VideoMAC: Video Masked Autoencoders Meet ConvNets
Open accessAbstract
Recently, the advancement of self-supervised learning techniques, like masked autoencoders (MAE), has greatly influenced visual representation learning for images and videos. Nevertheless, it is worth noting that the predomi-nant approaches in existing masked image / video modeling rely excessively on resource-intensive vision transformers (ViTs) as the feature encoder. In this paper, we propose a new approach termed as VideoMAC, which combines video masked autoencoders with resource-friendly Con-vNets. Specifically, VideoMAC employs symmetric masking on randomly sampled pairs of video frames. To prevent the issue of mask pattern dissipation, we utilize ConvNets which are implemented with sparse convolutional operators as en-coders. Simultaneously, we present a simple yet effective masked video modeling (MVM) approach, a dual encoder architecture comprising an online encoder and an exponential moving average target encoder, aimed to facilitate inter-frame reconstruction consistency in videos. Additionally, we demonstrate that VideoMAC, empowering classical (ResNet) / modern (ConvNeXt) convolutional encoders to harness the benefits of MVM, outperforms ViT-based approaches on downstream tasks, including video object segmentation (+5.2% /6.4% J&F), body part propagation (+6.3% /3.1% mIoU), and human pose tracking (+10.2% / 11.1% PCK@0.1).
The claims and statistics of this paper haven’t been checked yet.
Run the check — free, no account needed. It reads the open-access text and checks the sources it cites.
Integrity signals
Registry
Cites
Statistics
Claims the paper makes
The claims and statistics of this paper haven’t been checked yet.
Run the check — free, no account needed. It reads the open-access text and checks the sources it cites.
Claims other papers make about this one
Each is checked against this paper’s own text: does it say what the citing paper says it does?
Masked autoencoding can be combined with convolutional encoders, while requiring attention to the interaction between masking and convolutional architectures.
“Related developments in computer vision, such as ConvNeXt V2 and VideoMAC, also show that masked autoencoding can be combined with convolutional encoders, while requiring attention to the interaction between masking and convolutional architectures43,44.”
From Siamese masked reconstruction with temporal rotation consistency for inertial-based human activity recognition (2026) · cites this paper as [44]
Self-check
Percentages, sums and reported differences re-computed from the paper’s own numbers.
The self-check needs the paper’s full text.
Cites
Deep Residual Learning for Image Recognition
He, K., Zhang, X., Ren, S., & Sun, J. (2016) · 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR) · doi:10.1109/cvpr.2016.90 · Cited by 235,474 · cited 5× · see citing works
“Their performance is even inferior to MIM methods [23, 30], and it lags significantly behind supervised methods [21, 36].”
BackgroundMethodInfluentialImageNet classification with deep convolutional neural networks
Krizhevsky, A., Sutskever, I., & Hinton, G. E. (2017) · Communications of the ACM · doi:10.1145/3065386 · Cited by 112,155 · cited 1× · see citing works
“One might first consider ConvNets [21, 28, 29, 36, 50, 53], characterized by their low resource consumption and built-in hierarchical structure, as an ideal solution to tackle the aforementioned issues.”
BackgroundVery Deep Convolutional Networks for Large-Scale Image Recognition
Simonyan, K., & Zisserman, A. (2014) · arXiv · doi:10.48550/arxiv.1409.1556 · Cited by 74,248 · cited 1× · see citing works
“One might first consider ConvNets [21, 28, 29, 36, 50, 53], characterized by their low resource consumption and built-in hierarchical structure, as an ideal solution to tackle the aforementioned issues.”
BackgroundImageNet: A large-scale hierarchical image database
Deng, J., Dong, W., Socher, R., Li, L.-J., Li, K., & Fei-Fei, L. (2009) · 2009 IEEE Conference on Computer Vision and Pattern Recognition · doi:10.1109/cvpr.2009.5206848 · Cited by 65,333 · cited 3× · see citing works
“Specifically, our backbone is pre-trained on video data and then fine-tuned on ImageNet1K [11] (Table 5).”
BackgroundMethodResultInfluentialGradient-based learning applied to document recognition
Lecun, Y., Bottou, L., Bengio, Y., & Haffner, P. (1998) · Proceedings of the IEEE · doi:10.1109/5.726791 · Cited by 58,805 · cited 1× · see citing works
“One might first consider ConvNets [21, 28, 29, 36, 50, 53], characterized by their low resource consumption and built-in hierarchical structure, as an ideal solution to tackle the aforementioned issues.”
BackgroundBERT: Pre-training of Deep Bidirectional Transformers for Language Understanding
Devlin, J., Chang, M.-W., Lee, K., & Toutanova, K. (2019) · Proceedings of the 2019 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, Volume 1 (Long and Short Papers) · doi:10.18653/v1/n19-1423 · Cited by 48,293 · cited 2× · see citing works
“The success of masked language modeling in NLP [3, 10, 12, 13, 47] has catalyzed a rising interest in masked image modeling (MIM) for visual pre-training.”
BackgroundSwin Transformer: Hierarchical Vision Transformer using Shifted Windows
Liu, Z., Lin, Y., Cao, Y., Hu, H., Wei, Y., Zhang, Z., Lin, S., & Guo, B. (2021) · 2021 IEEE/CVF International Conference on Computer Vision (ICCV) · doi:10.1109/iccv48922.2021.00986 · Cited by 35,534 · cited 2× · see citing works
“However, previous MVM approaches [16, 56], confined by the plain vision transformer design devoid of a hierarchical structure tend to result in inadequate spatial modeling capabilities for local features [17, 35, 49, 62, 70].”
BackgroundMethodLanguage Models are Few-Shot Learners
Brown, T. B., Mann, B., Ryder, N., Subbiah, M., Kaplan, J., Dhariwal, P., Neelakantan, A., Shyam, P., Sastry, G., Askell, A., et al. (2020) · arXiv · doi:10.48550/arxiv.2005.14165 · Cited by 23,372 · cited 2× · see citing works
“The success of masked language modeling in NLP [3, 10, 12, 13, 47] has catalyzed a rising interest in masked image modeling (MIM) for visual pre-training.”
