{"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/end-to-end-learning-of-motion-representation","title":"End-to-End Learning of Motion Representation for Video Understanding","arxiv_id":"1804.00413","date":"2018-04-02","proceeding":"CVPR 2018 6","authors":["Lijie Fan","Wenbing Huang","Chuang Gan","Stefano Ermon","Boqing Gong","Junzhou Huang"],"abstract":"Despite the recent success of end-to-end learned representations,\nhand-crafted optical flow features are still widely used in video analysis\ntasks. To fill this gap, we propose TVNet, a novel end-to-end trainable neural\nnetwork, to learn optical-flow-like features from data. TVNet subsumes a\nspecific optical flow solver, the TV-L1 method, and is initialized by unfolding\nits optimization iterations as neural layers. TVNet can therefore be used\ndirectly without any extra learning. Moreover, it can be naturally concatenated\nwith other task-specific networks to formulate an end-to-end architecture, thus\nmaking our method more efficient than current multi-stage approaches by\navoiding the need to pre-compute and store features on disk. Finally, the\nparameters of the TVNet can be further fine-tuned by end-to-end training. This\nenables TVNet to learn richer and task-specific patterns beyond exact optical\nflow. Extensive experiments on two action recognition benchmarks verify the\neffectiveness of the proposed approach. Our TVNet achieves better accuracies\nthan all compared methods, while being competitive with the fastest counterpart\nin terms of features extraction time.","url_abs":"http://arxiv.org/abs/1804.00413v1","url_pdf":"http://arxiv.org/pdf/1804.00413v1.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":"end-to-end-learning-of-motion-representation","repo_url":"https://github.com/LijieFan/tvnet","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":0,"framework":"tf","reach":{"status":"ok","spdx":"MIT"}}],"tasks":[{"task_slug":"action-recognition-in-videos","task_name":"Action Recognition"},{"task_slug":"optical-flow-estimation","task_name":"Optical Flow Estimation"},{"task_slug":"video-understanding","task_name":"Video Understanding"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[{"leaderboard":"/sota/action-recognition-in-videos-on-hmdb-51","task":"Action Recognition","dataset":"HMDB-51","model":"TVNet+IDT","rank_in_archive_order":47,"of":77,"metrics":{"Average accuracy of 3 splits":"72.6"},"uses_additional_data":false},{"leaderboard":"/sota/action-recognition-in-videos-on-ucf101","task":"Action Recognition","dataset":"UCF101","model":"TVNet+IDT","rank_in_archive_order":45,"of":91,"metrics":{"3-fold Accuracy":"95.4"},"uses_additional_data":false}],"syntology":{"syntology_url":"https://syntology.ai/paper/1804.00413","atlas_url":"https://app.syntology.ai/?focus=1804.00413","mcp":{"get_harvested_code_for_paper":{"arxiv_id":"1804.00413"}},"developers":"https://syntology.ai/developers","read_at":"2026-09-25T09:33:49+00:00","read_at_is":"when the build read Syntology's graph, not when any sample ran","claim":"Per-sample execution status on synthesized fixtures; not a correctness claim about the paper. Samples come from repositories linked to the paper, official or community; repo_kind says which.","repos":[{"provenance":"external:paperswithcode_snapshot_2025-07-28","url":"https://github.com/LijieFan/tvnet","reach":{"status":"ok","spdx":"MIT"}}],"summary":{"unverified":2},"by_repo_kind":{"listed":{"samples":2,"ran":0,"repositories":1}},"repo_kind_vocabulary":{"official":"The archive marks this repository official for the paper","named_in_paper":"The archive records that the paper mentions this repository; it is not marked official","listed":"In the archive's code links for this paper, not marked official and not recorded as mentioned in the paper","found_in_text":"Syntology found this repository in the paper's own text; whether it is the authors' implementation is not asserted","community":"Not in the archive's code links for this paper; a community repository Syntology harvested"},"n_pointer_only_for_licence":0,"samples":[{"code_sha256_prefix":"63d1fd8578d2748b","entry":"batch_transformer","repo":"LijieFan/tvnet","repo_kind":"listed","path":"spatial_transformer.py","file_url":"https://github.com/LijieFan/tvnet/blob/HEAD/spatial_transformer.py","link_basis":"harvester_set","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"63d1fd8578d2748b"}},{"code_sha256_prefix":"51842b06389ec6fa","entry":"transformer","repo":"LijieFan/tvnet","repo_kind":"listed","path":"spatial_transformer.py","file_url":"https://github.com/LijieFan/tvnet/blob/HEAD/spatial_transformer.py","link_basis":"harvester_set","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"51842b06389ec6fa"}}]},"arxiv_metadata":null,"syntology_extracted_results":null}