Related Experiment Video
Updated: Jun 13, 2026
08:26
Deep-Tissue Three-Photon Fluorescence Microscopy in Intact Mouse and Zebrafish Brain
Published on: January 13, 2022
Remote Photoplethysmography Using Triple-Head Spatio-Temporal Transformer with Reaction-Driven Gating and
Ahmed Mehrez1, Abdelwahab Alsammak1, Shady Y El-Mashad1
1Department of Electrical Engineering, Faculty of Engineering at Shoubra, Benha University, Cairo 11614, Egypt.
Sensors (Basel, Switzerland)
|June 12, 2026
Summary
This study introduces the Triple-Head Spatio-Temporal Transformer (TH-STT) for accurate non-contact heart rate monitoring using remote photoplethysmography (rPPG). The novel method effectively isolates physiological signals from environmental interference, improving performance on benchmark datasets.
Area of Science:
- Biomedical Engineering
- Computer Vision
- Signal Processing
Background:
- Remote photoplethysmography (rPPG) offers non-contact heart rate monitoring via facial videos.
- rPPG is susceptible to illumination variations and environmental noise, degrading signal accuracy.
- Existing methods struggle to effectively isolate physiological signals from interference.
Purpose of the Study:
- To develop a robust rPPG method resilient to environmental interference.
- To improve the accuracy and stability of non-contact heart rate estimation.
- To separate physiological signals from background noise and illumination variations.
Main Methods:
- Proposed the Triple-Head Spatio-Temporal Transformer (TH-STT) architecture.
- Utilized a background anchor token for environmental reference alongside facial tokens.
- Incorporated Reaction-Driven Gating (RDG) for facial muscular activity tracking and Dynamic Anchor Locking (DAL) for illumination interference cancellation.
Main Results:
- TH-STT demonstrated improved and stable performance across three benchmark datasets.
- Achieved a Mean Absolute Error (MAE) of 0.42 bpm on the UBFC-rPPG dataset.
- Achieved a Mean Absolute Error (MAE) of 1.08 bpm on the COHFACE dataset.
Conclusions:
- The TH-STT effectively separates rPPG signals from environmental interference.
- The proposed RDG and DAL mechanisms enhance the robustness of rPPG estimation.
- This approach offers a promising solution for accurate and stable non-contact heart rate monitoring.
