New techniques in reversible watermarking of medical images for authentication and secure storage

Abstract

Telemedicine enables medical diagnosis and patient care using modern medical equipments. These equipments generate huge volumes of data every day. This huge data acquired should be stored, processed and managed which raises security issues like confidentiality, integrity and authenticity. Handling the data poses a very big challenge to the hospitals. There is a possibility of mishandling, mismatching and unauthorized modifications in these data. Also when these data is exchanged in public networks there is a chance for unauthorized tamper. Therefore it is essential to secure the patients data. Also Medical industry deals with data that are very sensitive, even a small distortion are not tolerated though the distortion is visually unnoticeable. Reversible watermarking has been found to be suitable and acceptable in such applications. It has also been applied in the field of intellectual property and newlineinformation security as the original cover data can be retrieved exactly. Considering the above scenario the present work focuses on reversible watermarking to ensure authenticity of the medical information. The first part of the thesis presents a reversible fragile watermarking technique on Least Significant Bit embedding for content authentication. Literature study reveals that conventional LSB scheme provides low embedding rate and is irreversible. Because of its irreversibility, the conventional LSB scheme cannot be used for critical applications like medical image processing where reversibility is mandatory. The proposed scheme presents a modified LSB embedding strategy that satisfies the reversibility. The proposed scheme is improved by utilizing two bits for embedding instead of a single bit. The newlinescheme combines hashing, compression, and digital signature techniques to create content dependent watermark making use of compressed region of interest (ROI) for recovery of ROI. newline newline