摘要
针对当前HEVC视频信息隐藏算法对视频码率和质量影响较大等问题,结合HEVC编码标准中帧内预测模式的特点,提出一种基于帧内预测模式差值的HEVC视频信息隐藏算法。对于连续两个帧内4×4亮度块预测编码过程中具有方向性的33种预测模式,建立隐秘信息与预测模式差值的映射关系,根据拉格朗日率失真模型调制预测模式完成隐秘信息的嵌入。为了保证隐秘信息的嵌入容量,对于连续两个无方向性的planar模式或DC模式,根据隐秘信息修改其值实现隐秘信息的嵌入。信息的提取过程,只需对码流中的预测模式解码即可。实验结果表明,隐秘信息的嵌入对视频序列的影响较小,亮度分量的峰值信噪比(PSNR)值平均下降约0.05dB,对码率影响仅在1.1%以下,隐秘信息嵌入前后的结构相似度(SSIM)值均在0.94以上。因此,本文算法能很好保证视频的主客观质量。
As existing high efficiency video coding （HEVC） video data hiding algorithms have large im- pact on the video quality and the bit rate,a new algorithm based on differences of intra prediction modes combining the prediction modes of HEVC is proposed. For 33 kinds of prediction modes with direction in process of 4 × 4 luminance prediction coding, we establish a mapping relationship between secret bits and differences of two consecutive intra predictions and complete the embedding of secret bits by modulating the prediction according to Lagrange rate distortion model. For two consecutive non-directional prediction modes-planar mode or DC mode, we achieve the embedding of secret bits by modifying the prediction modes directly in order to increase the embedding capacity. The extraction of secret information only re- quires to decode the intra prediction mode from the bit stream. Experimental results show that the aver- age decrease of peak signal to noise ratio （PSNR） is about 0. 05 dB. Increment of bit rate is bounded to 1.1%. The structural similarity （SSIM） is more than 0, 94. The experimental results indicate that the proposed algorithm has little impact on the video quality and bit rate. Key words：
出处
《光电子．激光》
EI
CAS
CSCD
北大核心
2015年第9期1753-1760,共8页
Journal of Optoelectronics·Laser
基金
国家自然科学基金(61170137
61301247
61300055)
浙江省自然科学基金(LY13F020013)
宁波市自然科学基金(2013A610059)
浙江省重中之重学科开放基金(xkx11405)
宁波大学优秀学位论文培育基金(py2013005)资助项目
关键词
信息隐藏
HEVC
预测模式差值
率失真
information hiding
high efficiency video coding （HEVC）
differences of prediction modes
rate distortion
