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_6d8be11da1
6d8be11da1
wacv
2,025
DiffuseKronA: A Parameter Efficient Fine-Tuning Method for Personalized Diffusion Models
In the realm of subject-driven text-to-image (T2I) generative models recent developments like DreamBooth and BLIP-Diffusion have led to impressive results yet encounter limitations due to their intensive fine-tuning demands and substantial parameter requirements. While the low-rank adaptation (LoRA) module within Dream...
Shyam Marjit; Harshit Singh; Nityanand Mathur; Sayak Paul; Chia-Mu Yu; Pin-Yu Chen
IIIT Guwahati; University of Maryland, College Park; Smallest.ai; Hugging Face; National Yang Ming Chiao Tung University; IBM Research
Poster
main
https://openaccess.thecvf.com/content/WACV2025/html/Marjit_DiffuseKronA_A_Parameter_Efficient_Fine-Tuning_Method_for_Personalized_Diffusion_Models_WACV_2025_paper.html
8
2402.17412
DiffuseKronA: A Parameter Efficient Fine-Tuning Method for Personalized Diffusion Models In the realm of subject-driven text-to-image (T2I) generative models recent developments like DreamBooth and BLIP-Diffusion have led to impressive results yet encounter limitations due to their intensive fine-tuning demands and sub...
[ -0.039374664425849915, -0.005826214328408241, -0.02203236147761345, 0.00573077704757452, -0.032085079699754715, -0.028103986755013466, -0.013888386078178883, 0.006948737893253565, -0.018632980063557625, 0.028994733467698097, -0.052426841109991074, 0.02195964753627777, -0.03295765072107315, ...
wacv_2025_afd5ea14b2
afd5ea14b2
wacv
2,025
Diffusion Model Guided Sampling with Pixel-Wise Aleatoric Uncertainty Estimation
Despite the remarkable progress in generative modelling current diffusion models lack a quantitative approach to assess image quality. To address this limitation we propose to estimate the pixel-wise aleatoric uncertainty during the sampling phase of diffusion models and utilise the uncertainty to improve the sample ge...
Michele De Vita; Vasileios Belagiannis
Friedrich-Alexander-Universit ¨at Erlangen-N ¨urnberg; Friedrich-Alexander-Universit ¨at Erlangen-N ¨urnberg
Poster
main
https://github.com/Michedev/diffusion-uncertainty
https://openaccess.thecvf.com/content/WACV2025/html/De_Vita_Diffusion_Model_Guided_Sampling_with_Pixel-Wise_Aleatoric_Uncertainty_Estimation_WACV_2025_paper.html
1
2412.00205
Diffusion Model Guided Sampling with Pixel-Wise Aleatoric Uncertainty Estimation Despite the remarkable progress in generative modelling current diffusion models lack a quantitative approach to assess image quality. To address this limitation we propose to estimate the pixel-wise aleatoric uncertainty during the sampli...
[ -0.09608473628759384, -0.016544271260499954, -0.023783531039953232, 0.0019891972187906504, -0.00003049207225558348, -0.022412458434700966, -0.0004204621072858572, 0.031132478266954422, -0.016187790781259537, 0.05027264356613159, -0.03226589784026146, -0.012860656715929508, -0.023600721731781...
wacv_2025_e24986d0aa
e24986d0aa
wacv
2,025
Diffusion-Based Conditional Image Editing through Optimized Inference with Guidance
We present a simple but effective training-free approach for text-driven image-to-image translation based on a pretrained text-to-image diffusion model. Our goal is to generate an image that aligns with the target task while preserving the structure and background of a source image. To this end we derive the representa...
Hyunsoo Lee; Minsoo Kang; Bohyung Han
ECE; ECE; ECE+IPAI
Poster
main
https://openaccess.thecvf.com/content/WACV2025/html/Lee_Diffusion-Based_Conditional_Image_Editing_through_Optimized_Inference_with_Guidance_WACV_2025_paper.html
1
2412.15798
Diffusion-Based Conditional Image Editing through Optimized Inference with Guidance We present a simple but effective training-free approach for text-driven image-to-image translation based on a pretrained text-to-image diffusion model. Our goal is to generate an image that aligns with the target task while preserving ...
[ -0.06856399029493332, -0.0006083786720409989, -0.01861695386469364, 0.0021528115030378103, -0.014931593090295792, -0.03580322489142418, -0.017883503809571266, 0.02051849104464054, 0.036600060760974884, 0.033050525933504105, -0.024792421609163284, -0.017430756241083145, -0.015900472179055214,...
wacv_2025_cda3080e35
cda3080e35
wacv
2,025
Diffusion-Based Generative Regularization for Supervised Discriminative Learning
Ensuring the quality and quantity of labeled training data has long been a challenge in training deep neural networks for discriminative tasks. One solution to this problem is to use a generative model to augment training data and learn a discriminative model with it. For image classification with the recent developmen...
Takuya Asakura; Nakamasa Inoue; Koichi Shinoda
Institute of Science Tokyo; Institute of Science Tokyo; Institute of Science Tokyo
Poster
main
https://openaccess.thecvf.com/content/WACV2025/html/Asakura_Diffusion-Based_Generative_Regularization_for_Supervised_Discriminative_Learning_WACV_2025_paper.html
0
Diffusion-Based Generative Regularization for Supervised Discriminative Learning Ensuring the quality and quantity of labeled training data has long been a challenge in training deep neural networks for discriminative tasks. One solution to this problem is to use a generative model to augment training data and learn a ...
[ -0.06787954270839691, -0.031269337981939316, -0.027155427262187004, 0.00006022713932907209, -0.019992167130112648, -0.0068835848942399025, 0.00007101093069650233, 0.030024338513612747, 0.01809760369360447, 0.006166356615722179, -0.023673035204410553, -0.020515428856015205, -0.030691947788000...
wacv_2025_5fd0cbc738
5fd0cbc738
wacv
2,025
Diffusion-Based Particle-DETR for BEV Perception
The Bird-Eye-View (BEV) is one of the most widely-used scene representations for visual perception in Autonomous Vehicles (AVs) due to its well suited compatibility to downstream tasks. For the enhanced safety of AVs modeling perception uncertainty in BEV is crucial. Recent diffusion-based methods offer a promising app...
Asen Nachkov; Danda Pani Paudel; Martin Danelljan; Luc Van Gool
INSAIT, Sofia University; INSAIT, Sofia University; ETH Zurich; INSAIT, Sofia University+ETH Zurich
Poster
main
https://github.com/insait-institute/ParticleDETR
https://openaccess.thecvf.com/content/WACV2025/html/Nachkov_Diffusion-Based_Particle-DETR_for_BEV_Perception_WACV_2025_paper.html
3
2312.11578
Diffusion-Based Particle-DETR for BEV Perception The Bird-Eye-View (BEV) is one of the most widely-used scene representations for visual perception in Autonomous Vehicles (AVs) due to its well suited compatibility to downstream tasks. For the enhanced safety of AVs modeling perception uncertainty in BEV is crucial. Rec...
[ -0.039443738758563995, -0.061903730034828186, -0.01045296061784029, 0.0132188331335783, -0.007275445852428675, 0.0032029536087065935, 0.04451296105980873, 0.02923128940165043, 0.05698251351714134, 0.035595569759607315, -0.022367486730217934, -0.02307051792740822, 0.01076747477054596, 0.014...
wacv_2025_48f3e21289
48f3e21289
wacv
2,025
Diffusion-Based Visual Anagram as Multi-Task Learning
Visual anagrams are images that change appearance upon transformation like flipping or rotation. With the advent of diffusion models generating such optical illusions can be achieved by averaging noise across multiple views during the reverse denoising process. However we observe two critical failure modes in this appr...
Zhiyuan Xu; Yinhe Chen; Huan-ang Gao; Weiyan Zhao; Guiyu Zhang; Hao Zhao
Institute for AI Industry Research (AIR), Tsinghua University; Institute for AI Industry Research (AIR), Tsinghua University + University of Chinese Academy of Sciences; Institute for AI Industry Research (AIR), Tsinghua University; Institute for AI Industry Research (AIR), Tsinghua University + Beijing Institute of Te...
Poster
main
https://github.com/Pixtella/Anagram-MTL
https://openaccess.thecvf.com/content/WACV2025/html/Xu_Diffusion-Based_Visual_Anagram_as_Multi-Task_Learning_WACV_2025_paper.html
4
2412.02693
Diffusion-Based Visual Anagram as Multi-Task Learning Visual anagrams are images that change appearance upon transformation like flipping or rotation. With the advent of diffusion models generating such optical illusions can be achieved by averaging noise across multiple views during the reverse denoising process. Howe...
[ -0.04633660987019539, -0.04012847691774368, -0.008942890912294388, 0.03412603214383125, -0.027469119057059288, 0.03631383925676346, -0.003197562647983432, 0.040801648050546646, -0.0018897874979302287, 0.009531915187835693, -0.03833335265517235, -0.05785531550645828, 0.0031367901246994734, ...
wacv_2025_798853d61f
798853d61f
wacv
2,025
DisCo: Discovering Common Affordance from Large Models for Actionable Part Perception
Actionable part perception for robotic object manipulation needs to perceive parts over open-world object categories within 3D space which is challenging as the appearance of the same part on different objects varies greatly. It is frequently observed that despite the huge intra-class difference in appearance the parts...
Youpeng Wen; Yi Zhu; Zhihao Zhan; Pengzhen Ren; Jianhua Han; Hang Xu; Shen Zhao; Xiaodan Liang
Shenzhen Campus of Sun Yat-Sen University; Huawei Noah’s Ark Lab; Peng Cheng Laboratory; Shenzhen Campus of Sun Yat-Sen University; Huawei Noah’s Ark Lab; Huawei Noah’s Ark Lab; Shenzhen Campus of Sun Yat-Sen University; Shenzhen Campus of Sun Yat-Sen University+Peng Cheng Laboratory
Poster
main
https://openaccess.thecvf.com/content/WACV2025/html/Wen_DisCo_Discovering_Common_Affordance_from_Large_Models_for_Actionable_Part_WACV_2025_paper.html
0
DisCo: Discovering Common Affordance from Large Models for Actionable Part Perception Actionable part perception for robotic object manipulation needs to perceive parts over open-world object categories within 3D space which is challenging as the appearance of the same part on different objects varies greatly. It is fr...
[ -0.02312331274151802, -0.03857555612921715, -0.012341609224677086, 0.02549069933593273, 0.004200276453047991, -0.017168141901493073, 0.02984008379280567, -0.003358385758474469, 0.0021012851502746344, 0.022829683497548103, -0.016369836404919624, -0.020168665796518326, 0.007487548515200615, ...
wacv_2025_5839586911
5839586911
wacv
2,025
DisFlowEm : One-Shot Emotional Talking Head Generation using Disentangled Pose and Expression Flow-Guidance
Generating realistic one-shot emotional talking head animation on arbitrary faces is a challenging problem as it requires realistic emotions head movements identity preservation and accurate lip sync. Existing emotional talking face generation methods either fail to retain the identity information of arbitrary subjects...
Sanjana Sinha; Brojeshwar Bhowmick; Lokender Tiwari; Sushovan Chanda
TCS Research, India; TCS Research, India; TCS Research, India; TCS Research, India
Poster
main
https://openaccess.thecvf.com/content/WACV2025/html/Sinha_DisFlowEm__One-Shot_Emotional_Talking_Head_Generation_using_Disentangled_Pose_WACV_2025_paper.html
0
DisFlowEm : One-Shot Emotional Talking Head Generation using Disentangled Pose and Expression Flow-Guidance Generating realistic one-shot emotional talking head animation on arbitrary faces is a challenging problem as it requires realistic emotions head movements identity preservation and accurate lip sync. Existing em...
[ -0.006541935261338949, -0.0290731992572546, -0.02370927669107914, -0.0015551646938547492, -0.028514457866549492, -0.003913520835340023, -0.00961967185139656, 0.02113906480371952, -0.019518714398145676, -0.0038390217814594507, -0.005918006878346205, -0.055054694414138794, 0.005093862768262625...
wacv_2025_7171889674
7171889674
wacv
2,025
Discriminative Score Suppression for Weakly Supervised Video Anomaly Detection
Weakly supervised video anomaly detection (WSVAD) often relies on Multiple Instance Learning (MIL). However selecting only the most discriminative segments for training limits the model's ability to comprehensively detect anomalous events particularly hard anomalies. To overcome this limitation we propose the Discrimin...
Chen Xu; Chunguo Li; Hongjie Xing
;;
Poster
main
https://openaccess.thecvf.com/content/WACV2025/html/Xu_Discriminative_Score_Suppression_for_Weakly_Supervised_Video_Anomaly_Detection_WACV_2025_paper.html
0
Discriminative Score Suppression for Weakly Supervised Video Anomaly Detection Weakly supervised video anomaly detection (WSVAD) often relies on Multiple Instance Learning (MIL). However selecting only the most discriminative segments for training limits the model's ability to comprehensively detect anomalous events pa...
[ -0.04089881852269173, -0.032730016857385635, -0.03695147857069969, 0.020869726315140724, -0.021015925332903862, -0.019663596525788307, 0.01578935608267784, 0.015743669122457504, 0.042799387127161026, 0.051790546625852585, -0.01955394819378853, -0.008314996026456356, -0.017534591257572174, ...
wacv_2025_099c9c96fa
099c9c96fa
wacv
2,025
Disentangle Source and Target Knowledge for Continual Test-Time Adaptation
Continual Test-Time Adaptation (CoTTA) task is proposed to tackle the challenges of constant domain shifts during testing. The goals are twofold: 1) to preserve the knowledge from the source domain without source data and 2) to effectively extract target knowledge using unlabeled target domain data. Existing works prim...
Tianyi Ma; Maoying Qiao
University of Technology Sydney; University of Technology Sydney
Poster
main
https://openaccess.thecvf.com/content/WACV2025/html/Ma_Disentangle_Source_and_Target_Knowledge_for_Continual_Test-Time_Adaptation_WACV_2025_paper.html
0
Disentangle Source and Target Knowledge for Continual Test-Time Adaptation Continual Test-Time Adaptation (CoTTA) task is proposed to tackle the challenges of constant domain shifts during testing. The goals are twofold: 1) to preserve the knowledge from the source domain without source data and 2) to effectively extra...
[ -0.05344625934958458, -0.02135293185710907, -0.016868632286787033, 0.03879694268107414, -0.00012058400898240507, -0.053154002875089645, -0.003454553661867976, 0.019562864676117897, 0.016028398647904396, 0.023435253649950027, -0.00977229978889227, -0.01928887516260147, -0.03583785519003868, ...
wacv_2025_82a67ed8f9
82a67ed8f9
wacv
2,025
Disentangling Disentangled Representations: Towards Improved Latent Units via Diffusion Models
Disentangled representation learning (DRL) aims to break down observed data into core intrinsic factors for a profound understanding of the data. In real-world scenarios manually defining and labeling these factors are non-trivial making unsupervised methods attractive. Recently there have been limited explorations of ...
Youngjun Jun; Jiwoo Park; Kyobin Choo; Tae Eun Choi; Seong Jae Hwang
Yonsei University; Yonsei University; Yonsei University; Yonsei University; Yonsei University
Poster
main
https://openaccess.thecvf.com/content/WACV2025/html/Jun_Disentangling_Disentangled_Representations_Towards_Improved_Latent_Units_via_Diffusion_Models_WACV_2025_paper.html
0
2410.23820
Disentangling Disentangled Representations: Towards Improved Latent Units via Diffusion Models Disentangled representation learning (DRL) aims to break down observed data into core intrinsic factors for a profound understanding of the data. In real-world scenarios manually defining and labeling these factors are non-tr...
[ -0.059971027076244354, -0.015277895145118237, -0.014477667398750782, 0.011975809000432491, -0.042310845106840134, -0.015719398856163025, 0.03154917433857918, -0.00892206933349371, 0.0008847335702739656, 0.036847230046987534, -0.03197228163480759, -0.013953381218016148, -0.012904807925224304,...
wacv_2025_f9951bbbd2
f9951bbbd2
wacv
2,025
Disentangling Spatio-Temporal Knowledge for Weakly Supervised Object Detection and Segmentation in Surgical Video
Weakly supervised video object segmentation (WSVOS) enables the identification of segmentation maps without requiring extensive annotations of object masks relying instead on coarse video labels indicating object presence. Weakly supervised semantic segmentation of objects in surgical videos is however more challenging...
Guiqiu Liao; Matjaz Jogan; Sai Koushik; Eric Eaton; Daniel A. Hashimoto
PCASO Laboratory, Department of Surgery, University of Pennsylvania; PCASO Laboratory, Department of Surgery, University of Pennsylvania; PCASO Laboratory, Department of Surgery, University of Pennsylvania + Department of Electrical and Systems Engineering, University of Pennsylvania; Department of Computer and Informa...
Poster
main
https://github.com/PCASOlab/VDST-net
https://openaccess.thecvf.com/content/WACV2025/html/Liao_Disentangling_Spatio-Temporal_Knowledge_for_Weakly_Supervised_Object_Detection_and_Segmentation_WACV_2025_paper.html
0
2407.15794
Disentangling Spatio-Temporal Knowledge for Weakly Supervised Object Detection and Segmentation in Surgical Video Weakly supervised video object segmentation (WSVOS) enables the identification of segmentation maps without requiring extensive annotations of object masks relying instead on coarse video labels indicating ...
[ -0.023858966305851936, -0.030751558020710945, -0.03219589218497276, 0.018812932074069977, -0.012002614326775074, -0.02044009603559971, 0.04252563789486885, 0.010823378339409828, -0.0011466701980680227, 0.06110089644789696, 0.01803591661155224, -0.0184655599296093, 0.00792099442332983, 0.01...
wacv_2025_1ed10224b3
1ed10224b3
wacv
2,025
Disentangling Subject-Irrelevant Elements in Personalized Text-to-Image Diffusion via Filtered Self-Distillation
Recent research has unveiled the development of customizing large-scale text-to-image models. These models bind a unique subject desired by a user to a specific token using the token to generate the subject in various contexts. However models from previous studies also bind elements unrelated to the subject's identity ...
Seunghwan Choi; Jooyeol Yun; Jeonghoon Park; Jaegul Choo
Korea Advanced Institute of Science and Technology (KAIST); Korea Advanced Institute of Science and Technology (KAIST); Korea Advanced Institute of Science and Technology (KAIST); Korea Advanced Institute of Science and Technology (KAIST)
Poster
main
https://openaccess.thecvf.com/content/WACV2025/html/Choi_Disentangling_Subject-Irrelevant_Elements_in_Personalized_Text-to-Image_Diffusion_via_Filtered_Self-Distillation_WACV_2025_paper.html
0
Disentangling Subject-Irrelevant Elements in Personalized Text-to-Image Diffusion via Filtered Self-Distillation Recent research has unveiled the development of customizing large-scale text-to-image models. These models bind a unique subject desired by a user to a specific token using the token to generate the subject ...