BackgroundAn Image is Worth 16x16 Words: Transformers for Image Recognition at Scale
Dosovitskiy, A., Beyer, L., Kolesnikov, A., Weissenborn, D., Zhai, X., Unterthiner, T., Dehghani, M., Minderer, M., Heigold, G., Gelly, S., et al. (2020) · arXiv · doi:10.48550/arxiv.2010.11929 · Cited by 21,793 · cited 3× · see citing works
“Importantly, all aforementioned methods utilize computationally intensive, non-hierarchical ViT [14] as their encoders.”
BackgroundMethodEfficientNet: Rethinking Model Scaling for Convolutional Neural Networks
Tan, M., & Le, Q. V. (2019) · arXiv · doi:10.48550/arxiv.1905.11946 · Cited by 17,072 · cited 1× · see citing works
“One might first consider ConvNets [21, 28, 29, 36, 50, 53], characterized by their low resource consumption and built-in hierarchical structure, as an ideal solution to tackle the aforementioned issues.”
BackgroundLearning Transferable Visual Models From Natural Language Supervision
Radford, A., Kim, J. W., Hallacy, C., Ramesh, A., Goh, G., Agarwal, S., Sastry, G., Askell, A., Mishkin, P., Clark, J., et al. (2021) · arXiv · doi:10.48550/arxiv.2103.00020 · Cited by 15,293 · cited 1× · see citing works
“A key advantage of SSL [2, 5–8, 18, 22, 48, 57, 69, 72] is its independence from costly human annotations, thereby providing robust prior knowledge across various visual tasks.”
BackgroundMomentum Contrast for Unsupervised Visual Representation Learning
He, K., Fan, H., Wu, Y., Xie, S., & Girshick, R. (2020) · 2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) · doi:10.1109/cvpr42600.2020.00975 · Cited by 13,347 · cited 2× · see citing works
“A key advantage of SSL [2, 5–8, 18, 22, 48, 57, 69, 72] is its independence from costly human annotations, thereby providing robust prior knowledge across various visual tasks.”
BackgroundA simple framework for contrastive learning of visual representations
Chen, T., Kornblith, S., Norouzi, M., & Hinton, G. (2020) · arXiv · doi:10.48550/arxiv.2002.05709 · Cited by 10,857 · see citing works
Self-citationLiu, Z., Mao, H., Wu, C.-Y., Feichtenhofer, C., Darrell, T., & Xie, S. (2022) · 2022 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) · doi:10.1109/cvpr52688.2022.01167 · Cited by 8,494 · cited 6× · see citing works
“There are two naive solutions: (i) the instigation of learnable masked markers at the input stage, and (ii) the transformation of the hierarchical convolutional network into an isotropic structure [36] that mirrors the ViT-style technique (Table 2).”
BackgroundMethodInfluentialMasked Autoencoders Are Scalable Vision Learners
He, K., Chen, X., Xie, S., Li, Y., Dollar, P., & Girshick, R. (2022) · 2022 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) · doi:10.1109/cvpr52688.2022.01553 · Cited by 8,489 · cited 6× · see citing works
“Their performance is even inferior to MIM methods [23, 30], and it lags significantly behind supervised methods [21, 36].”
BackgroundMethodInfluentialEmerging Properties in Self-Supervised Vision Transformers
Caron, M., Touvron, H., Misra, I., Jegou, H., Mairal, J., Bojanowski, P., & Joulin, A. (2021) · 2021 IEEE/CVF International Conference on Computer Vision (ICCV) · doi:10.1109/iccv48922.2021.00951 · Cited by 5,638 · cited 1× · see citing works
“Our evaluation methodology aligns with [6, 25], and we report the typical evaluation metrics [45], including the mean region similarity J m , the mean boundary accuracy F m , and their average values J & F m .”
MethodSGDR: Stochastic Gradient Descent with Warm Restarts
Loshchilov, I., & Hutter, F. (2016) · arXiv · doi:10.48550/arxiv.1608.03983 · Cited by 5,309 · cited 1× · see citing works
“Cosine decay [38] is applied to adjust the learning rate.”
MethodPyramid Vision Transformer: A Versatile Backbone for Dense Prediction without Convolutions
Wang, W., Xie, E., Li, X., Fan, D.-P., Song, K., Liang, D., Lu, T., Luo, P., & Shao, L. (2021) · 2021 IEEE/CVF International Conference on Computer Vision (ICCV) · doi:10.1109/iccv48922.2021.00061 · Cited by 5,259 · cited 2× · see citing works
“However, previous MVM approaches [16, 56], confined by the plain vision transformer design devoid of a hierarchical structure tend to result in inadequate spatial modeling capabilities for local features [17, 35, 49, 62, 70].”
BackgroundMethodBootstrap Your Own Latent: A New Approach to Self-Supervised Learning
Grill, J.-B., Strub, F., Altché, F., Tallec, C., Richemond, P. H., Buchatskaya, E., Doersch, C., Pires, B. A., Guo, Z. D., Azar, M. G., et al. (2020) · arXiv · doi:10.48550/arxiv.2006.07733 · Cited by 3,846 · cited 1× · see citing works
“A key advantage of SSL [2, 5–8, 18, 22, 48, 57, 69, 72] is its independence from costly human annotations, thereby providing robust prior knowledge across various visual tasks.”
BackgroundExploring Simple Siamese Representation Learning
Chen, X., & He, K. (2021) · 2021 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) · doi:10.1109/cvpr46437.2021.01549 · Cited by 3,725 · cited 1× · see citing works
“A key advantage of SSL [2, 5–8, 18, 22, 48, 57, 69, 72] is its independence from costly human annotations, thereby providing robust prior knowledge across various visual tasks.”
BackgroundThe Kinetics Human Action Video DatasetPreprint
Kay, W., Carreira, J., Simonyan, K., Zhang, B., Hillier, C., Vijayanarasimhan, S., Viola, F., Green, T., Back, T., Natsev, P., et al. (2017) · arXiv (Cornell University) · doi:10.48550/arxiv.1705.06950 · Cited by 3,711 · cited 3× · see citing works
“K400: Kinetics-400 [27].”
