Related Experiment Video
Updated: Jun 28, 2025
10:25
Brain Infarct Segmentation and Registration on MRI or CT for Lesion-symptom Mapping
Published on: September 25, 2019
48.0K
Multi-branch myocardial infarction detection and localization framework based on multi-instance learning and domain
Xinyue Li1, Yangcheng Huang1, Yixin Ning1
1School of Health Science and Engineering, University of Shanghai for Science and Technology, Shanghai, 200093, People's Republic of China.
Physiological Measurement
|April 10, 2024
Summary
This study introduces two models for detecting and localizing myocardial infarction (MI) using electrocardiograms (ECGs). The proposed methods show high accuracy, offering a promising tool for early cardiovascular disease diagnosis.
Area of Science:
- Cardiology
- Medical Imaging
- Artificial Intelligence
Background:
- Myocardial infarction (MI) causes irreversible heart damage, necessitating early detection.
- Automatic MI detection and localization from electrocardiograms (ECGs) are challenging clinical problems.
Purpose of the Study:
- To develop and evaluate novel deep learning models for accurate MI detection and localization.
- To improve the precision of MI diagnosis using ECG data.
Main Methods:
- Proposed two models: MFB-SENET for MI detection and MFB-DMIL for MI localization.
- MFB-DMIL integrates multi-instance learning with domain knowledge, handcrafted features, and a novel 'lead-loss' function.
- Utilized Grad-CAM for visualizing model decision-making processes.
Main Results:
- On the PTB database, MI detection accuracy reached 93.88% and localization accuracy reached 67.17%.
- On the PTB-XL database, MI detection accuracy was 94.89% and localization accuracy was 85.83%.
- Performance was comparable or superior to existing state-of-the-art algorithms.
Conclusions:
- The combined deep learning and domain knowledge approach is effective and reliable for MI diagnosis.
- The proposed models show potential as efficient tools to aid physicians in diagnosing MI.
- This work advances automated analysis of ECG for cardiovascular disease detection.
