Reducing input data via image categorization to improve the speed of copyright content management systems

• DOI: 10.1007/s10043-015-0034-4
• Publication Date: 2015-03-06T23:54:51Z
• Authors: Kanami Ikeda; Eriko Watanabe
• Keywords: 0103 physical sciences; 01 natural sciences

An optical correlator has the advantage of high data transfer speed and parallel operation. However, in copyright content management systems (CCMSs), the numerous video files that need to be downloaded from the Internet and input to the optical correlator constitute a bottleneck. This paper proposes an image categorization method for CCMSs that uses the difference in the color features between animation and live-action images to remove this bottleneck and increase the speed of CCMSs. The results of experiments conducted indicate that the proposed method achieves a live-action video true rejection rate of 86.7 % and an animation video false rejection rate of 13.3 %. This indicates that the proposed method can improve the overall speed of a CCMS more than twice the original speed.