A two-stage human brain MRI segmentation scheme using fuzzy logic

The authors have developed a two-stage image segmentation scheme using fuzzy logic. Based on the scheme, a two-stage fuzzy system has been built for segmenting human brain MR images. The first stage is a fuzzy rule-based system that assigns memberships to pixels, classifies the pixels that have only one high membership and calculates the initial conditions for the next stage. The second stage is the fuzzy c-means algorithm, which classifies the undetermined pixels. Preliminary segmentation of the human brain MR images shows that the two-stage fuzzy system could accurately determine white matter, gray matter, cerebrospinal fluid and HIV+lesion. The results were visually confirmed by expert observers. The satisfactory results achieved in this paper suggest the feasibility of developing similar segmentation systems for other types of images and the possibility of extending the two-stage scheme to multiple-stage schemes