The DOI matches its Crossref record; 56 of the 57 references it cites were found in a registry or an index.
Hide-and-Seek: Forcing a Network to be Meticulous for Weakly-Supervised Object and Action Localization
Open accessAbstract
We propose `Hide-and-Seek', a weakly-supervised framework that aims to improve object localization in images and action localization in videos. Most existing weakly-supervised methods localize only the most discriminative parts of an object rather than all relevant parts, which leads to suboptimal performance. Our key idea is to hide patches in a training image randomly, forcing the network to seek other relevant parts when the most discriminative part is hidden. Our approach only needs to modify the input image and can work with any network designed for object localization. During testing, we do not need to hide any patches. Our Hide-and-Seek approach obtains superior performance compared to previous methods for weakly-supervised object localization on the ILSVRC dataset. We also demonstrate that our framework can be easily extended to weakly-supervised action localization.
Overview
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
Verification
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?
In the citing paper
“However, various types of industrial image defects do not meet all the above assumptions [42,43].”
From An Improved VGG19 Transfer Learning Strip Steel Surface Defect Recognition Deep Neural Network Based on Few Samples and Imbalanced Datasets (2021) · cites this paper as [43]
Self-check
Percentages, sums and reported differences re-computed from the paper’s own numbers.
The self-check needs the paper’s full text.
Citations
Cites
ImageNet 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
Going deeper with convolutions
Szegedy, C., Liu, W., Jia, Y., Sermanet, P., Reed, S., Anguelov, D., Erhan, D., Vanhoucke, V., & Rabinovich, A. (2015) · 2015 IEEE Conference on Computer Vision and Pattern Recognition (CVPR) · doi:10.1109/cvpr.2015.7298594 · Cited by 48,562
ImageNet Large Scale Visual Recognition Challenge
Russakovsky, O., Deng, J., Su, H., Krause, J., Satheesh, S., Ma, S., Huang, Z., Karpathy, A., Khosla, A., Bernstein, M., et al. (2015) · International Journal of Computer Vision · doi:10.1007/s11263-015-0816-y · Cited by 39,380
Fully convolutional networks for semantic segmentation
Long, J., Shelhamer, E., & Darrell, T. (2015) · 2015 IEEE Conference on Computer Vision and Pattern Recognition (CVPR) · doi:10.1109/cvpr.2015.7298965 · Cited by 38,834
(2014) · Cited by 33,419
Rich Feature Hierarchies for Accurate Object Detection and Semantic Segmentation
Girshick, R., Donahue, J., Darrell, T., & Malik, J. (2014) · 2014 IEEE Conference on Computer Vision and Pattern Recognition · doi:10.1109/cvpr.2014.81 · Cited by 32,684
(2016) · Cited by 31,590
Girshick, R. (2015) · 2015 IEEE International Conference on Computer Vision (ICCV) · doi:10.1109/iccv.2015.169 · Cited by 28,105
Batch Normalization: Accelerating Deep Network Training by Reducing Internal Covariate Shift
Ioffe, S., & Szegedy, C. (2015) · arXiv · doi:10.48550/arxiv.1502.03167 · Cited by 27,818
Visualizing and understanding convolutional networks
Zeiler, M. D., & Fergus, R. (2014) · Lecture Notes in Computer Science · doi:10.1007/978-3-319-10590-1_53 · Cited by 15,839
Flagged first, then by citations
1 of the 57 references couldn’t be matched to a known paper.
Research works printed but not linked: 5
- jiang, 2014, THUMOS Challenge Action Recognition with A Large Number of Classes
- jhuang, 2013, Towards undestanding action recognition, ICCV
- 10.1109/CVPR.2016.311 · DOI · Paper page
- cinbis, 2015, Weakly supervised object localization with multi-fold multiple instance learning
- wan, 2013, Regularization of neural network using dropconnect, ICML
Cited by
Sorted by citations
Items 301–400 of 720
Su, H., Zhao, X., Lin, T., Liu, S., & Hu, Z. (2020) · IEEE Transactions on Multimedia · doi:10.1109/tmm.2020.2999184 · Cited by 18
“[10] present a data augmentation method by randomly hiding regions during the training phase to force the classification network to look for other discriminative areas.”
BackgroundMethodInfluentialMETAL: Minimum Effort Temporal Activity Localization in Untrimmed Videos
Zhang, D., Dai, X., & Wang, Y.-F. (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/cvpr42600.2020.00394 · Cited by 18
“Hide-and-Seek [35] addressed the challenge that weakly supervised detection models usually neglect some relevant parts of the target instance.”
BackgroundKeep CALM and Improve Visual Feature Attribution
Kim, J. M., Choe, J., Akata, Z., & Oh, S. J. (2021) · IEEE/CVF International Conference on Computer Vision (ICCV) · doi:10.1109/iccv48922.2021.00824 · Cited by 18
“We show the architecture-wise performances for CALM, CAM, and six previous WSOL methods (HaS [32], ACoL [66], SPG [67], ADL [14], CutMix [64], InCA [27]).”
