CLASSIFICATION TECHNIQUES ANALYSIS AND COMPARISON FOR DIAGNOSING THE BREAST CANCER BY USING MAGNETIC RESONANCE IMAGES

Abstract

Breast cancer is characterized by the uncontrolled growth of abnormal cells that invade surrounding tissues to support tumor progression. This study presents a framework for Breast cancer MRI image classsification ,  incorporating pre processing ,Segmentation, Feature extraction, Optimization, and classification stages. Image quality is enhanced using a Median filter, while OTSU’s  Thresholding is employed for accurate tumor segmentation.Texture features are extracted using the Gray Level Co-Occurrence Matrix and optimized through   Weighted Particle Swam Optimization (WPSO). The Optimized features were classified using Support Vector Machine (SVM), Random Forest, K-Nearest Neighbors (KNN), and Logistic Regression Models. Performance evaluation based on accuracy,Sensitivity, Specificity, F1-score and False negative rate showed that the Random classifier outperformed the other models, achieving an accuracy of 98.98%, demonstrating its effectiveness and reliability for differentiating Malignant and Normal Breast MRI Images.