Watermark-Based Image Authentication with Coefficient Value Differencing and Histogram Shifting

Image-based watermarking has become essential in the digital authentication system. This work uses coefficient value differencing and histogram shifting techniques for watermark embedding on the coefficients of integer wavelet transform to take advantages of both spatial and transform techniques. After some necessary preprocessing, the cover image is applied 3/1 integer wavelet transform (IWT) to get four sub-bands. Thereafter, SHA256 algorithm is executed on the watermark image to generate hash code for authentication. The watermark and the hash code are embedded in the HH and HH of the LH sub-band, respectively, using coefficient value differencing and histogram shifting techniques. In the authentication phase, the stego-image is processed to extract the watermark and SHA256 code, followed by computation of SHA256 code from extracted watermark. These two SHA256 codes are compared to establish authentication. The experimental results on standard dataset show average “mean square error” (MSE), “peak signal to noise ratio” (PSNR), and “structural similarity index measure” (SSIM) between the cover and stego-image to be around 0.10, 60 dB, and 0.999, respectively. The robustness of the system is verified with well know active attacks by computing normalized cross-correlation (NCC).