An Efficient Mathematical Method for Recovery of Tampered Image Content

To protect the integrity and reliability of images, we propose a watermarking-based approach for identifying whether the image has been tampered with during transmission and recovering the tampered content to its original form. The tampered pixel values are treated as unknowns in our proposed method and are calculated by solving a system of equations using Gauss-Jordan elimination. The recovery data is obtained by encoding the MSB matrix with a secret key and then embedded as part of the watermark for content self-recovery. The remainder of the watermark is authentication information for localizing the forged region, which is produced by MSB and recovery information of each block. Compared with existing works, the experimental results with our method show superior recovery capability.