{"about":{"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.","site":"https://codewithpapers.app","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","syntology":{"site":"https://syntology.ai","developers":"https://syntology.ai/developers","mcp":{"server":"https://syntology.ai/mcp","transport":"streamable-http","server_card":"https://syntology.ai/.well-known/mcp/server-card.json","auth":{"type":"trial token, no account","trial_token":"https://syntology.ai/api/oauth/trial/token","method":"POST","docs":"https://syntology.ai/developers"}},"have":"https://syntology.ai/api/graph/have?x=<method, arXiv id or title> (free, answers coverage only)","paper_base":"https://syntology.ai/paper/","atlas_base":"https://app.syntology.ai/?focus="},"machine_readable":[{"url":"https://codewithpapers.app/llms.txt","what":"the machine catalog: every machine-readable file, counted"},{"url":"https://codewithpapers.app/index/manifest.json","what":"paper-to-code index by arXiv id, with Syntology's counts"},{"url":"https://codewithpapers.app/search/manifest.json","what":"site search index (titles, authors) and its files"},{"url":"https://codewithpapers.app/download","what":"bulk files: Syntology's layer, described there"},{"url":"https://codewithpapers.app/build_manifest.json","what":"the build record: inputs, counts, exclusions, probes"}]},"url":"/paper/decafnet-delegate-and-conquer-for-efficient","title":"DeCafNet: Delegate and Conquer for Efficient Temporal Grounding in Long Videos","arxiv_id":"2505.16376","date":"2025-05-22","proceeding":"CVPR 2025 1","authors":["Zijia Lu","A S M Iftekhar","Gaurav Mittal","Tianjian Meng","Xiawei Wang","Cheng Zhao","Rohith Kukkala","Ehsan Elhamifar","Mei Chen"],"abstract":"Long Video Temporal Grounding (LVTG) aims at identifying specific moments within lengthy videos based on user-provided text queries for effective content retrieval. The approach taken by existing methods of dividing video into clips and processing each clip via a full-scale expert encoder is challenging to scale due to prohibitive computational costs of processing a large number of clips in long videos. To address this issue, we introduce DeCafNet, an approach employing ``delegate-and-conquer'' strategy to achieve computation efficiency without sacrificing grounding performance. DeCafNet introduces a sidekick encoder that performs dense feature extraction over all video clips in a resource-efficient manner, while generating a saliency map to identify the most relevant clips for full processing by the expert encoder. To effectively leverage features from sidekick and expert encoders that exist at different temporal resolutions, we introduce DeCaf-Grounder, which unifies and refines them via query-aware temporal aggregation and multi-scale temporal refinement for accurate grounding. Experiments on two LTVG benchmark datasets demonstrate that DeCafNet reduces computation by up to 47\\% while still outperforming existing methods, establishing a new state-of-the-art for LTVG in terms of both efficiency and performance. Our code is available at https://github.com/ZijiaLewisLu/CVPR2025-DeCafNet.","url_abs":"https://arxiv.org/abs/2505.16376v1","url_pdf":"https://arxiv.org/pdf/2505.16376v1.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":"decafnet-delegate-and-conquer-for-efficient","repo_url":"https://github.com/zijialewislu/cvpr2025-decafnet","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":0,"framework":"pytorch","reach":null}],"tasks":[{"task_slug":"natural-language-moment-retrieval","task_name":"Natural Language Moment Retrieval"},{"task_slug":"natural-language-queries","task_name":"Natural Language Queries"},{"task_slug":"temporal-sentence-grounding","task_name":"Temporal Sentence Grounding"},{"task_slug":"video-grounding","task_name":"Video Grounding"}],"methods":[{"method_slug":"clip","method_name":"CLIP"}],"datasets_introduced":[],"methods_introduced":[],"results":[{"leaderboard":"/sota/natural-language-moment-retrieval-on-mad","task":"Natural Language Moment Retrieval","dataset":"MAD","model":"DeCafNet","rank_in_archive_order":2,"of":8,"metrics":{"R@1,IoU=0.1":"13.25","R@1,IoU=0.3":"10.96","R@1,IoU=0.5":"7.06","R@5,IoU=0.1":"27.73","R@5,IoU=0.3":"23.68","R@5,IoU=0.5":"16.13"},"uses_additional_data":false},{"leaderboard":"/sota/natural-language-moment-retrieval-on-tacos","task":"Natural