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_b5925d1a46
b5925d1a46
wacv
2,025
Self-Aligning Depth-Regularized Radiance Fields for Asynchronous RGB-D Sequences
It has been shown that learning radiance fields with depth rendering and depth supervision can effectively promote the quality and convergence of view synthesis. However this paradigm requires input RGB-D sequences to be synchronized. In the UAV city modeling scenario there exists asynchrony between RGB images and dept...
Yuxin Huang; Andong Yang; Yuantao Chen; Runyi Yang; Zhenxin Zhu; Chao Hou; Hao Zhao; Guyue Zhou
;;;;;;;
Poster
main
github.com/saythe17/AsyncNeRF
https://openaccess.thecvf.com/content/WACV2025/html/Huang_Self-Aligning_Depth-Regularized_Radiance_Fields_for_Asynchronous_RGB-D_Sequences_WACV_2025_paper.html
0
Self-Aligning Depth-Regularized Radiance Fields for Asynchronous RGB-D Sequences It has been shown that learning radiance fields with depth rendering and depth supervision can effectively promote the quality and convergence of view synthesis. However this paradigm requires input RGB-D sequences to be synchronized. In t...
[ -0.051964666694402695, -0.010441324673593044, -0.013614668510854244, -0.025870662182569504, -0.005811593029648066, 0.03052365779876709, 0.041355833411216736, 0.05598485469818115, 0.006458359304815531, 0.06882712990045547, -0.0278435330837965, -0.0077705043368041515, 0.021478231996297836, 0...
wacv_2025_1e66576298
1e66576298
wacv
2,025
Self-Relaxed Joint Training: Sample Selection for Severity Estimation with Ordinal Noisy Labels
Severity level estimation is a crucial task in medical image diagnosis. However accurately assigning severity class labels to individual images is very costly and challenging. Consequently the attached labels tend to be noisy. In this paper we propose a new framework for training with "ordinal" noisy labels. Since seve...
Shumpei Takezaki; Kiyohito Tanaka; Seiichi Uchida
Kyushu University, Fukuoka, Japan; Kyoto Second Red Cross Hospital, Kyoto, Japan; Kyushu University, Fukuoka, Japan
Poster
main
https://github.com/shumpei-takezaki/Self-Relaxed-Joint-Training
https://openaccess.thecvf.com/content/WACV2025/html/Takezaki_Self-Relaxed_Joint_Training_Sample_Selection_for_Severity_Estimation_with_Ordinal_WACV_2025_paper.html
0
2410.21885
Self-Relaxed Joint Training: Sample Selection for Severity Estimation with Ordinal Noisy Labels Severity level estimation is a crucial task in medical image diagnosis. However accurately assigning severity class labels to individual images is very costly and challenging. Consequently the attached labels tend to be nois...
[ -0.05616394802927971, -0.011714835651218891, -0.05687393620610237, -0.003542943624779582, 0.0038395512383431196, -0.014040987007319927, 0.02711040899157524, 0.006497343070805073, 0.032958485186100006, 0.015619780868291855, -0.010332223027944565, 0.012835872359573841, 0.001372102415189147, ...
wacv_2025_6bec36f5ea
6bec36f5ea
wacv
2,025
Self-Supervised Anomaly Segmentation via Diffusion Models with Dynamic Transformer UNet
A robust anomaly detection mechanism should possess the capability to effectively remediate anomalies restoring them to a healthy state while preserving essential healthy information. Despite the efficacy of existing generative models in learning the underlying distribution of healthy reference data they face primary c...
Komal Kumar; Snehashis Chakraborty; Dwarikanath Mahapatra; Behzad Bozorgtabar; Sudipta Roy
Artificial Intelligence & Data Science, Jio Institute, Navi Mumbai, India; Artificial Intelligence & Data Science, Jio Institute, Navi Mumbai, India; Faculty of IT, Monash University, Australia; Swiss Federal Institute of Technology Lausanne (EPFL) University Hospital Center (CHUV), Lausanne, Switzerland; Artificial In...
Poster
main
https://openaccess.thecvf.com/content/WACV2025/html/Kumar_Self-Supervised_Anomaly_Segmentation_via_Diffusion_Models_with_Dynamic_Transformer_UNet_WACV_2025_paper.html
0
Self-Supervised Anomaly Segmentation via Diffusion Models with Dynamic Transformer UNet A robust anomaly detection mechanism should possess the capability to effectively remediate anomalies restoring them to a healthy state while preserving essential healthy information. Despite the efficacy of existing generative mode...
[ -0.05550609901547432, -0.0027739328797906637, -0.0270029678940773, 0.019776564091444016, -0.010729837231338024, 0.0119189927354455, 0.00045508050243370235, 0.01789220981299877, 0.0011657150462269783, 0.045041535049676895, -0.030058180913329124, 0.01316303201019764, -0.03649791330099106, 0....
wacv_2025_0cb1ca4ca4
0cb1ca4ca4
wacv
2,025
Self-Supervised Incremental Learning of Object Representations from Arbitrary Image Sets
Computing a comprehensive and robust visual representation of an arbitrary object or category of objects is a complex problem. The difficulty increases when one starts from a set of uncalibrated images obtained from different sources. We propose a self-supervised approach Multi-Image Latent Embedding (MILE) which compu...
George Leotescu; Alin-Ionut Popa; Diana-Nicoleta N Grigore; Daniel Voinea; Pietro Perona
Amazon Inc.+California Institute of Technology; Amazon Inc.+California Institute of Technology; Amazon Inc.+California Institute of Technology; Amazon Inc.+California Institute of Technology; Amazon Inc.+California Institute of Technology
Poster
main
https://github.com/amazon-science/mile
https://openaccess.thecvf.com/content/WACV2025/html/Leotescu_Self-Supervised_Incremental_Learning_of_Object_Representations_from_Arbitrary_Image_Sets_WACV_2025_paper.html
0
Self-Supervised Incremental Learning of Object Representations from Arbitrary Image Sets Computing a comprehensive and robust visual representation of an arbitrary object or category of objects is a complex problem. The difficulty increases when one starts from a set of uncalibrated images obtained from different sourc...
[ -0.061113398522138596, -0.02809702605009079, -0.019166523590683937, 0.032202787697315216, -0.010113037191331387, -0.01708526350557804, 0.03218386694788933, 0.020907215774059296, -0.0032921768724918365, 0.012298361398279667, -0.021853242069482803, -0.00717562111094594, 0.0010471345158293843, ...
wacv_2025_30034d9c9c
30034d9c9c
wacv
2,025
Self-Supervised Learning with Probabilistic Density Labeling for Rainfall Probability Estimation
Numerical weather prediction (NWP) models are fundamental in meteorology for simulating and forecasting the behavior of various atmospheric variables. The accuracy of precipitation forecasts and the acquisition of sufficient lead time are crucial for preventing hazardous weather events. However the performance of NWP m...
Junha Lee; Sojung An; Sujeong You; Namik Cho
Seoul National University; Korea Institute of Atmospheric Prediction Systems; Korea Institute of Industrial Technology; Seoul National University
Poster
main
https://github.com/joonha425/SSLPDL
https://openaccess.thecvf.com/content/WACV2025/html/Lee_Self-Supervised_Learning_with_Probabilistic_Density_Labeling_for_Rainfall_Probability_Estimation_WACV_2025_paper.html
0
2412.05825
Self-Supervised Learning with Probabilistic Density Labeling for Rainfall Probability Estimation Numerical weather prediction (NWP) models are fundamental in meteorology for simulating and forecasting the behavior of various atmospheric variables. The accuracy of precipitation forecasts and the acquisition of sufficien...
[ -0.1264614462852478, 0.013788117095828056, -0.0053931549191474915, 0.0033185116481035948, -0.04817579314112663, -0.02961415797472, 0.0013471412239596248, -0.01548638753592968, -0.02190309390425682, 0.01516509335488081, -0.0071006049402058125, -0.0054482342675328255, -0.008661177940666676, ...
wacv_2025_30c42b28d6
30c42b28d6
wacv
2,025
Self-Supervised Learning with Spectral Low-Rank Prior for Hyperspectral Image Reconstruction
Hyperspectral image (HSI) reconstruction from coded measurement is significant for acquiring images with higher spectral resolution than traditional RGB images. Current advanced neural networks have already shown impressive performance in some datasets like CAVE and KAIST. However these networks rely on a large amount ...
Zijun He; Lishun Wang; Ziyi Meng; Xin Yuan
Zhejiang University+Westlake University; Westlake University; Westlake Intelligent Vision; Westlake University
Poster
main
https://github.com/zjhe02/CASSI-SSL
https://openaccess.thecvf.com/content/WACV2025/html/He_Self-Supervised_Learning_with_Spectral_Low-Rank_Prior_for_Hyperspectral_Image_Reconstruction_WACV_2025_paper.html
0
Self-Supervised Learning with Spectral Low-Rank Prior for Hyperspectral Image Reconstruction Hyperspectral image (HSI) reconstruction from coded measurement is significant for acquiring images with higher spectral resolution than traditional RGB images. Current advanced neural networks have already shown impressive per...
[ -0.05025099962949753, -0.009350944310426712, 0.00025837833527475595, -0.007153151091188192, -0.006354787852615118, -0.017958583310246468, 0.04757143557071686, 0.001933966064825654, -0.021932046860456467, 0.02585962787270546, -0.05737203359603882, -0.033916670829057693, 0.00818092841655016, ...
wacv_2025_10c2c16499
10c2c16499
wacv
2,025
Self-Supervised Pre-Training with Diffusion Model for Few-Shot Landmark Detection in X-Ray Images
Deep neural networks have been extensively applied in the medical domain for various tasks including image classification segmentation and landmark detection. However their application is often hindered by data scarcity both in terms of available annotations and images. This study introduces a novel application of deno...
Roberto Di Via; Francesca Odone; Vito Paolo Pastore
MaLGa, DIBRIS, University of Genoa, Italy; MaLGa, DIBRIS, University of Genoa, Italy; MaLGa, DIBRIS, University of Genoa, Italy
Poster
main
https://github.com/Malga-Vision/DiffusionXray-FewShot-LandmarkDetection
https://openaccess.thecvf.com/content/WACV2025/html/Di_Via_Self-Supervised_Pre-Training_with_Diffusion_Model_for_Few-Shot_Landmark_Detection_in_WACV_2025_paper.html
2
Self-Supervised Pre-Training with Diffusion Model for Few-Shot Landmark Detection in X-Ray Images Deep neural networks have been extensively applied in the medical domain for various tasks including image classification segmentation and landmark detection. However their application is often hindered by data scarcity bo...
[ -0.07316341996192932, -0.03018530085682869, -0.02499270997941494, 0.020518817007541656, -0.017185857519507408, 0.00708815548568964, 0.023411575704813004, 0.034784965217113495, 0.021668734028935432, 0.01169680431485176, -0.040750157088041306, -0.028298718854784966, 0.0005440765526145697, 0....
wacv_2025_03283df41b
03283df41b
wacv
2,025
Semantic Clustering of Image Retrieval Databases used for Visual Localization
Accurate self-localization of unmanned aerial systems (UAS) is needed to reduce their dependency on global navigation satellite systems (GNSS). Image retrieval techniques comparing aerial images with a reference database can be used for visual localization (VL). But the search space may be vast and a full search not fe...
Henry Hölzemann; Torsten Fiolka
Fraunhofer Institute for Communication, Information Processing and Ergonomics FKIE; Fraunhofer Institute for Communication, Information Processing and Ergonomics FKIE
Poster
main
https://openaccess.thecvf.com/content/WACV2025/html/Holzemann_Semantic_Clustering_of_Image_Retrieval_Databases_used_for_Visual_Localization_WACV_2025_paper.html
0
Semantic Clustering of Image Retrieval Databases used for Visual Localization Accurate self-localization of unmanned aerial systems (UAS) is needed to reduce their dependency on global navigation satellite systems (GNSS). Image retrieval techniques comparing aerial images with a reference database can be used for visua...
[ -0.03596636652946472, 0.015606469474732876, -0.024695543572306633, 0.020907655358314514, -0.025809628888964653, 0.012013547122478485, 0.04166676849126816, 0.014678065665066242, 0.021408993750810623, 0.019793571904301643, -0.014752338640391827, -0.018113160505890846, 0.012245647609233856, 0...
wacv_2025_e35f209911
e35f209911
wacv
2,025
Semantic Prompt Learning for Weakly-Supervised Semantic Segmentation
Weakly-Supervised Semantic Segmentation (WSSS) aims to train segmentation models using image data with only image-level supervision. Since precise pixel-level annotations are not accessible existing methods typically focus on producing pseudo masks for training segmentation models by refining CAM-like heatmaps. However...
Ci-Siang Lin; Chien-Yi Wang; Yu-Chiang Frank Wang; Min-Hung Chen
;;;
Poster
main
https://openaccess.thecvf.com/content/WACV2025/html/Lin_Semantic_Prompt_Learning_for_Weakly-Supervised_Semantic_Segmentation_WACV_2025_paper.html
0
2401.11791
Semantic Prompt Learning for Weakly-Supervised Semantic Segmentation Weakly-Supervised Semantic Segmentation (WSSS) aims to train segmentation models using image data with only image-level supervision. Since precise pixel-level annotations are not accessible existing methods typically focus on producing pseudo masks fo...
[ -0.07099427282810211, -0.0011898793745785952, -0.013378586620092392, -0.0014630031073465943, -0.006554970517754555, -0.038434334099292755, 0.01912761852145195, -0.0013465897645801306, -0.0013868866953998804, 0.0333479642868042, -0.004746085032820702, -0.005234126001596451, 0.0007410161197185...
wacv_2025_9ad3809afa
9ad3809afa
wacv
2,025
Semantic Prompting with Image Token for Continual Learning
Continual learning aims to refine model parameters for new tasks while retaining knowledge from previous tasks. Recently prompt-based learning has emerged to leverage pre-trained models to be prompted to learn subsequent tasks without the reliance on the rehearsal buffer. Although this approach has demonstrated outstan...
Jisu Han; Jaemin Na; Wonjun Hwang
;;
Poster
main
https://github.com/pilsHan/I-Prompt
https://openaccess.thecvf.com/content/WACV2025/html/Han_Semantic_Prompting_with_Image_Token_for_Continual_Learning_WACV_2025_paper.html
2
2403.11537
Semantic Prompting with Image Token for Continual Learning Continual learning aims to refine model parameters for new tasks while retaining knowledge from previous tasks. Recently prompt-based learning has emerged to leverage pre-trained models to be prompted to learn subsequent tasks without the reliance on the rehear...
[ -0.08296941220760345, -0.032828304916620255, -0.015322932042181492, 0.009756792336702347, 0.002158369403332472, -0.026189284399151802, 0.004580373875796795, 0.003778005950152874, -0.012223499827086926, -0.014460960403084755, -0.033525217324495316, -0.04126463085412979, -0.010417026467621326,...
wacv_2025_4208a67211
4208a67211
wacv
2,025
Semantic Segmentation Method for Automated Indoor 3D Reconstruction Based on Architectural-Knowledge-Aware Features
3D point cloud semantic segmentation is an important step for 3D indoors reconstruction. In recent years many outstanding deep learning models have been proposed for semantic segmentation which can achieve remarkable performance. However it is found the index indicating semantic prediction accuracy in terms of structur...
Yahan Chen; Wenzheng Liu; Xiaowei Luo
;;
Poster
main
https://openaccess.thecvf.com/content/WACV2025/html/Chen_Semantic_Segmentation_Method_for_Automated_Indoor_3D_Reconstruction_Based_on_WACV_2025_paper.html
0
Semantic Segmentation Method for Automated Indoor 3D Reconstruction Based on Architectural-Knowledge-Aware Features 3D point cloud semantic segmentation is an important step for 3D indoors reconstruction. In recent years many outstanding deep learning models have been proposed for semantic segmentation which can achiev...
[ 0.0077786631882190704, -0.02262226864695549, -0.03151216730475426, -0.008975728414952755, 0.013118023984134197, -0.004085830878466368, 0.05330326780676842, 0.0016555633628740907, 0.00548391230404377, 0.015466981567442417, -0.022423509508371353, 0.021375512704253197, -0.0008435919298790395, ...
wacv_2025_0c4dc24d98
0c4dc24d98
wacv
2,025
Semantically Conditioned Prompts for Visual Recognition under Missing Modality Scenarios
This paper tackles the domain of multimodal prompting for visual recognition specifically when dealing with missing modalities through multimodal Transformers. It presents two main contributions: (i) we introduce a novel prompt learning module which is designed to produce sample-specific prompts and (ii) we show that m...
Vittorio Pipoli; Federico Bolelli; Sara Sarto; Marcella Cornia; Lorenzo Baraldi; Costantino Grana; Rita Cucchiara; Elisa Ficarra
University of Modena and Reggio Emilia, Italy + University of Pisa, Italy; University of Modena and Reggio Emilia, Italy; University of Modena and Reggio Emilia, Italy; University of Modena and Reggio Emilia, Italy; University of Modena and Reggio Emilia, Italy; University of Modena and Reggio Emilia, Italy; University...
Poster
main
https://github.com/vittoriopipoli/SCP_WACV2025
https://openaccess.thecvf.com/content/WACV2025/html/Pipoli_Semantically_Conditioned_Prompts_for_Visual_Recognition_under_Missing_Modality_Scenarios_WACV_2025_paper.html
0
Semantically Conditioned Prompts for Visual Recognition under Missing Modality Scenarios This paper tackles the domain of multimodal prompting for visual recognition specifically when dealing with missing modalities through multimodal Transformers. It presents two main contributions: (i) we introduce a novel prompt lea...
[ -0.05647933855652809, -0.029304295778274536, 0.008422277867794037, 0.0277885552495718, -0.021400796249508858, -0.01889260858297348, 0.02017376944422722, 0.009365104138851166, 0.03666645660996437, 0.005070509389042854, -0.006820827256888151, -0.04576089605689049, 0.011386089958250523, 0.037...
wacv_2025_1efd8399b4
1efd8399b4
wacv
2,025
Semiotic-Based Construction of a Large Emotional Image Dataset with Neutral Samples
Image Visual Sentiment Analysis (VSA) requires the availability of large annotated datasets whose construction presents many challenges. The necessity of gathering a large amount of labeled images contrasts with the rigorous but lengthy process required for manual annotation based on psychovisual experiments and with t...
