{"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/ms-tcn-multi-stage-temporal-convolutional","title":"MS-TCN: Multi-Stage Temporal Convolutional Network for Action Segmentation","arxiv_id":"1903.01945","date":"2019-03-05","proceeding":"CVPR 2019 6","authors":["Yazan Abu Farha","Juergen Gall"],"abstract":"Temporally locating and classifying action segments in long untrimmed videos\nis of particular interest to many applications like surveillance and robotics.\nWhile traditional approaches follow a two-step pipeline, by generating\nframe-wise probabilities and then feeding them to high-level temporal models,\nrecent approaches use temporal convolutions to directly classify the video\nframes. In this paper, we introduce a multi-stage architecture for the temporal\naction segmentation task. Each stage features a set of dilated temporal\nconvolutions to generate an initial prediction that is refined by the next one.\nThis architecture is trained using a combination of a classification loss and a\nproposed smoothing loss that penalizes over-segmentation errors. Extensive\nevaluation shows the effectiveness of the proposed model in capturing\nlong-range dependencies and recognizing action segments. Our model achieves\nstate-of-the-art results on three challenging datasets: 50Salads, Georgia Tech\nEgocentric Activities (GTEA), and the Breakfast dataset.","url_abs":"http://arxiv.org/abs/1903.01945v2","url_pdf":"http://arxiv.org/pdf/1903.01945v2.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":"ms-tcn-multi-stage-temporal-convolutional","repo_url":"https://github.com/yabufarha/ms-tcn","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":1,"framework":"pytorch","reach":{"status":"ok","spdx":"NOASSERTION"}},{"paper_slug":"ms-tcn-multi-stage-temporal-convolutional","repo_url":"https://github.com/MindSpore-paper-code-2/code3/tree/main/TCN","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":0,"framework":"mindspore","reach":null}],"tasks":[{"task_slug":"action-segmentation","task_name":"Action Segmentation"},{"task_slug":"segmentation","task_name":"Segmentation"},{"task_slug":"temporal-action-segmentation","task_name":"Temporal Action Segmentation"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[{"leaderboard":"/sota/action-segmentation-on-50-salads-1","task":"Action Segmentation","dataset":"50 Salads","model":"MS-TCN","rank_in_archive_order":27,"of":28,"metrics":{"Acc":"80.7","Edit":"67.9","F1@10%":"76.3","F1@25%":"74.0","F1@50%":"64.5"},"uses_additional_data":false},{"leaderboard":"/sota/action-segmentation-on-breakfast-1","task":"Action Segmentation","dataset":"Breakfast","model":"MS-TCN (IDT)","rank_in_archive_order":30,"of":37,"metrics":{"Acc":" 65.1","Average F1":"50.6","Edit":" 61.4","F1@10%":"58.2","F1@25%":"52.9","F1@50%":"40.8"},"uses_additional_data":false},{"leaderboard":"/sota/action-segmentation-on-breakfast-1","task":"Action Segmentation","dataset":"Breakfast","model":"MS-TCN (I3D)","rank_in_archive_order":31,"of":37,"metrics":{"Acc":"66.3","Average F1":"46.2","Edit":" 61.7","F1@10%":"52.6","F1@25%":"48.1","F1@50%":"37.9"},"uses_additional_data":false},{"leaderboard":"/sota/action-segmentation-on-gtea-1","task":"Action Segmentation","dataset":"GTEA","model":"MS-TCN","rank_in_archive_order":22,"of":28,"metrics":{"Acc":"79.2","Edit":"81.4","F1@10%":"87.5","F1@25%":" 85.4","F1@50%":"74.6"},"uses_additional_data":false}],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=1903.01945","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}