Language Moment Retrieval","dataset":"TACoS","model":"DeCafNet","rank_in_archive_order":3,"of":13,"metrics":{"R@1,IoU=0.3":"57.36","R@1,IoU=0.5":"46.79","R@5,IoU=0.1":"81.05","R@5,IoU=0.3":"71.13"},"uses_additional_data":false},{"leaderboard":"/sota/natural-language-queries-on-ego4d","task":"Natural Language Queries","dataset":"Ego4D","model":"DeCafNet-100%","rank_in_archive_order":2,"of":10,"metrics":{"R@1 IoU=0.3":"22.21","R@1 IoU=0.5":"15.52","R@1 Mean(0.3 and 0.5)":"18.86","R@5 IoU=0.3":"45.63","R@5 IoU=0.5":"33.93"},"uses_additional_data":false},{"leaderboard":"/sota/natural-language-queries-on-ego4d","task":"Natural Language Queries","dataset":"Ego4D","model":"DeCafNet-50%","rank_in_archive_order":3,"of":10,"metrics":{"R@1 IoU=0.3":"20.81","R@1 IoU=0.5":"15.04","R@1 Mean(0.3 and 0.5)":"17.93","R@5 IoU=0.3":"42.40","R@5 IoU=0.5":"31.68"},"uses_additional_data":false},{"leaderboard":"/sota/natural-language-queries-on-ego4d","task":"Natural Language Queries","dataset":"Ego4D","model":"DeCafNet-50% (no NaQ)","rank_in_archive_order":5,"of":10,"metrics":{"R@1 IoU=0.3":"18.10","R@1 IoU=0.5":"12.55","R@1 Mean(0.3 and 0.5)":"15.32","R@5 IoU=0.3":"38.85","R@5 IoU=0.5":"28.27"},"uses_additional_data":false},{"leaderboard":"/sota/temporal-sentence-grounding-on-charades-sta","task":"Temporal Sentence Grounding","dataset":"Charades-STA","model":"DeCafNet","rank_in_archive_order":1,"of":13,"metrics":{"R1@0.5":"68.79","R1@0.7":"47.55","R5@0.5":"91.53","R5@0.7":"72.96"},"uses_additional_data":false},{"leaderboard":"/sota/temporal-sentence-grounding-on-ego4d-goalstep","task":"Temporal Sentence Grounding","dataset":"Ego4D-Goalstep","model":"DeCafNet-100%","rank_in_archive_order":1,"of":3,"metrics":{"R@1,IoU=0.3":"23.20","R@1,IoU=0.5":"19.40","R@5,IoU=0.3":"51.38","R@5,IoU=0.5":"44.17"},"uses_additional_data":false},{"leaderboard":"/sota/temporal-sentence-grounding-on-ego4d-goalstep","task":"Temporal Sentence Grounding","dataset":"Ego4D-Goalstep","model":"DeCafNet-50%","rank_in_archive_order":2,"of":3,"metrics":{"R@1,IoU=0.3":"21.29","R@1,IoU=0.5":"17.46","R@5,IoU=0.3":"47.27","R@5,IoU=0.5":"40.40"},"uses_additional_data":false},{"leaderboard":"/sota/video-grounding-on-mad","task":"Video Grounding","dataset":"MAD","model":"DeCafNet","rank_in_archive_order":1,"of":2,"metrics":{"R@1,IoU=0.1":"13.25","R@1,IoU=0.3":"10.96","R@5,IoU=0.1":"27.73","R@5,IoU=0.3":"23.68"},"uses_additional_data":false}],"syntology":{"syntology_url":"https://syntology.ai/paper/2505.16376","atlas_url":"https://app.syntology.ai/?focus=2505.16376","mcp":null,"developers":"https://syntology.ai/developers","held":true,"reason":"arxiv_id_in_atlas","checked_against":{"atlas_ids":549449,"atlas_ids_sha256":"afd72399dafd9906e9062927f70655952385a3d75488558e1f0484e950ab09b7","read_at":"2026-10-05T01:46:21Z"},"agent_calls":[{"tool":"get_citation_path","arguments":{"paper_1":"2505.16376"},"arguments_in":{"paper_2":"another paper's arXiv id or title"},"call":"get_citation_path(paper_1=\"2505.16376\", paper_2=\"…\")"},{"tool":"get_concepts_for_paper","arguments":{"arxiv_id":"2505.16376"},"call":"get_concepts_for_paper(arxiv_id=\"2505.16376\")"}],"by_repo":[{"repo":"zijialewislu/cvpr2025-decafnet","kind":"official","samples":0,"ran":0,"instrument":0,"unverified":0}],"by_repo_is":"one row per repository in the archive's code links, Syntology's graph links or the samples; samples 0 means no sample is linked to this paper, a statement about Syntology's coverage, not about the repository: with harvested_for_other_papers present, Syntology harvested the repository for that many other papers, and without it, Syntology harvested nothing from it; placed_by_identical_code counts samples whose link records no repository, placed here because identical code was harvested from this repository for another paper; ran counts executions on a synthesized input and instrument counts failures of Syntology's instrument, not of the code"},"code_links_note":null,"arxiv_metadata":null,"syntology_extracted_results":null,"record_sha256":"ebffe69eae345399f346e35a4c6ead6eac7065c5cf56c71a80b94b56c78ed422","record_changed_at":"2026-09-28","record_changed_at_basis":"first_hashed"}