A non-invasive technique of early heart diseases prediction from photoplethysmography signal

Now a day heart diseases are the major cause of death throughout the world. So, early prediction of heart diseases is essential to reduce premature death. In this paper a non-invasive technique is proposed to detect the heart rate from photoplethysmography (PPG) signal. Non-invasive measurement of heart rate plays a vital role in medical diagnosis due to less discomfort to patients because tissues can be damaged due to traditional sensors during the extraction of cardiac signal. Photoplethysmogram is the blood volume measurement technique which is based on optic principle. The blood volume changes in accordance with heart rate. In this work, a convenient heart diseases estimation technique is proposed from PPG which is extracted from human facial video. It will overcome the high cost of conventional expensive ECG machine for cardiac diseases prediction. The heart rate is identified from FFT analysis and verified with the conventional heart rate measurement system using ECG. The percentage error is 2.3%, 3.2% and 5.8% for green, red and blue channels respectively. From heart rate the cardiac diseases have been predicted which has great potential for cardiac heart rate monitoring for computing the heart rate variability (HRV) without any contact of traditional sensors.