Algorithm To Calculate Pulse from PPG Signal After Eliminating Touch Errors from the Fingertip Video Captured by Smartphone Camera

TL;DR

This paper presents a novel algorithm for extracting pulse from fingertip PPG signals captured via smartphone camera, effectively eliminating touch errors and improving accuracy over existing methods.

Contribution

The research introduces a new approach to remove camera touch errors from PPG signals using pixel intensity analysis, enhancing pulse measurement accuracy from smartphone videos.

Findings

Effective removal of touch errors from PPG signals

Improved pulse detection accuracy compared to existing algorithms

Successful implementation using MATLAB with potential for mobile application

Abstract

With the ongoing heart problems of the population worldwide, the medical requirements of the people are expected to increase. Electrocardiogram (ECG) is one of the proven to capture the heart response signal to assess the electrical and muscular functions of the heart. The ECG setup is expensive and needs proper training, and of course, it is not instant. For fast, accurate heart parameter monitoring, scientists pay attention to the photoplethysmogram signal (PPG), based on the light intensity of a particular wavelength. Android smartphone with a good quality camera has come to ordinary people's reach and has become one of the most necessary and rugged devices for today and future generations. We can use its powerful features to solve or assess heart state monitoring by capturing the image's necessary data. The mobile camera has a photo emitting diode and a photodetector. The light source illuminates the tissue. The photodetector calculates the small variation in light intensity associated with blood volume change in the vessels (mainly fingertips, toes, and ears). We have captured unfocused contact video to capture PPG using an Android Smartphone. Then, we removed a certain percent of camera touch errors based on average pixel intensity count in the red plane, and it is a new approach that has been introduced in this research. We used a 2nd order Butterworth (IIR) band pass filter for noise removal, FFT Hann Window for frequency analysis and leakage reduction. We have developed an algorithm using MATLAB as a development platform, for accurate pulse (BPM) measurement. Moreover, we have done a comparative analysis of developed algorithm with other available algorithms for PPG-based pulse calculation. In this study, the fingertip video was captured when the body was at rest