A study of watermarking methods in medical application

• Main Author: Ng, Lee Ping
• Format: Thesis
• Language: English
• Published: 2004
• Subjects: QA Mathematics; QA71-90 Instruments and machines
• Online Access: http://eprints.uthm.edu.my/8165/1/24p%20NG%20LEE%20PING.pdf; http://eprints.uthm.edu.my/8165/

Nowadays, watermarking technology is playing an important role in medical image watermarking as it can hide the patient information and then get back the information by the owner itself using certain private key. The objective of this project is to address the specific requirements needs for watermarking of medical images and compare three appropriate schemes for watermarking of medical images. Four important requirements in medical image watermarking are imperceptible watermarking, integrity control, authentication and capacity hiding. The three methods of medical image watermarking that have been studied are Least Significant Bit (LSB), Rivest, Shamir and Adleman (RSA) encryption and decryption with LSB, RSA encryption and decryption with feature-based. These three methods of image watermarking are categorized in spatial domain. Performance for watermarking scheme is carried out and comparisons are made. There are three quality metrics used in this project which are Peak-to-Noise Ratio (PSNR), Mean Square Error (MSE) and Signal-to-Noise Ratio (SNR). Among these three methods, LSB method has the most capacity of watermark and fastest rate in embedding process. These three algorithms implemented in this project are shown to be sensitive for the attack. Attacks applied in this project are salt & pepper noise with 0.0001 noise density, median filtering, JPEG compression with index 100 and index 75, cropping and rotation. These three methods that being studied are better for integrity control. Meanwhile, a Graphic User Interface (GUI) is developed for embedding and extracting purpose. User can embed a watermark into their host image and extract the watermark from the watermarked image.