Related Experiment Video
Updated: May 12, 2026
12:54
Simultaneous Brightfield, Fluorescence, and Optical Coherence Tomographic Imaging of Contracting Cardiac Trabeculae Ex Vivo
Published on: October 2, 2021
Detecting multiple fiducial markers from a camera seismocardiogram.
Haozhe Li1, Dongmin Huang2, Lin Liu1
1Shandong University of Science and Technology, Qingdao, 266590, China.
Biomedical Optics Express
|May 11, 2026
Summary
This study introduces a new camera-based system and a deep learning model (DiAT) for analyzing seismocardiogram (SCG) signals. The method accurately detects key cardiac events, aiding in early cardiovascular event detection.
Area of Science:
- Biomedical Engineering
- Cardiovascular Monitoring
- Signal Processing
Background:
- Seismocardiogram (SCG) signals offer insights into cardiac contractions and valve function for early cardiovascular event detection.
- Accurate identification of SCG fiducial markers is essential for estimating hemodynamic parameters.
Purpose of the Study:
- To develop a novel framework for measuring SCG signals using a defocused camera-based speckle imaging system.
- To create a deep learning model (DiAT) for detecting multiple fiducial markers in SCG signals.
Main Methods:
- A defocused camera-based speckle imaging system was used to acquire SCG signals.
- A deep learning model, DiAT, employing dilated convolution and multi-head attention, was developed to detect seven fiducial markers (MC, IM, AO, IC, RE, AC, MO).
- Evaluation metrics included percentage of predictions within ±1 ms of true values and Mean Absolute Error (MAE).
Main Results:
- The DiAT model achieved 87.78% accuracy in detecting seven fiducial markers.
- The Mean Absolute Error (MAE) was 8.4106 ms.
- Experiments were conducted on 17 adult subjects using an ice water stimulation protocol.
Conclusions:
- The proposed camera-SCG measurement and DiAT deep learning model demonstrate reliable performance.
- This framework shows potential for accurate fiducial marker detection in SCG signals.
- The method supports enhanced monitoring for adverse cardiovascular events.