MethodZhang, J., Ma, K., Arnam, J. V., Gupta, R., Saltz, J. H., Vakalopoulou, M., & Samaras, D. (2021) · IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops (CVPRW) · doi:10.1109/cvprw53098.2021.00418 · Cited by 18
“However, such an attention estimation strategy encounters similar problems as other saliency detection models [33, 10, 25].”
ResultGhorbel, M., Ammar, S., Kessentini, Y., & Jmaïel, M. (2022) · Expert Systems with Applications · doi:10.1016/j.eswa.2022.116636 · Cited by 18
Occlusions for Effective Data Augmentation in Image Classification
Fong, R., & Vedaldi, A. (2019) · IEEE International Conference on Computer Vision Workshops · doi:10.1109/iccvw.2019.00511 · Cited by 18
“We first consider two existing ways of generating occlusions stochastically: Hide-andSeek [16] and Cutout [1].”
BackgroundMethodResultInfluentialBackMix: Regularizing Open Set Recognition by Removing Underlying Fore-Background Priors
Wang, Y., Mu, J., Huang, H., Wang, Q., Zhu, P., & Hu, Q. (2025) · IEEE Transactions on Pattern Analysis and Machine Intelligence · doi:10.1109/tpami.2025.3550703 · Cited by 20
“Masking-based methods: DeVries and Taylor [34] introduced Cutout to remove random image regions for occlusion-invariant training, while Singh and Lee [35] used random patch hiding to encourage learning from whole objects.”
Chen, Z., Liu, H., Zhang, L., & Liao, X. (2022) · IEEE Transactions on Multimedia · doi:10.1109/tmm.2022.3174344 · Cited by 17
Murtaza, S., Belharbi, S., Pedersoli, M., Sarraf, A., & Granger, É. (2023) · IEEE/CVF Winter Conference on Applications of Computer Vision Workshops (WACVW) · doi:10.1109/wacvw58289.2023.00021 · Cited by 17
“We compare our results with CAM [62], HaS [37], ACoL [59], SPG [60], ADL [14], and CutMix [56].”
BackgroundMethodResultInfluentialCLE-ViT: Contrastive Learning Encoded Transformer for Ultra-Fine-Grained Visual Categorization
Yu, X., Wang, J., & Gao, Y. (2023) · International Joint Conference on Artificial Intelligence · doi:10.24963/ijcai.2023/504 · Cited by 19
Bai, J., Ren, J., Xiao, Z. Q., Chen, Z., Gao, C., Ali, T. A. A., & Jiao, L. (2023) · IEEE Transactions on Neural Networks and Learning Systems · doi:10.1109/tnnls.2023.3309889 · Cited by 19
“CutMix [16] and Has [15] divide the training data images into patches and randomly select patches to be dropped or blended with other category patches to augment the data and force the model to focus on the extended region.”
InfluentialCategory-aware Allocation Transformer for Weakly Supervised Object Localization
Chen, Z., Ding, J., Cao, L., Shen, Y., Zhang, S., Jiang, G., & Ji, R. (2023) · IEEE/CVF International Conference on Computer Vision (ICCV) · doi:10.1109/iccv51070.2023.00611 · Cited by 17
“To capture more object parts, different techniques have been proposed to enhance CAM, such as graph propagation [40], data augmentation [33, 15], adversarial erasing [3, 6, 19] and spatial relation activation [32, 37, 9].”
BackgroundMethodXiang, T., Zhang, Y., Lu, Y., Yuille, A., Zhang, C., Cai, W., & Zhou, Z. (2024) · IEEE Transactions on Pattern Analysis and Machine Intelligence · doi:10.1109/tpami.2024.3382009 · Cited by 17
“It is understandable because, unlike [49], [114], [115], [116], [117], our SimSID is an unsupervised method, requiring zero manual annotation for normal/abnormal images.”
MethodDeep Grouping Model for Unified Perceptual Parsing
Li, Z., Bao, W., Zheng, J., & Xu, C. (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/cvpr42600.2020.00411 · Cited by 18
Chen, J., Li, H., Liang, J., Su, X., Zhai, Z., & Chai, X. (2022) · Neurocomputing · doi:10.1016/j.neucom.2022.06.041 · Cited by 17
D4: Text-guided diffusion model-based domain adaptive data augmentation for vineyard shoot detection
Hirahara, K., Nakane, C., Ebisawa, H., Kuroda, T., Iwaki, Y., Utsumi, T., Nomura, Y., Koike, M., & Mineno, H. (2024) · Computers and Electronics in Agriculture · doi:10.1016/j.compag.2024.109849 · Cited by 17
Seo, M., Cho, D., Lee, S.-W., Park, J., Kim, D., Lee, J., Ju, J., Noh, H., & Choi, D. (2021) · IEEE Robotics and Automation Letters · doi:10.1109/lra.2021.3139369 · Cited by 16
“These results are similar tendency with the previous study [29] that the deep learning model trained by solving difficult tasks improves performance.”
