{"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/a-closer-look-at-spatiotemporal-convolutions","title":"A Closer Look at Spatiotemporal Convolutions for Action Recognition","arxiv_id":"1711.11248","date":"2017-11-30","proceeding":"CVPR 2018 6","authors":["Du Tran","Heng Wang","Lorenzo Torresani","Jamie Ray","Yann Lecun","Manohar Paluri"],"abstract":"In this paper we discuss several forms of spatiotemporal convolutions for\nvideo analysis and study their effects on action recognition. Our motivation\nstems from the observation that 2D CNNs applied to individual frames of the\nvideo have remained solid performers in action recognition. In this work we\nempirically demonstrate the accuracy advantages of 3D CNNs over 2D CNNs within\nthe framework of residual learning. Furthermore, we show that factorizing the\n3D convolutional filters into separate spatial and temporal components yields\nsignificantly advantages in accuracy. Our empirical study leads to the design\nof a new spatiotemporal convolutional block \"R(2+1)D\" which gives rise to CNNs\nthat achieve results comparable or superior to the state-of-the-art on\nSports-1M, Kinetics, UCF101 and HMDB51.","url_abs":"http://arxiv.org/abs/1711.11248v3","url_pdf":"http://arxiv.org/pdf/1711.11248v3.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":"a-closer-look-at-spatiotemporal-convolutions","repo_url":"https://github.com/facebookresearch/R2Plus1D","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":1,"framework":"caffe2","reach":{"status":"ok","spdx":"Apache-2.0"}},{"paper_slug":"a-closer-look-at-spatiotemporal-convolutions","repo_url":"https://github.com/2023-MindSpore-1/ms-code-187","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"mindspore","reach":{"status":"unanswered"}},{"paper_slug":"a-closer-look-at-spatiotemporal-convolutions","repo_url":"https://github.com/2023-MindSpore-1/ms-code-68","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"mindspore","reach":{"status":"unanswered"}},{"paper_slug":"a-closer-look-at-spatiotemporal-convolutions","repo_url":"https://github.com/3dperceptionlab/visual-wetlandbirds","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":{"status":"ok","spdx":"NOASSERTION"}},{"paper_slug":"a-closer-look-at-spatiotemporal-convolutions","repo_url":"https://github.com/AD2605/Action-Recognition","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":{"status":"unanswered"}},{"paper_slug":"a-closer-look-at-spatiotemporal-convolutions","repo_url":"https://github.com/Bangbangbanana/r2plus1d_mindspore","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"mindspore","reach":{"status":"unanswered"}},{"paper_slug":"a-closer-look-at-spatiotemporal-convolutions","repo_url":"https://github.com/BelixRogner/SpeedChallenge","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":{"status":"unanswered"}},{"paper_slug":"a-closer-look-at-spatiotemporal-convolutions","repo_url":"https://github.com/JuliBaCSE/WeCare_makeathon","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":{"status":"unanswered"}},{"paper_slug":"a-closer-look-at-spatiotemporal-convolutions","repo_url":"https://github.com/anonymous-p/Flickering_Adversarial_Video","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":{"status":"ok"}},{"paper_slug":"a-closer-look-at-spatiotemporal-convolutions","repo_url":"https://github.com/facebookresearch/VMZ","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"caffe2","reach":null},{"paper_slug":"a-closer-look-at-spatiotemporal-convolutions","repo_url":"https://github.com/fmthoker/severe-benchmark","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":{"status":"ok"}},{"paper_slug":"a-closer-look-at-spatiotemporal-convolutions","repo_url":"https://github.com/juenkhaw/action_recognition_project","