Transferring pose and augmenting background for deep humanimage parsing and its applications

作     者：Takazumi Kikuchi Yuki Endo Yoshihiro Kanamori Taisuke Hashimoto Jun Mitani 

作者机构：University of Tsukuba1-1-1 TennohdaiTsukuba CityIbarakiJapan 

出 版 物：《Computational Visual Media》 (计算可视媒体（英文版）)

年 卷 期：2018年第4卷第1期

页      面：43-54页

核心收录：

学科分类：12[管理学] 1201[管理学-管理科学与工程(可授管理学、工学学位)] 08[工学] 0835[工学-软件工程] 0811[工学-控制科学与工程] 0812[工学-计算机科学与技术（可授工学、理学学位）] 

基　　金：Japan Society for the Promotion of Science  JSPS 

主　　题：image segmentation semantic segmentation human-image parsing deep convolutional neural network 

摘      要：Parsing of human images is a fundamental task for determining semantic parts such as the face,arms, and legs, as well as a hat or a dress. Recent deep-learning-based methods have achieved significant improvements, but collecting training datasets with pixel-wise annotations is labor-intensive. In this paper,we propose two solutions to cope with limited ***, to handle various poses, we incorporate a pose estimation network into an end-to-end humanimage parsing network, in order to transfer common features across the domains. The pose estimation network can be trained using rich datasets and can feed valuable features to the human-image parsing network. Secondly, to handle complicated backgrounds,we increase the variation in image backgrounds automatically by replacing the original backgrounds of human images with others obtained from large-scale scenery image datasets. Individually, each solution is versatile and beneficial to human-image parsing, while their combination yields further improvement. We demonstrate the effectiveness of our approach through comparisons and various applications such as garment recoloring, garment texture transfer, and visualization for fashion analysis.