Mean Square Error Optimization to Authenticate Multi-Modality Medical Images in Wavelet Domain

Abstract

Medical image watermarking which hides patient’s information in the medical image ensures its imperceptibility and at the same time provides robustness to the information that is embedded. Current research scenario embeds the authenticated message as watermark in the frequency domain to get watermarked image which has most of its values as real numbers. Since pixel intensity values must be represented as integers, the watermarked image is converted into spatial domain where the rounding error problem occurs and the embedded message cannot be extracted to prove its authenticity. In proposed work, a new watermarking scheme in Discrete Wavelet Transform (DWT) domain based on Genetic Algorithm (GA) is implemented to ensure correct retrieval of embedded watermark by optimizing Mean Square Error (MSE) measure.