{"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/deformable-part-models-are-convolutional","title":"Deformable Part Models are Convolutional Neural Networks","arxiv_id":"1409.5403","date":"2014-09-18","proceeding":"CVPR 2015 6","authors":["Ross Girshick","Forrest Iandola","Trevor Darrell","Jitendra Malik"],"abstract":"Deformable part models (DPMs) and convolutional neural networks (CNNs) are\ntwo widely used tools for visual recognition. They are typically viewed as\ndistinct approaches: DPMs are graphical models (Markov random fields), while\nCNNs are \"black-box\" non-linear classifiers. In this paper, we show that a DPM\ncan be formulated as a CNN, thus providing a novel synthesis of the two ideas.\nOur construction involves unrolling the DPM inference algorithm and mapping\neach step to an equivalent (and at times novel) CNN layer. From this\nperspective, it becomes natural to replace the standard image features used in\nDPM with a learned feature extractor. We call the resulting model DeepPyramid\nDPM and experimentally validate it on PASCAL VOC. DeepPyramid DPM significantly\noutperforms DPMs based on histograms of oriented gradients features (HOG) and\nslightly outperforms a comparable version of the recently introduced R-CNN\ndetection system, while running an order of magnitude faster.","url_abs":"http://arxiv.org/abs/1409.5403v2","url_pdf":"http://arxiv.org/pdf/1409.5403v2.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":"deformable-part-models-are-convolutional","repo_url":"https://github.com/rbgirshick/DeepPyramid","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"none","reach":{"status":"ok","spdx":"BSD-2-Clause"}}],"tasks":[{"task_slug":"object-detection","task_name":"Object Detection"},{"task_slug":"unrolling","task_name":"Rolling Shutter Correction"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[{"leaderboard":"/sota/object-detection-on-pascal-voc-2007","task":"Object Detection","dataset":"PASCAL VOC 2007","model":"Deformable Parts Model (DeepPyramid)","rank_in_archive_order":27,"of":30,"metrics":{"MAP":"45.2%"},"uses_additional_data":false}],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=1409.5403","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}