[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"detail-sidebar-cat-0-en-105":3,"doc-seo-280413-105":59,"doc-detail-280413-en":131},{"code":4,"msg":5,"data":6},0,"success",[7,13,18,23,28,33,38,43,48,51,55],{"id":8,"doc_module":4,"doc_module_name":9,"category_name":10,"show_sort_weight":11,"slug":12},1,"Document","Story & Novel",90,"story-novel",{"id":14,"doc_module":4,"doc_module_name":9,"category_name":15,"show_sort_weight":16,"slug":17},2,"Literature",80,"literature",{"id":19,"doc_module":4,"doc_module_name":9,"category_name":20,"show_sort_weight":21,"slug":22},4,"Exam",70,"exam",{"id":24,"doc_module":4,"doc_module_name":9,"category_name":25,"show_sort_weight":26,"slug":27},5,"Comic",60,"comic",{"id":29,"doc_module":4,"doc_module_name":9,"category_name":30,"show_sort_weight":31,"slug":32},6,"Technology",50,"technology",{"id":34,"doc_module":4,"doc_module_name":9,"category_name":35,"show_sort_weight":36,"slug":37},7,"Healthcare",40,"healthcare",{"id":39,"doc_module":4,"doc_module_name":9,"category_name":40,"show_sort_weight":41,"slug":42},8,"Research & Report",30,"research-report",{"id":44,"doc_module":4,"doc_module_name":9,"category_name":45,"show_sort_weight":46,"slug":47},9,"Religion & Spirituality",20,"religion-spirituality",{"id":46,"doc_module":4,"doc_module_name":9,"category_name":49,"show_sort_weight":46,"slug":50},"World Cup","world-cup",{"id":52,"doc_module":4,"doc_module_name":9,"category_name":53,"show_sort_weight":52,"slug":54},10,"Lifestyle","lifestyle",{"id":56,"doc_module":4,"doc_module_name":9,"category_name":57,"show_sort_weight":24,"slug":58},19,"General","general",{"code":4,"msg":60,"data":61},"ok",{"site_id":62,"language":63,"slug":64,"title":65,"keywords":66,"description":67,"schema_data":68,"social_meta":124,"head_meta":126,"extra_data":128,"updated_unix":130},105,"en","self-supervised-auxiliary-learning-for-texture-and-model-based-hybrid-robust-and-fair-featuring-in-face-analysis-arxiv-240919582v2","Self-supervised Auxiliary Learning for Texture and Model-based Hybrid Robust and Fair Featuring in Face Analysis - arXiv 2409.19582v2","","Self-supervised learning (SSL) is investigated as an auxiliary task to integrate texture-based local descriptors into feature modelling for efficient and robust face analysis. A mask auto-encoder (MAE) auxiliary objective is used to reconstruct texture features such as local patterns alongside the primary task. Experiments cover face attribute and face-based emotion analysis as well as deepfake detection, showing improved feature representation that supports fair, bias-less face analysis.",{"@graph":69,"@context":123},[70,84,106],{"@type":71,"itemListElement":72},"BreadcrumbList",[73,77,79,82],{"item":74,"name":75,"@type":76,"position":8},"https://docshare.wps.com","Home","ListItem",{"item":78,"name":9,"@type":76,"position":14},"https://docshare.wps.com/document/",{"item":80,"name":40,"@type":76,"position":81},"https://docshare.wps.com/document/research-report/",3,{"item":83,"name":65,"@type":76,"position":19},"https://docshare.wps.com/document/self-supervised-auxiliary-learning-for-texture-and-model-based-hybrid-robust-and-fair-featuring-in-face-analysis-arxiv-240919582v2/280413/",{"url":83,"name":65,"@type":85,"image":86,"author":91,"headline":65,"publisher":94,"fileFormat":97,"inLanguage":63,"description":67,"dateModified":98,"datePublished":99,"encodingFormat":97,"isAccessibleForFree":100,"interactionStatistic":101},"DigitalDocument",{"url":87,"@type":88,"width":89,"height":90},"https://docshare.wps.com/thumbnails/self-supervised-auxiliary-learning-for-texture-and-model-based-hybrid-robust-and-fair-featuring-in-face-analysis-arxiv-240919582v2/280413.png","ImageObject",300,407,{"name":92,"@type":93},"4398046744996","Person",{"url":74,"name":95,"@type":96},"DocShare","Organization","application/pdf","2026-10-22","2026-09-16",true,{"@type":102,"interactionType":103,"userInteractionCount":105},"InteractionCounter",{"@type":104},"ViewAction",13,{"@type":107,"mainEntity":108},"FAQPage",[109,115,119],{"name":110,"@type":111,"acceptedAnswer":112},"What auxiliary task is proposed for face analysis?","Question",{"text":113,"@type":114},"The method uses the SSL task of mask auto-encoder (MAE) to reconstruct texture features such as local patterns alongside the primary task.","Answer",{"name":116,"@type":111,"acceptedAnswer":117},"Which face analysis paradigms are evaluated in the experiments?",{"text":118,"@type":114},"Experiments evaluate face attribute analysis, face-based emotion analysis, and deepfake detection.",{"name":120,"@type":111,"acceptedAnswer":121},"Why does the proposed approach aim for fair and unbiased analysis?",{"text":122,"@type":114},"By blending texture-based local descriptors into feature modelling using a self-supervised auxiliary objective, the model produces better representations that support fair and bias-less face analysis.","https://schema.org",{"og:url":83,"og:type":125,"og:title":65,"og:site_name":95,"og:description":67},"article",{"robots":127,"canonical":83},"index,follow",{"doc_id":129,"site_id":62},280413,1789547949,{"code":4,"msg":5,"data":132},{"doc_id":129,"user_id":133,"nickname":92,"user_avatar":66,"doc_module":4,"category_id":39,"category_name":40,"doc_title":65,"doc_description":67,"doc_content":134,"file_id":135,"file_url":136,"file_type":137,"file_size":138,"view_count":105,"is_deleted":4,"is_public":8,"is_downloadable":8,"audit_status":8,"page_count":139,"language":140,"language_code":63,"site_id":62,"html_lang":63,"table_of_contents":141,"faqs":142,"seo_title":143,"seo_description":67,"update_tm":130,"read_time":144,"preview_page_count":4},4398046744996,"arXiv :2409 . 