A Brief Review of Non-Invasive Blood Pressure Estimation Techniques Using Photoplethysmographic Signals

광용적맥파 신호를 활용한 비침습적 혈압 측정 기술의 현황

In this study, we conducted a brief review of 254 studies on blood pressure estimation using photoplethysmography (PPG). The analysis revealed a growing trend in research on blood pressure prediction using PPG. Initially, simple methods like regression analysis were commonly used, but recent studies predominantly employ deep learning techniques. Additionally, many studies use other signals in conjunction with PPG for blood pressure estimation. There were numerous cases where PPG signals were measured using wearable devices or custom hardware, and a significant number of studies utilized publicly available datasets. The features used in blood pressure prediction were classified into six categories for analysis, and it was found that the combination of two feature categories was the most common approach. Overall, Temporal features were the most frequently used. Based on these findings, it is suggested that future research on blood pressure estimation using PPG should carefully consider factors such as signal characteristics, preprocessing, and feature extraction when developing models.