Video watermarking algorithm for resisting collusion attacks

A content dependent video watermark algorithm for resisting linear collusion attacks was presented. The image's visual features and DCT coefficients of sub-region were utilized to select the watermark embedding locations. Moreover the mechanism of JND(just-noticeable distortion) was used to adjust the embedding strength of proposed algorithm. These components may be then used as building blocks to construct a self-adaptive content dependent spread spectrum watermark solution. Experimental results have been shown that the performance of the proposed strategy can resist effectively two types of linear collusion attacks while still being robust, stable and imperceptible.