{"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/scaling-open-vocabulary-action-detection","title":"Scaling Open-Vocabulary Action Detection","arxiv_id":"2504.03096","date":"2025-04-04","proceeding":null,"authors":["Zhen Hao Sia","Yogesh Singh Rawat"],"abstract":"In this work, we focus on scaling open-vocabulary action detection. Existing approaches for action detection are predominantly limited to closed-set scenarios and rely on complex, parameter-heavy architectures. Extending these models to the open-vocabulary setting poses two key challenges: (1) the lack of large-scale datasets with many action classes for robust training, and (2) parameter-heavy adaptations to a pretrained vision-language contrastive model to convert it for detection, risking overfitting the additional non-pretrained parameters to base action classes. Firstly, we introduce an encoder-only multimodal model for video action detection, reducing the reliance on parameter-heavy additions for video action detection. Secondly, we introduce a simple weakly supervised training strategy to exploit an existing closed-set action detection dataset for pretraining. Finally, we depart from the ill-posed base-to-novel benchmark used by prior works in open-vocabulary action detection and devise a new benchmark to evaluate on existing closed-set action detection datasets without ever using them for training, showing novel results to serve as baselines for future work.","url_abs":"https://arxiv.org/abs/2504.03096v1","url_pdf":"https://arxiv.org/pdf/2504.03096v1.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":"scaling-open-vocabulary-action-detection","repo_url":"https://github.com/siatheindochinese/sia_act_placeholder","is_official":1,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"none","reach":null}],"tasks":[{"task_slug":"action-detection","task_name":"Action Detection"},{"task_slug":"multiple-action-detection","task_name":"Multiple Action Detection"},{"task_slug":"open-vocabulary-action-detection","task_name":"Open Vocabulary Action Detection"},{"task_slug":"spatio-temporal-action-localization","task_name":"Spatio-Temporal Action Localization"},{"task_slug":"video-action-detection","task_name":"Video Action Detection"},{"task_slug":"zero-shot-action-detection","task_name":"Zero-Shot Action Detection"}],"methods":[{"method_slug":"base","method_name":"BASE"},{"method_slug":"focus","method_name":"Focus"}],"datasets_introduced":[],"methods_introduced":[],"results":[{"leaderboard":"/sota/action-detection-on-j-hmdb","task":"Action Detection","dataset":"J-HMDB","model":"SiA","rank_in_archive_order":1,"of":18,"metrics":{"Frame-mAP 0.5":"88.5"},"uses_additional_data":false},{"leaderboard":"/sota/action-detection-on-multisports","task":"Action Detection","dataset":"MultiSports","model":"SiA","rank_in_archive_order":2,"of":2,"metrics":{"Frame-mAP 0.5":"28.8"},"uses_additional_data":false},{"leaderboard":"/sota/action-detection-on-ucf101-24","task":"Action Detection","dataset":"UCF101-24","model":"SiA","rank_in_archive_order":2,"of":19,"metrics":{"Frame-mAP 0.5":"88.5"},"uses_additional_data":false},{"leaderboard":"/sota/open-vocabulary-action-detection-on-jhmdb","task":"Open Vocabulary Action Detection","dataset":"JHMDB","model":"SiA","rank_in_archive_order":1,"of":1,"metrics":{"val mAP":"57.1"},"uses_additional_data":false},{"leaderboard":"/sota/open-vocabulary-action-detection-on","task":"Open Vocabulary Action Detection","dataset":"MultiSports","model":"SiA","rank_in_archive_order":1,"of":1,"metrics":{"val mAP":"1.3"},"uses_additional_data":false},{"leaderboard":"/sota/open-vocabulary-action-detection-on-ucf101-24","task":"Open Vocabulary Action Detection","dataset":"UCF101-24","model":"SiA","rank_in_archive_order":1,"of":1,"metrics":{"val mAP":"42.6"},"uses_additional_data":false}],"syntology":{"atlas_url":null,"mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}