Hybrid approaches for brain tumor detection in MR images

Abstract

The growing technology in medical image processing helps in quick as well as an accurate analysis of several life threatening diseases. Interestingly, domain of brain tumor analysis has effectively utilized this trend to automate core steps, i.e. extraction, detection, and the most important proximate segmentation for tumor examination. To diagnose neurological disorders magnetic resonance (MR) imaging methods are of great help. Discussing the MR image types this paper briefs the parameters influencing the process of brain tumor detection. Also, the study proposes a hybrid segmentation approach combining k-means with fuzzy c-means (FCM) and support vector machine (SVM) with fuzzy c-means. Experimentation performed show that fusion outperforms three of the base approaches in brain tumor identification on DICOM dataset using 200 T1W and T2W MR images. The evaluation parameters show that k-means combined with fuzzy c-means produce better accuracy. Results further prove applicability of the proposal in detecting ranges and shapes of brain tumor using MR images.