ResultTraffic Context Aware Data Augmentation for Rare Object Detection in Autonomous Driving
Li, N., Song, F., Zhang, Y., Liang, P., & Cheng, E. (2022) · Proceedings - IEEE International Conference on Robotics and Automation/Proceedings · doi:10.1109/icra46639.2022.9811724 · Cited by 17
Zhang, P., Guo, S., Zhang, W., Lin, C., Xia, Z., Zhang, X., Fang, H., & Du, P. (2023) · IEEE Transactions on Geoscience and Remote Sensing · doi:10.1109/tgrs.2023.3257293 · Cited by 16
“To alleviate this problem, one group of approaches used sub-category exploration [46], attention mechanism [49], or hiding and erasing operation [50] to force models to seek different parts of objects, thus directly improving the performance of initial location maps.”
MethodBoosting Positive Segments for Weakly-Supervised Audio-Visual Video Parsing
Rachavarapu, K. K., & Rajagopalan, A. N. (2023) · IEEE/CVF International Conference on Computer Vision (ICCV) · doi:10.1109/iccv51070.2023.00935 · Cited by 16
“Others aim to identify key frames by exploiting the complementary nature of RGB and flow modalities [47, 45], or using the most discriminative parts [37, 51, 27] of a video.”
BackgroundProgressive Representation Adaptation for Weakly Supervised Object Localization
Li, D., Huang, J., Li, Y., Wang, S., & Yang, M. (2019) · IEEE Transactions on Pattern Analysis and Machine Intelligence · doi:10.1109/tpami.2019.2899839 · Cited by 17
Action Coherence Network for Weakly Supervised Temporal Action Localization
Zhai, Y., Wang, L., Liu, Z., Zhang, Q., Hua, G., & Zheng, N. (2019) · Proceedings - International Conference on Image Processing · doi:10.1109/icip.2019.8803447 · Cited by 16
“Hide-and-Seek [14] randomly hides parts of the input to guide the network to learn the most relevant parts.”
BackgroundWeakly-supervised action localization via embedding-modeling iterative optimization
Zhang, X., Shi, H.-C., Li, C., Li, P., Li, Z., & Ren, P.-S. (2021) · Pattern Recognition · doi:10.1016/j.patcog.2021.107831 · Cited by 15
KeepOriginalAugment: Single Image-Based Better Information-Preserving Data Augmentation Approach
Kumar, T., Mileo, A., & Bendechache, M. (2024) · IFIP advances in information and communication technology · doi:10.1007/978-3-031-63223-5_3 · Cited by 15
Injecting Text Clues for Improving Anomalous Event Detection From Weakly Labeled Videos
Liu, T., Lam, K., & Bao, B. (2024) · IEEE Transactions on Image Processing · doi:10.1109/tip.2024.3477351 · Cited by 15
“(2) The erasing-based [43], [44], [45], multi-attention-based [46], [47], [48] and multi-branch-based [49], [50], [51] methods force the network to excavate activity clips from less-discriminative regions, thus improving the completeness of localization results.”
Guan, W., Wang, S., Chen, Z., Wang, G., Liu, Z., Cui, D., & Mao, Y. (2025) · Reliability Engineering & System Safety · doi:10.1016/j.ress.2025.111055 · Cited by 15
A pixel-level coarse-to-fine image segmentation labelling algorithm
Lee, J.-H., Ilyas, T., Jin, H., Lee, J., Won, O., Kim, H., & Lee, S. J. (2022) · Scientific Reports · doi:10.1038/s41598-022-12532-7 · Cited by 15
“To improve localization performance, Singh et al. 29 forced their network to locate more than one discriminative region of an object by blocking out the portion of images at random.”
BackgroundMethodZhang, X., Su, Q., Xiao, P., Wang, W., Li, Z., & He, G. (2024) · IEEE Transactions on Geoscience and Remote Sensing · doi:10.1109/tgrs.2024.3360276 · Cited by 16
“However, these methods only enhance the original images through random masks [89] or heatmap-based masks [90], lacking further integration of high-dimensional feature information extracted by the network.”
MethodWeakly supervised target detection based on spatial attention
Zhao, W., & Xu, L. (2024) · Visual Intelligence · doi:10.1007/s44267-024-00037-y · Cited by 15
“The erasure mechanism has been widely utilized in computer vision tasks such as weakly supervised target localization [18, 19] and weakly supervised semantic segmentation [20].”
Traffic Anomaly Detection based on Spatio-Temporal Hypergraph Convolution Neural Networks
Feng, J., Zhang, Y., Piao, X., Hu, Y., & Yin, B. (2024) · Physica A Statistical Mechanics and its Applications · doi:10.1016/j.physa.2024.129891 · Cited by 16
Wei, J., Wang, S., Zhou, S., Cui, S. R., & Li, Z. (2022) · Lecture notes in computer science · doi:10.1007/978-3-031-20056-4_12 · Cited by 15
“Hence, lots of approaches have been proposed to improve CAMs, such as erasing based methods [32, 19, 3, 26, 14], feature refining based methods [24, 27, 16, 29, 4] and regression based methods [31, 12, 6].”
