Protecting Patients’ Privacy using Medical Images Watermarking

Abstract

As a result of modern communication technology, the transmission of medical images among different specialists in different medical institutes has become popular. Accordingly, protecting patients’ data and authenticity against any unauthorized access or modification is a must. Watermarking the images technique before transmission has became a main step to protect patient’s information integrity, copyright, authentication and to protect patients’ information against any signal processing or geometric attacks. This thesis proposes a fragile reversible watermarking scheme. The proposed scheme is based on embedding a Quick Response (QR) code that contains the patient’s data into the medical image followed by encrypting the image using Rivest-Shamir-Adleman (RSA) and finally compressing the encrypted image using Huffman encoding algorithm. The proposed scheme can detect various types of geometric and signal processing attacks and localize tampering caused by copy-paste, text addition and content removal attacks in the extraction steps. For evaluating the proposed scheme, two different datasets were used which are OPENi and MURA as host images and QR code as a watermark image. Peak Signal to Noise Ratio (PSNR), Mean Squared Error (MSE), Structural Similarity Index Measure (SSIM) and Bit Error Rate (BER) were used as evaluation metrics .