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":{"status":"unanswered"}},{"paper_slug":"a-closer-look-at-spatiotemporal-convolutions","repo_url":"https://github.com/juenkhaw/action_recognition_project1","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":{"status":"unanswered"}},{"paper_slug":"a-closer-look-at-spatiotemporal-convolutions","repo_url":"https://github.com/karatuno/Action-Recognition","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"tf","reach":{"status":"unanswered"}},{"paper_slug":"a-closer-look-at-spatiotemporal-convolutions","repo_url":"https://github.com/kietngt00/hmdb51-recognition","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":null},{"paper_slug":"a-closer-look-at-spatiotemporal-convolutions","repo_url":"https://github.com/leftthomas/r2plus1d-c3d","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":null},{"paper_slug":"a-closer-look-at-spatiotemporal-convolutions","repo_url":"https://github.com/wasilone11/ICPR-RIP-2024","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"none","reach":{"status":"unanswered"}},{"paper_slug":"a-closer-look-at-spatiotemporal-convolutions","repo_url":"https://github.com/2023-MindSpore-1/ms-code-216/tree/main/r2plus1d","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":0,"framework":"mindspore","reach":null},{"paper_slug":"a-closer-look-at-spatiotemporal-convolutions","repo_url":"https://github.com/ZJUT-ERCISS/zjut_mindvideo/blob/master/msvideo/models/r2plus1d.py","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":0,"framework":"mindspore","reach":null},{"paper_slug":"a-closer-look-at-spatiotemporal-convolutions","repo_url":"https://github.com/kingcong/r2plus1d","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":0,"framework":"mindspore","reach":{"status":"unanswered"}},{"paper_slug":"a-closer-look-at-spatiotemporal-convolutions","repo_url":"https://github.com/microsoft/computervision-recipes","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":0,"framework":"pytorch","reach":{"status":"ok","spdx":"MIT"}},{"paper_slug":"a-closer-look-at-spatiotemporal-convolutions","repo_url":"https://github.com/mindspore-ai/models/tree/master/research/cv/r2plus1d","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":0,"framework":"mindspore","reach":null},{"paper_slug":"a-closer-look-at-spatiotemporal-convolutions","repo_url":"https://github.com/open-mmlab/mmaction2","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":0,"framework":"pytorch","reach":null},{"paper_slug":"a-closer-look-at-spatiotemporal-convolutions","repo_url":"https://github.com/xiuyu0000/vision/blob/main/mindvision/msvideo/models/r2plus1d.py","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":0,"framework":"mindspore","reach":null}],"tasks":[{"task_slug":"action-classification","task_name":"Action Classification"},{"task_slug":"action-recognition-in-videos","task_name":"Action Recognition"},{"task_slug":"action-recognition","task_name":"Temporal Action Localization"}],"methods":[{"method_slug":"2-1-d-convolution","method_name":"(2+1)D Convolution"},{"method_slug":"average-pooling","method_name":"Average Pooling"},{"method_slug":"batch-normalization","method_name":"Batch Normalization"},{"method_slug":"dense-connections","method_name":"Dense Connections"},{"method_slug":"global-average-pooling","method_name":"Global Average Pooling"},{"method_slug":"linear-warmup","method_name":"Linear Warmup"},{"method_slug":"r-2-1-d","method_name":"R(2+1)D"},{"method_slug":"random-resized-crop","method_name":"Random Resized Crop"},{"method_slug":"relu","method_name":"ReLU"},{"method_slug":"residual-connection","method_name":"Residual Connection"},{"method_slug":"sgd-with-momentum","method_name":"SGD with Momentum"},{"method_slug":"temporal-jittering","method_name":"Temporal