摘要
针对现有基于H.264的视频水印算法存在误差漂移的问题,在分析了误差漂移产生原因的基础上,提出了一种抑制误差漂移的鲁棒性视频水印新算法。通过调制H.264编码中Ⅰ帧量化后的4×4 DCT系数来实现水印的嵌入,并对帧内预测量化后满足特定条件的DCT子块的最后一行、列进行补偿和对帧间预测进行补偿,有效地抑制了帧内、帧间误差漂移。另外,该算法结合CAVLC编码来调制量化后的DCT残差系数,使整体码流变动较小,并且算法能够根据不同视频序列自适应地调节嵌入容量及嵌入位置,具有较好的灵活性和鲁棒性。在水印提取过程中,不需要原始视频的参与,即可实现盲提取。仿真实验结果表明,该算法在具有良好不可见性的前提下,对压缩码率影响较小,能够较好地抵抗噪声、帧剪切、重编码等攻击。
Based on the presence of the existing H. 264 video watermarking algorithm error drift problems, the system based on the error drift causes is analyzed, the robustness of the proposed video watermarking algorithm suppress error drift. By modulating the H. 264 coding coefficients in 4 ×4 DCT quantized I frame to the watermark embedding, and specific conditions are met after the last line of the intra prediction quantized DCT subblock column and compensating inter - frame prediction compensation, frame ef- fectively suppressed interframe error drift. In addition, the algorithm combines CAVLC residual coding modulation DCT coefficients after quantization, so that changes in the overall stream is small, and the algorithm can adaptively adjust the embedding capacity and embedding position according to different video sequences, it has good flexibility and Lu stick properties. In the watermark ex- traction process, is not involved in the original video, you can realize blind extraction. Simulation results show that the algorithm has good invisibility under the premise of the compression rate has little effect, better able to resist noise, frame cropping, re - en- coding and other attacks.
出处
《电视技术》
北大核心
2016年第10期15-21,共7页
Video Engineering
基金
国家自然科学基金项目(61363076)
江苏省自然科学基金项目(120142BAB207020)
关键词
视频水印
误差漂移
鲁棒性
DCT变换
帧内预测
video watermarking
error drift
robustness
DCT transform
intra prediction
