A blind watermarking algorithm using low density parity check coding for Mpeg-4 videos

A new digital watermarking algorithm for MPEG-4 video based on LDPC coding is proposed. Which embeds into and retrieves the digital watermark information out of the MPEG-4 video using LDPC coding technology and human visual characteristics. The watermark is encoded by LDPC at first and then embedded into the intermediate frequency coefficient of the luminance of the video´s I-VOP. Furthermore, it is retrieved from the watermarked video by the LDPC iterative methods. Making full use of LDPC codes´ error correction features can reduce the bit error rate in transmission process of the watermarking, and improve the anti-attack capability of the watermarking. Experimental results show that this algorithm greatly decrease the bit error rate of the retrieved watermark and is robust to the video attacking such as frame deleting ,frame switching ,frame average, recompression,et al. The watermark can be extracted blindly. As compared without LDPC watermarking, the performance of the algorithm is improved.