Jittering"},{"method_slug":"weight-decay","method_name":"Weight Decay"}],"datasets_introduced":[],"methods_introduced":[{"slug":"2-1-d-convolution","name":"(2+1)D Convolution","full_name":"(2+1)D Convolution"},{"slug":"r-2-1-d","name":"R(2+1)D","full_name":"R(2+1)D"}],"results":[{"leaderboard":"/sota/action-classification-on-kinetics-400","task":"Action Classification","dataset":"Kinetics-400","model":"R[2+1]D-Flow (Sports-1M pretrain)","rank_in_archive_order":156,"of":207,"metrics":{"Acc@1":"75.4","Acc@5":"91.9"},"uses_additional_data":false},{"leaderboard":"/sota/action-classification-on-kinetics-400","task":"Action Classification","dataset":"Kinetics-400","model":"R[2+1]D-RGB (Sports-1M pretrain)","rank_in_archive_order":163,"of":207,"metrics":{"Acc@1":"74.3","Acc@5":"91.4"},"uses_additional_data":false},{"leaderboard":"/sota/action-classification-on-kinetics-400","task":"Action Classification","dataset":"Kinetics-400","model":"R[2+1]D-Two-Stream","rank_in_archive_order":165,"of":207,"metrics":{"Acc@1":"73.9","Acc@5":"90.9"},"uses_additional_data":false},{"leaderboard":"/sota/action-classification-on-kinetics-400","task":"Action Classification","dataset":"Kinetics-400","model":"R[2+1]D","rank_in_archive_order":175,"of":207,"metrics":{"Acc@1":"72","Acc@5":"90"},"uses_additional_data":false},{"leaderboard":"/sota/action-classification-on-kinetics-400","task":"Action Classification","dataset":"Kinetics-400","model":"R[2+1]D-RGB","rank_in_archive_order":176,"of":207,"metrics":{"Acc@1":"72","Acc@5":"90"},"uses_additional_data":false},{"leaderboard":"/sota/action-classification-on-kinetics-400","task":"Action Classification","dataset":"Kinetics-400","model":"R[2+1]D-Flow","rank_in_archive_order":185,"of":207,"metrics":{"Acc@1":"67.5","Acc@5":"87.2"},"uses_additional_data":false},{"leaderboard":"/sota/action-recognition-in-videos-on-hmdb-51","task":"Action Recognition","dataset":"HMDB-51","model":"R[2+1]D-TwoStream (Kinetics pretrained)","rank_in_archive_order":28,"of":77,"metrics":{"Average accuracy of 3 splits":"78.7"},"uses_additional_data":true},{"leaderboard":"/sota/action-recognition-in-videos-on-hmdb-51","task":"Action Recognition","dataset":"HMDB-51","model":"R[2+1]D-Flow (Kinetics pretrained)","rank_in_archive_order":36,"of":77,"metrics":{"Average accuracy of 3 splits":"76.4"},"uses_additional_data":true},{"leaderboard":"/sota/action-recognition-in-videos-on-hmdb-51","task":"Action Recognition","dataset":"HMDB-51","model":"R[2+1]D-RGB (Kinetics pretrained)","rank_in_archive_order":41,"of":77,"metrics":{"Average accuracy of 3 splits":"74.5"},"uses_additional_data":true},{"leaderboard":"/sota/action-recognition-in-videos-on-hmdb-51","task":"Action Recognition","dataset":"HMDB-51","model":"R[2+1D]D-TwoStream (Sports1M pretrained)","rank_in_archive_order":46,"of":77,"metrics":{"Average accuracy of 3 splits":"72.7"},"uses_additional_data":true},{"leaderboard":"/sota/action-recognition-in-videos-on-hmdb-51","task":"Action Recognition","dataset":"HMDB-51","model":"R[2+1]D-Flow (Sports1M pretrained)","rank_in_archive_order":55,"of":77,"metrics":{"Average accuracy of 3 splits":"70.1"},"uses_additional_data":true},{"leaderboard":"/sota/action-recognition-in-videos-on-hmdb-51","task":"Action Recognition","dataset":"HMDB-51","model":"R[2+1]D-RGB (Sports1M pretrained)","rank_in_archive_order":59,"of":77,"metrics":{"Average accuracy of 3 splits":"66.6"},"uses_additional_data":true},{"leaderboard":"/sota/action-recognition-in-videos-on-sports-1m","task":"Action Recognition","dataset":"Sports-1M","model":"R[2+1]D-Two-Stream-32frame","rank_in_archive_order":3,"of":9,"metrics":{"Video hit@1 ":"73.3","Video hit@5":"91.9"},"uses_additional_data":false},{"leaderboard":"/sota/action-recognition-in-videos-on-sports-1m","task":"Action