{"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/improving-facial-attribute-prediction-using","title":"Improving Facial Attribute Prediction using Semantic Segmentation","arxiv_id":"1704.08740","date":"2017-04-27","proceeding":"CVPR 2017 7","authors":["Mahdi M. Kalayeh","Boqing Gong","Mubarak Shah"],"abstract":"Attributes are semantically meaningful characteristics whose applicability\nwidely crosses category boundaries. They are particularly important in\ndescribing and recognizing concepts where no explicit training example is\ngiven, \\textit{e.g., zero-shot learning}. Additionally, since attributes are\nhuman describable, they can be used for efficient human-computer interaction.\nIn this paper, we propose to employ semantic segmentation to improve facial\nattribute prediction. The core idea lies in the fact that many facial\nattributes describe local properties. In other words, the probability of an\nattribute to appear in a face image is far from being uniform in the spatial\ndomain. We build our facial attribute prediction model jointly with a deep\nsemantic segmentation network. This harnesses the localization cues learned by\nthe semantic segmentation to guide the attention of the attribute prediction to\nthe regions where different attributes naturally show up. As a result of this\napproach, in addition to recognition, we are able to localize the attributes,\ndespite merely having access to image level labels (weak supervision) during\ntraining. We evaluate our proposed method on CelebA and LFWA datasets and\nachieve superior results to the prior arts. Furthermore, we show that in the\nreverse problem, semantic face parsing improves when facial attributes are\navailable. That reaffirms the need to jointly model these two interconnected\ntasks.","url_abs":"http://arxiv.org/abs/1704.08740v1","url_pdf":"http://arxiv.org/pdf/1704.08740v1.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":[],"tasks":[{"task_slug":"attribute","task_name":"Attribute"},{"task_slug":"face-parsing","task_name":"Face Parsing"},{"task_slug":"facial-attribute-classification","task_name":"Facial Attribute Classification"},{"task_slug":"prediction","task_name":"Prediction"},{"task_slug":"segmentation","task_name":"Segmentation"},{"task_slug":"semantic-segmentation","task_name":"Semantic Segmentation"},{"task_slug":"zero-shot-learning","task_name":"Zero-Shot Learning"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[{"leaderboard":"/sota/facial-attribute-classification-on-lfwa","task":"Facial Attribute Classification","dataset":"LFWA","model":"SSP + SSG","rank_in_archive_order":2,"of":7,"metrics":{"Error Rate":"12.87"},"uses_additional_data":false}],"syntology":{"syntology_url":null,"atlas_url":null,"mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}