Related Experiment Videos
Non-Contact Blood Pressure Prediction Using Radar with a Lightweight Temporal Multi-Scale Feature Fusion Network
Yuhan Liu1, Tianlin Zhang1, Yonggang Luo2
1School of Optoelectronic Information and Computer Engineering, University of Shanghai for Science and Technology, Shanghai 200093, China.
Sensors (Basel, Switzerland)
|June 12, 2026
Summary
A new lightweight network (LULMNet) enables accurate, non-contact blood pressure monitoring by effectively analyzing radar signals. This method shows promise for continuous hypertension screening and management.
Area of Science:
- Biomedical Engineering
- Signal Processing
- Artificial Intelligence
Background:
- Hypertension is a critical global health concern requiring continuous monitoring for early detection and intervention.
- Existing radar-based non-contact blood pressure methods struggle with multi-scale temporal feature utilization and model complexity.
Purpose of the Study:
- To develop a lightweight network, LULMNet, for accurate non-contact blood pressure prediction and waveform reconstruction.
- To address limitations in existing methods concerning temporal feature extraction and model complexity.
Main Methods:
- A two-stage training strategy using a lightweight one-dimensional U-Net (1D U-Net) for feature extraction and waveform reconstruction.
- Multi-scale feature fusion from the 1D U-Net encoder, followed by LSTM-based temporal modeling and regression for SBP and DBP estimation.
- Utilized Global Average Pooling (GAP) and a two-layer fully connected prediction head.
Main Results:
- Achieved high accuracy with errors of 3.21 ± 4.94 mmHg for SBP and 2.25 ± 3.39 mmHg for DBP, meeting the BHS Grade A standard.
- Obtained a low Normalized Mean Absolute Error (NMAE) of 0.044 for blood pressure waveform reconstruction.
- Maintained low model complexity with 3.0 M parameters and 0.37 G FLOPs.
Conclusions:
- LULMNet offers a computationally efficient and accurate solution for non-contact continuous blood pressure monitoring.
- The model demonstrates significant potential for practical application in hypertension screening and management.
- The proposed method effectively fuses multi-scale temporal features for improved prediction accuracy.
Related Concept Videos
01:19Assessing Blood pressure using a doppler ultrasound
To obtain accurate blood pressure measurements in clinical settings, especially when traditional methods are insufficient, healthcare professionals utilize the Doppler ultrasound technique. This method uses high-frequency sound waves to detect blood flow within the arteries, which is crucial for patients with conditions that complicate circulatory system assessment.
Pre-Procedural Guidelines for Doppler Ultrasound Blood Pressure Assessment:
Preparation of Equipment:
Pre-Procedural Guidelines for Doppler Ultrasound Blood Pressure Assessment:
Preparation of Equipment:
01:17Measurement of Blood Pressure
Assessing blood pressure is a standard procedure executed in virtually all medical environments. The method utilized today was established over a hundred years ago by an innovative Russian doctor, Dr. Nikolai Korotkoff. The soft ticking noise, known as Korotkoff sounds, heard while taking blood pressure readings results from turbulent blood flow within the vessels. The apparatus required for this procedure includes a sphygmomanometer, a blood pressure cuff attached to a gauge, and a stethoscope.