Recognition","dataset":"Sports-1M","model":"R[2+1]D-RGB-32frame","rank_in_archive_order":4,"of":9,"metrics":{"Clip Hit@1":"57","Video hit@1 ":"73","Video hit@5":"91.5"},"uses_additional_data":false},{"leaderboard":"/sota/action-recognition-in-videos-on-sports-1m","task":"Action Recognition","dataset":"Sports-1M","model":"R[2+1]D-Flow-32frame","rank_in_archive_order":6,"of":9,"metrics":{"Clip Hit@1":"46.4","Video hit@1 ":"68.4","Video hit@5":"88.7"},"uses_additional_data":false},{"leaderboard":"/sota/action-recognition-in-videos-on-ucf101","task":"Action Recognition","dataset":"UCF101","model":"R[2+1]D-TwoStream (Kinetics pretrained)","rank_in_archive_order":20,"of":91,"metrics":{"3-fold Accuracy":"97.3"},"uses_additional_data":true},{"leaderboard":"/sota/action-recognition-in-videos-on-ucf101","task":"Action Recognition","dataset":"UCF101","model":"R[2+1]D-RGB (Kinetics pretrained)","rank_in_archive_order":30,"of":91,"metrics":{"3-fold Accuracy":"96.8"},"uses_additional_data":true},{"leaderboard":"/sota/action-recognition-in-videos-on-ucf101","task":"Action Recognition","dataset":"UCF101","model":"R[2+1]D-Flow (Kinetics pretrained)","rank_in_archive_order":44,"of":91,"metrics":{"3-fold Accuracy":"95.5"},"uses_additional_data":true},{"leaderboard":"/sota/action-recognition-in-videos-on-ucf101","task":"Action Recognition","dataset":"UCF101","model":"R[2+1]D-TwoStream (Sports-1M pretrained)","rank_in_archive_order":49,"of":91,"metrics":{"3-fold Accuracy":"95"},"uses_additional_data":true},{"leaderboard":"/sota/action-recognition-in-videos-on-ucf101","task":"Action Recognition","dataset":"UCF101","model":"R[2+1]D-RGB (Sports-1M pretrained)","rank_in_archive_order":59,"of":91,"metrics":{"3-fold Accuracy":"93.6"},"uses_additional_data":true},{"leaderboard":"/sota/action-recognition-in-videos-on-ucf101","task":"Action Recognition","dataset":"UCF101","model":"R[2+1]D-Flow (Sports-1M pretrained)","rank_in_archive_order":61,"of":91,"metrics":{"3-fold Accuracy":"93.3"},"uses_additional_data":true}],"syntology":{"syntology_url":"https://syntology.ai/paper/1711.11248","atlas_url":"https://app.syntology.ai/?focus=1711.11248","mcp":{"get_harvested_code_for_paper":{"arxiv_id":"1711.11248"}},"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/facebookresearch/VMZ","reach":null},{"provenance":"external:paperswithcode_snapshot_2025-07-28","url":"https://github.com/wasilone11/ICPR-RIP-2024","reach":{"status":"unanswered"}},{"provenance":"external:paperswithcode_snapshot_2025-07-28","url":"https://github.com/AD2605/Action-Recognition","reach":{"status":"unanswered"}},{"provenance":"external:paperswithcode_snapshot_2025-07-28","url":"https://github.com/ZJUT-ERCISS/zjut_mindvideo/blob/master/msvideo/models/r2plus1d.py","reach":null},{"provenance":"external:paperswithcode_snapshot_2025-07-28","url":"https://github.com/JuliBaCSE/WeCare_makeathon","reach":{"status":"unanswered"}},{"provenance":"external:paperswithcode_snapshot_2025-07-28","url":"https://github.com/juenkhaw/action_recognition_project1","reach":{"status":"unanswered"}},{"provenance":"external:paperswithcode_snapshot_2025-07-28","url":"https://github.com/karatuno/Action-Recognition","reach":{"status":"unanswered"}},{"provenance":"external:paperswithcode_snapshot_2025-07-28","url":"https://github.com/2023-MindSpore-1/ms-code-68","reach":{"status":"unanswered"}},{"provenance":"external:paperswithcode_snapshot_2025-07-28","url":"https://github.com/anonymous-p/Flickering_Adversarial_Video","reach":{"status":"ok"}},{"provenance":"external:paperswithcode_snapshot_2025-07-28","url":"https://github.com/3dperceptionlab/visual-wetlandbirds","reach":{"status":"ok","spdx":"NOASSERTION"}},{"provenance":"external:paperswithcode_snapshot_2025-07-28","url":"https://github.com/mindspore