Table 5.

Segmentation evaluation by mIoU and mDice for each combination of modules when divided by the ventricular systolic state.

U-Net	YOLO	ED	VGG	mIoU		mDice
				Systole	Diastole	Systole	Diastole
✓				0.1397 ± 0.0686	0.1631 ± 0.0528	0.2072 ± 0.0914	0.2388 ± 0.0677
✓	✓			0.5255 ± 0.0158	0.5491 ± 0.0114	0.6567 ± 0.0235	0.6882 ± 0.0146
✓	✓	✓		0.5413 ± 0.0196	0.5655 ± 0.0065	0.6733 ± 0.0186	0.7037 ± 0.0067
✓	✓	✓	✓	0.5435 ± 0.0102	0.5648 ± 0.0073	0.6755 ± 0.0127	0.7026 ± 0.0073

The values are the mean ± standard deviation of the three datasets for cross-validation. The diastolic group yielded higher values than the systolic group. The addition of the YOLO significantly improved the segmentation accuracy, and the addition of the ED further improved it, irrespective of the ventricular systolic state. The addition of the VGG contributed to the higher values in the systolic group.