MethodMulti-Scale Low-Discriminative Feature Reactivation for Weakly Supervised Object Localization
Wang, B., Yuan, C., Li, B., Ding, X., Li, Z., Wu, Y. N., & Hu, W. (2021) · IEEE Transactions on Image Processing · doi:10.1109/tip.2021.3091833 · Cited by 15
UDD: An Underwater Open-sea Farm Object Detection Dataset for Underwater Robot Picking.Preprint
Wang, Z., Liu, C., Wang, S., Tang, T., Tao, Y., Yang, C., Li, H., Liu, X., & Fan, X. (2020) · arXiv (Cornell University) · Cited by 15
Accurate Landmark Localization for Medical Images Using Perturbations
Kang, J., Oh, K., & Oh, I.-S. (2021) · Applied Sciences · doi:10.3390/app112110277 · Cited by 14
On Label Granularity and Object Localization
Cole, E., Wilber, M. J., Horn, G. V., Yang, X., Fornoni, M., Perona, P., Belongie, S., Howard, A., & Aodha, O. M. (2022) · Lecture notes in computer science · doi:10.1007/978-3-031-20080-9_35 · Cited by 15
TCAM: Temporal Class Activation Maps for Object Localization in Weakly-Labeled Unconstrained Videos
Belharbi, S., Ayed, I. B., McCaffrey, L. M., & Granger, É. (2023) · IEEE Winter Conference on Applications of Computer Vision · doi:10.1109/wacv56688.2023.00022 · Cited by 14
“It is motivated by the success of weaklysupervised object localization (WSOL) methods for still images (Belharbi et al., 2022c; Choe and Shim, 2019; Durand et al., 2017; Gao et al., 2021; Jiang et al., 2021; Singh and Lee, 2017; Wei et al., 2021; Zhou et al., 2016).”
ICEv2: Interpretability, Comprehensiveness, and Explainability in Vision Transformer
Choi, H.-Y., Jin, S., & Han, K. (2024) · International Journal of Computer Vision · doi:10.1007/s11263-024-02290-6 · Cited by 14
“Second, prior studies optimized the threshold by repeatedly using localization labels in test data (Kumar Singh & Jae Lee, 2017; Zhang et al., 2018; Choe & Shim, 2019).”
DAP: Detection-Aware Pre-training with Weak Supervision
Zhong, Y., Wang, J., Wang, L., Peng, J., Wang, Y.-X., & Zhang, L. (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/cvpr46437.2021.00451 · Cited by 14
“WSOL refers to a class of object localization methods that rely on weak supervision (image-level labels) [27, 47, 35, 44, 45, 7, 6], which is exactly what we need for the pre-training data.”
BackgroundMulti-Hierarchical Category Supervision for Weakly-Supervised Temporal Action Localization
Li, G., Li, J., Wang, N., Ding, X., Li, Z., & Gao, X. (2021) · IEEE Transactions on Image Processing · doi:10.1109/tip.2021.3124671 · Cited by 14
Detecting and Mapping Video Impairments
Goodall, T., & Bovik, A. C. (2018) · IEEE Transactions on Image Processing · doi:10.1109/tip.2018.2889276 · Cited by 14
“This is a phenomenon similar to that observed by Singh and Lee [44], who proposed randomly hiding the most discriminative data during training.”
ResultUncertainty-Aware Graph-Guided Weakly Supervised Object Detection
Zhu, Y., Zhang, Y., Ding, M., & Zuo, W. (2022) · IEEE Transactions on Circuits and Systems for Video Technology · doi:10.1109/tcsvt.2022.3232487 · Cited by 13
“…supervised learning explores how to construct predictive models by learning with weak supervision, including weakly supervised object localization [34], [35], weakly supervised salience detection [36], [37], weakly supervised object detection [10], [11], [12], [13], [14], [15], [16], [17], etc. .”
BackgroundA Weakly Supervised Approach for Disease Segmentation of Maize Northern Leaf Blight from UAV Images
Chen, S., Zhang, K., Wu, S., Tang, Z., Zhao, Y., Sun, Y., & Shi, Z. (2023) · Drones · doi:10.3390/drones7030173 · Cited by 13
Hierarchical Semantic Contrast for Weakly Supervised Semantic Segmentation
Wu, Y., Li, X., Dai, S., Li, J., Liu, T., & Xie, S. (2023) · International Joint Conference on Artificial Intelligence · doi:10.24963/ijcai.2023/171 · Cited by 13
ContextMix: A context-aware data augmentation method for industrial visual inspection systems
Kim, H., Kim, D., Ahn, P., Suh, S., Cho, H., & Kim, J. (2024) · Engineering Applications of Artificial Intelligence · doi:10.1016/j.engappai.2023.107842 · Cited by 15
“Hide and Seek (HaS) [34] is a method motivated by Cutout that removes multiple random regions.”
MethodKarkehabadi, A., Latibari, B. S., Homayoun, H., & Sasan, A. (2024) · International Conference on Information Science and Technology · doi:10.1109/icist63249.2024.10805345 · Cited by 13
“Singh et al. [7] introduced methods that use higher-order backward gradients, enhancing saliency map insights.”
