{"about":{"site":"https://codewithpapers.app","non_affiliation":"Code with Papers and Syntology are not affiliated with, endorsed by, or sponsored by Papers with Code, Meta, or the pwc-archive mirror.","licence":"CC BY-SA 4.0","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","attribution":"https://codewithpapers.app/attribution","modified":"archive material modified by Syntology; see the attribution page"},"url":"/paper/weakly-supervised-action-localization-by","title":"Weakly Supervised Action Localization by Sparse Temporal Pooling Network","arxiv_id":"1712.05080","date":"2017-12-14","proceeding":"CVPR 2018 6","authors":["Phuc Nguyen","Ting Liu","Gautam Prasad","Bohyung Han"],"abstract":"We propose a weakly supervised temporal action localization algorithm on\nuntrimmed videos using convolutional neural networks. Our algorithm learns from\nvideo-level class labels and predicts temporal intervals of human actions with\nno requirement of temporal localization annotations. We design our network to\nidentify a sparse subset of key segments associated with target actions in a\nvideo using an attention module and fuse the key segments through adaptive\ntemporal pooling. Our loss function is comprised of two terms that minimize the\nvideo-level action classification error and enforce the sparsity of the segment\nselection. At inference time, we extract and score temporal proposals using\ntemporal class activations and class-agnostic attentions to estimate the time\nintervals that correspond to target actions. The proposed algorithm attains\nstate-of-the-art results on the THUMOS14 dataset and outstanding performance on\nActivityNet1.3 even with its weak supervision.","url_abs":"http://arxiv.org/abs/1712.05080v2","url_pdf":"http://arxiv.org/pdf/1712.05080v2.pdf","source":{"archive":"pwc-archive (Hugging Face), CC BY-SA 4.0","snapshot":"2025-07-28","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","row_kind":"abstracts"},"code_links":[{"paper_slug":"weakly-supervised-action-localization-by","repo_url":"https://github.com/HoganZhang/weakly-action-localization","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"none","reach":null},{"paper_slug":"weakly-supervised-action-localization-by","repo_url":"https://github.com/Lin716/STPN-lin","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"none","reach":null},{"paper_slug":"weakly-supervised-action-localization-by","repo_url":"https://github.com/demianzhang/weakly-action-localization","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"none","reach":null}],"tasks":[{"task_slug":"action-classification","task_name":"Action Classification"},{"task_slug":"action-localization","task_name":"Action Localization"},{"task_slug":"action-recognition","task_name":"Temporal Action Localization"},{"task_slug":"temporal-localization","task_name":"Temporal Localization"},{"task_slug":"weakly-supervised-action-localization","task_name":"Weakly Supervised Action Localization"},{"task_slug":"weakly-supervised-temporal-action","task_name":"Weakly-supervised Temporal Action Localization"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[{"leaderboard":"/sota/weakly-supervised-action-localization-on-1","task":"Weakly Supervised Action Localization","dataset":"ActivityNet-1.3","model":"STPN","rank_in_archive_order":17,"of":17,"metrics":{"mAP@0.5":"29.3"},"uses_additional_data":false},{"leaderboard":"/sota/weakly-supervised-action-localization-on","task":"Weakly Supervised Action Localization","dataset":"THUMOS 2014","model":"STPN","rank_in_archive_order":22,"of":30,"metrics":{"mAP@0.1:0.7":"27.0","mAP@0.5":"16.9"},"uses_additional_data":false}],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=1712.05080","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}