Automated segmentation of the lateral ventricle of MR brain by fuzzy inference

The cerebral ventricle is a part of the brain region filled with cerebrospinal fluid (CSF). Some brain diseases may cause a change of the shape of the cerebral ventricle. Information of the shape and volume is useful for diagnosing hydrocephalus and Alzheimer´s disease. The authors propose a fully automated procedure for segmenting the lateral ventricle from human brain MR volumes. In this method, we first decompose a segmented CSF into small parts (primitives) using watershed segmentation and then determine the primitives which belong to the lateral ventricle by fuzzy if-then rules. We applied our method to 10 MR volumes (10 different persons). As a result, our method can segment the lateral ventricle with high accuracy and therefore provide the three dimensional shape