An Intelligent Framework for Automatic Breast Cancer Classification Using Novel Feature Extraction and Machine Learning Techniques

Breast cancer is one of the most significant medical problems of our time. Determining the appropriate methodologies for its early detection is still an open research problem in the scientific community. This research proposes a novel framework for automatically identifying and classifying breast cancer using MRI (Magnetic Resonance Imaging) images. The proposed approach utilizes automatic segmentation methods to detect suspicious areas in MRI images, features new feature extraction, and utilizes a variety of classification methods to create an automatic decision-making system that is able to classify the MRI images as benign or malign cancers. This research used MRI images of 56 patients from the medical imaging department of King Abdullah Medical City (KAMC), Saudi Arabia to assess the performance of the proposed framework. Our framework was able to achieve a classification accuracy of over 98% for its optimal configuration (SVM -linear kernel), while demonstrating excellent false-positive and false negative rates, sensitivity and specificity (0%,15%, 97%, 100% respectively).