Image tamper detection and correction using Merkle tree and remainder value differencing

Abstract

This article proposes an image tamper detection and correction technique. It uses the principle of differencing and addresses the fall off boundary problem (FOBP). It operates on 2 × 2 non-overlapped blocks. From each pixel of a block the decimal value for 4 most significant bits (MSBs), is called as quotient, and the decimal value for 4 least significant bits (LSBs) is called as remainder. The 4 watermark bits (WBs) are computed from 4 quotients as the root of Merkle tree. These 4 WBs are XORed with the 4 bits generated from logistic map sequence to generate the 4 recovery bits (RBs). The 4 WBs and 4 RBs are stored in 4 remainders by LSB alteration and remainder value differencing (RVD). While extracting the watermark, the tampered blocks can be identified and a correction logic can be applied to get the true value of the pixels in the block. The peak signal-to-noise ratio (PSNR) value is 42.02 dB and structural similarity (SSIM) index value is 0.9765. The tampered blocks can be accurately identified and corrected.