Marco Blanchini; Giovanna Dimitri; Lydia Abady; Benedetta Tondi; Tarcisio Lancioni; Mauro Barni
University of Siena, Italy; University of Siena, Italy; University of Siena, Italy; University of Siena, Italy; University of Siena, Italy; University of Siena, Italy
Poster
main
https://openaccess.thecvf.com/content/WACV2025/html/Blanchini_Semiotic-Based_Construction_of_a_Large_Emotional_Image_Dataset_with_Neutral_WACV_2025_paper.html
0
Semiotic-Based Construction of a Large Emotional Image Dataset with Neutral Samples Image Visual Sentiment Analysis (VSA) requires the availability of large annotated datasets whose construction presents many challenges. The necessity of gathering a large amount of labeled images contrasts with the rigorous but lengthy...
[ -0.09179585427045822, -0.054955728352069855, -0.015308824367821217, 0.035032372921705246, -0.004643074236810207, -0.01952374540269375, 0.027344660833477974, 0.03769643232226372, 0.004507492296397686, 0.00744985044002533, -0.0035726933274418116, 0.005061711650341749, 0.016003381460905075, 0...
wacv_2025_8e33b5f31d
8e33b5f31d
wacv
2,025
SenCLIP: Enhancing Zero-Shot Land-Use Mapping for Sentinel-2 with Ground-Level Prompting
Pre-trained vision-language models (VLMs) such as CLIP demonstrate impressive zero-shot classification capabilities with free-form prompts and even show some generalization in specialized domains. However their performance on satellite imagery is limited due to the under representation of such data in their training se...
Pallavi Jain; Dino Ienco; Roberto Interdonato; Tristan Berchoux; Diego Marcos
Mediterranean Agronomic Institute of Montpellier - CIHEAM-IAMM+Inria+Univ. of Montpellier; Inria+INRAE+UMR TETIS+Univ. of Montpellier; Inria+Cirad+UMR TETIS+Univ. of Montpellier; Mediterranean Agronomic Institute of Montpellier - CIHEAM-IAMM; Inria+Univ. of Montpellier
Poster
main
https://github.com/pallavijain-pj/SenCLIP
https://openaccess.thecvf.com/content/WACV2025/html/Jain_SenCLIP_Enhancing_Zero-Shot_Land-Use_Mapping_for_Sentinel-2_with_Ground-Level_Prompting_WACV_2025_paper.html
4
SenCLIP: Enhancing Zero-Shot Land-Use Mapping for Sentinel-2 with Ground-Level Prompting Pre-trained vision-language models (VLMs) such as CLIP demonstrate impressive zero-shot classification capabilities with free-form prompts and even show some generalization in specialized domains. However their performance on satel...
[ -0.08976687490940094, 0.011439803056418896, -0.00756723852828145, -0.01185944490134716, 0.0010593677870929241, -0.0212740208953619, -0.0047711459919810295, 0.02503255195915699, -0.022605927661061287, 0.014915533363819122, -0.0055100806057453156, -0.0025155707262456417, 0.004570447839796543, ...
wacv_2025_065d977b7d
065d977b7d
wacv
2,025
SensorFlow: Sensor and Image Fused Video Stabilization
We present SensorFlow a novel image and sensor fusion framework for robust high-quality video stabilization. We start with sensor-based pre-stabilization that smooths out large-scale camera motion. A new angular velocity domain optimization has been introduced to achieve frame rate invariance. We then feed the stabiliz...
Jiyang Yu; Tianhao Zhang; Fuhao Shi; Lei He; Chia-Kai Liang
Google; Google; Google; Google; Google
Poster
main
https://openaccess.thecvf.com/content/WACV2025/html/Yu_SensorFlow_Sensor_and_Image_Fused_Video_Stabilization_WACV_2025_paper.html
0
SensorFlow: Sensor and Image Fused Video Stabilization We present SensorFlow a novel image and sensor fusion framework for robust high-quality video stabilization. We start with sensor-based pre-stabilization that smooths out large-scale camera motion. A new angular velocity domain optimization has been introduced to a...
[ -0.05850336328148842, -0.01208463590592146, -0.023427315056324005, -0.015618209727108479, 0.0017157772090286016, 0.003160275984555483, -0.03129206597805023, 0.03943505138158798, 0.03928666189312935, 0.0327945314347744, 0.017500927671790123, -0.026691928505897522, 0.026729026809334755, 0.01...
wacv_2025_4d65e68cd2
4d65e68cd2
wacv
2,025
Separating Direct and Global Components from Novel Viewpoints
Separating an image of a scene illuminated by a light source into direct components such as specular reflection and diffuse reflection and global components such as inter-reflection and subsurface scattering is important as preprocessing for various computer vision and graphics applications. Conventional methods cannot...
Kengo Matsufuji; Lin Shi; Ryo Kawahara; Takahiro Okabe
Kyushu Institute of Technology*; Kyushu Institute of Technology†; Kyoto University‡; Okayama University§
Poster
main
https://openaccess.thecvf.com/content/WACV2025/html/Matsufuji_Separating_Direct_and_Global_Components_from_Novel_Viewpoints_WACV_2025_paper.html
0
Separating Direct and Global Components from Novel Viewpoints Separating an image of a scene illuminated by a light source into direct components such as specular reflection and diffuse reflection and global components such as inter-reflection and subsurface scattering is important as preprocessing for various computer...
[ -0.011237566359341145, 0.0026204087771475315, 0.00043673478648997843, 0.01404926273971796, -0.007047678343951702, -0.029094142839312553, 0.02568323351442814, 0.02496417611837387, 0.028670083731412888, 0.024871990084648132, -0.01601284183561802, -0.0028462663758546114, -0.010269605554640293, ...
wacv_2025_603d0c0193
603d0c0193
wacv
2,025
Shadow Removal Refinement via Material-Consistent Shadow Edges
Shadow boundaries can be confused with material boundaries as both exhibit sharp changes in luminance or contrast within a scene. However shadows do not modify the intrinsic color or texture of surfaces. Therefore on both sides of shadow edges traversing regions with the same material the original color and texture sho...
Shilin Hu; Hieu Le; ShahRukh Athar; Sagnik Das; Dimitris Samaras
;;;;
Poster
main
https://github.com/cvlab-stonybrook/ShadowRemovalRefine
https://openaccess.thecvf.com/content/WACV2025/html/Hu_Shadow_Removal_Refinement_via_Material-Consistent_Shadow_Edges_WACV_2025_paper.html
0
2409.06848
Shadow Removal Refinement via Material-Consistent Shadow Edges Shadow boundaries can be confused with material boundaries as both exhibit sharp changes in luminance or contrast within a scene. However shadows do not modify the intrinsic color or texture of surfaces. Therefore on both sides of shadow edges traversing re...
[ -0.037576824426651, -0.002596191829070449, -0.05897294729948044, -0.024621205404400826, -0.016484782099723816, -0.020198233425617218, 0.010670416988432407, 0.053075652569532394, 0.014641876332461834, 0.031992822885513306, -0.015157890506088734, -0.01490909792482853, -0.025984954088926315, ...
wacv_2025_572d3a048a
572d3a048a
wacv
2,025
Shape-Biased Texture Agnostic Representations for Improved Textureless and Metallic Object Detection and 6D Pose Estimation
Recent advances in machine learning have greatly benefited object detection and 6D pose estimation. However textureless and metallic objects still pose a significant challenge due to few visual cues and the texture bias of CNNs. To address this issue we propose a strategy for inducing a shape bias to CNN training. In p...
Peter Hönig; Stefan Thalhammer; Jean-Baptiste Weibel; Matthias Hirschmanner; Markus Vincze
TU Wien, Austria; UAS Technikum Vienna, Austria; TU Wien, Austria; TU Wien, Austria; TU Wien, Austria
Poster
main
https://github.com/hoenigpeter/randomized_texturing
https://openaccess.thecvf.com/content/WACV2025/html/Honig_Shape-Biased_Texture_Agnostic_Representations_for_Improved_Textureless_and_Metallic_Object_WACV_2025_paper.html
0
Shape-Biased Texture Agnostic Representations for Improved Textureless and Metallic Object Detection and 6D Pose Estimation Recent advances in machine learning have greatly benefited object detection and 6D pose estimation. However textureless and metallic objects still pose a significant challenge due to few visual cu...
[ -0.05114622786641121, -0.022457139566540718, -0.02738001011312008, 0.009725895710289478, 0.006333355791866779, -0.004881385713815689, 0.004667046945542097, 0.04546741023659706, 0.018880221992731094, 0.037188876420259476, -0.004487279336899519, -0.049191828817129135, 0.004968964494764805, 0...
wacv_2025_6207511fff
6207511fff
wacv
2,025
ShapeMorph: 3D Shape Completion via Blockwise Discrete Diffusion
We introduce ShapeMorph a diffusion-based method specifically designed for generating precise and diverse 3D shape completions. By integrating an irregular discrete representation with a novel blockwise discrete diffusion model ShapeMorph can produce multiple high-quality shape completions while maintaining fidelity to...
Jiahui Li; Pourya Shamsolmoali; Yue Lu; Masoumeh Zareapoor
East China Normal University; East China Normal University; East China Normal University; Shanghai Jiao Tong University
Poster
main
https://openaccess.thecvf.com/content/WACV2025/html/Li_ShapeMorph_3D_Shape_Completion_via_Blockwise_Discrete_Diffusion_WACV_2025_paper.html
0
ShapeMorph: 3D Shape Completion via Blockwise Discrete Diffusion We introduce ShapeMorph a diffusion-based method specifically designed for generating precise and diverse 3D shape completions. By integrating an irregular discrete representation with a novel blockwise discrete diffusion model ShapeMorph can produce mult...
[ -0.011980381794273853, -0.01907750591635704, 0.0020430423319339752, -0.005015361588448286, -0.01330141257494688, -0.012936990708112717, 0.024452734738588333, -0.008390822447836399, 0.024543840438127518, 0.013647614046931267, -0.02068096399307251, -0.011205984279513359, -0.01443112175911665, ...
wacv_2025_cf8554ac2c
cf8554ac2c
wacv
2,025
Shapley Consensus Deep Learning for Ensemble Pruning
This paper targets a new foundation for designing general-purpose learning systems by establishing a consensus method that facilitates self-adaptation and flexibility to deal with different learning tasks and different data distribution. We present the Shapely Consensus Deep Learning (SCDL) as a consensus method for ge...
Youcef Djenouri; Ahmed Nabil Belbachir; Asma Belhadi; Nassim Belmecheri; Tomasz Michalak
IDEAS NCBR, Warsaw, Poland + NORCE Norwegian Research Center, Oslo, Norway + University of South-Eastern Norway, Kongsberg, Norway; NORCE Norwegian Research Center, Grimstad, Norway; OsloMet University, Oslo, Norway; Simula Laboratory Research, Oslo, Norway; IDEAS NCBR, Warsaw, Poland
Poster
main
https://openaccess.thecvf.com/content/WACV2025/html/Djenouri_Shapley_Consensus_Deep_Learning_for_Ensemble_Pruning_WACV_2025_paper.html
0
Shapley Consensus Deep Learning for Ensemble Pruning This paper targets a new foundation for designing general-purpose learning systems by establishing a consensus method that facilitates self-adaptation and flexibility to deal with different learning tasks and different data distribution. We present the Shapely Consen...
[ -0.08170074969530106, -0.05249033868312836, -0.05285869166254997, -0.02488226257264614, -0.014282897114753723, -0.007661747746169567, 0.051495783030986786, -0.013242299668490887, 0.027829088270664215, 0.0017577356193214655, -0.017367856577038765, 0.0020316983573138714, -0.0017646423075348139...
wacv_2025_1e0ed818d6
1e0ed818d6
wacv
2,025
Shift Equivariant Pose Network
Human pose estimation has been greatly advanced in recent years. However even the best-performing models are not shift equivariant. In particular a small change in input images often results in drastic alterations in output which are problematic especially in video applications. The prevalence of top-down approaches wh...
Pengxiao Wang; Tzu-Heng Lin; Chunyu Wang; Yizhou Wang
School of Computer Science, Peking University; School of Computer Science, Peking University; Independent Researcher; School of Computer Science, Peking University + Center on Frontiers of Computing Studies, Peking University
Poster
main
https://openaccess.thecvf.com/content/WACV2025/html/Wang_Shift_Equivariant_Pose_Network_WACV_2025_paper.html
0
Shift Equivariant Pose Network Human pose estimation has been greatly advanced in recent years. However even the best-performing models are not shift equivariant. In particular a small change in input images often results in drastic alterations in output which are problematic especially in video applications. The preva...
[ -0.018321851268410683, -0.008099058642983437, -0.04743002727627754, 0.01882483996450901, 0.006278051529079676, 0.03057057037949562, 0.0039517260156571865, 0.03532102704048157, 0.029657738283276558, 0.03234034776687622, -0.01009704452008009, -0.008378497324883938, -0.005467679817229509, 0.0...
wacv_2025_ef9fbfa5a3
ef9fbfa5a3
wacv
2,025
Sifting through the Haystack - Efficiently Finding Rare Animal Behaviors in Large-Scale Datasets
In the study of animal behavior researchers often record long continuous videos accumulating into large-scale datasets. However the behaviors of interest are often rare compared to routine behaviors. This incurs a heavy cost on manual annotation forcing users to sift through many samples before finding their needles. W...
Shir Bar; Or Hirschorn; Roi Holzman; Shai Avidan
School of Zoology, the Faculty of Life Sciences, Tel Aviv University, Israel+The Interuniversity Institute for Marine Sciences in Eilat, Israel; School of Electrical Engineering, the Faculty of Engineering, Tel Aviv University, Israel; School of Zoology, the Faculty of Life Sciences, Tel Aviv University, Israel+The Int...
Poster
main
https://github.com/shir3bar/SiftingTheHaystack
https://openaccess.thecvf.com/content/WACV2025/html/Bar_Sifting_through_the_Haystack_-_Efficiently_Finding_Rare_Animal_Behaviors_WACV_2025_paper.html
0
Sifting through the Haystack - Efficiently Finding Rare Animal Behaviors in Large-Scale Datasets In the study of animal behavior researchers often record long continuous videos accumulating into large-scale datasets. However the behaviors of interest are often rare compared to routine behaviors. This incurs a heavy cos...
[ -0.12147102504968643, -0.044464390724897385, -0.0416150726377964, 0.038953207433223724, -0.008683865889906883, -0.009822656400501728, 0.01293441466987133, 0.011659718118607998, -0.030161553993821144, 0.02073255367577076, -0.03554151952266693, -0.01617738977074623, -0.0033320176880806684, 0...
wacv_2025_d7a13e67d6
d7a13e67d6
wacv
2,025
Sigma: Siamese Mamba Network for Multi-Modal Semantic Segmentation
Multi-modal semantic segmentation significantly enhances AI agents' perception and scene understanding especially under adverse conditions like low-light or overexposed environments. Leveraging additional modalities (X-modality) like thermal and depth alongside traditional RGB provides complementary information enablin...
Zifu Wan; Pingping Zhang; Yuhao Wang; Silong Yong; Simon Stepputtis; Katia Sycara; Yaqi Xie
Robotics Institute, Carnegie Mellon University, USA; School of Future Technology, Dalian University of Technology, China; School of Future Technology, Dalian University of Technology, China; Robotics Institute, Carnegie Mellon University, USA; Robotics Institute, Carnegie Mellon University, USA; Robotics Institute, Car...
Poster
main
https://github.com/zifuwan/Sigma
https://openaccess.thecvf.com/content/WACV2025/html/Wan_Sigma_Siamese_Mamba_Network_for_Multi-Modal_Semantic_Segmentation_WACV_2025_paper.html
53
2404.04256
Sigma: Siamese Mamba Network for Multi-Modal Semantic Segmentation Multi-modal semantic segmentation significantly enhances AI agents' perception and scene understanding especially under adverse conditions like low-light or overexposed environments. Leveraging additional modalities (X-modality) like thermal and depth a...
[ -0.03847500681877136, -0.01518701296299696, -0.004436917137354612, 0.02875382825732231, -0.021826092153787613, -0.057023461908102036, 0.027170877903699875, 0.02344628982245922, 0.034303467720746994, 0.037078287452459335, 0.006243342533707619, -0.043652184307575226, -0.010419538244605064, 0...
wacv_2025_90bd43dceb
90bd43dceb
wacv
2,025
Sign Language Recognition: A Large-Scale Multi-View Dataset and Comprehensive Evaluation
Vision-based sign language recognition is an extensively researched problem aimed at advancing communication between deaf and hearing individuals. Numerous Sign Language Recognition (SLR) datasets have been introduced to promote research in this field spanning multiple languages vocabulary sizes and signers. However mo...
Nguyen Son Dinh; Tuan Dung Nguyen; Duc Tri Tran; Nguyen Dang Huy Pham; Thuan Hieu Tran; Ngoc Anh Tong; Quang Huy Hoang; Phi Le Nguyen
Hanoi University of Science and Technology, Vietnam; Hanoi University of Science and Technology, Vietnam; Hanoi University of Science and Technology, Vietnam; Hanoi - Amsterdam High School for the Gifted, Vietnam; Hanoi - Amsterdam High School for the Gifted, Vietnam; Hanoi University of Science and Technology, Vietnam...
Poster
main
https://github.com/Etdihatthoc/Multi-VSL WACV 2025
https://openaccess.thecvf.com/content/WACV2025/html/Dinh_Sign_Language_Recognition_A_Large-Scale_Multi-View_Dataset_and_Comprehensive_Evaluation_WACV_2025_paper.html
0
Sign Language Recognition: A Large-Scale Multi-View Dataset and Comprehensive Evaluation Vision-based sign language recognition is an extensively researched problem aimed at advancing communication between deaf and hearing individuals. Numerous Sign Language Recognition (SLR) datasets have been introduced to promote re...
[ -0.05092298239469528, -0.00009757827501744032, -0.026331812143325806, -0.032072219997644424, -0.01672125607728958, 0.028646493330597878, -0.025424456223845482, 0.03166483715176582, 0.012702969834208488, -0.02314681001007557, 0.008203229866921902, 0.0031780570279806852, 0.004071523901075125, ...
wacv_2025_b83c162589
b83c162589
wacv
2,025
Similarity over Factuality: Are we Making Progress on Multimodal Out-of-Context Misinformation Detection?
Out-of-context (OOC) misinformation poses a significant challenge in multimodal fact-checking where images are paired with texts that misrepresent their original context to support false narratives. Recent research in evidence-based OOC detection has seen a trend towards increasingly complex architectures incorporating...
