Improved Brain Tumor Segmentation and Classification in Brain MRI with FCM-SVM: A Diagnostic Approach

• Published in: IEEE access 2024-01, Vol.12, p.1-1
• Main Authors: Alqhtani, Samar M., Soomro, Toufique A., Ali, Ahmed, Aziz, Abdul, Irfan, Muhammad, Rahman, Saifur, Jalalah, Mohammed, Almawgani, Abdulkarem H. M., Eljak, Ladon Ahmed Bade
• Format: Article
• Language: English
• Subjects: Abnormalities; Brain; Brain cancer; Brain Tumor; Classification; Classifiers; Clustering; Computed tomography; Contrast Limited Adaptive Histogram Equalization; Diffusion Filtering; Fuzzy C-Means; Image contrast; Image enhancement; Image segmentation; Magnetic resonance imaging; Medical imaging; Nervous system; Neuroimaging; Pipelines; Segmentation; Sensitivity; Support Vector Machine; Support vector machines; Tumors
• ISSN: 2169-3536; 2169-3536
• DOI: 10.1109/ACCESS.2024.3394541

Cancer associated with the nervous system and brain tumors ranks among the leading causes of death in various countries. Magnetic resonance imaging (MRI) and computed tomography (CT) capture brain images. MRI is pivotal in diagnosing brain tumors and analyzing other brain disorders. Typically, radiologists or experts manually assess MRI images to detect brain tumors and abnormalities in the early stages for appropriate treatment. However, early brain tumor diagnosis is complex, necessitating computerized methods. This research introduces a novel approach for the automated segmentation of brain tumors and a framework for classifying brain tumor regions. The proposed methods comprise several stages: preprocessing, enhancing the coherence of MRI brain images using Contrast Limited Adaptive Histogram Equalization (CLAHE) and diffusion filtering in the first two steps, followed by the segmentation of the region of interest using the Fuzzy C-Means (FCM) clustering technique in the third step. The last step involves classification using the Support Vector Machine (SVM) classifier. The classifier is applied to different brain tumor types, from meningioma to pituitary tumors, utilizing the CE-MRI database. The proposed method exhibits significantly improved contrast and proves the effectiveness of the classification framework, achieving an average sensitivity of 0.977, specificity of 0.979, accuracy of 0.982, and a Dice score (DSC) of 0.961. Furthermore, this method demonstrates a shorter processing time of 0.42 seconds compared to existing approaches. The performance of this method underscores its significance when compared to state-of-the-art methods in terms of sensitivity, specificity, accuracy, and DSC. For future enhancements, it is possible to standardize the approach by incorporating a set of classifiers to increase the robustness of the brain classification method.