, -0.04959140345454216, ...
wacv_2025_747c5a7a70
747c5a7a70
wacv
2,025
ReinDiffuse: Crafting Physically Plausible Motions with Reinforced Diffusion Model
Generating human motion from textual descriptions is a challenging task. Existing methods either struggle with physical credibility or are limited by the complexities of physics simulations. In this paper we present ReinDiffuse that combines reinforcement learning with motion diffusion model to generate physically cred...
Gaoge Han; Mingjiang Liang; Jinglei Tang; Yongkang Cheng; Wei Liu; Shaoli Huang
Northwest A&F University; University of Technology Sydney; Northwest A&F University + Tencent AI Lab; Tencent AI Lab; University of Technology Sydney; Tencent AI Lab
Poster
main
https://openaccess.thecvf.com/content/WACV2025/html/Han_ReinDiffuse_Crafting_Physically_Plausible_Motions_with_Reinforced_Diffusion_Model_WACV_2025_paper.html
7
2410.07296
ReinDiffuse: Crafting Physically Plausible Motions with Reinforced Diffusion Model Generating human motion from textual descriptions is a challenging task. Existing methods either struggle with physical credibility or are limited by the complexities of physics simulations. In this paper we present ReinDiffuse that comb...
[ -0.08371564745903015, 0.0006778219249099493, -0.008430990390479565, -0.0027553923428058624, -0.026295775547623634, 0.008310282602906227, -0.0030618051532655954, 0.018681885674595833, 0.037419483065605164, 0.008152433671057224, -0.03988935425877571, -0.02183886431157589, -0.014958508312702179...
wacv_2025_c350d1e21a
c350d1e21a
wacv
2,025
Relational Self-Supervised Distillation with Compact Descriptors for Image Copy Detection
Image copy detection is the task of detecting edited copies of any image within a reference database. While previous approaches have shown remarkable progress the large size of their networks and descriptors remains a dis-advantage complicating their practical application. In this paper we propose a novel method that a...
Juntae Kim; Sungwon Woo; Jongho Nang
Sogang University; Sogang University; Sogang University + mAy-I
Poster
main
https://github.com/juntae9926/RDCD
https://openaccess.thecvf.com/content/WACV2025/html/Kim_Relational_Self-Supervised_Distillation_with_Compact_Descriptors_for_Image_Copy_Detection_WACV_2025_paper.html
1
2405.17928
Relational Self-Supervised Distillation with Compact Descriptors for Image Copy Detection Image copy detection is the task of detecting edited copies of any image within a reference database. While previous approaches have shown remarkable progress the large size of their networks and descriptors remains a dis-advantag...
[ -0.07948857545852661, 0.0036347536370158195, -0.0009912964887917042, 0.019835172221064568, -0.005106678698211908, 0.0009762768168002367, -0.0035746749490499496, 0.03161057084798813, 0.03684202954173088, 0.020648542791604996, -0.020075485110282898, -0.012348457239568233, -0.02364322915673256,...
wacv_2025_57ce1bad36
57ce1bad36
wacv
2,025
Relaxing Binary Constraints in Contrastive Vision-Language Medical Representation Learning
By aligning paired image and caption embeddings as input contrastive vision-language representation learning has witnessed significant advances as illustrated by CLIP allowing visual encoders to learn from textual supervision and vice versa. Benefiting from millions of image-caption pairs collected from the Internet CL...
Xiaoyang Wei; Camille Kurtz; Florence Cloppet
Universit ´e Paris Cit ´e, LIPADE, F-75006, Paris, France; Universit ´e Paris Cit ´e, LIPADE, F-75006, Paris, France; Universit ´e Paris Cit ´e, LIPADE, F-75006, Paris, France
Poster
main
https://openaccess.thecvf.com/content/WACV2025/html/Wei_Relaxing_Binary_Constraints_in_Contrastive_Vision-Language_Medical_Representation_Learning_WACV_2025_paper.html
0
Relaxing Binary Constraints in Contrastive Vision-Language Medical Representation Learning By aligning paired image and caption embeddings as input contrastive vision-language representation learning has witnessed significant advances as illustrated by CLIP allowing visual encoders to learn from textual supervision and...
[ -0.08156360685825348, -0.04244712367653847, -0.01915116421878338, 0.006254196632653475, 0.01051001064479351, -0.008058291859924793, 0.030105261132121086, 0.02044641226530075, -0.015496714971959591, 0.02040940523147583, -0.04259515181183815, -0.04773913323879242, -0.017328565940260887, 0.01...
wacv_2025_4550a27d42
4550a27d42
wacv
2,025
Remote Blood Pressure Estimation from Facial Videos using Transfer Learning: Leveraging PPG to rPPG Conversion
Blood pressure (BP) monitoring is crucial for health assessment but existing contact-based methods face cost and comfort barriers. Remote photoplethysmography (rPPG) offers a promising contactless solution yet research is hampered by limited rPPG datasets with corresponding BP labels. This paper presents a transfer lea...
Chun-Hong Cheng; Jing Wei Chin; Kwan Long Wong; Tsz Tai Chan; Hau Ching Lo; Kwan Lok Pang; Richard So; Bryan Yan
PanopticAI; PanopticAI + Hong Kong University of Science and Technology; PanopticAI + Hong Kong University of Science and Technology; PanopticAI + Hong Kong University of Science and Technology; Hong Kong University of Science and Technology; Hong Kong University of Science and Technology; Hong Kong University of Scien...
Poster
main
https://openaccess.thecvf.com/content/WACV2025/html/Cheng_Remote_Blood_Pressure_Estimation_from_Facial_Videos_using_Transfer_Learning_WACV_2025_paper.html
0
Remote Blood Pressure Estimation from Facial Videos using Transfer Learning: Leveraging PPG to rPPG Conversion Blood pressure (BP) monitoring is crucial for health assessment but existing contact-based methods face cost and comfort barriers. Remote photoplethysmography (rPPG) offers a promising contactless solution yet...
[ -0.008916985243558884, -0.022581009194254875, -0.04464077576994896, 0.004975082352757454, -0.011243964545428753, -0.02816575951874256, 0.03466268628835678, -0.012537765316665173, 0.038683708757162094, -0.006832011975347996, -0.028147144243121147, 0.02675095573067665, 0.009321879595518112, ...
wacv_2025_bfbe260ede
bfbe260ede
wacv
2,025
Removing Geometric Bias in One-Class Anomaly Detection with Adaptive Feature Perturbation
One-class anomaly detection aims to detect objects that do not belong to a predefined normal class. In practice training data lack those anomalous samples; hence state-of-the-art methods are trained to discriminate between normal and synthetically-generated pseudo-anomalous data. Most methods use data augmentation tech...
Romain Hermary; Vincent Gaudilliere; Abd El Rahman Shabayek; Djamila Aouada
University of Luxembourg, Esch-sur-Alzette, Luxembourg; University of Luxembourg, Esch-sur-Alzette, Luxembourg; University of Luxembourg, Esch-sur-Alzette, Luxembourg; University of Luxembourg, Esch-sur-Alzette, Luxembourg
Poster
main
https://openaccess.thecvf.com/content/WACV2025/html/Hermary_Removing_Geometric_Bias_in_One-Class_Anomaly_Detection_with_Adaptive_Feature_WACV_2025_paper.html
0
Removing Geometric Bias in One-Class Anomaly Detection with Adaptive Feature Perturbation One-class anomaly detection aims to detect objects that do not belong to a predefined normal class. In practice training data lack those anomalous samples; hence state-of-the-art methods are trained to discriminate between normal ...
[ -0.05497585982084274, -0.03033086098730564, -0.03390805795788765, 0.012426051311194897, 0.009818462654948235, -0.045863423496484756, 0.00627421448007226, 0.046089351177215576, 0.016389207914471626, 0.02285640314221382, -0.06216790899634361, -0.02046532928943634, -0.02163262479007244, 0.003...
wacv_2025_12ec89edb0
12ec89edb0
wacv
2,025
RendBEV: Semantic Novel View Synthesis for Self-Supervised Bird's Eye View Segmentation
Bird's Eye View (BEV) semantic maps have recently garnered a lot of attention as a useful representation of the environment to tackle assisted and autonomous driving tasks. However most of the existing work focuses on the fully supervised setting training networks on large annotated datasets. In this work we present Re...
Henrique Piñeiro Monteagudo; Leonardo Taccari; Aurel Pjetri; Francesco Sambo; Samuele Salti
Verizon Connect, Italy+University of Bologna, Italy; Verizon Connect, Italy; University of Bologna, Italy+University of Florence, Italy; Verizon Connect, Italy; University of Bologna, Italy
Poster
main
https://henriquepm.github.io/RendBEV/
https://openaccess.thecvf.com/content/WACV2025/html/Monteagudo_RendBEV_Semantic_Novel_View_Synthesis_for_Self-Supervised_Birds_Eye_View_WACV_2025_paper.html
0
RendBEV: Semantic Novel View Synthesis for Self-Supervised Bird's Eye View Segmentation Bird's Eye View (BEV) semantic maps have recently garnered a lot of attention as a useful representation of the environment to tackle assisted and autonomous driving tasks. However most of the existing work focuses on the fully supe...
[ -0.07139787077903748, -0.0064206719398498535, -0.006975601427257061, -0.01775774359703064, 0.008025839924812317, -0.00412757508456707, 0.03742334619164467, 0.05349336937069893, 0.017904503270983696, 0.017620159313082695, -0.0391477532684803, -0.04945752024650574, 0.01220844965428114, 0.043...
wacv_2025_b09b0d2c70
b09b0d2c70
wacv
2,025
Retaining and Enhancing Pre-Trained Knowledge in Vision-Language Models with Prompt Ensembling
The advancement of vision-language models particularly the Contrastive Language-Image Pre-training (CLIP) model has revolutionized the field of machine learning by enabling robust zero-shot learning capabilities. These capabilities allow models to understand and respond to previously unseen data without task-specific t...