Stefanos-Iordanis Papadopoulos; Christos Koutlis; Symeon Papadopoulos; Panagiotis C. Petrantonakis
Centre for Research & Technology Hellas, Greece+Aristotle University of Thessaloniki, Greece; Centre for Research & Technology Hellas, Greece; Centre for Research & Technology Hellas, Greece; Aristotle University of Thessaloniki, Greece
Poster
main
https://github.com/stevejpapad/outcontext-misinfo-progress
https://openaccess.thecvf.com/content/WACV2025/html/Papadopoulos_Similarity_over_Factuality_Are_we_Making_Progress_on_Multimodal_Out-of-Context_WACV_2025_paper.html
7
2407.13488
Similarity over Factuality: Are we Making Progress on Multimodal Out-of-Context Misinformation Detection? Out-of-context (OOC) misinformation poses a significant challenge in multimodal fact-checking where images are paired with texts that misrepresent their original context to support false narratives. Recent research...
[ -0.0701751708984375, -0.02025502175092697, -0.04658468812704086, 0.03328010067343712, 0.00018080681911669672, 0.006978386081755161, -0.01498163491487503, 0.057988617569208145, 0.018028274178504944, 0.018410269170999527, -0.04949157312512398, -0.0687217265367508, -0.0023513620253652334, 0.0...
wacv_2025_37457179ad
37457179ad
wacv
2,025
SimuScope: Realistic Endoscopic Synthetic Dataset Generation through Surgical Simulation and Diffusion Models
Computer-assisted surgical (CAS) systems enhance surgical execution and outcomes by providing advanced support to surgeons. These systems often rely on deep learning models trained on complex challenging-to-annotate data. While synthetic data generation can address these challenges enhancing the realism of such data is...
Sabina Martyniak; Joanna Kaleta; Diego Dall'Alba; Michał Naskręt; Szymon Płotka; Przemysław Korzeniowski
;;;;;
Poster
main
https://openaccess.thecvf.com/content/WACV2025/html/Martyniak_SimuScope_Realistic_Endoscopic_Synthetic_Dataset_Generation_through_Surgical_Simulation_and_WACV_2025_paper.html
1
2412.02332
SimuScope: Realistic Endoscopic Synthetic Dataset Generation through Surgical Simulation and Diffusion Models Computer-assisted surgical (CAS) systems enhance surgical execution and outcomes by providing advanced support to surgeons. These systems often rely on deep learning models trained on complex challenging-to-ann...
[ -0.055798664689064026, -0.01446699257940054, -0.035414181649684906, 0.02083827741444111, -0.0029133653733879328, -0.05009899660944939, 0.015447190031409264, 0.06937621533870697, 0.020529696717858315, 0.06730691343545914, 0.019585803151130676, -0.012996695935726166, 0.006353133358061314, 0....
wacv_2025_042978796f
042978796f
wacv
2,025
Single-Layer Distillation with Fourier Convolutions for Texture Anomaly Detection
In industrial quality control detecting anomalies in visual textures is essential for ensuring product quality and operational efficiency. Early identification of defects prevents faulty items from reaching consumers reduces waste and maintains high standards of production. Numerous unsupervised anomaly detection metho...
Simon Thomine; Hichem Snoussi
University of Technology of Troyes; University of Technology of Troyes
Poster
main
https://openaccess.thecvf.com/content/WACV2025/html/Thomine_Single-Layer_Distillation_with_Fourier_Convolutions_for_Texture_Anomaly_Detection_WACV_2025_paper.html
0
Single-Layer Distillation with Fourier Convolutions for Texture Anomaly Detection In industrial quality control detecting anomalies in visual textures is essential for ensuring product quality and operational efficiency. Early identification of defects prevents faulty items from reaching consumers reduces waste and mai...
[ -0.04365621134638786, -0.025549013167619705, -0.019433461129665375, 0.016698040068149567, 0.00240155216306448, -0.005880693905055523, -0.05371371656656265, 0.04166681319475174, -0.003034751396626234, 0.04671398550271988, -0.015445456840097904, -0.018991371616721153, -0.03433551639318466, 0...
wacv_2025_bf76cb94c6
bf76cb94c6
wacv
2,025
Situational Scene Graph for Structured Human-Centric Situation Understanding
Graph based representation has been widely used in modelling spatio-temporal relationships in video understanding. Although effective existing graph-based approaches focus on capturing the human-object relationships while ignoring fine-grained semantic properties of the action components. These semantic properties are ...
Chinthani Sugandhika; Chen Li; Deepu Rajan; Basura Fernando
College of Computing and Data Science, Nanyang Technological University, Singapore + Centre for Frontier AI Research, Agency for Science, Technology and Research, Singapore + Institute of High-Performance Computing, Agency for Science, Technology and Research, Singapore; Centre for Frontier AI Research, Agency for Scie...
Poster
main
https://github.com/LUNAProject22/SSG
https://openaccess.thecvf.com/content/WACV2025/html/Sugandhika_Situational_Scene_Graph_for_Structured_Human-Centric_Situation_Understanding_WACV_2025_paper.html
1
2410.22829
Situational Scene Graph for Structured Human-Centric Situation Understanding Graph based representation has been widely used in modelling spatio-temporal relationships in video understanding. Although effective existing graph-based approaches focus on capturing the human-object relationships while ignoring fine-grained...
[ -0.04629925638437271, -0.018022840842604637, -0.03697613254189491, -0.013165885582566261, 0.008322887122631073, -0.05221692472696304, 0.0033961471635848284, 0.017362220212817192, 0.0030635108705610037, 0.009499907493591309, 0.002419173251837492, 0.015575754456222057, 0.005564098712056875, ...
wacv_2025_9a31b8fdcd
9a31b8fdcd
wacv
2,025
Skip-and-Play: Depth-Driven Pose-Preserved Image Generation for Any Objects
The emergence of diffusion models has enabled the generation of diverse high-quality images solely from text prompting subsequent efforts to enhance the controllability of these models. Despite the improvement in controllability pose control remains limited to specific objects (e.g. humans) or poses (e.g. frontal view)...
Kyungmin Jo; Jaegul Choo
KAIST; KAIST
Poster
main
https://openaccess.thecvf.com/content/WACV2025/html/Jo_Skip-and-Play_Depth-Driven_Pose-Preserved_Image_Generation_for_Any_Objects_WACV_2025_paper.html
1
Skip-and-Play: Depth-Driven Pose-Preserved Image Generation for Any Objects The emergence of diffusion models has enabled the generation of diverse high-quality images solely from text prompting subsequent efforts to enhance the controllability of these models. Despite the improvement in controllability pose control re...
[ -0.04198809713125229, -0.023449817672371864, -0.04870914667844772, 0.015242381021380424, -0.020938655361533165, -0.0034805438481271267, 0.014457643032073975, 0.025905584916472435, -0.0068226042203605175, -0.011346387676894665, -0.012297306209802628, 0.025019293650984764, -0.01962768100202083...
wacv_2025_e12eb94e39
e12eb94e39
wacv
2,025
Skyeyes: Ground Roaming using Aerial View Images
Integrating aerial imagery-based scene generation into applications like autonomous driving and gaming enhances realism in 3D environments but challenges remain in creating detailed content for occluded areas and ensuring real-time consistent rendering. In this paper we introduce Skyeyes a novel framework that can gene...
Zhiyuan Gao; Wenbin Teng; Gonglin Chen; Jinsen Wu; Ningli Xu; Rongjun Qin; Andrew Feng; Yajie Zhao
University of Southern California+Institute for Creative Technologies; University of Southern California+Institute for Creative Technologies; University of Southern California+Institute for Creative Technologies; University of Southern California+Institute for Creative Technologies; The Ohio State University; The Ohio ...
Poster
main
https://openaccess.thecvf.com/content/WACV2025/html/Gao_Skyeyes_Ground_Roaming_using_Aerial_View_Images_WACV_2025_paper.html
2
2409.16685
Skyeyes: Ground Roaming using Aerial View Images Integrating aerial imagery-based scene generation into applications like autonomous driving and gaming enhances realism in 3D environments but challenges remain in creating detailed content for occluded areas and ensuring real-time consistent rendering. In this paper we ...
[ -0.03816966712474823, 0.0043013859540224075, -0.019322024658322334, -0.00832452718168497, -0.01904834248125553, -0.009245926514267921, 0.0011055655777454376, 0.04878856986761093, 0.006322079338133335, 0.03309740871191025, -0.0007589003653265536, -0.008288036100566387, 0.017990101128816605, ...
wacv_2025_54b7a5333e
54b7a5333e
wacv
2,025
Sli2Vol+: Segmenting 3D Medical Images Based on an Object Estimation Guided Correspondence Flow Network
Deep learning (DL) methods have shown remarkable successes in medical image segmentation often using large amounts of annotated data for model training. However acquiring a large number of diverse labeled 3D medical image datasets is highly difficult and expensive. Recently mask propagation DL methods were developed to...
Delin An; Pengfei Gu; Milan Sonka; Chaoli Wang; Danny Z. Chen
University of Notre Dame; The University of Texas Rio Grande Valley; University of Iowa; University of Notre Dame; University of Notre Dame
Poster
main
https://github.com/adlsn/Sli2Volplus
https://openaccess.thecvf.com/content/WACV2025/html/An_Sli2Vol_Segmenting_3D_Medical_Images_Based_on_an_Object_Estimation_WACV_2025_paper.html
2
Sli2Vol+: Segmenting 3D Medical Images Based on an Object Estimation Guided Correspondence Flow Network Deep learning (DL) methods have shown remarkable successes in medical image segmentation often using large amounts of annotated data for model training. However acquiring a large number of diverse labeled 3D medical ...
[ -0.07684939354658127, -0.0231916643679142, -0.023119639605283737, 0.01656547375023365, 0.01372053474187851, 0.017303718253970146, 0.02481219917535782, -0.03401324152946472, -0.007206881884485483, 0.001878245733678341, 0.0062840767204761505, 0.012775221839547157, -0.010542484000325203, 0.07...
wacv_2025_10aa0019ed
10aa0019ed
wacv
2,025
SmartKC++: Improving Performance of Smartphone-Based Corneal Topographers
Keratoconus an ocular condition marked by progressive corneal thinning and outward bulging presents diagnostic challenges due to the high cost and lack of portability in conventional corneal topographers. These limitations restrict accessibility for many necessitating affordable and mobile alternatives. Innovations lik...
Vaibhav Ganatra; Siddhartha Gairola; Pallavi Joshi; Anand Balasubramaniam; Kaushik Murali; Arivunithi Varadharajan; Bellamkonda Mallikarjuna; Nipun Kwatra; Mohit Jain
Microsoft Research, India; Max Planck Institute of Informatics, Saarbr ¨ucken, Germany; Sankara Eye Hospital, India; Remidio Innovative Solutions Private Limited, Bengaluru, India; Sankara Eye Hospital, India; Remidio Innovative Solutions Private Limited, Bengaluru, India; Remidio Innovative Solutions Private Limited, ...
Poster
main
https://github.com/microsoft/SmartKC-A-Smartphone-based-Corneal-Topographer
https://openaccess.thecvf.com/content/WACV2025/html/Ganatra_SmartKC_Improving_Performance_of_Smartphone-Based_Corneal_Topographers_WACV_2025_paper.html
0
SmartKC++: Improving Performance of Smartphone-Based Corneal Topographers Keratoconus an ocular condition marked by progressive corneal thinning and outward bulging presents diagnostic challenges due to the high cost and lack of portability in conventional corneal topographers. These limitations restrict accessibility ...
[ -0.051368411630392075, 0.021036408841609955, -0.016201956197619438, -0.03507312759757042, 0.00611306494101882, -0.021204402670264244, 0.0074570057913661, 0.016453944146633148, 0.007918985560536385, 0.03546511009335518, 0.001796587835997343, -0.01024754997342825, -0.01431670505553484, 0.020...
wacv_2025_7c0c82380e
7c0c82380e
wacv
2,025
Social EgoMesh Estimation
Accurately estimating the 3D pose of the camera wearer in egocentric video sequences is crucial to modeling human behavior in virtual and augmented reality applications. The task presents unique challenges due to the limited visibility of the user's body caused by the front-facing camera mounted on their head. Recent r...
Luca Scofano; Alessio Sampieri; Edoardo De Matteis; Indro Spinelli; Fabio Galasso
Sapienza University of Rome, Italy; Sapienza University of Rome, Italy; Sapienza University of Rome, Italy; Sapienza University of Rome, Italy; Sapienza University of Rome, Italy
Poster
main
https://openaccess.thecvf.com/content/WACV2025/html/Scofano_Social_EgoMesh_Estimation_WACV_2025_paper.html
0
2411.04598
Social EgoMesh Estimation Accurately estimating the 3D pose of the camera wearer in egocentric video sequences is crucial to modeling human behavior in virtual and augmented reality applications. The task presents unique challenges due to the limited visibility of the user's body caused by the front-facing camera mount...
[ -0.040659014135599136, -0.021831326186656952, -0.022930219769477844, 0.0181408803910017, -0.015320389531552792, 0.013314911164343357, 0.015302075073122978, 0.020036470144987106, 0.0037682848051190376, 0.03642828017473221, -0.006140060722827911, -0.008374474942684174, -0.0007589224842377007, ...
wacv_2025_a0668d0fe8
a0668d0fe8
wacv
2,025
Socially-Informed Reconstruction for Pedestrian Trajectory Forecasting
Pedestrian trajectory prediction remains a challenge for autonomous systems particularly due to the intricate dynamics of social interactions. Accurate forecasting requires a comprehensive understanding not only of each pedestrian's previous trajectory but also of their interaction with the surrounding environment an i...
Haleh Damirchi; Ali Etemad; Michael Greenspan
Dept. ECE & Ingenuity Labs Research Institute, Queen’s University, Kingston, Canada; Dept. ECE & Ingenuity Labs Research Institute, Queen’s University, Kingston, Canada; Dept. ECE & Ingenuity Labs Research Institute, Queen’s University, Kingston, Canada
Poster
main
https://openaccess.thecvf.com/content/WACV2025/html/Damirchi_Socially-Informed_Reconstruction_for_Pedestrian_Trajectory_Forecasting_WACV_2025_paper.html
1
2412.04673
Socially-Informed Reconstruction for Pedestrian Trajectory Forecasting Pedestrian trajectory prediction remains a challenge for autonomous systems particularly due to the intricate dynamics of social interactions. Accurate forecasting requires a comprehensive understanding not only of each pedestrian's previous traject...
[ -0.050873786211013794, -0.004010405857115984, -0.04104546457529068, 0.028430575504899025, -0.00946587510406971, -0.03264809027314186, -0.003824477316811681, -0.009522359818220139, 0.005714357364922762, 0.04906627908349037, -0.048501431941986084, 0.004951815586537123, 0.009122260846197605, ...
wacv_2025_32c91aa43c
32c91aa43c
wacv
2,025
Solar Multimodal Transformer: Intraday Solar Irradiance Predictor using Public Cameras and Time Series
Accurate intraday solar irradiance forecasting is crucial for optimizing dispatch planning and electricity trading. For this purpose we introduce a novel and effective approach that includes three distinguishing components from the literature: 1) the uncommon use of single-frame public camera imagery; 2) solar irradian...
Yanan Niu; Roy Sarkis; Demetri Psaltis; Mario Paolone; Christophe Moser; Luisa Lambertini
EPFL, Lausanne, 1015, Vaud, Switzerland; EPFL, Lausanne, 1015, Vaud, Switzerland; EPFL, Lausanne, 1015, Vaud, Switzerland; EPFL, Lausanne, 1015, Vaud, Switzerland; EPFL, Lausanne, 1015, Vaud, Switzerland; EPFL, Lausanne, 1015, Vaud, Switzerland + Universita’ della Svizzera Italiana (USI), Lugano, 6900, Ticino, Switzerl...
Poster
main
https://openaccess.thecvf.com/content/WACV2025/html/Niu_Solar_Multimodal_Transformer_Intraday_Solar_Irradiance_Predictor_using_Public_Cameras_WACV_2025_paper.html
0
Solar Multimodal Transformer: Intraday Solar Irradiance Predictor using Public Cameras and Time Series Accurate intraday solar irradiance forecasting is crucial for optimizing dispatch planning and electricity trading. For this purpose we introduce a novel and effective approach that includes three distinguishing compo...
[ -0.06814306229352951, 0.009441949427127838, -0.05346778407692909, 0.0037443002220243216, -0.026331327855587006, -0.0006450169021263719, -0.01453804038465023, 0.027813494205474854, 0.02369636483490467, 0.027776896953582764, -0.03196721896529198, 0.02104310318827629, 0.0206405408680439, -0.0...
wacv_2025_cd386fb2e1
cd386fb2e1
wacv
2,025
SoundLoc3D: Invisible 3D Sound Source Localization and Classification using a Multimodal RGB-D Acoustic Camera
Accurately localizing 3D sound sources and estimating their semantic labels - where the sources may not be visible but are assumed to lie on the physical surface of objects in the scene - have many real applications including detecting gas leak and machinery malfunction. The audio-visual weak- correlation in such setti...
Yuhang He; Sangyun Shin; Anoop Cherian; Niki Trigoni; Andrew Markham
Department of Computer Science, University of Oxford, UK; Department of Computer Science, University of Oxford, UK; Mitsubishi Electric Research Labs, Cambridge, MA, US; Department of Computer Science, University of Oxford, UK; Department of Computer Science, University of Oxford, UK
Poster
main
https://openaccess.thecvf.com/content/WACV2025/html/He_SoundLoc3D_Invisible_3D_Sound_Source_Localization_and_Classification_using_a_WACV_2025_paper.html
0
SoundLoc3D: Invisible 3D Sound Source Localization and Classification using a Multimodal RGB-D Acoustic Camera Accurately localizing 3D sound sources and estimating their semantic labels - where the sources may not be visible but are assumed to lie on the physical surface of objects in the scene - have many real applic...
[ -0.05033703148365021, 0.009617968462407589, -0.03338423743844032, 0.0023325819056481123, -0.006719094701111317, 0.02475503273308277, 0.016440434381365776, 0.020817957818508148, 0.009456170722842216, -0.007869655266404152, -0.017878634855151176, -0.009096620604395866, 0.03431906923651695, -...
wacv_2025_7244f8f0ac
7244f8f0ac
wacv
2,025
SoundSil-DS: Deep Denoising and Segmentation of Sound-Field Images with Silhouettes
Development of optical technology has enabled imaging of two-dimensional (2D) sound fields. This acousto-optic sensing enables understanding of the interaction between sound and objects such as reflection and diffraction. Moreover it is expected to be used an advanced measurement technology for sonars in self-driving v...