-ai/models/tree/master/research/cv/r2plus1d","reach":null},{"provenance":"external:paperswithcode_snapshot_2025-07-28","url":"https://github.com/facebookresearch/R2Plus1D","reach":{"status":"ok","spdx":"Apache-2.0"}},{"provenance":"external:paperswithcode_snapshot_2025-07-28","url":"https://github.com/2023-MindSpore-1/ms-code-187","reach":{"status":"unanswered"}},{"provenance":"external:paperswithcode_snapshot_2025-07-28","url":"https://github.com/2023-MindSpore-1/ms-code-216/tree/main/r2plus1d","reach":null},{"provenance":"external:paperswithcode_snapshot_2025-07-28","url":"https://github.com/kingcong/r2plus1d","reach":{"status":"unanswered"}},{"provenance":"external:paperswithcode_snapshot_2025-07-28","url":"https://github.com/fmthoker/severe-benchmark","reach":{"status":"ok"}},{"provenance":"external:paperswithcode_snapshot_2025-07-28","url":"https://github.com/juenkhaw/action_recognition_project","reach":{"status":"unanswered"}},{"provenance":"external:paperswithcode_snapshot_2025-07-28","url":"https://github.com/leftthomas/r2plus1d-c3d","reach":null},{"provenance":"external:paperswithcode_snapshot_2025-07-28","url":"https://github.com/kietngt00/hmdb51-recognition","reach":null},{"provenance":"external:paperswithcode_snapshot_2025-07-28","url":"https://github.com/BelixRogner/SpeedChallenge","reach":{"status":"unanswered"}},{"provenance":"external:paperswithcode_snapshot_2025-07-28","url":"https://github.com/xiuyu0000/vision/blob/main/mindvision/msvideo/models/r2plus1d.py","reach":null},{"provenance":"external:paperswithcode_snapshot_2025-07-28","url":"https://github.com/Bangbangbanana/r2plus1d_mindspore","reach":{"status":"unanswered"}},{"provenance":"external:paperswithcode_snapshot_2025-07-28","url":"https://github.com/open-mmlab/mmaction2","reach":null},{"provenance":"external:paperswithcode_snapshot_2025-07-28","url":"https://github.com/microsoft/computervision-recipes","reach":{"status":"ok","spdx":"MIT"}}],"summary":{"ran_draft_wrong":1,"unverified":3},"by_repo_kind":{"listed":{"samples":4,"ran":1,"repositories":2}},"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":4,"samples":[{"code_sha256_prefix":"3c34bc53cbb7d79d","entry":"r3d_18","repo":"kietngt00/hmdb51-recognition","repo_kind":"listed","path":"model/R2plus1D.py","file_url":"https://github.com/kietngt00/hmdb51-recognition/blob/HEAD/model/R2plus1D.py","link_basis":"first_harvest_node","language":"python","status":"ran_draft_wrong","verification_level":1,"contract_check":"OUTPUT_MISDECLARED","metamorphic_tier":"deterministic","behaviour_fingerprint":false,"licence":"NONE","inline_ok":false,"mcp_get_code":{"code_sha256":"3c34bc53cbb7d79d"}},{"code_sha256_prefix":"b09d9df7736daaa3","entry":"center_crop","repo":"leftthomas/r2plus1d-c3d","repo_kind":"listed","path":"inference.py","file_url":"https://github.com/leftthomas/r2plus1d-c3d/blob/HEAD/inference.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"NONE","inline_ok":false,"mcp_get_code":{"code_sha256":"b09d9df7736daaa3"}},{"code_sha256_prefix":"b0848b1c8d35150b","entry":"mc3_18","repo":"kietngt00/hmdb51-recognition","repo_kind":"listed","path":"model/R2plus1D.py","file_url":"https://github.com/kietngt00/hmdb51-recognition/blob/HEAD/model/R2plus1D.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"NONE","inline_ok":false,"mcp_get_code":{"code_sha256":"b0848b1c8d35150b"}},{"code_sha256_prefix":"8fb2663dede3c507","entry":"r2plus1d_18","repo":"kietngt00/hmdb51-recognition","repo_kind":"listed","path":"model/R2plus1D.py","file_url":"https://github.com/kietngt00/hmdb51-recognition/blob/HEAD/model/R2plus1D.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"NONE","inline_ok":false,"mcp_get_code":{"code_sha256":"8fb2663dede3c507"}}]},"arxiv_metadata":null,"syntology_extracted_results":null}