Hemorrhagic Stroke Classification on Computerized Tomography Scan Images of the Brain using CNN Algorithm

Stroke is the number 2 deadly disease and the third cause of disability worldwide, with symptoms appearing suddenly and quickly in hours or even seconds. There are 23,809 potential and 2,243 possible misdiagnosed strokes, 101 million stroke cases that have occurred, 12.2 million new stroke cases each year with an increased rate of about 1 case per 3 seconds, and 6,552,724 cases of death and disability caused by stroke each year. The death rate for stroke is so high that it is essential to take preventive measures to reduce the risk of stroke. One way of prevention can be predictions using machine learning algorithms. This study designed a hemorrhagic stroke classification system through brain CT scan images using the Convolutional Neural Network (CNN) algorithm. Six datasets are used, with 13,568 CT Scan slice images in PNG format. The designed system correctly classifies CT Scan images into regular or hemorrhagic classes. The average accuracy value of the designed system testing phase reaches 87.67%.