Risako Tanigawa; Kenji Ishikawa; Noboru Harada; Yasuhiro Oikawa
NTT Communication Science Laboratories, NTT Corporation, Atsugi, Kanagawa 243-0198, Japan+Intermedia Art and Science, Waseda University, Shinjuku, Tokyo 169-8555, Japan; NTT Communication Science Laboratories, NTT Corporation, Atsugi, Kanagawa 243-0198, Japan; NTT Communication Science Laboratories, NTT Corporation, At...
Poster
main
https://github.com/nttcslab/soundsil-ds
https://openaccess.thecvf.com/content/WACV2025/html/Tanigawa_SoundSil-DS_Deep_Denoising_and_Segmentation_of_Sound-Field_Images_with_Silhouettes_WACV_2025_paper.html
0
SoundSil-DS: Deep Denoising and Segmentation of Sound-Field Images with Silhouettes Development of optical technology has enabled imaging of two-dimensional (2D) sound fields. This acousto-optic sensing enables understanding of the interaction between sound and objects such as reflection and diffraction. Moreover it is...
[ -0.0532388873398304, -0.005886636208742857, -0.031358372420072556, 0.008325518108904362, -0.013596649281680584, 0.01365218311548233, 0.012819169089198112, 0.019233381375670433, 0.006867742165923119, 0.04265034571290016, -0.001030276995152235, 0.0048361122608184814, 0.012782146222889423, 0....
wacv_2025_df327a6b88
df327a6b88
wacv
2,025
SpaGBOL: Spatial-Graph-Based Orientated Localisation
Cross-View Geo-Localisation within urban regions is challenging in part due to the lack of geo-spatial structuring within current datasets and techniques. We propose utilising graph representations to model sequences of local observations and the connectivity of the target location. Modelling as a graph enables generat...
Tavis Shore; Oscar Mendez; Simon Hadfield
University of Surrey1; Locus Robotics2; University of Surrey1
Poster
main
github.com/tavisshore/SpaGBOL
https://openaccess.thecvf.com/content/WACV2025/html/Shore_SpaGBOL_Spatial-Graph-Based_Orientated_Localisation_WACV_2025_paper.html
1
2409.15514
SpaGBOL: Spatial-Graph-Based Orientated Localisation Cross-View Geo-Localisation within urban regions is challenging in part due to the lack of geo-spatial structuring within current datasets and techniques. We propose utilising graph representations to model sequences of local observations and the connectivity of the ...
[ -0.08208499848842621, -0.020576022565364838, -0.013939479365944862, -0.011219135485589504, -0.0005939340917393565, -0.025176871567964554, 0.0393628254532814, -0.01771874912083149, 0.008786344900727272, 0.03061755932867527, -0.023716285824775696, 0.030106354504823685, 0.0019387013744562864, ...
wacv_2025_8da7ef54bb
8da7ef54bb
wacv
2,025
Sparse-View 3D Reconstruction of Clothed Humans via Normal Maps
We present a novel deep learning-based approach to the 3D reconstruction of clothed humans using weak supervision via 2D normal maps. Given a single RGB image or multiview images our network is optimized to infer a person-specific signed distance function (SDF) discretized on a tetrahedral mesh surrounding the body in ...
Jane Wu; Diego Thomas; Ronald Fedkiw
Stanford University; Kyushu University; Stanford University + Epic Games
Poster
main
https://openaccess.thecvf.com/content/WACV2025/html/Wu_Sparse-View_3D_Reconstruction_of_Clothed_Humans_via_Normal_Maps_WACV_2025_paper.html
0
Sparse-View 3D Reconstruction of Clothed Humans via Normal Maps We present a novel deep learning-based approach to the 3D reconstruction of clothed humans using weak supervision via 2D normal maps. Given a single RGB image or multiview images our network is optimized to infer a person-specific signed distance function ...
[ -0.00788346491754055, -0.0211031436920166, -0.051874689757823944, 0.0022346556652337313, -0.010570164769887924, 0.013926215469837189, 0.06671196222305298, -0.010746799409389496, 0.005550033878535032, 0.024747386574745178, -0.03646102547645569, 0.01843503676354885, -0.006465742830187082, 0....
wacv_2025_0a592458db
0a592458db
wacv
2,025
Spatially-Adaptive Hash Encodings for Neural Surface Reconstruction
Positional encodings are a common component of neural scene reconstruction methods and provide a way to bias the learning of neural fields towards coarser or finer representations. Current neural surface reconstruction methods use a "one-size-fits-all" approach to encoding choosing a fixed set of encoding functions and...
Thomas Walker; Octave Mariotti; Amir Vaxman; Hakan Bilen
;;;
Poster
main
https://openaccess.thecvf.com/content/WACV2025/html/Walker_Spatially-Adaptive_Hash_Encodings_for_Neural_Surface_Reconstruction_WACV_2025_paper.html
0
2412.05179
Spatially-Adaptive Hash Encodings for Neural Surface Reconstruction Positional encodings are a common component of neural scene reconstruction methods and provide a way to bias the learning of neural fields towards coarser or finer representations. Current neural surface reconstruction methods use a "one-size-fits-all"...
[ -0.01584506407380104, -0.009939778596162796, -0.020578742027282715, -0.012216856703162193, -0.0008137485128827393, -0.022355053573846817, 0.03206807002425194, -0.000508740427903831, 0.02819420024752617, 0.014295517466962337, -0.03941896930336952, 0.017082812264561653, 0.012292444705963135, ...
wacv_2025_8d270023e1
8d270023e1
wacv
2,025
Spatio-Temporal Context Prompting for Zero-Shot Action Detection
Spatio-temporal action detection encompasses the tasks of localizing and classifying individual actions within a video. Recent works aim to enhance this process by incorporating interaction modeling which captures the relationship between people and their surrounding context. However these approaches have primarily foc...
Wei-Jhe Huang; Min-Hung Chen; Shang-Hong Lai
;;
Poster
main
https://openaccess.thecvf.com/content/WACV2025/html/Huang_Spatio-Temporal_Context_Prompting_for_Zero-Shot_Action_Detection_WACV_2025_paper.html
0
2408.15996
Spatio-Temporal Context Prompting for Zero-Shot Action Detection Spatio-temporal action detection encompasses the tasks of localizing and classifying individual actions within a video. Recent works aim to enhance this process by incorporating interaction modeling which captures the relationship between people and their...
[ -0.03569302707910538, -0.024030771106481552, -0.019485387951135635, 0.03685200959444046, 0.002954046241939068, 0.0002310324489371851, 0.023034770041704178, 0.014541604556143284, -0.011571712791919708, 0.02069869637489319, -0.0020610415376722813, -0.022962333634495735, -0.013391676358878613, ...
wacv_2025_78caa0dee7
78caa0dee7
wacv
2,025
SpectFormer: Frequency and Attention is What You Need in a Vision Transformer
Vision transformers have been applied successfully for image recognition tasks. There have been either multiheaded self-attention based (ViT [12] DeIT [54]) similar to the original work in textual models or more recently based on spectral layers (Fnet [29] GFNet [46] AFNO [15]). We hypothesize that spectral layers capt...
Badri N. Patro; Vinay P. Namboodiri; Vijay S. Agneeswaran
Microsoft; University of Bath; Microsoft
Poster
main
https://github.com/badripatro/SpectFormers
https://openaccess.thecvf.com/content/WACV2025/html/Patro_SpectFormer_Frequency_and_Attention_is_What_You_Need_in_a_WACV_2025_paper.html
86
2304.06446
SpectFormer: Frequency and Attention is What You Need in a Vision Transformer Vision transformers have been applied successfully for image recognition tasks. There have been either multiheaded self-attention based (ViT [12] DeIT [54]) similar to the original work in textual models or more recently based on spectral lay...
[ -0.01067472342401743, -0.018073316663503647, -0.05194486305117607, 0.0024661978241056204, -0.004213467240333557, -0.012522096745669842, 0.022914709523320198, 0.009018458425998688, 0.011211644858121872, 0.0010761177400127053, -0.01819162257015705, -0.011630261316895485, -0.011948774568736553,...
wacv_2025_d4ae3043a8
d4ae3043a8
wacv
2,025
SpiralMLP: A Lightweight Vision MLP Architecture
We present SpiralMLP a novel architecture introduces a Spiral FC layer as a replacement for the conventional Token Mixing approach. Differing from several existing MLP-based models that primarily emphasize axes our Spiral FC layer is designed as a deformable convolution layer with spiral-like offsets. We further adapt ...
Haojie Mu; Burhan Ul Tayyab; Nicholas Chua
Kookree; Kookree; Kookree
Poster
main
https://github.com/Kookree/SpiralMLP
https://openaccess.thecvf.com/content/WACV2025/html/Mu_SpiralMLP_A_Lightweight_Vision_MLP_Architecture_WACV_2025_paper.html
0
2404.00648
SpiralMLP: A Lightweight Vision MLP Architecture We present SpiralMLP a novel architecture introduces a Spiral FC layer as a replacement for the conventional Token Mixing approach. Differing from several existing MLP-based models that primarily emphasize axes our Spiral FC layer is designed as a deformable convolution ...
[ -0.03372238948941231, 0.006372771691530943, -0.015962054952979088, 0.0048757498152554035, 0.01459944061934948, 0.00482476782053709, -0.02098611742258072, 0.000341812294209376, 0.026547808200120926, 0.012309878133237362, -0.014775561168789864, -0.005464361980557442, -0.020188940688967705, 0...
wacv_2025_a78d851076
a78d851076
wacv
2,025
Spk2ImgMamba: Spiking Camera Image Reconstruction with Multi-Scale State Space Models
As a bio-inspired vision sensor the spiking camera has showcased remarkable capability in high-speed imaging with a sampling rate of 40000 Hz. Reconstructing clear images from continuous spike streams which is obtained by each photosensor continuously detecting photons and firing them asynchronously has garnered signif...
Jiaoyang Yin; Bin Fan; Chao Xu; Tiejun Huang; Boxin Shi
State Key Lab of Multimedia Info. Processing, School of Computer Science, Peking University + Nat’l Eng. Research Ctr. of Visual Technology, School of Computer Science, Peking University; Nat’l Key Lab of General AI, School of Intelligence Science and Technology, Peking University; Nat’l Key Lab of General AI, School o...
Poster
main
https://github.com/interstellarH/Spk2ImgMamba
https://openaccess.thecvf.com/content/WACV2025/html/Yin_Spk2ImgMamba_Spiking_Camera_Image_Reconstruction_with_Multi-Scale_State_Space_Models_WACV_2025_paper.html
0
Spk2ImgMamba: Spiking Camera Image Reconstruction with Multi-Scale State Space Models As a bio-inspired vision sensor the spiking camera has showcased remarkable capability in high-speed imaging with a sampling rate of 40000 Hz. Reconstructing clear images from continuous spike streams which is obtained by each photose...
[ -0.036968398839235306, -0.00534287141636014, -0.03683842346072197, -0.010648607276380062, -0.005960249342024326, -0.006679749581962824, 0.03384901583194733, 0.027851631864905357, 0.02844580076634884, 0.029504161328077316, -0.02215133048593998, 0.013777273707091808, -0.03312487527728081, 0....
wacv_2025_0eecb7928f
0eecb7928f
wacv
2,025
SplatFace: Gaussian Splat Face Reconstruction Leveraging an Optimizable Surface
We present SplatFace a novel Gaussian splatting framework designed for 3D human face reconstruction without reliance on accurate pre-determined geometry. Our method is designed to simultaneously deliver both high-quality novel view rendering and accurate 3D mesh reconstructions. We incorporate a generic 3D Morphable Mo...
Jiahao Luo; Jing Liu; James Davis
University of California, Santa Cruz; ByteDance Inc.; University of California, Santa Cruz
Poster
main
https://openaccess.thecvf.com/content/WACV2025/html/Luo_SplatFace_Gaussian_Splat_Face_Reconstruction_Leveraging_an_Optimizable_Surface_WACV_2025_paper.html
4
2403.18784
SplatFace: Gaussian Splat Face Reconstruction Leveraging an Optimizable Surface We present SplatFace a novel Gaussian splatting framework designed for 3D human face reconstruction without reliance on accurate pre-determined geometry. Our method is designed to simultaneously deliver both high-quality novel view renderin...
[ -0.023526109755039215, 0.010877055115997791, -0.016862263903021812, -0.07216181606054306, -0.02111317589879036, 0.044940900057554245, 0.03464822843670845, -0.010009907186031342, 0.014157137833535671, 0.04267877712845802, -0.02271551452577114, 0.0254112146794796, 0.02982236072421074, 0.0280...
wacv_2025_9997a0cdf6
9997a0cdf6
wacv
2,025
SpotDiffusion: A Fast Approach for Seamless Panorama Generation Over Time
Generating high-resolution images with generative models has recently been made widely accessible by leveraging diffusion models pre-trained on large-scale datasets. Various techniques such as MultiDiffusion and SyncDiffusion have further pushed image generation beyond training resolutions i.e. from square images to pa...
Stanislav Frolov; Brian B. Moser; Andreas Dengel
German Research Center for Artificial Intelligence (DFKI), Germany; RPTU Kaiserslautern-Landau, Germany; German Research Center for Artificial Intelligence (DFKI), Germany
Poster
main
https://openaccess.thecvf.com/content/WACV2025/html/Frolov_SpotDiffusion_A_Fast_Approach_for_Seamless_Panorama_Generation_Over_Time_WACV_2025_paper.html
5
2407.15507
SpotDiffusion: A Fast Approach for Seamless Panorama Generation Over Time Generating high-resolution images with generative models has recently been made widely accessible by leveraging diffusion models pre-trained on large-scale datasets. Various techniques such as MultiDiffusion and SyncDiffusion have further pushed ...
[ -0.06536485999822617, -0.010165650397539139, -0.03079620935022831, 0.031993232667446136, -0.007073333952575922, -0.026388976722955704, -0.02450275421142578, 0.03979203850030899, 0.02881930209696293, 0.016576992347836494, -0.023432686924934387, -0.0060168681666255, -0.01322169415652752, 0.0...
wacv_2025_22f4e76c8a
22f4e76c8a
wacv
2,025
Stable Autofocus with Focal Consistency Loss
Autofocus aims to accurately position the camera lens to bring the desired region of interest into focus. Conventional works search for the sharpest frame within the lens movement. However sharpness measure in many real-world settings is ambiguous and may cause a focus hunting problem where the lens continuously moves ...
Sangwon Lee; Myungsub Choi; Nagyeong Lee; Hyong-Euk Lee
Samsung Advanced Institute of Technology (SAIT)+*; Samsung Advanced Institute of Technology (SAIT); Samsung Advanced Institute of Technology (SAIT); Samsung Advanced Institute of Technology (SAIT)
Poster
main
https://openaccess.thecvf.com/content/WACV2025/html/Lee_Stable_Autofocus_with_Focal_Consistency_Loss_WACV_2025_paper.html
0
Stable Autofocus with Focal Consistency Loss Autofocus aims to accurately position the camera lens to bring the desired region of interest into focus. Conventional works search for the sharpest frame within the lens movement. However sharpness measure in many real-world settings is ambiguous and may cause a focus hunti...
[ -0.06467696279287338, -0.04305648431181908, -0.051431652158498764, -0.017229972407221794, -0.005953933112323284, -0.002522696740925312, 0.023428335785865784, -0.002331303898245096, -0.0018343741539865732, 0.009804850444197655, -0.041765160858631134, -0.016704218462109566, -0.0059769926592707...
wacv_2025_68bf8f18ee
68bf8f18ee
wacv
2,025
Strategic Base Representation Learning via Feature Augmentations for Few-Shot Class Incremental Learning
Few-shot class incremental learning implies the model to learn new classes while retaining knowledge of previously learned classes with a small number of training instances. Existing frameworks typically freeze the parameters of the previously learned classes during the incorporation of new classes. However this approa...
Parinita Nema; Vinod K Kurmi
Indian Institute of Science Education and Research Bhopal, India; Indian Institute of Science Education and Research Bhopal, India
Poster
main
https://openaccess.thecvf.com/content/WACV2025/html/Nema_Strategic_Base_Representation_Learning_via_Feature_Augmentations_for_Few-Shot_Class_WACV_2025_paper.html
0
2501.09361
Strategic Base Representation Learning via Feature Augmentations for Few-Shot Class Incremental Learning Few-shot class incremental learning implies the model to learn new classes while retaining knowledge of previously learned classes with a small number of training instances. Existing frameworks typically freeze the ...
[ -0.07678106427192688, -0.052975259721279144, -0.011535528115928173, 0.01014869287610054, 0.0010297938715666533, -0.02792038582265377, 0.021693404763936996, 0.043423548340797424, -0.013638741336762905, 0.013693847693502903, -0.061608538031578064, -0.02397112175822258, -0.0634821429848671, 0...
wacv_2025_035a54d68b
035a54d68b
wacv
2,025
Stratified Domain Adaptation: A Progressive Self-Training Approach for Scene Text Recognition
Unsupervised domain adaptation (UDA) has become increasingly prevalent in scene text recognition (STR) especially where training and testing data reside in different domains. The efficacy of existing UDA approaches tends to degrade when there is a large gap between the source and target domains. To deal with this probl...
Kha Nhat Le; Hoang-Tuan Nguyen; Hung Tien Tran; Thanh Duc Ngo
University of Information Technology, VNU-HCM, Vietnam; University of Information Technology, VNU-HCM, Vietnam; University of Information Technology, VNU-HCM, Vietnam; Vietnam National University, Ho Chi Minh City, Vietnam
Poster
main
https://github.com/KhaLee2307/StrDA
https://openaccess.thecvf.com/content/WACV2025/html/Le_Stratified_Domain_Adaptation_A_Progressive_Self-Training_Approach_for_Scene_Text_WACV_2025_paper.html
0
2410.09913
Stratified Domain Adaptation: A Progressive Self-Training Approach for Scene Text Recognition Unsupervised domain adaptation (UDA) has become increasingly prevalent in scene text recognition (STR) especially where training and testing data reside in different domains. The efficacy of existing UDA approaches tends to de...
[ -0.06670917570590973, -0.016242705285549164, -0.02902681939303875, -0.01647810824215412, -0.0027727584820240736, -0.012231825850903988, -0.006134020630270243, 0.011380759067833424, 0.046790581196546555, 0.024807700887322426, -0.014450032263994217, -0.023087458685040474, -0.02428257279098034,...
wacv_2025_582ba68d70
582ba68d70
wacv
2,025
Street TryOn: Learning In-the-Wild Virtual Try-On from Unpaired Person Images
Most virtual try-on research is motivated to serve the fashion business by generating images to demonstrate garments on studio models at a lower cost. However virtual try-on should be a broader application that also allows customers to visualize garments on themselves using their own casual photos known as in-the-wild ...
