摘要:Since video products compressed by H.264/AVC become increasing popular, issues of authentication for H.264/AVC-based video will be very important. In this paper, a content-based authentication semi-fragile watermarking scheme for H.264/AVC video is proposed. The content-based authentication code is generated using the invariance of the relation between DCT DC coefficients of the two 4×4 sub-blocks. The authentication code is then embedded into the DCT coefficients in diagonal positions in I frames. Spatial tampering can be located by comparing the extracted watermarking and the content-based authentication code. The experimental results show that the proposed scheme can distinguish the malicious spatial tampering from the common signal processing. The tampered areas can also be located.