2011 | OriginalPaper | Chapter
A New Video Feature Extraction Method Based on Local Class Information Preserving
Authors : Yongliang Xiao, Shaoping Zhu, Weizhong Luo, Xiangbao Li, Wenbin Liu, Gelan Yang
Published in: Informatics in Control, Automation and Robotics
Publisher: Springer Berlin Heidelberg
Activate our intelligent search to find suitable subject content or patents.
Select sections of text to find matching patents with Artificial Intelligence. powered by
Select sections of text to find additional relevant content using AI-assisted search. powered by
Video feature extraction is the first step of video shot boundary detection. In this paper, a more useful and discriminating video feature extraction method based on local class information preserving is proposed. Maximum margin criterion is a very famous feature extraction method, which seeks to preserve global structure of samples, and can resolve small sample size problem. But this method ignores the local structure information of samples. To address the issue, we develop a new method namely local class information preserving (LCIP). We redefine the local between-class scatter matrix and within-class scatter matrix with the local structure information of each sample. Experimental results show the effectiveness of the proposed method.