能量差值信息隐藏(differential energy watermarking,DEW)算法通过有选择地舍弃视频图像中某些区域的部分高频DCT系数,形成能量差异而达到信息嵌入的目的.为此,利用视频序列时间域的强相关性建立针对性统计估计模型,侦测DEW算法所造成的载体信息统计特性变化,并结合K-S统计假设检验方法设计一种视频信息隐藏分析算法.结果表明,此算法能够有效地检测压缩视频码流中是否有DEW算法嵌入的特定信息.
Differential energy watermarking (DEW) steganography algorithm embeds bits in compressed videos by selectively discarding high frequenxy DCT coefficients in certain image regions. Based on analyzing the altered statistical characteristics introduced by DEW embedding process, a video steganalysis technique is proposed, which exploits temporal statistics of video sequence to estimate a statistical model and utilizes K-S hypothesis test to detect the presence of the hidden message. The simulation results show that the proposed steganalysis method can effectively detect whether specific message embedded by DEW algorithm is present in compressed video sequence.