[ -0.010779221542179585, -0.01665298454463482, 0.0020370902493596077, -0.0011687006335705519, -0.01865353435277939, -0.050205595791339874, 0.016853952780365944, 0.004818677436560392, 0.001766467816196382, 0.004250028170645237, -0.0011007594875991344, -0.0126336133107543, -0.032008808106184006,...
wacv_2025_58811c21a5
58811c21a5
wacv
2,025
Distillation of Diffusion Features for Semantic Correspondence
Semantic correspondence the task of determining relationships between different parts of images underpins various applications including 3D reconstruction image-to-image translation object tracking and visual place recognition. Recent studies have begun to explore representations learned in large generative image model...
Frank Fundel; Johannes Schusterbauer; Vincent Tao Hu; Björn Ommer
CompVis@LMU Munich; CompVis@LMU Munich; CompVis@LMU Munich; CompVis@LMU Munich
Poster
main
https://compvis.github.io/distilldift
https://openaccess.thecvf.com/content/WACV2025/html/Fundel_Distillation_of_Diffusion_Features_for_Semantic_Correspondence_WACV_2025_paper.html
1
2412.03512
Distillation of Diffusion Features for Semantic Correspondence Semantic correspondence the task of determining relationships between different parts of images underpins various applications including 3D reconstruction image-to-image translation object tracking and visual place recognition. Recent studies have begun to ...
[ -0.05755101144313812, -0.014897957444190979, 0.007305193692445755, 0.0012465211329981685, -0.004350648727267981, -0.034100182354450226, 0.0010575137566775084, -0.012745823711156845, 0.035009272396564484, -0.0010720081627368927, -0.020705007016658783, 0.009030611254274845, -0.0069897952489554...
wacv_2025_79f2c6f217
79f2c6f217
wacv
2,025
Distilling Aggregated Knowledge for Weakly-Supervised Video Anomaly Detection
Video anomaly detection aims to develop automated models capable of identifying abnormal events in surveillance videos. The benchmark setup for this task is extremely challenging due to: i) the limited size of the training sets ii) weak supervision provided in terms of video-level labels and iii) intrinsic class imbala...
Jash Dalvi; Ali Dabouei; Gunjan Dhanuka; Min Xu
K J Somaiya Institute of Technology; Carnegie Mellon University; Carnegie Mellon University; Carnegie Mellon University
Poster
main
https://openaccess.thecvf.com/content/WACV2025/html/Dalvi_Distilling_Aggregated_Knowledge_for_Weakly-Supervised_Video_Anomaly_Detection_WACV_2025_paper.html
1
2406.02831
Distilling Aggregated Knowledge for Weakly-Supervised Video Anomaly Detection Video anomaly detection aims to develop automated models capable of identifying abnormal events in surveillance videos. The benchmark setup for this task is extremely challenging due to: i) the limited size of the training sets ii) weak super...
[ -0.038708750158548355, -0.07577666640281677, -0.031343646347522736, 0.006968880072236061, -0.024761665612459183, -0.011933333240449429, 0.013005468063056469, 0.020174792036414146, -0.014068281278014183, 0.03294718638062477, -0.029646873474121094, -0.005784869659692049, -0.019354375079274178,...
wacv_2025_d49a3c4c14
d49a3c4c14
wacv
2,025
Distribution Optimization under Gaussian Hypothesis for Domain Adaptive Semantic Segmentation
Domain adaptive semantic segmentation aims to transfer a model proficient in dense image classification from a source domain to a target domain. While various transfer methods have been explored in previous studies we argue that the modeling of categories within the model significantly affects its transferability. Buil...
Chen Liang; Weihua Chen; Xin Zhao; Junyan Wang; Lijun Cao; Junge Zhang
Institute of Automation, Chinese Academy of Sciences; Alibaba Group; University of Science and Technology Beijing; University of New South Wales; Jiangsu Jiyuan Medical Technology Co., Ltd; Institute of Automation, Chinese Academy of Sciences
Poster
main
https://openaccess.thecvf.com/content/WACV2025/html/Liang_Distribution_Optimization_under_Gaussian_Hypothesis_for_Domain_Adaptive_Semantic_Segmentation_WACV_2025_paper.html
0
Distribution Optimization under Gaussian Hypothesis for Domain Adaptive Semantic Segmentation Domain adaptive semantic segmentation aims to transfer a model proficient in dense image classification from a source domain to a target domain. While various transfer methods have been explored in previous studies we argue th...
[ -0.05097115412354469, -0.05371547490358353, -0.011882902123034, 0.003508154535666108, -0.0022103204391896725, -0.011855458840727806, -0.00098967005033046, -0.016145743429660797, 0.04394569620490074, 0.028650691732764244, -0.055179111659526825, 0.007519434206187725, -0.008795542642474174, 0...
wacv_2025_503caa90ca
503caa90ca
wacv
2,025
DivAvatar: Diverse 3D Avatar Generation with a Single Prompt
Text-to-Avatar generation has recently made significant strides due to advancements in diffusion models. However most existing work remains constrained by limited diversity producing avatars with subtle differences in appearance for a given text prompt. We design DivAvatar a novel framework that generates diverse avata...
Weijing Tao; Biwen Lei; Kunhao Liu; Shijian Lu; Miaomiao Cui; Xuansong Xie
Nanyang Technological University; Alibaba Group; Nanyang Technological University; Nanyang Technological University; Alibaba Group; Alibaba Group
Poster
main
https://openaccess.thecvf.com/content/WACV2025/html/Tao_DivAvatar_Diverse_3D_Avatar_Generation_with_a_Single_Prompt_WACV_2025_paper.html
1
2402.17292
DivAvatar: Diverse 3D Avatar Generation with a Single Prompt Text-to-Avatar generation has recently made significant strides due to advancements in diffusion models. However most existing work remains constrained by limited diversity producing avatars with subtle differences in appearance for a given text prompt. We de...
[ -0.06334584951400757, -0.01765643246471882, -0.01645217277109623, 0.018833529204130173, -0.013473216444253922, -0.024936316534876823, -0.0070082456804811954, 0.02087986283004284, 0.00012931073433719575, -0.0012099184095859528, 0.011933937668800354, -0.037485964596271515, -0.02495442517101764...
wacv_2025_9ca8acb4f9
9ca8acb4f9
wacv
2,025
Divergent Domains Convergent Grading: Enhancing Generalization in Diabetic Retinopathy Grading
Diabetic Retinopathy (DR) constitutes 5% of global blindness cases. While numerous deep learning approaches have sought to enhance traditional DR grading methods they often falter when confronted with new out-of-distribution data thereby impeding their widespread application. In this study we introduce a novel deep lea...
Sharon Chokuwa; Muhammad Haris Khan
;
Poster
main
https://openaccess.thecvf.com/content/WACV2025/html/Chokuwa_Divergent_Domains_Convergent_Grading_Enhancing_Generalization_in_Diabetic_Retinopathy_Grading_WACV_2025_paper.html
1
2411.02614
Divergent Domains Convergent Grading: Enhancing Generalization in Diabetic Retinopathy Grading Diabetic Retinopathy (DR) constitutes 5% of global blindness cases. While numerous deep learning approaches have sought to enhance traditional DR grading methods they often falter when confronted with new out-of-distribution ...
[ -0.03810039907693863, -0.02601439319550991, -0.019805574789643288, -0.021426867693662643, -0.01919759064912796, 0.016157664358615875, -0.022016430273652077, 0.02105839177966118, 0.03323651850223541, 0.012141278013586998, -0.03657122328877449, -0.010768705978989601, -0.023490333929657936, 0...
wacv_2025_0a080afcaf
0a080afcaf
wacv
2,025
DocMatcher: Document Image Dewarping via Structural and Textual Line Matching
Document image dewarping is a crucial step in the digitization of physical documents as it aims to remove the distortions induced by challenging environment settings and document sheet deformations often encountered when using smartphone cameras for image capture. Recently deep learning-based methods were combined with...
Felix Hertlein; Alexander Naumann; York Sure-Vetter
FZI and KIT; FZI and KIT; FZI and KIT
Poster
main
https://felixhertlein.github.io/doc-matcher
https://openaccess.thecvf.com/content/WACV2025/html/Hertlein_DocMatcher_Document_Image_Dewarping_via_Structural_and_Textual_Line_Matching_WACV_2025_paper.html
0
DocMatcher: Document Image Dewarping via Structural and Textual Line Matching Document image dewarping is a crucial step in the digitization of physical documents as it aims to remove the distortions induced by challenging environment settings and document sheet deformations often encountered when using smartphone came...
[ -0.02124006859958172, -0.04452034831047058, -0.01961882784962654, -0.030275294557213783, 0.00043178044143132865, 0.023444226011633873, -0.0013423048658296466, 0.013908054679632187, 0.009536171332001686, 0.004713437054306269, -0.028471892699599266, -0.02752465195953846, -0.0048910449258983135...
wacv_2025_28c5103d06
28c5103d06
wacv
2,025
DocTTT: Test-Time Training for Handwritten Document Recognition using Meta-Auxiliary Learning
Despite recent significant advancements in Handwritten Document Recognition (HDR) the efficient and accurate recognition of text against complex backgrounds diverse handwriting styles and varying document layouts remains a practical challenge. Moreover this issue is seldom addressed in academic research particularly in...
Wenhao Gu; Li Gu; Ziqiang Wang; Ching Y Suen; Yang Wang
Department of Computer Science and Software Engineering, Concordia University; Department of Computer Science and Software Engineering, Concordia University; Department of Computer Science and Software Engineering, Concordia University; Department of Computer Science and Software Engineering, Concordia University; Depa...
Poster
main
https://openaccess.thecvf.com/content/WACV2025/html/Gu_DocTTT_Test-Time_Training_for_Handwritten_Document_Recognition_using_Meta-Auxiliary_Learning_WACV_2025_paper.html
1
2501.12898
DocTTT: Test-Time Training for Handwritten Document Recognition using Meta-Auxiliary Learning Despite recent significant advancements in Handwritten Document Recognition (HDR) the efficient and accurate recognition of text against complex backgrounds diverse handwriting styles and varying document layouts remains a pra...
[ -0.02835266850888729, -0.03568994253873825, -0.006961350329220295, -0.002113994909450412, 0.0022498704493045807, -0.022174866870045662, 0.008292929269373417, 0.026522880420088768, 0.024149589240550995, 0.025236593559384346, -0.046306341886520386, -0.05387913063168526, -0.014420910738408566, ...
wacv_2025_08a257a1cc
08a257a1cc
wacv
2,025
Domain Generalization using Large Pretrained Models with Mixture-of-Adapters
Learning robust vision models that perform well in out-of-distribution (OOD) situations is an important task for model deployment in real-world settings. Despite extensive research in this field many proposed methods have only shown minor performance improvements compared to the simplest empirical risk minimization (ER...
Gyuseong Lee; Wooseok Jang; Jinhyeon Kim; Jaewoo Jung; Seungryong Kim
Korea University; Korea University; Korea University; KAIST; KAIST
Poster
main
https://openaccess.thecvf.com/content/WACV2025/html/Lee_Domain_Generalization_using_Large_Pretrained_Models_with_Mixture-of-Adapters_WACV_2025_paper.html
4
2310.11031
Domain Generalization using Large Pretrained Models with Mixture-of-Adapters Learning robust vision models that perform well in out-of-distribution (OOD) situations is an important task for model deployment in real-world settings. Despite extensive research in this field many proposed methods have only shown minor perf...
[ -0.01613616570830345, -0.017636563628911972, -0.03821868821978569, -0.0266205333173275, -0.020434848964214325, -0.01529852207750082, 0.02201808989048004, 0.004080065991729498, 0.014902711845934391, 0.025773683562874794, -0.06778478622436523, -0.023895885795354843, -0.03310077264904976, 0.0...
wacv_2025_ba11284fa4
ba11284fa4
wacv
2,025
Domain-Generalized Object Anti-Spoofing: Bridging Gaps and Patch Selection for Robust Detection Across Domains
In online applications significant risks exist in peer-to-peer transactions due to malicious behaviors of arbitrary users such as taking advantage of manipulated images or impersonating others using recaptured images. Moreover recent advancements in display screens and imaging devices have made it increasingly challeng...
Geonu Lee; Yonghyun Jeong; Haneol Jang; Youngjoon Yoo
NA VER Cloud + SNUAILAB; NA VER Cloud + SNUAILAB; Hanbat National University; SNUAILAB + ChungAng University
Poster
main
https://openaccess.thecvf.com/content/WACV2025/html/Lee_Domain-Generalized_Object_Anti-Spoofing_Bridging_Gaps_and_Patch_Selection_for_Robust_WACV_2025_paper.html
0
Domain-Generalized Object Anti-Spoofing: Bridging Gaps and Patch Selection for Robust Detection Across Domains In online applications significant risks exist in peer-to-peer transactions due to malicious behaviors of arbitrary users such as taking advantage of manipulated images or impersonating others using recaptured...
[ -0.05324043333530426, -0.03635433688759804, -0.041351430118083954, -0.020471373572945595, -0.001820503268390894, -0.04060836881399155, 0.014053170569241047, -0.0013026815140619874, 0.025914305821061134, 0.006822242867201567, -0.03410657122731209, 0.015799367800354958, -0.012446298263967037, ...
wacv_2025_25580ffaf2
25580ffaf2
wacv
2,025
Domain-Guided Weight Modulation for Semi-Supervised Domain Generalization
Unarguably deep learning models capable of generalizing to unseen domain data while leveraging a few labels are of great practical significance due to low developmental costs. In search of this endeavor we study the challenging problem of semi-supervised domain generalization (SSDG) where the goal is to learn a domain-...
Chamuditha Jayanga Galappaththige; Zachary Izzo; Xilin He; Honglu Zhou; Muhammad Haris Khan
MBZUAI, UAE; NEC Labs, USA; Shenzhen University, China; Salesforce AI Research, USA; MBZUAI, UAE
Poster
main
github.com/DGWM
https://openaccess.thecvf.com/content/WACV2025/html/Galappaththige_Domain-Guided_Weight_Modulation_for_Semi-Supervised_Domain_Generalization_WACV_2025_paper.html
3
2409.03509
Domain-Guided Weight Modulation for Semi-Supervised Domain Generalization Unarguably deep learning models capable of generalizing to unseen domain data while leveraging a few labels are of great practical significance due to low developmental costs. In search of this endeavor we study the challenging problem of semi-su...
[ -0.08579578995704651, -0.01985611394047737, -0.05078966170549393, -0.037793587893247604, -0.011747145093977451, -0.012968921102583408, 0.0034277590457350016, 0.011991499923169613, -0.002182226860895753, 0.04405631870031357, -0.019222602248191833, 0.005850946996361017, -0.007253726478666067, ...
wacv_2025_488a0e9dde
488a0e9dde
wacv
2,025
DrIFT: Autonomous Drone Dataset with Integrated Real and Synthetic Data Flexible Views and Transformed Domains
Dependable visual drone detection is crucial for the secure integration of drones into the airspace. However drone detection accuracy is significantly affected by domain shifts due to environmental changes varied points of view and background shifts. To address these challenges we present the DrIFT dataset specifically...
Fardad Dadboud; Hamid Azad; Varun Mehta; Miodrag Bolic; Iraj Mantegh
University of Ottawa; University of Ottawa; National Research Council Canada; University of Ottawa; National Research Council Canada
Poster
main
https://github.com/CARG-uOttawa/DrIFT.git
https://openaccess.thecvf.com/content/WACV2025/html/Dadboud_DrIFT_Autonomous_Drone_Dataset_with_Integrated_Real_and_Synthetic_Data_WACV_2025_paper.html
0
2412.04789
DrIFT: Autonomous Drone Dataset with Integrated Real and Synthetic Data Flexible Views and Transformed Domains Dependable visual drone detection is crucial for the secure integration of drones into the airspace. However drone detection accuracy is significantly affected by domain shifts due to environmental changes var...
[ -0.07002005726099014, -0.028662551194429398, -0.01319843903183937, -0.006244084797799587, -0.04290391877293587, -0.013872746378183365, -0.02289048209786415, 0.005776565056294203, 0.04657214879989624, 0.02319616824388504, -0.03605295717716217, -0.012946697883307934, -0.022440943866968155, 0...
wacv_2025_438df2ca7a
438df2ca7a
wacv
2,025
DragText: Rethinking Text Embedding in Point-Based Image Editing
Point-based image editing enables accurate and flexible control through content dragging. However the role of text embedding during the editing process has not been thoroughly investigated. A significant aspect that remains unexplored is the interaction between text and image embeddings. During the progressive editing ...
Gayoon Choi; Taejin Jeong; Sujung Hong; Seong Jae Hwang
Yonsei University; Yonsei University; Yonsei University; Yonsei University
Poster
main
https://openaccess.thecvf.com/content/WACV2025/html/Choi_DragText_Rethinking_Text_Embedding_in_Point-Based_Image_Editing_WACV_2025_paper.html
1
2407.17843
DragText: Rethinking Text Embedding in Point-Based Image Editing Point-based image editing enables accurate and flexible control through content dragging. However the role of text embedding during the editing process has not been thoroughly investigated. A significant aspect that remains unexplored is the interaction b...
[ -0.017636137083172798, -0.01809733361005783, -0.013042625039815903, -0.011935754679143429, -0.004976305644959211, -0.05995548889040947, 0.014490780420601368, 0.004625796806067228, 0.02684161253273487, 0.01786673627793789, -0.01786673627793789, -0.03508779779076576, -0.004787215497344732, -...
wacv_2025_a0f0e6dd04
a0f0e6dd04
wacv
2,025
DragonTrack: Transformer-Enhanced Graphical Multi-Person Tracking in Complex Scenarios
This paper introduces the dynamic robust adaptive graph-based tracker (DragonTrack) as a novel end-to-end framework for multi-person tracking (MPT) by integrating a detection transformer model for object detection and feature extraction with a graph convolutional network for re-identification. DragonTrack leverages enc...
Bishoy Galoaa; Somaieh Amraee; Sarah Ostadabbas
Augmented Cognition Lab (ACLab), Department of Electrical and Computer Engineering, Northeastern University, Boston, MA, USA; Augmented Cognition Lab (ACLab), Department of Electrical and Computer Engineering, Northeastern University, Boston, MA, USA; Augmented Cognition Lab (ACLab), Department of Electrical and Comput...
