Search research articles

ABOUT JoVE

Overview Leadership Blog JoVE Help Center

AUTHORS

Publishing Process Editorial Board Scope & Policies Peer Review FAQ Submit

LIBRARIANS

Testimonials Subscriptions Access Resources Library Advisory Board FAQ

RESEARCH

JoVE Journal Methods Collections JoVE Encyclopedia of Experiments Archive

EDUCATION

JoVE Core JoVE Business JoVE Science Education JoVE Lab Manual Faculty Resource Center Faculty Site

Terms & Conditions of Use

Search research articles

Related Experiment Video

Updated: May 14, 2026

BrainBeats as an Open-Source EEGLAB Plugin to Jointly Analyze EEG and Cardiovascular Signals

BrainBeats as an Open-Source EEGLAB Plugin to Jointly Analyze EEG and Cardiovascular Signals

Published on: April 26, 2024

Parameter-Efficient Deep Learning Models for Vital Sign Estimation From PPG.

Taha Samavati¹, Mahdi Farvardin¹, Aboozar Ghaffari²

¹Department of Computer Engineering Iran University of Science and Technology Tehran Iran.

Healthcare Technology Letters

|May 13, 2026

Summary

This summary is machine-generated.

Related Concept Videos

You might also read

Related Articles

Articles linked to this work by shared authors, journal, and citation graph.

Sort by

Same author

A High-resolution dataset for AI-driven segmentation and analysis of drug-treated breast tumor spheroids.

Computer methods and programs in biomedicine·2025

Same author

SCTD-ICA: A ICA-Based Approach for Fetal ECG Extraction From Single Channel Abdominal ECG.

IEEE journal of biomedical and health informatics·2025

Same author

ISRSL0: compressed sensing MRI with image smoothness regularized-smoothed [Formula: see text] norm.

Scientific reports·2024

Same author

AECG-DecompNet: abdominal ECG signal decomposition through deep-learning model.

Physiological measurement·2021

Same author

Image compression-encryption method based on two-dimensional sparse recovery and chaotic system.

Scientific reports·2021

Same author

Image Registration Based on Low Rank Matrix: Rank-Regularized SSD.

IEEE transactions on medical imaging·2017

Same journal

Driving Innovation: Transatlantic Attitudes to the <i>Bionics Bus</i> as a Vehicle for Health Transformation and STEM Engagement.

Healthcare technology letters·2026

Same journal

Gamified Digital Solutions for Tinnitus Health Literacy: The Erasmus+ Project TinWise.

Healthcare technology letters·2026

Same journal

Effect of Technology-Supported Measures Used for Care Transition Decisions for Chronic Disease Patients: A Systematic Review and Meta-Analysis.

Healthcare technology letters·2026

Same journal

Bibliometric Trends in the Integration of Computer Vision With Healthcare.

Healthcare technology letters·2026

Same journal

Machine Learning-Based Depression Recognition With Preserved Efficacy From Compressed EEG Signals Using Wavelet Transform and Adaptive Filtering.

Healthcare technology letters·2026

Same journal

Correction to 'In-Vivo Hand Mass Determination and an Anthropometric Investigation on Segment Length and Radius for Prosthetic Segment Design'.

Healthcare technology letters·2026

See all related articles

Researchers developed efficient deep learning models for estimating vital signs like heart rate (HR) and blood oxygen saturation (SpO2) using photoplethysmography (PPG) signals. These compact models offer accurate, generalizable, and deployable solutions for various sensors, including smartphones.

Area of Science:

Biomedical Engineering
Signal Processing
Machine Learning

Background:

Photoplethysmography (PPG) is a non-invasive technique for estimating vital signs.
Existing methods often require complex models or extensive preprocessing.
There is a need for efficient and accurate PPG-based vital sign estimation models.

Purpose of the Study:

To develop parameter-efficient, end-to-end deep learning models for estimating heart rate (HR), blood oxygen saturation (SpO2), and respiratory rate (RR) from PPG signals.
To evaluate the cross-dataset performance and generalizability of these models.
To demonstrate the feasibility of on-device deployment for real-time vital sign monitoring.

Main Methods:

Proposed a family of fully convolutional architectures, including baseline FCN, residual FCN, ConvNeXt-inspired FCN, and a compact DCT-based variant.

Keywords:

ML‐healthcare deep learning photoplethysmography signal processing vital sign estimation

Related Experiment Videos

Last Updated: May 14, 2026

BrainBeats as an Open-Source EEGLAB Plugin to Jointly Analyze EEG and Cardiovascular Signals

BrainBeats as an Open-Source EEGLAB Plugin to Jointly Analyze EEG and Cardiovascular Signals

Published on: April 26, 2024

Integrated signal preprocessing as a layer within the deep models.

Evaluated models on PPG-DaLiA and BIDMC datasets, and introduced the MTHS smartphone-based dataset.

Main Results:

Achieved strong cross-dataset performance with models using only 3.3K-53K parameters.
Reported MAE of 6.07 ± 2.70 bpm for HR on PPG-DaLiA.
On the BIDMC dataset, the best model achieved error rates of 2.2 bpm for HR, 3.14% for SpO2, and 1.49 breaths/min for RR.

Conclusions:

Compact, fully convolutional, and DCT-based designs provide accurate and generalizable PPG vital sign estimators.
The developed models are suitable for deployment across various sensors, including smartphones.
Demonstrated practical feasibility for on-device deployment using TFLite.