基于JSeg和显著性检测的服装图像分割方法  

Garment Image Segmentation Based on JSeg and Salient Region

作　　者：刘正东[1] 

机构地区：[1]北京服装学院计算机信息中心,北京100029

出　　处：《软件导刊》2015年第2期136-139,共4页Software Guide

基　　金：国家科技支撑计划项目(2013BAH41F00;2013BAH41F01);数字与交互媒体北京市重点实验室开放课题(KF2013-08)

摘　　要：针对服装图像分割问题,提出将JSeg算法应用于图像显著性区域的分割算法,利用人脸检测辅助判别背景区域与人体着装区域。通过网络商城实际图像样本实验,验证了方法的有效性,可为服装结构和特征提取,以及可视化检索提供参考借鉴。Internet marketing of apparel products are generally widely adopted model pictures ,because this is the most in‐tuitive form .It is easy to determine the clothing styles by human eyes ,but it is difficult by computer automatic method . For clothing image segmentation problem ,an algorithm applied JSeg algorism and salient region determination is pro‐posed ,which identify background region and dressing region by help of face detection .With actual images from online shopping mall the experiment show s the effectiveness of the method and lays a foundation for further cognize clothing structure and characteristics .

关 键 词：服装 图像分割 JSEG 显著性计算 人脸检测 

分 类 号：TP317.4[自动化与计算机技术—计算机软件与理论]