Computer Aided Diagnosis (CAD) tool for MS lesions exploration In multimodal brain MRI

• Published in: International Conference on Advanced Technologies for Signal and Image Processing (Online) pp. 1 - 6
• Main Authors: Nass, Marwa, Mzoughi, Hiba, Njeh, Ines, Benslima, Mohamed, BenHamida, Ahmed
• Format: Conference Proceeding
• Language: English
• Published: IEEE 24.05.2022
• Subjects: 3D reconstruction; Discrete Wavelet Transform; Discrete wavelet transforms; Feature extraction; Fuzzy C-Means; GLCM; Image segmentation; Magnetic resonance imaging; MRI; MS lesions; Multiple sclerosis; Planning; Three-dimensional displays; Traditional Gamma correction (TGC)
• ISSN: 2687-878X
• DOI: 10.1109/ATSIP55956.2022.9805933

Multiple sclerosis (MS) is a chronic inflammatory disease of the central nervous system (CNS) responsible for neuronal degeneration and loss of myelin sheaths. Magnetic resonance imaging (MRI) could provide detailed information about the brain tissues. During clinical routines, manual and accurate diagnosis of brain MS lesions is considered as a harmful and time-consuming critical task. Therefore, Computer-Aided Diagnosis (CAD) tools are of great interest and highly recommended by Neuro-radiologists. In this paper, an automatic CAD tools for MS lesions detection and exploration is proposed using the fusion of different MRI modalities.The proposed CAD consists of four basic steps. The first step is MR images preprocessing based on applying Discrete Wavelet Transform (DWT) combined with Traditional Gamma correction (TGC) to enhance the contrast of MR images brain tissues. The second step is the detection and extraction of the region of interest (ROI) using the Fuzzy-C-Means (FCM) algorithm. The third step is the morphological and regional features extraction using GLCM matrix and the final step is the 3D images reconstruction.Experimental results have been carried out using a benchmark dataset. Even the simplicity of proposed approach, competitive results have been obtained confirming that brain MRI preprocessing constitutes a key factor to boost segmentation 'accuracy of MS lesions.