Automatic MR Brain Tumor Detection using Possibilistic C-Means and K-Means Clustering with Color Segmentation

Magnetic resonance imaging is often the medical imaging method of choice when soft-tissue delineation is necessary. This paper presents a new approach for automated detection of brain tumor based on k-means and possibilistic c-means clustering with color segmentation, which separates brain tumor from healthy tissues in magnetic resonance images. The magnetic resonance feature images used for the tumor detection consist of T1-weighted and T2-weighted images for each axial slice through the head. The proposed method consists of three stages namely pre-processing, segmentation and feature extraction. In the first stage, we have suppressed the noise using image filtering. In the second stage, segmentation is computed using an unsupervised k-means and possibilistic c-means clustering algorithm with color conversion. The segmentation accuracy is obtained using the silhouette method. The experimental results show the superiority of the possibilistic c-means clustering method. In the third stage, the key features are extracted using the threshold. The application of the proposed method for tracking tumor is demon¬strated to help pathologists distinguish exactly tumor size and region.