Weakly-supervised temporal action localization: a survey
Baraka, A. M., & Noor, M. H. M. (2022) · Neural Computing and Applications · doi:10.1007/s00521-022-07102-x · Cited by 13
“Where most models generate a frame class-score of each action category, and the frames that have low scores are the candidates of background frames such as UntrimmedNet [21], Hide-and-Seek [30], and W-TALC [24] models.”
BackgroundXu, J., Xie, H., Xu, H. Y., Wang, Y., Liu, S.-A., & Zhang, Y. (2022) · ACM International Conference on Multimedia (ACM MM) · doi:10.1145/3503161.3548201 · Cited by 13
“Based on the common observation that CAMs only focus on the most discriminative regions, some extensive methods [8, 23, 43] aim to expand object coverage to capture the full extent of objects by erasing or hiding some high response areas.”
BackgroundLocation-Free Human Pose Estimation
Xu, X., Gao, Y., Yan, K., Lin, X., & Zou, Q. (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/cvpr52688.2022.01279 · Cited by 14
“After that, a series of works [20, 46, 48] are proposed to improve the quality of CAM.”
BackgroundDeep joint learning of pathological region localization and Alzheimer’s disease diagnosis
Park, C., Jung, W., & Suk, H. (2023) · Scientific Reports · doi:10.1038/s41598-023-38240-4 · Cited by 13
AAN-Face: Attention Augmented Networks for Face Recognition
Wang, Q., & Guo, G. (2021) · IEEE Transactions on Image Processing · doi:10.1109/tip.2021.3107238 · Cited by 15
E2Net: Excitative-Expansile Learning for Weakly Supervised Object Localization
Chen, Z., Cao, L., Shen, Y., Lian, F., Wu, Y., & Ji, R. (2021) · ACM International Conference on Multimedia (ACM MM) · doi:10.1145/3474085.3475211 · Cited by 13
“Previous works randomly erase image patches [18, 30] or intermediate feature maps [2] to capture less discriminative regions during training.”
BackgroundMethodInfluentialCrossing the Format Boundary of Text and Boxes: Towards Unified Vision-Language ModelingPreprint
Yang, Z., Gan, Z., Wang, J., Hu, X., Ahmed, F., Liu, Z., Lu, Y., & Wang, L. (2021) · arXiv (Cornell University) · Cited by 13
Moniruzzaman, M., & Yin, Z. (2023) · IEEE Transactions on Multimedia · doi:10.1109/tmm.2023.3263965 · Cited by 12
“Motivated by the object completeness modeling in object detection task [38], [39], [40], Hide-and-Seek [41] first hid random frame sequences to force the network to discover different action parts for the action completeness modeling.”
MethodFu, J., Gao, J., & Xu, C. (2023) · IEEE Transactions on Pattern Analysis and Machine Intelligence · doi:10.1109/tpami.2023.3287208 · Cited by 12
“[51] P. Lee, J. Wang, Y. Lu, and H. Byun, “Weakly-supervised temporal action localization by uncertainty modeling,” in Proc.”
InfluentialWeakly supervised semantic segmentation via self-supervised destruction learning
Li, J., Jie, Z., Wang, X., Zhou, Y., Ma, L., & Jiang, J. (2023) · Neurocomputing · doi:10.1016/j.neucom.2023.126821 · Cited by 14
Exploring Intrinsic Discrimination and Consistency for Weakly Supervised Object Localization
Wang, C., Xu, R., Xu, S., Meng, W., Wang, R., & Zhang, X. (2024) · IEEE Transactions on Image Processing · doi:10.1109/tip.2024.3356174 · Cited by 12
“A series of follow-up works ( e.g. [15], [19], [20], [21], [24], [25], [26], [27], [28], [31], [32], [34], [44], and [45]) emerge to improve incomplete activation of CAM [16].”
BackgroundMethodInfluentialAn Attention-based Semi-supervised Neural Network for Thyroid Nodules Segmentation
Wang, J., Zhang, R., Wei, X., Li, X., Yu, M., Zhu, J., Gao, J., Liu, Z., & Yu, R. (2019) · IEEE International Conference on Bioinformatics and Biomedicine (BIBM) · doi:10.1109/bibm47256.2019.8983288 · Cited by 12
“Some other similar weakly supervised segmentation methods also have been proposed such as [18].”
Zhong, C., Reibman, A. R., Mina, H. A., & Deering, A. J. (2021) · Journal of Imaging · doi:10.3390/jimaging7090170 · Cited by 12
“Hiding patches of an image, either during training [69], testing [70], or both [71], has been shown to be a valuable tool.”
Huynh, D., & Elhamifar, E. (2021) · IEEE/CVF International Conference on Computer Vision (ICCV) · doi:10.1109/iccv48922.2021.00836 · Cited by 12
“Recent works [79, 85, 86] focus on capturing the full spatial extents of objects by regularizing predictions to prevent localizing only the most discriminative parts.”
