A Video Copyright Registration And Identification System Based On Copy Detection

With the rapid development of internet and video processing technology, the spread of digital video is becoming more convenient. However, the technology and low of video copyright protection lag in China, and the piracy becomes increasingly serious. Therefore, how to certificate the copyright ownership of digital videos has become very important.Related technologies involved in the process of video copyright registration are analyzed, and combining with the copy detection technology and related technologies of public-key authentication, a system of video copyright registration and identification is designed and implemented. First, by using feature extraction technology, the SIFT feature information is extracted from the video used as a copyright notice, which could prevent watermark attacks and against various geometric attacks and signal processing. Second, video copyright is detected by using copy detection technology, which has strong robustness and can prevent repeat registration and copy registration. At last, by using the RSA digital signature technology and combining SHA-1 hash algorithm and RSA encryption algorithm, the video feature information is registered by signature, to ensure the authority and unforgeability of the copyright. The copyright identification online and offline are implemented by developing relatively complete protocols and procedures. And the copyright certificate of registered video is issued, by which the copyright owner can determine whether the owners of the video and certificate are the same, and implement the identification offline in the local.System consists of copyright server, client and copyright management system. The server and client are based of C/S architecture, in which the services of copyright registration and identification are provided. The copyright management system is based of B/S architecture, and in which the services of copyright information inquiry, copyright certificate download and copyright transfer for users. The experimental results show that the accuracy of registration and identification is very high. And the original videos are successfully registered, but the copy videos almost can't, so system can well protect the copyright of videos.