Abstract:

Due to the exponential growth of video data on the World Wide Web, efficient and fast video retrieval algorithm has attracted a lot of attentions. Because the traditional video image features, such as color histogram and scale invariant feature transform could not obtain promising results on the retrieval speed and detection precision in video copy detection, a video retrieval algorithm using multiple feature fusion was proposed in this paper. Using the temporal and spatial characteristics of the two frames before and after, the time alignment algorithm based on sliding window was applied to reduce the retrieval range and improve retrieval speed. In order to improve the detection precision, this algorithm performed the global feature, the color correlation graph, and the local feature extraction of the SIFT, and then combined the advantages of both global and local features. Experimental results showed that the proposed method could achieve better performance.

Key words: video retrieval, sliding window, multiple feature fusion, color correlogram, temporal and spatial characteristics, key frame