Donggeun Kim; Yujin Jo; Myungjoo Lee; Taesup Kim
Graduate School of Data Science, Seoul National University+Nota Inc.; Graduate School of Data Science, Seoul National University; Graduate School of Data Science, Seoul National University; Graduate School of Data Science, Seoul National University
Poster
main
https://openaccess.thecvf.com/content/WACV2025/html/Kim_Retaining_and_Enhancing_Pre-Trained_Knowledge_in_Vision-Language_Models_with_Prompt_WACV_2025_paper.html
0
2412.07077
Retaining and Enhancing Pre-Trained Knowledge in Vision-Language Models with Prompt Ensembling The advancement of vision-language models particularly the Contrastive Language-Image Pre-training (CLIP) model has revolutionized the field of machine learning by enabling robust zero-shot learning capabilities. These capabi...
[ -0.07832448184490204, -0.028943628072738647, 0.0189253818243742, 0.014016439206898212, 0.023005539551377296, -0.056211479008197784, 0.008515510708093643, 0.03489993140101433, -0.013561064377427101, 0.02089259959757328, -0.028998274356126785, -0.031056568026542664, -0.018743231892585754, -0...
wacv_2025_edecbe631a
edecbe631a
wacv
2,025
Rethinking Cluster-Conditioned Diffusion Models for Label-Free Image Synthesis
Diffusion-based image generation models can enhance image quality when conditioned on ground truth labels. Here we conduct a comprehensive experimental study on image-level conditioning for diffusion models using cluster assignments. We investigate how individual clustering determinants such as the number of clusters a...
Nikolaos Adaloglou; Tim Kaiser; Felix Michels; Markus Kollmann
Heinrich Heine University of Dusseldorf; Heinrich Heine University of Dusseldorf; Heinrich Heine University of Dusseldorf; Heinrich Heine University of Dusseldorf
Poster
main
https://github.com/HHU-MMBS/cedm-official-wavc2025
https://openaccess.thecvf.com/content/WACV2025/html/Adaloglou_Rethinking_Cluster-Conditioned_Diffusion_Models_for_Label-Free_Image_Synthesis_WACV_2025_paper.html
0
2403.00570
Rethinking Cluster-Conditioned Diffusion Models for Label-Free Image Synthesis Diffusion-based image generation models can enhance image quality when conditioned on ground truth labels. Here we conduct a comprehensive experimental study on image-level conditioning for diffusion models using cluster assignments. We inve...
[ -0.08182203024625778, -0.03930523246526718, -0.05499812960624695, 0.012390089221298695, -0.026477200910449028, -0.023466354236006737, -0.019871586933732033, 0.02436048537492752, -0.0030473414808511734, 0.009561718441545963, -0.05160408467054367, -0.015829753130674362, -0.03103908896446228, ...
wacv_2025_3d11aefb0f
3d11aefb0f
wacv
2,025
Rethinking Low-Rank Adaptation in Vision: Exploring Head-Level Responsiveness Across Diverse Tasks
Low-rank adaptation (LoRA) has shifted the paradigm of adapting pre-trained Vision Transformers (ViT) achieving great efficiency by updating only a subset of tailored parameters to approximate weight updates. However the multi-head design of the self-attention mechanism with the heads working in parallel in the computa...
Yibo Zhong; Yao Zhou
Sichuan University; Sichuan University
Poster
main
https://openaccess.thecvf.com/content/WACV2025/html/Zhong_Rethinking_Low-Rank_Adaptation_in_Vision_Exploring_Head-Level_Responsiveness_Across_Diverse_WACV_2025_paper.html
0
2404.08894
Rethinking Low-Rank Adaptation in Vision: Exploring Head-Level Responsiveness Across Diverse Tasks Low-rank adaptation (LoRA) has shifted the paradigm of adapting pre-trained Vision Transformers (ViT) achieving great efficiency by updating only a subset of tailored parameters to approximate weight updates. However the ...
[ -0.010526635684072971, -0.014825993217527866, -0.015378120355308056, 0.044785723090171814, -0.029597677290439606, -0.0020851881708949804, 0.044785723090171814, 0.021143784746527672, 0.033326804637908936, -0.03006834350526333, -0.03528188169002533, -0.027751216664910316, -0.019098196178674698...
wacv_2025_700d590958
700d590958
wacv
2,025
Retrieval Augmented Recipe Generation
The growing interest in generating recipes from food images has drawn substantial research attention in recent years. Existing works for recipe generation primarily utilize a two-stage training method--first predicting ingredients from a food image and then generating instructions from both the image and ingredients. L...
Guoshan Liu; Hailong Yin; Bin Zhu; Jingjing Chen; Chong-Wah Ngo; Yu-Gang Jiang
Shanghai Key Lab of Intelligent Information Processing, School of Computer Science, Fudan University+Shanghai Collaborative Innovation Center on Intelligent Visual Computing; Shanghai Key Lab of Intelligent Information Processing, School of Computer Science, Fudan University+Shanghai Collaborative Innovation Center on ...
Poster
main
https://openaccess.thecvf.com/content/WACV2025/html/Liu_Retrieval_Augmented_Recipe_Generation_WACV_2025_paper.html
4
2411.08715
Retrieval Augmented Recipe Generation The growing interest in generating recipes from food images has drawn substantial research attention in recent years. Existing works for recipe generation primarily utilize a two-stage training method--first predicting ingredients from a food image and then generating instructions ...
[ -0.03907683864235878, 0.00034803387825377285, -0.00188198231626302, 0.02504688687622547, -0.03719139099121094, 0.011377385817468166, 0.02238507755100727, 0.019427509978413582, -0.019797205924987793, -0.017421912401914597, -0.019945085048675537, 0.009293226525187492, -0.037893813103437424, ...
wacv_2025_0edac77902
0edac77902
wacv
2,025
Reversing the Damage: A QP-Aware Transformer-Diffusion Approach for 8K Video Restoration under Codec Compression
In this paper we introduce DiQP; a novel Transformer-Diffusion model for restoring 8K video quality degraded by codec compression. To the best of our knowledge our model is the first to consider restoring the artifacts introduced by various codecs (AV1 HEVC) by Denoising Diffusion without considering additional noise. ...
Ali Mollaahmadi Dehaghi; Reza Razavi; Mohammad Moshirpour
University of Calgary; Userful Corporation; University of California, Irvine
Poster
main
https://github.com/alimd94/DiQP
https://openaccess.thecvf.com/content/WACV2025/html/Dehaghi_Reversing_the_Damage_A_QP-Aware_Transformer-Diffusion_Approach_for_8K_Video_WACV_2025_paper.html
1
2412.08912
Reversing the Damage: A QP-Aware Transformer-Diffusion Approach for 8K Video Restoration under Codec Compression In this paper we introduce DiQP; a novel Transformer-Diffusion model for restoring 8K video quality degraded by codec compression. To the best of our knowledge our model is the first to consider restoring th...
[ -0.04310713708400726, -0.031108621507883072, -0.02314000017940998, 0.017933955416083336, -0.027844585478305817, 0.01011121366173029, 0.058752622455358505, 0.028573978692293167, -0.028081638738512993, 0.026057573035359383, -0.010485026985406876, -0.037235524505376816, -0.017241032794117928, ...
wacv_2025_cd7fb60021
cd7fb60021
wacv
2,025
Revisiting Deep Archetypal Analysis for Phenotype Discovery in High Content Imaging
The discovery of unique treatment candidates for complex diseases is a challenging task for current drug discovery programs. Biopharma research has developed automated and scalable screening assays of cell culture models to screen thousands of drug candidates in parallel e.g. by considering bio-image based assays. Howe...
Mario Wieser; Daniel Siegismund; Stephan Steigele
Genedata AG; Genedata AG; Genedata AG
Poster
main
https://openaccess.thecvf.com/content/WACV2025/html/Wieser_Revisiting_Deep_Archetypal_Analysis_for_Phenotype_Discovery_in_High_Content_WACV_2025_paper.html
0
Revisiting Deep Archetypal Analysis for Phenotype Discovery in High Content Imaging The discovery of unique treatment candidates for complex diseases is a challenging task for current drug discovery programs. Biopharma research has developed automated and scalable screening assays of cell culture models to screen thous...
[ -0.011059897020459175, -0.04818689078092575, -0.0270353052765131, 0.010696819983422756, -0.0036493937950581312, -0.019997188821434975, 0.010501316748559475, 0.0051017035730183125, 0.005744071211665869, 0.01769769750535488, -0.03662427142262459, -0.012028103694319725, -0.018898645415902138, ...
wacv_2025_0d205785d7
0d205785d7
wacv
2,025
Revisiting Disparity from Dual-Pixel Images: Physics-Informed Lightweight Depth Estimation
In this study we propose a high-performance disparity (depth) estimation method using dual-pixel (DP) images with few parameters. Conventional end-to-end deep-learning methods have many parameters but do not fully exploit disparity constraints which limits their performance. Therefore we propose a lightweight disparity...
Teppei Kurita; Yuhi Kondo; Legong Sun; Takayuki Sasaki; Sho Nitta; Yasuhiro Hashimoto; Yoshinori Muramatsu; Yusuke Moriuchi
;;;;;;;
Poster
main
https://github.com/sony/dual-pixel-disparity
https://openaccess.thecvf.com/content/WACV2025/html/Kurita_Revisiting_Disparity_from_Dual-Pixel_Images_Physics-Informed_Lightweight_Depth_Estimation_WACV_2025_paper.html
0
2411.04714
Revisiting Disparity from Dual-Pixel Images: Physics-Informed Lightweight Depth Estimation In this study we propose a high-performance disparity (depth) estimation method using dual-pixel (DP) images with few parameters. Conventional end-to-end deep-learning methods have many parameters but do not fully exploit dispari...