Poster
main
https://github.com/ostadabbas/DragonTrack
https://openaccess.thecvf.com/content/WACV2025/html/Galoaa_DragonTrack_Transformer-Enhanced_Graphical_Multi-Person_Tracking_in_Complex_Scenarios_WACV_2025_paper.html
2
DragonTrack: Transformer-Enhanced Graphical Multi-Person Tracking in Complex Scenarios This paper introduces the dynamic robust adaptive graph-based tracker (DragonTrack) as a novel end-to-end framework for multi-person tracking (MPT) by integrating a detection transformer model for object detection and feature extract...
[ -0.03179366886615753, -0.04187004640698433, -0.020599765703082085, -0.019202854484319687, -0.01642766036093235, -0.04198179766535759, -0.010607202537357807, -0.027900947257876396, 0.010793457739055157, 0.0014551144558936357, -0.03279944509267807, 0.007566595450043678, -0.015971336513757706, ...
wacv_2025_08d319eb2c
08d319eb2c
wacv
2,025
DreaMo: Articulated 3D Reconstruction from a Single Casual Video
Articulated 3D reconstruction has valuable applications in various domains yet it remains costly and demands intensive work from domain experts. Recent advancements in template-free learning methods show promising results with monocular videos. Nevertheless these approaches necessitate a comprehensive coverage of all v...
Tao Tu; Ming-Feng Li; Chieh Hubert Lin; Yen-Chi Cheng; Min Sun; Ming-Hsuan Yang
National Tsing Hua University; Carnegie Mellon University; University of California, Merced; University of Illinois Urbana-Champaign; National Tsing Hua University+Amazon; University of California, Merced
Poster
main
https://openaccess.thecvf.com/content/WACV2025/html/Tu_DreaMo_Articulated_3D_Reconstruction_from_a_Single_Casual_Video_WACV_2025_paper.html
4
2312.02617
DreaMo: Articulated 3D Reconstruction from a Single Casual Video Articulated 3D reconstruction has valuable applications in various domains yet it remains costly and demands intensive work from domain experts. Recent advancements in template-free learning methods show promising results with monocular videos. Neverthele...
[ -0.030544569715857506, -0.0036401101388037205, 0.00757142947986722, -0.043856970965862274, -0.015540381893515587, 0.04455957189202309, 0.04733298718929291, 0.021318335086107254, -0.022982385009527206, -0.0024151841644197702, 0.019155068323016167, -0.023222748190164566, -0.013978023082017899,...
wacv_2025_c285da48c5
c285da48c5
wacv
2,025
DreamBlend: Advancing Personalized Fine-Tuning of Text-to-Image Diffusion Models
Given a small number of images of a subject personalized image generation techniques can fine-tune large pre-trained text-to-image diffusion models to generate images of the subject in novel contexts conditioned on text prompts. In doing so a trade-off is made between prompt fidelity subject fidelity and diversity. As ...
Shwetha Ram; Tal Neiman; Qianli Feng; Andrew M Stuart; Son Tran; Trishul A Chilimbi
Amazon; Amazon; Amazon; Amazon; Amazon; Amazon
Poster
main
https://openaccess.thecvf.com/content/WACV2025/html/Ram_DreamBlend_Advancing_Personalized_Fine-Tuning_of_Text-to-Image_Diffusion_Models_WACV_2025_paper.html
0
2411.19390
DreamBlend: Advancing Personalized Fine-Tuning of Text-to-Image Diffusion Models Given a small number of images of a subject personalized image generation techniques can fine-tune large pre-trained text-to-image diffusion models to generate images of the subject in novel contexts conditioned on text prompts. In doing s...
[ -0.0826767235994339, -0.020249594002962112, -0.022803597152233124, -0.023952897638082504, -0.016482440754771233, -0.027017701417207718, -0.0049027735367417336, 0.01328081451356411, -0.011638956144452095, -0.00007696214015595615, -0.028185246512293816, -0.022986026480793953, -0.02099755220115...
wacv_2025_7e37400bef
7e37400bef
wacv
2,025
Dropout Connects Transformers and CNNs: Transfer General Knowledge for Knowledge Distillation
Thanks to their long-range dependencies transformers obtain state-of-the-art performance in diverse research fields such as computer vision and audio processing. In practical scenarios convolutional neural networks (CNNs) are used more than Transformers due to their low complexity. So Transformer-to-CNN knowledge disti...
Bokyeung Lee; Jonghwan Hong; Hyunuk Shin; Bonwha Ku; Hanseok Ko
Department of Electrical and Computer Engineering, Korea University; Department of Electrical and Computer Engineering, Korea University; Department of Electrical and Computer Engineering, Korea University; Department of Electrical and Computer Engineering, Korea University; Department of Electrical and Computer Engine...
Poster
main
https://openaccess.thecvf.com/content/WACV2025/html/Lee_Dropout_Connects_Transformers_and_CNNs_Transfer_General_Knowledge_for_Knowledge_WACV_2025_paper.html
0
Dropout Connects Transformers and CNNs: Transfer General Knowledge for Knowledge Distillation Thanks to their long-range dependencies transformers obtain state-of-the-art performance in diverse research fields such as computer vision and audio processing. In practical scenarios convolutional neural networks (CNNs) are ...
[ -0.06104210391640663, -0.03709791600704193, -0.03704295679926872, -0.02040843479335308, -0.011569053865969181, -0.02343122847378254, -0.03220648691058159, -0.01319953054189682, -0.01426208857446909, 0.004987608641386032, -0.005051728803664446, -0.028505856171250343, -0.03227976709604263, 0...
wacv_2025_5cf38ff6f3
5cf38ff6f3
wacv
2,025
Dropout the High-Rate Downsampling: A Novel Design Paradigm for UHD Image Restoration
With the popularization of high-end mobile devices Ultra-high-definition (UHD) images have become ubiquitous in our lives. The restoration of UHD images is a highly challenging problem due to the exaggerated pixel count which often leads to memory overflow during processing. Existing methods either downsample UHD image...
Chen Wu; Ling Wang; Long Peng; Dianjie Lu; Zhuoran Zheng
University of Science and Technology of China; The Hong Kong University of Science and Technology (Guangzhou); University of Science and Technology of China; Shandong Normal University; Sun Yat-sen University
Poster
main
https://openaccess.thecvf.com/content/WACV2025/html/Wu_Dropout_the_High-Rate_Downsampling_A_Novel_Design_Paradigm_for_UHD_WACV_2025_paper.html
2
2411.06456
Dropout the High-Rate Downsampling: A Novel Design Paradigm for UHD Image Restoration With the popularization of high-end mobile devices Ultra-high-definition (UHD) images have become ubiquitous in our lives. The restoration of UHD images is a highly challenging problem due to the exaggerated pixel count which often le...
[ -0.06588158011436462, -0.0017726297955960035, -0.045972514897584915, 0.01578451506793499, -0.02267112396657467, -0.008420570753514767, 0.012540486641228199, 0.015256201848387718, -0.012234620749950409, -0.0019186110002920032, -0.04200553148984909, 0.025192197412252426, 0.00029254183755256236...
wacv_2025_f30e4bd39d
f30e4bd39d
wacv
2,025
Dual-Representation Interaction Driven Image Quality Assessment with Restoration Assistance
No-Reference Image Quality Assessment for distorted images has always been a challenging problem due to image content variance and distortion diversity. Previous IQA models mostly encode explicit single-quality features of synthetic images to obtain quality-aware representations for quality score prediction. However pe...
Jingtong Yue; Xin Lin; Zijiu Yang; Chao Ren
Sichuan University; Sichuan University; Sichuan University; Sichuan University
Poster
main
https://github.com/Jingtong0527/DRI-IQA
https://openaccess.thecvf.com/content/WACV2025/html/Yue_Dual-Representation_Interaction_Driven_Image_Quality_Assessment_with_Restoration_Assistance_WACV_2025_paper.html
0
2411.17390
Dual-Representation Interaction Driven Image Quality Assessment with Restoration Assistance No-Reference Image Quality Assessment for distorted images has always been a challenging problem due to image content variance and distortion diversity. Previous IQA models mostly encode explicit single-quality features of synth...
[ -0.03956208750605583, -0.0037171642761677504, 0.004304208792746067, -0.013487929478287697, -0.03272419422864914, 0.0009069833904504776, 0.018306387588381767, 0.04230475798249245, 0.007838214747607708, 0.03888581320643425, -0.04256775230169296, 0.011421533301472664, 0.003071415703743696, 0....
wacv_2025_94c8782e9f
94c8782e9f
wacv
2,025
Dual-Schedule Inversion: Training- and Tuning-Free Inversion for Real Image Editing
Text-conditional image editing is a practical AIGC task that has recently emerged with great commercial and academic value. For real image editing most diffusion model-based methods use DDIM Inversion as the first stage before editing. However DDIM Inversion often results in reconstruction failure leading to unsatisfac...
Jiancheng Huang; Yi Huang; Jianzhuang Liu; Donghao Zhou; Yifan Liu; Shifeng Chen
Shenzhen Institute of Advanced Technology, Chinese Academy of Sciences; Shenzhen Institute of Advanced Technology, Chinese Academy of Sciences; Shenzhen Institute of Advanced Technology, Chinese Academy of Sciences; Shenzhen Institute of Advanced Technology, Chinese Academy of Sciences; Shenzhen Institute of Advanced T...
Poster
main
https://openaccess.thecvf.com/content/WACV2025/html/Huang_Dual-Schedule_Inversion_Training-_and_Tuning-Free_Inversion_for_Real_Image_Editing_WACV_2025_paper.html
0
Dual-Schedule Inversion: Training- and Tuning-Free Inversion for Real Image Editing Text-conditional image editing is a practical AIGC task that has recently emerged with great commercial and academic value. For real image editing most diffusion model-based methods use DDIM Inversion as the first stage before editing. ...
[ -0.06916461884975433, -0.013998042792081833, -0.006697539240121841, 0.002379389014095068, -0.012402507476508617, 0.013088958337903023, -0.0016187266446650028, 0.03710548207163811, 0.028292929753661156, 0.026363443583250046, -0.028886616230010986, -0.03584389388561249, 0.011716055683791637, ...
wacv_2025_f594897f35
f594897f35
wacv
2,025
DualCIR: Enhancing Training-Free Composed Image Retrieval via Dual-Directional Descriptions
Integrating language and images for Composed Image Retrieval (CIR) allows for a flexible representation of search intent making it a focal point in multi-modal research. Traditional CIR methods train models on complex triplet datasets. In contrast Zero-Shot Composed Image Retrieval (ZS-CIR) eliminates the need for cons...
Jingjiao Zhao; Jiaju Li; Dongze Lian; Liguo Sun; Pin Lv
;;;;
Poster
main
https://openaccess.thecvf.com/content/WACV2025/html/Zhao_DualCIR_Enhancing_Training-Free_Composed_Image_Retrieval_via_Dual-Directional_Descriptions_WACV_2025_paper.html
0
DualCIR: Enhancing Training-Free Composed Image Retrieval via Dual-Directional Descriptions Integrating language and images for Composed Image Retrieval (CIR) allows for a flexible representation of search intent making it a focal point in multi-modal research. Traditional CIR methods train models on complex triplet da...
[ -0.07552085816860199, -0.0330541655421257, 0.009964046999812126, 0.043312352150678635, -0.0008031444740481675, 0.039120834320783615, 0.03520507365465164, -0.0025002036709338427, 0.0024036886170506477, 0.009095409885048866, -0.009798592887818813, -0.02343940921127796, -0.01803455874323845, ...
wacv_2025_03084e39b0
03084e39b0
wacv
2,025
DyRoNet: Dynamic Routing and Low-Rank Adapters for Autonomous Driving Streaming Perception
The advancement of autonomous driving systems hinges on the ability to achieve low-latency and high-accuracy perception. To address this critical need this paper introduces Dynamic Routering Network (DyRoNet) a low-rank enhanced dynamic routing framework designed for streaming perception in autonomous driving systems. ...
Xiang Huang; Zhi-Qi Cheng; Jun-Yan He; Chenyang Li; Wangmeng Xiang; Baigui Sun
Institute of Artificial Intelligence, Southwest Jiaotong University, Chengdu, China; School of Engineering and Technology, University of Washington, Tacoma, WA, USA + CMU; Institute for Intelligent Computing, Alibaba Group, China; Institute for Intelligent Computing, Alibaba Group, China; Institute for Intelligent Comp...
Poster
main
https://openaccess.thecvf.com/content/WACV2025/html/Huang_DyRoNet_Dynamic_Routing_and_Low-Rank_Adapters_for_Autonomous_Driving_Streaming_WACV_2025_paper.html
1
2403.05050
DyRoNet: Dynamic Routing and Low-Rank Adapters for Autonomous Driving Streaming Perception The advancement of autonomous driving systems hinges on the ability to achieve low-latency and high-accuracy perception. To address this critical need this paper introduces Dynamic Routering Network (DyRoNet) a low-rank enhanced ...
[ -0.050331152975559235, -0.04786180332303047, 0.0012369443429633975, 0.01611432060599327, -0.04230576753616333, -0.007362655829638243, 0.031193694099783897, 0.000022678495952277444, 0.023876430466771126, 0.022060731425881386, -0.01786646991968155, -0.022151516750454903, -0.015215550549328327,...
wacv_2025_e4ff319801
e4ff319801
wacv
2,025
Dynamic Adapter Tuning for Long-Tailed Class-Incremental Learning
Long-tailed class-incremental learning (LT-CIL) aims to learn new classes continuously from a long-tailed data stream while simultaneously dealing with challenges such as imbalanced learning of tail classes and catastrophic forgetting. To address these challenges most existing methods employ a two-stage strategy by ini...
Yanan Gu; Muli Yang; Xu Yang; Kun Wei; Hongyuan Zhu; Gabriel James Goenawan; Cheng Deng
Norinco Group Testing and Research Institute, Xi’an, China; Institute for Infocomm Research (I2R), A*STAR, Singapore; School of Electronic Engineering, Xidian University, Xi’an, China; School of Electronic Engineering, Xidian University, Xi’an, China; Institute for Infocomm Research (I2R), A*STAR, Singapore; Institute ...
Poster
main
https://openaccess.thecvf.com/content/WACV2025/html/Gu_Dynamic_Adapter_Tuning_for_Long-Tailed_Class-Incremental_Learning_WACV_2025_paper.html
0
Dynamic Adapter Tuning for Long-Tailed Class-Incremental Learning Long-tailed class-incremental learning (LT-CIL) aims to learn new classes continuously from a long-tailed data stream while simultaneously dealing with challenges such as imbalanced learning of tail classes and catastrophic forgetting. To address these c...
[ -0.029067644849419594, -0.06098966300487518, -0.019998684525489807, 0.02202203869819641, -0.04270721599459648, -0.01368473656475544, 0.02231108956038952, 0.04357437044382095, -0.06265170127153397, 0.0247860848903656, -0.026809437200427055, -0.03743204474449158, -0.03974444791674614, 0.0437...
wacv_2025_1c397e7d0a
1c397e7d0a
wacv
2,025
Dynamic Attention-Guided Diffusion for Image Super-Resolution
Diffusion models in image Super-Resolution (SR) treat all image regions uniformly which risks compromising the overall image quality by potentially introducing artifacts during denoising of less-complex regions. To address this we propose "You Only Diffuse Areas" (YODA) a dynamic attention-guided diffusion process for ...
Brian B. Moser; Stanislav Frolov; Federico Raue; Sebastian Palacio; Andreas Dengel
German Research Center for Artificial Intelligence, Germany + RPTU Kaiserslautern-Landau, Germany + Equal Contribution; German Research Center for Artificial Intelligence, Germany + RPTU Kaiserslautern-Landau, Germany + Equal Contribution; German Research Center for Artificial Intelligence, Germany; German Research Cen...
Poster
main
https://openaccess.thecvf.com/content/WACV2025/html/Moser_Dynamic_Attention-Guided_Diffusion_for_Image_Super-Resolution_WACV_2025_paper.html
1
2308.07977
Dynamic Attention-Guided Diffusion for Image Super-Resolution Diffusion models in image Super-Resolution (SR) treat all image regions uniformly which risks compromising the overall image quality by potentially introducing artifacts during denoising of less-complex regions. To address this we propose "You Only Diffuse A...
[ -0.0803823471069336, -0.002999945543706417, -0.028227373957633972, 0.009840901009738445, -0.028659150004386902, 0.006166304927319288, -0.027705645188689232, 0.01836848258972168, 0.023837648332118988, 0.01023669634014368, -0.00355990557000041, -0.012944293208420277, -0.018332501873373985, 0...
wacv_2025_be782f64ae
be782f64ae
wacv
2,025
ECF-YOLOv7-Tiny: Improving Feature Fusion and the Receptive Field for Lightweight Object Detectors
In this work we aim to increase the efficiency and the detection performance of lightweight object detectors with focus on feature fusion and receptive field of the models. For improved feature fusion we introduce the Convolutional Squeeze-and-Excitation (CSE) module which requires only minimal additional computation. ...
Dan-Sebastian Bacea; Florin Oniga
Technical University of Cluj-Napoca, Romania+Stratec Biomedical, Romania; Technical University of Cluj-Napoca, Romania
Poster
main
https://github.com/dbacea/ecf-yolov7-tiny
https://openaccess.thecvf.com/content/WACV2025/html/Bacea_ECF-YOLOv7-Tiny_Improving_Feature_Fusion_and_the_Receptive_Field_for_Lightweight_WACV_2025_paper.html
0
ECF-YOLOv7-Tiny: Improving Feature Fusion and the Receptive Field for Lightweight Object Detectors In this work we aim to increase the efficiency and the detection performance of lightweight object detectors with focus on feature fusion and receptive field of the models. For improved feature fusion we introduce the Con...
[ -0.010613745078444481, -0.016664400696754456, -0.010367338545620441, 0.025516794994473457, 0.024677185341715813, -0.031120266765356064, -0.02285194955766201, 0.03575636446475983, 0.04139634594321251, 0.05563318356871605, -0.015058194287121296, 0.008765693753957748, -0.030809978023171425, 0...
wacv_2025_885e1d7830
885e1d7830
wacv
2,025
EDMB: Edge Detector with Mamba
Transformer-based models have made significant progress in edge detection but their high computational cost is prohibitive. Recently vision Mamba have shown excellent ability in efficiently capturing long-range dependencies. Drawing inspiration from this we propose a novel edge detector with Mamba termed EDMB to effici...
