Reversible data hiding technique exploiting spatial redundancy and correlation between sub-sampled images

Abstract

We present a simple and efficient reversible data hiding method that enables host media to be restored from marked media without any loss of host information. Also, a region-based tampering detection and restoring method exploiting both the improved reversible data hiding algorithm and image homogeneity analysis for image authentication and integrity verification are addressed. Since the reversibility helps to make right decision during image understanding and analysis, it is highly desired in quality-sensitive imagery where even the minimal distortion introduced by embedding data is unacceptable. To insert useful information, the presented method modifies the difference histogram between sub-sampled images. The advantages of the presented method are that (1) it exploits the high spatial correlation inherent in neighboring pixels to achieve high capacity and imperceptible embedding and that (2) it is capable of detecting a tampered region and then recovering the tampered region by using the embedded information selected as the recovery feature. On various test images including 8-bit, 12-bit, and 16-bit images, we demonstrate the validity and effectiveness of the proposed method by comparing to other existing reversible data hiding algorithms. Through the experiments, the presented method can be deployed for applications in the areas of image authentication and verification, tamper proofing, medical and military imaging.