Study and development of optimised Denoising technique for medical images

Abstract

In medical imaging, the accurate diagnosis and evaluation of a disease relies upon newlineboth image procurement and image interpretation. But medical images are degraded by newlinenoises through the transmission and reception procedure. Therefore, noise reduction has been newlinea conventional problem in medical image processing. Denoising images is an image newlineprocessing technique that attempts to remove noise, although it is still a challenge for newlinescientists. newlineIn order to reduce noise, spatial filters lead to misleading and flatten the data. newlineAlthough several denoising algorithms based on the wavelet framework have been developed newlinesuccessfully, they all suffered from a lack of directionality, aliasing, shift variance, and newlineoscillations. newlineIn the field of denoising, there is still a great deal of room for development. This newlineresearch work designs a high capacity denoising algorithm using adaptive subband newlinethresholding technique. In this work, the original medical diagnostic information is assessed newlinefor the noise variance and based on the decomposition level, the Adaptive optimum threshold newlinemethod is applied to it. newlineThe proposed algorithm gives a complete denoising procedure, which suits for newlinedifferent noises which may corrupt the data. The Gaussian Noise (GN), Salt and Pepper newlineNoise (SPN), and Speckle Noise (SN) inputs have all been used to corrupt these images. A newlinestatistical model dependent upon the wavelet coefficient magnitude and noise variance for newlineeach coefficient is assessed in light of the subband. Adaptive thresholding utilizing optimum newlinethresholding method is performed on each subband coefficient relying upon the decay level. newlineThe proposed algorithm yields a predominant quality in medical images by producing a newlinesuperior PSNR value. newlineIt has been shown that the Discrete Wavelet Transform (DWT) is an excellent newlinetechnique for decomposing an ultrasound image into its constituent components and newlineapproximation coefficients. To enhance the noisy image and obtain images with lower and newlinehigher frequency content, the DWT is performed on the noisy image.