A Novel Method of Medical Image Denoising Using Bilateral and NLm Filtering

Abstract

One of the bottleneck in medical imaging is due to low Signal-to-Noise Ratio(SNR) which requires long and repeated acquisition of the same subject to reduce noise and blur. To obtain a high SNR image without lengthy repeated scans, post processing of data such as denoising plays a critical role. Bilateral and Nonlocal means (NLm) filtering are commonly used procedures for medical image denoising. In this paper we propose a new thresholding scheme to wavelet and contour let based denoising by introducing a scaling factor to universal threshold. Also we propose the aforementioned novel contour let thresholding scheme as a preprocessing step on bilateral filtering and NLm denoising. Simulation results show that the proposed single entity comprising the novel preprocessing step and bilateral or NLm denoising is superior in terms of PSNR and perceptual quality compared to Bilateral filtering or NLm denoising used alone.