19582v2 [ cs .CV] 5 Nov 2024  \nSelf-supervised Auxiliary Learning for Texture and Model-based Hybrid Robust and Fair Featuring in Face Analysis  \nShukesh Reddy 1 , Nishit Poddar 1 , Srijan Das2 , and Abhijit Das 1  \n1 Machine Intelligence Group, Department of CSIS, Birla Institute of Technology and Science, Pilani, Hyderabad Campus, India  \n2 University of North Carolina at Charlotte, United States [abhijit.das@hyderabad.bits-pilani.ac.in](abhijit.das@hyderabad.bits-pilani.ac.in)  \nAbstract. In this work, we explore Self-supervised Learning (SSL) asan auxiliary task to blend the texture-based local descriptors into feature modelling for efficient face analysis. Combining a primary task anda self-supervised auxiliary task is beneficial for robust representation.  \nTherefore, we used the SSL task of mask auto-encoder (MAE) as an auxiliary task to reconstruct texture features such as local patterns along with the primary task for robust and unbiased face analysis. We experimented with our hypothesis on three major paradigms of face analysis:  \nface attribute and face-based emotion analysis, and deepfake detection.  \nOur experiment results exhibit that better feature representation can begleaned from our proposed model for fair and bias-less face analysis.  \nKeywords: Self-supervised Auxiliary Task, Texture analysis, Model  \nbased Featuring, Local pattern feature, Local Directional Pattern  \n1 Introduction  \nAchieving accuracy in face analysis is heavily dependent on using efficient face description methods. The research of facial representation techniques has been comprehensive, encompassing principal component analysis (PCA)[76], linear discriminant analysis (LDA)[21], local binary patterns (LBP)[2], and local directional patterns (LDP)[33] . Such holistic approaches like PCA, LDA, focus on deriving global information. On the other hand, local descriptor techniques like LBP and LDP have gained popularity because of their superior capability to accurately handle certain local features. Such techniques may face challenges when dealing with complex changes such as facial expressions, posture, and illumination.  \nIn the past few years, Deep learning models such as AlexNet[36], GoogLeNet[72], and ResNet[31] have revolutionized the field of facial recognition. They have produced limited results in major competitions such as the ILSVRC. However, the application of these models in real-world, data-rich environments is difficult to achieve. This is sometimes hampered by the need for large amounts of training data and fast processing resources. FaceNet[68] has solved this problem by implementing an optimal algorithm that uses deep variable networks to generate  \n2 S. Reddy et al.  \nfacial input with a consistent learning curve. Local features for CNN have been explored in a few works [35] and also for face analysis [64] . Further in the recent past, the Vision Transformer (ViT) architecture-based featuring has been explored for faces[18] . This work also utilizes SSL as an auxiliary task to learn from limited. However, a more robust path for bleeding local and model features for face analysis is not explored in the literature. This motivates us to propose a hybrid approach that combines the advantages of texture-based and modelbased features with SSL for robust and fair face analysis.  \nThe specific contribution of this work is as follows:  \n1. We proposed a Self-supervised Learning (SSL) based auxiliary task to blend the texture-based local descriptors into feature modelling for efficient face analysis.  \n2. We made a deep analysis of the different ways to combine the texture-based local descriptors into feature modelling for efficient and unbiased face analysis.  \n3. A detailed benchmark experiments on different types of task-on-face analysis with different ways to combine the texture-based local descriptors into feature modelling.  \n2 Related Work  \nVision Transformers-Recent advances in Vision Transformers (ViTs) have le","cbCaid9whHhrcSTy","https://ap.wps.com/l/cbCaid9whHhrcSTy","pdf",1260322,15,"English","# 1 Introduction\n## Face representation methods\n## Deep learning and transformer-based featuring\n# 2 Related Work\n## Vision Transformers\n## Self-supervised Learning","[{\"question\":\"What auxiliary task is proposed for face analysis?\",\"answer\":\"The method uses the SSL task of mask auto-encoder (MAE) to reconstruct texture features such as local patterns alongside the primary task.\"},{\"question\":\"Which face analysis paradigms are evaluated in the experiments?\",\"answer\":\"Experiments evaluate face attribute analysis, face-based emotion analysis, and deepfake detection.\"},{\"question\":\"Why does the proposed approach aim for fair and unbiased analysis?\",\"answer\":\"By blending texture-based local descriptors into feature modelling using a self-supervised auxiliary objective, the model produces better representations that support fair and bias-less face analysis.\"}]","Self-supervised Auxiliary Learning for Texture and Model-based Hybrid Robust and Fair Featuring in Face Analysis - arXiv 2409.19582v2 | PDF",38]