A relevance feedback algorithm for motion data retrieval

作     者：Chen, Song-Le Sun, Zheng-Xing Zhang, Yan Li, Qian 

作者机构：State Key Laboratory for Novel Software Technology Nanjing University NanjingJiangsu210046 China Key Laboratory of Ministry of Education of China for Broadband Wireless Communication and Sensor Network Technology Nanjing University of Posts and Telecommunications NanjingJiangsu210003 China 

出 版 物：《Tien Tzu Hsueh Pao/Acta Electronica Sinica》 (Tien Tzu Hsueh Pao)

年 卷 期：2016年第44卷第4期

页      面：868-872页

核心收录：

学科分类：080804[工学-电力电子与电力传动] 080805[工学-电工理论与新技术] 0808[工学-电气工程] 08[工学] 0811[工学-控制科学与工程] 

基　　金：江苏省科技计划 计算机软件新技术国家重点实验室创新基金重点项目 国家教育部新世纪优秀人才支持计划 国家自然科学基金 国家863计划 

主　　题：运动捕获数据 相关反馈 RankBoost 排序损失 

摘      要：A relevance feedback algorithm based on RankBoost for content-based motion data retrieval (CBMR) is presented and has two characteristics. First, KNN-DTW is employed as the weak ranker for RankBoost ensemble learning. While adapting to variable-length multivariate time series (VLMTS) data, by taking the advantage of the ensemble and efficiency of RankBoost, it can resolve the conflict between the real-time requirement of relevance feedback and the high computational complexity of VLMTS data. Second, minimizing ranking experience loss and generalization loss risk proposed in this paper are used as the learning objective for RankBoost ensemble learning, which can effectively solve the over-fitting problem caused by small-sample training in relevance feedback. Experimental results on CMU action library verify the effectiveness of the proposed algorithm. © 2016, Chinese Institute of Electronics. All right reserved.