BackgroundMethodUnsupervised Learning of Visual Features by Contrasting Cluster Assignments
Caron, M., Misra, I., Mairal, J., Goyal, P., Bojanowski, P., & Joulin, A. (2020) · arXiv · doi:10.48550/arxiv.2006.09882 · Cited by 2,499 · cited 1× · see citing works
“A key advantage of SSL [2, 5–8, 18, 22, 48, 57, 69, 72] is its independence from costly human annotations, thereby providing robust prior knowledge across various visual tasks.”
BackgroundLiu, Z., Ning, J., Cao, Y., Wei, Y., Zhang, Z., Lin, S., & Hu, H. (2022) · 2022 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) · doi:10.1109/cvpr52688.2022.00320 · Cited by 2,294 · cited 1× · see citing works
“Furthermore, these ViT-based MVM strategies may encounter challenges with inadequate local feature modeling when dealing with dense tasks such as segmentation and detection due to their non-hierarchical structure [35, 37, 62, 70].”
Method4D Spatio-Temporal ConvNets: Minkowski Convolutional Neural Networks
Choy, C., Gwak, J., & Savarese, S. (2019) · IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings - IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings · doi:10.1109/cvpr.2019.00319 · Cited by 2,067 · cited 1× · see citing works
“ViT Backbone MAE-ST [16] 14 with the MinkowskiEngine sparse convolution library [9].”
MethodA Benchmark Dataset and Evaluation Methodology for Video Object Segmentation
Perazzi, F., Pont-Tuset, J., McWilliams, B., Gool, L. V., Groß, M. H., & Sorkine‐Hornung, A. (2016) · IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings - IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings · doi:10.1109/cvpr.2016.85 · Cited by 2,039 · cited 1× · see citing works
“Our evaluation methodology aligns with [6, 25], and we report the typical evaluation metrics [45], including the mean region similarity J m , the mean boundary accuracy F m , and their average values J & F m .”
MethodConvNeXt V2: Co-designing and Scaling ConvNets with Masked Autoencoders
Woo, S., Debnath, S., Hu, R., Chen, X., Liu, Z., Kweon, I. S., & Xie, S. (2023) · 2023 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) · doi:10.1109/cvpr52729.2023.01548 · Cited by 1,791 · cited 7× · see citing works
“To efficiently incorporate mask modeling into ConvNets, we opt for sparse convolution over dense convolution to re-structure the original ConvNets, an approach validated in recent studies [54, 64] to effectively maintain the structure of the masked region when utilizing ConvNets.”
BackgroundMethodInfluentialBEiT: BERT Pre-Training of Image Transformers
Bao, H., Dong, L., Piao, S., & Wei, F. (2021) · arXiv · doi:10.48550/arxiv.2106.08254 · Cited by 1,392 · cited 3× · see citing works
“A key advantage of SSL [2, 5–8, 18, 22, 48, 57, 69, 72] is its independence from costly human annotations, thereby providing robust prior knowledge across various visual tasks.”
BackgroundSimMIM: a Simple Framework for Masked Image Modeling
Xie, Z., Zhang, Z., Cao, Y., Lin, Y., Bao, J., Yao, Z., Dai, Q., & Hu, H. (2022) · IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings - IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings · doi:10.1109/cvpr52688.2022.00943 · Cited by 1,348 · cited 2× · see citing works
“This strategy involves masking or corrupting a portion of information within an image and then requiring the model to recover the masked portion, encouraging the model to learn useful image representations [2, 23, 67].”
BackgroundUnified Language Model Pre-training for Natural Language Understanding and Generation
Dong, L., Yang, N., Wang, W., Wei, F., Liu, X., Wang, Y., Gao, J., Zhou, M., & Hon, H.-W. (2019) · arXiv · doi:10.48550/arxiv.1905.03197 · Cited by 1,225 · cited 1× · see citing works
“The success of masked language modeling in NLP [3, 10, 12, 13, 47] has catalyzed a rising interest in masked image modeling (MIM) for visual pre-training.”
BackgroundOne-Shot Video Object Segmentation
Caelles, S., Maninis, K.-K., Pont-Tuset, J., Leal-Taixe, L., Cremers, D., & Van Gool, L. (2017) · 2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR) · doi:10.1109/cvpr.2017.565 · Cited by 927 · cited 1× · see citing works
“Recently, researchers have embarked on exploring MVM approaches [16, 19, 65, 66] to comprehend unsupervised video representations, exhibiting effectiveness across a multitude of downstream tasks [4, 25, 31, 32, 40–44, 60, 61, 71, 74, 75].”
BackgroundVideoMAE: Masked Autoencoders are Data-Efficient Learners for Self-Supervised Video Pre-Training
Tong, Z., Song, Y., Wang, J., & Wang, L. (2022) · arXiv · doi:10.48550/arxiv.2203.12602 · Cited by 909 · cited 10× · see citing works
“These models solely employ the visible patches as the exclusive input to the encoder, thereby allowing ViT-based approaches [16, 20, 56, 65] to amplify the efficiency and efficacy of model pre-training by concentrating dedicatedly on unmasked regions during this phase.”
BackgroundMethodInfluentialTowards Understanding Action Recognition
Jhuang, H., Gall, J., Zuffi, S., Schmid, C., & Black, M. J. (2013) · IEEE/CVF International Conference on Computer Vision (ICCV) · doi:10.1109/iccv.2013.396 · Cited by 906 · cited 4× · see citing works
“Notably, VideoMAC exhibits superior performance on a range of down-stream tasks, e.g ., video object segmentation on DAVIS17 [46], body part propagation on VIP [73], and human pose tracking on JHMDB [26], compared to ViT-based methods [16, 20].”
MethodInfluentialFlow-Guided Feature Aggregation for Video Object Detection
Zhu, X., Wang, Y., Dai, J., Yuan, L., & Wei, Y. (2017) · 2017 IEEE International Conference on Computer Vision (ICCV) · doi:10.1109/iccv.2017.52 · Cited by 729 · cited 1× · see citing works
“Recently, researchers have embarked on exploring MVM approaches [16, 19, 65, 66] to comprehend unsupervised video representations, exhibiting effectiveness across a multitude of downstream tasks [4, 25, 31, 32, 40–44, 60, 61, 71, 74, 75].”
