Block Truncation Coding based effective watermarking scheme for image authentication with recovery capability

Abstract

Due to the advancement of photo-editing software, powerful computers and high resolution capturing devices, it has become tough to prevent the digital image from tampering. So, in these days just by looking a digital image we cannot say whether it is a genuine or not. This is why digital image authentication, as well as restoration, has become the essential issues, especially when it is utilized in medical science, evidence of court, and forensic science. This paper proposes an effective self-embedding fragile watermarking technique for the digital image authentication as well as recovery. The watermark is generated by quantization, and block truncation coding (BTC) of each 2 × 2 non-overlapping block and embedded in three least significant bits (LSBs) of the corresponding mapped block. The recovery bits are derived from most significant bits (MSBs) of the host image, and the authentication bits are derived from recovery bits, the spatial location of pixels and watermark keys. Even if tempering rate is 50%, the reconstruction of tampered image is achieved with high peak signal-to-noise ratio (PSNR) and normalized correlation coefficient (NCC). The experimental results demonstrate that the proposed scheme not only outperforms high-quality recovery fidelity but also negotiate the blocking artifacts additionally it improves the accuracy of tamper localization due to the use of very small size blocks.