Yachuan Li; Xavier Soria Poma; Yun Bai; Qian Xiao; Chaozhi Yang; Guanlin Li; Zongmin Li
China University of Petroleum (East China); Polytechnic School of Chimborazo (ESPOCH); China University of Petroleum (East China); China University of Petroleum (East China); China University of Petroleum (East China); China University of Petroleum (East China); China University of Petroleum (East China)
Poster
main
https://github.com/Li-yachuan/EDMB
https://openaccess.thecvf.com/content/WACV2025/html/Li_EDMB_Edge_Detector_with_Mamba_WACV_2025_paper.html
0
2501.04846
EDMB: Edge Detector with Mamba Transformer-based models have made significant progress in edge detection but their high computational cost is prohibitive. Recently vision Mamba have shown excellent ability in efficiently capturing long-range dependencies. Drawing inspiration from this we propose a novel edge detector w...
[ -0.051008567214012146, -0.028230922296643257, -0.03700439631938934, -0.015367490239441395, -0.024688147008419037, -0.04566657915711403, -0.03650358319282532, 0.01955019310116768, 0.021534889936447144, 0.015534427016973495, -0.05523763969540596, -0.014059815555810928, -0.005253882147371769, ...
wacv_2025_92dc84a9d0
92dc84a9d0
wacv
2,025
EI-Nexus: Towards Unmediated and Flexible Inter-Modality Local Feature Extraction and Matching for Event-Image Data
Event cameras with high temporal resolution and high dynamic range have limited research on the inter-modality local feature extraction and matching of event-image data. We propose EI-Nexus an unmediated and flexible framework that integrates two modality-specific keypoint extractors and a feature matcher. To achieve k...
Zhonghua Yi; Hao Shi; Qi Jiang; Kailun Yang; Ze Wang; Diyang Gu; Yufan Zhang; Kaiwei Wang
State Key Laboratory of Extreme Photonics and Instrumentation, Zhejiang University + Jiaxing Research Institute, Zhejiang University; State Key Laboratory of Extreme Photonics and Instrumentation, Zhejiang University + Ataraxia Technologies; State Key Laboratory of Extreme Photonics and Instrumentation, Zhejiang Univer...
Poster
main
https://openaccess.thecvf.com/content/WACV2025/html/Yi_EI-Nexus_Towards_Unmediated_and_Flexible_Inter-Modality_Local_Feature_Extraction_and_WACV_2025_paper.html
0
EI-Nexus: Towards Unmediated and Flexible Inter-Modality Local Feature Extraction and Matching for Event-Image Data Event cameras with high temporal resolution and high dynamic range have limited research on the inter-modality local feature extraction and matching of event-image data. We propose EI-Nexus an unmediated ...
[ -0.038626618683338165, -0.06259284913539886, -0.02274760603904724, 0.020901210606098175, -0.0348045788705349, -0.0077040898613631725, 0.004546751733869314, 0.023929301649332047, 0.00007508198905270547, 0.03153645619750023, -0.02607112005352974, -0.024944819509983063, -0.026200369000434875, ...
wacv_2025_df79bf6f55
df79bf6f55
wacv
2,025
ELBA: Learning by Asking for Embodied Visual Navigation and Task Completion
The research community has shown increasing interest in designing intelligent embodied agents that can assist humans in accomplishing tasks. Although there have been significant advancements in related vision-language benchmarks most prior work has focused on building agents that follow instructions rather than endowin...
Ying Shen; Daniel Bis; Cynthia Lu; Ismini Lourentzou
University of Illinois Urbana - Champaign; Amazon; Amazon; University of Illinois Urbana - Champaign
Poster
main
https://github.com/PLAN-Lab/ELBA
https://openaccess.thecvf.com/content/WACV2025/html/Shen_ELBA_Learning_by_Asking_for_Embodied_Visual_Navigation_and_Task_WACV_2025_paper.html
0
2302.04865
ELBA: Learning by Asking for Embodied Visual Navigation and Task Completion The research community has shown increasing interest in designing intelligent embodied agents that can assist humans in accomplishing tasks. Although there have been significant advancements in related vision-language benchmarks most prior work...
[ -0.03362836316227913, -0.01827428676187992, 0.017383530735969543, 0.014913287945091724, -0.013398082926869392, -0.04316037520766258, 0.04073604941368103, -0.011240064166486263, -0.016309112310409546, 0.013388900086283684, 0.005509835202246904, -0.03498745337128639, 0.0034872666001319885, 0...
wacv_2025_0f5cce53e9
0f5cce53e9
wacv
2,025
ELMGS: Enhancing Memory and Computation Scalability through Compression for 3D Gaussian Splatting
3D models have recently been popularized by the potentiality of end-to-end training offered first by Neural Radiance Fields and most recently by 3D Gaussian Splatting models. The latter has the big advantage of naturally providing fast training convergence and high editability. However as the research around these is s...
Muhammad Salman Ali; Sung-Ho Bae; Enzo Tartaglione
LTCI, T´el´ecom Paris, Institut Polytechnique de Paris, France+Kyung Hee University, Republic of Korea; Kyung Hee University, Republic of Korea; LTCI, T´el´ecom Paris, Institut Polytechnique de Paris, France
Poster
main
https://openaccess.thecvf.com/content/WACV2025/html/Ali_ELMGS_Enhancing_Memory_and_Computation_Scalability_through_Compression_for_3D_WACV_2025_paper.html
4
2410.23213
ELMGS: Enhancing Memory and Computation Scalability through Compression for 3D Gaussian Splatting 3D models have recently been popularized by the potentiality of end-to-end training offered first by Neural Radiance Fields and most recently by 3D Gaussian Splatting models. The latter has the big advantage of naturally p...
[ -0.0757049098610878, -0.004900704137980938, -0.009920268319547176, -0.03699300065636635, -0.02294919267296791, -0.05822329223155975, 0.037596445530653, -0.008091648109257221, 0.006880186498165131, 0.0367918536067009, -0.04571552202105522, 0.009005958214402199, 0.011639172211289406, 0.00646...
wacv_2025_1a7917fee5
1a7917fee5
wacv
2,025
ENAF: A Multi-Exit Network with an Adaptive Patch Fusion for Large Image Super Resolution
To accelerate single image super-resolution (SISR) networks on large images (2K-8K) many recent approaches decompose an image into small patches and dynamically determine an execution path according to its difficulty (referred to as a dynamic network). To quantify the hardness of a patch they mainly rely on a handcraft...
Manh Duong Nguyen; Tuan Nghia Nguyen; Xuan Truong Nguyen
Hanoi University of Science and Technology; Seoul National University; Seoul National University
Poster
main
https://github.com/nmduonggg/ENAF.git
https://openaccess.thecvf.com/content/WACV2025/html/Nguyen_ENAF_A_Multi-Exit_Network_with_an_Adaptive_Patch_Fusion_for_WACV_2025_paper.html
0
ENAF: A Multi-Exit Network with an Adaptive Patch Fusion for Large Image Super Resolution To accelerate single image super-resolution (SISR) networks on large images (2K-8K) many recent approaches decompose an image into small patches and dynamically determine an execution path according to its difficulty (referred to ...
[ -0.02010604925453663, -0.02413846179842949, 0.01211590226739645, 0.0037313804496079683, -0.015046867541968822, 0.000018960266970680095, 0.008214171044528484, 0.02533324994146824, 0.011677190661430359, 0.04278836399316788, -0.004914501216262579, 0.013852079398930073, -0.03573164343833923, 0...
wacv_2025_1ff5246acd
1ff5246acd
wacv
2,025
ERM++: An Improved Baseline for Domain Generalization
Domain Generalization (DG) aims to develop classifiers that can generalize to new unseen data distributions a critical capability when collecting new domain-specific data is impractical. A common DG baseline minimizes the empirical risk on the source domains. Recent studies have shown that this approach known as Empiri...
Piotr Teterwak; Kuniaki Saito; Theodoros Tsiligkaridis; Kate Saenko; Bryan Plummer
Boston University‡; Boston University‡; MIT Lincoln Laboratory†; Boston University‡; Boston University‡
Poster
main
https://github.com/piotr-teterwak/erm_plusplus
https://openaccess.thecvf.com/content/WACV2025/html/Teterwak_ERM_An_Improved_Baseline_for_Domain_Generalization_WACV_2025_paper.html
10
ERM++: An Improved Baseline for Domain Generalization Domain Generalization (DG) aims to develop classifiers that can generalize to new unseen data distributions a critical capability when collecting new domain-specific data is impractical. A common DG baseline minimizes the empirical risk on the source domains. Recent...
[ -0.06944956630468369, -0.03816108778119087, -0.040910135954618454, -0.009567401371896267, -0.01909863017499447, -0.0405484177172184, -0.0027128737419843674, 0.017018761485815048, 0.0019623120315372944, 0.013528196141123772, -0.046118851751089096, 0.019315659999847412, -0.046769943088293076, ...
wacv_2025_44d147284e
44d147284e
wacv
2,025
ERUP-YOLO: Enhancing Object Detection Robustness for Adverse Weather Condition by Unified Image-Adaptive Processing
We propose an image-adaptive object detection method for adverse weather conditions such as fog and low-light. Our framework employs differentiable preprocessing filters to perform image enhancement suitable for later-stage object detections. Our framework introduces two differentiable filters: a Bezier curve-based pix...
Yuka Ogino; Yuho Shoji; Takahiro Toizumi; Atsushi Ito
NEC Corporation; NEC Corporation; NEC Corporation; NEC Corporation
Poster
main
https://openaccess.thecvf.com/content/WACV2025/html/Ogino_ERUP-YOLO_Enhancing_Object_Detection_Robustness_for_Adverse_Weather_Condition_by_WACV_2025_paper.html
1
ERUP-YOLO: Enhancing Object Detection Robustness for Adverse Weather Condition by Unified Image-Adaptive Processing We propose an image-adaptive object detection method for adverse weather conditions such as fog and low-light. Our framework employs differentiable preprocessing filters to perform image enhancement suita...
[ -0.042361751198768616, -0.03888947516679764, -0.02134535275399685, 0.028637127950787544, -0.00916040875017643, -0.02130880206823349, 0.0015522438334301114, 0.04082664102315903, -0.004290086682885885, 0.021601203829050064, -0.03357141464948654, -0.03470447286963463, 0.011102141812443733, 0....
wacv_2025_5429d4a7c3
5429d4a7c3
wacv
2,025
EasyRet3D: Uncalibrated Multi-View Multi-Human 3D Reconstruction and Tracking
Current methods performing 3D human pose estimation from multi-view still bear several key limitations. First most methods require manual intrinsic and extrinsic camera calibration which is laborious and difficult in many settings. Second more accurate models rely on further training on the same datasets they evaluate ...
Junjie Oscar Yin; Ting Li; Jiahao Wang; Yi Zhang; Alan Yuille
Johns Hopkins University; Johns Hopkins University; Johns Hopkins University; ;
Poster
main
https://openaccess.thecvf.com/content/WACV2025/html/Yin_EasyRet3D_Uncalibrated_Multi-View_Multi-Human_3D_Reconstruction_and_Tracking_WACV_2025_paper.html
0
EasyRet3D: Uncalibrated Multi-View Multi-Human 3D Reconstruction and Tracking Current methods performing 3D human pose estimation from multi-view still bear several key limitations. First most methods require manual intrinsic and extrinsic camera calibration which is laborious and difficult in many settings. Second mor...
[ -0.006801102310419083, -0.023761991411447525, -0.0339403860270977, -0.007754744961857796, -0.0015455993125215173, 0.005661382805556059, 0.026981117203831673, -0.014030180871486664, -0.020784763619303703, -0.007996644824743271, -0.03140974044799805, -0.008126898668706417, 0.01579790934920311,...
wacv_2025_92858c95f5
92858c95f5
wacv
2,025
EchoDFKD: Data-Free Knowledge Distillation for Cardiac Ultrasound Segmentation using Synthetic Data
The application of machine learning to medical ultrasound videos of the heart i.e. echocardiography has recently gained traction with the availability of large public datasets. Traditional supervised tasks such as ejection fraction regression are now making way for approaches focusing more on the latent structure of da...
Grégoire Petit; Nathan Palluau; Axel Bauer; Clemens Dlaska
Digital Cardiology Lab, Medical University of Innsbruck, A-6020 Innsbruck, Austria + University Clinic of Internal Medicine III, Cardiology and Angiology, Medical University of Innsbruck, A-6020 Innsbruck, Austria; Digital Cardiology Lab, Medical University of Innsbruck, A-6020 Innsbruck, Austria + University Clinic of...
Poster
main
https://github.com/GregoirePetit/EchoDFKD
https://openaccess.thecvf.com/content/WACV2025/html/Petit_EchoDFKD_Data-Free_Knowledge_Distillation_for_Cardiac_Ultrasound_Segmentation_using_Synthetic_WACV_2025_paper.html
0
2409.07566
EchoDFKD: Data-Free Knowledge Distillation for Cardiac Ultrasound Segmentation using Synthetic Data The application of machine learning to medical ultrasound videos of the heart i.e. echocardiography has recently gained traction with the availability of large public datasets. Traditional supervised tasks such as ejecti...
[ -0.04086184874176979, -0.06174016743898392, -0.029043231159448624, 0.01620865799486637, 0.01224736962467432, -0.04041445627808571, 0.02803659625351429, 0.030198995023965836, 0.024904849007725716, 0.032324109226465225, -0.0023732774425297976, -0.004082456696778536, -0.005499199498444796, 0....
wacv_2025_97fc925d0c
97fc925d0c
wacv
2,025
EdgeGaussians - 3D Edge Mapping via Gaussian Splatting
With their meaningful geometry and omnipresence in the 3D world edges are extremely useful primitives in computer vision. Methods for 3D edge reconstruction have 1) either focused on reconstructing 3D edges by triangulating tracks of 2D line segments across images or 2) more recently learning a 3D edge distance field f...
Kunal Chelani; Assia Benbihi; Torsten Sattler; Fredrik Kahl
Chalmers University of Technology; Czech Institute of Informatics, Robotics and Cybernetics, Czech Technical University in Prague; Czech Institute of Informatics, Robotics and Cybernetics, Czech Technical University in Prague; Chalmers University of Technology
Poster
main
https://github.com/kunalchelani/EdgeGaussians
https://openaccess.thecvf.com/content/WACV2025/html/Chelani_EdgeGaussians_-_3D_Edge_Mapping_via_Gaussian_Splatting_WACV_2025_paper.html
2
EdgeGaussians - 3D Edge Mapping via Gaussian Splatting With their meaningful geometry and omnipresence in the 3D world edges are extremely useful primitives in computer vision. Methods for 3D edge reconstruction have 1) either focused on reconstructing 3D edges by triangulating tracks of 2D line segments across images ...
[ -0.04649294912815094, 0.002170102670788765, -0.026059741154313087, -0.05437749996781349, 0.0036970407236367464, -0.03751640394330025, 0.04349459707736969, 0.024097857996821404, -0.01991497352719307, 0.037201762199401855, -0.05197141692042351, -0.001618322916328907, 0.021636247634887695, 0....
wacv_2025_7ad168267e
7ad168267e
wacv
2,025
Effective Backdoor Learning on Open-Set Face Recognition Systems
Backdoor attacks pose a serious threat to the security of face recognition systems. These involve the insertion of poisoned inputs into the training data to manipulate the model's behavior at inference time and can cause severe consequences such as unauthorized access to secure systems or impersonation of legitimate us...
Diana Voth; Leonidas Dane; Jonas Grebe; Sebastian Peitz; Philipp Terhörst
Paderborn University; TU Darmstadt; TU Darmstadt; TU Dortmund; Paderborn University
Poster
main
https://openaccess.thecvf.com/content/WACV2025/html/Voth_Effective_Backdoor_Learning_on_Open-Set_Face_Recognition_Systems_WACV_2025_paper.html
0
Effective Backdoor Learning on Open-Set Face Recognition Systems Backdoor attacks pose a serious threat to the security of face recognition systems. These involve the insertion of poisoned inputs into the training data to manipulate the model's behavior at inference time and can cause severe consequences such as unauth...
[ -0.02224782295525074, -0.027481501922011375, -0.05008574202656746, 0.009534111246466637, -0.005561955738812685, -0.029075993224978447, 0.04261977598071098, 0.009027626365423203, 0.008802521973848343, 0.01900256797671318, -0.041831910610198975, -0.005918371491134167, -0.017229869961738586, ...
wacv_2025_f9488a2a30
f9488a2a30
wacv
2,025
Effective Scene Graph Generation by Statistical Relation Distillation
Annotating scene graphs for images is a time-consuming task resulting in many instances of missing relations within existing datasets. In this paper we introduce the Statistical Relation Distillation (SRD) method to enhance scenegraph datasets. SRD leverages human-annotated relations alongside object-to-object and pred...
Thanh-Son Nguyen; Hong Yang; Basura Fernando
Institute of High Performance Computing, Agency for Science, Technology and Research, Singapore + Centre for Frontier AI Research (CFAR), Agency for Science, Technology and Research, Singapore; Institute of High Performance Computing, Agency for Science, Technology and Research, Singapore + Centre for Frontier AI Resea...
Poster
main
https://github.com/LUNAProject22/SRD
https://openaccess.thecvf.com/content/WACV2025/html/Nguyen_Effective_Scene_Graph_Generation_by_Statistical_Relation_Distillation_WACV_2025_paper.html
0
Effective Scene Graph Generation by Statistical Relation Distillation Annotating scene graphs for images is a time-consuming task resulting in many instances of missing relations within existing datasets. In this paper we introduce the Statistical Relation Distillation (SRD) method to enhance scenegraph datasets. SRD l...
[ -0.07981420308351517, -0.015020283870398998, -0.03200652450323105, 0.006416922435164452, -0.03174929693341255, -0.04413299262523651, 0.012181587517261505, 0.02572280913591385, 0.005263989325612783, 0.012686857022345066, -0.020559873431921005, 0.0013343708124011755, -0.019916804507374763, 0...
wacv_2025_e21658da9b
e21658da9b
wacv
2,025
Effective and Efficient Medical Image Segmentation with Hierarchical Context Interaction
The U-Net models have become the predominant architecture within the domain of medical image segmentation. Recent advancements have showcased the potential of incorporating attention-based techniques into U-Net structures. Nevertheless the inclusion of attention mechanisms often leads to a substantial increase in both ...
