Deep Learning and Ensemble Models for Accurate Heart Rate Estimation from PPG
作者
Dhatri Shankari T M,Rashmi N Ugarakhod
出处
期刊:IEEE India Conference日期:2024-12-19卷期号:: 1-6
标识
DOI:10.1109/indicon63790.2024.10958345
摘要
Photoplethysmography (PPG) is a non-invasive, optical technique for measuring blood volume changes in peripheral tissues. In this paper, deep learning and ensemble methods are explored for an accurate estimation of heart rate directly from the PPG signals. Different models such as Convolutional Neural Networks (CNN), Long Short-Term Memory (LSTM), Hybrid CNN-LSTM, Siamese network, Extreme Gradient Boosting Trees (XGBoost), and hybrid models of XGBoost-CNN-LSTM, have been evaluated for their performance using the PPG-DaLiA dataset. The performance of these models were evaluated using mean absolute error (MAE), standard deviation (SD) and the Percentage of Data within an Error range (PDE). The results show that the proposed CNN-LSTM architecture is effective for practical heart rate prediction. The paper also provides valuable insights about the strengths and weaknesses of these models and their possible applications in wearable health monitoring.