{"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/panda-pose-aligned-networks-for-deep","title":"PANDA: Pose Aligned Networks for Deep Attribute Modeling","arxiv_id":"1311.5591","date":"2013-11-21","proceeding":"CVPR 2014 6","authors":["Ning Zhang","Manohar Paluri","Marc'Aurelio Ranzato","Trevor Darrell","Lubomir Bourdev"],"abstract":"We propose a method for inferring human attributes (such as gender, hair\nstyle, clothes style, expression, action) from images of people under large\nvariation of viewpoint, pose, appearance, articulation and occlusion.\nConvolutional Neural Nets (CNN) have been shown to perform very well on large\nscale object recognition problems. In the context of attribute classification,\nhowever, the signal is often subtle and it may cover only a small part of the\nimage, while the image is dominated by the effects of pose and viewpoint.\nDiscounting for pose variation would require training on very large labeled\ndatasets which are not presently available. Part-based models, such as poselets\nand DPM have been shown to perform well for this problem but they are limited\nby shallow low-level features. We propose a new method which combines\npart-based models and deep learning by training pose-normalized CNNs. We show\nsubstantial improvement vs. state-of-the-art methods on challenging attribute\nclassification tasks in unconstrained settings. Experiments confirm that our\nmethod outperforms both the best part-based methods on this problem and\nconventional CNNs trained on the full bounding box of the person.","url_abs":"http://arxiv.org/abs/1311.5591v2","url_pdf":"http://arxiv.org/pdf/1311.5591v2.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":"panda-pose-aligned-networks-for-deep","repo_url":"https://github.com/FanjieLUO/matlab","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"tf","reach":null}],"tasks":[{"task_slug":"attribute","task_name":"Attribute"},{"task_slug":"facial-attribute-classification","task_name":"Facial Attribute Classification"},{"task_slug":"classification","task_name":"General Classification"},{"task_slug":"object-recognition","task_name":"Object Recognition"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[{"leaderboard":"/sota/facial-attribute-classification-on-lfwa","task":"Facial Attribute Classification","dataset":"LFWA","model":"PANDA","rank_in_archive_order":7,"of":7,"metrics":{"Error Rate":"18.97"},"uses_additional_data":false}],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=1311.5591","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}