Gait depicts individuals' unique and distinguishing walking patterns and has become one of the most promising biometric features for human identification. As a fine-grained recognition task, gait recognition is easily affected by many factors and usually requires a large amount of completely annotated data that is costly and insatiable. This paper proposes a large-scale self-supervised benchmark for gait recognition with contrastive learning, aiming to learn the general gait representation from massive unlabelled walking videos for practical applications via offering informative walking priors and diverse real-world variations. Specifically, we collect a large-scale unlabelled gait dataset GaitLU-1M consisting of 1.02M walking sequences and propose a conceptually simple yet empirically powerful baseline model GaitSSB. Experimentally, we evaluate the pre-trained model on four widely-used gait benchmarks, CASIA-B, OU-MVLP, GREW and Gait3D with or without transfer learning. The unsupervised results are comparable to or even better than the early model-based and GEI-based methods. After transfer learning, GaitSSB outperforms existing methods by a large margin in most cases, and also showcases the superior generalization capacity. Further experiments indicate that the pre-training can save about 50% and 80% annotation costs of GREW and Gait3D. Theoretically, we discuss the critical issues for gait-specific contrastive framework and present some insights for further study. As far as we know, GaitLU-1M is the first large-scale unlabelled gait dataset, and GaitSSB is the first method that achieves remarkable unsupervised results on the aforementioned benchmarks. The source code of GaitSSB and anonymous data of GaitLU-1M is available at https://github.com/ShiqiYu/OpenGait.
{"title":"Learning Gait Representation from Massive Unlabelled Walking Videos: A Benchmark","authors":"Chao Fan, Saihui Hou, Jilong Wang, Yongzhen Huang, Shiqi Yu","doi":"10.48550/arXiv.2206.13964","DOIUrl":"https://doi.org/10.48550/arXiv.2206.13964","url":null,"abstract":"Gait depicts individuals' unique and distinguishing walking patterns and has become one of the most promising biometric features for human identification. As a fine-grained recognition task, gait recognition is easily affected by many factors and usually requires a large amount of completely annotated data that is costly and insatiable. This paper proposes a large-scale self-supervised benchmark for gait recognition with contrastive learning, aiming to learn the general gait representation from massive unlabelled walking videos for practical applications via offering informative walking priors and diverse real-world variations. Specifically, we collect a large-scale unlabelled gait dataset GaitLU-1M consisting of 1.02M walking sequences and propose a conceptually simple yet empirically powerful baseline model GaitSSB. Experimentally, we evaluate the pre-trained model on four widely-used gait benchmarks, CASIA-B, OU-MVLP, GREW and Gait3D with or without transfer learning. The unsupervised results are comparable to or even better than the early model-based and GEI-based methods. After transfer learning, GaitSSB outperforms existing methods by a large margin in most cases, and also showcases the superior generalization capacity. Further experiments indicate that the pre-training can save about 50% and 80% annotation costs of GREW and Gait3D. Theoretically, we discuss the critical issues for gait-specific contrastive framework and present some insights for further study. As far as we know, GaitLU-1M is the first large-scale unlabelled gait dataset, and GaitSSB is the first method that achieves remarkable unsupervised results on the aforementioned benchmarks. The source code of GaitSSB and anonymous data of GaitLU-1M is available at https://github.com/ShiqiYu/OpenGait.","PeriodicalId":13426,"journal":{"name":"IEEE Transactions on Pattern Analysis and Machine Intelligence","volume":" ","pages":""},"PeriodicalIF":23.6,"publicationDate":"2022-06-28","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"46310622","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2022-06-19DOI: 10.48550/arXiv.2206.09479
Minguk Kang, Joonghyuk Shin, Jaesik Park
Generative Adversarial Network (GAN) is one of the state-of-the-art generative models for realistic image synthesis. While training and evaluating GAN becomes increasingly important, the current GAN research ecosystem does not provide reliable benchmarks for which the evaluation is conducted consistently and fairly. Furthermore, because there are few validated GAN implementations, researchers devote considerable time to reproducing baselines. We study the taxonomy of GAN approaches and present a new open-source library named StudioGAN. StudioGAN supports 7 GAN architectures, 9 conditioning methods, 4 adversarial losses, 12 regularization modules, 3 differentiable augmentations, 7 evaluation metrics, and 5 evaluation backbones. With our training and evaluation protocol, we present a large-scale benchmark using various datasets (CIFAR10, ImageNet, AFHQv2, FFHQ, and Baby/Papa/Granpa-ImageNet) and 3 different evaluation backbones (InceptionV3, SwAV, and Swin Transformer). Unlike other benchmarks used in the GAN community, we train representative GANs, including BigGAN and StyleGAN series in a unified training pipeline and quantify generation performance with 7 evaluation metrics. The benchmark evaluates other cutting-edge generative models (e.g., StyleGAN-XL, ADM, MaskGIT, and RQ-Transformer). StudioGAN provides GAN implementations, training, and evaluation scripts with the pre-trained weights. StudioGAN is available at https://github.com/POSTECH-CVLab/PyTorch-StudioGAN.