Zehua Cheng; Di Yuan; Wenhu Zhang; Thomas Lukasiewicz
Department of Computer Science, University of Oxford; Department of Computer Science, University of Oxford; FLock.io; Department of Computer Science, University of Oxford
Poster
main
https://openaccess.thecvf.com/content/WACV2025/html/Cheng_Effective_and_Efficient_Medical_Image_Segmentation_with_Hierarchical_Context_Interaction_WACV_2025_paper.html
0
Effective and Efficient Medical Image Segmentation with Hierarchical Context Interaction The U-Net models have become the predominant architecture within the domain of medical image segmentation. Recent advancements have showcased the potential of incorporating attention-based techniques into U-Net structures. Neverthe...
[ -0.031003223732113838, -0.023088667541742325, -0.03269530087709427, 0.038717638701200485, 0.00515355821698904, -0.012972592376172543, -0.00915177259594202, 0.032549746334552765, 0.06295255571603775, 0.003802625462412834, -0.031039612367749214, 0.004785121884196997, 0.035242512822151184, 0....
wacv_2025_4030017485
4030017485
wacv
2,025
Efficient Progressive Image Compression with Variance-Aware Masking
Learned progressive image compression is gaining momentum as it allows improved image reconstruction as more bits are decoded at the receiver. We propose a progressive image compression method in which an image is first represented as a pair of base-quality and top-quality latent representations. Next a residual latent...
Alberto Presta; Enzo Tartaglione; Attilio Fiandrotti; Marco Grangetto; Pamela Cosman
University of Turin, Italy; LTCI, T ´el´ecom Paris, Institut Polytechnique de Paris; University of Turin, Italy + LTCI, T ´el´ecom Paris, Institut Polytechnique de Paris; University of Turin, Italy; Dept. of Electrical and Computer Engineering, UC San Diego, CA, USA
Poster
main
https://openaccess.thecvf.com/content/WACV2025/html/Presta_Efficient_Progressive_Image_Compression_with_Variance-Aware_Masking_WACV_2025_paper.html
0
2411.10185
Efficient Progressive Image Compression with Variance-Aware Masking Learned progressive image compression is gaining momentum as it allows improved image reconstruction as more bits are decoded at the receiver. We propose a progressive image compression method in which an image is first represented as a pair of base-qu...
[ -0.014121199026703835, -0.023936131969094276, -0.018399503082036972, -0.007806273642927408, 0.013226390816271305, -0.009926783852279186, 0.03271644189953804, 0.03722776845097542, -0.00704661849886179, 0.03310792148113251, -0.02604266069829464, -0.019909493625164032, -0.021792320534586906, ...
wacv_2025_8c5d189d1c
8c5d189d1c
wacv
2,025
Efficient Video Object Segmentation via Modulated Cross-Attention Memory
Recently transformer-based approaches have shown promising results for semi-supervised video object segmentation. However these approaches typically struggle on long videos due to increased GPU memory demands as they frequently expand the memory bank every few frames. We propose a transformer-based approach named MAVOS...
Abdelrahman Shaker; Syed Talal Wasim; Martin Danelljan; Salman Khan; Ming-Hsuan Yang; Fahad Shahbaz Khan
Mohamed Bin Zayed University of Artificial Intelligence; University of Bonn; ETH Zürich; Mohamed Bin Zayed University of Artificial Intelligence; University of California, Merced + Yonsei University + Google Research; Mohamed bin Zayed University of Artificial Intelligence + Linköping University
Poster
main
https://github.com/Amshaker/MAVOS
https://openaccess.thecvf.com/content/WACV2025/html/Shaker_Efficient_Video_Object_Segmentation_via_Modulated_Cross-Attention_Memory_WACV_2025_paper.html
2
2403.17937
Efficient Video Object Segmentation via Modulated Cross-Attention Memory Recently transformer-based approaches have shown promising results for semi-supervised video object segmentation. However these approaches typically struggle on long videos due to increased GPU memory demands as they frequently expand the memory b...
[ 0.00903312861919403, -0.028584346175193787, -0.05150299519300461, 0.020286843180656433, -0.016786906868219376, 0.008356899954378605, 0.010234803892672062, 0.027268443256616592, -0.009330120868980885, 0.04082955792546272, -0.01781952567398548, -0.0199213158339262, 0.019208533689379692, 0.02...
wacv_2025_a518585db6
a518585db6
wacv
2,025
EfficientCrackNet: A Lightweight Model for Crack Segmentation
Crack detection particularly from pavement images presents a formidable challenge in computer vision due to inherent complexities such as intensity inhomogeneity intricate topologies low contrast and noisy backgrounds. Automated crack detection is crucial for maintaining the structural integrity of essential infrastruc...
Abid Hasan Zim; Aquib Iqbal; Zaid Al-Huda; Asad Malik; Minoru Kuribayashi
Aligarh Muslim University; University of Massachusetts Amherst; Stirling College, Chengdu University; Monash University Malaysia; Tohoku University
Poster
main
https://openaccess.thecvf.com/content/WACV2025/html/Zim_EfficientCrackNet_A_Lightweight_Model_for_Crack_Segmentation_WACV_2025_paper.html
2
2409.18099
EfficientCrackNet: A Lightweight Model for Crack Segmentation Crack detection particularly from pavement images presents a formidable challenge in computer vision due to inherent complexities such as intensity inhomogeneity intricate topologies low contrast and noisy backgrounds. Automated crack detection is crucial fo...
[ -0.036268752068281174, -0.018033934757113457, -0.0208189208060503, 0.015906386077404022, 0.003914961591362953, -0.015979435294866562, -0.014061902649700642, 0.015395044349133968, 0.015504617244005203, 0.03467993810772896, -0.02982219122350216, 0.013440988026559353, -0.02454441227018833, 0....
wacv_2025_b01e57025a
b01e57025a
wacv
2,025
EfficientMorph: Parameter-Efficient Transformer-Based Architecture for 3D Image Registration
Transformers have emerged as the state-of-the-art architecture in medical image registration outperforming convolutional neural networks (CNNs) by addressing their limited receptive fields and overcoming gradient instability in deeper models. Despite their success transformer-based models require substantial resources ...
Abu Zahid Bin Aziz; Mokshagna Sai Teja Karanam; Tushar Kataria; Shireen Y. Elhabian
Scientific Computing and Imaging Institute & Kahlert School of Computing, University of Utah, Salt Lake City, UT, USA; Scientific Computing and Imaging Institute & Kahlert School of Computing, University of Utah, Salt Lake City, UT, USA; Scientific Computing and Imaging Institute & Kahlert School of Computing, Universi...
Poster
main
https://github.com/MedVIC-Lab/Efficient-Morph-Registration
https://openaccess.thecvf.com/content/WACV2025/html/Bin_Aziz_EfficientMorph_Parameter-Efficient_Transformer-Based_Architecture_for_3D_Image_Registration_WACV_2025_paper.html
1
2403.11026
EfficientMorph: Parameter-Efficient Transformer-Based Architecture for 3D Image Registration Transformers have emerged as the state-of-the-art architecture in medical image registration outperforming convolutional neural networks (CNNs) by addressing their limited receptive fields and overcoming gradient instability in...
[ -0.029715385288000107, -0.026057826355099678, -0.013447440229356289, 0.0006749877939000726, -0.015403596684336662, -0.0005348154809325933, -0.0159767959266901, 0.008775410242378712, 0.024037981405854225, 0.007055811583995819, -0.043672334402799606, -0.0007563047693111002, 0.02855078876018524...
wacv_2025_f7a1bbab7d
f7a1bbab7d
wacv
2,025
Ego-VPA: Egocentric Video Understanding with Parameter-Efficient Adaptation
Video understanding typically requires fine-tuning the large backbone when adapting to new domains. In this paper we leverage the egocentric video foundation models (Ego-VFMs) based on video-language pre-training and propose a parameter-efficient adaptation for egocentric video tasks namely Ego-VPA. It employs a local ...
Tz-Ying Wu; Kyle Min; Subarna Tripathi; Nuno Vasconcelos
Intel Labs; Intel Labs; Intel Labs; UC San Diego
Poster
main
https://openaccess.thecvf.com/content/WACV2025/html/Wu_Ego-VPA_Egocentric_Video_Understanding_with_Parameter-Efficient_Adaptation_WACV_2025_paper.html
0
Ego-VPA: Egocentric Video Understanding with Parameter-Efficient Adaptation Video understanding typically requires fine-tuning the large backbone when adapting to new domains. In this paper we leverage the egocentric video foundation models (Ego-VFMs) based on video-language pre-training and propose a parameter-efficie...
[ -0.018261900171637535, -0.03372419625520706, -0.025988442823290825, 0.049287792295217514, -0.023188836872577667, -0.020812856033444405, 0.04737227410078049, 0.00999201275408268, 0.010001222603023052, 0.014707137830555439, -0.029874738305807114, -0.020315557718276978, 0.008711930364370346, ...
wacv_2025_1aee3c9249
1aee3c9249
wacv
2,025
EgoCast: Forecasting Egocentric Human Pose in the Wild
Accurately estimating and forecasting human body pose is important for enhancing the user's sense of immersion in Augmented Reality. Addressing this need our paper introduces EgoCast a bimodal method for 3D human pose forecasting using egocentric videos and proprioceptive data. We study the task of human pose forecasti...
Maria Escobar; Juanita Puentes; Cristhian Forigua; Jordi Pont-Tuset; Kevis-Kokitsi Maninis; Pablo Arbelaez
Universidad de Los Andes; Universidad de Los Andes; Universidad de Los Andes; Google DeepMind; Google DeepMind; Universidad de Los Andes
Poster
main
https://openaccess.thecvf.com/content/WACV2025/html/Escobar_EgoCast_Forecasting_Egocentric_Human_Pose_in_the_Wild_WACV_2025_paper.html
2
2412.02903
EgoCast: Forecasting Egocentric Human Pose in the Wild Accurately estimating and forecasting human body pose is important for enhancing the user's sense of immersion in Augmented Reality. Addressing this need our paper introduces EgoCast a bimodal method for 3D human pose forecasting using egocentric videos and proprio...
[ -0.06585034728050232, -0.0156030198559165, -0.046080294996500015, 0.005484414752572775, -0.03442007303237915, -0.0075866179540753365, 0.031953487545251846, 0.009978458285331726, 0.007306323852390051, 0.014407099224627018, -0.03541044518351555, -0.00005941643757978454, 0.004863096866756678, ...
wacv_2025_30f1080749
30f1080749
wacv
2,025
EgoPoints: Advancing Point Tracking for Egocentric Videos
We introduce EgoPoints a benchmark for point tracking in egocentric videos. We annotate 4.7K challenging tracks in egocentric sequences. Compared to the popular TAP-Vid-DAVIS evaluation benchmark we include 9x more points that go out-of-view and 59x more points that require re-identification (ReID) after returning to v...
Ahmad Darkhalil; Rhodri Guerrier; Adam W. Harley; Dima Damen
University of Bristol; University of Bristol; Stanford University; University of Bristol
Poster
main
http://ahmaddarkhalil.github.io/EgoPoints
https://openaccess.thecvf.com/content/WACV2025/html/Darkhalil_EgoPoints_Advancing_Point_Tracking_for_Egocentric_Videos_WACV_2025_paper.html
0
2412.04592
EgoPoints: Advancing Point Tracking for Egocentric Videos We introduce EgoPoints a benchmark for point tracking in egocentric videos. We annotate 4.7K challenging tracks in egocentric sequences. Compared to the popular TAP-Vid-DAVIS evaluation benchmark we include 9x more points that go out-of-view and 59x more points ...
[ -0.05347428843379021, -0.05504376441240311, -0.0035967156291007996, 0.016105812042951584, -0.024980824440717697, 0.010491198860108852, -0.0010317201958969235, -0.009253368712961674, -0.008576065301895142, 0.02423345483839512, -0.03107188642024994, -0.035798996686935425, 0.006889812648296356,...
wacv_2025_74112133f4
74112133f4
wacv
2,025
EgoSonics: Generating Synchronized Audio for Silent Egocentric Videos
We introduce EgoSonics a method to generate semantically meaningful and synchronized audio tracks conditioned on silent egocentric videos. Generating audio for silent egocentric videos could open new applications in virtual reality assistive technologies or for augmenting existing datasets. Existing work has been limit...
Aashish Rai; Srinath Sridhar
Brown University; Brown University
Poster
main
https://openaccess.thecvf.com/content/WACV2025/html/Rai_EgoSonics_Generating_Synchronized_Audio_for_Silent_Egocentric_Videos_WACV_2025_paper.html
3
2407.20592
EgoSonics: Generating Synchronized Audio for Silent Egocentric Videos We introduce EgoSonics a method to generate semantically meaningful and synchronized audio tracks conditioned on silent egocentric videos. Generating audio for silent egocentric videos could open new applications in virtual reality assistive technolo...
[ -0.07179774343967438, -0.008774555288255215, -0.01166527159512043, 0.022660231217741966, -0.014681669883430004, -0.0172046460211277, -0.0035237870179116726, 0.0371277742087841, 0.012614878825843334, 0.04379364475607872, -0.006917235907167196, -0.019699692726135254, -0.012195934541523457, -...
wacv_2025_ead48ca69c
ead48ca69c
wacv
2,025
ElasticLaneNet: An Efficient Geometry-Flexible Lane Detection Framework
The task of lane detection involves identifying the boundaries of driving areas in real-time. Recognizing lanes with variable and complex geometric structures remains a challenge. In this paper we explore a novel and flexible way of implicit lanes representation named Elastic Lane map (ELM) and introduce an efficient p...
Yaxin Feng; Yuan Lan; Luchan Zhang; Yang Xiang
The Hong Kong University of Science and Technology, Hong Kong SAR, China; The Hong Kong University of Science and Technology, Hong Kong SAR, China; Shenzhen University, Shenzhen, China; The Hong Kong University of Science and Technology, Hong Kong SAR, China+HKUST Shenzhen-Hong Kong Collaborative Innovation Research In...
Poster
main
https://openaccess.thecvf.com/content/WACV2025/html/Feng_ElasticLaneNet_An_Efficient_Geometry-Flexible_Lane_Detection_Framework_WACV_2025_paper.html
0
ElasticLaneNet: An Efficient Geometry-Flexible Lane Detection Framework The task of lane detection involves identifying the boundaries of driving areas in real-time. Recognizing lanes with variable and complex geometric structures remains a challenge. In this paper we explore a novel and flexible way of implicit lanes ...
[ -0.03877941519021988, -0.005711553618311882, 0.007666235789656639, 0.0039902315475046635, -0.020702071487903595, -0.019759386777877808, -0.0074952589347958565, 0.009907419793307781, 0.01665407605469227, 0.03650588542222977, -0.01972241885960102, 0.02800324745476246, 0.015859264880418777, 0...
wacv_2025_0a53681f87
0a53681f87
wacv
2,025
Elemental Composite Prototypical Network: Few-Shot Object Detection on Outdoor 3D Point Cloud Scenes
This paper introduces the Elemental Composite Prototypical Network (ECPN) a novel approach to few-shot learning (FSL) in outdoor 3D point cloud object detection. Such point clouds are inherently non-uniformly packed and show marked intra-class variations due to aberrations in lidar scanning methods. Due to the limited ...
Arkadipta De; Vartika Sengar; Daksh Thapar; Mahesh Chandran; Manohar Kaul
Fujitsu Research India, Bangalore; Fujitsu Research India, Bangalore; Fujitsu Research India, Bangalore; Fujitsu Research India, Bangalore; Fujitsu Research India, Bangalore
Poster
main
https://openaccess.thecvf.com/content/WACV2025/html/De_Elemental_Composite_Prototypical_Network_Few-Shot_Object_Detection_on_Outdoor_3D_WACV_2025_paper.html
0
Elemental Composite Prototypical Network: Few-Shot Object Detection on Outdoor 3D Point Cloud Scenes This paper introduces the Elemental Composite Prototypical Network (ECPN) a novel approach to few-shot learning (FSL) in outdoor 3D point cloud object detection. Such point clouds are inherently non-uniformly packed and...
[ -0.052170995622873306, -0.04329543188214302, -0.014963984489440918, -0.020781807601451874, 0.003567363368347287, -0.019086070358753204, 0.007436892483383417, 0.028123991563916206, 0.003409515367820859, 0.024534078314900398, -0.01863507553935051, 0.018337419256567955, -0.014675348065793514, ...
wacv_2025_a81751c9ea
a81751c9ea
wacv
2,025
Elucidating Optimal Reward-Diversity Tradeoffs in Text-to-Image Diffusion Models
Text-to-image (T2I) diffusion models have become prominent tools for generating high-fidelity images from text prompts. However when trained on unfiltered internet data these models can produce unsafe incorrect or stylistically undesirable images that are not aligned with human preferences. To address this recent appro...
Rohit Jena; Ali Taghibakhshi; Sahil Jain; Gerald Shen; Nima Tajbakhsh; Arash Vahdat
;;;;;
Poster
main
https://openaccess.thecvf.com/content/WACV2025/html/Jena_Elucidating_Optimal_Reward-Diversity_Tradeoffs_in_Text-to-Image_Diffusion_Models_WACV_2025_paper.html
2
2409.06493
Elucidating Optimal Reward-Diversity Tradeoffs in Text-to-Image Diffusion Models Text-to-image (T2I) diffusion models have become prominent tools for generating high-fidelity images from text prompts. However when trained on unfiltered internet data these models can produce unsafe incorrect or stylistically undesirable...
[ -0.08604054898023605, -0.017688771709799767, 0.012909765355288982, 0.028563546016812325, -0.011952121742069721, -0.013222840614616871, -0.03410682454705238, 0.011906081810593605, -0.01788214221596718, 0.006657460704445839, -0.04751382768154144, -0.024714557453989983, -0.053186021745204926, ...
wacv_2025_ad336c229b
ad336c229b
wacv
2,025
Elucidating the Solution Space of Extended Reverse-Time SDE for Diffusion Models
Sampling from Diffusion Models can alternatively be seen as solving differential equations where there is a challenge in balancing speed and image visual quality. ODE-based samplers offer rapid sampling time but reach a performance limit whereas SDE-based samplers achieve superior quality albeit with longer iterations....
Qinpeng Cui; Xinyi Zhang; Qiqi Bao; Qingmin Liao
Tsinghua University, China; Tsinghua University, China; Zhejiang University of Science and Technology, China; Tsinghua University, China
Poster
main
https://github.com/QinpengCui/ER-SDE-Solver
https://openaccess.thecvf.com/content/WACV2025/html/Cui_Elucidating_the_Solution_Space_of_Extended_Reverse-Time_SDE_for_Diffusion_WACV_2025_paper.html
5
Elucidating the Solution Space of Extended Reverse-Time SDE for Diffusion Models Sampling from Diffusion Models can alternatively be seen as solving differential equations where there is a challenge in balancing speed and image visual quality. ODE-based samplers offer rapid sampling time but reach a performance limit w...
