HUC Academic Repository

Brain Tumour Isolation in MRI Images Based on Statistical Properties and Morphological Process Techniques

Abstract

With the great development in medical image processing that has been achieved recently, it has become an integral part of diagnostic tools and treatment tools attached to medical imaging devices such as the magnetic resonance imaging (MRI) device. In this paper, we propose an algorithm to extract tumours from MRI images using the MATLAB environment. MRI images were enhanced using different filters, such as the median filter and Wiener's filter. The best results were obtained via Wiener's filter. The k-mean method was used as a key step in the proposed algorithm, then morphology processes were applied and a threshold was proposed based on the statistical properties of the image (standard deviation, maximum and minimum values) to obtain the tumour area. The experimental results obtained by applying the proposed algorithm and DICE coefficient test on different MRI angles (axial, sagittal and coronal) indicate high accuracy was achieved in tumour isolation.