{"title":"StudioGAN: A Taxonomy and Benchmark of GANs for Image Synthesis","authors":"Minguk Kang, Joonghyuk Shin, Jaesik Park","doi":"10.48550/arXiv.2206.09479","DOIUrl":"https://doi.org/10.48550/arXiv.2206.09479","url":null,"abstract":"Generative Adversarial Network (GAN) is one of the state-of-the-art generative models for realistic image synthesis. While training and evaluating GAN becomes increasingly important, the current GAN research ecosystem does not provide reliable benchmarks for which the evaluation is conducted consistently and fairly. Furthermore, because there are few validated GAN implementations, researchers devote considerable time to reproducing baselines. We study the taxonomy of GAN approaches and present a new open-source library named StudioGAN. StudioGAN supports 7 GAN architectures, 9 conditioning methods, 4 adversarial losses, 12 regularization modules, 3 differentiable augmentations, 7 evaluation metrics, and 5 evaluation backbones. With our training and evaluation protocol, we present a large-scale benchmark using various datasets (CIFAR10, ImageNet, AFHQv2, FFHQ, and Baby/Papa/Granpa-ImageNet) and 3 different evaluation backbones (InceptionV3, SwAV, and Swin Transformer). Unlike other benchmarks used in the GAN community, we train representative GANs, including BigGAN and StyleGAN series in a unified training pipeline and quantify generation performance with 7 evaluation metrics. The benchmark evaluates other cutting-edge generative models (e.g., StyleGAN-XL, ADM, MaskGIT, and RQ-Transformer). StudioGAN provides GAN implementations, training, and evaluation scripts with the pre-trained weights. StudioGAN is available at https://github.com/POSTECH-CVLab/PyTorch-StudioGAN.","PeriodicalId":13426,"journal":{"name":"IEEE Transactions on Pattern Analysis and Machine Intelligence","volume":" ","pages":""},"PeriodicalIF":23.6,"publicationDate":"2022-06-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"44616657","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2022-06-12DOI: 10.48550/arXiv.2206.05707
Zexi Chen, Yiyi Liao, Haozhe Du, Haodong Zhang, Xuecheng Xu, Haojian Lu, R. Xiong, Yue Wang
Pose registration is critical in vision and robotics. This paper focuses on the challenging task of initialization-free pose registration up to 7DoF for homogeneous and heterogeneous measurements. While recent learning-based methods show promise using differentiable solvers, they either rely on heuristically defined correspondences or require initialization. Phase correlation seeks solutions in the spectral domain and is correspondence-free and initialization-free. Following this, we propose a differentiable solver and combine it with simple feature extraction networks, namely DPCN++. It can perform registration for homo/hetero inputs and generalizes well on unseen objects. Specifically, the feature extraction networks first learn dense feature grids from a pair of homogeneous/heterogeneous measurements. These feature grids are then transformed into a translation and scale invariant spectrum representation based on Fourier transform and spherical radial aggregation, decoupling translation and scale from rotation. Next, the rotation, scale, and translation are independently and efficiently estimated in the spectrum step-by-step. The entire pipeline is differentiable and trained end-to-end. We evaluate DCPN++ on a wide range of tasks taking different input modalities, including 2D bird's-eye view images, 3D object and scene measurements, and medical images. Experimental results demonstrate that DCPN++ outperforms both classical and learning-based baselines, especially on partially observed and heterogeneous measurements.