[ -0.05965572968125343, -0.038081999868154526, -0.02317718043923378, -0.002170583000406623, -0.030374495312571526, -0.021938148885965347, 0.06461185216903687, 0.03256102278828621, 0.02574634924530983, -0.0038719738367944956, -0.031340211629867554, -0.01308271661400795, -0.03618701174855232, ...
wacv_2025_350e928640
350e928640
wacv
2,025
Revisiting Machine Unlearning with Dimensional Alignment
Machine unlearning an emerging research topic focusing on data privacy compliance enables trained models to erase information learned from specific data. While many existing methods indirectly address this issue by intentionally injecting incorrect supervision they often result in drastic and unpredictable changes to d...
Seonguk Seo; Dongwan Kim; Bohyung Han
ECE; ECE; ECE+IPAI
Poster
main
https://openaccess.thecvf.com/content/WACV2025/html/Seo_Revisiting_Machine_Unlearning_with_Dimensional_Alignment_WACV_2025_paper.html
1
2407.17710
Revisiting Machine Unlearning with Dimensional Alignment Machine unlearning an emerging research topic focusing on data privacy compliance enables trained models to erase information learned from specific data. While many existing methods indirectly address this issue by intentionally injecting incorrect supervision th...
[ -0.04241213575005531, -0.03031584806740284, -0.046694669872522354, -0.015017051249742508, -0.024568233639001846, 0.005047945771366358, 0.0387682244181633, 0.01233107503503561, 0.029602093622088432, 0.024117441847920418, -0.03585685044527054, -0.0341663733124733, -0.008729424327611923, 0.02...
wacv_2025_c6804d9e3f
c6804d9e3f
wacv
2,025
Reviving Poor Object Segmentations in OOD Medical Images using Variational-Deep-PCA Modeling on Segmentation Maps with Sampling-Free Learning
For object segmentation in medical images deep neural networks (DNNs) typically perform poorly on out-of-distribution (OOD) images stemming from the large variability in image-acquisition equipment and protocols across sites. However compared to such variability in the acquired medical images we observe that the variab...
Jimut B. Pal; Shantanu Welling; Himali Saini; Suyash P. Awate
;;;
Poster
main
https://openaccess.thecvf.com/content/WACV2025/html/Pal_Reviving_Poor_Object_Segmentations_in_OOD_Medical_Images_using_Variational-Deep-PCA_WACV_2025_paper.html
0
Reviving Poor Object Segmentations in OOD Medical Images using Variational-Deep-PCA Modeling on Segmentation Maps with Sampling-Free Learning For object segmentation in medical images deep neural networks (DNNs) typically perform poorly on out-of-distribution (OOD) images stemming from the large variability in image-ac...
[ -0.04784734919667244, -0.03084464743733406, -0.04886460676789284, -0.02245228737592697, -0.021144388243556023, -0.010917333886027336, 0.042034462094306946, 0.014523142017424107, -0.002883738372474909, 0.02417798899114132, -0.028701145201921463, -0.02257944457232952, 0.014732043258845806, 0...
wacv_2025_619e3191dd
619e3191dd
wacv
2,025
RiemStega: Covariance-Based Loss for Print-Proof Transmission of Data in Images
Covariance matrices outperform first-order features in many tasks attracting considerable attention from the computer vision research community. Covariance matrices encode second-order statistics between features at the same time it is robust to noise. Based on this we propose representing images by covariance matrices...
Aniana Cruz; Guilherme Schardong; Luiz Schirmer; João Marcos; Farhad Shadmand; Nuno Gonçalves
Institute of Systems and Robotics, University of Coimbra; Institute of Systems and Robotics, University of Coimbra; University of the Sinos River Valley; Institute of Systems and Robotics, University of Coimbra; Institute of Systems and Robotics, University of Coimbra; Institute of Systems and Robotics, University of C...
Poster
main
https://openaccess.thecvf.com/content/WACV2025/html/Cruz_RiemStega_Covariance-Based_Loss_for_Print-Proof_Transmission_of_Data_in_Images_WACV_2025_paper.html
0
RiemStega: Covariance-Based Loss for Print-Proof Transmission of Data in Images Covariance matrices outperform first-order features in many tasks attracting considerable attention from the computer vision research community. Covariance matrices encode second-order statistics between features at the same time it is robu...
[ -0.040200550109148026, -0.0070060607977211475, -0.029953718185424805, 0.0008886363357305527, 0.004598897881805897, 0.009357024915516376, -0.009239944629371166, 0.07170908898115158, 0.01976308599114418, 0.019444627687335014, -0.033662810921669006, 0.027574654668569565, -0.007530578412115574, ...
wacv_2025_94ad84a225
94ad84a225
wacv
2,025
Robot Instance Segmentation with Few Annotations for Grasping
The ability of robots to manipulate objects relies heavily on their aptitude for visual perception. In domains characterized by cluttered scenes and high object variability most methods call for vast labeled datasets laboriously hand-annotated with the aim of training capable models. Once deployed the challenge of gene...
Moshe Kimhi; David Vainshtein; Chaim Baskin; Dotan Di Castro
;;;
Poster
main
https://openaccess.thecvf.com/content/WACV2025/html/Kimhi_Robot_Instance_Segmentation_with_Few_Annotations_for_Grasping_WACV_2025_paper.html
1
2407.01302
Robot Instance Segmentation with Few Annotations for Grasping The ability of robots to manipulate objects relies heavily on their aptitude for visual perception. In domains characterized by cluttered scenes and high object variability most methods call for vast labeled datasets laboriously hand-annotated with the aim o...
[ -0.10409779101610184, -0.03883421793580055, 0.009846590459346771, 0.024441629648208618, -0.0031012159306555986, -0.031159397214651108, 0.016334297135472298, 0.001367709832265973, 0.011088917031884193, 0.031969211995601654, -0.03206123411655426, -0.0345274843275547, 0.0034992205910384655, 0...
wacv_2025_d1b28e143a
d1b28e143a
wacv
2,025
Robust Long-Range Perception Against Sensor Misalignment in Autonomous Vehicles
Advances in machine learning algorithms for sensor fusion have significantly improved the detection and prediction of other road users thereby enhancing safety. However even a small angular displacement in the sensor's placement can cause significant degradation in output especially at long range. In this paper we demo...
Zi-Xiang Xia; Sudeep Fadadu; Yi Shi; Louis Foucard
Aurora Innovation, Inc.; Aurora Innovation, Inc.; Aurora Innovation, Inc.; Aurora Innovation, Inc.
Poster
main
https://openaccess.thecvf.com/content/WACV2025/html/Xia_Robust_Long-Range_Perception_Against_Sensor_Misalignment_in_Autonomous_Vehicles_WACV_2025_paper.html
0
2408.11196
Robust Long-Range Perception Against Sensor Misalignment in Autonomous Vehicles Advances in machine learning algorithms for sensor fusion have significantly improved the detection and prediction of other road users thereby enhancing safety. However even a small angular displacement in the sensor's placement can cause s...
[ -0.040856655687093735, -0.023975428193807602, -0.025962546467781067, 0.006026356481015682, -0.01710408180952072, 0.0075166961178183556, -0.018552634865045547, 0.01615695096552372, 0.007252056151628494, 0.018209068104624748, -0.0337253101170063, -0.013556981459259987, 0.008802752010524273, ...
wacv_2025_98d1e66318
98d1e66318
wacv
2,025
Robust Novelty Detection through Style-Conscious Feature Ranking
Novelty detection seeks to identify samples deviating from a known distribution yet data shifts in a multitude of ways and only a few consist of relevant changes. Aligned with out-of-distribution generalization literature we advocate for a formal distinction between task-relevant semantic or content changes and irrelev...
Stefan Smeu; Elena Burceanu; Emanuela Haller; Andrei Liviu Nicolicioiu
Bitdefender, Bucharest, Romania; Bitdefender, Bucharest, Romania; Bitdefender, Bucharest, Romania; Mila and Université de Montréal, Montréal, Canada
Poster
main
https://github.com/bit-ml/Stylist
https://openaccess.thecvf.com/content/WACV2025/html/Smeu_Robust_Novelty_Detection_through_Style-Conscious_Feature_Ranking_WACV_2025_paper.html
0
2310.03738
Robust Novelty Detection through Style-Conscious Feature Ranking Novelty detection seeks to identify samples deviating from a known distribution yet data shifts in a multitude of ways and only a few consist of relevant changes. Aligned with out-of-distribution generalization literature we advocate for a formal distinct...
[ -0.06903870403766632, -0.026721693575382233, -0.0009620637283660471, 0.0364871583878994, -0.002616077894344926, -0.03582509234547615, -0.03273545578122139, 0.0017574616940692067, 0.04575607553124428, 0.0170481838285923, -0.03095155581831932, 0.008675815537571907, -0.04130552336573601, 0.02...
wacv_2025_a047a8aa94
a047a8aa94
wacv
2,025
Robust Portrait Image Matting and Depth-of-Field Synthesis via Multiplane Images
High-quality portrait photography has become an essential function in our daily lives. However due to the limited aperture and focal length of a smartphone camera images captured by a smartphone cannot match the same level of bokeh effect by a digital single-lens reflex camera. A typical solution on a smartphone is to ...
Zhefan Rao; Tianjia Zhang; Yuen Fui Lau; Qifeng Chen
Hong Kong University of Science and Technology; Hong Kong University of Science and Technology; Hong Kong University of Science and Technology; Hong Kong University of Science and Technology
Poster
main
https://openaccess.thecvf.com/content/WACV2025/html/Rao_Robust_Portrait_Image_Matting_and_Depth-of-Field_Synthesis_via_Multiplane_Images_WACV_2025_paper.html
0
Robust Portrait Image Matting and Depth-of-Field Synthesis via Multiplane Images High-quality portrait photography has become an essential function in our daily lives. However due to the limited aperture and focal length of a smartphone camera images captured by a smartphone cannot match the same level of bokeh effect ...
