Brain Tumor Segmentation Using Improved Kernel Weighted FCM

S. Sadagopan

Magnetic Resonance Image segmentation is a challenging task because of the presence of intensity inhomogeneity and bias field. We propose a modified fuzzy clustering means method for segmenting tissue regions like grey matter, white matter, cerebrospinal fluid and abnormal tissue like tumor from the brain image automatically. The proposed method consists of pre processing using spectral subtraction de-noising to remove noise and modified kernel weighted Fuzzy C-Means clustering algorithm segments the normal tissues by considering spatial information because neighboring pixels are highly correlated and also construct initial membership matrix randomly. A weighted fuzzy factor is introduced to the objective function to reduce the membership function. The Jaccard similarity is also integrated with the objective function to correct intensity in homogeneity simultaneously. The results show that the proposed algorithm yields better results in segmenting the brain tumor.

Volume 11 | 09-Special Issue

Pages: 899-905

DOI: 10.5373/JARDCS/V11/20192649