Improvements of nuclear magnetic resonance image quality using iterations of adaptive nonlinear filtering

To improve nuclear magnetic resonance (NMR) image quality, a method using iterations of adaptive nonlinear filtering is presented. The method utilizes the properties of an NMR image's noises which are spike-like in the local region. An NMR image is iteratively filtered by a mode-change-type nonlinear filter which has two modes. The first mode reduces small amplitude noises in the NMR image, so that the S/N ratio of the image is improved, and then the residual noises become more spike-like. The second mode mainly reduces large amplitude spike-like noises of the image improved by the first mode, which cannot reduce spike-like noises. Using the proposed iterative filter, it is found by simulations that the S/N ratio of the NMR image is improved by as much as a factor of 2.7. A significant improvement in the image quality when this filter is used is demonstrated by applying it to an NMR image of a human head obtained from an actual NMR imaging apparatus.