A High Capacity Framework for Reversible Information Embedding in Medical Images

Abstract

In this paper, a scheme of reversible data hiding (RDH) with increased capacity for medical images has been presented. The proposed method focusses on increasing the medical image's visual quality and simultaneously improving the payload. This algorithm presents a block based approach and modifies only pixels in a given range and not in the entire image. The histogram of each individual block is generated with the objective of computing the peak value i.e. the value with highest frequency of occurrence. For medical images, an additional preprocessing step is first done to make computation simpler. Stego-image is formed by embedding particular values of secret data bits in a pre-determined range. The cover-image is recovered in its original form by concatenating a location tag to the stego-image. Peak Signal to Noise Ratio (PSNR), Structural Similarity Index Matrix (SSIM) and capacity are the metrics that assess the visual quality, structural similarity and payload. Since the process of embedding of data is done only in particular pixels of a defined range, the proposed idea results in increased payload with less distortion.