基于H.264/AVC标准的脆弱视频水印算法研究

The fragile video watermarking algorithm based on the H.264/AVC standard

针对H.264/AVC的编码特征,本文提出一种基于H.264/AVC低比特率视频流的脆弱水印算法,该算法在I帧的量化后的4 4DCT块中结合系数值的符号特性与奇偶性,首先用倒数第二位非零系数与水印值一起生成标志信息,然后根据标志信息的值,利用最后一位非零系数的符号特性,修改系数值的奇偶性,从而将水印嵌入到最后一位非零系数中。在解码端根据4 4DCT块中最后两位非零系数的奇偶性关系提取水印,检测时不需要原始视频的参与。该算法具有较小的码率变化,以及较强的脆弱性。

Aiming at new feature of the H.264/AVC video coding standard,base on H.264/AVC low bit—rate video stream a fragile watermark algorithm is proposed.In the algorithm,we uses the symbols and the parity of the quantized DCT coefficient in I frame,depending on the value of the flag information which is generated by the second-lowest non-zero coefficients and the watermark value and the last symbol of a non-zero coefficient,we modify the parity of the last a non-zero coefficients in order to embed into the coefficients.At the decoder,we use the parity relations of the last two non-zero coefficient in DCT blocks to extract the watermark.Not need the original video to participate in the detection.Finally,the algorithm has a smaller rate changes,as well as a strong vulnerability.