A BrainNet (BrN) based New Approach to Classify Brain Stroke from CT Scan Images

Abstract

Worldwide, brain stroke is known as the 2nd leading cause of death, and based on Indian history, three people have suffered every minute. There are mainly two different types of brain stroke: ischemic stroke and Hemorrhagic stroke used to train the proposed models. Ischemic stroke is the most common and it contributes mostly to 80% of the brain stroke and Hemorrhagic stroke contributes mostly to 20% of the brain stroke. In the proposed model, there has been used a hybrid model called BrainNet (BrN) as CNN(Convolutional Neural Network) and SVM(Support Vector Machine)to classify brain stroke disease. After applying the required proposed model, it has produced a smart score of 91.91% accuracy, and compared to the existing model it performs pretty well. The BrainNet (BrN) model is mainly designed based on a deep neural network with dataset collection, preprocessing, and feature extraction with the desired model and make the classification concerning SVM. With compare to the existing model, it is an acceptable performance that belongs to the collected dataset designed with Ischemic stroke and Hemorrhagic stroke disease within the total number of 2515 data.