人体行为识别中具有判别力的时空特征选择方法

张洪博; 李绍滋; 苏松志

0
0
浏览
下载

摘要
关键词
基金信息
论文图表
同行评议
相关论文
评论

Selecting Effective and Discriminative Spatio-Temporal Interest Points for Recognizing Human Action

首发时间：2012-11-05

ZHANG Hongbo ¹
Zhang Hongbo received the B.S. degree in Computer Science and Technology from Shenyang Normal University, Shenyang, China, in 2008. He is currently working toward Ph.D. degree in artificial intelligent in Cognitive Science Department of Xiamen University, China. During Sept. 2011 and July 2012, he went to Yuan Ze University as an exchange student. His research interests include human action analysis, object detection/recognition, and image/video retrieval.
LI Shaozi ²
LI Shaozi received the B.S. degree from the Computer Science Department, Hunan University in 1983, and the M.S. degree from the Institute of System Engineering, Xi'an Jiaotong University in 1988, and the Ph.D. degree from the College of Computer Science, National University of Defense Technology in 2009. He currently serves as the Professor and Chair of Cognitive Science Department of Xiamen University, the Vice Director of Fujian Key Lab of the Brain-like Intelligence System, and the Vice Director and General Secretary concurrently of the Council of Fujian Artificial Intelligence Society. His research interests cover Artificial Intelligence and Its Applications, Moving Objects Detection and Recognition, Machine Learning, Computer Vision, Natural Language Processing and Multimedia Information Retrieval, Network Multimedia and CSCW Technology and others.
SU Songzhi ³

1、School of Information Science and Technology, Xiamen University, FuJian XiaMen 361005
2、 Fujian Key Laboratory of the Brain-like Intelligent Systems (Xiamen University), Xiamen, China, 361005
3、School of Information Science and Technology, Xiamen University, China, 361005

Abstract：Many successful methods for recognizing human action are spatio-temporal interest point (STIP) based methods. Given a test video sequence, for matching-based method using voting mechanism, each test STIP casts a vote for each action class based on its mutual information with respect to the respective class, which is measured in terms of class likelihood probability. Therefore, two issues should be addressed to improve the accuracy of action recognition. First, effective STIPs in the training set must be selected as references for accurately estimating probability. Second, discriminative STIPs in test set must be selected for voting. This work uses e-nearest neighbors as effective STIPs for estimating the class probability and uses a variance filter for selecting discriminative STIPs. Experimental results verify that the proposed method is more accurate than existing action recognition methods.paper.

keywords： Artificial Intelligence Human action recognition discriminative power class likelihood probability variance filter.

点击查看论文中文信息

人体行为识别中具有判别力的时空特征选择方法

张洪博 ^{1
2}
Zhang Hongbo received the B.S. degree in Computer Science and Technology from Shenyang Normal University, Shenyang, China, in 2008. He is currently working toward Ph.D. degree in artificial intelligent in Cognitive Science Department of Xiamen University, China. During Sept. 2011 and July 2012, he went to Yuan Ze University as an exchange student. His research interests include human action analysis, object detection/recognition, and image/video retrieval.
李绍滋 ^{3
4}
LI Shaozi received the B.S. degree from the Computer Science Department, Hunan University in 1983, and the M.S. degree from the Institute of System Engineering, Xi'an Jiaotong University in 1988, and the Ph.D. degree from the College of Computer Science, National University of Defense Technology in 2009. He currently serves as the Professor and Chair of Cognitive Science Department of Xiamen University, the Vice Director of Fujian Key Lab of the Brain-like Intelligence System, and the Vice Director and General Secretary concurrently of the Council of Fujian Artificial Intelligence Society. His research interests cover Artificial Intelligence and Its Applications, Moving Objects Detection and Recognition, Machine Learning, Computer Vision, Natural Language Processing and Multimedia Information Retrieval, Network Multimedia and CSCW Technology and others.
苏松志 ^{3
5}

1、厦门大学信息科学与技术学院，福建厦门 361005
2、厦门大学福建省仿脑智能系统重点实验室，福建厦门 361005
3、信息科学与技术学院，厦门大学，厦门，361005
4、福建省仿脑智能系统重点实验室，厦门大学，厦门，361005
5、福建省仿脑智能系统重点实验室，厦门大学，厦门，361005

摘要：针对人体行为识别中时空特征的类条件概率估计问题和特征选择问题，本文提出基于改进的类条件概率估计的人体行为识别方法，以及基于互信息方差的特征选择方法。实验结果表明本文提出的方法能有效提高时空特征的类条件概率估计的准确性以及选择判别性强的时空特征，从而提高人体行为识别的准确率。

关键词：人工智能人体行为识别判别力类条件概率方差筛选

基金：

论文图表：

引用

导出参考文献

.txt

.ris

.doc

ZHANG Hongbo,LI Shaozi,SU Songzhi. Selecting Effective and Discriminative Spatio-Temporal Interest Points for Recognizing Human Action[EB/OL]. Beijing:Sciencepaper Online[2012-11-05]. https://www.paper.edu.cn/releasepaper/content/201211-55.

No.****

同行评议

未申请同行评议

全部评论

0/1000

论文编号	201211-55
论文题目	人体行为识别中具有判别力的时空特征选择方法
文献类型
收录期刊	上传封面中文期刊英文期刊期刊名称（中文）期刊名称（英文）年，卷（）上传封面中文专著英文专著书名（中文）书名（英文）出版地出版社出版年上传封面中文译著英文译著书名（中文）书名（英文）出版地出版社出版年上传封面中文论文集英文论文集编者.论文集名称（中文） [c]. 出版地出版社出版年， - 编者.论文集名称（英文） [c]. 出版地出版社出版年，- 上传封面中文文献英文文献期刊名称（中文）期刊名称（英文）日期-- 在线地址http:// 上传封面中文文献英文文献文题（中文）文题（英文）出版地出版社,出版日期-- 上传封面中文文献英文文献文题（中文）文题（英文）出版地出版社,出版日期--
英文作者写法：中外文作者均姓前名后，姓大写，名的第一个字母大写，姓全称写出，名可只写第一个字母，其后不加实心圆点“.”, 作者之间用逗号“，”分隔，最后为实心圆点“.”, 示例1：原姓名写法：Albert Einstein,编入参考文献时写法：Einstein A. 示例2：原姓名写法：李时珍；编入参考文献时写法：LI S Z. 示例3：YELLAND R L,JONES S C,EASTON K S,et al.