BackgroundA new weakly supervised strategy for surgical tool detection
Xue, Y., Liu, S., Li, Y., Wang, P., & Qian, X. (2022) · Knowledge-Based Systems · doi:10.1016/j.knosys.2021.107860 · Cited by 12
Entropy regularization for weakly supervised object localization
Hwang, D., Ha, J.-W., Shim, H., & Choe, J. (2023) · Pattern Recognition Letters · doi:10.1016/j.patrec.2023.03.018 · Cited by 12
Bilateral Relation Distillation for Weakly Supervised Temporal Action Localization
Xu, Z., Wei, K., Yang, E., Deng, C. D., & Liu, W. (2023) · IEEE Transactions on Pattern Analysis and Machine Intelligence · doi:10.1109/tpami.2023.3284853 · Cited by 12
“, actionness learning [13], [14], erasing-based strategies [15], [16], uncertainty modeling [9], [17], contrastive learning [18], [19], and pseudo-label generation [20], [21].”
BackgroundSemantic Segmentation of Surgical Hyperspectral Images Under Geometric Domain Shifts
Sellner, J., Seidlitz, S., Studier‐Fischer, A., Motta, A., Özdemir, B., Müller-Stich, B. P., Nickel, F., & Maier-Hein, L. (2023) · Lecture notes in computer science · doi:10.1007/978-3-031-43996-4_59 · Cited by 12
Improve CAM with Auto-adapted Segmentation and Co-supervised Augmentation
Kou, Z., Cui, G., Wang, S., Zhao, W., & Xu, C. (2021) · IEEE Winter Conference on Applications of Computer Vision · doi:10.1109/wacv48630.2021.00364 · Cited by 11
“(b)CUB-200-2011 err AlexNet-GAP [26] 45.01 AlexNet-HaS [10] 41.26 GoogLeNet-GAP [26] 41.34 GoogLeNet-HaS [10] 39.43 VGGnet-ACoL [24] 37.04 SPG [25] 35.31 ADL [4] 34.59 GoogLeNet-CSoA 33.80 different blocks to reduce the receptive differences.”
BackgroundInfluentialWatanabe, K., Ito, K., & Aoki, T. (2022) · Asia-Pacific Signal and Information Processing Association Annual Summit and Conference (APSIPA ASC) · doi:10.23919/apsipaasc55919.2022.9979996 · Cited by 11
“It is known that masking a part of the input image with pixel value 0 during training improves the generalization ability of the model and improves its accuracy in image classification and segmentation [20], [21].”
BackgroundMethodInfluentialChen, Q., Jiao, L., Wang, F., Du, J., Liu, H., Wang, R., & Wang, R. (2024) · Pattern Recognition · doi:10.1016/j.patcog.2024.110339 · Cited by 11
EntAugment: Entropy-Driven Adaptive Data Augmentation Framework for Image Classification
Yang, S., Shen, F., & Zhao, J. (2024) · Lecture notes in computer science · doi:10.1007/978-3-031-72848-8_12 · Cited by 11
Causal Explanation of Convolutional Neural Networks
Debbi, H. (2021) · Lecture notes in computer science · doi:10.1007/978-3-030-86520-7_39 · Cited by 11
“Kumar and Lee [31] for instance tried to tackle this issue by modifying the input image by randomly removing grid patches, thus forcing the network to be meticulous for WSOL, by identifying all the object’s features not only the most discriminative ones.”
BackgroundWeakly Supervised Temporal Action Localization with Segment-Level Labels
Ding, X., Wang, N., Li, J., & Gao, X. (2021) · Lecture notes in computer science · doi:10.1007/978-3-030-88004-0_4 · Cited by 11
Weakly Supervised Temporal Action Detection With Temporal Dependency Learning
Li, B., Liu, R., Chen, T., & Zhu, Y. (2021) · IEEE Transactions on Circuits and Systems for Video Technology · doi:10.1109/tcsvt.2021.3125701 · Cited by 13
Weakly supervised foreground learning for weakly supervised localization and detection
Zhang, C.-L., Li, Y., & Wu, J. (2022) · Pattern Recognition · doi:10.1016/j.patcog.2022.109279 · Cited by 11
Baek, K., & Shim, 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/cvpr52688.2022.00770 · Cited by 11
“Focusing on the visual effects, inserting the monotone object can be similar to the regional dropout [2,43] in the weaklysupervised object localization (WSOL) task.”
BackgroundComplementary adversarial mechanisms for weakly-supervised temporal action localization
Wang, C., Wang, J., & Liu, P. (2023) · Pattern Recognition · doi:10.1016/j.patcog.2023.109426 · Cited by 11
MTHM: Self-Supervised Multitask Anomaly Detection With Hard Example Mining
Zhang, C., Wang, Y., & Tan, W. (2023) · IEEE Transactions on Instrumentation and Measurement · doi:10.1109/tim.2023.3276529 · Cited by 11
Object-centric Contour-aware Data Augmentation Using Superpixels of Varying Granularity
Dornaika, F., Sun, D., Hammoudi, K., Charafeddine, J., Cabani, A., & Zhang, C. (2023) · Pattern Recognition · doi:10.1016/j.patcog.2023.109481 · Cited by 11
SF-Net: Single-Frame Supervision for Temporal Action LocalizationPreprint
Ma, F., Zhu, L., Yang, Y., Zha, S., Kundu, G., Feiszli, M., & Shou, Z. (2020) · arXiv (Cornell University) · doi:10.48550/arxiv.2003.06845 · Cited by 11
“Weak Hide-and-Seek [32] 36.”
