To develop an automated segmentation model for mouse cardiac slice images with myocardial infarction, addressing the complexities of irregular U-shaped structures.
Key Findings:
The proposed method achieved an average Dice coefficient of 80.68%, outperforming existing algorithms by 1.7%.
For infarct size segmentation, it reached 80.13%, surpassing the best current method by 2.43%.
The approach quantifies the ratio of infarcted to risk areas, aiding in the assessment of myocardial injury severity.
Interpretation:
Limitations:
The dataset consists of 243 images, which may limit generalizability.
The focus is primarily on mouse cardiac images, which may not directly translate to human applications.
Conclusion:
The DUCNet algorithm provides a method for automated quantitative analysis of infarct size in mouse cardiac slices.
