Authentication Watermarking with Tamper Detection and Recovery (AW-TDR)

Abstract

This paper will study and evaluates watermarking technique by Zain and Fauzi (2006). We proposed a watermarking scheme that can detect and localise tampered and recovered images. The purpose is to verify the integrity and authenticity of medical images especially to minimize the distortion of the Region of Interest (ROI). Recommendations will then be made to enhance the technique especially in the aspect of recovery or reconstruction rate for medical images. A proposal will also be made for a better distribution of watermark to minimize the distortion of the Region of Interest (ROI). The proposal will enhance AW-TDR in three aspects: Firstly, image quality in the ROI will be improved as the maximum change is only 2 (two) bits in every 4 (four) pixels, or embedding rate of 0.5 bits/pixel; Secondly, recovery rate will also be better since the recovery bits are located outside the region of interest. The disadvantage is that, only manipulation done in the ROI will be detected; and lastly, the quality of the reconstructed image will be enhanced since the average of 2 x 2 pixels would be used to reconstruct the tampered image. We will present our final proposal for AW-TDR that include preparation of blocks, the embedding algorithm and the location plan for authentication bits and recovery bits. The experimental results demonstrate that the precision of tamper detection and localisation is close to 100% after level-2 detection. The tamper recovery rate is better than 86% for a less than half a tampered image.