{"title":"DPCN++: Differentiable Phase Correlation Network for Versatile Pose Registration","authors":"Zexi Chen, Yiyi Liao, Haozhe Du, Haodong Zhang, Xuecheng Xu, Haojian Lu, R. Xiong, Yue Wang","doi":"10.48550/arXiv.2206.05707","DOIUrl":"https://doi.org/10.48550/arXiv.2206.05707","url":null,"abstract":"Pose registration is critical in vision and robotics. This paper focuses on the challenging task of initialization-free pose registration up to 7DoF for homogeneous and heterogeneous measurements. While recent learning-based methods show promise using differentiable solvers, they either rely on heuristically defined correspondences or require initialization. Phase correlation seeks solutions in the spectral domain and is correspondence-free and initialization-free. Following this, we propose a differentiable solver and combine it with simple feature extraction networks, namely DPCN++. It can perform registration for homo/hetero inputs and generalizes well on unseen objects. Specifically, the feature extraction networks first learn dense feature grids from a pair of homogeneous/heterogeneous measurements. These feature grids are then transformed into a translation and scale invariant spectrum representation based on Fourier transform and spherical radial aggregation, decoupling translation and scale from rotation. Next, the rotation, scale, and translation are independently and efficiently estimated in the spectrum step-by-step. The entire pipeline is differentiable and trained end-to-end. We evaluate DCPN++ on a wide range of tasks taking different input modalities, including 2D bird's-eye view images, 3D object and scene measurements, and medical images. Experimental results demonstrate that DCPN++ outperforms both classical and learning-based baselines, especially on partially observed and heterogeneous measurements.","PeriodicalId":13426,"journal":{"name":"IEEE Transactions on Pattern Analysis and Machine Intelligence","volume":" ","pages":""},"PeriodicalIF":23.6,"publicationDate":"2022-06-12","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"49215308","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2022-06-10DOI: 10.48550/arXiv.2206.05184
Zhong-Yu Li, Shanghua Gao, Ming-Ming Cheng
Learning representations with self-supervision for convolutional networks (CNN) has been validated to be effective for vision tasks. As an alternative to CNN, vision transformers (ViT) have strong representation ability with spatial self-attention and channel-level feedforward networks. Recent works reveal that self-supervised learning helps unleash the great potential of ViT. Still, most works follow self-supervised strategies designed for CNN, e.g., instance-level discrimination of samples, but they ignore the properties of ViT. We observe that relational modeling on spatial and channel dimensions distinguishes ViT from other networks. To enforce this property, we explore the feature SElf-RElation (SERE) for training self-supervised ViT. Specifically, instead of conducting self-supervised learning solely on feature embeddings from multiple views, we utilize the feature self-relations, i.e., spatial/channel self-relations, for self-supervised learning. Self-relation based learning further enhances the relation modeling ability of ViT, resulting in stronger representations that stably improve performance on multiple downstream tasks. Our source code is publicly available.
{"title":"Exploring Feature Self-relation for Self-supervised Transformer","authors":"Zhong-Yu Li, Shanghua Gao, Ming-Ming Cheng","doi":"10.48550/arXiv.2206.05184","DOIUrl":"https://doi.org/10.48550/arXiv.2206.05184","url":null,"abstract":"Learning representations with self-supervision for convolutional networks (CNN) has been validated to be effective for vision tasks. As an alternative to CNN, vision transformers (ViT) have strong representation ability with spatial self-attention and channel-level feedforward networks. Recent works reveal that self-supervised learning helps unleash the great potential of ViT. Still, most works follow self-supervised strategies designed for CNN, e.g., instance-level discrimination of samples, but they ignore the properties of ViT. We observe that relational modeling on spatial and channel dimensions distinguishes ViT from other networks. To enforce this property, we explore the feature SElf-RElation (SERE) for training self-supervised ViT. Specifically, instead of conducting self-supervised learning solely on feature embeddings from multiple views, we utilize the feature self-relations, i.e., spatial/channel self-relations, for self-supervised learning. Self-relation based learning further enhances the relation modeling ability of ViT, resulting in stronger representations that stably improve performance on multiple downstream tasks. Our source code is publicly available.","PeriodicalId":13426,"journal":{"name":"IEEE Transactions on Pattern Analysis and Machine Intelligence","volume":" ","pages":""},"PeriodicalIF":23.6,"publicationDate":"2022-06-10","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"44320055","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2022-05-14DOI: 10.48550/arXiv.2205.07019
Yihao Liu, Hengyuan Zhao, Jinjin Gu, Y. Qiao, Chao Dong
Performance and generalization ability are two important aspects to evaluate the deep learning models. However, research on the generalization ability of Super-Resolution (SR) networks is currently absent. Assessing the generalization ability of deep models not only helps us to understand their intrinsic mechanisms, but also allows us to quantitatively measure their applicability boundaries, which is important for unrestricted real-world applications. To this end, we make the first attempt to propose a Generalization Assessment Index for SR networks, namely SRGA. SRGA exploits the statistical characteristics of the internal features of deep networks to measure the generalization ability. Specially, it is a non-parametric and non-learning metric. To better validate our method, we collect a patch-based image evaluation set (PIES) that includes both synthetic and real-world images, covering a wide range of degradations. With SRGA and PIES dataset, we benchmark existing SR models on the generalization ability. This work provides insights and tools for future research on model generalization in low-level vision.
