An automated method for brain tumor segmentation on MRI images using SVM algorithm

Automatically segmentation of brain tumor onMRI images is a very difficult and important task for theexperts in brain surgery planning. If experts do thesegmentation manually according to their own knowledgeand experience, it will be time-consuming and may not reachany result. In this article, in order to minimize the time andreduce expert’s diagnosis errors, an automatic method forsegmentation of brain tumor on MRI images based on SVMalgorithm is presented. In this research, 30 T1-weighted MRIimages are used from www.oasis-brains.org , because thistype of images has higher contrast than T2-weighted and CTimages. The proposed method segmented 26 tumor MRIimages out of the target data and practically it achieved to86.66% of segmentation among these 30 data.

Keywords:

Digital image processing , Image segmentation , Medical images processing , MRI images , SVM