Aiyu Cui; Jay Mahajan; Viraj Shah; Preeti Gomathinayagam; Chang Liu; Svetlana Lazebnik
University of Illinois Urbana-Champaign; University of Illinois Urbana-Champaign; University of Illinois Urbana-Champaign; University of Illinois Urbana-Champaign; University of Illinois Urbana-Champaign; University of Illinois Urbana-Champaign
Poster
main
https://openaccess.thecvf.com/content/WACV2025/html/Cui_Street_TryOn_Learning_In-the-Wild_Virtual_Try-On_from_Unpaired_Person_Images_WACV_2025_paper.html
17
2311.16094
Street TryOn: Learning In-the-Wild Virtual Try-On from Unpaired Person Images Most virtual try-on research is motivated to serve the fashion business by generating images to demonstrate garments on studio models at a lower cost. However virtual try-on should be a broader application that also allows customers to visual...
[ -0.05900442600250244, 0.008365876972675323, -0.03863205388188362, 0.016250738874077797, 0.0011023245751857758, 0.045649204403162, 0.02027805522084236, 0.0450078509747982, 0.013836235739290714, 0.03614209592342377, -0.028709953650832176, 0.00891291256994009, -0.007300099823623896, 0.0006519...
wacv_2025_5bb17936fd
5bb17936fd
wacv
2,025
Structure-Aware Human Body Reshaping with Adaptive Affinity-Graph Network
Given a source portrait the automatic human body reshaping task aims at editing it to an aesthetic body shape. As the technology has been widely used in media several methods have been proposed mainly focusing on generating optical flow to warp the body shape. However those previous works only consider the local transf...
Qiwen Deng; Yangcen Liu
University of Electronic Science and Technology of China; Georgia Institute of Technology
Poster
main
https://github.com/Randle-Github/AGGN
https://openaccess.thecvf.com/content/WACV2025/html/Deng_Structure-Aware_Human_Body_Reshaping_with_Adaptive_Affinity-Graph_Network_WACV_2025_paper.html
0
2404.13983
Structure-Aware Human Body Reshaping with Adaptive Affinity-Graph Network Given a source portrait the automatic human body reshaping task aims at editing it to an aesthetic body shape. As the technology has been widely used in media several methods have been proposed mainly focusing on generating optical flow to warp t...
[ -0.046503014862537384, -0.029105806723237038, -0.04002278298139572, -0.00808648020029068, -0.010916978120803833, -0.05935301631689072, 0.0046875812113285065, -0.03420530632138252, 0.017636535689234734, 0.017581306397914886, -0.037997711449861526, 0.00047721510054543614, -0.016835711896419525...
wacv_2025_d1e014bd26
d1e014bd26
wacv
2,025
Structured Human Assessment of Text-to-Image Generative Models
Following the great progress in text-conditioned image generation there is a dire need for establishing clear comparison benchmarks. Unfortunately assessing performance of such models is highly subjective and notoriously difficult. Current automatic assessment of generated images quality and their alignment to text are...
Ciprian A. Corneanu; Qianli Feng; Aleix M. Martinez
Amazon; Amazon; Amazon
Poster
main
https://openaccess.thecvf.com/content/WACV2025/html/Corneanu_Structured_Human_Assessment_of_Text-to-Image_Generative_Models_WACV_2025_paper.html
0
Structured Human Assessment of Text-to-Image Generative Models Following the great progress in text-conditioned image generation there is a dire need for establishing clear comparison benchmarks. Unfortunately assessing performance of such models is highly subjective and notoriously difficult. Current automatic assessm...
[ -0.040841925889253616, -0.013738235458731651, -0.01286538876593113, -0.014147381298244, 0.015574849210679531, -0.07557392865419388, 0.02316679246723652, 0.013710958883166313, -0.01428376417607069, 0.01623857580125332, -0.026367228478193283, -0.021439284086227417, -0.048952121287584305, 0.0...
wacv_2025_52fe3d17b9
52fe3d17b9
wacv
2,025
Style-Pro: Style-Guided Prompt Learning for Generalizable Vision-Language Models
Pre-trained Vision-language (VL) models such as CLIP have shown significant generalization ability to downstream tasks even with minimal fine-tuning. While prompt learning has emerged as an effective strategy to adapt pre-trained VL models for downstream tasks current approaches frequently encounter severe overfitting ...
Niloufar Alipour Talemi; Hossein Kashiani; Fatemeh Afghah
Clemson University; Clemson University; Clemson University
Poster
main
https://openaccess.thecvf.com/content/WACV2025/html/Talemi_Style-Pro_Style-Guided_Prompt_Learning_for_Generalizable_Vision-Language_Models_WACV_2025_paper.html
0
Style-Pro: Style-Guided Prompt Learning for Generalizable Vision-Language Models Pre-trained Vision-language (VL) models such as CLIP have shown significant generalization ability to downstream tasks even with minimal fine-tuning. While prompt learning has emerged as an effective strategy to adapt pre-trained VL models...
[ -0.07413720339536667, -0.04604502022266388, -0.020351389423012733, -0.00832680705934763, 0.004345112480223179, -0.001980626257136464, -0.029364148154854774, 0.02591167949140072, -0.007636313792318106, -0.005973678082227707, 0.00852214451879263, -0.018061857670545578, 0.003073150757700205, ...
wacv_2025_8703b1ca80
8703b1ca80
wacv
2,025
Sun Off Lights On: Photorealistic Monocular Nighttime Simulation for Robust Semantic Perception
Nighttime scenes are hard to semantically perceive with learned models and annotate for humans. Thus realistic synthetic nighttime data become all the more important for learning robust semantic perception at night thanks to their accurate and cheap semantic annotations. However existing data-driven or hand-crafted tec...
Konstantinos Tzevelekakis; Shutong Zhang; Luc Van Gool; Christos Sakaridis
ETH Zürich; University of Toronto; ETH Zürich+KU Leuven+INSAIT; ETH Zürich
Poster
main
https://github.com/ktzevel/SOLO
https://openaccess.thecvf.com/content/WACV2025/html/Tzevelekakis_Sun_Off_Lights_On_Photorealistic_Monocular_Nighttime_Simulation_for_Robust_WACV_2025_paper.html
0
2407.20336
Sun Off Lights On: Photorealistic Monocular Nighttime Simulation for Robust Semantic Perception Nighttime scenes are hard to semantically perceive with learned models and annotate for humans. Thus realistic synthetic nighttime data become all the more important for learning robust semantic perception at night thanks to...
[ -0.03847381845116615, 0.005870222579687834, -0.00896517839282751, -0.017689431086182594, -0.01611415483057499, -0.007135077379643917, 0.0227951817214489, 0.07324110716581345, -0.010100304149091244, -0.0027799000963568687, -0.01875505968928337, -0.016846194863319397, -0.0028540308121591806, ...
wacv_2025_f20a4e9082
f20a4e9082
wacv
2,025
Survival Prediction in Lung Cancer through Multi-Modal Representation Learning
Survival prediction is a crucial task associated with cancer diagnosis and treatment planning. This paper presents a novel approach to survival prediction by harnessing comprehensive information from CT and PET scans along with associated Genomic data. Current methods rely on either a single modality or the integration...
Aiman Farooq; Deepak Mishra; Santanu Chaudhury
Indian Institute of Technology Jodhpur; Indian Institute of Technology Jodhpur; Indian Institute of Technology Delhi
Poster
main
https://openaccess.thecvf.com/content/WACV2025/html/Farooq_Survival_Prediction_in_Lung_Cancer_through_Multi-Modal_Representation_Learning_WACV_2025_paper.html
3
2409.20179
Survival Prediction in Lung Cancer through Multi-Modal Representation Learning Survival prediction is a crucial task associated with cancer diagnosis and treatment planning. This paper presents a novel approach to survival prediction by harnessing comprehensive information from CT and PET scans along with associated Ge...
[ -0.04483319818973541, -0.023956097662448883, -0.036816541105508804, -0.002642492763698101, -0.010476099327206612, -0.05981514975428581, 0.013282868079841137, -0.01018509641289711, 0.027842393144965172, 0.016849998384714127, -0.014521976932883263, 0.02384345233440399, 0.008547033183276653, ...
wacv_2025_041a622e28
041a622e28
wacv
2,025
Swap Path Network for Robust Person Search Pre-Training
In person search we detect and rank matches to a query person image within a set of gallery scenes. Most person search models make use of a feature extraction backbone followed by separate heads for detection and re-identification. While pre-training methods for vision backbones are well-established pre-training additi...
Lucas Jaffe; Avideh Zakhor
Lawrence Livermore National Laboratory + University of California, Berkeley; University of California, Berkeley
Poster
main
https://openaccess.thecvf.com/content/WACV2025/html/Jaffe_Swap_Path_Network_for_Robust_Person_Search_Pre-Training_WACV_2025_paper.html
0
2412.05433
Swap Path Network for Robust Person Search Pre-Training In person search we detect and rank matches to a query person image within a set of gallery scenes. Most person search models make use of a feature extraction backbone followed by separate heads for detection and re-identification. While pre-training methods for v...
[ -0.02652723714709282, -0.04792376980185509, -0.06469151377677917, 0.010493781417608261, -0.012956892140209675, -0.02232600748538971, 0.051158346235752106, -0.007166258059442043, 0.018412914127111435, -0.004545136354863644, 0.002725687576457858, -0.005618680734187365, -0.024705465883016586, ...
wacv_2025_7f20f970ad
7f20f970ad
wacv
2,025
Swin-: Gradient-Based Image Restoration from Image Sequences using Video Swin-Transformers
Most deep-learning models for vision tasks rely on RGB images as their primary input layer assuming the model inherently discovers an optimal representation. In this work we challenge this assumption and show that image gradients offer a straightforward yet robust representation for multi-frame image restoration. We de...
Monika Kwiatkowski; Simon Matern; Olaf Hellwich
Technische Universit ¨at Berlin, Germany; Technische Universit ¨at Berlin, Germany; Technische Universit ¨at Berlin, Germany
Poster
main
https://openaccess.thecvf.com/content/WACV2025/html/Kwiatkowski_Swin-_Gradient-Based_Image_Restoration_from_Image_Sequences_using_Video_Swin-Transformers_WACV_2025_paper.html
0
Swin-: Gradient-Based Image Restoration from Image Sequences using Video Swin-Transformers Most deep-learning models for vision tasks rely on RGB images as their primary input layer assuming the model inherently discovers an optimal representation. In this work we challenge this assumption and show that image gradients...
[ -0.03670483082532883, -0.02847757376730442, -0.0301121287047863, 0.006965018808841705, -0.0068923719227313995, 0.018089070916175842, 0.004785612691193819, 0.010016187094151974, 0.01829793117940426, 0.014365918934345245, -0.02233891375362873, 0.00935328472405672, -0.04475047439336777, 0.040...
wacv_2025_2dc4172671
2dc4172671
wacv
2,025
SwinIA: Self-Supervised Blind-Spot Image Denoising without Convolutions
Self-supervised image denoising implies restoring the signal from a noisy image without access to the ground truth. State-of-the-art solutions for this task rely on predicting masked pixels with a fully-convolutional neural network. This most often requires multiple forward passes information about the noise model or i...
Mikhail Papkov; Pavel Chizhov; Leopold Parts
Institute of Computer Science, University of Tartu, Estonia; Institute of Computer Science, University of Tartu, Estonia + CAIRO, Technical University of Applied Sciences Würzburg-Schweinfurt, Germany; Institute of Computer Science, University of Tartu, Estonia
Poster
main
https://openaccess.thecvf.com/content/WACV2025/html/Papkov_SwinIA_Self-Supervised_Blind-Spot_Image_Denoising_without_Convolutions_WACV_2025_paper.html
0
2305.05651
SwinIA: Self-Supervised Blind-Spot Image Denoising without Convolutions Self-supervised image denoising implies restoring the signal from a noisy image without access to the ground truth. State-of-the-art solutions for this task rely on predicting masked pixels with a fully-convolutional neural network. This most often...
[ -0.03193967416882515, 0.0020908431615680456, -0.019754379987716675, 0.021922316402196884, -0.01473635621368885, -0.011185427196323872, -0.0006710555171594024, 0.03283675014972687, 0.032818060368299484, 0.036462437361478806, -0.016810845583677292, -0.002586104441434145, -0.006802832707762718,...
wacv_2025_14d9ad4f8d
14d9ad4f8d
wacv
2,025
SynDRA: Synthetic Dataset for Railway Applications
The use of deep learning techniques in railway environments faces significant obstacles especially for computer vision tasks. Such obstacles are mainly due to the inherent safety concerns required for installing the proper equipment on a train and the substantial effort required to precisely annotate large datasets esp...
Gianluca D'Amico; Federico Nesti; Giulio Rossolini; Mauro Marinoni; Salvatore Sabina; Giorgio Buttazzo
Scuola Superiore Sant’Anna, Italy; Scuola Superiore Sant’Anna, Italy; Scuola Superiore Sant’Anna, Italy; Scuola Superiore Sant’Anna, Italy; Scuola Superiore Sant’Anna, Italy; Scuola Superiore Sant’Anna, Italy
Poster
main
https://openaccess.thecvf.com/content/WACV2025/html/DAmico_SynDRA_Synthetic_Dataset_for_Railway_Applications_WACV_2025_paper.html
0
SynDRA: Synthetic Dataset for Railway Applications The use of deep learning techniques in railway environments faces significant obstacles especially for computer vision tasks. Such obstacles are mainly due to the inherent safety concerns required for installing the proper equipment on a train and the substantial effor...
[ -0.0012676562182605267, -0.026186589151620865, -0.01970512792468071, 0.0232438575476408, -0.019928626716136932, -0.003545712446793914, 0.01976100355386734, 0.05691764876246452, 0.02665221132338047, -0.003880960401147604, -0.002214033156633377, -0.008176324889063835, 0.007366142235696316, 0...
wacv_2025_160ea2fc8c
160ea2fc8c
wacv
2,025
SynDroneVision: A Synthetic Dataset for Image-Based Drone Detection
Developing robust drone detection systems is often constrained by the limited availability of large-scale annotated training data and the high costs associated with real-world data collection. However leveraging synthetic data generated via game engine-based simulations provides a promising and cost-effective solution ...
Tamara R. Lenhard; Andreas Weinmann; Kai Franke; Tobias Koch
Institute for the Protection of Terrestrial Infrastructures, German Aerospace Center (DLR), Germany+Working Group Algorithms for Computer Vision, Imaging and Data Analysis, University of Applied Sciences Darmstadt, Germany; Working Group Algorithms for Computer Vision, Imaging and Data Analysis, University of Applied S...
Poster
main
https://openaccess.thecvf.com/content/WACV2025/html/Lenhard_SynDroneVision_A_Synthetic_Dataset_for_Image-Based_Drone_Detection_WACV_2025_paper.html
4
2411.05633
SynDroneVision: A Synthetic Dataset for Image-Based Drone Detection Developing robust drone detection systems is often constrained by the limited availability of large-scale annotated training data and the high costs associated with real-world data collection. However leveraging synthetic data generated via game engine...
[ -0.03614509105682373, -0.009960021823644638, 0.0007631214684806764, -0.009324372746050358, -0.024328427389264107, -0.018118275329470634, 0.014770830981433392, 0.06233930587768555, 0.03202937915921211, 0.024163799360394478, -0.013188567943871021, -0.01775243505835533, -0.01988345943391323, ...
wacv_2025_798eb5d3f1
798eb5d3f1
wacv
2,025
SyncDiff: Diffusion-Based Talking Head Synthesis with Bottlenecked Temporal Visual Prior for Improved Synchronization
Talking head synthesis also known as speech-to-lip synthesis reconstructs the facial motions that align with the given audio tracks. The synthesized videos are evaluated on mainly two aspects lip-speech synchronization and image fidelity. Recent studies demonstrate that GAN-based and diffusion-based models achieve stat...
Xulin Fan; Heting Gao; Ziyi Chen; Peng Chang; Mei Han; Mark Hasegawa-Johnson
University of Illinois at Urbana-Champaign; University of Illinois at Urbana-Champaign; PAII Inc.; PAII Inc.; PAII Inc.; University of Illinois at Urbana-Champaign
Poster
main
https://openaccess.thecvf.com/content/WACV2025/html/Fan_SyncDiff_Diffusion-Based_Talking_Head_Synthesis_with_Bottlenecked_Temporal_Visual_Prior_WACV_2025_paper.html
0
SyncDiff: Diffusion-Based Talking Head Synthesis with Bottlenecked Temporal Visual Prior for Improved Synchronization Talking head synthesis also known as speech-to-lip synthesis reconstructs the facial motions that align with the given audio tracks. The synthesized videos are evaluated on mainly two aspects lip-speech...
[ -0.047249164432287216, -0.008531604893505573, 0.0011705361539497972, -0.007521462626755238, -0.013896278105676174, -0.00024215537996497005, -0.0358554907143116, 0.01636248081922531, 0.003742530709132552, 0.028866400942206383, -0.009964914061129093, -0.02411600388586521, -0.04240776225924492,...
wacv_2025_cf2d043abf
cf2d043abf
wacv
2,025
SyncViolinist: Music-Oriented Violin Motion Generation Based on Bowing and Fingering
Automatically generating realistic musical performance motion can greatly enhance digital media production often involving collaboration between professionals and musicians. However capturing the intricate body hand and finger movements required for accurate musical performances is challenging. Existing methods often f...
Hiroki Nishizawa; Keitaro Tanaka; Asuka Hirata; Shugo Yamaguchi; Qi Feng; Masatoshi Hamanaka; Shigeo Morishima
Waseda University; Waseda University; Waseda University; Waseda University; Waseda Research Institute for Science and Engineering+RIKEN; RIKEN; Waseda Research Institute for Science and Engineering
Poster
main
https://github.com/Kakanat/SyncViolinist
https://openaccess.thecvf.com/content/WACV2025/html/Nishizawa_SyncViolinist_Music-Oriented_Violin_Motion_Generation_Based_on_Bowing_and_Fingering_WACV_2025_paper.html
0
2412.08343
SyncViolinist: Music-Oriented Violin Motion Generation Based on Bowing and Fingering Automatically generating realistic musical performance motion can greatly enhance digital media production often involving collaboration between professionals and musicians. However capturing the intricate body hand and finger movement...
