Two watermarks are embedded into the original image. One is the authentication water- mark generated by secret key, which is embedded into the sub-LSB (Least Significant Bit) of the original image for tamper localization; the other is the recovery watermark for tamper recovering. The original image is divided into 8×8 blocks and each block is transformed by Discrete Cosine Transform (DCT). For each block, some lower frequency DCT coefficients are chosen to be quantized and binary encoded so as to gain the recovery watermark of each block, and the recovery watermark is embedded into the LSB of another block by chaos encryption and authentication chain technology. After the two watermarks being detected, the location of any minute changes in image can be detected, and the tampered image data can be recovered effectively. In the paper, the number of coefficients and their bit lengths are carefully chosen in order to satisfy with the payload of each block and gain the capability of self-recovering. The proposed algorithm can well resist against possible forged attacks. Experimental results show that the watermark generated by the proposed algorithm is sensitive to tiny changes in images, and it has higher accuracy of tamper localization and good capability of the tamper recovery.
展开▼
机译：Two watermarks are embedded into the original image. One is the authentication watermark generated by secret key, which is embedded into the sub-LSB (Least Significant Bit) of the original image for tamper localization; the other is the recovery watermark for tamper recovering. The original image is divided into 8×8 blocks and each block is transformed by Discrete Cosine Transform(DCT). For each block, some lower frequency DCT coefficients are chosen to be quantized and binary encoded so as to gain the recovery watermark of each block, and the recovery watermark is embedded into the LSB of another block by chaos encryption and authentication chain technology. After the two watermarks being detected, the location of any minute changes in image can be detected, and the tampered image data can be recovered effectively. In the paper, the number of coefficients and their bit lengths are carefully chosen in order to satisfy with the payload of each block and gain the capability of self-recovering. The proposed algorithm can well resist against possible forged attacks. Experimental results show that the watermark generated by the proposed algorithm is sensitive to tiny changes in images, and it has higher accuracy of tamper localization and good capability of the tamper recovery.
展开▼
