Robustness of Input Features from Noisy Silhouettes in Human Pose Estimation

摘要：

　　Silhouettes are frequently extracted and described to compose inputs for learning methods in solving human pose estimation problem.Although silhouettes extracted from background subtraction methods are usually noisy,the effect of noisy inputs to pose estimation accuracies is seldom studied.In this paper,we explore this problem.First,We compare performances of several image features widely used for human pose estimation and explore their performances against each other and select one with best performance.Second,iterative closest point algorithm is introduced for a new quantitative measurement of noisy inputs.The proposed measurement is able to automatically discard noise,like camouflage from the background or shadows.With the proposed measurement,we split inputs into different noise levels and assess their pose estimation accuracies.Furthermore,we explore performances of silhouette samples of different noise levels and compare with the selected feature on a public dataset: HumanEva dataset.

关键词： Human pose estimation silhouette iterative closest point

作者: Wenjuan Gong Preben Fihl Jordi Gonzàlez Thomas B. Moueslund Weishan Zhang Zhen Li Yan Ren

作者单位: China University of Petroleum,Qingdao,China Faculty of Engineering and Science,Aalborg University,Aalborg,Denmark Computer Vision Center & Universitat Autònoma de Barcelona,Spain Sichuan Water Conservancy Vocational College,Chongzhou,Sichuan,China

会议类型: 国际会议

会议名称: 2014International Conference on Identification,Information and Knowledge in The Internet of Things(2014信息与知识物联网国际会议)

会议地点: 北京

会议语种:英文

页码: 126-131

在线出版日期: 2014-10-17（万方平台首次上网日期，不代表论文的发表时间）