[ -0.05711285024881363, -0.010770798660814762, -0.038222525268793106, 0.02691779099404812, 0.009270251728594303, 0.005251915194094181, 0.0046259197406470776, 0.025868328288197517, 0.023990342393517494, 0.04532941058278084, -0.020436899736523628, 0.013891568407416344, -0.01733454130589962, 0....
wacv_2025_b68712feb2
b68712feb2
wacv
2,025
EmoVOCA: Speech-Driven Emotional 3D Talking Heads
A notable challenge in 3D talking head generation consists in blending speech-related motions with expression dynamics. This is primarily caused by the lack of comprehensive 3D datasets that combine diversity in spoken sentences with a variety of facial expressions. Some literature works attempted to overcome such lack...
Federico Nocentini; Claudio Ferrari; Stefano Berretti
University of Florence, Italy; University of Parma, Italy; University of Florence, Italy
Poster
main
https://github.com/miccunifi/EmoVOCA
https://openaccess.thecvf.com/content/WACV2025/html/Nocentini_EmoVOCA_Speech-Driven_Emotional_3D_Talking_Heads_WACV_2025_paper.html
4
2403.12886
EmoVOCA: Speech-Driven Emotional 3D Talking Heads A notable challenge in 3D talking head generation consists in blending speech-related motions with expression dynamics. This is primarily caused by the lack of comprehensive 3D datasets that combine diversity in spoken sentences with a variety of facial expressions. Som...
[ -0.004050226882100105, -0.013802806846797466, 0.021418148651719093, -0.03807670623064041, -0.022058861330151558, -0.006635964382439852, -0.012457307428121567, 0.021344922482967377, 0.007578729186207056, -0.014315376989543438, 0.007532963994890451, 0.01338176615536213, -0.008301820605993271, ...
wacv_2025_94a3fb25e9
94a3fb25e9
wacv
2,025
Endoscopic Scoring and Localization in Unconstrained Clinical Trial Videos
Endoscopic assessment using the Mayo clinic score (or Mayo score 4 categories) is currently the standard for diagnosing and evaluating mucosal disease activities. However annotating Mayo scores is time-consuming and often relies on weakly labeled evaluations from central and local readers (doctors) leading to a large n...
Jinlin Xiang; Hillol Sarker; Bozhao Qi; Ruisu Zhang; Roger Trullo; Salvatore Badalamenti; Maria Wiekowski; Annie Kruger; Etienne Pochet; Qi Tang; Wei Zhao
Sanofi; Sanofi; Sanofi; Sanofi; Sanofi; Sanofi; Sanofi; Sanofi; Sanofi; Sanofi; Sanofi
Poster
main
https://openaccess.thecvf.com/content/WACV2025/html/Xiang_Endoscopic_Scoring_and_Localization_in_Unconstrained_Clinical_Trial_Videos_WACV_2025_paper.html
0
Endoscopic Scoring and Localization in Unconstrained Clinical Trial Videos Endoscopic assessment using the Mayo clinic score (or Mayo score 4 categories) is currently the standard for diagnosing and evaluating mucosal disease activities. However annotating Mayo scores is time-consuming and often relies on weakly labele...
[ -0.047530777752399445, -0.06733527034521103, -0.028583211824297905, -0.015519865788519382, 0.01401548646390438, 0.016624346375465393, -0.03294400870800018, 0.019747363403439522, -0.01640535518527031, 0.02172781154513359, 0.005746158771216869, -0.029402052983641624, 0.03863780200481415, -0....
wacv_2025_dc357a432a
dc357a432a
wacv
2,025
Enhancing Embodied Object Detection with Spatial Feature Memory
Deep-learning and large scale language-image training have produced image object detectors that generalise well to diverse environments and semantic classes. However existing object detection paradigms are not optimally tailored for the embodied conditions inherent in robotics where the same objects are repeatedly obse...
Nicolas Harvey Chapman; Christopher Lehnert; Will Browne; Feras Dayoub
Queensland University of Technology; Queensland University of Technology; Queensland University of Technology; University of Adelaide
Poster
main
https://openaccess.thecvf.com/content/WACV2025/html/Chapman_Enhancing_Embodied_Object_Detection_with_Spatial_Feature_Memory_WACV_2025_paper.html
2
Enhancing Embodied Object Detection with Spatial Feature Memory Deep-learning and large scale language-image training have produced image object detectors that generalise well to diverse environments and semantic classes. However existing object detection paradigms are not optimally tailored for the embodied conditions...
[ -0.059331320226192474, -0.04009571671485901, 0.004859667271375656, 0.0009253212483599782, -0.008842471055686474, -0.0621003620326519, 0.005293483380228281, 0.010439285077154636, 0.045707639306783676, -0.0008964770822785795, -0.021598519757390022, -0.028373440727591515, -0.003608983475714922,...
wacv_2025_5de4dfa0b1
5de4dfa0b1
wacv
2,025
Enhancing Image Layout Control with Loss-Guided Diffusion Models
Diffusion models are a powerful class of generative models capable of producing high-quality images from pure noise using a simple text prompt. While most methods which introduce additional spatial constraints into the generated images (e.g. bounding boxes) require fine-tuning a smaller and more recent subset of these ...
Zakaria Patel; Kirill Serkh
LeapTools; University of Toronto
Poster
main
https://openaccess.thecvf.com/content/WACV2025/html/Patel_Enhancing_Image_Layout_Control_with_Loss-Guided_Diffusion_Models_WACV_2025_paper.html
1
2405.14101
Enhancing Image Layout Control with Loss-Guided Diffusion Models Diffusion models are a powerful class of generative models capable of producing high-quality images from pure noise using a simple text prompt. While most methods which introduce additional spatial constraints into the generated images (e.g. bounding boxe...
[ -0.06707831472158432, -0.0375073216855526, -0.03080311417579651, 0.0021245088428258896, 0.0009110700339078903, -0.02817578986287117, 0.0005379220237955451, 0.04214590787887573, 0.01581830531358719, 0.012828591279685497, -0.07023110240697861, -0.01507540699094534, -0.03359351307153702, 0.00...
wacv_2025_c61c4a91da
c61c4a91da
wacv
2,025
Enhancing Monocular Depth Estimation with Multi-Source Auxiliary Tasks
Monocular depth estimation (MDE) is a challenging task in computer vision often hindered by the cost and scarcity of high-quality labeled datasets. We tackle this challenge using auxiliary datasets from related vision tasks for an alternating training scheme with a shared decoder built on top of a pre-trained vision fo...
Alessio Quercia; Erenus Yildiz; Zhuo Cao; Kai Krajsek; Abigail Morrison; Ira Assent; Hanno Scharr
IAS-8+Dept. of Computer Science, RWTH Aachen University; IAS-8+Dept. of Computer Science, Aarhus University; IAS-8; JSC, Forschungszentrum Juelich; IAS-8+Dept. of Computer Science, RWTH Aachen University; IAS-8+Dept. of Computer Science, Aarhus University; IAS-8
Poster
main
https://openaccess.thecvf.com/content/WACV2025/html/Quercia_Enhancing_Monocular_Depth_Estimation_with_Multi-Source_Auxiliary_Tasks_WACV_2025_paper.html
0
2501.12824
Enhancing Monocular Depth Estimation with Multi-Source Auxiliary Tasks Monocular depth estimation (MDE) is a challenging task in computer vision often hindered by the cost and scarcity of high-quality labeled datasets. We tackle this challenge using auxiliary datasets from related vision tasks for an alternating traini...
[ -0.03160824999213219, -0.053713608533144, -0.0469205267727375, -0.015238034538924694, 0.003315339796245098, 0.0060599492862820625, 0.025464778766036034, 0.001883873948827386, -0.002670368179678917, 0.0009245366672985256, -0.05185757577419281, -0.0500386618077755, -0.02886132150888443, 0.03...
wacv_2025_c134462317
c134462317
wacv
2,025
Enhancing Novel Object Detection via Cooperative Foundational Models
In this work we address the challenging and emergent problem of novel object detection (NOD) focusing on the accurate detection of both known and novel object categories during inference. Traditional object detection algorithms are inherently closed-set limiting their capability to handle NOD. We present a novel approa...
Rohit Bharadwaj; Muzammal Naseer; Salman Khan; Fahad Shahbaz Khan
MBZUAI; Khalifa University; MBZUAI + Australian National University; MBZUAI + Linkoping University
Poster
main
https://rohit901.github.io/coop-foundation-models/
https://openaccess.thecvf.com/content/WACV2025/html/Bharadwaj_Enhancing_Novel_Object_Detection_via_Cooperative_Foundational_Models_WACV_2025_paper.html
2
2311.12068
Enhancing Novel Object Detection via Cooperative Foundational Models In this work we address the challenging and emergent problem of novel object detection (NOD) focusing on the accurate detection of both known and novel object categories during inference. Traditional object detection algorithms are inherently closed-s...
[ -0.039053793996572495, -0.045195020735263824, -0.016244834288954735, 0.037840258330106735, -0.0012261769734323025, -0.00784661341458559, -0.015573712065815926, 0.020060118287801743, 0.043246008455753326, 0.008333866484463215, -0.02826067805290222, -0.03513738140463829, -0.010526505298912525,...
wacv_2025_7bd26f22e2
7bd26f22e2
wacv
2,025
Enhancing Predictive Imaging Biomarker Discovery through Treatment Effect Analysis
Identifying predictive covariates which forecast individual treatment effectiveness is crucial for decision-making across different disciplines such as personalized medicine. These covariates referred to as biomarkers are extracted from pre-treatment data often within randomized controlled trials and should be distingu...
Shuhan Xiao; Lukas Klein; Jens Petersen; Philipp Vollmuth; Paul F. Jaeger; Klaus H. Maier-Hein
;;;;;
Poster
main
https://github.com/MIC-DKFZ/predictive_image_biomarker_analysis
https://openaccess.thecvf.com/content/WACV2025/html/Xiao_Enhancing_Predictive_Imaging_Biomarker_Discovery_through_Treatment_Effect_Analysis_WACV_2025_paper.html
0
2406.02534
Enhancing Predictive Imaging Biomarker Discovery through Treatment Effect Analysis Identifying predictive covariates which forecast individual treatment effectiveness is crucial for decision-making across different disciplines such as personalized medicine. These covariates referred to as biomarkers are extracted from ...
[ -0.024240588769316673, -0.04098561406135559, -0.07107865065336227, 0.04142870008945465, 0.0315515361726284, -0.024092894047498703, 0.017797356471419334, 0.01601577363908291, -0.012563382275402546, 0.05117662996053696, -0.015969619154930115, -0.007684802170842886, -0.02863454259932041, 0.04...
wacv_2025_ad179d9f75
ad179d9f75
wacv
2,025
Enhancing Scene Graph Generation with Hierarchical Relationships and Commonsense Knowledge
This work introduces an enhanced approach to generating scene graphs by incorporating both a relationship hierarchy and commonsense knowledge. Specifically we begin by proposing a hierarchical relation head that exploits an informative hierarchical structure. It jointly predicts the relation super-category between obje...
Bowen Jiang; Zhijun Zhuang; Shreyas S. Shivakumar; Camillo J. Taylor
;;;
Poster
main
https://github.com/bowen-upenn/scene graph commonsense
https://openaccess.thecvf.com/content/WACV2025/html/Jiang_Enhancing_Scene_Graph_Generation_with_Hierarchical_Relationships_and_Commonsense_Knowledge_WACV_2025_paper.html
8
2311.12889
Enhancing Scene Graph Generation with Hierarchical Relationships and Commonsense Knowledge This work introduces an enhanced approach to generating scene graphs by incorporating both a relationship hierarchy and commonsense knowledge. Specifically we begin by proposing a hierarchical relation head that exploits an infor...
[ -0.03829170763492584, -0.016991477459669113, -0.0016345164040103555, 0.019483059644699097, -0.011867806315422058, -0.05118050053715706, -0.017918795347213745, 0.02542164735496044, 0.011408830061554909, -0.012617154978215694, -0.03087315708398819, 0.002840499160811305, -0.023211069405078888, ...
wacv_2025_f3829c3750
f3829c3750
wacv
2,025
Enhancing Skin Disease Diagnosis: Interpretable Visual Concept Discovery with SAM
Current AI-assisted skin image diagnosis has achieved dermatologist-level performance in classifying skin cancer driven by rapid advancements in deep learning architectures. However unlike traditional vision tasks skin images in general present unique challenges due to the limited availability of well-annotated dataset...
Xin Hu; Janet Wang; Jihun Hamm; Rie R Yotsu; Zhengming Ding
Department of Computer Science, Tulane University; Department of Computer Science, Tulane University; Department of Computer Science, Tulane University; Department of Tropical Medicine, Tulane University; Department of Computer Science, Tulane University
Poster
main
https://openaccess.thecvf.com/content/WACV2025/html/Hu_Enhancing_Skin_Disease_Diagnosis_Interpretable_Visual_Concept_Discovery_with_SAM_WACV_2025_paper.html
0
2409.09520
Enhancing Skin Disease Diagnosis: Interpretable Visual Concept Discovery with SAM Current AI-assisted skin image diagnosis has achieved dermatologist-level performance in classifying skin cancer driven by rapid advancements in deep learning architectures. However unlike traditional vision tasks skin images in general p...
[ -0.009037462994456291, -0.038633428514003754, -0.024596616625785828, 0.006627473048865795, -0.03534039109945297, -0.046580877155065536, -0.009888318367302418, 0.018120918422937393, 0.027834465727210045, 0.06523530930280685, 0.00273883412592113, 0.03039163164794445, -0.0026054568588733673, ...
wacv_2025_44e5d8aa15
44e5d8aa15
wacv
2,025
Enhancing Vision-Language Few-Shot Adaptation with Negative Learning
Large-scale pre-trained Vision-Language Models (VLMs) have exhibited impressive zero-shot performance and transferability allowing them to adapt to downstream tasks in a data-efficient manner. However when only a few labeled samples are available adapting VLMs to distinguish subtle differences between similar classes i...
Ce Zhang; Simon Stepputtis; Katia Sycara; Yaqi Xie
School of Computer Science, Carnegie Mellon University; School of Computer Science, Carnegie Mellon University; School of Computer Science, Carnegie Mellon University; School of Computer Science, Carnegie Mellon University
Poster
main
https://github.com/zhangce01/SimNL
https://openaccess.thecvf.com/content/WACV2025/html/Zhang_Enhancing_Vision-Language_Few-Shot_Adaptation_with_Negative_Learning_WACV_2025_paper.html
3
2403.12964
Enhancing Vision-Language Few-Shot Adaptation with Negative Learning Large-scale pre-trained Vision-Language Models (VLMs) have exhibited impressive zero-shot performance and transferability allowing them to adapt to downstream tasks in a data-efficient manner. However when only a few labeled samples are available adap...
[ -0.08058664947748184, -0.01716531626880169, 0.015602369792759418, 0.0308252964168787, 0.00009055552072823048, -0.02666948176920414, 0.030608471482992172, 0.03967898711562157, 0.010570221580564976, 0.03328264504671097, -0.030210958793759346, -0.03658922761678696, -0.02441088855266571, 0.026...
wacv_2025_f42b9dddd7
f42b9dddd7
wacv
2,025
Enhancing Visual Classification using Comparative Descriptors
The performance of vision-language models (VLMs) such as CLIP in visual classification tasks has been enhanced by leveraging semantic knowledge from large language models (LLMs) including GPT. Recent studies have shown that in zero-shot classification tasks descriptors incorporating additional cues high-level concepts ...
Hankyeol Lee; Gawon Seo; Wonseok Choi; Geunyoung Jung; Kyungwoo Song; Jiyoung Jung
Department of Artificial Intelligence, University of Seoul; Department of Artificial Intelligence, University of Seoul; Department of Artificial Intelligence, University of Seoul; Department of Artificial Intelligence, University of Seoul; Department of Applied Statistics, Yonsei University; Department of Artificial In...
Poster
main
https://github.com/hk1ee/Comparative-CLIP
https://openaccess.thecvf.com/content/WACV2025/html/Lee_Enhancing_Visual_Classification_using_Comparative_Descriptors_WACV_2025_paper.html
0
2411.05357
Enhancing Visual Classification using Comparative Descriptors The performance of vision-language models (VLMs) such as CLIP in visual classification tasks has been enhanced by leveraging semantic knowledge from large language models (LLMs) including GPT. Recent studies have shown that in zero-shot classification tasks ...
[ -0.06874094158411026, -0.04515155777335167, 0.0008454399066977203, 0.031532373279333115, 0.005238502286374569, 0.023128654807806015, -0.006464044563472271, 0.008951987139880657, 0.0016436550067737699, 0.0017715075518935919, -0.025929894298315048, -0.06365448236465454, -0.015489748679101467, ...
wacv_2025_d962b72db7
d962b72db7
wacv
2,025
Enhancing Zero-Shot Facial Expression Recognition by LLM Knowledge Transfer
Current facial expression recognition (FER) models are often designed in a supervised learning manner and thus are constrained by the lack of large-scale facial expression images with high-quality annotations. Consequently these models often fail to generalize well performing poorly on unseen images in inference. Visio...
Zengqun Zhao; Yu Cao; Shaogang Gong; Ioannis Patras
Queen Mary University of London; Queen Mary University of London; Queen Mary University of London; Queen Mary University of London
Poster
main
https://github.com/zengqunzhao/Exp-CLIP
https://openaccess.thecvf.com/content/WACV2025/html/Zhao_Enhancing_Zero-Shot_Facial_Expression_Recognition_by_LLM_Knowledge_Transfer_WACV_2025_paper.html
10
2405.19100
Enhancing Zero-Shot Facial Expression Recognition by LLM Knowledge Transfer Current facial expression recognition (FER) models are often designed in a supervised learning manner and thus are constrained by the lack of large-scale facial expression images with high-quality annotations. Consequently these models often fa...
[ -0.08158054947853088, -0.04897718131542206, 0.014804957434535027, 0.024326294660568237, -0.007028297521173954, 0.03227877616882324, -0.01668938808143139, 0.015643484890460968, 0.009359041228890419, 0.003058819565922022, -0.003973986022174358, 0.009692649357020855, -0.03496566787362099, 0.0...
wacv_2025_da6bf05fe8
da6bf05fe8
wacv
2,025
Enriching Local Patterns with Multi-Token Attention for Broad-Sight Neural Networks
In neural networks recognizing visual patterns is challenging because global average pooling disregards local patterns and solely relies on over-concentrated activation. Global average pooling enforces the network to learn objects regardless of their location so features tend to be activated only in specific regions. T...
