Integer Wavelet Transform Based Medical Image Watermarking for Tamper Detection

摘要

Watermarking is used to hide sensitive information from being misused and to reserve integrity. The security of medical images is a prime concern now-a-days and tampers may be introduced by any third party. In this paper, we propose a new integer wavelet transform (IWT) based fragile medical image watermarking for tamper detection. The medical image is first segmented into two regions: Region of interest (ROI), region of non-interest (RONI). The watermark image is scrambled using logistic mapping. Hash value of ROI is generated using SHA-512 algorithm. Last bits of average value of each 5×5 mask of ROI is generated. Then, scrambled watermark; hash value of ROI and compressed last bits are embedded in specific regions of RONI. The specific regions of RONI are chosen based on the position and size of the RONI. Logistic mapping is used to increase security of the proposed technique so that one cannot just easily gain access of the information. Hash value and last bits are used for two level tamper detection. If the hash values of ROI changes, then last bits are compared to detect tampered regions. Results of experiments applied on medical images shows that the proposed method produces high quality watermarked medical images, identifies the presence of tampers inside ROI. Results are evaluated using Peak signal to noise ratio (PSNR), Mean squared error(MSE), Normalized correlation(NC).