BackgroundFrame-Level Label Refinement for Skeleton-Based Weakly-Supervised Action Recognition
Yu, Q., & Fujiwara, K. (2023) · Proceedings of the AAAI Conference on Artificial Intelligence · doi:10.1609/aaai.v37i3.25439 · Cited by 10
“Hide-and-seek (Singh and Lee 2017) uses class activation maps to identify the relevant frames of each action.”
BackgroundMethodABC-Learning: Attention-Boosted Contrastive Learning for unsupervised person re-identification
Zhang, S., Wang, C., & Peng, J. (2024) · Engineering Applications of Artificial Intelligence · doi:10.1016/j.engappai.2024.108344 · Cited by 10
Adaptive Zone Learning for Weakly Supervised Object Localization
Chen, Z., Wang, S., Cao, L., Shen, Y., & Ji, R. (2024) · IEEE Transactions on Neural Networks and Learning Systems · doi:10.1109/tnnls.2024.3392948 · Cited by 10
“To mitigate this challenge, numerous preceding WSOL techniques [39], [40], [41], [42], such as graph propagation [43], data augmentation [44], [45], adversarial erasing [2], [3], [46], spatial relation activation [34], [41], [47], and self-attention [48], [49], concentrate on broadening activation…”
Kabiraj, A., Meena, T., Tadepalli, K., & Roy, S. S. (2024) · Applied Soft Computing · doi:10.1016/j.asoc.2024.112139 · Cited by 13
CAM Back Again: Large Kernel CNNs from a Weakly Supervised Object Localization Perspective
Yasuki, S., & Taki, M. (2024) · 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/cvpr52733.2024.00040 · Cited by 12
“Improvements to solve some problems in CAM include data augmentation strategies [10, 32, 38, 42, 45], devising training methods [8, 15, 40, 46, 47], specialized post-processing methods us-ing only trained classifiers [5, 18, 31], and higher resolution heatmaps [4, 33, 41], Classification and…”
MethodCLIP-Driven Transformer for Weakly Supervised Object Localization
Chen, Z., Shen, Y., Cao, L., Zhang, S., & Ji, R. (2025) · IEEE Transactions on Pattern Analysis and Machine Intelligence · doi:10.1109/tpami.2025.3548704 · Cited by 11
“Singh et al. [29] experimented with hiding random image patches during training, while Yun et al. [28] introduced the innovative cut-and-paste technique among training images.”
Coarse2Fine: Local Consistency Aware Re-prediction for Weakly Supervised Object Localization
Pan, Y., Yao, Y., Cao, Y., Chen, C., & Lu, X. (2023) · Proceedings of the AAAI Conference on Artificial Intelligence · doi:10.1609/aaai.v37i2.25292 · Cited by 10
Yang, H., Jeong, Y., Choi, M., & Lee, J. (2018) · doi:10.1145/3267471.3267482 · Cited by 10
“Inspired by a hide-and-seek idea [13], we ignore either playlists or contents at a time during training.”
MethodKumar, N., & Sukavanam, N. (2019) · The Visual Computer · doi:10.1007/s00371-019-01777-5 · Cited by 10
“Training of deep network from the deep features can be supervised [20], weakly supervised [3,21] and unsupervised [22,23] but this is noticed that weakly supervised or fully unsupervised action localization is highly challenging.”
BackgroundDual-attention Guided Dropblock Module for Weakly Supervised Object Localization
Yin, J., Zhang, S., Chang, D., Ma, Z., & Guo, J. H. (2021) · Proceedings - International Conference on Pattern Recognition/Proceedings/International Conference on Pattern Recognition · doi:10.1109/icpr48806.2021.9412181 · Cited by 10
“2) Evaluation metrics: Three evaluation metrics are used for WSOL evaluation [4].”
BackgroundMethodInfluentialA deep learning method for video‐based action recognition
Zhang, G., Rao, Y., Wang, C., Zhou, W., & Ji, X. (2021) · IET Image Processing · doi:10.1049/ipr2.12303 · Cited by 11
Chang, Y.-H., Lin, M., Hsieh, M., Ou, M.-C., Huang, C.-R., & Sheu, B. (2023) · IEEE Journal of Translational Engineering in Health and Medicine · doi:10.1109/jtehm.2023.3286423 · Cited by 10
“Many weakly-supervised methods [17], [18], [25], [26], [27] are proposed in the computer vision domain.”
Kumar, T., Mileo, A., & Bendechache, M. (2024) · Signal Image and Video Processing · doi:10.1007/s11760-024-03623-1 · Cited by 9
Saliency Aware: Weakly Supervised Object Localization
Chen, Y.-C., & Hsu, W. H. (2019) · IEEE International Conference on Acoustics Speech and Signal Processing · doi:10.1109/icassp.2019.8682756 · Cited by 9
“We compare our proposed approach with two existing weaklysupervised object localization methods: AlexNet-GAP [8] (adopts the CAM [8] method) and AlexNet-HaS [10] (adopts the Grad-CAM [8] method).”