[ -0.03199910745024681, -0.038200199604034424, -0.02879735827445984, -0.032643139362335205, -0.017683234065771103, 0.0024450153578072786, 0.03707774728536606, 0.033047959208488464, 0.05137522146105766, 0.010672503150999546, -0.07029131799936295, 0.020774580538272858, -0.0034409621730446815, ...
wacv_2025_798f52d9a6
798f52d9a6
wacv
2,025
RopeTP: Global Human Motion Recovery via Integrating Robust Pose Estimation with Diffusion Trajectory Prior
We present RopeTP a novel framework that combines Robust pose estimation with a diffusion Trajectory Prior to reconstruct global human motion from videos. At the heart of RopeTP is a hierarchical attention mechanism that significantly improves context awareness which is essential for accurately inferring the posture of...
Mingjiang Liang; Yongkang Cheng; Hualin Liang; Shaoli Huang; Wei Liu
;;;;
Poster
main
https://openaccess.thecvf.com/content/WACV2025/html/Liang_RopeTP_Global_Human_Motion_Recovery_via_Integrating_Robust_Pose_Estimation_WACV_2025_paper.html
4
2410.20358
RopeTP: Global Human Motion Recovery via Integrating Robust Pose Estimation with Diffusion Trajectory Prior We present RopeTP a novel framework that combines Robust pose estimation with a diffusion Trajectory Prior to reconstruct global human motion from videos. At the heart of RopeTP is a hierarchical attention mechan...
[ -0.028691651299595833, -0.02630068175494671, -0.006616236642003059, 0.04183286800980568, -0.013232473284006119, 0.005991116166114807, -0.00016654664068482816, 0.013643136247992516, 0.008149378001689911, 0.01976657658815384, -0.0344591848552227, -0.013743520714342594, -0.016572531312704086, ...
wacv_2025_d98ddc9fdd
d98ddc9fdd
wacv
2,025
Rubric-Constrained Figure Skating Scoring
Figure skating automatic scoring is the task of estimating the competition score of a performance video. The technical element score (TES) aggregates the technical quality (grade of execution) and difficulty (base value) scores for each element. Most prior work adapted from short-term action quality assessment entangle...
Arushi Rai; Adriana Kovashka
University of Pittsburgh; University of Pittsburgh
Poster
main
https://arushirai1.github.io/rcs-project
https://openaccess.thecvf.com/content/WACV2025/html/Rai_Rubric-Constrained_Figure_Skating_Scoring_WACV_2025_paper.html
0
Rubric-Constrained Figure Skating Scoring Figure skating automatic scoring is the task of estimating the competition score of a performance video. The technical element score (TES) aggregates the technical quality (grade of execution) and difficulty (base value) scores for each element. Most prior work adapted from sho...
[ -0.07006201148033142, -0.016269788146018982, 0.00011914506467292085, 0.03127498924732208, -0.0006771214539185166, -0.015958359465003014, 0.012995068915188313, 0.00608701491728425, 0.003555477363988757, 0.04522321745753288, -0.021705634891986847, -0.018355417996644974, 0.014467277564108372, ...
wacv_2025_a2af271396
a2af271396
wacv
2,025
S3PT: Scene Semantics and Structure Guided Clustering to Boost Self-Supervised Pre-Training for Autonomous Driving
Recent self-supervised clustering-based pre-training techniques like DINO and CriBo have shown impressive results for downstream detection and segmentation tasks. However real-world applications such as autonomous driving face challenges with imbalanced object class and size distributions and complex scene geometries. ...
Maciej K. Wozniak; Hariprasath Govindarajan; Marvin Klingner; Camille Maurice; B Ravi Kiran; Senthil Yogamani
Automated Driving, Qualcomm Technologies International GmbH + KTH Royal Institute of Technology, Sweden; Arriver Software AB and Link ¨oping University, Sweden; Automated Driving, Qualcomm Technologies International GmbH; Automated Driving, Qualcomm Technologies International GmbH; Qualcomm SARL, France; Automated Driv...
Poster
main
https://openaccess.thecvf.com/content/WACV2025/html/Wozniak_S3PT_Scene_Semantics_and_Structure_Guided_Clustering_to_Boost_Self-Supervised_WACV_2025_paper.html
2
2410.23085
S3PT: Scene Semantics and Structure Guided Clustering to Boost Self-Supervised Pre-Training for Autonomous Driving Recent self-supervised clustering-based pre-training techniques like DINO and CriBo have shown impressive results for downstream detection and segmentation tasks. However real-world applications such as au...
[ -0.07271185517311096, -0.026529010385274887, -0.036979276686906815, 0.0258139930665493, -0.029242413118481636, -0.006572667974978685, -0.0006857987609691918, 0.01627124845981598, -0.011366956867277622, 0.020533857867121696, -0.02062552608549595, 0.00885064247995615, -0.0038065556436777115, ...
wacv_2025_820b2c3950
820b2c3950
wacv
2,025
SADA: Semantic Adversarial Unsupervised Domain Adaptation for Temporal Action Localization
Temporal Action Localization (TAL) is a complex task that poses relevant challenges particularly when attempting to generalize on new - unseen - domains in real-world applications. These scenarios despite realistic are often neglected in the literature exposing these solutions to important performance degradation. In t...
David Pujol-Perich; Albert Clapés; Sergio Escalera
Universitat de Barcelona and Computer Vision Center, Barcelona, Spain; Universitat de Barcelona and Computer Vision Center, Barcelona, Spain; Universitat de Barcelona and Computer Vision Center, Barcelona, Spain
Poster
main
https://github.com/davidpujol/SADA
https://openaccess.thecvf.com/content/WACV2025/html/Pujol-Perich_SADA_Semantic_Adversarial_Unsupervised_Domain_Adaptation_for_Temporal_Action_Localization_WACV_2025_paper.html
0
SADA: Semantic Adversarial Unsupervised Domain Adaptation for Temporal Action Localization Temporal Action Localization (TAL) is a complex task that poses relevant challenges particularly when attempting to generalize on new - unseen - domains in real-world applications. These scenarios despite realistic are often negl...
[ -0.07480902969837189, 0.019746921956539154, -0.027678972110152245, -0.00113826768938452, -0.02653261460363865, 0.020449528470635414, 0.031709711998701096, 0.040455304086208344, 0.023925578221678734, 0.0602022260427475, -0.0076038590632379055, 0.00045790718286298215, -0.01396891288459301, -...
wacv_2025_2b3b2415b1
2b3b2415b1
wacv
2,025
SADDLe: Sharpness-Aware Decentralized Deep Learning with Heterogeneous Data
Decentralized training enables learning with distributed datasets generated at different locations without relying on a central server. In realistic scenarios the data distribution across these sparsely connected learning agents can be significantly heterogeneous leading to local model over-fitting and poor global mode...
Sakshi Choudhary; Sai Aparna Aketi; Kaushik Roy
Purdue University; Purdue University; Purdue University
Poster
main
https://openaccess.thecvf.com/content/WACV2025/html/Choudhary_SADDLe_Sharpness-Aware_Decentralized_Deep_Learning_with_Heterogeneous_Data_WACV_2025_paper.html
0
2405.13961
SADDLe: Sharpness-Aware Decentralized Deep Learning with Heterogeneous Data Decentralized training enables learning with distributed datasets generated at different locations without relying on a central server. In realistic scenarios the data distribution across these sparsely connected learning agents can be signific...
[ -0.024233439937233925, -0.017956886440515518, -0.030658552423119545, -0.03673084080219269, -0.014326512813568115, -0.008240298368036747, 0.04456724599003792, 0.009368406608700752, 0.009953351691365242, 0.011596769094467163, -0.0061372811906039715, -0.026461802423000336, -0.016406318172812462...
wacv_2025_5142fc27c1
5142fc27c1
wacv
2,025
SALVE: A 3D Reconstruction Benchmark of Wounds from Consumer-Grade Videos
Managing chronic wounds is a global challenge that can be alleviated by the adoption of automatic systems for clinical wound assessment from consumer-grade videos. While 2D image analysis approaches are insufficient for handling the 3D features of wounds existing approaches utilizing 3D reconstruction methods have not ...
Remi Chierchia; Leo Lebrat; David Ahmedt-Aristizabal; Olivier Salvado; Clinton Fookes; Rodrigo Santa Cruz
;;;;;
Poster
main
https://remichierchia.github.io/SALVE/
https://openaccess.thecvf.com/content/WACV2025/html/Chierchia_SALVE_A_3D_Reconstruction_Benchmark_of_Wounds_from_Consumer-Grade_Videos_WACV_2025_paper.html
0
2407.19652
SALVE: A 3D Reconstruction Benchmark of Wounds from Consumer-Grade Videos Managing chronic wounds is a global challenge that can be alleviated by the adoption of automatic systems for clinical wound assessment from consumer-grade videos. While 2D image analysis approaches are insufficient for handling the 3D features o...
[ -0.040269769728183746, -0.04023285582661629, -0.03209400922060013, -0.030488384887576103, -0.003603424411267042, 0.01776335947215557, 0.057654790580272675, 0.046286240220069885, -0.02087310142815113, 0.01760648936033249, 0.011848391965031624, 0.021223755553364754, 0.012097540311515331, 0.0...
wacv_2025_00dcdce1fc
00dcdce1fc
wacv
2,025
SAM-DA: Decoder Adapter for Efficient Medical Domain Adaptation
This paper addresses the domain adaptation challenge for semantic segmentation in medical imaging. Despite the impressive performance of recent foundational segmentation models like SAM on natural images they struggle with medical domain images. Beyond this recent approaches that perform end-to-end fine-tuning of model...
Javier Gamazo Tejero; Moritz J Schmid; Pablo Márquez Neila; Martin Zinkernagel; Sebastian Wolf; Raphael Sznitman
University of Bern; University of Bern; University of Bern; Inselspital Bern; Inselspital Bern; University of Bern
Poster
main
https://openaccess.thecvf.com/content/WACV2025/html/Tejero_SAM-DA_Decoder_Adapter_for_Efficient_Medical_Domain_Adaptation_WACV_2025_paper.html
0
SAM-DA: Decoder Adapter for Efficient Medical Domain Adaptation This paper addresses the domain adaptation challenge for semantic segmentation in medical imaging. Despite the impressive performance of recent foundational segmentation models like SAM on natural images they struggle with medical domain images. Beyond thi...
