Removal of impulse noise using Histogram-based Localized Wiener Filter for MR brain image restoration

Medical Imaging is one of the important technique which plays a major role in diagnosis of diseases which are present inside of our body. These medical images are produced by many imaging modalities such as, Computed Tomography (CT), Magnetic Resonance Imaging (MRI), Ultrasound(US) and X-ray. Mostly these medical images are corrupted by noise during image acquisition and transmission process. The noise present in the image degrades visual quality and it affects the accuracy of segmentation result. In this paper, we present a method called Histogram-based Localized Wiener Filter (HLWF) to denoise MR brain image. This method is tested with the brain images obtained from IBSR and IDEA GROUP database. The performance of this method is quantatively evaluated by calculating the Peak signal to noise ratio (PSNR) value. The output of the proposed method is then compared with the existing method such as Anisotropic Diffusion Filter (ADF), Bilateral Filter (BLF), Non-Local Mean Filter (NLMF) and Wiener Filter (WF). Experimental results show that the proposed method gives better result on denoising of impulsive noises in medical images compared to the existing methods.

I. Introduction

Medical Images are the pictured representations of the internal part of a body which is used for clinical and medical diagnosis. These medical images are generated by imaging methods namely CT, MRI, US, and X-ray. In general, medical images are affected by noise during image sending and receiving process.