{"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/excl-extractive-clip-localization-using","title":"ExCL: Extractive Clip Localization Using Natural Language Descriptions","arxiv_id":"1904.02755","date":"2019-04-04","proceeding":"NAACL 2019 6","authors":["Soham Ghosh","Anuva Agarwal","Zarana Parekh","Alexander Hauptmann"],"abstract":"The task of retrieving clips within videos based on a given natural language\nquery requires cross-modal reasoning over multiple frames. Prior approaches\nsuch as sliding window classifiers are inefficient, while text-clip similarity\ndriven ranking-based approaches such as segment proposal networks are far more\ncomplicated. In order to select the most relevant video clip corresponding to\nthe given text description, we propose a novel extractive approach that\npredicts the start and end frames by leveraging cross-modal interactions\nbetween the text and video - this removes the need to retrieve and re-rank\nmultiple proposal segments. Using recurrent networks we encode the two\nmodalities into a joint representation which is then used in different variants\nof start-end frame predictor networks. Through extensive experimentation and\nablative analysis, we demonstrate that our simple and elegant approach\nsignificantly outperforms state of the art on two datasets and has comparable\nperformance on a third.","url_abs":"http://arxiv.org/abs/1904.02755v1","url_pdf":"http://arxiv.org/pdf/1904.02755v1.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":"excl-extractive-clip-localization-using","repo_url":"https://github.com/jayleicn/TVRetrieval","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":{"status":"ok","spdx":"MIT"}}],"tasks":[],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"syntology_url":"https://syntology.ai/paper/1904.02755","atlas_url":"https://app.syntology.ai/?focus=1904.02755","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}