Two-dimensional fast magnetic resonance brain segmentation

Jasjit S. Suri · IEEE Engineering in Medicine and Biology Magazine · 2001

Region-based level-set snakes are a very powerful technique for segmenting white matter/grey matter in MR slices of human brain. We showed how one can apply the region-based level-set technique for segmenting the brain using fast techniques. The system used the fuzzy clustering method for computing the fuzzy membership values, which were used in the regional speed computation. Recently, the authors have developed a mathematical morphology-based speed control function that acts as a regularizer for making the propagation more robust and leak free. It would also be worth exploring how either the neural network or learning models would do in terms of the performance evaluation if clustering was to be replaced.

Read the paper · More papers on PaperTik