Semantic Segmentation of Brain Tumor from MRI Images and SVM Classification using GLCM Features

Brain tumors are a life-threatening condition that occurs due to the presence of abnormal or unwanted tissues in the brain, posing severe risks to a patient’s health. Early detection of brain tumors is crucial for timely treatment and improving survival rates. This study focuses on semantic segmentation of brain tumors from MRI images and their classification using Support Vector Machines (SVM) with Gray Level Co-occurrence Matrix (GLCM) features. The process involves a comprehensive approach starting with pre-processing, including resizing the input images, converting them into JPEG format, applying median filtering for noise removal, and performing skull stripping to isolate the brain region. Watershed segmentation is utilized to identify and extract the tumor region accurately. Post-segmentation, GLCM-based feature extraction is applied to derive texture features from the segmented regions. These features, along with Labeled data, are then fed into an SVM classifier to categorize the tumor into one of four classes: glioma, meningioma, pituitary tumor, or no tumor. The proposed method demonstrates high accuracy, making it an effective tool for automated tumor detection and classification, assisting radiologists in early diagnosis and reducing manual interpretation errors. This approach leverages the strengths of semantic segmentation and machine learning, providing a robust and efficient solution for brain tumor analysis.

Keywords: Brain tumor Dataset, Pre-Processing GLCM Feature Extraction, Machine learning, SVM, Classification, Accuracy.