Intelligent selection of embedding locations for reversible watermarking based on 2D-difference expansion

Fragile watermarking is frequently adopted to authenticate multimedia and to detect tampering. In many particular applications such as medical or military discipline, the use of reversible scheme as data hiding tools is required. In the literature, we have noted the frequent use of difference expansion scheme for data embedding. In this paper, we propose a two dimensional difference expansion scheme (2D-DE) which can control alteration of modified image using Particle swarm Optimization approach. The developed method consist of dividing original image into blocks of 4×4 pixels, and embed 16 bits of payload in only expandable blocks. The distortion control operation is based on statistical information like standard deviation. This propriety can be exploited as separation measure for enhancing the perceptual quality. It´s a threshold-based which is adaptive metaheuristic by PSO. The experimental results prove that the proposed approach outperforms many approaches based on regular difference expansion.