Adaptive Multi Scale Products Threshold-Based MRI Denoising

Abstract

Denoising an image has become an extremely important step in medical imaging, and it is performed throughout the entire diagnostic process. In medical imaging, it is imperative that a balance be maintained between the elimination of distracting noise and the maintenance of diagnostically relevant information. Imaging modalities have many objectives, one of the most important of which is to supply the doctor with the most reliable information possible so that they can make an precise diagnosis. The utilization of multiresolution noise filters in a wide range of medical imaging applications is garnering an increasing amount of attention. This study discusses some of the possible uses of new wavelet denoising algorithms for medical magnetic resonance images and reviews some of the techniques that have been used recently. These techniques were used to investigate various areas of the human body. The goal of this project is to demonstrate and evaluate various approaches of noise suppression that are based on both image processing and clinical experience. Rician noise is a phenomenon that is frequently observed in magnetic resonance imaging (MRI). In the field of medical image processing, edge-preserving denoising is becoming an increasingly important technique. In this paper, a wavelet-based multi scale products thresholding system is presented for the purpose of eliminating noise in magnetic resonance pictures. A dyadic wavelet transform that works similarly to an edge detector is used. As a consequence of this, significant features in images will continue to evolve with high magnitude throughout wavelet scales, whereas noise will quickly fade away. The wavelet sub bands that are next to one another are multiplied in order to improve edge structures while simultaneously reducing noise in order to take advantage of wavelet inter scale dependencies. When using the multi scale products, it is possible to differentiate edges from noise in an efficient manner. After that, an adaptive threshold is computed and applied to the products rather than the wavelet coefficients so that relevant features can be identified. Experiments have demonstrated that adaptive multi scale products thresholding is superior to conventional wavelet-thresholding denoising approaches in terms of its ability to reduce noise and retain edges. The fact that the wavelet transform can recreate an image without any noticeable loss of quality is the primary benefit of using this technique.