Main Article Content
With advances in multimedia signal processing and abundance of data, multimedia authentication becomes more and more challenging and demanding. Although many prior techniques were put forward to authenticate individual video or audio stream, the cross-layer authentication of audio-visual content as a whole remains unexplored. In this work, an authentication system of the dominant audio-visual content file format (MP4) is presented. Specifically, the perceptual hashes of I-frames from video stream are encrypted and embedded into the audio stream, and the audio hash vector is encrypted and embedded in the synchronization information, i.e., stts box, in the MP4 file. The proposed system is implemented and tested on various videos downloaded from YouTube. Results show that the proposed system can authenticate MP4 files correctly without degrading the video quality while being able to provide indication of tampering.