Real Time Heart Beat Monitoring Using Computer Vision

Among the different biological parameters, heart rate (HR) plays a very significant part in the disease diagnosis. It confirms the state of living of a person. Heart rate gives an idea about the person's health and hence it should be observed. Nowadays, a non-contact based method is more comfortable, low cost, simple and easy to use. In this era, with an increase in automobiles, people who are engaged in driving are also on the surge. Such people are also victims of cardiac-related diseases. The proposed method demonstrates real-time non-contact heart rate monitoring systems that enable the detection of cardiac activity cycles without any cutaneous contact using machine learning. The methodology involves the extraction of heart beats signals from face images using independent component analysis. Neural network is employed to generate the BPM from the samples. As part of preparing the dataset, a heartbeat sensor is developed with a microcontroller. The algorithm is successfully tested with real time face images and the non-contact based detection method more conveniently helps to measure HR with more acceptable accuracy. It can apply to the drivers for the prevention of accidents or manual labourers who work in.