Reversible image authentication scheme based on prediction error expansion

Abstract

Reversible image authentication scheme is a technique that detects tampered areas in images and allows them to be reconstructed to their original version without any distortion. In this article, a new, reversible, image authentication scheme based on prediction error expansion is proposed for digital images. The proposed scheme classifies the host image into smooth blocks and complex blocks. Then, an authentication code that is created randomly with a seed is embedded adaptively into each image block. Experimental results showed that our proposed scheme achieves the high accuracy of tamper detection and preserved high image quality. Moreover, the proposed scheme achieved the reversibility, which is needed for some special applications, such as fine artwork, military images, and medical images.