Acute Coronary Syndrome Disorders Classification Based On ECG Images Using Adaptive Neuro-Fuzzy Inference System (ANFIS)

Abstract

The heart is one of the most vital human organs that acts as a blood-pumping tool to supply oxygen and essential nutrients throughout the human body. Abnormalities in the heart greatly affect the work of the heart which results in the heart not being able to carry out its duties properly. Heart defect is one of the most common causes of death in many countries, including Indonesia. Electrocardiogram (ECG) is one of the most important examination models used to diagnose various abnormal heart rhythms. An ECG records the electrical activity of the heart by showing waveforms on a monitor or printing them on paper to classify cardiac abnormalities from the electrocardiogram image using image processing and artificial neural networks. The method used for the classification is the Adaptive Neuro-Fuzzy Inference System (ANFIS) and using the Chain code to take the value of the ECG feature. There were 92 ECG images to be used which were partitioned to 70 images for training data 22 images for test data with 3 types of abnormalities, namely coronary heart disease, angina, and myocardial infarction. The test was carried out using 4 choices of ANFIS functions. The parameters used to classify coronary heart disease, angina pectoris, and myocardial infarction reached 95.23% (DR), and 29.41% (DER), using the GBell function with the number of MFs (3) and epoch (100).