BackgroundDense Contrastive Learning for Self-Supervised Visual Pre-Training
Wang, X., Zhang, R., Shen, C., Kong, T., & Li, L. (2021) · 2021 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) · doi:10.1109/cvpr46437.2021.00304 · Cited by 660 · see citing works
YouTube-VOS: Sequence-to-Sequence Video Object Segmentation
Xu, N., Yang, L., Fan, Y., Yang, J., Yue, D., Liang, Y., Price, B. L., Cohen, S. D., & Huang, T. S. (2018) · Lecture notes in computer science · doi:10.1007/978-3-030-01228-1_36 · Cited by 618 · cited 4× · see citing works
“YT18: YouTube-VOS 2018 [68].”
BackgroundMethodInfluentialEVA: Exploring the Limits of Masked Visual Representation Learning at Scale
Fang, Y., Wang, W., Xie, B., Sun, Q., Wu, L., Wang, X., Huang, T., Wang, X., & Cao, Y. (2023) · IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings - IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings · doi:10.1109/cvpr52729.2023.01855 · Cited by 540 · see citing works
VideoMAE V2: Scaling Video Masked Autoencoders with Dual Masking
Wang, L., Huang, B., Zhao, Z., Tong, Z., He, Y., Wang, Y., Wang, Y., & Qiao, Y. (2023) · 2023 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) · doi:10.1109/cvpr52729.2023.01398 · Cited by 536 · cited 2× · see citing works
“Moreover, the uniform reliance on ViT in existing MVM methods has resulted in significant computational resource consumption [16, 56, 58].”
BackgroundMethodHiera: A Hierarchical Vision Transformer without the Bells-and-Whistles
Ryali, C. K., Hu, Y.-T., Bolya, D., et al. (2023) · International Conference on Machine Learning · doi:10.48550/arxiv.2306.00989 · Cited by 482 · cited 1×
“However, previous MVM approaches [16, 56], confined by the plain vision transformer design devoid of a hierarchical structure tend to result in inadequate spatial modeling capabilities for local features [17, 35, 49, 62, 70].”
BackgroundMasked Siamese Networks for Label-Efficient Learning
Assran, M., Caron, M., Misra, I., et al. (2022) · European Conference on Computer Vision · doi:10.48550/arxiv.2204.07141 · Cited by 456 · cited 1×
“Spurred by the success of masked language modeling in nat-ural language processing [3, 10, 12], transformer-based visual models, notably vision transformers (ViT) [14] adopt masked image modeling (MIM) are attracting attention [1, 2, 15, 23, 34, 67].”
BackgroundMasked Autoencoders As Spatiotemporal Learners
Feichtenhofer, C., Fan, H., Li, Y., & He, K. (2022) · arXiv · doi:10.48550/arxiv.2205.09113 · Cited by 345 · cited 10× · see citing works
“However, previous MVM approaches [16, 56], confined by the plain vision transformer design devoid of a hierarchical structure tend to result in inadequate spatial modeling capabilities for local features [17, 35, 49, 62, 70].”
BackgroundMethodInfluentialZero-Shot Video Object Segmentation via Attentive Graph Neural Networks
Wang, W., Lu, X., Shen, J., Crandall, D., & Shao, L. (2019) · IEEE/CVF International Conference on Computer Vision (ICCV) · doi:10.1109/iccv.2019.00933 · Cited by 312 · cited 1× · see citing works
“Recently, researchers have embarked on exploring MVM approaches [16, 19, 65, 66] to comprehend unsupervised video representations, exhibiting effectiveness across a multitude of downstream tasks [4, 25, 31, 32, 40–44, 60, 61, 71, 74, 75].”
BackgroundVideo Object Segmentation with Episodic Graph Memory Networks
Lu, X., Wang, W., Danelljan, M., Zhou, T., Shen, J., & Gool, L. V. (2020) · Lecture notes in computer science · doi:10.1007/978-3-030-58580-8_39 · Cited by 288 · cited 1× · see citing works
“Recently, researchers have embarked on exploring MVM approaches [16, 19, 65, 66] to comprehend unsupervised video representations, exhibiting effectiveness across a multitude of downstream tasks [4, 25, 31, 32, 40–44, 60, 61, 71, 74, 75].”
BackgroundA Survey on Deep Learning Technique for Video Segmentation
Zhou, T., Porikli, F., Crandall, D., Gool, L. V., & Wang, W. (2022) · IEEE Transactions on Pattern Analysis and Machine Intelligence · doi:10.1109/tpami.2022.3225573 · Cited by 273 · cited 1× · see citing works
“Recently, researchers have embarked on exploring MVM approaches [16, 19, 65, 66] to comprehend unsupervised video representations, exhibiting effectiveness across a multitude of downstream tasks [4, 25, 31, 32, 40–44, 60, 61, 71, 74, 75].”
BackgroundLearning Unsupervised Video Object Segmentation Through Visual Attention
Wang, W., Song, H., Shuyang, Z., Shen, J., Zhao, S., Hoi, S. C. H., & Ling, H. (2019) · IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings - IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings · doi:10.1109/cvpr.2019.00318 · Cited by 240 · see citing works
Non-Salient Region Object Mining for Weakly Supervised Semantic Segmentation
Yao, Y., Chen, T., Xie, G.-S., Zhang, C., Shen, F., Wu, Q., Tang, Z., & Zhang, J. (2021) · IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings - IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings · doi:10.1109/cvpr46437.2021.00265 · Cited by 222 · cited 1× · see citing works
“A key advantage of SSL [2, 5–8, 18, 22, 48, 57, 69, 72] is its independence from costly human annotations, thereby providing robust prior knowledge across various visual tasks.”
