DCA-YOLOv8: A New Framework for Detecting Coronary Artery Stenosis
Author Information
Author(s): Duan Hualin, Yi Sanli, Ren Yanyou
Primary Institution: Kunming University of Science and Technology
Hypothesis
Can a new deep-learning framework improve the detection of coronary artery stenosis?
Conclusion
The DCA-YOLOv8 framework significantly improves the accuracy of coronary artery stenosis detection compared to existing algorithms.
Supporting Evidence
- The DCA-YOLOv8 framework achieved a precision of 96.62%, recall of 95.06%, F1-score of 95.83%, and mean average precision of 97.6%.
- The framework outperformed existing object detection algorithms in coronary artery stenosis detection.
- The AICI loss function improved convergence speed and accuracy for small target detection.
Takeaway
This study created a new computer program that helps doctors find narrow spots in heart arteries more accurately using special images.
Methodology
The study used a deep-learning framework that includes preprocessing, feature extraction, and a detection head with a new loss function.
Limitations
The framework can only detect the presence of stenosis and cannot classify the types of stenosis or identify arteries with severe blockage.
Participant Demographics
Patients with confirmed single-branch coronary artery disease.
Statistical Information
P-Value
p<0.05
Statistical Significance
p<0.05
Digital Object Identifier (DOI)
Want to read the original?
Access the complete publication on the publisher's website