A Study on Computer-Aided Diagnosis of Human Brain Tumor Detection and Classification using Binary Genetic Algorithm and Deep Learning Techniques

Computer-aided detection (CAD) frameworks enable doctors diagnose faster and more accurately. The objective of paper is to present a hybrid learning approach for a CAD system for automatically detecting brain tumors using magnetic resonance (MR) images. The methodology focuses on four computational techniques: (1) the Conventional and deep learning (DL) for Feature Extraction (2) select a deep features by last fully connected layer and conventional optimal feature sets are from Genetic Algorithm (GA) (3) combine the features that were chosen by employing a fusion approach and (4) the support vector machine (SVM) is used to classify inputs as either normal or abnormal. A total of 125 flair MR images, 50 normal and 75 abnormal, were used in the experiments. These images were taken from a real-world of brain MRI dataset. The results demonstrated that the proposed method efficiency compared with the other MR images state of the art techniques. The outcomes demonstrated how specific and robust the proposed hybrid technique. At last, some probable next steps are proposed