Cerebral vasculature extraction using classifier fusion

The detection of cerebral aneurysms is of a paramount importance in the prevention of intracranial subarachnoid hemorrhage. The segmentation of intracranial vasculature presents a crucial step in the detection scheme. We propose in this paper, a new approach to extract cerebral vasculature in 2D-DSA images based on multiple classifier fusion. The classifiers used here are the FCM and the Fuzzy KNN. The main advantage of multiple classifier fusion is increasing classification efficiency and accuracy. The proposed method demostrates the contribution of fusing FCM-FKNN over the use of the individual classifier. Our method succeeded in classifying 6.23% of pixels rejected by the FCM method and 8.36% of pixels rejected by the FKNN method.