Enhanced MRA Images Quality Using Structure Adaptive Noise Filter And Edge Sharpening Methods

Abstract

MR imaging is an emerging and fast growing medical imaging technique which gives high quality images of the soft tissues. There are certain kinds of noise which contaminates these images and thus makes their interpretation difficult for both human and machine. Filtering is a mathematical technique in which intensities of each pixel of the input image are combined with the intensities of its neighboring pixels, to remove the noise and smooth the image. Filtering could be used with MR images for noise removal. The ordinary image filters blur the image and also remove important structural information like lines and edges. This loss of structural information could be dangerous in a clinical environment and could leads to incorrect diagnosis. To address this problem, a structure preserving noise filter is required. When images are processed for human vision, it is also desirable to make them pleasing by sharpening their edges. Such a structure adaptive noise (SAN) filter and an edge sharpening method is designed and implemented on our MRI Visualization toolkit. Our results show that the methods are effective in removing noise, preserving structure, and sharpening edges.