典型文献
Pedestrian Attribute Recognition in Video Surveillance Scenarios Based on View-attribute Attention Localization
文献摘要：
Pedestrian attribute recognition in surveillance scenarios is still a challenging task due to the inaccurate localization of spe-cific attributes.In this paper,we propose a novel view-attribute localization method based on attention(VALA),which utilizes view in-formation to guide the recognition process to focus on specific attributes and attention mechanism to localize specific attribute-corres-ponding areas.Concretely,view information is leveraged by the view prediction branch to generate four view weights that represent the confidences for attributes from different views.View weights are then delivered back to compose specific view-attributes,which will par-ticipate and supervise deep feature extraction.In order to explore the spatial location of a view-attribute,regional attention is intro-duced to aggregate spatial information and encode inter-channel dependencies of the view feature.Subsequently,a fine attentive attrib-ute-specific region is localized,and regional weights for the view-attribute from different spatial locations are gained by the regional at-tention.The final view-attribute recognition outcome is obtained by combining the view weights with the regional weights.Experi-ments on three wide datasets(richly annotated pedestrian(RAP),annotated pedestrian v2(RAPv2),and PA-100K)demonstrate the effectiveness of our approach compared with state-of-the-art methods.
文献关键词：
中图分类号：
作者姓名：
Wei-Chen Chen;Xin-Yi Yu;Lin-Lin Ou
作者机构：
College of Information Engineering,Zhejiang University of Technology,Hangzhou 310023,China
文献出处：
引用格式：
[1]Wei-Chen Chen;Xin-Yi Yu;Lin-Lin Ou-.Pedestrian Attribute Recognition in Video Surveillance Scenarios Based on View-attribute Attention Localization)[J].机器智能研究（英文）,2022(02):153-168
A类：
VALA,confidences,ticipate,ute,richly,RAPv2
B类：
Pedestrian,Attribute,Recognition,Video,Surveillance,Scenarios,Based,View,Attention,Localization,recognition,surveillance,scenarios,still,challenging,task,due,inaccurate,localization,attributes,In,this,paper,propose,novel,attention,which,utilizes,guide,process,focus,specific,mechanism,corres,ponding,areas,Concretely,information,leveraged,by,prediction,branch,generate,four,weights,that,represent,from,different,views,then,delivered,back,compose,will,supervise,deep,feature,extraction,order,explore,spatial,regional,intro,duced,aggregate,encode,inter,channel,dependencies,Subsequently,fine,attentive,localized,locations,gained,final,outcome,obtained,combining,Experi,ments,three,wide,datasets,annotated,pedestrian,PA,100K,demonstrate,effectiveness,approach,compared,state,art,methods
AB值：
0.496509
相似文献
机标中图分类号，由域田数据科技根据网络公开资料自动分析生成，仅供学习研究参考。
