SVD-DCT Based Medical Image Watermarking in NSCT Domain

Abstract

Medical images are of high importance and patient data must be kept confidential. In this chapter, we discuss a new hybrid transform domain technique for medical image watermarking and provide a detailed analysis of existing image watermarking methods. The proposed method uses a combination of nonsubsampled contourlet transform (NSCT), discrete cosine transform (DCT) and singular value decomposition (SVD) to achieve high capacity, robustness and imperceptibility. This method is non blind which requires cover image in receiver to extract watermarked image. Cover and watermark images are pre-processed in order to ensure accurate extraction of watermark. In this approach, we have considered medical images as cover and electronic patient record (EPR) is used as secret message. EPR message is embedded into selected sub band of cover image with selected gain factor so that there should be a good trade off among imperceptibility, robustness and capacity. NSCT increases hiding capacity and is more resistant to geometrical attacks. The combination of NSCT with DCT and SVD enhanced the perceptual quality and security of watermarked image. Experimental demonstration proved that the proposed method provides high robustness against geometrical and signal processing attacks in terms of peak signal to noise ratio (PSNR) and correlation coefficient (CC).