{"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/delving-deeper-into-convolutional-networks","title":"Delving Deeper into Convolutional Networks for Learning Video Representations","arxiv_id":"1511.06432","date":"2015-11-19","proceeding":null,"authors":["Nicolas Ballas","Li Yao","Chris Pal","Aaron Courville"],"abstract":"We propose an approach to learn spatio-temporal features in videos from\nintermediate visual representations we call \"percepts\" using\nGated-Recurrent-Unit Recurrent Networks (GRUs).Our method relies on percepts\nthat are extracted from all level of a deep convolutional network trained on\nthe large ImageNet dataset. While high-level percepts contain highly\ndiscriminative information, they tend to have a low-spatial resolution.\nLow-level percepts, on the other hand, preserve a higher spatial resolution\nfrom which we can model finer motion patterns. Using low-level percepts can\nleads to high-dimensionality video representations. To mitigate this effect and\ncontrol the model number of parameters, we introduce a variant of the GRU model\nthat leverages the convolution operations to enforce sparse connectivity of the\nmodel units and share parameters across the input spatial locations.\n  We empirically validate our approach on both Human Action Recognition and\nVideo Captioning tasks. In particular, we achieve results equivalent to\nstate-of-art on the YouTube2Text dataset using a simpler text-decoder model and\nwithout extra 3D CNN features.","url_abs":"http://arxiv.org/abs/1511.06432v4","url_pdf":"http://arxiv.org/pdf/1511.06432v4.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":"delving-deeper-into-convolutional-networks","repo_url":"https://github.com/bionick87/ConvGRUCell-pytorch","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":{"status":"ok"}},{"paper_slug":"delving-deeper-into-convolutional-networks","repo_url":"https://github.com/jacobkimmel/pytorch_convgru","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":{"status":"ok","spdx":"MIT"}},{"paper_slug":"delving-deeper-into-convolutional-networks","repo_url":"https://github.com/yaoli/arctic-capgen-vid","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"none","reach":{"status":"ok","spdx":"NOASSERTION"}},{"paper_slug":"delving-deeper-into-convolutional-networks","repo_url":"https://github.com/MindCode-4/code-11/tree/main/delving-deeper-into-convolutional","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":0,"framework":"mindspore","reach":null},{"paper_slug":"delving-deeper-into-convolutional-networks","repo_url":"https://github.com/MindCode-4/code-6/tree/main/dice-loss-for-data-imbalanced-nlp-tasks","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":0,"framework":"mindspore","reach":null},{"paper_slug":"delving-deeper-into-convolutional-networks","repo_url":"https://github.com/MindSpore-scientific-2/code-8/tree/main/pytorch_convgru","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":0,"framework":"mindspore","reach":null}],"tasks":[{"task_slug":"action-recognition-in-videos","task_name":"Action Recognition"},{"task_slug":"decoder","task_name":"Decoder"},{"task_slug":"action-recognition","task_name":"Temporal Action Localization"},{"task_slug":"video-captioning","task_name":"Video Captioning"}],"methods":[{"method_slug":"cgru","method_name":"CGRU"},{"method_slug":"convolution","method_name":"Convolution"},{"method_slug":"gru","method_name":"GRU"},{"method_slug":"relu","method_name":"ReLU"},{"method_slug":"sigmoid-activation","method_name":"Sigmoid Activation"}],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=1511.06432","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}