Fig. 6

Segmentation results for myocardial infarction detection using the proposed model. Each row corresponds to a different case and displays three images: the original cine-MRI image showing the myocardial infarction (left), the expert-annotated ground truth segmentation mask (middle), and the model-predicted segmentation mask (right). The model accurately identifies infarct regions; however, minor differences in boundary sharpness compared to expert annotations are observed.