Automatic Diagnosis System Identification of Coronary Heart Disease Using GLCM Feature Extraction and SVM Classification


Coronary heart disease is one of the heart disease that is very much attacking humans. The cause of coronary heart disease is the accumulation of fat and cholesterol in the walls of blood vessels. The main cause of coronary heart disease is due to unhealthy human lifestyles. Poor diet, obesity and lack of exercise are the main causes of coronary heart disease. Narrowing of blood vessels is closely related to blood flow that occurs to the eye, especially the iris. So the need for a method with a more efficient and inexpensive method to find out and identify heart disease since yourself by using iridology. The system is designed by using Circle Hough Transform (CHT) as an automatic detection of the iris, Gray Level Co-occurrence Matrix (GLCM) as feature extraction and Support Vector Machine (SVM) as a classification. Tests have been carried out on 40 image data that obtained an identification success rate of 87.5%.



The conclusions obtained in this study is:

1. It is known that the iris of a person who have heart problems has damaged tissue in the iris. While the iris of someone who does not have problems with his heart tend to have good network.

2. The GLCM feature extraction method has been obtained used for testing on image data similar.

3. The classification method with SVM can be selected as a method of determination success in classifying between data normal and abnormal data.

4. Testing has been carried out automatically cropping the iris using CHT, extraction characteristics using GLCM and classification using SVM on the iris of the eye left. Tests were carried out on 40 test data which consists of 20 normal data and 20 data abnormal. Test results obtained with the highest success rate of 87.5%.

Authors: Vincentius Abdi Gunawan, Ignatia Imelda Fitriani, Leonardus Sandy Ade Putra

Mulawarman University, Kalimantan Timur – Indonesia

Reference link:

Download Full Abstract: System Identification of Coronary Heart Disease