Wavelet-based digital watermarking for image authentication

We present a novel watermarking scheme to ensure the credibility of digital images. The proposed technique is able to detect malicious tampering of images even if they have been incidentally distorted by basic image processing operations. Our system is based on the quantization of wavelet packet coefficients and uses characteristics of the human visual system to maximize the embedding weights while keeping good perceptual transparency. We develop an image-dependant method to evaluate, in the discrete wavelet domain, the optimal quantization steps allowing the tamper proofing of still images. The nature of multiresolution discrete wavelet decomposition allows the spatial and frequency localization of image tampering. Experimental results are presented to demonstrate the capacity of our system to detect unauthorized modification of images while staying robust to image compression.