Motion consideration in H.264/AVC compressed video watermarking

Abstract

Motion and time dependency are inherent in video signals. In video watermarking algorithms, analyzing P frames can reduce the visible artifacts and preserve the video quality. In a watermarked video stream, two visual artifacts may occur considered as spatial noise and temporal flicker. Even by embedding the watermark in I frames, some visible artifacts may occur in P frames. In this paper, statistical analysis of the coefficients in each block is utilized to provide the information to select the best positions for embedding the watermark in order to prevent noticeable spatial noise. On the other hand, motion information as one of the available data in every coding standard is exploited for analyzing the temporal variation of video signals. The inter mode information extracted from P frames are employed to decrease the temporal quality degradation. Compromising between perceptual quality and bit rate increment is another advantage of the proposed method. © 2009 Springer-Verlag Berlin Heidelberg.