{"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/convolutional-channel-features","title":"Convolutional Channel Features","arxiv_id":"1504.07339","date":"2015-04-28","proceeding":"ICCV 2015 12","authors":["Bin Yang","Junjie Yan","Zhen Lei","Stan Z. Li"],"abstract":"Deep learning methods are powerful tools but often suffer from expensive\ncomputation and limited flexibility. An alternative is to combine light-weight\nmodels with deep representations. As successful cases exist in several visual\nproblems, a unified framework is absent. In this paper, we revisit two widely\nused approaches in computer vision, namely filtered channel features and\nConvolutional Neural Networks (CNN), and absorb merits from both by proposing\nan integrated method called Convolutional Channel Features (CCF). CCF transfers\nlow-level features from pre-trained CNN models to feed the boosting forest\nmodel. With the combination of CNN features and boosting forest, CCF benefits\nfrom the richer capacity in feature representation compared with channel\nfeatures, as well as lower cost in computation and storage compared with\nend-to-end CNN methods. We show that CCF serves as a good way of tailoring\npre-trained CNN models to diverse tasks without fine-tuning the whole network\nto each task by achieving state-of-the-art performances in pedestrian\ndetection, face detection, edge detection and object proposal generation.","url_abs":"http://arxiv.org/abs/1504.07339v3","url_pdf":"http://arxiv.org/pdf/1504.07339v3.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":"convolutional-channel-features","repo_url":"https://github.com/byangderek/CCF","is_official":1,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"none","reach":null}],"tasks":[{"task_slug":"edge-detection","task_name":"Edge Detection"},{"task_slug":"face-detection","task_name":"Face Detection"},{"task_slug":"object-proposal-generation","task_name":"Object Proposal Generation"},{"task_slug":"pedestrian-detection","task_name":"Pedestrian Detection"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"atlas_url":null,"mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}