An audio-video mixed fingerprint generation algorithm based on key frames

Recent years, digital fingerprint technology is more and more used in multimedia content security protection. But the current digital fingerprint technology still exist many disadvantages, such as complex extraction method, huge fingerprint size and difficulty in real-time monitoring of multimedia contents. This paper proposed a new audio-video mixed fingerprint (AVMF) generation algorithm, which indicates the feature information of audio and video content. Thus, it can authenticate audio and video data simultaneously, and avoid the flaws of content security monitoring by one type of fingerprint. The first work of this paper is the extraction of video key I frames based on the graph theory knowledge - minimum vertex cover set. And then, we present the methods of formation and mix of the audio and video fingerprint, which are based on DCT-coefficient characteristics and XOR respectively. Experiment and simulation results illustrate that the proposed algorithm can simplify the generation process of digital fingerprints and reduce the fingerprint size,, and it have ability of error-detection and error correction. Therefore, this algorithm provides theoretical basis and technical support for fast and real-time detection of multimedia contents security in internet.