{"title":"Evaluating the Generalization Ability of Super-Resolution Networks","authors":"Yihao Liu, Hengyuan Zhao, Jinjin Gu, Y. Qiao, Chao Dong","doi":"10.48550/arXiv.2205.07019","DOIUrl":"https://doi.org/10.48550/arXiv.2205.07019","url":null,"abstract":"Performance and generalization ability are two important aspects to evaluate the deep learning models. However, research on the generalization ability of Super-Resolution (SR) networks is currently absent. Assessing the generalization ability of deep models not only helps us to understand their intrinsic mechanisms, but also allows us to quantitatively measure their applicability boundaries, which is important for unrestricted real-world applications. To this end, we make the first attempt to propose a Generalization Assessment Index for SR networks, namely SRGA. SRGA exploits the statistical characteristics of the internal features of deep networks to measure the generalization ability. Specially, it is a non-parametric and non-learning metric. To better validate our method, we collect a patch-based image evaluation set (PIES) that includes both synthetic and real-world images, covering a wide range of degradations. With SRGA and PIES dataset, we benchmark existing SR models on the generalization ability. This work provides insights and tools for future research on model generalization in low-level vision.","PeriodicalId":13426,"journal":{"name":"IEEE Transactions on Pattern Analysis and Machine Intelligence","volume":" ","pages":""},"PeriodicalIF":23.6,"publicationDate":"2022-05-14","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"49321791","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2022-05-03DOI: 10.48550/arXiv.2205.01679
Fangzhou Mu, Sicheng Mo, Jiayong Peng, Xiaochun Liu, J. Nam, S. Raghavan, A. Velten, Yin Li
Computational approach to imaging around the corner, or non-line-of-sight (NLOS) imaging, is becoming a reality thanks to major advances in imaging hardware and reconstruction algorithms. A recent development towards practical NLOS imaging, Nam et al. [1] demonstrated a high-speed non-confocal imaging system that operates at 5 Hz, 100x faster than the prior art. This enormous gain in acquisition rate, however, necessitates numerous approximations in light transport, breaking many existing NLOS reconstruction methods that assume an idealized image formation model. To bridge the gap, we present a novel deep model that incorporates the complementary physics priors of wave propagation and volume rendering into a neural network for high-quality and robust NLOS reconstruction. This orchestrated design regularizes the solution space by relaxing the image formation model, resulting in a deep model that generalizes well on real captures despite being exclusively trained on synthetic data. Further, we devise a unified learning framework that enables our model to be flexibly trained using diverse supervision signals, including target intensity images or even raw NLOS transient measurements. Once trained, our model renders both intensity and depth images at inference time in a single forward pass, capable of processing more than 5 captures per second on a high-end GPU. Through extensive qualitative and quantitative experiments, we show that our method outperforms prior physics and learning based approaches on both synthetic and real measurements. We anticipate that our method along with the fast capturing system will accelerate future development of NLOS imaging for real world applications that require high-speed imaging.