BackgroundSelf-citationTraining data-efficient image transformers & distillation through attention
Touvron, H., Cord, M., Douze, M., Massa, F., Sablayrolles, A., & Jégou, H. (2020) · arXiv · doi:10.48550/arxiv.2012.12877 · Cited by 182 · cited 1× · see citing works
“A key advantage of SSL [2, 5–8, 18, 22, 48, 57, 69, 72] is its independence from costly human annotations, thereby providing robust prior knowledge across various visual tasks.”
BackgroundLearning Video Object Segmentation From Unlabeled Videos
Lu, X., Wang, W., Shen, J., Tai, Y., Crandall, D., & Hoi, S. C. H. (2020) · IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings - IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings · doi:10.1109/cvpr42600.2020.00898 · Cited by 182 · see citing works
Space-Time Correspondence as a Contrastive Random WalkPreprint
Jabri, A., Owens, A., & Efros, A. A. (2020) · arXiv (Cornell University) · doi:10.48550/arxiv.2006.14613 · Cited by 167 · cited 4× · see citing works
“We adhere to the same parameters as [25], adjusting the video frames to 560 × 560 dimensions.”
BackgroundMethodInfluentialMaskViT: Masked Visual Pre-Training for Video Prediction
Gupta, A., Tian, S., Zhang, Y.-Z., et al. (2022) · International Conference on Learning Representations · doi:10.48550/arxiv.2206.11894 · Cited by 160 · cited 2×
“It is worth noting that while the isotropic ViT is well suited for the implementation of MVM methods [16, 19, 56], this is exceptionally difficult for ConvNets.”
BackgroundMethodThin-Slicing Network: A Deep Structured Model for Pose Estimation in Videos
Song, J., Wang, L., Gool, L. V., & Hilliges, O. (2017) · IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings - IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings · doi:10.1109/cvpr.2017.590 · Cited by 124 · cited 1× · see citing works
“Conforming to the protocol delineated in [25], we resize video frames to a resolution of 320 × 320 , and employ the Possibility of Correct Keypoints (PCK) metric [51] for quantitative evaluation.”
MethodGupta, A., Wu, J., Deng, J., & Fei-Fei, L. (2023) · Neural Information Processing Systems · doi:10.48550/arxiv.2305.14344 · Cited by 116 · cited 9×
“These models solely employ the visible patches as the exclusive input to the encoder, thereby allowing ViT-based approaches [16, 20, 56, 65] to amplify the efficiency and efficacy of model pre-training by concentrating dedicatedly on unmasked regions during this phase.”
BackgroundMethodInfluentialObject discovery and representation networks
H'enaff, O. J., Koppula, S., Shelhamer, E., Zoran, D., Jaegle, A., Zisserman, A., Carreira, J., & Arandjelovi'c, R. (2022) · European Conference on Computer Vision · doi:10.48550/arxiv.2203.08777 · Cited by 101
MixMAE: Mixed and Masked Autoencoder for Efficient Pretraining of Hierarchical Vision Transformers
Liu, J., Huang, X., Zheng, J., Liu, Y., & Li, H. (2023) · IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings - IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings · doi:10.1109/cvpr52729.2023.00605 · Cited by 93 · cited 1× · see citing works
“Unlike previous methods that utilize hierarchical decoders [54, 55] or transformer decoders[16, 33, 56], our VideoMAC implements a lightweight decoder derived from the ConvNeXt block [36].”
MethodMCMAE: Masked Convolution Meets Masked Autoencoders
Gao, P., Ma, T., Li, H., et al. (2022) · Neural Information Processing Systems · doi:10.52202/068431-2582 · Cited by 86 · cited 1×
“However, previous MVM approaches [16, 56], confined by the plain vision transformer design devoid of a hierarchical structure tend to result in inadequate spatial modeling capabilities for local features [17, 35, 49, 62, 70].”
BackgroundWang, R., Chen, D., Wu, Z., Chen, Y., Dai, X., Liu, M., Yuan, L., & Jiang, Y. (2023) · IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings - IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings · doi:10.1109/cvpr52729.2023.00611 · Cited by 84 · cited 1× · see citing works
“However, current MVM methodologies [20, 56, 58, 59] primarily depend on the isotropic ViT, a visual backbone compatible with masking modeling.”
MethodHierarchical Feature Alignment Network for Unsupervised Video Object Segmentation
Pei, G.-S., Shen, F.-M., Yao, Y., et al. (2022) · European Conference on Computer Vision · doi:10.48550/arxiv.2207.08485 · Cited by 75
Self-citationLocality-Aware Inter-and Intra-Video Reconstruction for Self-Supervised Correspondence Learning
Li, L.-L., Zhou, T.-F., Wang, W.-G., Yang, L., Li, J.-W., & Yang, Y. (2022) · Computer Vision and Pattern Recognition · doi:10.1109/cvpr52688.2022.00852 · Cited by 63 · cited 1× · see citing works
“Recently, researchers have embarked on exploring MVM approaches [16, 19, 65, 66] to comprehend unsupervised video representations, exhibiting effectiveness across a multitude of downstream tasks [4, 25, 31, 32, 40–44, 60, 61, 71, 74, 75].”
BackgroundDesigning BERT for Convolutional Networks: Sparse and Hierarchical Masked ModelingPreprint
Tian, K., Jiang, Y., Diao, Q., Chen, L., Wang, L., & Yuan, Z. (2023) · arXiv (Cornell University) · doi:10.48550/arxiv.2301.03580 · Cited by 58 · cited 4× · see citing works
“To efficiently incorporate mask modeling into ConvNets, we opt for sparse convolution over dense convolution to re-structure the original ConvNets, an approach validated in recent studies [54, 64] to effectively maintain the structure of the masked region when utilizing ConvNets.”
MethodInfluentialScalable Video Object Segmentation with Simplified Framework
Wu, Q., Yang, T.-Y., Wu, W., & Chan, A. B. (2023) · IEEE International Conference on Computer Vision · doi:10.1109/iccv51070.2023.01276 · Cited by 52 · cited 1× · see citing works
“Recently, researchers have embarked on exploring MVM approaches [16, 19, 65, 66] to comprehend unsupervised video representations, exhibiting effectiveness across a multitude of downstream tasks [4, 25, 31, 32, 40–44, 60, 61, 71, 74, 75].”
