Performance Analysis of Medical Image Denoising Techniques

Abstract

Noise reduction in digital image is a daunting assignment for the analysts in Digital Image Processing. As per survey, images are affected by some kind of noises such as gaussian noise, speckle noise and salt and pepper noise. Noise destroys the quality of the active radar, synthetic aperture radar (SAR), medical images. As per the literature, medical images are affected by low resolution, low contrast and geometric deformations, thereby reducing the diagnostic value of medical image. Predominantly, ultrasound images are influenced by speckle noise. Denoising is used to improve visual appearance of an organ and helps in better understanding of disease and accordingly decide the line of treatment for medical field. The proposed work focuses the importance of Spatial filtering techniques that improves resolution, contrast, edge preservation. This paper emphasis on preprocessing techniques that improves the statistical parameter peak signal to noise ratio (PSNR) which are the integral attributes of image quality. The generated result shows that the Wiener and Noise Adaptive Fuzzy switching median filter performs better PNSR values for reduction of low (10%), medium (50%) and high (80%) densities of speckle noise while Adaptive median and Noise Adaptive fuzzy switching median filters achieves good PSNR value for low (10%), medium (50%) and high (80%) noise densities of Gaussian noise among various types of filters. Hence Noise Adaptive Fuzzy switching median filter performs well for low, medium and high noise densities of both types speckle and gaussian noise.