{"title":"Physics to the Rescue: Deep Non-line-of-sight Reconstruction for High-speed Imaging","authors":"Fangzhou Mu, Sicheng Mo, Jiayong Peng, Xiaochun Liu, J. Nam, S. Raghavan, A. Velten, Yin Li","doi":"10.48550/arXiv.2205.01679","DOIUrl":"https://doi.org/10.48550/arXiv.2205.01679","url":null,"abstract":"Computational approach to imaging around the corner, or non-line-of-sight (NLOS) imaging, is becoming a reality thanks to major advances in imaging hardware and reconstruction algorithms. A recent development towards practical NLOS imaging, Nam et al. [1] demonstrated a high-speed non-confocal imaging system that operates at 5 Hz, 100x faster than the prior art. This enormous gain in acquisition rate, however, necessitates numerous approximations in light transport, breaking many existing NLOS reconstruction methods that assume an idealized image formation model. To bridge the gap, we present a novel deep model that incorporates the complementary physics priors of wave propagation and volume rendering into a neural network for high-quality and robust NLOS reconstruction. This orchestrated design regularizes the solution space by relaxing the image formation model, resulting in a deep model that generalizes well on real captures despite being exclusively trained on synthetic data. Further, we devise a unified learning framework that enables our model to be flexibly trained using diverse supervision signals, including target intensity images or even raw NLOS transient measurements. Once trained, our model renders both intensity and depth images at inference time in a single forward pass, capable of processing more than 5 captures per second on a high-end GPU. Through extensive qualitative and quantitative experiments, we show that our method outperforms prior physics and learning based approaches on both synthetic and real measurements. We anticipate that our method along with the fast capturing system will accelerate future development of NLOS imaging for real world applications that require high-speed imaging.","PeriodicalId":13426,"journal":{"name":"IEEE Transactions on Pattern Analysis and Machine Intelligence","volume":" ","pages":""},"PeriodicalIF":23.6,"publicationDate":"2022-05-03","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"42504585","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2022-04-26DOI: 10.48550/arXiv.2204.12139
Youjian Zhang, Chaoyue Wang, D. Tao
Real-world dynamic scene deblurring has long been a challenging task since paired blurry-sharp training data is unavailable. Conventional Maximum A Posteriori estimation and deep learning-based deblurring methods are restricted by handcrafted priors and synthetic blurry-sharp training pairs respectively, thereby failing to generalize to real dynamic blurriness. To this end, we propose a Neural Maximum A Posteriori (NeurMAP) estimation framework for training neural networks to recover blind motion information and sharp content from unpaired data. The proposed NeruMAP consists of a motion estimation network and a deblurring network which are trained jointly to model the (re)blurring process (i.e. likelihood function). Meanwhile, the motion estimation network is trained to explore the motion information in images by applying implicit dynamic motion prior, and in return enforces the deblurring network training (i.e. providing sharp image prior). The proposed NeurMAP is an orthogonal approach to existing deblurring neural networks, and is the first framework that enables training image deblurring networks on unpaired datasets. Experiments demonstrate our superiority on both quantitative metrics and visual quality over State-of-the-art methods. Codes are available on https://github.com/yjzhang96/NeurMAP-deblur.
{"title":"Neural Maximum A Posteriori Estimation on Unpaired Data for Motion Deblurring","authors":"Youjian Zhang, Chaoyue Wang, D. Tao","doi":"10.48550/arXiv.2204.12139","DOIUrl":"https://doi.org/10.48550/arXiv.2204.12139","url":null,"abstract":"Real-world dynamic scene deblurring has long been a challenging task since paired blurry-sharp training data is unavailable. Conventional Maximum A Posteriori estimation and deep learning-based deblurring methods are restricted by handcrafted priors and synthetic blurry-sharp training pairs respectively, thereby failing to generalize to real dynamic blurriness. To this end, we propose a Neural Maximum A Posteriori (NeurMAP) estimation framework for training neural networks to recover blind motion information and sharp content from unpaired data. The proposed NeruMAP consists of a motion estimation network and a deblurring network which are trained jointly to model the (re)blurring process (i.e. likelihood function). Meanwhile, the motion estimation network is trained to explore the motion information in images by applying implicit dynamic motion prior, and in return enforces the deblurring network training (i.e. providing sharp image prior). The proposed NeurMAP is an orthogonal approach to existing deblurring neural networks, and is the first framework that enables training image deblurring networks on unpaired datasets. Experiments demonstrate our superiority on both quantitative metrics and visual quality over State-of-the-art methods. Codes are available on https://github.com/yjzhang96/NeurMAP-deblur.","PeriodicalId":13426,"journal":{"name":"IEEE Transactions on Pattern Analysis and Machine Intelligence","volume":" ","pages":""},"PeriodicalIF":23.6,"publicationDate":"2022-04-26","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"44976215","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2022-04-21DOI: 10.48550/arXiv.2204.10149