BackgroundWu, Q., Yang, T.-Y., Liu, Z., Lin, W., Wu, B., & Chan, A. B. (2023) · doi:10.48550/arxiv.2304.00571 · Cited by 50 · cited 6×
“These models solely employ the visible patches as the exclusive input to the encoder, thereby allowing ViT-based approaches [16, 20, 56, 65] to amplify the efficiency and efficacy of model pre-training by concentrating dedicatedly on unmasked regions during this phase.”
BackgroundMethodInfluentialIntegrally Pre-Trained Transformer Pyramid Networks
Tian, Y., Xie, L.-X., Wang, Z.-Z., Wei, L.-H., Zhang, X.-P., Jiao, J.-B., Wang, Y., Tian, Q., & Ye, Q.-X. (2023) · Computer Vision and Pattern Recognition · doi:10.1109/cvpr52729.2023.01785 · Cited by 49 · cited 1× · see citing works
“Unlike previous methods that utilize hierarchical decoders [54, 55] or transformer decoders[16, 33, 56], our VideoMAC implements a lightweight decoder derived from the ConvNeXt block [36].”
MethodBoosting Video Object Segmentation via Space-Time Correspondence Learning
Zhang, Y., Li, L., Wang, W., Xie, R., Song, L., & Zhang, W. (2023) · IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings - IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings · doi:10.1109/cvpr52729.2023.00223 · Cited by 48 · cited 1× · see citing works
“Recently, researchers have embarked on exploring MVM approaches [16, 19, 65, 66] to comprehend unsupervised video representations, exhibiting effectiveness across a multitude of downstream tasks [4, 25, 31, 32, 40–44, 60, 61, 71, 74, 75].”
BackgroundAdaptive Temporal Encoding Network for Video Instance-level Human Parsing
Zhou, Q., Liang, X., Gong, K., & Lin, L. (2018) · ACM International Conference on Multimedia (ACM MM) · doi:10.1145/3240508.3240660 · Cited by 46 · cited 5× · see citing works
“…(RN50), respectively, in comparison with the top-performing SiamMAE [20] in MVM. Significantly, our RN50-based model outperforms CNXv2-S on the VIP dataset [73] due to the higher dimensionality of RN50 features (2048 vs. 768), which benefits complex tasks and improves accuracy in mask propagation.”
MethodInfluentialImproving Pixel-based MIM by Reducing Wasted Modeling Capability
Liu, Y., Zhang, S., Chen, J., Yu, Z., Chen, K., & Lin, D. (2023) · IEEE/CVF International Conference on Computer Vision (ICCV) · doi:10.1109/iccv51070.2023.00494 · Cited by 39 · cited 1× · see citing works
“Spurred by the success of masked language modeling in nat-ural language processing [3, 10, 12], transformer-based visual models, notably vision transformers (ViT) [14] adopt masked image modeling (MIM) are attracting attention [1, 2, 15, 23, 34, 67].”
BackgroundCrOC: Cross-View Online Clustering for Dense Visual Representation Learning
Stegmüller, T., Lebailly, T., Bozorgtabar, B., Tuytelaars, T., & Thiran, J. (2023) · Computer Vision and Pattern Recognition · doi:10.1109/cvpr52729.2023.00676 · Cited by 27 · see citing works
Unified Mask Embedding and Correspondence Learning for Self-Supervised Video Segmentation
Li, L., Wang, W., Zhou, T., Li, J., & Yang, Y. (2023) · IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings - IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings · doi:10.1109/cvpr52729.2023.01794 · Cited by 26 · cited 1× · see citing works
“Recently, researchers have embarked on exploring MVM approaches [16, 19, 65, 66] to comprehend unsupervised video representations, exhibiting effectiveness across a multitude of downstream tasks [4, 25, 31, 32, 40–44, 60, 61, 71, 74, 75].”
BackgroundExploring The Role of Mean Teachers in Self-supervised Masked Auto-Encoders
Lee, Y., Willette, J., Kim, J., et al. (2022) · arXiv.org · doi:10.48550/arxiv.2210.02077 · Cited by 23 · cited 1×
“Their performance is even inferior to MIM methods [23, 30], and it lags significantly behind supervised methods [21, 36].”
BackgroundHierarchical Co-Attention Propagation Network for Zero-Shot Video Object Segmentation
Pei, G.-S., Yao, Y., Shen, F.-M., et al. (2023) · IEEE Transactions on Image Processing · doi:10.1109/tip.2023.3267244 · Cited by 18 · cited 1× · see citing works
“Recently, researchers have embarked on exploring MVM approaches [16, 19, 65, 66] to comprehend unsupervised video representations, exhibiting effectiveness across a multitude of downstream tasks [4, 25, 31, 32, 40–44, 60, 61, 71, 74, 75].”
BackgroundSelf-citationHierarchical Graph Pattern Understanding for Zero-Shot Video Object Segmentation
Pei, G.-S., Shen, F.-M., Yao, Y., et al. (2023) · IEEE Transactions on Image Processing · doi:10.1109/tip.2023.3326395 · Cited by 16 · see citing works
Self-citationThe 2017 DAVIS Challenge on Video Object Segmentation
Pont-Tuset, J., Perazzi, F., Caelles, S., Arbeláez, P., Sorkine-Hornung, A., & Van Gool, L. (2017) · arXiv · doi:10.48550/arxiv.1704.00675 · Cited by 13 · cited 5× · see citing works
“We evaluate VideoMAC on the popular DAVIS17 [46] dataset, which includes a validation set comprising 30 videos.”
BackgroundMethodInfluential
Flagged first, then by citations
6 of the 76 references couldn’t be matched to a known paper.