[ 0.0046083517372608185, -0.02088080532848835, -0.0384984128177166, -0.008639226667582989, -0.003817759221419692, -0.014033587649464607, 0.004298989195376635, 0.010092084296047688, 0.05580436438322067, 0.058114271610975266, -0.025408951565623283, 0.007406361401081085, 0.02491397224366665, 0....
wacv_2025_abb0427f7f
abb0427f7f
wacv
2,025
SAM-Mamba: Mamba Guided SAM Architecture for Generalized Zero-Shot Polyp Segmentation
Polyp segmentation in colonoscopy images is crucial for detecting colorectal cancer but is challenging due to variations in the structure color and size of polyps as well as the lack of clear boundaries with surrounding tissues. Traditional segmentation models based on Convolutional Neural Networks (CNNs) struggle to c...
Tapas Kumar Dutta; Snehashis Majhi; Deepak Ranjan Nayak; Debesh Jha
University of Surrey, United Kingdom; Côte d’Azur University, France; Malaviya National Institute of Technology Jaipur, India; University of South Dakota, USA
Poster
main
https://github.com/TapasKumarDutta1/SAM_Mamba_2025
https://openaccess.thecvf.com/content/WACV2025/html/Dutta_SAM-Mamba_Mamba_Guided_SAM_Architecture_for_Generalized_Zero-Shot_Polyp_Segmentation_WACV_2025_paper.html
1
SAM-Mamba: Mamba Guided SAM Architecture for Generalized Zero-Shot Polyp Segmentation Polyp segmentation in colonoscopy images is crucial for detecting colorectal cancer but is challenging due to variations in the structure color and size of polyps as well as the lack of clear boundaries with surrounding tissues. Tradi...
[ -0.015905536711215973, -0.0517534539103508, -0.02362576685845852, 0.00615293113514781, 0.012538399547338486, -0.044014617800712585, 0.0349549725651741, 0.00730166444554925, -0.008050433360040188, 0.02334672212600708, 0.006734273862093687, -0.0077620879746973515, -0.02662084437906742, 0.017...
wacv_2025_67d0ab5d3a
67d0ab5d3a
wacv
2,025
SAND: Enhancing Open-Set Neuron Descriptions through Spatial Awareness
We propose Spatially-Aware open-set Network Dissection (SAND) a technique to identify and label the learned representation of the neurons of deep vision networks. Contrary to earlier open-vocabulary neuron explanation methods we also leverage a neuron's spatial pattern of activation to guide our predictions towards mor...
Anvita Agarwal Srinivas; Tuomas Oikarinen; Divyansh Srivastava; Wei-Hung Weng; Tsui-Wei Weng
University of California San Diego; University of California San Diego; University of California San Diego; Google Research; University of California San Diego
Poster
main
https://github.com/Trustworthy-ML-Lab/SAND
https://openaccess.thecvf.com/content/WACV2025/html/Srinivas_SAND_Enhancing_Open-Set_Neuron_Descriptions_through_Spatial_Awareness_WACV_2025_paper.html
0
SAND: Enhancing Open-Set Neuron Descriptions through Spatial Awareness We propose Spatially-Aware open-set Network Dissection (SAND) a technique to identify and label the learned representation of the neurons of deep vision networks. Contrary to earlier open-vocabulary neuron explanation methods we also leverage a neur...
[ -0.04047837853431702, -0.030778490006923676, -0.015967508777976036, 0.008091012015938759, -0.007223618682473898, -0.03684091940522194, 0.0248279832303524, 0.010585935786366463, 0.016881536692380905, 0.059691619127988815, 0.005955171771347523, 0.0059598349034786224, -0.0338749922811985, 0.0...
wacv_2025_6946f9ad02
6946f9ad02
wacv
2,025
SANPO: A Scene Understanding Accessibility and Human Navigation Dataset
Vision is essential for human navigation. The World Health Organization (WHO) estimates that 43.3 million people were blind in 2020 and this number is projected to reach 61 million by 2050. Modern scene understanding models could empower these people by assisting them with navigation obstacle avoidance and visual recog...
Sagar M. Waghmare; Kimberly Wilber; Dave Hawkey; Xuan Yang; Matthew Wilson; Stephanie Debats; Cattalyya Nuengsigkapian; Astuti Sharma; Lars Pandikow; Huisheng Wang; Hartwig Adam; Mikhail Sirotenko
Google Research; Google Research; Google Research; Google Research; Google Research; Google Research; Google Research; Google Research; Parallel Domain; Google Research; Google Research; Google Research
Poster
main
https://openaccess.thecvf.com/content/WACV2025/html/Waghmare_SANPO_A_Scene_Understanding_Accessibility_and_Human_Navigation_Dataset_WACV_2025_paper.html
1
2309.12172
SANPO: A Scene Understanding Accessibility and Human Navigation Dataset Vision is essential for human navigation. The World Health Organization (WHO) estimates that 43.3 million people were blind in 2020 and this number is projected to reach 61 million by 2050. Modern scene understanding models could empower these peop...
[ -0.015669558197259903, -0.04360388219356537, -0.0005469883908517659, -0.013572132214903831, -0.021482154726982117, -0.011004431173205376, 0.007293962873518467, 0.02590273879468441, 0.004293609410524368, 0.020654471591114998, 0.015631934627890587, -0.010035664774477482, 0.04507113993167877, ...
wacv_2025_56fbcc6ed8
56fbcc6ed8
wacv
2,025
SCOT: Self-Supervised Contrastive Pretraining for Zero-Shot Compositional Retrieval
Compositional image retrieval (CIR) is a multimodal learning task where a model combines a query image with a user-provided text modification to retrieve a target image. CIR finds applications in a variety of domains including product retrieval (e-commerce) and web search. Existing methods primarily focus on fully-supe...
Bhavin Jawade; João V. B. Soares; Kapil Thadani; Deen Dayal Mohan; Amir Erfan Eshratifar; Benjamin Culpepper; Paloma de Juan; Srirangaraj Setlur; Venu Govindaraju
Yahoo Research+University at Buffalo, SUNY; Yahoo Research; Yahoo Research; Yahoo Research; Yahoo Research; Yahoo Research; Yahoo Research; University at Buffalo, SUNY; University at Buffalo, SUNY
Poster
main
https://github.com/yahoo/SCOT
https://openaccess.thecvf.com/content/WACV2025/html/Jawade_SCOT_Self-Supervised_Contrastive_Pretraining_for_Zero-Shot_Compositional_Retrieval_WACV_2025_paper.html
0
2501.08347
SCOT: Self-Supervised Contrastive Pretraining for Zero-Shot Compositional Retrieval Compositional image retrieval (CIR) is a multimodal learning task where a model combines a query image with a user-provided text modification to retrieve a target image. CIR finds applications in a variety of domains including product r...
[ -0.07470504939556122, -0.014462603256106377, -0.02877800352871418, 0.043424610048532486, 0.029550815001130104, -0.0049910699017345905, -0.011205757036805153, 0.009834937751293182, 0.006145686376839876, 0.027324384078383446, -0.01908106729388237, -0.027765989303588867, -0.029661215841770172, ...
wacv_2025_ac5b67c17f
ac5b67c17f
wacv
2,025
SEED4D: A Synthetic Ego-Exo Dynamic 4D Data Generator Driving Dataset and Benchmark
Models for egocentric 3D and 4D reconstruction including few-shot interpolation and extrapolation settings can benefit from having images from exocentric viewpoints as supervision signals. No existing dataset provides the necessary mixture of complex dynamic and multi-view data. To facilitate the development of 3D and ...
Marius Kästingschäfer; Théo Gieruc; Sebastian Bernhard; Dylan Campbell; Eldar Insafutdinov; Eyvaz Najafli; Thomas Brox
Continental + University of Freiburg; Continental; Continental; Australian National University; University of Oxford; Continental + University of T ¨ubingen; University of Freiburg
Poster
main
https://openaccess.thecvf.com/content/WACV2025/html/Kastingschafer_SEED4D_A_Synthetic_Ego-Exo_Dynamic_4D_Data_Generator_Driving_Dataset_WACV_2025_paper.html
2
SEED4D: A Synthetic Ego-Exo Dynamic 4D Data Generator Driving Dataset and Benchmark Models for egocentric 3D and 4D reconstruction including few-shot interpolation and extrapolation settings can benefit from having images from exocentric viewpoints as supervision signals. No existing dataset provides the necessary mixt...
[ -0.07987888902425766, -0.014226550236344337, -0.016169093549251556, -0.009975477121770382, -0.02648240514099598, -0.013475809246301651, -0.010951440781354904, 0.05656836926937103, 0.014442388899624348, 0.013194280676543713, -0.0012422425206750631, 0.013119206763803959, 0.005269266664981842, ...
wacv_2025_7bf5b2c0ba
7bf5b2c0ba
wacv
2,025
SEM-Net: Efficient Pixel Modelling for Image Inpainting with Spatially Enhanced SSM
Image inpainting aims to repair a partially damaged image based on the information from known regions of the images. Achieving semantically plausible inpainting results is particularly challenging because it requires the reconstructed regions to exhibit similar patterns to the semantically consistent regions. This requ...
Shuang Chen; Haozheng Zhang; Amir Atapour-Abarghouei; Hubert P. H. Shum
Durham University; Durham University; Durham University; Durham University
Poster
main
https://github.com/ChrisChen1023/SEM-Net
https://openaccess.thecvf.com/content/WACV2025/html/Chen_SEM-Net_Efficient_Pixel_Modelling_for_Image_Inpainting_with_Spatially_Enhanced_WACV_2025_paper.html
0
SEM-Net: Efficient Pixel Modelling for Image Inpainting with Spatially Enhanced SSM Image inpainting aims to repair a partially damaged image based on the information from known regions of the images. Achieving semantically plausible inpainting results is particularly challenging because it requires the reconstructed r...
