Segmentation of low-grade gliomas in MRI : Phase based method
Abstract
Segmentation of gliomas in magnetic resonance imaging (MRI) images is a crucial task for early tumor diagnosis and surgical planning. Although many methods for brain tumor segmentation exist, the improvement of this process is still difficult. Indeed, MRI images show complex characteristics and the different tumor tissues are difficult to distinguish from the normal brain tissues; especially the low-grade glioma (LGG), distinguished by their infiltrating character. In fact, it is difficult to extract the tumor from the surrounding healthy parenchyma tissue without any risk of neurological functional sequelae. The purpose of this paper is to provide an overview about a new MRI brain tumor segmentation method based on the local phase information. We applied the proposed method on a set of selected images (Flair, T1 and T1c). Those images were from patients with low-grade glioma. The preliminary results obtained seem to be interesting.