Research works printed but not linked: 9
- Kevin Clark, Minh-Thang Luong, Quoc V Le, and Christo- pher D Manning. Electra: Pre-training text encoders as dis- criminators rather than generators. In ICLR, 2020. 1, 2
- Hueihan Jhuang, Juergen Gall, Silvia Zuffi, Cordelia Schmid, and Michael J Black. Towards understanding ac- tion recognition. In ICCV, pages 3192-3199, 2013. 1, 2, 5, 6
- Ilya Loshchilov and Frank Hutter. Decoupled weight decay regularization. In ICLR, 2019. 5
- Alec Radford, Jeffrey Wu, Rewon Child, David Luan, Dario Amodei, Ilya Sutskever, et al. Language models are unsu- pervised multitask learners. OpenAI blog, 1(8):9, 2019. 2
- Alec Radford, Jong Wook Kim, Chris Hallacy, Aditya Ramesh, Gabriel Goh, Sandhini Agarwal, Girish Sastry, Amanda Askell, Pamela Mishkin, Jack Clark, et al. Learn- ing transferable visual models from natural language super- vision. In ICML, pages 8748-8763, 2021. 2
- Hugo Touvron, Matthieu Cord, Matthijs Douze, Francisco Massa, Alexandre Sablayrolles, and Hervé Jégou. Training data-efficient image transformers & distillation through at- tention. In ICML, pages 10347-10357, 2021. 2
- Qiangqiang Wu, Tianyu Yang, Ziquan Liu, Baoyuan Wu, Ying Shan, and Antoni B Chan. Dropmae: Masked autoen- coders with spatial-attention dropout for tracking tasks. In CVPR, pages 14561-14571, 2023. 1, 2, 3, 5, 7
- Xiaosong Zhang, Yunjie Tian, Lingxi Xie, Wei Huang, Qi Dai, Qixiang Ye, and Qi Tian. Hivit: A simpler and more efficient design of hierarchical vision transformer. In ICLR, 2023. 2, 3
- Jinghao Zhou, Chen Wei, Huiyu Wang, Wei Shen, Cihang Xie, Alan Yuille, and Tao Kong. Image bert pre-training with online tokenizer. In ICLR, 2022. 2
Cited by
Sorted by citations
Masked Image Modeling: A Survey
Hondru, V., Croitoru, F.-A., Minaee, S., Ionescu, R. T., & Sebe, N. (2024) · International Journal of Computer Vision · doi:10.1007/s11263-025-02524-1 · Cited by 65 · see citing works
“Furthermore, the authors employed a third contrastive objective between the fully visible encoded text and a decoded representation (using an additional decoder) of all the image embeddings (masked and unmasked tokens).”
MGKsite: Multi-Modal Knowledge-Driven Site Selection via Intra and Inter-Modal Graph Fusion
Liang, K., Meng, L., Li, H., Liu, M., Wang, S., Zhou, S., Liu, X., & He, K. (2024) · IEEE Transactions on Multimedia · doi:10.1109/tmm.2024.3521742 · Cited by 21 · see citing works
“As described in [50], [51], [52], the fusion strategy plays an important role in multi-modal learning to generate high-quality fused features, which are investigated in different fields [53], [54], [55], [56], [57], [58], [59], [60], [61], [62], [63], [64], [65].”
Knowledge Transfer with Simulated Inter-image Erasing for Weakly Supervised Semantic Segmentation
Chen, T., Jiang, X., Pei, G., et al. (2024) · Lecture notes in computer science · doi:10.1007/978-3-031-72946-1_25 · Cited by 16 · see citing works
Self-citationSemi-Supervised Semantic Segmentation With Multi-Constraint Consistency Learning
Yin, J., Chen, T., Pei, G., et al. (2025) · IEEE Transactions on Multimedia · doi:10.1109/tmm.2025.3586111 · Cited by 15 · see citing works
Self-citationMao, J., Shen, Y., Guo, J., Yao, Y., Hua, X., & Shen, H. T. (2025) · IEEE Transactions on Multimedia · doi:10.1109/tmm.2025.3535405 · Cited by 15 · see citing works
“R ECENTLY, Vision Transformer (ViT) has gained prominence within numerous computer vision tasks [1], [2], such as image recognition [3], [4], [5], [6], [7], object detection [8], [9], [10], [11], [12], [13], semantic segmentation [14], [15], [16], [17], [18], [19], and video analysis [20], [21].”
Self-citationAbout Time: Advances, Challenges, and Outlooks of Action Understanding
Stergiou, A., & Poppe, R. W. (2025) · International Journal of Computer Vision · doi:10.1007/s11263-025-02478-4 · Cited by 15 · see citing works
ChangeTitans: Toward Remote Sensing Change Detection With Neural Memory
Yang, Z.-Y., Pei, G.-S., Yao, Y.-Z., Zhou, T.-F., Ding, L., & Shen, F.-M. (2025) · IEEE Transactions on Geoscience and Remote Sensing · doi:10.1109/tgrs.2025.3636902 · Cited by 15 · see citing works
“Achieving this goal in high-resolution (HR) imagery demands models that are simultaneously context-aware and computationally e ffi cient, a balance that remains elusive despite recent progress in deep learning [4], [5], [6], [7], [8].”
Self-citationUniAlign: Scaling Multimodal Alignment within One Unified Model
Zhou, B., Li, L., Wang, Y., Liu, H., Yao, Y., & Wang, W. (2025) · IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings - IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings · doi:10.1109/cvpr52734.2025.02760 · Cited by 10 · see citing works
“While our dependence on semantic similarity between labels may appear restrictive, it effectively harnesses available data to align modalities without requiring extensive paired datasets, showcasing U NI A LIGN ’s innovative use of soft bindings.”
Self-citationSeeing What Matters: Empowering CLIP with Patch Generation-to-Selection
Pei, G., Chen, T., Wang, Y., Cai, X., Shu, X., Zhou, T., & Yao, Y. (2025) · IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings - IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings · doi:10.1109/cvpr52734.2025.02315 · Cited by 8 · see citing works
“Masked image modeling [1, 2, 17, 23, 40, 53, 70] is a pivotal technique in computer vision that trains models to reconstruct masked image regions, reducing computational load and enhancing data efficiency by guiding models to concentrate on essential information within constrained inputs.”
