An optimal prediction based reversible image watermarking in Hadamard domain

Abstract

This paper presents a reversible watermarking scheme based on a reversible Hadamarh Transform. In the proposed method, the watermark is embedded using the prediction error of Hadamard coefficients. To achieve a more accurate prediction, a Gravitational Search Algorithm (GSA) is used to optimize the prediction coefficients. The proposed method does not need any location map. This property leads to increase the capacity as well as the quality of the watermarked image. To evaluate the performance of the proposed method, a comparative experiment with some well-known reversible methods is performed. The obtained results confirm the efficiency of the proposed method.