Fig. 8 | Scientific Reports

Fig. 8

From: A hybrid segmentation and classification CAD framework for automated myocardial infarction prediction from MRI images

Fig. 8

The convergence rates of both U-Net and ResU-Net segmentation models during the training and validation process for Fold 3 are assessed based on (a) F1-score, (b) MIoU, and (c) loss functions. The convergence curves clearly demonstrate that the AI models began with lower values and progressively improved, ultimately reaching optimal learning. This indicates that the AI models do not encounter overfitting issues.

Back to article page