Zheng Hua Zhu, Guan Huang, Jiankang Deng, Yun Ye, Junjie Huang, Xinze Chen, Jiagang Zhu, Tian Yang, Dalong Du, Jiwen Lu, Jie Zhou
In this paper, we contribute a new million-scale recognition benchmark, containing uncurated 4M identities/260M faces (WebFace260M) and cleaned 2M identities/42M faces (WebFace42M) training data, as well as an elaborately designed time-constrained evaluation protocol. Firstly, we collect 4M name lists and download 260M faces from the Internet. Then, a Cleaning Automatically utilizing Self-Training pipeline is devised to purify the tremendous WebFace260M, which is efficient and scalable. To our best knowledge, the cleaned WebFace42M is the largest public face recognition training set in the community. Referring to practical deployments, Face Recognition under Inference Time conStraint (FRUITS) protocol and a new test set with rich attributes are constructed. Moreover, we gather a large-scale masked face sub-set for biometrics assessment under COVID-19. For a comprehensive evaluation of face matchers, three recognition tasks are performed under standard, masked and unbiased settings, respectively. Equipped with this benchmark, we delve into million-scale face recognition problems. Enabled by WebFace42M, we reduce 40% failure rate on the challenging IJB-C set and rank the 3rd among 430 entries on NIST-FRVT. Even 10% data (WebFace4M) shows superior performance compared with the public training set. The proposed benchmark shows enormous potential on standard, masked and unbiased face recognition scenarios.
{"title":"WebFace260M: A Benchmark for Million-Scale Deep Face Recognition","authors":"Zheng Hua Zhu, Guan Huang, Jiankang Deng, Yun Ye, Junjie Huang, Xinze Chen, Jiagang Zhu, Tian Yang, Dalong Du, Jiwen Lu, Jie Zhou","doi":"10.48550/arXiv.2204.10149","DOIUrl":"https://doi.org/10.48550/arXiv.2204.10149","url":null,"abstract":"In this paper, we contribute a new million-scale recognition benchmark, containing uncurated 4M identities/260M faces (WebFace260M) and cleaned 2M identities/42M faces (WebFace42M) training data, as well as an elaborately designed time-constrained evaluation protocol. Firstly, we collect 4M name lists and download 260M faces from the Internet. Then, a Cleaning Automatically utilizing Self-Training pipeline is devised to purify the tremendous WebFace260M, which is efficient and scalable. To our best knowledge, the cleaned WebFace42M is the largest public face recognition training set in the community. Referring to practical deployments, Face Recognition under Inference Time conStraint (FRUITS) protocol and a new test set with rich attributes are constructed. Moreover, we gather a large-scale masked face sub-set for biometrics assessment under COVID-19. For a comprehensive evaluation of face matchers, three recognition tasks are performed under standard, masked and unbiased settings, respectively. Equipped with this benchmark, we delve into million-scale face recognition problems. Enabled by WebFace42M, we reduce 40% failure rate on the challenging IJB-C set and rank the 3rd among 430 entries on NIST-FRVT. Even 10% data (WebFace4M) shows superior performance compared with the public training set. The proposed benchmark shows enormous potential on standard, masked and unbiased face recognition scenarios.","PeriodicalId":13426,"journal":{"name":"IEEE Transactions on Pattern Analysis and Machine Intelligence","volume":"PP 1","pages":"1-1"},"PeriodicalIF":23.6,"publicationDate":"2022-04-21","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"41630541","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2022-04-17DOI: 10.48550/arXiv.2204.08090
Samarth Mishra, Pengkai Zhu, Venkatesh Saligrama
We propose Recognition as Part Composition (RPC), an image encoding approach inspired by human cognition. It is based on the cognitive theory that humans recognize complex objects by components, and that they build a small compact vocabulary of concepts to represent each instance with. RPC encodes images by first decomposing them into salient parts, and then encoding each part as a mixture of a small number of prototypes, each representing a certain concept. We find that this type of learning inspired by human cognition can overcome hurdles faced by deep convolutional networks in low-shot generalization tasks, like zero-shot learning, few-shot learning and unsupervised domain adaptation. Furthermore, we find a classifier using an RPC image encoder is fairly robust to adversarial attacks, that deep neural networks are known to be prone to. Given that our image encoding principle is based on human cognition, one would expect the encodings to be interpretable by humans, which we find to be the case via crowd-sourcing experiments. Finally, we propose an application of these interpretable encodings in the form of generating synthetic attribute annotations for evaluating zero-shot learning methods on new datasets.
