Research Article
Enhanced Video Indexing and Retrieval Based on Face Recognition through Combined Detection and Fast LDA
@INPROCEEDINGS{10.1007/978-3-642-35615-5_56, author={Loganathan D. and Jamal J. and Nijanthan P. and Kalichy V.}, title={Enhanced Video Indexing and Retrieval Based on Face Recognition through Combined Detection and Fast LDA}, proceedings={Third International conference on advances in communication, network and computing}, proceedings_a={CNC}, year={2012}, month={12}, keywords={Video Indexing and Retrieval Face Detection Face Recognition Fast LDA Wavelet Transformation}, doi={10.1007/978-3-642-35615-5_56} }- Loganathan D.
Jamal J.
Nijanthan P.
Kalichy V.
Year: 2012
Enhanced Video Indexing and Retrieval Based on Face Recognition through Combined Detection and Fast LDA
CNC
Springer
DOI: 10.1007/978-3-642-35615-5_56
Abstract
The content based indexing and retrieval of videos plays a key role in helping the Internet today to move towards semantic web. The exponential growth of multimedia data has increased the demand for video search based on the query image rather than the traditional text annotation. The best possible method to index most videos is by the people featured in the video. The paper proposes combined face detection approach with high detection efficiency and low computational complexity. The fast LDA method proposed performs wavelet decomposition as a pre-processing stage over the face image. The preprocessing stage introduced reduces the retrieval time by a factor of 1/4 where n is the level of decomposition as well as improving the face recognition rate. Experimental results demonstrate the effectiveness of the proposed method reducing the retrieval time by 64 times over the direct LDA implementation.