Compressed video copy detection based on texture analysis

This paper presents a new copy detection method for compressed videos, based on texture information in the compressed domains. We firstly obtain texture information from discrete cosine transform (DCT) alternating current (AC) coefficients of I-frames, then extract video signatures from the texture information, and finally measure the similarity of videos by the video signatures. The experimental results show that the presented method outperforms the previous ones in accuracy and efficiency.