{"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/umt-unified-multi-modal-transformers-for","title":"UMT: Unified Multi-modal Transformers for Joint Video Moment Retrieval and Highlight Detection","arxiv_id":"2203.12745","date":"2022-03-23","proceeding":"CVPR 2022 1","authors":["Ye Liu","Siyuan Li","Yang Wu","Chang Wen Chen","Ying Shan","XiaoHu Qie"],"abstract":"Finding relevant moments and highlights in videos according to natural language queries is a natural and highly valuable common need in the current video content explosion era. Nevertheless, jointly conducting moment retrieval and highlight detection is an emerging research topic, even though its component problems and some related tasks have already been studied for a while. In this paper, we present the first unified framework, named Unified Multi-modal Transformers (UMT), capable of realizing such joint optimization while can also be easily degenerated for solving individual problems. As far as we are aware, this is the first scheme to integrate multi-modal (visual-audio) learning for either joint optimization or the individual moment retrieval task, and tackles moment retrieval as a keypoint detection problem using a novel query generator and query decoder. Extensive comparisons with existing methods and ablation studies on QVHighlights, Charades-STA, YouTube Highlights, and TVSum datasets demonstrate the effectiveness, superiority, and flexibility of the proposed method under various settings. Source code and pre-trained models are available at https://github.com/TencentARC/UMT.","url_abs":"https://arxiv.org/abs/2203.12745v2","url_pdf":"https://arxiv.org/pdf/2203.12745v2.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":"umt-unified-multi-modal-transformers-for","repo_url":"https://github.com/tencentarc/umt","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":1,"framework":"pytorch","reach":{"status":"ok","spdx":"NOASSERTION"}},{"paper_slug":"umt-unified-multi-modal-transformers-for","repo_url":"https://github.com/MS-P3/code7/tree/main/umt5","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":0,"framework":"mindspore","reach":{"status":"ok","spdx":"Apache-2.0"}},{"paper_slug":"umt-unified-multi-modal-transformers-for","repo_url":"https://github.com/MindCode-4/code-5/tree/main/umt5","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":0,"framework":"mindspore","reach":null}],"tasks":[{"task_slug":"decoder","task_name":"Decoder"},{"task_slug":"highlight-detection","task_name":"Highlight Detection"},{"task_slug":"moment-retrieval","task_name":"Moment Retrieval"},{"task_slug":"natural-language-queries","task_name":"Natural Language Queries"},{"task_slug":"retrieval","task_name":"Retrieval"},{"task_slug":"video-grounding","task_name":"Video Grounding"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[{"leaderboard":"/sota/highlight-detection-on-qvhighlights","task":"Highlight Detection","dataset":"QVHighlights","model":"UMT (w. PT)","rank_in_archive_order":12,"of":21,"metrics":{"mAP":"39.12"},"uses_additional_data":false},{"leaderboard":"/sota/highlight-detection-on-qvhighlights","task":"Highlight Detection","dataset":"QVHighlights","model":"UMT","rank_in_archive_order":17,"of":21,"metrics":{"mAP":"38.18"},"uses_additional_data":false},{"leaderboard":"/sota/highlight-detection-on-tvsum","task":"Highlight Detection","dataset":"TvSum","model":"UMT","rank_in_archive_order":7,"of":7,"metrics":{"mAP":"83.1"},"uses_additional_data":false},{"leaderboard":"/sota/highlight-detection-on-youtube-highlights","task":"Highlight Detection","dataset":"YouTube Highlights","model":"UMT","rank_in_archive_order":7,"of":7,"metrics":{"mAP":"74.9"},"uses_additional_data":false},{"leaderboard":"/sota/moment-retrieval-on-charades-sta","task":"Moment Retrieval","dataset":"Charades-STA","model":"UMT (VO)","rank_in_archive_order":22,"of":25,"metrics":{"R@1 IoU=0.5":"49.35","R@1 IoU=0.7":"26.16","R@5 IoU=0.5":"89.41","R@5 IoU=0.7":"54.95"},"uses_additional_data":false},{"leaderboard":"/sota/moment-retrieval-on-charades-sta","task":"Moment Retrieval","dataset":"Charades-STA","model":"UMT (VA)","rank_in_archive_order":24,"of":25,"metrics":{"R@1 IoU=0.5":"48.31","R@1 IoU=0.7":"29.25","R@5 IoU=0.5":"88.79","R@5 IoU=0.7":"56.08"},"uses_additional_data":false},{"leaderboard":"/sota/moment-retrieval-on-qvhighlights","task":"Moment Retrieval","dataset":"QVHighlights","model":"UMT (w/ audio + PT ASR Cpations)","rank_in_archive_order":25,"of":32,"metrics":{"mAP":"38.08"},"uses_additional_data":false},{"leaderboard":"/sota/moment-retrieval-on-qvhighlights","task":"Moment Retrieval","dataset":"QVHighlights","model":"UMT","rank_in_archive_order":27,"of":32,"metrics":{"mAP":"36.12"},"uses_additional_data":false},{"leaderboard":"/sota/video-grounding-on-qvhighlights","task":"Video Grounding","dataset":"QVHighlights","model":"UMT","rank_in_archive_order":6,"of":7,"metrics":{"R@1,IoU=0.5":"56.23","R@1,IoU=0.7":"41.18"},"uses_additional_data":false}],"syntology":{"syntology_url":null,"atlas_url":"https://app.syntology.ai/?focus=2203.12745","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}