[ -0.046936385333538055, 0.009732932783663273, -0.05894336476922035, 0.060689836740493774, 0.0004980168887414038, -0.03984135016798973, -0.0322733111679554, 0.039186421781778336, -0.008636840619146824, 0.042788516730070114, 0.02223110944032669, 0.013034852221608162, 0.006954043637961149, 0.0...
wacv_2025_cf6d714771
cf6d714771
wacv
2,025
TACLE: Task and Class-Aware Exemplar-Free Semi-Supervised Class Incremental Learning
We propose a novel TACLE (TAsk and CLass-awarE) framework for the relatively unexplored and challenging problem of exemplar-free semi-supervised class incremental learning. In this scenario at each new task the model has to learn new classes from both (few) labeled and unlabeled data without access to exemplars from pr...
Jayateja Kalla; Rohit Kumar; Soma Biswas
Department of Electrical Engineering, Indian Institute of Science, Bangalore, India; Department of Electrical Engineering, Indian Institute of Science, Bangalore, India; Department of Electrical Engineering, Indian Institute of Science, Bangalore, India
Poster
main
https://github.com/rokmr/TACLE
https://openaccess.thecvf.com/content/WACV2025/html/Kalla_TACLE_Task_and_Class-Aware_Exemplar-Free_Semi-Supervised_Class_Incremental_Learning_WACV_2025_paper.html
0
2407.08041
TACLE: Task and Class-Aware Exemplar-Free Semi-Supervised Class Incremental Learning We propose a novel TACLE (TAsk and CLass-awarE) framework for the relatively unexplored and challenging problem of exemplar-free semi-supervised class incremental learning. In this scenario at each new task the model has to learn new c...
[ -0.056483831256628036, -0.013545151799917221, -0.00480752345174551, 0.02184407040476799, -0.0043025510385632515, 0.009761284105479717, 0.07183866202831268, 0.03374406322836876, -0.023635467514395714, 0.02440320886671543, -0.0016497300239279866, -0.038094598799943924, -0.03518814966082573, ...
wacv_2025_76eeb3627c
76eeb3627c
wacv
2,025
TAM-VT: Transformation-Aware Multi-Scale Video Transformer for Segmentation and Tracking
Video Object Segmentation (VOS) has emerged as an increasingly important problem with availability of larger datasets and more complex and realistic settings which involve long videos with global motion (e.g. in egocentric settings) depicting small objects undergoing both rigid and non-rigid (including state) deformati...
Raghav Goyal; Wan-Cyuan Fan; Mennatullah Siam; Leonid Sigal
UBC; UBC; UBC; UBC+Vector Institute for AI+CIFAR AI Chair
Poster
main
https://openaccess.thecvf.com/content/WACV2025/html/Goyal_TAM-VT_Transformation-Aware_Multi-Scale_Video_Transformer_for_Segmentation_and_Tracking_WACV_2025_paper.html
0
TAM-VT: Transformation-Aware Multi-Scale Video Transformer for Segmentation and Tracking Video Object Segmentation (VOS) has emerged as an increasingly important problem with availability of larger datasets and more complex and realistic settings which involve long videos with global motion (e.g. in egocentric settings...
[ -0.014924143441021442, -0.0387471579015255, -0.04089771583676338, 0.05828758701682091, -0.024064023047685623, 0.03437187895178795, -0.017779197543859482, 0.02061571180820465, 0.005186371970921755, 0.03594772145152092, 0.015359818004071712, -0.029792670160531998, 0.02986682765185833, 0.0190...
wacv_2025_20478599d6
20478599d6
wacv
2,025
TFM^2: Training-Free Mask Matching for Open-Vocabulary Semantic Segmentation
The potential of Open-Vocabulary Semantic Segmentation (OVSS) in few-shot scenarios is not fully explored due to the complexity of extending few-shot concepts to semantic segmentation tasks. To address this challenge we propose Training-Free Mask Matching (TFM^2) an efficient mask-based adapter method that enhances OVS...
Yaoxin Zhuo; Zachary Bessinger; Lichen Wang; Naji Khosravan; Baoxin Li; Sing Bing Kang
Arizona State University; Zillow Group; Zillow Group; Zillow Group; Arizona State University; Zillow Group
Poster
main
https://openaccess.thecvf.com/content/WACV2025/html/Zhuo_TFM2_Training-Free_Mask_Matching_for_Open-Vocabulary_Semantic_Segmentation_WACV_2025_paper.html
0
TFM^2: Training-Free Mask Matching for Open-Vocabulary Semantic Segmentation The potential of Open-Vocabulary Semantic Segmentation (OVSS) in few-shot scenarios is not fully explored due to the complexity of extending few-shot concepts to semantic segmentation tasks. To address this challenge we propose Training-Free M...
[ -0.0009502823231741786, -0.00221125571988523, -0.043373845517635345, 0.01890094392001629, 0.013629438355565071, -0.029717091470956802, 0.012764510698616505, 0.02013915590941906, 0.02223319187760353, 0.03186575323343277, -0.043264590203762054, -0.015586906112730503, 0.01078883372247219, 0.0...
wacv_2025_482fd4129c
482fd4129c
wacv
2,025
TLDR: Text Based Last-Layer Retraining for Debiasing Image Classifiers
An image classifier may depend on incidental features stemming from a strong correlation between the feature and the classification target in the training dataset. Recently Last Layer Retraining (LLR) with group-balanced datasets is shown to be efficient in mitigating the spurious correlation of classifiers. However th...
Juhyeon Park; Seokhyeon Jeong; Taesup Moon
IPAI, Seoul National University; ECE, Seoul National University; ASRI / INMC / AIIS, Seoul National University
Poster
main
https://github.com/beotborry/TLDR
https://openaccess.thecvf.com/content/WACV2025/html/Park_TLDR_Text_Based_Last-Layer_Retraining_for_Debiasing_Image_Classifiers_WACV_2025_paper.html
0
2311.18291
TLDR: Text Based Last-Layer Retraining for Debiasing Image Classifiers An image classifier may depend on incidental features stemming from a strong correlation between the feature and the classification target in the training dataset. Recently Last Layer Retraining (LLR) with group-balanced datasets is shown to be effi...
[ -0.04171478748321533, -0.01789219304919243, -0.006866782903671265, 0.0023111130576580763, -0.012631943449378014, 0.002462586387991905, -0.00033220817567780614, -0.03247951716184616, -0.017965633422136307, -0.016983354464173317, -0.029450053349137306, -0.0377122238278389, -0.03093724511563778...
wacv_2025_1281ef1eb4
1281ef1eb4
wacv
2,025
TORE: Token Recycling in Vision Transformers for Efficient Active Visual Exploration
Active Visual Exploration (AVE) optimizes the utilization of robotic resources in real-world scenarios by sequentially selecting the most informative observations. However modern methods require a high computational budget due to processing the same observations multiple times through the autoencoder transformers. As a...
Jan Olszewski; Dawid Damian Rymarczyk; Piotr Wojcik; Mateusz Pach; Bartosz Zielinski
University of Warsaw; Faculty of Mathematics and Computer Science, Jagiellonian University + Ardigen SA; Center for Molecular Medicine Cologne (CMMC), University of Cologne; Faculty of Mathematics and Computer Science, Jagiellonian University; IDEAS NCBR
Poster
main
https://openaccess.thecvf.com/content/WACV2025/html/Olszewski_TORE_Token_Recycling_in_Vision_Transformers_for_Efficient_Active_Visual_WACV_2025_paper.html
0
2311.15335
TORE: Token Recycling in Vision Transformers for Efficient Active Visual Exploration Active Visual Exploration (AVE) optimizes the utilization of robotic resources in real-world scenarios by sequentially selecting the most informative observations. However modern methods require a high computational budget due to proce...
[ 0.022528504952788353, -0.033615145832300186, 0.0008833791362121701, 0.06177110597491264, 0.0019478744361549616, -0.0034166674595326185, -0.00033740175422281027, 0.029296405613422394, 0.03490516170859337, 0.02574419230222702, -0.008824443444609642, -0.022397633641958237, -0.006239741574972868...
wacv_2025_7ab5b860f0
7ab5b860f0
wacv
2,025
TPD-STR: Text Polygon Detection with Split Transformers
Regressing text in natural scenes with polygonal representations is challenging due to shape prediction difficulties. To address this we introduce Text Polygon Detection with Split Transformers (TPD-STR) which directly regresses polygonal points. TPD-STR incorporates the Decoder Split (DS) architecture to separate poly...
Sangyeon Kim; Sangkuk Lee; Jeesoo Kim; Nojun Kwak
NAVER WEBTOON AI; NAVER WEBTOON AI + Seoul National University; NAVER WEBTOON AI; Seoul National University
Poster
main
https://openaccess.thecvf.com/content/WACV2025/html/Kim_TPD-STR_Text_Polygon_Detection_with_Split_Transformers_WACV_2025_paper.html
0
TPD-STR: Text Polygon Detection with Split Transformers Regressing text in natural scenes with polygonal representations is challenging due to shape prediction difficulties. To address this we introduce Text Polygon Detection with Split Transformers (TPD-STR) which directly regresses polygonal points. TPD-STR incorpora...
[ -0.06243792176246643, -0.042020536959171295, -0.010662618093192577, 0.027068784460425377, 0.003640668001025915, -0.005206248257309198, 0.014896168373525143, -0.017499292269349098, 0.0658099427819252, 0.04639304429292679, -0.0468377061188221, 0.009361056610941887, -0.01547052338719368, 0.02...
wacv_2025_0cebaa9557
0cebaa9557
wacv
2,025
TPP-Gaze: Modelling Gaze Dynamics in Space and Time with Neural Temporal Point Processes
Attention guides our gaze to fixate the proper location of the scene and holds it in that location for the deserved amount of time given current processing demands before shifting to the next one. As such gaze deployment crucially is a temporal process. Existing computational models have made significant strides in pre...
Alessandro D'Amelio; Giuseppe Cartella; Vittorio Cuculo; Manuele Lucchi; Marcella Cornia; Rita Cucchiara; Giuseppe Boccignone
University of Milan, Italy; University of Modena and Reggio Emilia, Italy; University of Modena and Reggio Emilia, Italy; University of Milan, Italy; University of Modena and Reggio Emilia, Italy; University of Modena and Reggio Emilia, Italy; University of Milan, Italy
Poster
main
https://github.com/phuselab/tppgaze
https://openaccess.thecvf.com/content/WACV2025/html/DAmelio_TPP-Gaze_Modelling_Gaze_Dynamics_in_Space_and_Time_with_Neural_WACV_2025_paper.html
3
TPP-Gaze: Modelling Gaze Dynamics in Space and Time with Neural Temporal Point Processes Attention guides our gaze to fixate the proper location of the scene and holds it in that location for the deserved amount of time given current processing demands before shifting to the next one. As such gaze deployment crucially ...
[ -0.019777992740273476, -0.04186310991644859, -0.001454742974601686, 0.013200833462178707, 0.00014543069119099528, -0.028169220313429832, 0.017266206443309784, 0.0034653351176530123, 0.02256886661052704, 0.04882168769836426, -0.019852416589856148, -0.015582378953695297, -0.029136722907423973,...
wacv_2025_e2d8f3248d
e2d8f3248d
wacv
2,025
TRH2TQA: Table Recognition with Hierarchical Relationships to Table Question-Answering on Business Table Images
Despite advancements in visual question answering challenges persist with documents like financial reports often structured in complicated tabular structures with complex numerical computations. An alternative approach the pipeline-driven methodology includes table recognition (TR) and table question-answering (TQA). R...
Pongsakorn Jirachanchaisiri; Nam Tuan Ly; Atsuhiro Takasu
National Institute of Informatics, Tokyo, Japan + The Graduate University for Advanced Studies (SOKENDAI), Kanagawa, Japan; Tokyo University of Agriculture and Technology, Tokyo, Japan; National Institute of Informatics, Tokyo, Japan + The Graduate University for Advanced Studies (SOKENDAI), Kanagawa, Japan
Poster
main
https://openaccess.thecvf.com/content/WACV2025/html/Jirachanchaisiri_TRH2TQA_Table_Recognition_with_Hierarchical_Relationships_to_Table_Question-Answering_on_WACV_2025_paper.html
0
TRH2TQA: Table Recognition with Hierarchical Relationships to Table Question-Answering on Business Table Images Despite advancements in visual question answering challenges persist with documents like financial reports often structured in complicated tabular structures with complex numerical computations. An alternativ...
[ -0.02965524047613144, 0.011611095629632473, 0.010346797294914722, 0.036367177963256836, -0.005968233570456505, -0.01051413081586361, -0.0016303410520777106, 0.015915289521217346, 0.017105218023061752, 0.03614406660199165, -0.03937918320298195, -0.012801024131476879, -0.009486888535320759, ...
wacv_2025_a9952fa4b1
a9952fa4b1
wacv
2,025
TRNeRF: Restoring Blurry Rolling Shutter and Noisy Thermal Images with Neural Radiance Fields
Thermal cameras offer unique detection capabilities in building inspections search and rescue operations and autonomous vehicle perception. Of the different types of thermal cameras uncooled microbolometers are often chosen due to their relative affordability small size and low power consumption. However microbolometer...
Spencer Carmichael; Manohar Bhat; Mani Ramanagopal; Austin Buchan; Ram Vasudevan; Katherine A. Skinner
University of Michigan, Ann Arbor, MI USA; University of Michigan, Ann Arbor, MI USA; Carnegie Mellon University, Pittsburgh, PA USA; University of Michigan, Ann Arbor, MI USA; University of Michigan, Ann Arbor, MI USA; University of Michigan, Ann Arbor, MI USA
Poster
main
https://umautobots.github.io/trnerf
https://openaccess.thecvf.com/content/WACV2025/html/Carmichael_TRNeRF_Restoring_Blurry_Rolling_Shutter_and_Noisy_Thermal_Images_with_WACV_2025_paper.html
0
TRNeRF: Restoring Blurry Rolling Shutter and Noisy Thermal Images with Neural Radiance Fields Thermal cameras offer unique detection capabilities in building inspections search and rescue operations and autonomous vehicle perception. Of the different types of thermal cameras uncooled microbolometers are often chosen du...
[ -0.045363061130046844, -0.0068756816908717155, -0.006693060509860516, 0.01746770180761814, -0.01691070757806301, 0.022955462336540222, 0.030662069097161293, 0.010226777754724026, 0.014381405897438526, 0.03155691176652908, -0.023539850488305092, -0.031027311459183693, -0.023959878832101822, ...
wacv_2025_553d8eccf1
553d8eccf1
wacv
2,025
TRUST: Time-Domain Residual Unsupervised Stability Technique for Improved Heart Rate Estimation
Camera-based estimation of vital signs is a promising method for non-contact health monitoring which analyzes minute changes in video data. However the creation of accurate models for this task is challenging due to the scarcity of datasets that possess synchronized vital sign recordings. Our research enhances an exist...
Shahzad Ahmad; Sania Bano; Sukalpa Chanda; Santosh Kumar Vipparthi; Subrahmanyam Murala
Østfold University College, Norway; Indian Institute of Technology Ropar, India; Østfold University College, Norway; Indian Institute of Technology Ropar, India; Trinity College Dublin, Ireland
Poster
main
https://openaccess.thecvf.com/content/WACV2025/html/Ahmad_TRUST_Time-Domain_Residual_Unsupervised_Stability_Technique_for_Improved_Heart_Rate_WACV_2025_paper.html
0
TRUST: Time-Domain Residual Unsupervised Stability Technique for Improved Heart Rate Estimation Camera-based estimation of vital signs is a promising method for non-contact health monitoring which analyzes minute changes in video data. However the creation of accurate models for this task is challenging due to the scar...
[ -0.03800588473677635, -0.040678173303604126, -0.025646548718214035, 0.0067595914006233215, 0.010958239436149597, 0.014326437376439571, 0.044055648148059845, 0.019578225910663605, -0.012479959987103939, 0.03091689571738243, -0.03449850529432297, -0.005892025772482157, 0.0023185352329164743, ...
wacv_2025_cac9b53ebc
cac9b53ebc
wacv
2,025
TaCOS: Task-Specific Camera Optimization with Simulation
The performance of perception tasks is heavily influenced by imaging systems. However designing cameras with high task performance is costly requiring extensive camera knowledge and experimentation with physical hardware. Additionally cameras and perception tasks are mostly designed in isolation whereas recent methods ...
Chengyang Yan; Donald G. Dansereau
;
Poster
main
https://openaccess.thecvf.com/content/WACV2025/html/Yan_TaCOS_Task-Specific_Camera_Optimization_with_Simulation_WACV_2025_paper.html
0
2404.11031
TaCOS: Task-Specific Camera Optimization with Simulation The performance of perception tasks is heavily influenced by imaging systems. However designing cameras with high task performance is costly requiring extensive camera knowledge and experimentation with physical hardware. Additionally cameras and perception tasks...
[ -0.04227227345108986, -0.005032626911997795, -0.016081061214208603, 0.007573632523417473, -0.01987910084426403, 0.018640024587512016, -0.011789186857640743, 0.01443793810904026, 0.010092190466821194, 0.05627922713756561, -0.02456604316830635, -0.009064115583896637, 0.03340119868516922, 0.0...
wacv_2025_5b350ae4c2
5b350ae4c2
wacv
2,025
Talking Head Anime 4: Distillation for Real-Time Performance
We study the problem of creating a character model that can be controlled in real time from a single image of an anime character. A solution would greatly reduce the cost of creating avatars computer games and other interactive applications. Talking Head Anime 3 (THA3) is an open source project that attempts to directl...
Pramook Khungurn
pixiv Inc., Tokyo, Japan
Poster
main
https://openaccess.thecvf.com/content/WACV2025/html/Khungurn_Talking_Head_Anime_4_Distillation_for_Real-Time_Performance_WACV_2025_paper.html
0
Talking Head Anime 4: Distillation for Real-Time Performance We study the problem of creating a character model that can be controlled in real time from a single image of an anime character. A solution would greatly reduce the cost of creating avatars computer games and other interactive applications. Talking Head Anim...
[ 0.0032285472843796015, -0.05504347383975983, 0.009108783677220345, 0.014095818623900414, -0.024283885955810547, -0.03773771598935127, 0.004198506940156221, 0.009620512835681438, 0.0002215555723523721, -0.0018271065782755613, -0.00012509748921729624, -0.03170861303806305, 0.005982581991702318...
wacv_2025_a63a7ff49e
a63a7ff49e
wacv
2,025
Task Configuration Impacts Annotation Quality and Model Training Performance in Crowdsourced Image Segmentation
Many industrial image segmentation systems require training on large annotated datasets but there is little standardization for producing training annotations. Data is often obtained via crowdsourcing but many labeling task configurations are set by convenience and may have unintended effects on data quality. In this w...
