A hybridized methodology of different wavelet transformations targeting medical images in IoT infrastructure

Abstract

The Internet of Things (IoT) paradigm has become a vital part of all significant scientific sectors, including the healthcare domain. Medical images in the healthcare sector are indispensable items that are usually susceptible to distortion once they are shared and transferred via the Internet. The sector faces the distinct and constant challenge of preserving medical data, which can be manipulated by various malicious attacks, in turn potentially compromising the patients’ diagnostic data. In this situation, such medical data ought to be private, with access only granted to patients and physicians. This paper elaborates on a hybrid measurement technique for digital image watermarking that utilizes medical images (X-ray, MRA, and CT), which are an extremely robust method for protecting clinical information. The authors explore various different wavelet families, in addition to hybridization between these wavelets. These are carried out on three levels decomposition of Discrete wavelet transformation (biorthogonal 6.8 wavelets, biorthogonal 3.5 wavelets, biorthogonal 5.5 wavelets, reverse biorthogonal 6.8, reverse biorthogonal 3.5, reverse biorthogonal 5.5, discrete meyer, symlets 5, symlets 8 coiflets 4 wavelet, and coiflets 5 wavelet transform). Each level uses various types of wavelet transformation to present the watermarked image, and then extracts the medical watermark from the original watermarked image. The results of diverse types of attack have been compared, while the proposed measurement technique's performance is evaluated using statistical parameters (MSE, PSNR, SSIM, and NC). This in turn measures the quality of the image, which so far shows promising results.