Beat classification of an ecg signal using photoplethysmography and neural network

Pramod R. Bokde and Choudhari N.K

This paper presents a simple method to indirectly estimate the range of certain important electrocardiogram (ECG) parameters using photoplethysmography (PPG). The proposed method, termed as PhotoECG, extracts a set of time and frequency domain features from fingertip PPG signal. A feature selection algorithm utilizing the concept of Maximal Information Coefficient (MIC) is presented to rank the PPG features according to their relevance to create training models for different ECG parameters. The proposed method yields above 90% accuracy in estimating ECG parameters on a benchmark hospital dataset having clean PPG signal. The same method results an average of 80% accuracy on noisy PPG signal captured by iPhone, indicating its feasibility to create phone applications for preventive ECG monitoring at home.
An abnormal respiratory rate is often the earliest sign of critical illness. A reliable estimate of respiratory rate is vital in the application of remote tele-monitoring systems, which may facilitate early supported discharge from hospital or prompt recognition of physiological deterioration in high-risk patient groups. Traditional approaches use analysis of respiratory sinus arrhythmia from the electrocardiogram (ECG), but this phenomenon is predominantly limited to the young and healthy. Analysis of the photoplethysmogram (PPG) waveform offers an alternative means of non-invasive respiratory rate monitoring, but further development is required to enable reliable estimates. This review conceptualizes the challenge by discussing the effect of respiration on the PPG waveform and the key physiological mechanisms that underpin the derivation of respiratory rate from the PPG.


Download PDF: