A Survey of AI-based Approaches for Processing Photoplethysmography Signals

Abstract

Photoplethysmography (PPG) is a non-invasive optical technique that measures physiological parameters like heart rate, blood oxygen saturation, and blood volume. However, PPG signals are often noisy and contaminated with artifacts, posing challenges to inaccurate measurements. To address this, artificial intelligence (AI) techniques have been employed by many researchers to improve  PPG signal processing. This paper presents a comprehensive survey of AI-based approaches for processing PPG signals in recent years. Various AI techniques, including machine learning, deep learning, and natural language processing, are discussed in relation to their application in PPG signal analysis.  The limitations and challenges associated with AI-based approaches in this context are also explored. Furthermore, future research directions are highlighted to leverage AI’s potential for revolutionizing PPG signal processing and expanding its applications. By examining the latest advancements, this survey aims to guide researchers and practitioners in understanding and harnessing AI-based methods for enhanced PPG signal processing, contributing to improved healthcare monitoring and diagnosis.