Multitask Interactive Attention Learning Model Based on Hand Images for Assisting Chinese Medicine in Predicting
Qida Wang1, Chenqi Zhao1, Yan Qiang1
1College of Information and Computer, Taiyuan University of Technology, Taiyuan, China.
Insights
This study introduces a novel deep learning model, MTIALM, for noninvasive diagnosis of acute myocardial infarction (AMI) using palm images. The model shows improved accuracy in detecting key indicators, offering a convenient auxiliary diagnostic tool.
Area of Science:
- Cardiology
- Artificial Intelligence
- Traditional Chinese Medicine
Background:
- Acute myocardial infarction (AMI) is a critical cardiovascular disease with increasing incidence, particularly in younger populations.
- Current diagnostic methods for AMI are often invasive, complex, and carry risks of complications.
- There is a need for noninvasive and convenient auxiliary diagnostic approaches for AMI.
Purpose of the Study:
- To develop a noninvasive auxiliary diagnostic method for acute myocardial infarction (AMI) by integrating Traditional Chinese Medicine (TCM) principles with deep learning.
- To propose a novel deep learning model, MTIALM (Myocardial Infarction Traditional Chinese Medicine Palm Image), for predicting AMI based on palm features.
- To enhance diagnostic accuracy by utilizing shared networks, task-specific attention branches, and an information interaction module (IIM).
Main Methods:
- A depth model, MTIALM, was developed using traditional palm images for AMI detection.
- A shared network architecture was employed to learn comprehensive information across all diagnostic tasks.
- Task-specific attention branch networks and an information interaction module (IIM) were integrated to refine feature learning and information integration.
Main Results:
- The MTIALM model achieved an accuracy of 83.16% in detecting metacarpophalangeal joint swelling and 84.15% in identifying palmar thenar hypertrophy.
- These accuracies represent significant improvements compared to traditional classification methods for AMI auxiliary diagnosis.
- The model effectively learned and integrated features from different palm regions for improved prediction.
Conclusions:
- The proposed MTIALM model offers a promising noninvasive and convenient auxiliary diagnostic tool for acute myocardial infarction (AMI).
- Integrating deep learning with TCM palm diagnosis provides a valuable approach for early detection and risk assessment of AMI.
- Further research and validation are warranted to establish this method in clinical practice for cardiovascular disease management.
Abstract:
Acute myocardial infarction (AMI) is one of the most serious and dangerous cardiovascular diseases. In recent years, the number of patients around the world has been increasing significantly, among which people under the age of 45 have become the high-risk group for sudden death of AMI. AMI occurs quickly and does not show obvious symptoms before onset. In addition, postonset clinical testing is also a complex and invasive test, which may cause some postoperative complications. Therefore, it is necessary to propose a noninvasive and convenient auxiliary diagnostic method. In traditional Chinese medicine (TCM), it is an effective auxiliary diagnostic strategy to complete the disease diagnosis through some body surface features. It is helpful to observe whether the palmar thenar undergoes hypertrophy and whether the metacarpophalangeal joint is swelling in detecting acute myocardial infarction. Combined with deep learning, we propose a depth model based on traditional palm image (MTIALM), which can help doctors of traditional Chinese medicine to predict myocardial infarction. By building the shared network, the model learns information that covers all the tasks. In addition, task-specific attention branch networks are built to simultaneously detect the symptoms of different parts of the palm. The information interaction module (IIM) is proposed to further integrate the information between task branches to ensure that the model learns as many features as possible. Experimental results show that the accuracy of our model in the detection of metacarpophalangeal joints and palmar thenar is 83.16% and 84.15%, respectively, which are significantly improved compared with the traditional classification methods.
