A Novel Video Content Authentication Algorithm Combined Semi-fragile Watermarking with Compressive Sensing

For the shortcomings of current video watermarking algorithms on tamper detection, combined semi-fragile watermarking with Compressive Sensing theory, the paper proposes a novel video content authentication algorithm for MPEG-2. The image features of I-frame are extracted by Compressed Sensing algorithm, and semi-fragile content authentication watermarking are generated, in which watermarking bits are embedded into the low-frequency DCT coefficients of I-frame. Simulation results show that the algorithm has better detection ability and detection accuracy than invertible semi-fragile watermarking algorithm distinguishing MPEG-2 compression from malicious manipulation the algorithm. For video inner frames tampering, accuracy of the algorithm can reach to the sub-block of image.