Automated Detection of Brain Abnormality using Deep-Learning-Scheme: A Study
Seifedine Kadry, Yunyoung Nam, Hafiz Tayyab Rauf, V. Rajinikanth, Isah A. Lawal
Abstract
Brain is the vital organ in human physiology; which is conscientious for sensory signal handling and judgment making. The irregularity in brain severely influence entire decision making procedure and the unrecognized and untreated defect will lead to various harsh conditions. This research aims to implement pre-trained Deep-Learning-Scheme (DLS) to classify the brain MRI slices using a multi-class classifier. In this research, the brain MRI slices with classes; normal, stroke, Low-Grade-Glioma (LGG) and High-Grade-Glioma (HGG) are considered for the experimental study. In this work every test picture is resized into 224x224x3 pixels and these imagery are then considered to validate the performance of DLS, such as VGG16, VGG19 and ResNet50 using different classifiers. The attained classification accuracy of every DLS with classifiers, SoftMax, SVM-RBF and SVM-Cubic are presented and discussed.