Deep-Learning and Geometric Mean Optimizer Combined Scheme for Brain MRI Classification

Disease in brain is one of the medical emergencies and appropriate diagnoses and treatment is necessary. Medical imaging-assisted Brain Condition Monitoring (BCM) is one of the common clinical procedures, and Magnetic Resonance Imaging (MRI) is a standard modality for BCM. This research presents a unique BCM scheme utilizing the deep-learning procedure. Several stages of this arrangement includes the following; MRI slice resizing, feature extraction using deep-learning procedure, feature optimization with Geometric Mean Optimizer (GMO), feature fusion and binary classification. In this work, brain tumour detection is presented. This work is implemented using individual, fused and ensemble of features. The investigational outcome authorizes that this arrangement works fine on the brain tumour detection task using the benchmark image datasets. The proposed approach helped to accomplish a BT recognition accuracy of 100% with the Support Vector Machine (SVM) using ensemble features. This confirms that the proposed MRI-supported BCM arrangement worked well on the chosen image database.