| 基于Fisher鉴别字典学习的人体行为识别 |
| |
| 引用本文： | 季冲,王胜,陆建峰.基于Fisher鉴别字典学习的人体行为识别[J].计算机科学,2017,44(7):270-274. |
| |
| 作者姓名： | 季冲 王胜 陆建峰 |
| |
| 作者单位： | 南京理工大学计算机科学与工程学院 南京210094,南京理工大学计算机科学与工程学院 南京210094,南京理工大学计算机科学与工程学院 南京210094 |
| |
| 基金项目： | 本文受江苏省科技支撑计划——社会发展项目:面向治安防控的监控视频目标检索关键技术研究(BE2014714),国家科技重大专项:无人装备智能控制支撑软件系统(2015ZX01041101)资助 |
| |
| 摘 要： | 人体行为识别是计算机视觉中的一个重要研究领域,具有广阔的应用前景。研究了基于Fisher鉴别的字典学习方法在人体行为识别上的应用。首先对人体行为的视频序列提取了局部时空特征,并通过随机投影法降维；然后把降维后的特征作为待分类的信号进行Fisher鉴别字典学习,从而增强字典和编码系数的鉴别能力；最后同时利用重构误差和稀疏表示系数进行分类。实验结果验证了所提方法在人体行为识别上的有效性与鲁棒性。
|
| 关 键 词： | 稀疏表示 人体行为识别 运动特征 Fisher鉴别准则 |
| 收稿时间： | 2016/7/12 0:00:00 |
| 修稿时间： | 2016/10/28 0:00:00 |
Human Action Recognition Based on Fisher Discrimination Dictionary Learning |
| |
| JI Chong,WANG Sheng and LU Jian-feng.Human Action Recognition Based on Fisher Discrimination Dictionary Learning[J].Computer Science,2017,44(7):270-274. |
| |
| Authors: | JI Chong WANG Sheng and LU Jian-feng |
| |
| Affiliation: | Department of Computer Science and Engineering,Nanjing University of Science and Technology,Nanjing 210094,China,Department of Computer Science and Engineering,Nanjing University of Science and Technology,Nanjing 210094,China and Department of Computer Science and Engineering,Nanjing University of Science and Technology,Nanjing 210094,China |
| |
| Abstract: | Human action recognition is a hot computer vision research field,and has broad application prospects.This paper explored the application of Fisher discrimination based dictionary learning on human action recognition.First,local spatial-temporal features are extracted from the video sequences and random projection is used to reduce dimension.Then,Fisher discrimination based dictionary learning is performed on the reduced motion features.Last,a new classification scheme is proposed using both reconstruction error and representation coefficients.Experimental results confirm the efficiency and robustness of the proposed scheme. |
| |
| Keywords: | Sparse representation Human action recognition Motion features Fisher discrimination criterion |
|
|
点击此处可从《计算机科学》下载全文 |
|
