Fig. 9 | Scientific Reports

Fig. 9

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

Fig. 9

Visual examples of the segmentation evaluation performance using three cases: normal (1st row), abnormal with MI (2nd row), and abnormal with MI and MVO (3rd row). (a) represents the original MRI image superimposed by the ground-truth (GT) of the segmented regions (LV, Myo, MI, and MVO). (b) and (c) represent the prediction outcomes by U-Net and ResU-Net, respectively. (d) depicts the segmentation evaluation metrics for each region in terms of F1-score and MIoU. The quantitative and qualitative evaluation results show the superiority of ResU-Net, providing better details of each segmented region.

Back to article page