KLASIFIKASI PENYAKIT STROKE DENGAN CITRA CT SCAN OTAK MENGGUNAKAN ALGORITMA BACKPROPAGATION

Stroke is the second leading cause of death in the world and is the leading cause of disability in the productive age. From the results of basic health research conducted by the Indonesian Ministry of Health in 2018 which showed an increase in the prevalence of stroke from 7 percent to 10.9 percent. Viewed from the cause of a stroke divided into two namely ischemic or hemorrhagic. General examination is carried out in order to obtain a picture of the part of the brain that has had a stroke using Computerized Tomography (CT) Scan. The image produced by the CT Scan is manually checked and requires good lighting by an experienced neurology doctor. In the examination the doctor will look for a section called a lesion to determine the type of stroke suffered by the patient. In general, the quality of scanning images, in the form of gray level digital images, has decreased due to external factors (noise) and the medical equipment used. with the need for an image processing with a method that can classify types of stroke. The method used in this study for classification is backpropagation neural networks. CT brain scan images are used as input for image processing. The image stages before prior classification are pre-processing (Grayscale, CLAHE, open morphology) and extraction of the Gray-Level Co-Occurrence Matrix (GLCM) feature. In this study, the GLCM feature extraction process uses an average of 0o, 45o, 90o, 135o angles, and uses a neighboring distance 1. Features used in GLCM are contrast, correlation, energy, homogeneity which will be used as input parameters in the classification process using backpropagation. After conducting several testing processes, it can be concluded that the backpropagation neural network can classify types of stroke through brain CT-scan images with an accuracy rate of up to 94,12%.