MethodInfluentialGao, S., Guo, G., Huang, H., & Chen, C. L. P. (2023) · IEEE Transactions on Neural Networks and Learning Systems · doi:10.1109/tnnls.2022.3225180 · Cited by 9
“Hide and Seek (HaS) [17] aims to randomly erase some regions in the image to impose network to be meticulous for WSOL.”
InfluentialAPSL: Action-positive separation learning for unsupervised temporal action localization
Liu, Y., Zhou, N., Zhang, F., Wang, W., Wang, Y., Liu, K., & Liu, Z. (2023) · Information Sciences · doi:10.1016/j.ins.2023.02.047 · Cited by 10
Privacy-Preserving Age Estimation for Content Rating
Ye, L., Li, B., Mohammed, N., Wang, Y., & Liang, J. (2018) · IEEE International Workshop on Multimedia Signal Processing (MMSP) · doi:10.1109/mmsp.2018.8547144 · Cited by 10
A Generic Visualization Approach for Convolutional Neural Networks
Taha, A., Yang, X., Shrivastava, A., & Davis, L. S. (2020) · Lecture notes in computer science · doi:10.1007/978-3-030-58520-4_43 · Cited by 9
Dynamic proposal sampling for weakly supervised object detection
Jiang, W., Zhao, Z., Su, F., & Fang, Y. (2021) · Neurocomputing · doi:10.1016/j.neucom.2021.02.018 · Cited by 9
Weakly Supervised Object Localization on grocery shelves using simple FCN and Synthetic Dataset
Varadarajan, S., & Srivastava, M. M. (2018) · Indian Conference on Computer Vision, Graphics and Image Processing · doi:10.1145/3293353.3293367 · Cited by 9
“There have been previous works [6] [5] [9] [15] [13] which does the weakly supervised localization (WSL) task on the standard datasets like PASCAL VOC and COCO.”
BackgroundMethodForced Spatial Attention for Driver Foot Activity Classification
Rangesh, A., & Trivedi, M. M. (2019) · IEEE International Conference on Computer Vision Workshops · doi:10.1109/iccvw.2019.00308 · Cited by 9
“[25] improve the localization in CAMs by randomly hiding patches in the input image, thereby forcing the network to pay attention to other relevant parts that contribute to an accurate classification.”
BackgroundWang, H., Wang, G., Li, G., & Lin, L. (2019) · ACM International Conference on Information and Knowledge Management (CIKM) · doi:10.1145/3357384.3357999 · Cited by 9
“Instead of attaching dropout to the internal features, there are also some data augmentation methods which drop a fixed-size block from the raw images, like Cutout[5], [33] and RandomErasing[47].”
MethodFundus Image Classification and Retinal Disease Localization with Limited Supervision
Meng, Q., Hashimoto, Y., & Satoh, S. (2020) · Lecture notes in computer science · doi:10.1007/978-3-030-41404-7_33 · Cited by 9
Min-max Entropy for Weakly Supervised Pointwise LocalizationPreprint
Belharbi, S., Rony, J., Dolz, J., Ayed, I. B., McCaffrey, L. M., & Granger, É. (2019) · arXiv (Cornell University) · doi:10.48550/arxiv.1907.12934 · Cited by 9
“Among the cited literature, [34, 47, 18, 22] combined with [46] is probably the closest work to our proposal.”
BackgroundMethodInfluential
Metadata
- Authors
- Singh, K. K., & Lee, Y. J.
- Affiliations
- Intel (United States) (US)
- University of California, Davis (US)
- Venue
- 2017 IEEE International Conference on Computer Vision (ICCV)
- Publisher
- IEEE
- Published
- 2017-10
- Pages
- 3544-3553
- OpenAlex
- W2606520630
- Cited by
- 724
- References
- 57
- Citations by year
- 2026: 19 · 2025: 49 · 2024: 74 · 2023: 121 · 2022: 129 · 2021: 137 · 2020: 98 · 2019: 70 · 2018: 25 · 2017: 2
- FWCI
- 15.59
- Type
- conference paper
- Language
- en
- Open access
- green
- arxiv.org
- Locations
- IEEE/CVF International Conference on Computer Vision (ICCV) (publishedVersion)
- arXiv (Cornell University) (submittedVersion)
- arXiv (Cornell University)
- Funders
- Nvidia
- Amazon Web Services
- Topics
- Physical Sciences › Computer Science › Computer Vision and Pattern Recognition › Advanced Neural Network Applications
- Physical Sciences › Computer Science › Computer Vision and Pattern Recognition › Advanced Image and Video Retrieval Techniques
- Physical Sciences › Computer Science › Computer Vision and Pattern Recognition › Visual Attention and Saliency Detection
- Keywords
- weakly supervised object localization, data augmentation, image classification, video action recognition
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
