Related Experiment Video
Updated: Apr 28, 2026
Quantification of Mouse Heart Left Ventricular Function, Myocardial Strain, and Hemodynamic Forces by Cardiovascular Magnetic Resonance Imaging
Published on: May 24, 2021
Multimodal Phase-Space Dynamics Fusion for Robust Ischemia Screening: An Edge-AI Paradigm with SERF
Keyi Li1,2, Xiangyang Zhou1,3, Yifan Jia1,2
1Key Laboratory of Ultra-Weak Magnetic Field Measurement Technology, Ministry of Education, School of Instrumentation and Optoelectronic Engineering, Beihang University, Beijing 100191, China.
This study introduces a novel Edge-AI framework for detecting myocardial ischemia (MI) using quantum sensing and computer vision. The approach significantly improves diagnostic sensitivity for early MI detection in portable screening scenarios.
Area of Science:
- Cardiovascular diagnostics
- Quantum sensing
- Artificial Intelligence in Medicine
Background:
- Myocardial ischemia (MI) poses a significant global health burden, necessitating advanced detection methods.
- Spin-Exchange Relaxation-Free (SERF) magnetocardiography (MCG) offers high sensitivity but faces computational challenges on edge devices.
- Current limitations hinder the clinical deployment of sensitive MCG for real-time MI diagnosis.
Purpose of the Study:
- To develop a computationally efficient Edge-AI framework for MI detection using SERF-MCG.
- To integrate quantum sensing with computer vision for enhanced diagnostic capabilities.
- To overcome computational bottlenecks for portable, real-time MI screening.
Main Methods:
- A "Sensor-to-Image" framework converting single-channel SERF-MCG signals into phase-space images (RP, GASF, MTF).
- Utilizing a streamlined MobileNetV3-Small architecture for low-latency image analysis.
- Implementing an adaptive weighted fusion mechanism combining Recurrence Plots and Gramian Angular Summation Fields for improved accuracy.
Main Results:
- The fusion model achieved an Area Under the Curve (AUC) of 0.865, outperforming baseline models.
- Achieved high sensitivity (88.3%) for MI detection, prioritizing false negative reduction.
- Demonstrated an average inference time of 4.7 ms, indicating suitability for real-time Point-of-Care applications.
Conclusions:
- Multi-view phase-space image fusion effectively captures subtle ischemic changes.
- The lightweight AI framework supports portable SERF-MCG systems with embedded screening capabilities.
- This approach holds promise for advancing pre-hospital and Point-of-Care MI diagnostics.