[ -0.0843048170208931, -0.006062389351427555, -0.02767971344292164, -0.011647858656942844, -0.0177102480083704, -0.00022799879661761224, -0.002554274396970868, 0.010189583525061607, -0.005878958385437727, 0.009208228439092636, 0.006786941085010767, 0.050663597881793976, -0.00025780629948712885...
wacv_2025_9d1e81f544
9d1e81f544
wacv
2,025
SEMU-Net: A Segmentation-Based Corrector for Fabrication Process Variations of Nanophotonics with Microscopic Images
Integrated silicon photonic devices which manipulate light to transmit and process information on a silicon-on-insulator chip are highly sensitive to structural variations. Minor deviations during nanofabrication--the precise process of building structures at the nanometer scale--such as over- or under-etching corner r...
Rambod Azimi; Yijian Kong; Dusan Gostimirovic; James J. Clark; Odile Liboiron-Ladouceur
McGill University; McGill University; McGill University; McGill University; McGill University
Poster
main
https://openaccess.thecvf.com/content/WACV2025/html/Azimi_SEMU-Net_A_Segmentation-Based_Corrector_for_Fabrication_Process_Variations_of_Nanophotonics_WACV_2025_paper.html
0
SEMU-Net: A Segmentation-Based Corrector for Fabrication Process Variations of Nanophotonics with Microscopic Images Integrated silicon photonic devices which manipulate light to transmit and process information on a silicon-on-insulator chip are highly sensitive to structural variations. Minor deviations during nanofa...
[ -0.05202646553516388, -0.03275740519165993, -0.031886592507362366, 0.0381305105984211, 0.02080688066780567, -0.017221724614501, -0.02369724027812481, 0.044059451669454575, -0.007346329744905233, 0.037333808839321136, -0.018639111891388893, 0.002324098953977227, 0.012163595296442509, 0.0201...
wacv_2025_a9e6b10498
a9e6b10498
wacv
2,025
SGD: Street View Synthesis with Gaussian Splatting and Diffusion Prior
Novel View Synthesis (NVS) for street scenes plays a critical role in the autonomous driving simulation. Current mainstream methods such as Neural Radiance Fields (NeRF) and 3D Gaussian Splatting (3DGS) struggle to maintain rendering quality at the viewpoint that deviates significantly from the training viewpoints. Thi...
Zhongrui Yu; Haoran Wang; Jinze Yang; Hanzhang Wang; Jiale Cao; Zhong Ji; Mingming Sun
ETH Zürich; Baidu Research; University of Chinese Academy of Sciences; Harbin Institute of Technology; Tianjin University; Tianjin University; BIMSA
Poster
main
https://openaccess.thecvf.com/content/WACV2025/html/Yu_SGD_Street_View_Synthesis_with_Gaussian_Splatting_and_Diffusion_Prior_WACV_2025_paper.html
18
2403.20079
SGD: Street View Synthesis with Gaussian Splatting and Diffusion Prior Novel View Synthesis (NVS) for street scenes plays a critical role in the autonomous driving simulation. Current mainstream methods such as Neural Radiance Fields (NeRF) and 3D Gaussian Splatting (3DGS) struggle to maintain rendering quality at the ...
[ -0.0625336766242981, -0.028436021879315376, -0.01608087122440338, -0.004545289091765881, -0.010053968988358974, -0.01040097326040268, 0.011871172115206718, 0.011195427738130093, 0.031248576939105988, 0.050297241657972336, -0.024637246504426003, -0.0028878909070044756, -0.006136482581496239, ...
wacv_2025_dbe70f86d6
dbe70f86d6
wacv
2,025
SHIP: Structural Hierarchies for Instance-Dependent Partial Labels
Partial label learning (PLL) aims to train classification models under conditions where each training sample is associated with a candidate set of labels. This set contains multiple labels among which only one is correct. This work addresses instance-dependent noise in PLL by leveraging hierarchical structures within t...
Tushar Kadam; Utkarsh Mishra; Aakarsh Malhotra
AI Garage, Mastercard, Gurugram, Haryana, India; AI Garage, Mastercard, Gurugram, Haryana, India; AI Garage, Mastercard, Gurugram, Haryana, India
Poster
main
https://openaccess.thecvf.com/content/WACV2025/html/Kadam_SHIP_Structural_Hierarchies_for_Instance-Dependent_Partial_Labels_WACV_2025_paper.html
0
SHIP: Structural Hierarchies for Instance-Dependent Partial Labels Partial label learning (PLL) aims to train classification models under conditions where each training sample is associated with a candidate set of labels. This set contains multiple labels among which only one is correct. This work addresses instance-de...
[ -0.07239368557929993, -0.021167203783988953, -0.012265886180102825, 0.0181169081479311, -0.006752245593816042, 0.008845857344567776, 0.022054562345147133, 0.0067152720876038074, 0.008369826711714268, 0.01601867377758026, -0.018736209720373154, -0.013892710208892822, -0.03937654569745064, 0...
wacv_2025_48bfc75dba
48bfc75dba
wacv
2,025
SIGNN - Star Identification using Graph Neural Networks
As a solution for the lost-in-space star identification problem we present Star Identification using Graph Neural Network (SIGNN) a novel approach using Graph Attention Networks. By representing the celestial sphere as a graph data structure created from the ESA's Hipparcos catalogue we are able to accurately capture t...
Floyd Hepburn-Dickins; Mark W. Jones; Mike Edwards; Jay Paul Morgan; Steve Bell
Swansea University, United Kingdom; Swansea University, United Kingdom; Swansea University, United Kingdom; Swansea University, United Kingdom; UK Hydrographic Office, Taunton, Somerset, United Kingdom
Poster
main
https://github.com/FloydHepburn/SIGNN
https://openaccess.thecvf.com/content/WACV2025/html/Hepburn-Dickins_SIGNN_-_Star_Identification_using_Graph_Neural_Networks_WACV_2025_paper.html
0
SIGNN - Star Identification using Graph Neural Networks As a solution for the lost-in-space star identification problem we present Star Identification using Graph Neural Network (SIGNN) a novel approach using Graph Attention Networks. By representing the celestial sphere as a graph data structure created from the ESA's...
[ -0.06351335346698761, -0.032168611884117126, 0.02473500557243824, -0.02181399054825306, -0.01152490172535181, -0.035333044826984406, 0.015382139943540096, -0.005739045329391956, 0.036138199269771576, 0.024866076186299324, -0.014717421494424343, 0.008280890062451363, -0.007354029919952154, ...
wacv_2025_315bda727b
315bda727b
wacv
2,025
SMDAF: A Scalable Sidewalk Material Data Acquisition Framework with Bidirectional Cross-Modal Knowledge Distillation
Ensuring safe and independent navigation poses considerable difficulties for individuals who are blind or have low vision (BLV) as it requires detailed knowledge of their immediate environment. Our research highlights the critical need for accessible data on sidewalk materials and objects which is currently lacking in ...
Jiawei Liu; Wayne Lam; Zhigang Zhu; Hao Tang
The Graduate Center - CUNY; The City College of New York - CUNY; The Graduate Center - CUNY + The City College of New York - CUNY; The Graduate Center - CUNY + Borough of Manhattan Community College - CUNY
Poster
main
https://github.com/FgSurewin/SMDAF-CMKD
https://openaccess.thecvf.com/content/WACV2025/html/Liu_SMDAF_A_Scalable_Sidewalk_Material_Data_Acquisition_Framework_with_Bidirectional_WACV_2025_paper.html
1
SMDAF: A Scalable Sidewalk Material Data Acquisition Framework with Bidirectional Cross-Modal Knowledge Distillation Ensuring safe and independent navigation poses considerable difficulties for individuals who are blind or have low vision (BLV) as it requires detailed knowledge of their immediate environment. Our resea...
[ -0.06455723196268082, -0.0166131891310215, -0.06288041919469833, -0.02260049805045128, 0.0030346636194735765, -0.012311804108321667, 0.004030270501971245, 0.01844492368400097, 0.03184118494391441, 0.044289685785770416, -0.001466070651076734, -0.01166477333754301, -0.003934582695364952, 0.0...
wacv_2025_0307063f49
0307063f49
wacv
2,025
SODA: Spectral Orthogonal Decomposition Adaptation for Diffusion Models
Adapting large-scale pre-trained generative models in a parameter-efficient manner is gaining traction. Traditional methods like low rank adaptation achieve parameter efficiency by imposing constraints but may not be optimal for tasks requiring high representation capacity. We propose a novel spectrum-aware adaptation ...
Xinxi Zhang; Song Wen; Ligong Han; Felix Juefei-Xu; Akash Srivastava; Junzhou Huang; Vladimir Pavlovic; Hao Wang; Molei Tao; Dimitris Metaxas
Rutgers University; Rutgers University; Rutgers University+New York University; New York University; MIT-IBM Watson AI Lab; UT Arlington; Rutgers University; Rutgers University; Georgia Tech; Rutgers University
Poster
main
https://openaccess.thecvf.com/content/WACV2025/html/Zhang_SODA_Spectral_Orthogonal_Decomposition_Adaptation_for_Diffusion_Models_WACV_2025_paper.html
0
SODA: Spectral Orthogonal Decomposition Adaptation for Diffusion Models Adapting large-scale pre-trained generative models in a parameter-efficient manner is gaining traction. Traditional methods like low rank adaptation achieve parameter efficiency by imposing constraints but may not be optimal for tasks requiring hig...
[ -0.024667594581842422, 0.017329998314380646, -0.042624421417713165, -0.015449509024620056, -0.01761575974524021, -0.06788197159767151, 0.022178711369633675, 0.02599499747157097, 0.05136316642165184, 0.02381953038275242, -0.03266889229416847, 0.01967139169573784, -0.011430424638092518, -0.0...
wacv_2025_2bb6693de8
2bb6693de8
wacv
2,025
SPACE: SPAtial-Aware Consistency rEgularization for Anomaly Detection in Industrial Applications
In this paper we propose SPACE a novel anomaly detection methodology that integrates a Feature Encoder (FE) into the structure of the Student-Teacher method. The proposed method has two key elements: Spatial Consistency regularization Loss (SCL) and Feature converter Module (FM). SCL prevents overfitting in student mod...
