Similarity Measurement and Detection of Video Sequences
Hoi Chu-Hong · 2003
Efficient technique to detect the similar video sequences on the web has become one of the most important and challenging issues in multimedia and database related areas. In this paper, an original two-phase scheme for video similarity detection is proposed. For each video sequence, we extract two kinds of signatures with different granularities: coarse and fine. Coarse signature is based on the Pyramid Density Histogram technique and fine signature is based on the Nearest Feature Trajectory technique. In the first phase, most of unrelated video data are filtered out with respect to the similarity measure of the coarse signature. In the second phase, the query video example is compared with the results of the first phase according to the similarity measure of the fine signature. Different from the conventional nearest neighbor and Hausdorff distance measure methods, our proposed similarity measurement method well incorporates the temporal order of video sequences. Experimental results show that our scheme achieves better quality results than the conventional approaches.