MRIにおける異なる非線形フィルタ処理画像のノイズパワースペクトル比較
概要
Nonlinear filters that reduce noise components while preserving edges are often used to improve MR image quality. Effects of nonlinear filters change adaptively according to the target images. Therefore, the image quality evaluation is important, but no appropriate evaluation method has been established. We have started a basic investigation to apply noise power spectra(NPS)to image quality evaluation of nonlinear filtered images, such as Non-Local Means(NLM)and Bilateral filters. However, the noise reduction effects of nonlinear filters change complicatedly with the filter parameters and the target images, and it is difficult to characterize differenc- es in NPSs among nonlinear filters. In this study, we adjusted filter parameters so that the noise reduction ef- fects of nonlinear filters were the same. Results showed that differences of NPSs between two nonlinear filters were apparent for low spatial frequency region. Results suggest that differences in filtered images such as im- age blur and edge preservation were reflected in NPS.