Daehwan Kim; Hyungmin Kim; Daun Jeong; Sungho Suh; Hansang Cho
Samsung Electro-Mechanics, Suwon, Republic of Korea; Samsung Electro-Mechanics, Suwon, Republic of Korea; Samsung Electro-Mechanics, Suwon, Republic of Korea; German Research Center for Artificial Intelligence (DFKI), Kaiserslautern, Germany+Department of Computer Science, RPTU Kaiserslautern-Landau, Kaiserslautern, Ge...
Poster
main
https://openaccess.thecvf.com/content/WACV2025/html/Kim_SPACE_SPAtial-Aware_Consistency_rEgularization_for_Anomaly_Detection_in_Industrial_Applications_WACV_2025_paper.html
0
2411.05822
SPACE: SPAtial-Aware Consistency rEgularization for Anomaly Detection in Industrial Applications In this paper we propose SPACE a novel anomaly detection methodology that integrates a Feature Encoder (FE) into the structure of the Student-Teacher method. The proposed method has two key elements: Spatial Consistency reg...
[ -0.05124020203948021, -0.021393340080976486, -0.045382060110569, 0.007656368892639875, 0.006710909307003021, -0.009320192039012909, 0.0007455922313965857, 0.03691000118851662, 0.03731784597039223, 0.046679750084877014, -0.01566496677696705, 0.013153010979294777, -0.010464940220117569, -0.0...
wacv_2025_29d86ca6d0
29d86ca6d0
wacv
2,025
STAY Diffusion: Styled Layout Diffusion Model for Diverse Layout-to-Image Generation
In layout-to-image (L2I) synthesis controlled complex scenes are generated from coarse information like bounding boxes. Such a task is exciting to many downstream applications because the input layouts offer strong guidance to the generation process while remaining easily reconfigurable by humans. In this paper we prop...
Ruyu Wang; Xuefeng Hou; Sabrina Schmedding; Marco Huber
Bosch Center for Artificial Intelligence, Renningen, Germany+Institute of Industrial Manufacturing and Management IFF, University of Stuttgart, Stuttgart, Germany; Bosch Center for Artificial Intelligence, Renningen, Germany; Bosch Center for Artificial Intelligence, Renningen, Germany; Institute of Industrial Manufact...
Poster
main
https://openaccess.thecvf.com/content/WACV2025/html/Wang_STAY_Diffusion_Styled_Layout_Diffusion_Model_for_Diverse_Layout-to-Image_Generation_WACV_2025_paper.html
0
STAY Diffusion: Styled Layout Diffusion Model for Diverse Layout-to-Image Generation In layout-to-image (L2I) synthesis controlled complex scenes are generated from coarse information like bounding boxes. Such a task is exciting to many downstream applications because the input layouts offer strong guidance to the gene...
[ -0.06402704864740372, -0.01970490999519825, -0.031123746186494827, 0.00903682317584753, -0.012373496778309345, -0.024357715621590614, -0.03720390796661377, 0.04871543124318123, 0.01889854669570923, 0.02070591039955616, -0.030437875539064407, 0.0275831650942564, -0.03803807497024536, 0.0446...
wacv_2025_26a292a8b5
26a292a8b5
wacv
2,025
STLight: A Fully Convolutional Approach for Efficient Predictive Learning by Spatio-Temporal Joint Processing
Spatio-Temporal predictive Learning is a self-supervised learning paradigm that enables models to identify spatial and temporal patterns by predicting future frames based on past frames. Traditional methods which use recurrent neural networks to capture temporal patterns have proven their effectiveness but come with hi...
Andrea Alfarano; Alberto Alfarano; Linda Friso; Andrea Bacciu; Irene Amerini; Fabrizio Silvestri
DVS, University of Zurich+Max Planck Society; Meta; Google; Sapienza, University of Rome; Sapienza, University of Rome; Sapienza, University of Rome
Poster
main
https://github.com/AlfaranoAndrea/STLight/
https://openaccess.thecvf.com/content/WACV2025/html/Alfarano_STLight_A_Fully_Convolutional_Approach_for_Efficient_Predictive_Learning_by_WACV_2025_paper.html
1
2411.10198
STLight: A Fully Convolutional Approach for Efficient Predictive Learning by Spatio-Temporal Joint Processing Spatio-Temporal predictive Learning is a self-supervised learning paradigm that enables models to identify spatial and temporal patterns by predicting future frames based on past frames. Traditional methods whi...
[ -0.055456336587667465, -0.009576317854225636, -0.039039790630340576, 0.018969004973769188, -0.0039480505511164665, -0.004944244399666786, -0.02532261237502098, 0.007276348769664764, 0.018216121941804886, 0.02234780229628086, -0.03404504805803299, -0.007014675531536341, -0.004689457826316357,...
wacv_2025_5d1bd97253
5d1bd97253
wacv
2,025
STRIDE: Single-Video Based Temporally Continuous Occlusion-Robust 3D Pose Estimation
The capability to accurately estimate 3D human poses is crucial for diverse fields such as action recognition gait recognition and virtual/augmented reality. However a persistent and significant challenge within this field is the accurate prediction of human poses under conditions of severe occlusion. Traditional image...
Rohit Lal; Saketh Bachu; Yash Garg; Arindam Dutta; Calvin-Khang Ta; Hannah Dela Cruz; Dripta S. Raychaudhuri; M. Salman Asif; Amit Roy-Chowdhury
University of California, Riverside‡; University of California, Riverside∗; University of California, Riverside; University of California, Riverside; University of California, Riverside; University of California, Riverside; University of California, Riverside†; University of California, Riverside; University of Califor...
Poster
main
https://openaccess.thecvf.com/content/WACV2025/html/Lal_STRIDE_Single-Video_Based_Temporally_Continuous_Occlusion-Robust_3D_Pose_Estimation_WACV_2025_paper.html
0
2312.16221
STRIDE: Single-Video Based Temporally Continuous Occlusion-Robust 3D Pose Estimation The capability to accurately estimate 3D human poses is crucial for diverse fields such as action recognition gait recognition and virtual/augmented reality. However a persistent and significant challenge within this field is the accur...
[ -0.04434030130505562, -0.0324096754193306, -0.05693993344902992, -0.021222390234470367, 0.005417098756879568, -0.001475067576393485, -0.0054867868311703205, -0.0004814295389223844, 0.03662813454866409, 0.01228371262550354, -0.011410286650061607, -0.03820773586630821, -0.009268535301089287, ...
wacv_2025_bd6e5d3142
bd6e5d3142
wacv
2,025
SUM: Saliency Unification through Mamba for Visual Attention Modeling
Visual attention modeling important for interpreting and prioritizing visual stimuli plays a significant role in applications such as marketing multimedia and robotics. Traditional saliency prediction models especially those based on Convolutional Neural Networks (CNNs) or Transformers achieve notable success by levera...
Alireza Hosseini; Amirhossein Kazerouni; Saeed Akhavan; Michael Brudno; Babak Taati
University of Tehran; University of Toronto+Vector Institute+University Health Network; University of Tehran; University of Toronto+Vector Institute+University Health Network; University of Toronto+Vector Institute+University Health Network
Poster
main
https://arhosseini77.github.io/sum_page/
https://openaccess.thecvf.com/content/WACV2025/html/Hosseini_SUM_Saliency_Unification_through_Mamba_for_Visual_Attention_Modeling_WACV_2025_paper.html
11
2406.17815
SUM: Saliency Unification through Mamba for Visual Attention Modeling Visual attention modeling important for interpreting and prioritizing visual stimuli plays a significant role in applications such as marketing multimedia and robotics. Traditional saliency prediction models especially those based on Convolutional Ne...
[ -0.01386236771941185, -0.01860252395272255, -0.05183074250817299, 0.009838417172431946, -0.0072515918873250484, -0.004009814467281103, -0.005494058597832918, 0.01067713275551796, 0.034264832735061646, 0.00657308055087924, -0.0033525065518915653, 0.0017057026270776987, 0.016218310222029686, ...
wacv_2025_51360bd67e
51360bd67e
wacv
2,025
SV-data2vec: Guiding Video Representation Learning with Latent Skeleton Targets
Recent advancements in action recognition leverage both skeleton and video modalities to achieve state-of-the-art performance. However due to the challenges of early fusion which tends to underutilize the strengths of each modality existing methods often resort to late fusion consequently leading to more complex design...
Zorana Doždor; Tomislav Hrkac; Zoran Kalafatic
University of Zagreb, Faculty of Electrical Engineering and Computing, Croatia; University of Zagreb, Faculty of Electrical Engineering and Computing, Croatia; University of Zagreb, Faculty of Electrical Engineering and Computing, Croatia
Poster
main
github.com/zoranadozdor/SVdata2vec
https://openaccess.thecvf.com/content/WACV2025/html/Dozdor_SV-data2vec_Guiding_Video_Representation_Learning_with_Latent_Skeleton_Targets_WACV_2025_paper.html
0
SV-data2vec: Guiding Video Representation Learning with Latent Skeleton Targets Recent advancements in action recognition leverage both skeleton and video modalities to achieve state-of-the-art performance. However due to the challenges of early fusion which tends to underutilize the strengths of each modality existing...
[ -0.01907811313867569, -0.011469137854874134, -0.02707681618630886, 0.022325847297906876, -0.031530849635601044, 0.0022409360390156507, 0.021249454468488693, 0.02856149524450302, 0.02471989020705223, 0.029155366122722626, 0.030974095687270164, 0.015765424817800522, -0.023829083889722824, 0....
wacv_2025_dceb639235
dceb639235
wacv
2,025
Scene-LLM: Extending Language Model for 3D Visual Reasoning
This paper introduces Scene-LLM a 3D-visual-language model that enhances embodied agents' abilities in interactive 3D indoor environments by integrating the reasoning strengths of Large Language Models (LLMs). Scene-LLM adopts a unified 3D visual feature representation that incorporates dense spatial information and su...