Benjamin R Bauchwitz; Mary Cummings
Duke University Department of Computer Science; George Mason University College of Engineering and Computing
Poster
main
https://openaccess.thecvf.com/content/WACV2025/html/Bauchwitz_Task_Configuration_Impacts_Annotation_Quality_and_Model_Training_Performance_in_WACV_2025_paper.html
0
Task Configuration Impacts Annotation Quality and Model Training Performance in Crowdsourced Image Segmentation Many industrial image segmentation systems require training on large annotated datasets but there is little standardization for producing training annotations. Data is often obtained via crowdsourcing but man...
[ -0.07632608711719513, -0.03837926685810089, -0.05499226227402687, 0.01319850329309702, -0.01382914837449789, -0.011297558434307575, 0.027802444994449615, 0.026505116373300552, -0.010847097262740135, 0.015414770692586899, -0.0012939488515257835, -0.03470350429415703, 0.0032005244866013527, ...
wacv_2025_91d1ab7ef6
91d1ab7ef6
wacv
2,025
TaxaBind: A Unified Embedding Space for Ecological Applications
We present TaxaBind a unified embedding space for characterizing any species of interest. TaxaBind is a multimodal embedding space across six modalities: ground-level images of species geographic location satellite image text audio and environmental features useful for solving ecological problems. To learn this joint e...
Srikumar Sastry; Subash Khanal; Aayush Dhakal; Adeel Ahmad; Nathan Jacobs
Washington University in St. Louis; Washington University in St. Louis; Washington University in St. Louis; Washington University in St. Louis; Washington University in St. Louis
Poster
main
https://github.com/mvrl/TaxaBind
https://openaccess.thecvf.com/content/WACV2025/html/Sastry_TaxaBind_A_Unified_Embedding_Space_for_Ecological_Applications_WACV_2025_paper.html
7
2411.00683
TaxaBind: A Unified Embedding Space for Ecological Applications We present TaxaBind a unified embedding space for characterizing any species of interest. TaxaBind is a multimodal embedding space across six modalities: ground-level images of species geographic location satellite image text audio and environmental featur...
[ -0.03893326222896576, -0.04101249948143959, -0.054327256977558136, 0.014354375191032887, -0.019180497154593468, -0.02939547225832939, 0.055433642119169235, -0.01674836128950119, -0.019762303680181503, 0.021574482321739197, -0.04498022049665451, -0.008889220654964447, -0.0005683927447535098, ...
wacv_2025_1485cadd5f
1485cadd5f
wacv
2,025
TempA-VLP: Temporal-Aware Vision-Language Pretraining for Longitudinal Exploration in Chest X-ray Image
Longitudinal medical image processing is a significant task to understand the dynamic changes of disease by taking and comparing image series over time providing insights into how conditions evolve and enabling more accurate diagnosis and treatment planning. While recent advancements in biomedical Vision-Language Pre-t...
Zhuoyi Yang; Liyue Shen
University of Michigan; University of Michigan
Poster
main
https://openaccess.thecvf.com/content/WACV2025/html/Yang_TempA-VLP_Temporal-Aware_Vision-Language_Pretraining_for_Longitudinal_Exploration_in_Chest_X-ray_WACV_2025_paper.html
1
TempA-VLP: Temporal-Aware Vision-Language Pretraining for Longitudinal Exploration in Chest X-ray Image Longitudinal medical image processing is a significant task to understand the dynamic changes of disease by taking and comparing image series over time providing insights into how conditions evolve and enabling more ...
[ -0.07605059444904327, -0.026088379323482513, -0.02284758724272251, 0.015852876007556915, -0.05196070671081543, 0.03055347129702568, -0.036134835332632065, 0.032587967813014984, 0.019840851426124573, -0.002381082158535719, 0.009470315650105476, -0.012288005091249943, 0.02270355075597763, 0....
wacv_2025_6a42bf9d99
6a42bf9d99
wacv
2,025
Temporal Dynamics in Visual Data: Analyzing the Impact of Time on Classification Accuracy
Visual datasets are generally constructed from the samples available at the time of their collection and are not further updated. However these static datasets do not reflect the distribution changes that occur in real data. We analyze how different collection times lead to a shift in class distribution by collecting a...
Tom Pégeot; Eva Feillet; Adrian Popescu; Inna Kucher; Bertrand Delezoide
Université Paris-Saclay, CEA, LIST; Université Paris-Saclay, CEA, LIST + Université Paris-Saclay, CentraleSupélec, MICS; Université Paris-Saclay, CEA, LIST; Université Paris-Saclay, CEA, LIST; Amanda
Poster
main
https://openaccess.thecvf.com/content/WACV2025/html/Pegeot_Temporal_Dynamics_in_Visual_Data_Analyzing_the_Impact_of_Time_WACV_2025_paper.html
0
Temporal Dynamics in Visual Data: Analyzing the Impact of Time on Classification Accuracy Visual datasets are generally constructed from the samples available at the time of their collection and are not further updated. However these static datasets do not reflect the distribution changes that occur in real data. We an...
[ -0.06769631057977676, -0.04913562163710594, -0.03792109340429306, 0.027766646817326546, -0.0020353063009679317, -0.02893831394612789, -0.00325463037006557, 0.021071406081318855, -0.006769631057977676, 0.024902570992708206, 0.004038066603243351, -0.03193257376551628, -0.03820006176829338, 0...
wacv_2025_0d3f0ec46e
0d3f0ec46e
wacv
2,025
Temporally Grounding Instructional Diagrams in Unconstrained Videos
We study the challenging problem of simultaneously localizing a sequence of queries in the form of instructional diagrams in a video. This requires understanding not only the individual queries but also their interrelationships. However most existing methods focus on grounding one query at a time ignoring the inherent ...
Jiahao Zhang; Frederic Z. Zhang; Cristian Rodriguez; Yizhak Ben-Shabat; Anoop Cherian; Stephen Gould
;;;;;
Poster
main
https://openaccess.thecvf.com/content/WACV2025/html/Zhang_Temporally_Grounding_Instructional_Diagrams_in_Unconstrained_Videos_WACV_2025_paper.html
2
2407.12066
Temporally Grounding Instructional Diagrams in Unconstrained Videos We study the challenging problem of simultaneously localizing a sequence of queries in the form of instructional diagrams in a video. This requires understanding not only the individual queries but also their interrelationships. However most existing m...
[ -0.03019043616950512, -0.038615237921476364, -0.021634289994835854, -0.011361286044120789, -0.006829905789345503, 0.018153665587306023, 0.009180031716823578, -0.007317756302654743, -0.05317569524049759, 0.04308094456791878, -0.003841822035610676, -0.04867246001958847, 0.031729042530059814, ...
wacv_2025_babf74e881
babf74e881
wacv
2,025
Temporally Streaming Audio-Visual Synchronization for Real-World Videos
We introduce RealSync a novel dataset designed to significantly enhance the training and evaluation of models for audio-visual synchronization (AV Sync) tasks. Sourced from high-quality YouTube channels RealSync covers a wide range of content domains providing an improved scale diversity and alignment with broadcast co...
Jordan G Voas; Wei-Cheng Tseng; Layne Berry; Xixi Hu; Puyuan Peng; James Stuedemann; David Harwath
The University of Texas at Austin; The University of Texas at Austin; The University of Texas at Austin; The University of Texas at Austin; The University of Texas at Austin; The University of Texas at Austin; The University of Texas at Austin
Poster
main
https://github.com/jvoas655/StreamSync
https://openaccess.thecvf.com/content/WACV2025/html/Voas_Temporally_Streaming_Audio-Visual_Synchronization_for_Real-World_Videos_WACV_2025_paper.html
0
Temporally Streaming Audio-Visual Synchronization for Real-World Videos We introduce RealSync a novel dataset designed to significantly enhance the training and evaluation of models for audio-visual synchronization (AV Sync) tasks. Sourced from high-quality YouTube channels RealSync covers a wide range of content domai...
[ -0.032644595950841904, -0.04969334974884987, -0.030382463708519936, 0.013627969659864902, -0.018906278535723686, 0.004119656048715115, 0.02431332692503929, 0.024386892095208168, -0.0012977375881746411, 0.048553090542554855, 0.015301579609513283, -0.02832263521850109, -0.00768757238984108, ...
wacv_2025_9e1bb15a0e
9e1bb15a0e
wacv
2,025
Test-Time Adaptation in Point Clouds: Leveraging Sampling Variation with Weight Averaging
Test-Time Adaptation (TTA) addresses distribution shifts during testing by adapting a pretrained model without access to source data. In this work we propose a novel TTA approach for 3D point cloud classification combining sampling variation with weight averaging. Our method leverages Farthest Point Sampling (FPS) and ...
Ali Bahri; Moslem Yazdanpanah; Mehrdad Noori; Sahar Dastani Oghani; Milad Cheraghalikhani; David Osowiechi; Farzad Beizaee; Gustavo A. Vargas Hakim; Ismail Ben Ayed; Christian Desrosiers
;;;;;;;;;
Poster
main
https://github.com/AliBahri94/SVWA_TTA
https://openaccess.thecvf.com/content/WACV2025/html/Bahri_Test-Time_Adaptation_in_Point_Clouds_Leveraging_Sampling_Variation_with_Weight_WACV_2025_paper.html
0
2411.01116
Test-Time Adaptation in Point Clouds: Leveraging Sampling Variation with Weight Averaging Test-Time Adaptation (TTA) addresses distribution shifts during testing by adapting a pretrained model without access to source data. In this work we propose a novel TTA approach for 3D point cloud classification combining samplin...
[ -0.011268001049757004, -0.003722304245457053, -0.029904231429100037, -0.0003375233500264585, -0.029508864507079124, -0.0228594858199358, 0.010809732601046562, 0.027603905647993088, 0.01763882488012314, 0.028790012001991272, -0.018941743299365044, 0.017045771703124046, -0.026723314076662064, ...
wacv_2025_0b3378ae6d
0b3378ae6d
wacv
2,025
Test-Time Adaptation of 3D Point Clouds via Denoising Diffusion Models
Test-time adaptation (TTA) of 3D point clouds is crucial for mitigating discrepancies between training and testing samples in real-world scenarios particularly when handling corrupted point clouds. LiDAR data for instance can be affected by sensor failures or environmental factors causing domain gaps. Adapting models t...
Hamidreza Dastmalchi; Aijun An; Ali Cheraghian; Shafin Rahman; Sameera Ramasinghe
York University, Canada; York University, Canada; Data61-CSIRO, Australia+Australian National University; North South University, Bangladesh; University of Adelaide, Australia
Poster
main
https://github.com/hamidreza-dastmalchi/3DD-TTA
https://openaccess.thecvf.com/content/WACV2025/html/Dastmalchi_Test-Time_Adaptation_of_3D_Point_Clouds_via_Denoising_Diffusion_Models_WACV_2025_paper.html
1
2411.14495
Test-Time Adaptation of 3D Point Clouds via Denoising Diffusion Models Test-time adaptation (TTA) of 3D point clouds is crucial for mitigating discrepancies between training and testing samples in real-world scenarios particularly when handling corrupted point clouds. LiDAR data for instance can be affected by sensor f...
[ -0.009822795167565346, 0.002388725057244301, -0.030522102490067482, 0.001990232151001692, -0.01272498443722725, -0.006657176185399294, -0.007206359878182411, 0.02253885008394718, -0.0004498393682297319, 0.05093565583229065, -0.024003339931368828, -0.010510390624403954, -0.016421927139163017,...
wacv_2025_4fa5a6b9e9
4fa5a6b9e9
wacv
2,025
Test-Time Low Rank Adaptation via Confidence Maximization for Zero-Shot Generalization of Vision-Language Models
The conventional modus operandi for adapting pre-trained vision-language models (VLMs) during test-time involves tuning learnable prompts i.e. test-time prompt tuning. This paper introduces Test-Time Low-rank adaptation (TTL) as an alternative to prompt tuning for zero-shot generalization of large-scale VLMs. Taking in...
Raza Imam; Hanan Gani; Muhammad Huzaifa; Karthik Nandakumar
;;;
Poster
main
https://github.com/Razaimam45/TTL-Test-Time-Low-Rank-Adaptation
https://openaccess.thecvf.com/content/WACV2025/html/Imam_Test-Time_Low_Rank_Adaptation_via_Confidence_Maximization_for_Zero-Shot_Generalization_WACV_2025_paper.html
4
2407.15913
Test-Time Low Rank Adaptation via Confidence Maximization for Zero-Shot Generalization of Vision-Language Models The conventional modus operandi for adapting pre-trained vision-language models (VLMs) during test-time involves tuning learnable prompts i.e. test-time prompt tuning. This paper introduces Test-Time Low-ran...
[ -0.0435967892408371, -0.014121255837380886, -0.0025695464573800564, 0.04204876348376274, 0.012996235862374306, -0.0034335621166974306, 0.02910652756690979, -0.0019856609869748354, 0.01525527611374855, 0.024624446406960487, -0.025758467614650726, -0.025236457586288452, -0.01879234053194523, ...
wacv_2025_656c3c9f0c
656c3c9f0c
wacv
2,025
Text Change Detection in Multilingual Documents using Image Comparison
Document comparison typically relies on optical character recognition (OCR) as its core technology. However OCR requires the selection of appropriate language models for each document and the performance of multilingual or hybrid models remains limited. To overcome these challenges we propose text change detection (TCD...
Doyoung Park; Naresh Reddy Yarram; Sunjin Kim; MinKyu Kim; Seongho Joe; Taehee Lee
Samsung SDS; Samsung SDS; Samsung SDS; Samsung SDS; Samsung SDS; Samsung SDS
Poster
main
https://openaccess.thecvf.com/content/WACV2025/html/Park_Text_Change_Detection_in_Multilingual_Documents_using_Image_Comparison_WACV_2025_paper.html
1
2412.04137
Text Change Detection in Multilingual Documents using Image Comparison Document comparison typically relies on optical character recognition (OCR) as its core technology. However OCR requires the selection of appropriate language models for each document and the performance of multilingual or hybrid models remains limi...
[ -0.08818170428276062, -0.024297183379530907, -0.020175015553832054, 0.01924888975918293, 0.01413703802973032, 0.0005010840250179172, -0.0046669477596879005, 0.021900150924921036, 0.03708134964108467, 0.0010191923938691616, -0.02998105250298977, -0.08157170563936234, -0.019721031188964844, ...
wacv_2025_36723b8a63
36723b8a63
wacv
2,025
Text-to-Image Synthesis for Domain Generalization in Face Anti-Spoofing
This paper addresses the challenge of developing robust Face Anti-Spoofing (FAS) models for face recognition systems. Traditional FAS protocols are limited by a lack of diversity in subject identities and environmental conditions restricting generalization to real-world scenarios. Recent advancements in spoof image syn...
Naeun Ko; Yonghyun Jeong; Jong Chul Ye
NA VER Cloud, South Korea + Kim Jaechul Graduate School of AI, KAIST, South Korea; NA VER Cloud, South Korea; Kim Jaechul Graduate School of AI, KAIST, South Korea
Poster
main
https://openaccess.thecvf.com/content/WACV2025/html/Ko_Text-to-Image_Synthesis_for_Domain_Generalization_in_Face_Anti-Spoofing_WACV_2025_paper.html
0
Text-to-Image Synthesis for Domain Generalization in Face Anti-Spoofing This paper addresses the challenge of developing robust Face Anti-Spoofing (FAS) models for face recognition systems. Traditional FAS protocols are limited by a lack of diversity in subject identities and environmental conditions restricting genera...
[ -0.04254808649420738, -0.01056385226547718, -0.0241276565939188, -0.007335246540606022, -0.013902212493121624, -0.053194254636764526, -0.021621599793434143, 0.048365067690610886, 0.04298710450530052, 0.02734711579978466, -0.02639591135084629, 0.0056569199077785015, -0.005958744324743748, 0...
wacv_2025_31acd9b8d0
31acd9b8d0
wacv
2,025
Texture Shape and Order Matter: A New Transformer Design for Sequential DeepFake Detection
Sequential DeepFake detection is an emerging task that predicts the manipulation sequence in order. Existing methods typically formulate it as an image-to-sequence problem employing conventional Transformer architectures. However these methods lack dedicated design and consequently result in limited performance. As suc...
Yunfei Li; Yuezun Li; Xin Wang; Baoyuan Wu; Jiaran Zhou; Junyu Dong
;;;;;
Poster
main
https://openaccess.thecvf.com/content/WACV2025/html/Li_Texture_Shape_and_Order_Matter_A_New_Transformer_Design_for_WACV_2025_paper.html
0
2404.13873
Texture Shape and Order Matter: A New Transformer Design for Sequential DeepFake Detection Sequential DeepFake detection is an emerging task that predicts the manipulation sequence in order. Existing methods typically formulate it as an image-to-sequence problem employing conventional Transformer architectures. However...
[ -0.032778628170490265, -0.04776690527796745, -0.032985489815473557, 0.015467826277017593, 0.00619653332978487, 0.01985899545252323, -0.015749914571642876, 0.01494126208126545, 0.013634254224598408, 0.06070534139871597, -0.026365825906395912, -0.01302306354045868, 0.02433479204773903, -0.00...
wacv_2025_d9b40f3b13
d9b40f3b13
wacv
2,025
The FineView Dataset:A 3D Scanned Multi-View Object Dataset of Fine-Grained Category Instances
In the past decade state-of-the-art deep learning models have shown impressive performance in many computer vision tasks by learning from large and diverse image datasets. Most of these datasets consist of web-scraped image collections. This approach however makes it very challenging to obtain desirable data such as mu...
Suguru Onda; Ryan Farrell
Brigham Young University; Brigham Young University
Poster
main
https://github.com/byu-vision/fineview
https://openaccess.thecvf.com/content/WACV2025/html/Onda_The_FineView_DatasetA_3D_Scanned_Multi-View_Object_Dataset_of_Fine-Grained_WACV_2025_paper.html
0
The FineView Dataset:A 3D Scanned Multi-View Object Dataset of Fine-Grained Category Instances In the past decade state-of-the-art deep learning models have shown impressive performance in many computer vision tasks by learning from large and diverse image datasets. Most of these datasets consist of web-scraped image c...
[ -0.052854590117931366, -0.04956260696053505, -0.03840645030140877, -0.00039035125519149005, -0.0033102701418101788, 0.02719542384147644, -0.014823060482740402, 0.021690499037504196, -0.0189288929104805, 0.02050172910094261, -0.009318135678768158, 0.012372363358736038, 0.0007829886162653565, ...
wacv_2025_714bf845ee
714bf845ee
wacv
2,025
Through the Curved Cover: Synthesizing Cover Aberrated Scenes with Refractive Field
Recent extended reality headsets and field robots have adopted covers to protect the front-facing cameras from environmental hazards and falls. The surface irregularities on the cover can lead to optical aberrations like blurring and non-parametric distortions. Novel view synthesis methods like NeRF and 3D Gaussian Spl...