Self-citationLiu, Y., Xu, Q., Wen, P., Dai, S., & Huang, Q. (2025) · IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings - IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings · doi:10.1109/cvpr52734.2025.02238 · Cited by 7 · see citing works
DepMatch: Boosting Semi-Supervised Semantic Segmentation by Exploring Depth Difference Knowledge
Yin, J., Jiang, X., Chen, T., Pei, G., Yao, Y., Shen, F., & Shen, H. T. (2026) · IEEE Transactions on Image Processing · doi:10.1109/tip.2026.3675487 · Cited by 5 · see citing works
“The significant advancement of deep learning [30], [31], [32], [33], [34], [35], [36] has laid the foundation for the development of semantic segmentation [7], [17], [37].”
Self-citationCycleACR: Cycle Modeling of Actor-Context Relations for Video Action Detection
Chen, L., Zhan, T., Song, Y., Wu, G., & Wang, L. (2025) · IEEE Transactions on Pattern Analysis and Machine Intelligence · doi:10.1109/tpami.2025.3595393 · Cited by 5 · see citing works
“Computational efficiency advances through state-space modeling [54], token redundancy reduction [55], and ConvNet-compatible masked autoencoding [56].”
EchoCardMAE: Video Masked Auto-Encoders Customized for Echocardiography
Yang, X., Xu, R., Ye, X., Wang, Z., Zhang, M., Wang, Y., Fan, X., Wang, H., Yue, Q., He, X., et al. (2025) · Lecture notes in computer science · doi:10.1007/978-3-032-05169-1_17 · Cited by 3 · see citing works
Self-Supervised Cross-View Correspondence with Predictive Cycle Consistency
Baade, A., & Chen, C. (2025) · IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings - IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings · doi:10.1109/cvpr52734.2025.01561 · Cited by 1 · see citing works
“Recently, predictive approaches have seen competitive results in object correspondence and representation learning through tasks such as image-conditioned grayscale coloration [56] and conditional masked autoen-coders [18, 38].”
Video is Worth a Thousand Images: Exploring the Latest Trends in Long Video Generation
Waseem, F., & Shahzad, M. S. (2025) · ACM Computing Surveys · doi:10.1145/3771724 · Cited by 3 · see citing works
“Transformers, introduced in 2017 [57], made inroads into image and generation via autoregressive and masked encoding.”
MAE-Based Dual Cross-Modality Relationship Learning for 6D Pose Estimation
Fu, H., Lu, H., & Li, Y. (2026) · IEEE Transactions on Consumer Electronics · doi:10.1109/tce.2026.3672868
“To address the issue of low e ff ective feature density after masking, we employ sparse convolution [19], [26], [27] to e ff ectively extract feature information from masked images.”
Low-Channel EMG Gesture Recognition via Sparse-To-Dense Representation Learning
Kim, K. K., & Bao, Z. (2026) · IEEE International Conference on Acoustics Speech and Signal Processing · doi:10.1109/icassp55912.2026.11463598
Lou, J., Wang, S., Lü, S., Yang, J., Hou, B., & Zhang, R. (2026) · IEEE Sensors Journal · doi:10.1109/jsen.2026.3714089
“Recent advances have further extended this idea to video and time-series modeling [42], [43], [44], [45].”
Multimodal Deep Learning-Based Anomaly Action Recognition in Multiple Physical Security Domains
Kim, D., & Yang, C. M. (2026) · doi:10.1109/icufn69619.2026.11628503
“Masked video modeling has demonstrated strong generalization capability for video understanding tasks, and this motivates exploring efficient ConvNet-based architectures for robust spatio-temporal representation learning [1].”
Calatrava-Nicolás, F. M., Stoyanov, T., & Martinez Mozos, O. (2026) · Scientific Reports · doi:10.1038/s41598-026-64846-5
“Related developments in computer vision, such as ConvNeXt V2 and VideoMAC, also show that masked autoencoding can be combined with convolutional encoders, while requiring attention to the interaction between masking and convolutional architectures 43,44 .”
M3R++: Masked Token Mixup and Dual Cross-Modal Masking for Zero-Shot Learning
Wang, Q., Xiao, Y., Lan, R., Yin, Y., & Lu, H. (2026) · IEEE Transactions on Pattern Analysis and Machine Intelligence · doi:10.1109/tpami.2026.3724025
“Moreover, MAEs have been successfully extended to other domains, such as videos [56], [57], [58], audio [59], [60], and point clouds [61], [62].”
Beyond Attention: Convolutional Global Context for Remote Sensing Change Detection
Yang, Z., Pei, G., Mao, J., et al. (2026) · Lecture notes in computer science · doi:10.1007/978-3-032-37261-1_27
Self-citation
- Authors
- Pei, G., Chen, T., Jiang, X., Liu, H., Sun, Z., & Yao, Y.
- Affiliations
- Nanjing University of Science and Technology (CN)
- Venue
- 2024 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)
- Publisher
- IEEE
- Published
- 2024-06-16
- Pages
- 22733-22743
- ISSN
- 1063-6919 (ISSN-L)
- OpenAlex
- W4402780389
- Cited by
- 23
- References
- 76
- Citations by year
- 2026: 9 · 2025: 12 · 2024: 2
- FWCI
- 4.72
- Type
- conference paper
- Language
- en
- Open access
- green
- arxiv.org
- Locations
- IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings - IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings (publishedVersion)
- arXiv (Cornell University) (submittedVersion)
- arXiv (Cornell University)
- Funders
- National Natural Science Foundation of China (62102182,62202227,62302217)
- Natural Science Foundation of Jiangsu Province (BK20220934,BK20220938,BK20220936)
- Fundamental Research Funds for the Central Universities (30923010303)
- Topics
- Physical Sciences › Computer Science › Computer Vision and Pattern Recognition › Advanced Neural Network Applications
- Physical Sciences › Computer Science › Artificial Intelligence › Domain Adaptation and Few-Shot Learning
- Physical Sciences › Computer Science › Computer Vision and Pattern Recognition › Human Pose and Action Recognition
- Keywords
- convolutional neural networks, self-supervised learning, video representation learning, video object segmentation, human pose tracking, ResNet, ConvNeXt
About these checks
Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.
Statistics and arithmetic are recomputed, not judged.
Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.
What counts as a claim ›How we measure confidence ›How we check this ›
Are you an author, or do you see a problem with this page? Request a review
