Geometrically robust image content authentication using Zernike moments

A geometrically robust image content authentication algorithm using modified Zernike moments is proposed in this paper. Firstly the characteristic of image Zernike moments magnitude is quantified and coded, and then embedded in the original image as watermark. When verifying the image, the variance of Zernike moments magnitude and watermark of test image is computed to reconstruct the difference image with the complex angle of test image ZMS. The difference image indicates whether test image has been tampered and the tamper location. This algorithm is robust to geometry distortion of rotation, scale and translation. It can resist the processes such as compression, enhance, filter and noise, and locate the tamper position correctly.