Rao Fu; Jingyu Liu; Xilun Chen; Yixin Nie; Wenhan Xiong
Brown University; University of Chicago; Meta AI; Meta AI; Meta AI
Poster
main
https://openaccess.thecvf.com/content/WACV2025/html/Fu_Scene-LLM_Extending_Language_Model_for_3D_Visual_Reasoning_WACV_2025_paper.html
2
Scene-LLM: Extending Language Model for 3D Visual Reasoning This paper introduces Scene-LLM a 3D-visual-language model that enhances embodied agents' abilities in interactive 3D indoor environments by integrating the reasoning strengths of Large Language Models (LLMs). Scene-LLM adopts a unified 3D visual feature repre...
[ -0.061762694269418716, -0.00399787025526166, -0.011964264325797558, -0.023549284785985947, -0.02085845358669758, 0.008131184615194798, -0.014925983734428883, 0.010158337652683258, 0.03469185158610344, -0.0067902845330536366, 0.008645874448120594, 0.02833499014377594, -0.0025012085679918528, ...
wacv_2025_1bea982769
1bea982769
wacv
2,025
SeCo-INR: Semantically Conditioned Implicit Neural Representations for Improved Medical Image Super-Resolution
Implicit Neural Representations (INRs) have recently advanced the field of deep learning due to their ability to learn continuous representations of signals without the need for large training datasets. Although INR methods have been studied for medical image super-resolution their adaptability to localized priors in m...
Mevan Ekanayake; Zhifeng Chen; Gary Egan; Mehrtash Harandi; Zhaolin Chen
Monash University, Melbourne, Australia; Monash University, Melbourne, Australia; Monash University, Melbourne, Australia; Monash University, Melbourne, Australia; Monash University, Melbourne, Australia
Poster
main
https://openaccess.thecvf.com/content/WACV2025/html/Ekanayake_SeCo-INR_Semantically_Conditioned_Implicit_Neural_Representations_for_Improved_Medical_Image_WACV_2025_paper.html
0
SeCo-INR: Semantically Conditioned Implicit Neural Representations for Improved Medical Image Super-Resolution Implicit Neural Representations (INRs) have recently advanced the field of deep learning due to their ability to learn continuous representations of signals without the need for large training datasets. Althou...
[ -0.06762836873531342, -0.0038266051560640335, 0.01986287720501423, 0.02439204975962639, 0.005393171217292547, -0.00820344127714634, 0.01287812925875187, 0.01720721833407879, 0.014324191026389599, 0.01979012042284012, 0.00442913081496954, -0.013814886100590229, -0.0038266051560640335, 0.058...
wacv_2025_6383f8c4a1
6383f8c4a1
wacv
2,025
Secrets of Edge-Informed Contrast Maximization for Event-Based Vision
Event cameras capture the motion of intensity gradients (edges) in the image plane in the form of rapid asynchronous events. When accumulated in 2D histograms these events depict overlays of the edges in motion consequently obscuring the spatial structure of the generating edges. Contrast maximization (CM) is an optimi...
Pritam P. Karmokar; Quan H. Nguyen; William J. Beksi
The University of Texas at Arlington; The University of Texas at Arlington; The University of Texas at Arlington
Poster
main
https://openaccess.thecvf.com/content/WACV2025/html/Karmokar_Secrets_of_Edge-Informed_Contrast_Maximization_for_Event-Based_Vision_WACV_2025_paper.html
1
2409.14611
Secrets of Edge-Informed Contrast Maximization for Event-Based Vision Event cameras capture the motion of intensity gradients (edges) in the image plane in the form of rapid asynchronous events. When accumulated in 2D histograms these events depict overlays of the edges in motion consequently obscuring the spatial stru...
[ -0.025315847247838974, -0.03289952874183655, -0.014807555824518204, 0.0054155983962118626, -0.008497046306729317, -0.03310249745845795, -0.02430099807679653, 0.041627220809459686, 0.007053194101899862, 0.029818082228302956, -0.0321430042386055, 0.018405653536319733, -0.00974254123866558, 0...
wacv_2025_de039c8de2
de039c8de2
wacv
2,025
Seeing Eye to AI: Comparing Human Gaze and Model Attention in Video Memorability
Understanding what makes a video memorable has important applications in advertising and education technology. Towards this goal we investigate spatio-temporal attention mechanisms underlying video memorability. Different from previous works that fuse multiple features we adopt a simple CNN+Transformer architecture tha...
Prajneya Kumar; Eshika Khandelwal; Makarand Tapaswi; Vishnu Sreekumar
Cognitive Science Lab; CVIT; CVIT; Cognitive Science Lab
Poster
main
https://github.com/katha-ai/video-memorability
https://openaccess.thecvf.com/content/WACV2025/html/Kumar_Seeing_Eye_to_AI_Comparing_Human_Gaze_and_Model_Attention_WACV_2025_paper.html
1
2311.16484
Seeing Eye to AI: Comparing Human Gaze and Model Attention in Video Memorability Understanding what makes a video memorable has important applications in advertising and education technology. Towards this goal we investigate spatio-temporal attention mechanisms underlying video memorability. Different from previous wor...
[ -0.04957692325115204, -0.022870196029543877, -0.002605023328214884, 0.00043096955050714314, -0.008245748467743397, -0.042127348482608795, 0.002774966647848487, 0.024862956255674362, 0.007165560033172369, 0.00757528655230999, -0.007081752177327871, -0.06343313306570053, -0.0016191186150535941...
wacv_2025_2261cba6c4
2261cba6c4
wacv
2,025
SegBuilder: A Semi-Automatic Annotation Tool for Segmentation
This paper addresses the problem of image annotation for segmentation tasks. Semantic segmentation involves labeling each pixel in an image with predefined categories such as sky cars roads and humans. Deep learning models require numerous annotated images for effective training but manual annotation is slow and time-c...
Md Alimoor Reza; Eric Manley; Sean Chen; Sameer Chaudhary; Jacob Elafros
Drake University, Des Moines, Iowa, USA; Drake University, Des Moines, Iowa, USA; Drake University, Des Moines, Iowa, USA; Drake University, Des Moines, Iowa, USA; Drake University, Des Moines, Iowa, USA
Poster
main
https://github.com/alimoorreza/segbuilder-v1
https://openaccess.thecvf.com/content/WACV2025/html/Reza_SegBuilder_A_Semi-Automatic_Annotation_Tool_for_Segmentation_WACV_2025_paper.html
0
SegBuilder: A Semi-Automatic Annotation Tool for Segmentation This paper addresses the problem of image annotation for segmentation tasks. Semantic segmentation involves labeling each pixel in an image with predefined categories such as sky cars roads and humans. Deep learning models require numerous annotated images f...
[ -0.060739997774362564, -0.04540051519870758, -0.03344153240323067, 0.017883948981761932, 0.006315723992884159, -0.029115941375494003, 0.030460873618721962, -0.0008366062538698316, 0.005338831339031458, 0.026462430134415627, 0.013767370022833347, -0.016021037474274635, 0.005470598582178354, ...
wacv_2025_6bee39bf60
6bee39bf60
wacv
2,025
SegDesicNet: Lightweight Semantic Segmentation in Remote Sensing with Geo-Coordinate Embeddings for Domain Adaptation
Semantic segmentation is essential for analyzing high-definition remote sensing images (HRSIs) because it allows the precise classification of objects and regions at the pixel level. However remote sensing data present challenges owing to geographical location weather and environmental variations making it difficult fo...
Sachin Verma; Frank Lindseth; Gabriel Kiss
Norwegian University of Science and Technology (NTNU), Norway; Norwegian University of Science and Technology (NTNU), Norway; Norwegian University of Science and Technology (NTNU), Norway
Poster
main
https://openaccess.thecvf.com/content/WACV2025/html/Verma_SegDesicNet_Lightweight_Semantic_Segmentation_in_Remote_Sensing_with_Geo-Coordinate_Embeddings_WACV_2025_paper.html
0
SegDesicNet: Lightweight Semantic Segmentation in Remote Sensing with Geo-Coordinate Embeddings for Domain Adaptation Semantic segmentation is essential for analyzing high-definition remote sensing images (HRSIs) because it allows the precise classification of objects and regions at the pixel level. However remote sens...
[ -0.0939573273062706, 0.009562192484736443, -0.019697746261954308, -0.0058168466202914715, -0.011125065386295319, -0.03486408665776253, -0.003393931081518531, -0.050529807806015015, 0.05356307700276375, 0.031997278332710266, -0.06170111149549484, 0.0014692392433062196, -0.01159670203924179, ...
wacv_2025_20086d7921
20086d7921
wacv
2,025
Segment Anything Meets Point Tracking
Foundation models have marked a significant stride toward addressing generalization challenges in deep learning. While the Segment Anything Model (SAM) has established a strong foothold in image segmentation existing video segmentation methods still require extensive mask labeling for fine-tuning or face performance dr...
Frano Rajič; Lei Ke; Yu-Wing Tai; Chi-Keung Tang; Martin Danelljan; Fisher Yu
ETH Z ¨urich; ETH Z ¨urich; HKUST; HKUST; ETH Z ¨urich; ETH Z ¨urich
Poster
main
https://github.com/SysCV/sam-pt
https://openaccess.thecvf.com/content/WACV2025/html/Rajic_Segment_Anything_Meets_Point_Tracking_WACV_2025_paper.html
80
Segment Anything Meets Point Tracking Foundation models have marked a significant stride toward addressing generalization challenges in deep learning. While the Segment Anything Model (SAM) has established a strong foothold in image segmentation existing video segmentation methods still require extensive mask labeling ...
[ 0.0015932864043861628, -0.0613844059407711, -0.04296908527612686, 0.029771437868475914, -0.010308969765901566, -0.020906807854771614, 0.0027239329647272825, -0.0028322585858404636, 0.0020751089323312044, 0.01975133642554283, -0.010164535604417324, 0.007817484438419342, 0.00582700502127409, ...