6 April 2012 Robust video copy detection approach based on local tangent space alignment
Author Affiliations +
Abstract
We propose a robust content-based video copy detection approach based on local tangent space alignment (LTSA), which is an efficient dimensionality reduction algorithm. The idea is motivated by the fact that the content of video becomes richer and the dimension of content becomes higher. It does not give natural tools for video analysis and understanding because of the high dimensionality. The proposed approach reduces the dimensionality of video content using LTSA, and then generates video fingerprints in low dimensional space for video copy detection. Furthermore, a dynamic sliding window is applied to fingerprint matching. Experimental results show that the video copy detection approach has good robustness and discrimination.
© 2012 Society of Photo-Optical Instrumentation Engineers (SPIE)
Xiushan Nie, Qianping Qiao, "Robust video copy detection approach based on local tangent space alignment," Optical Engineering 51(4), 047001 (6 April 2012). https://doi.org/10.1117/1.OE.51.4.047001 . Submission:
JOURNAL ARTICLE
9 PAGES


SHARE
Back to Top