Hankyul Kang; Jongbin Ryu
Ajou University; Ajou University
Poster
main
https://github.com/Lab-LVM/imagenet-models
https://openaccess.thecvf.com/content/WACV2025/html/Kang_Enriching_Local_Patterns_with_Multi-Token_Attention_for_Broad-Sight_Neural_Networks_WACV_2025_paper.html
0
Enriching Local Patterns with Multi-Token Attention for Broad-Sight Neural Networks In neural networks recognizing visual patterns is challenging because global average pooling disregards local patterns and solely relies on over-concentrated activation. Global average pooling enforces the network to learn objects regar...
[ 0.0032415888272225857, -0.04485756903886795, -0.024618737399578094, 0.029436638578772545, -0.017930252477526665, -0.036590490490198135, -0.0022663746494799852, -0.011296515353024006, -0.00747778220102191, -0.004498531576246023, -0.025056729093194008, 0.00982742104679346, -0.02469173632562160...
wacv_2025_51d2ead470
51d2ead470
wacv
2,025
Epipolar Attention Field Transformers for Bird's Eye View Semantic Segmentation
Spatial understanding of the semantics of the surroundings is a key capability needed by autonomous cars to enable safe driving decisions. Recently purely vision-based solutions have gained increasing research interest. In particular approaches extracting a bird's eye view (BEV) from multiple cameras have demonstrated ...
Christian Witte; Jens Behley; Cyrill Stachniss; Marvin Raaijmakers
;;;
Poster
main
https://openaccess.thecvf.com/content/WACV2025/html/Witte_Epipolar_Attention_Field_Transformers_for_Birds_Eye_View_Semantic_Segmentation_WACV_2025_paper.html
1
Epipolar Attention Field Transformers for Bird's Eye View Semantic Segmentation Spatial understanding of the semantics of the surroundings is a key capability needed by autonomous cars to enable safe driving decisions. Recently purely vision-based solutions have gained increasing research interest. In particular approa...
[ -0.05825827270746231, -0.06207302585244179, -0.011240550316870213, 0.01318496186286211, -0.000009675039109424688, -0.03859193995594978, 0.03135132044553757, 0.021055199205875397, 0.03974006697535515, 0.03370313346385956, -0.036166053265333176, -0.012231274507939816, 0.006356373894959688, 0...
wacv_2025_fa9e06aad9
fa9e06aad9
wacv
2,025
Evaluating Sensitivity Consistency of Explanations
While the performance of deep neural networks is rapidly developing their reliability is increasingly receiving more attention. Explainability methods are one of the most relevant tools to enhance reliability mainly by highlighting important input features for the explanation purpose. Although numerous explainability m...
Hanxiao Tan
AI Group, TU Dortmund
Poster
main
https://openaccess.thecvf.com/content/WACV2025/html/Tan_Evaluating_Sensitivity_Consistency_of_Explanations_WACV_2025_paper.html
0
Evaluating Sensitivity Consistency of Explanations While the performance of deep neural networks is rapidly developing their reliability is increasingly receiving more attention. Explainability methods are one of the most relevant tools to enhance reliability mainly by highlighting important input features for the expl...
[ -0.04315987974405289, -0.04597955569624901, -0.02554628811776638, 0.005926024168729782, -0.011109533719718456, -0.08037963509559631, 0.02966301701962948, 0.014530743472278118, 0.02678694576025009, -0.00028710789047181606, -0.027463668957352638, -0.016720693558454514, 0.005075421184301376, ...
wacv_2025_f6357a222d
f6357a222d
wacv
2,025
Event-Guided Fusion-Mamba for Context-Aware 3D Human Pose Estimation
3D human pose estimation (3D HPE) is an important computer vision task with various practical applications. Researchers have proposed various deep learning-based methods for 3D HPE. However the majority of such methods rely on lifting 2D pose sequence to 3D which do not perform well in challenging scenarios and are oft...
Bo Lang; Mooi Choo Chuah
Lehigh University; Lehigh University
Poster
main
https://openaccess.thecvf.com/content/WACV2025/html/Lang_Event-Guided_Fusion-Mamba_for_Context-Aware_3D_Human_Pose_Estimation_WACV_2025_paper.html
0
Event-Guided Fusion-Mamba for Context-Aware 3D Human Pose Estimation 3D human pose estimation (3D HPE) is an important computer vision task with various practical applications. Researchers have proposed various deep learning-based methods for 3D HPE. However the majority of such methods rely on lifting 2D pose sequence...
[ 0.02502773143351078, -0.04553215205669403, -0.04085931181907654, 0.0006512770778499544, -0.02785012684762478, -0.02461652085185051, 0.040971461683511734, 0.03846681863069534, 0.022485706955194473, 0.028971608728170395, -0.01585027389228344, -0.017252124845981598, 0.022018423303961754, 0.02...
wacv_2025_ecde745561
ecde745561
wacv
2,025
Event-Guided Low-Light Video Semantic Segmentation
Recent video semantic segmentation (VSS) methods have demonstrated promising results in well-lit environments. However their performance significantly drops in low-light scenarios due to limited visibility and reduced contextual details. In addition unfavorable low-light conditions make it harder to incorporate tempora...
Zhen Yao; Mooi Choo Chuah
Lehigh University; Lehigh University
Poster
main
https://openaccess.thecvf.com/content/WACV2025/html/Yao_Event-Guided_Low-Light_Video_Semantic_Segmentation_WACV_2025_paper.html
6
2411.00639
Event-Guided Low-Light Video Semantic Segmentation Recent video semantic segmentation (VSS) methods have demonstrated promising results in well-lit environments. However their performance significantly drops in low-light scenarios due to limited visibility and reduced contextual details. In addition unfavorable low-lig...
[ -0.06421145796775818, -0.02961905486881733, -0.045128535479307175, 0.03173733502626419, -0.013695129193365574, -0.002353130839765072, -0.017894845455884933, 0.03610282763838768, 0.0081093218177557, 0.03790796920657158, -0.008394828997552395, 0.02438782900571823, 0.00010972738527925685, 0.0...
wacv_2025_e2b73a06e8
e2b73a06e8
wacv
2,025
Event-Guided Video Transformer for End-to-End 3D Human Pose Estimation
3D human pose estimation (3D HPE) is an important computer vision task with various practical applications. However 3D pose estimation for multi-person from a monocular video (3DMPPE) is particularly challenging. Recent transformer-based approaches focus on capturing the spatial-temporal information from sequential 2D ...
Bo Lang; Mooi Choo Chuah
Lehigh University; Lehigh University
Poster
main
https://openaccess.thecvf.com/content/WACV2025/html/Lang_Event-Guided_Video_Transformer_for_End-to-End_3D_Human_Pose_Estimation_WACV_2025_paper.html
0
Event-Guided Video Transformer for End-to-End 3D Human Pose Estimation 3D human pose estimation (3D HPE) is an important computer vision task with various practical applications. However 3D pose estimation for multi-person from a monocular video (3DMPPE) is particularly challenging. Recent transformer-based approaches ...
[ -0.003243317361921072, -0.048435989767313004, -0.026774032041430473, 0.016375189647078514, -0.0223055649548769, -0.005700515117496252, 0.003091610036790371, -0.0027215362060815096, 0.011768805794417858, -0.007893375121057034, 0.005429280921816826, 0.007589960005134344, 0.02245267480611801, ...
wacv_2025_a6e92626d6
a6e92626d6
wacv
2,025
EvoCL: Continual Learning over Evolving Domains
Continual Learning aspires to build models capable of learning new tasks without forgetting previously learnt tasks. In real-world settings the distributions underlying the tasks are prone to shift. This necessitates a model capable of observing how the task distributions drift with time and adapt proactively. We prese...
Vishnuprasadh Kumaravelu; P.K. Srijith; Sunil Gupta
Indian Institute of Technology Hyderabad + Deakin University; Indian Institute of Technology Hyderabad; Deakin University
Poster
main
https://openaccess.thecvf.com/content/WACV2025/html/Kumaravelu_EvoCL_Continual_Learning_over_Evolving_Domains_WACV_2025_paper.html
0
EvoCL: Continual Learning over Evolving Domains Continual Learning aspires to build models capable of learning new tasks without forgetting previously learnt tasks. In real-world settings the distributions underlying the tasks are prone to shift. This necessitates a model capable of observing how the task distributions...
[ -0.0981614962220192, -0.034712936729192734, -0.01743071898818016, 0.02984941191971302, -0.019806792959570885, -0.012502226047217846, 0.019658287987113, -0.013254029676318169, -0.02066069282591343, -0.01479476410895586, -0.03085181675851345, -0.011230655945837498, -0.039762090891599655, 0.0...
wacv_2025_b54129cc19
b54129cc19
wacv
2,025
Exo2EgoDVC: Dense Video Captioning of Egocentric Procedural Activities using Web Instructional Videos
We propose a novel benchmark for cross-view knowledge transfer of dense video captioning adapting models from web instructional videos with exocentric views to an egocentric view. While dense video captioning (predicting time segments and their captions) is primarily studied with exocentric videos (e.g. YouCook2) bench...
Takehiko Ohkawa; Takuma Yagi; Taichi Nishimura; Ryosuke Furuta; Atsushi Hashimoto; Yoshitaka Ushiku; Yoichi Sato
The University of Tokyo + OMRON SINIC X Corp.; National Institute of Advanced Industrial Science and Technology (AIST); LY Corporation; The University of Tokyo; OMRON SINIC X Corp.; OMRON SINIC X Corp.; The University of Tokyo
Poster
main
https://openaccess.thecvf.com/content/WACV2025/html/Ohkawa_Exo2EgoDVC_Dense_Video_Captioning_of_Egocentric_Procedural_Activities_using_Web_WACV_2025_paper.html
9
2311.16444
Exo2EgoDVC: Dense Video Captioning of Egocentric Procedural Activities using Web Instructional Videos We propose a novel benchmark for cross-view knowledge transfer of dense video captioning adapting models from web instructional videos with exocentric views to an egocentric view. While dense video captioning (predicti...
[ -0.09515762329101562, -0.04465600848197937, -0.0032465325202792883, 0.019997159019112587, -0.03126290440559387, 0.00016215319919865578, -0.006317326799035072, -0.010794026777148247, -0.0063543240539729595, 0.005471008364111185, -0.009050518274307251, -0.032816801220178604, -0.002476521534845...
wacv_2025_ac6df8ef34
ac6df8ef34
wacv
2,025
Explicit Guidance for Robust Video Frame Interpolation against Discontinuous Motions
Nowadays many videos contain graphic elements such as logos subtitles and user interfaces. These overlayed elements exhibit discontinuous motions characterized by static or instantaneous motions that are neither spatially nor temporally coherent. As existing Video Frame Interpolation (VFI) methods rely on motion-compen...
JaeHyun Park; Nam Ik Cho
IPAI, Seoul National University, Korea; IPAI, Seoul National University, Korea + Department of Electrical and Computer Engineering, INMC, Seoul National University, Korea
Poster
main
https://openaccess.thecvf.com/content/WACV2025/html/Park_Explicit_Guidance_for_Robust_Video_Frame_Interpolation_against_Discontinuous_Motions_WACV_2025_paper.html
0
Explicit Guidance for Robust Video Frame Interpolation against Discontinuous Motions Nowadays many videos contain graphic elements such as logos subtitles and user interfaces. These overlayed elements exhibit discontinuous motions characterized by static or instantaneous motions that are neither spatially nor temporall...
[ -0.048454176634550095, -0.016944266855716705, -0.014023791998624802, 0.022978078573942184, -0.027331240475177765, -0.013041116297245026, 0.01570444367825985, 0.0398213230073452, 0.03713962808251381, 0.020810682326555252, -0.014923812821507454, -0.03363138437271118, 0.022702563554048538, 0....
wacv_2025_e2b91ed1ed
e2b91ed1ed
wacv
2,025
Exploiting Inter-Sample Information for Long-Tailed Out-of-Distribution Detection
Detecting out-of-distribution (OOD) data is essential for safe deployment of deep neural networks (DNNs). This problem becomes particularly challenging in the presence of long-tailed in-distribution (ID) datasets often leading to high false positive rates (FPR) and low tail-class ID classification accuracy. In this pap...
Nimeshika Udayangani; Hadi Mohaghegh Dolatabadi; Sarah Erfani; Christopher Leckie
The University of Melbourne; The University of Melbourne; The University of Melbourne; The University of Melbourne
Poster
main
https://openaccess.thecvf.com/content/WACV2025/html/Udayangani_Exploiting_Inter-Sample_Information_for_Long-Tailed_Out-of-Distribution_Detection_WACV_2025_paper.html
0
Exploiting Inter-Sample Information for Long-Tailed Out-of-Distribution Detection Detecting out-of-distribution (OOD) data is essential for safe deployment of deep neural networks (DNNs). This problem becomes particularly challenging in the presence of long-tailed in-distribution (ID) datasets often leading to high fal...
[ -0.06512534618377686, -0.033328425139188766, -0.03382069244980812, 0.010146208107471466, -0.031140564009547234, -0.02142281085252762, -0.004336979705840349, 0.019636057317256927, -0.01674625650048256, 0.05462361127138138, -0.053274430334568024, -0.02723887749016285, -0.022115634754300117, ...
wacv_2025_6e9fd42a91
6e9fd42a91
wacv
2,025
Exploiting VLM Localizability and Semantics for Open Vocabulary Action Detection
Action detection aims to detect (recognize and localize) human actions spatially and temporally in videos. Existing approaches focus on the closed-set setting where an action detector is trained and tested on videos from a fixed set of action categories. However this constrained setting is not viable in an open world w...
Wentao Bao; Kai Li; Yuxiao Chen; Deep A Patel; Martin Renqiang Min; Yu Kong
Department of Computer Science and Engineering, Michigan State University; Machine Learning Department, NEC Laboratories America; Department of Computer Science, Rutgers University; Machine Learning Department, NEC Laboratories America; Machine Learning Department, NEC Laboratories America; Department of Computer Scien...
Poster
main
https://github.com/Cogito2012/OpenMixer
https://openaccess.thecvf.com/content/WACV2025/html/Bao_Exploiting_VLM_Localizability_and_Semantics_for_Open_Vocabulary_Action_Detection_WACV_2025_paper.html
3
2411.10922
Exploiting VLM Localizability and Semantics for Open Vocabulary Action Detection Action detection aims to detect (recognize and localize) human actions spatially and temporally in videos. Existing approaches focus on the closed-set setting where an action detector is trained and tested on videos from a fixed set of act...
[ -0.0054755243472754955, -0.049413494765758514, -0.032290372997522354, 0.03498431295156479, -0.021108677610754967, 0.05594537779688835, -0.006697945762425661, -0.003102182177826762, 0.02430081181228161, 0.0007980334921739995, -0.020776547491550446, -0.046571943908929825, -0.01902364194393158,...
wacv_2025_bf8d26db13
bf8d26db13
wacv
2,025
Exploring Scalability of Self-Training for Open-Vocabulary Temporal Action Localization
The vocabulary size in temporal action localization (TAL) is limited by the scarcity of large-scale annotated datasets. To overcome this recent works integrate vision-language models (VLMs) such as CLIP for open-vocabulary TAL (OV-TAL). However despite the success of VLMs trained on extensive datasets existing OV-TAL m...
Jeongseok Hyun; Su Ho Han; Hyolim Kang; Joon-Young Lee; Seon Joo Kim
Yonsei University; Yonsei University; Yonsei University; Adobe Research; Yonsei University
Poster
main
https://github.com/HYUNJS/STOV-TAL
https://openaccess.thecvf.com/content/WACV2025/html/Hyun_Exploring_Scalability_of_Self-Training_for_Open-Vocabulary_Temporal_Action_Localization_WACV_2025_paper.html
1
2407.07024
Exploring Scalability of Self-Training for Open-Vocabulary Temporal Action Localization The vocabulary size in temporal action localization (TAL) is limited by the scarcity of large-scale annotated datasets. To overcome this recent works integrate vision-language models (VLMs) such as CLIP for open-vocabulary TAL (OV-T...
[ -0.046994760632514954, -0.018149573355913162, -0.031207744032144547, 0.03325895592570305, -0.011263358406722546, 0.053514689207077026, 0.01622656174004078, 0.015100225806236267, -0.023057835176587105, 0.036976780742406845, -0.011354930698871613, -0.0398155152797699, 0.007366968784481287, 0...
wacv_2025_1f45adc276
1f45adc276
wacv
2,025
Exploring the Stability Gap in Continual Learning: The Role of the Classification Head
Continual learning (CL) has emerged as a critical area in machine learning enabling neural networks to learn from evolving data distributions while mitigating catastrophic forgetting. However recent research has identified the stability gap - a phenomenon where models initially lose performance on previously learned ta...
Wojciech Łapacz; Daniel Marczak; Filip Szatkowski; Tomasz Trzciński
Warsaw University of Technology; Warsaw University of Technology+IDEAS NCBR; Warsaw University of Technology+IDEAS NCBR; Warsaw University of Technology+IDEAS NCBR+Tooploox
Poster
main
https://openaccess.thecvf.com/content/WACV2025/html/Lapacz_Exploring_the_Stability_Gap_in_Continual_Learning_The_Role_of_WACV_2025_paper.html
1
Exploring the Stability Gap in Continual Learning: The Role of the Classification Head Continual learning (CL) has emerged as a critical area in machine learning enabling neural networks to learn from evolving data distributions while mitigating catastrophic forgetting. However recent research has identified the stabil...
[ -0.059674862772226334, -0.014509360305964947, -0.05017782747745514, -0.012198778800666332, 0.009187745861709118, -0.0263624619692564, -0.002512984909117222, 0.00850093923509121, 0.007259228732436895, 0.013426844030618668, -0.029109688475728035, -0.06480544805526733, -0.03518633544445038, 0...
wacv_2025_851fe2c22b
851fe2c22b
wacv
2,025
F2FLDM: Latent Diffusion Models with Histopathology Pre-Trained Embeddings for Unpaired Frozen Section to FFPE Translation
Frozen Section (FS) technique is a rapid method taking only 15-30 minutes to prepare slides for pathologists' evaluation during surgery enabling immediate surgical decisions. However the FS process often introduces artifacts and distortions like folds and ice-crystal effects. In contrast these artifacts are absent in h...
Man M. Ho; Shikha Dubey; Yosep Chong; Beatrice Knudsen; Tolga Tasdizen
Scientific Computing and Imaging Institute, University of Utah, USA; Scientific Computing and Imaging Institute, University of Utah, USA; The Catholic University of Korea College of Medicine, Korea; Department of Pathology, University of Utah, USA+Huntsman Cancer Institute, University of Utah Health, USA; Scientific Co...