Liuyue Xie; Jiancong Guo; László A. Jeni; Zhiheng Jia; Mingyang Li; Yunwen Zhou; Chao Guo
Carnegie Mellon University; Google; Carnegie Mellon University; Google; Google; Google; Google
Poster
main
https://openaccess.thecvf.com/content/WACV2025/html/Xie_Through_the_Curved_Cover_Synthesizing_Cover_Aberrated_Scenes_with_Refractive_WACV_2025_paper.html
0
2411.06365
Through the Curved Cover: Synthesizing Cover Aberrated Scenes with Refractive Field Recent extended reality headsets and field robots have adopted covers to protect the front-facing cameras from environmental hazards and falls. The surface irregularities on the cover can lead to optical aberrations like blurring and no...
[ -0.043996408581733704, 0.0025010639801621437, -0.000369580986443907, -0.005908647086471319, -0.01522022858262062, -0.017897948622703552, -0.02761397697031498, 0.03971949592232704, 0.01628945767879486, 0.038678161799907684, -0.043103836476802826, 0.004318751394748688, 0.0042559923604130745, ...
wacv_2025_05b858d354
05b858d354
wacv
2,025
TimberVision: A Multi-Task Dataset and Framework for Log-Component Segmentation and Tracking in Autonomous Forestry Operations
Timber represents an increasingly valuable and versatile resource. However forestry operations such as harvesting handling and measuring logs still require substantial human labor in remote environments posing significant safety risks. Progressively automating these tasks has the potential of increasing their efficienc...
Daniel Steininger; Julia Simon; Andreas Trondl; Markus Murschitz
AIT Austrian Institute of Technology (Center for Vision, Automation & Control); AIT Austrian Institute of Technology (Center for Vision, Automation & Control); AIT Austrian Institute of Technology (Center for Vision, Automation & Control); AIT Austrian Institute of Technology (Center for Vision, Automation & Control)
Poster
main
https://openaccess.thecvf.com/content/WACV2025/html/Steininger_TimberVision_A_Multi-Task_Dataset_and_Framework_for_Log-Component_Segmentation_and_WACV_2025_paper.html
0
2501.07360
TimberVision: A Multi-Task Dataset and Framework for Log-Component Segmentation and Tracking in Autonomous Forestry Operations Timber represents an increasingly valuable and versatile resource. However forestry operations such as harvesting handling and measuring logs still require substantial human labor in remote env...
[ -0.01724540814757347, -0.009381097741425037, -0.016528382897377014, 0.012842129915952682, -0.016592729836702347, -0.03704637289047241, -0.004221723880618811, 0.012814552523195744, 0.010810554958879948, 0.039344534277915955, 0.004775580950081348, -0.020977623760700226, 0.010258995927870274, ...
wacv_2025_c77ed30595
c77ed30595
wacv
2,025
To Ask or Not to Ask? Detecting Absence of Information in Vision and Language Navigation
Recent research in Vision Language Navigation (VLN) has overlooked the development of agents' inquisitive abilities which allow them to ask clarifying questions when instructions are incomplete. This paper addresses how agents can recognize "when" they lack sufficient information without focusing on "what" is missing p...
Savitha Sam Abraham; Sourav Garg; Feras Dayoub
Australian Institute for Machine Learning; Australian Institute for Machine Learning; Australian Institute for Machine Learning
Poster
main
https://openaccess.thecvf.com/content/WACV2025/html/Abraham_To_Ask_or_Not_to_Ask_Detecting_Absence_of_Information_WACV_2025_paper.html
0
2411.05831
To Ask or Not to Ask? Detecting Absence of Information in Vision and Language Navigation Recent research in Vision Language Navigation (VLN) has overlooked the development of agents' inquisitive abilities which allow them to ask clarifying questions when instructions are incomplete. This paper addresses how agents can ...
[ -0.03658434376120567, -0.04609627276659012, -0.010133863426744938, 0.013325847685337067, -0.002590628806501627, -0.04188907518982887, 0.032852742820978165, -0.011487483978271484, 0.009274131618440151, 0.00800282508134842, -0.008574455976486206, -0.04675479233264923, -0.01774340681731701, 0...
wacv_2025_5ed3e41560
5ed3e41560
wacv
2,025
Token Turing Machines are Efficient Vision Models
We propose Vision Token Turing Machines (ViTTM) an efficient low-latency memory-augmented Vision Transformer (ViT). Our approach builds on Neural Turing Machines (NTM) and Token Turing Machines (TTM) which were applied to NLP and sequential visual understanding tasks. ViTTMs are designed for non-sequential computer vis...
Purvish Jajal; Nick Eliopoulous; Benjamin Shiue-Hal Chou; George K. Thiravathukal; James C. Davis; Yung-Hsiang Lu
;;;;;
Poster
main
https://openaccess.thecvf.com/content/WACV2025/html/Jajal_Token_Turing_Machines_are_Efficient_Vision_Models_WACV_2025_paper.html
0
Token Turing Machines are Efficient Vision Models We propose Vision Token Turing Machines (ViTTM) an efficient low-latency memory-augmented Vision Transformer (ViT). Our approach builds on Neural Turing Machines (NTM) and Token Turing Machines (TTM) which were applied to NLP and sequential visual understanding tasks. V...
[ -0.0134030282497406, -0.030699988827109337, -0.003242592327296734, 0.04206199571490288, -0.005920846480876207, 0.03256230428814888, 0.006692108698189259, 0.03848785534501076, 0.00893535278737545, 0.030587121844291687, -0.014992580749094486, -0.018529098480939865, -0.00617950176820159, 0.03...
wacv_2025_e3d82a5856
e3d82a5856
wacv
2,025
TokenBinder: Text-Video Retrieval with One-to-Many Alignment Paradigm
Text-Video Retrieval (TVR) methods typically match query-candidate pairs by aligning text and video features in coarse-grained fine-grained or combined (coarse-to-fine) manners. However these frameworks predominantly employ a one(query)-to-one(candidate) alignment paradigm which struggles to discern nuanced differences...
Bingqing Zhang; Zhuo Cao; Heming Du; Xin Yu; Xue Li; Jiajun Liu; Sen Wang
The University of Queensland, Australia+CSIRO Data61, Australia; The University of Queensland, Australia+CSIRO Data61, Australia; The University of Queensland, Australia; The University of Queensland, Australia; The University of Queensland, Australia; CSIRO Data61, Australia+The University of Queensland, Australia; Th...
Poster
main
https://github.com/bingqingzhang/TokenBinder
https://openaccess.thecvf.com/content/WACV2025/html/Zhang_TokenBinder_Text-Video_Retrieval_with_One-to-Many_Alignment_Paradigm_WACV_2025_paper.html
0
2409.19865
TokenBinder: Text-Video Retrieval with One-to-Many Alignment Paradigm Text-Video Retrieval (TVR) methods typically match query-candidate pairs by aligning text and video features in coarse-grained fine-grained or combined (coarse-to-fine) manners. However these frameworks predominantly employ a one(query)-to-one(candid...
[ -0.024699131026864052, -0.042178232222795486, -0.024735968559980392, 0.028290728107094765, -0.0018153677228838205, 0.04066792130470276, 0.031237678602337837, 0.01060902327299118, -0.020554982125759125, 0.0048394459299743176, -0.016825247555971146, -0.01765407808125019, -0.03993118554353714, ...
wacv_2025_162a071be1
162a071be1
wacv
2,025
Towards Accurate Unified Anomaly Segmentation
Unsupervised anomaly detection (UAD) from images strives to model normal data distributions creating discriminative representations to distinguish and precisely localize anomalies. Despite recent advancements in the efficient and unified one-for-all scheme challenges persist in accurately segmenting anomalies for furth...
Wenxin Ma; Qingsong Yao; Xiang Zhang; Zhelong Huang; Zihang Jiang; S.Kevin Zhou
;;;;;
Poster
main
https://github.com/Mwxinnn/UniAS
https://openaccess.thecvf.com/content/WACV2025/html/Ma_Towards_Accurate_Unified_Anomaly_Segmentation_WACV_2025_paper.html
1
2501.12295
Towards Accurate Unified Anomaly Segmentation Unsupervised anomaly detection (UAD) from images strives to model normal data distributions creating discriminative representations to distinguish and precisely localize anomalies. Despite recent advancements in the efficient and unified one-for-all scheme challenges persis...
[ -0.0684976652264595, -0.002105009974911809, -0.06937819719314575, -0.005764700472354889, -0.038743190467357635, 0.003941734321415424, -0.0033914048690348864, 0.011923803947865963, -0.020967550575733185, 0.06769051402807236, -0.032744601368904114, -0.017417926341295242, 0.016546571627259254, ...
wacv_2025_a0523a565e
a0523a565e
wacv
2,025
Towards Generalized Face Anti-Spoofing from a Frequency Shortcut View
The generalization capability of a face anti-spoofing (FAS) model is critical to its practicality in the real world. Recent studies have theoretically and empirically uncovered that neural networks tend to exploit easy-to-learn frequency sets for decisions. These simplicity-biased representations depending on what best...
Junyi Cao; Chao Ma
MoE Key Lab of Artificial Intelligence, AI Institute, Shanghai Jiao Tong University; MoE Key Lab of Artificial Intelligence, AI Institute, Shanghai Jiao Tong University
Poster
main
https://github.com/VISION-SJTU/UniDefense
https://openaccess.thecvf.com/content/WACV2025/html/Cao_Towards_Generalized_Face_Anti-Spoofing_from_a_Frequency_Shortcut_View_WACV_2025_paper.html
0
Towards Generalized Face Anti-Spoofing from a Frequency Shortcut View The generalization capability of a face anti-spoofing (FAS) model is critical to its practicality in the real world. Recent studies have theoretically and empirically uncovered that neural networks tend to exploit easy-to-learn frequency sets for dec...
[ -0.057159386575222015, -0.037945836782455444, -0.0485706701874733, -0.004845035262405872, 0.005830701440572739, -0.062009047716856, 0.03054177016019821, 0.03405870124697685, 0.012022350914776325, 0.060083989053964615, -0.053494371473789215, -0.01799187809228897, -0.010245375335216522, -0.0...
wacv_2025_adaf5925ec
adaf5925ec
wacv
2,025
Towards High-Fidelity Head Blending with Chroma Keying for Industrial Applications
We introduce an industrial Head Blending pipeline for the task of seamlessly integrating an actor's head onto a target body in digital content creation. The key challenge stems from discrepancies in head shape and hair structure which lead to unnatural boundaries and blending artifacts. Existing methods treat foregroun...
Hah Min Lew; Sahng-Min Yoo; Hyunwoo Kang; Gyeong-Moon Park
Klleon AI Research; Samsung Research + Klleon AI Research; Hyperconnect + Klleon AI Research; Kyung Hee University
Poster
main
https://openaccess.thecvf.com/content/WACV2025/html/Lew_Towards_High-Fidelity_Head_Blending_with_Chroma_Keying_for_Industrial_Applications_WACV_2025_paper.html
0
2411.00652
Towards High-Fidelity Head Blending with Chroma Keying for Industrial Applications We introduce an industrial Head Blending pipeline for the task of seamlessly integrating an actor's head onto a target body in digital content creation. The key challenge stems from discrepancies in head shape and hair structure which le...
[ -0.06305290758609772, -0.030883805826306343, -0.04076221585273743, 0.014064800925552845, -0.009065921418368816, -0.012026689946651459, -0.017480013892054558, 0.03955036774277687, -0.00917149893939495, -0.00016912532737478614, -0.02772565186023712, -0.01988535188138485, -0.0178105179220438, ...
wacv_2025_8d9fdca6f6
8d9fdca6f6
wacv
2,025
Towards On-the-Fly Novel Category Discovery in Dynamic Long-Tailed Distributions
As the diversity of real-world object categories increases the need for sophisticated classification methods also grows. However Novel Category Discovery (NCD) which aims to predict unseen categories often falls short in scenarios where new categories are constantly updated and data distributions are potentially biased...
Hoin Jung; Xiaoqian Wang
;
Poster
main
https://github.com/HoinJung/NCD-DLT
https://openaccess.thecvf.com/content/WACV2025/html/Jung_Towards_On-the-Fly_Novel_Category_Discovery_in_Dynamic_Long-Tailed_Distributions_WACV_2025_paper.html
0
Towards On-the-Fly Novel Category Discovery in Dynamic Long-Tailed Distributions As the diversity of real-world object categories increases the need for sophisticated classification methods also grows. However Novel Category Discovery (NCD) which aims to predict unseen categories often falls short in scenarios where ne...
[ -0.0725504457950592, -0.03907667472958565, -0.03720904141664505, 0.041590798646211624, -0.016943402588367462, 0.005306598264724016, -0.007304429076611996, 0.0072325970977544785, -0.023040153086185455, 0.027511702850461006, -0.029540961608290672, -0.028194108977913857, -0.04313519224524498, ...
wacv_2025_2c82ed6d28
2c82ed6d28
wacv
2,025
Towards Privacy-Preserving Split Learning for ControlNet
With the emerging trend of large generative models ControlNet is introduced to enable users to fine-tune pre-trained models with their own data for various use cases. A natural question arises: how can we train ControlNet models while ensuring users' data privacy across distributed devices? We first propose a new distr...
Dixi Yao
Poster
main
https://openaccess.thecvf.com/content/WACV2025/html/Yao_Towards_Privacy-Preserving_Split_Learning_for_ControlNet_WACV_2025_paper.html
0
Towards Privacy-Preserving Split Learning for ControlNet With the emerging trend of large generative models ControlNet is introduced to enable users to fine-tune pre-trained models with their own data for various use cases. A natural question arises: how can we train ControlNet models while ensuring users' data privacy...
[ -0.044526029378175735, -0.007678336463868618, -0.04667097330093384, 0.007867868058383465, -0.046966828405857086, -0.019914673641324043, 0.014025329612195492, 0.014154765754938126, 0.012342660687863827, 0.03160090744495392, -0.04090182110667229, 0.0020848463755100965, -0.038165170699357986, ...
wacv_2025_f2f6ecaebe
f2f6ecaebe
wacv
2,025
Towards Real-Time Open-Vocabulary Video Instance Segmentation
In this paper we address the challenge of performing open-vocabulary video instance segmentation (OV-VIS) in real-time. We analyze the computational bottlenecks of state-of-the-art foundation models that performs OV-VIS and propose a new method TROY-VIS that significantly improves processing speed while maintaining hig...
Bin Yan; Martin Sundermeyer; David Joseph Tan; Huchuan Lu; Federico Tombari
Dalian University of Technology; Google; Google; Dalian University of Technology; Google+TU Munich
Poster
main
https://github.com/google-research/troyvis
https://openaccess.thecvf.com/content/WACV2025/html/Yan_Towards_Real-Time_Open-Vocabulary_Video_Instance_Segmentation_WACV_2025_paper.html
2
2412.04434
Towards Real-Time Open-Vocabulary Video Instance Segmentation In this paper we address the challenge of performing open-vocabulary video instance segmentation (OV-VIS) in real-time. We analyze the computational bottlenecks of state-of-the-art foundation models that performs OV-VIS and propose a new method TROY-VIS that...
[ -0.002477461937814951, -0.026078546419739723, -0.008299383334815502, 0.03175177425146103, 0.007343169767409563, 0.04106684774160385, 0.01771511137485504, 0.009507231414318085, 0.033490344882011414, 0.041213251650333405, -0.018620997667312622, -0.03098314255475998, 0.04190868139266968, 0.01...
wacv_2025_94c3cd8758
94c3cd8758
wacv
2,025
Towards Robust Training via Gradient-Diversified Backpropagation
Neural networks are prone to be vulnerable to adversarial attacks and domain shifts. Adversarial-driven methods including adversarial training and adversarial augmentation have been frequently proposed to improve the model's robustness against adversarial attacks and distribution-shifted samples. Nonetheless recent res...
Xilin He; Cheng Luo; Qinliang Lin; Weicheng Xie; Muhammad Haris Khan; Siyang Song; Linlin Shen
Shenzhen University, China; Monash University, Australia; Shenzhen University, China; Shenzhen University, China; MBZUAI, UAE; University of Exeter, England; Shenzhen University, China
Poster
main
https://openaccess.thecvf.com/content/WACV2025/html/He_Towards_Robust_Training_via_Gradient-Diversified_Backpropagation_WACV_2025_paper.html
0
Towards Robust Training via Gradient-Diversified Backpropagation Neural networks are prone to be vulnerable to adversarial attacks and domain shifts. Adversarial-driven methods including adversarial training and adversarial augmentation have been frequently proposed to improve the model's robustness against adversarial...
[ -0.03835572674870491, -0.018993813544511795, -0.031177829951047897, -0.03226371854543686, -0.010555186308920383, -0.017963141202926636, 0.029742252081632614, -0.0033427823800593615, 0.018469275906682014, 0.018340442329645157, -0.0277545265853405, 0.025895636528730392, -0.037048980593681335, ...
wacv_2025_f358f97804
f358f97804
wacv
2,025
Towards Secure and Usable 3D Assets: A Novel Framework for Automatic Visible Watermarking
3D models particularly AI-generated ones have witnessed a recent surge across various industries such as entertainment. Hence there is an alarming need to protect the intellectual property and avoid the misuse of these valuable assets. As a viable solution to address these concerns we rigorously define the novel task o...
Gursimran Singh; Tianxi Hu; Mohammad Akbari; Qiang Tang; Yong Zhang
Huawei Technologies Canada Co. Ltd.; Huawei Technologies Canada Co. Ltd.; Huawei Technologies Canada Co. Ltd.; Huawei Technologies Canada Co. Ltd.; Huawei Technologies Canada Co. Ltd.
Poster
main
https://openaccess.thecvf.com/content/WACV2025/html/Singh_Towards_Secure_and_Usable_3D_Assets_A_Novel_Framework_for_WACV_2025_paper.html
1
2409.00314
Towards Secure and Usable 3D Assets: A Novel Framework for Automatic Visible Watermarking 3D models particularly AI-generated ones have witnessed a recent surge across various industries such as entertainment. Hence there is an alarming need to protect the intellectual property and avoid the misuse of these valuable as...
[ -0.022763609886169434, 0.0002519169938750565, -0.022468458861112595, -0.0016809838125482202, -0.007517156656831503, -0.008259648457169533, -0.0014227256178855896, 0.015246453694999218, 0.0332968533039093, 0.025788920000195503, 0.0024788170121610165, -0.0014261844335123897, 0.0232247859239578...