中文摘要：

提出了一种基于帧内预测模式调制的H.264/AVC视频信息改进隐藏算法。选择参考像素亮度值接近的帧内4×4块,通过对预测模式值进行调制以实现隐秘信息的嵌入,而预测模式调制过程是将最优预测模式变为满足嵌入条件的具有率失真开销最小的预测模式,以尽量减小预测模式的改变对视频客观质量的影响。信息的提取过程不需要原始载体视频,也不需要完全解码,只需对视频流中的预测模式进行解码即可。实验结果表明,该算法能很好地保持视频的感知质量,并且对视频流的比特率影响很小。

英文摘要：

An information hiding algorithm with prediction mode modulating for H.264/AVC is proposed. According to the luminance values of the reference pixels, the luminance blocks for information embedding are selected. The information is embedded by modulating the prediction modes of 4×4 luminance blocks. If the best mode does not match the information bit, the prediction modes can be modulated by replacing the best mode with the substitute mode. The substitute mode is the one with the least Lagrangian cost among those having different parity with the best mode. The extraction of information can be performed directly from the encoded stream without resorting to the original video, and merely requires decoding the intra prediction mode from bit stream rather than decoding the whole video. Experimental results show that the proposed hiding scheme can effectively embed information with little bit rate increase and almost no quality degradation.