Poster
main
https://minhmanho.github.io/f2f_ldm/
https://openaccess.thecvf.com/content/WACV2025/html/Ho_F2FLDM_Latent_Diffusion_Models_with_Histopathology_Pre-Trained_Embeddings_for_Unpaired_WACV_2025_paper.html
4
2404.12650
F2FLDM: Latent Diffusion Models with Histopathology Pre-Trained Embeddings for Unpaired Frozen Section to FFPE Translation Frozen Section (FS) technique is a rapid method taking only 15-30 minutes to prepare slides for pathologists' evaluation during surgery enabling immediate surgical decisions. However the FS process...
[ -0.03904270753264427, -0.04337269812822342, -0.0075456490740180016, -0.014536404050886631, 0.0190301351249218, 0.026835039258003235, -0.001731769647449255, 0.034967418760061264, -0.00986073911190033, 0.040752872824668884, -0.0386788435280323, 0.031110450625419617, -0.004602891393005848, 0....
wacv_2025_19a6fd46cf
19a6fd46cf
wacv
2,025
F2former: When Fractional Fourier Meets Deep Wiener Deconvolution and Selective Frequency Transformer for Image Deblurring
Recent progress in image deblurring techniques focuses mainly on operating in both frequency and spatial domains using the Fourier transform (FT) properties. However their performance is limited due to the dependency of FT on stationary signals and its lack of capability to extract spatial-frequency properties. In this...
Subhajit Paul; Sahil Kumawat; Ashutosh Gupta; Deepak Mishra
Space Applications Centre (SAC), Ahmedabad; Indian Institute of Space Science and Technology (IIST), Trivandrum; Space Applications Centre (SAC), Ahmedabad; Indian Institute of Space Science and Technology (IIST), Trivandrum
Poster
main
https://openaccess.thecvf.com/content/WACV2025/html/Paul_F2former_When_Fractional_Fourier_Meets_Deep_Wiener_Deconvolution_and_Selective_WACV_2025_paper.html
4
2409.02056
F2former: When Fractional Fourier Meets Deep Wiener Deconvolution and Selective Frequency Transformer for Image Deblurring Recent progress in image deblurring techniques focuses mainly on operating in both frequency and spatial domains using the Fourier transform (FT) properties. However their performance is limited du...
[ -0.00932376366108656, -0.007128583267331123, -0.013198694214224815, 0.040314000099897385, 0.002283769892528653, 0.0027888454496860504, 0.02764914743602276, -0.003511368064209819, 0.017625870183110237, 0.05478286370635033, -0.027280982583761215, 0.0038588230963796377, -0.013327551074326038, ...
wacv_2025_8eb214ec10
8eb214ec10
wacv
2,025
FAIR-TAT: Improving Model Fairness using Targeted Adversarial Training
Deep neural networks are susceptible to adversarial attacks and common corruptions which undermine their robustness. In order to enhance model resilience against such challenges Adversarial Training (AT) has emerged as a prominent solution. Nevertheless adversarial robustness is often attained at the expense of model f...
Tejaswini Medi; Steffen Jung; Margret Keuper
University of Mannheim, Germany+MPI for Informatics, Saarland Informatics Campus, Germany; University of Mannheim, Germany+MPI for Informatics, Saarland Informatics Campus, Germany; University of Mannheim, Germany+MPI for Informatics, Saarland Informatics Campus, Germany
Poster
main
https://openaccess.thecvf.com/content/WACV2025/html/Medi_FAIR-TAT_Improving_Model_Fairness_using_Targeted_Adversarial_Training_WACV_2025_paper.html
0
FAIR-TAT: Improving Model Fairness using Targeted Adversarial Training Deep neural networks are susceptible to adversarial attacks and common corruptions which undermine their robustness. In order to enhance model resilience against such challenges Adversarial Training (AT) has emerged as a prominent solution. Neverthe...
[ -0.02669166773557663, -0.033885035663843155, -0.027397993952035904, -0.0023501585237681866, -0.0251117292791605, 0.000650562928058207, 0.024535516276955605, 0.0388479009270668, 0.007541883271187544, 0.049814533442258835, -0.015446222387254238, -0.020855188369750977, -0.04620855674147606, 0...
wacv_2025_34d0009d3e
34d0009d3e
wacv
2,025
FALCON: Fair Face Recognition via Local Optimal Feature Normalization
Face recognition systems are widely used for identity verification in various fields. However recent studies have highlighted bias issues related to demographic and non-demographic attributes such as accessories haircolor ethnicity or gender. These biases lead to higher error rates for specific attribute subgroups. Thi...
Rouqaiah Al-Refai; Philipp Hempel; Clara Biagi; Philipp Terhörst
Paderborn University, Germany; Technical University of Darmstadt, Germany; Paderborn University, Germany; Paderborn University, Germany
Poster
main
https://openaccess.thecvf.com/content/WACV2025/html/Al-Refai_FALCON_Fair_Face_Recognition_via_Local_Optimal_Feature_Normalization_WACV_2025_paper.html
0
FALCON: Fair Face Recognition via Local Optimal Feature Normalization Face recognition systems are widely used for identity verification in various fields. However recent studies have highlighted bias issues related to demographic and non-demographic attributes such as accessories haircolor ethnicity or gender. These b...
[ -0.008393005467951298, -0.035479091107845306, -0.021789714694023132, -0.009724179282784462, -0.025717150419950485, 0.04342837259173393, 0.03597002103924751, 0.024942992255091667, 0.017503522336483, 0.058722712099552155, -0.027661984786391258, 0.010639951564371586, 0.009922439232468605, 0.0...
wacv_2025_53b03c29aa
53b03c29aa
wacv
2,025
FASTER: A Font-Agnostic Scene Text Editing and Rendering Framework
Scene Text Editing (STE) is a challenging research problem that primarily aims towards modifying existing texts in an image while preserving the background and the font style of the original text. Despite its utility in numerous real-world applications existing style-transfer-based approaches have shown sub-par editing...
Alloy Das; Sanket Biswas; Prasun Roy; Subhankar Ghosh; Umapada Pal; Michael Blumenstein; Josep Lladós; Saumik Bhattacharya
CVPRU, Indian Statistical Institute, Kolkata; CVC, Universitat Autf1noma de Barcelona; FEIT, University of Technology Sydney, Australia; FEIT, University of Technology Sydney, Australia; CVPRU, Indian Statistical Institute, Kolkata; FEIT, University of Technology Sydney, Australia; CVC, Universitat Autf1noma de Barce...
Poster
main
https://openaccess.thecvf.com/content/WACV2025/html/Das_FASTER_A_Font-Agnostic_Scene_Text_Editing_and_Rendering_Framework_WACV_2025_paper.html
0
2308.02905
FASTER: A Font-Agnostic Scene Text Editing and Rendering Framework Scene Text Editing (STE) is a challenging research problem that primarily aims towards modifying existing texts in an image while preserving the background and the font style of the original text. Despite its utility in numerous real-world applications ...
[ -0.011074566282331944, -0.027714522555470467, -0.01937580667436123, 0.022898679599165916, 0.011964653618633747, -0.02321723662316799, -0.0052515179850161076, 0.05407986044883728, 0.03863918036222458, 0.042499348521232605, -0.006830252707004547, -0.017305180430412292, -0.0007173405028879642, ...
wacv_2025_bb3c332494
bb3c332494
wacv
2,025
FDS: Feedback-Guided Domain Synthesis with Multi-Source Conditional Diffusion Models for Domain Generalization
Domain Generalization techniques aim to enhance model robustness by simulating novel data distributions during training typically through various augmentation or stylization strategies. However these methods frequently suffer from limited control over the diversity of generated images and lack assurance that these imag...
Mehrdad Noori; Milad Cheraghalikhani; Ali Bahri; Gustavo A Vargas Hakim; David Osowiechi; Moslem Yazdanpanah; Ismail Ben Ayed; Christian Desrosiers
ÉTS Montreal, Canada; ÉTS Montreal, Canada; ÉTS Montreal, Canada; ÉTS Montreal, Canada; ÉTS Montreal, Canada; ÉTS Montreal, Canada; ÉTS Montreal, Canada; ÉTS Montreal, Canada
Poster
main
https://github.com/Mehrdad-Noori/FDS
https://openaccess.thecvf.com/content/WACV2025/html/Noori_FDS_Feedback-Guided_Domain_Synthesis_with_Multi-Source_Conditional_Diffusion_Models_for_WACV_2025_paper.html
2
2407.03588
FDS: Feedback-Guided Domain Synthesis with Multi-Source Conditional Diffusion Models for Domain Generalization Domain Generalization techniques aim to enhance model robustness by simulating novel data distributions during training typically through various augmentation or stylization strategies. However these methods f...
[ -0.06570792943239212, -0.0436369888484478, -0.06859301775693893, 0.004665731452405453, -0.014957388862967491, -0.023441357538104057, -0.009998640045523643, 0.020628394559025764, 0.028941059485077858, 0.036135751754045486, -0.0773204118013382, 0.006045165471732616, -0.013866464607417583, 0....
wacv_2025_7b9cb19a55
7b9cb19a55
wacv
2,025
FLAIR: A Conditional Diffusion Framework with Applications to Face Video Restoration
Face video restoration (FVR) is a challenging but important problem where one seeks to recover a perceptually realistic face videos from a low-quality input. While diffusion probabilistic models (DPMs) have been shown to achieve remarkable performance for face image restoration they often fail to preserve temporally co...
Zihao Zou; Jiaming Liu; Shirin Shoushtari; Yubo Wang; Ulugbek S. Kamilov
University of North Carolina, Chapel Hill, NC, USA; Washington University in St. Louis, MO, USA; Washington University in St. Louis, MO, USA; Washington University in St. Louis, MO, USA; Washington University in St. Louis, MO, USA
Poster
main
https://openaccess.thecvf.com/content/WACV2025/html/Zou_FLAIR_A_Conditional_Diffusion_Framework_with_Applications_to_Face_Video_WACV_2025_paper.html
2
2311.15445
FLAIR: A Conditional Diffusion Framework with Applications to Face Video Restoration Face video restoration (FVR) is a challenging but important problem where one seeks to recover a perceptually realistic face videos from a low-quality input. While diffusion probabilistic models (DPMs) have been shown to achieve remark...
[ -0.06638620793819427, -0.009002207778394222, -0.009909196756780148, -0.015964817255735397, -0.04133344441652298, 0.04259691387414932, 0.003957361448556185, 0.039745084941387177, -0.008939034305512905, 0.0508275032043457, 0.005496084690093994, -0.02922220341861248, -0.006998708937317133, 0....
wacv_2025_cc8491fc2f
cc8491fc2f
wacv
2,025
FMD: Comprehensive Data Compression in Medical Domain via Fused Matching Distillation
Medical datasets are often large and contain sensitive information presenting significant challenges for data sharing and storage. To address these issues this paper introduces a novel method called Fused Matching Distillation (FMD) which combines multiple dataset distillation techniques to achieve both data compressio...
Ju Heon Son; Jang-Hwan Choi
Ewha Womans University; Ewha Womans University
Poster
main
https://github.com/juheonewha/FMD.git
https://openaccess.thecvf.com/content/WACV2025/html/Son_FMD_Comprehensive_Data_Compression_in_Medical_Domain_via_Fused_Matching_WACV_2025_paper.html
0
FMD: Comprehensive Data Compression in Medical Domain via Fused Matching Distillation Medical datasets are often large and contain sensitive information presenting significant challenges for data sharing and storage. To address these issues this paper introduces a novel method called Fused Matching Distillation (FMD) w...
[ -0.027498090639710426, -0.057526007294654846, -0.057012710720300674, -0.007722380571067333, -0.014244011603295803, -0.036994099617004395, 0.024564960971474648, 0.015875564888119698, 0.02461995743215084, 0.04901992902159691, -0.042933687567710876, -0.005073397886008024, 0.0019374697003513575,...
wacv_2025_e9661c3b17
e9661c3b17
wacv
2,025
FOR: Finetuning for Object Level Open Vocabulary Image Retrieval
As working with large datasets becomes standard the task of accurately retrieving images containing objects of interest by an open set textual query gains practical importance. The current leading approach utilizes a pre-trained CLIP model without any adaptation to the target domain balancing accuracy and efficiency th...
Hila Levi; Guy Heller; Dan Levi
General Motors, RND, Israel; General Motors, RND, Israel; General Motors, RND, Israel
Poster
main
https://openaccess.thecvf.com/content/WACV2025/html/Levi_FOR_Finetuning_for_Object_Level_Open_Vocabulary_Image_Retrieval_WACV_2025_paper.html
0
2412.18806
FOR: Finetuning for Object Level Open Vocabulary Image Retrieval As working with large datasets becomes standard the task of accurately retrieving images containing objects of interest by an open set textual query gains practical importance. The current leading approach utilizes a pre-trained CLIP model without any ada...
[ -0.025395458564162254, -0.04304376244544983, -0.017367741093039513, 0.02993876487016678, -0.006452944129705429, -0.03140493109822273, 0.008136320859193802, 0.020128117874264717, -0.0041315131820738316, -0.007873859256505966, -0.005918969865888357, -0.015856323763728142, -0.006855687592178583...
wacv_2025_81f9f63528
81f9f63528
wacv
2,025
FRAUD-Net: Fraud News Detection using Sample Uncertainty & Domain Aware Generalized Network
Due to the widespread impact of social media efficiently detecting out-of-context misinformation where a real image is paired with a fake caption has become imperative. Towards this goal we propose a novel framework FRAUD-Net which incorporates several unexplored aspects of this task in the model formulation. Keeping i...
Devendra Patel; Vikas Verma; Shreyas Kumar Tah; Shwetabh Biswas; Soma Biswas
Indian Institute of Science, Bangalore; Indian Institute of Science, Bangalore; Indian Institute of Science, Bangalore; Indian Institute of Science, Bangalore; Indian Institute of Science, Bangalore
Poster
main
https://openaccess.thecvf.com/content/WACV2025/html/Patel_FRAUD-Net_Fraud_News_Detection_using_Sample_Uncertainty__Domain_Aware_WACV_2025_paper.html
0
FRAUD-Net: Fraud News Detection using Sample Uncertainty & Domain Aware Generalized Network Due to the widespread impact of social media efficiently detecting out-of-context misinformation where a real image is paired with a fake caption has become imperative. Towards this goal we propose a novel framework FRAUD-Net wh...
[ -0.07432431727647781, -0.025722043588757515, -0.022443028166890144, -0.011567632667720318, 0.018234960734844208, 0.01617646962404251, 0.015793917700648308, 0.0427364818751812, 0.009973667562007904, 0.0158850010484457, -0.04328298568725586, -0.016704754903912544, -0.03927529975771904, 0.009...
wacv_2025_0f6dca959d
0f6dca959d
wacv
2,025
FT2TF: First-Person Statement Text-To-Talking Face Generation
Talking face generation has gained immense popularity in the computer vision community with various applications including AR VR teleconferencing digital assistants and avatars. Traditional methods are mainly audio-driven which have to deal with the inevitable resource-intensive nature of audio storage and processing. ...
Xingjian Diao; Ming Cheng; Wayner Barrios; SouYoung Jin
Dartmouth College; Dartmouth College; Dartmouth College; Dartmouth College
Poster
main
https://openaccess.thecvf.com/content/WACV2025/html/Diao_FT2TF_First-Person_Statement_Text-To-Talking_Face_Generation_WACV_2025_paper.html
14
2312.05430
FT2TF: First-Person Statement Text-To-Talking Face Generation Talking face generation has gained immense popularity in the computer vision community with various applications including AR VR teleconferencing digital assistants and avatars. Traditional methods are mainly audio-driven which have to deal with the inevitab...
[ -0.03530072793364525, 0.009819954633712769, -0.006767190061509609, 0.005740475375205278, -0.01809755526483059, -0.003082425333559513, -0.012849903665482998, 0.02256034128367901, 0.012859029695391655, 0.010714337229728699, -0.002559941727668047, -0.019347865134477615, -0.004606525879353285, ...
wacv_2025_ddc96e17bc
ddc96e17bc
wacv
2,025
FUN-AD: Fully Unsupervised Learning for Anomaly Detection with Noisy Training Data
While the mainstream research in anomaly detection has mainly followed the one-class classification practical industrial environments often incur noisy training data due to annotation errors or lack of labels for new or refurbished products. To address these issues we propose a novel learning-based approach for fully u...
Jiin Im; Yongho Son; Je Hyeong Hong
;;
Poster
main
https://openaccess.thecvf.com/content/WACV2025/html/Im_FUN-AD_Fully_Unsupervised_Learning_for_Anomaly_Detection_with_Noisy_Training_WACV_2025_paper.html
0
FUN-AD: Fully Unsupervised Learning for Anomaly Detection with Noisy Training Data While the mainstream research in anomaly detection has mainly followed the one-class classification practical industrial environments often incur noisy training data due to annotation errors or lack of labels for new or refurbished produ...
[ -0.07392159104347229, -0.01947578601539135, -0.04526195302605629, 0.0011873046169057488, -0.0009320017998106778, -0.008230726234614849, 0.007404366973787546, 0.018564913421869278, -0.001182609354145825, 0.036284685134887695, -0.03853839263319969, 0.001497776247560978, -0.019015653058886528, ...
wacv_2025_454fffb973
454fffb973
wacv
2,025
FaVoR: Features via Voxel Rendering for Camera Relocalization
Camera relocalization methods range from dense image alignment to direct camera pose regression from a query image. Among these sparse feature matching stands out as an efficient versatile and generally lightweight approach with numerous applications. However feature-based methods often struggle with significant viewpo...
Vincenzo Polizzi; Marco Cannici; Davide Scaramuzza; Jonathan Kelly
University of Toronto; University of Zurich; University of Zurich; University of Toronto
Poster
main
https://openaccess.thecvf.com/content/WACV2025/html/Polizzi_FaVoR_Features_via_Voxel_Rendering_for_Camera_Relocalization_WACV_2025_paper.html
0
2409.07571
FaVoR: Features via Voxel Rendering for Camera Relocalization Camera relocalization methods range from dense image alignment to direct camera pose regression from a query image. Among these sparse feature matching stands out as an efficient versatile and generally lightweight approach with numerous applications. Howeve...
[ -0.043835729360580444, -0.02282731980085373, -0.009667506441473961, 0.024828121066093445, -0.023882286623120308, 0.04565463960170746, 0.0457637719810009, 0.01515151932835579, 0.027501918375492096, 0.006079706363379955, -0.0214085690677166, 0.005302122328430414, -0.0018916663248091936, 0.05...