De-Noising X-Ray Images BY Fast Non-Local Mean FOR Pulmonary Tuberculosis Detection – An Application

Abstract

There are numerous facets of our existence, including the medical field, that need the use of digital photographs. However, medical pictures may get distorted due to noise, particularly Poisson noise, which can dramatically lower the quality of X-ray images. This can be especially problematic in emergency situations. Denoising an image is a basic image processing method with the goal of reducing the amount of noise present in a picture. The Peak Signal-to-Noise Ratio, abbreviated as PSNR, is a statistic that is often used to evaluate the quality of denoised photographs. This research indicates that the Non-Local Means (NLM) approach is successful in decreasing Poisson noise in X-ray lung images taken from the publicly accessible Shenzhen (SH) and Montgomery (MC) datasets. The datasets were obtained from Shenzhen and Montgomery, respectively. Images of the lungs are often blurry, thus in order to denoise them, the NLM method, more especially the Fast Local Means approach, is used. The following three important metrics—peak signal-to-noise ratio (PSNR), structural similarity index (SSIM), and root mean square error (RMSE)—are used in the process of performance analysis of the method. The results of the experiments make it abundantly evident that the suggested method is capable of effectively preserving picture edges while simultaneously lowering noise. As a consequence, the X-ray images are upgraded to have a higher PSNR, an increased SSIM, and a lowered RMSE.