{"title":"Learning Compositional Representations for Effective Low-Shot Generalization","authors":"Samarth Mishra, Pengkai Zhu, Venkatesh Saligrama","doi":"10.48550/arXiv.2204.08090","DOIUrl":"https://doi.org/10.48550/arXiv.2204.08090","url":null,"abstract":"We propose Recognition as Part Composition (RPC), an image encoding approach inspired by human cognition. It is based on the cognitive theory that humans recognize complex objects by components, and that they build a small compact vocabulary of concepts to represent each instance with. RPC encodes images by first decomposing them into salient parts, and then encoding each part as a mixture of a small number of prototypes, each representing a certain concept. We find that this type of learning inspired by human cognition can overcome hurdles faced by deep convolutional networks in low-shot generalization tasks, like zero-shot learning, few-shot learning and unsupervised domain adaptation. Furthermore, we find a classifier using an RPC image encoder is fairly robust to adversarial attacks, that deep neural networks are known to be prone to. Given that our image encoding principle is based on human cognition, one would expect the encodings to be interpretable by humans, which we find to be the case via crowd-sourcing experiments. Finally, we propose an application of these interpretable encodings in the form of generating synthetic attribute annotations for evaluating zero-shot learning methods on new datasets.","PeriodicalId":13426,"journal":{"name":"IEEE Transactions on Pattern Analysis and Machine Intelligence","volume":" ","pages":""},"PeriodicalIF":23.6,"publicationDate":"2022-04-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"46674884","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
There has been rapid progress recently on 3D human rendering, including novel view synthesis and pose animation, based on the advances of neural radiance fields (NeRF). However, most existing methods focus on person-specific training and their training typically requires multi-view videos. This paper deals with a new challenging task - rendering novel views and novel poses for a person unseen in training, using only multiview still images as input without videos. For this task, we propose a simple yet surprisingly effective method to train a generalizable NeRF with multiview images as conditional input. The key ingredient is a dedicated representation combining a canonical NeRF and a volume deformation scheme. Using a canonical space enables our method to learn shared properties of human and easily generalize to different people. Volume deformation is used to connect the canonical space with input and target images and query image features for radiance and density prediction. We leverage the parametric 3D human model fitted on the input images to derive the deformation, which works quite well in practice when combined with our canonical NeRF. The experiments on both real and synthetic data with the novel view synthesis and pose animation tasks collectively demonstrate the efficacy of our method.
{"title":"MPS-NeRF: Generalizable 3D Human Rendering from Multiview Images","authors":"Xiangjun Gao, Jiaolong Yang, Jongyoo Kim, Sida Peng, Zicheng Liu, Xin Tong","doi":"10.48550/arXiv.2203.16875","DOIUrl":"https://doi.org/10.48550/arXiv.2203.16875","url":null,"abstract":"There has been rapid progress recently on 3D human rendering, including novel view synthesis and pose animation, based on the advances of neural radiance fields (NeRF). However, most existing methods focus on person-specific training and their training typically requires multi-view videos. This paper deals with a new challenging task - rendering novel views and novel poses for a person unseen in training, using only multiview still images as input without videos. For this task, we propose a simple yet surprisingly effective method to train a generalizable NeRF with multiview images as conditional input. The key ingredient is a dedicated representation combining a canonical NeRF and a volume deformation scheme. Using a canonical space enables our method to learn shared properties of human and easily generalize to different people. Volume deformation is used to connect the canonical space with input and target images and query image features for radiance and density prediction. We leverage the parametric 3D human model fitted on the input images to derive the deformation, which works quite well in practice when combined with our canonical NeRF. The experiments on both real and synthetic data with the novel view synthesis and pose animation tasks collectively demonstrate the efficacy of our method.","PeriodicalId":13426,"journal":{"name":"IEEE Transactions on Pattern Analysis and Machine Intelligence","volume":" ","pages":""},"PeriodicalIF":23.6,"publicationDate":"2022-03-31","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"42950994","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
京公网安备 11010802042870号
京ICP备2023020795号-1
扫码关注我们
求助内容：
应助结果提醒方式：