Evaluating the performance of convolutional neural networks with direct acyclic graph architectures in automatic segmentation of breast lesion in US images

Table 6 Metrics values for cross-validation with 5 folders for dataset B, using original images resized for 320 × 320 pixels

Folder	Global Accuracy	Accuracy	IoU	Weighted IoU	BF Score	Dice Coefficient
1	0.918	0.903	0.811	0.853	0.501	0.820
2	0.887	0.875	0.756	0.806	0.443	0.799
3	0.921	0.909	0.811	0.811	0.490	0.838
4	0.900	0.867	0.772	0.773	0.476	0.789
5	0.918	0.911	0.805	0.773	0.507	0.871
Mean	0.909	0.893	0.791	0.803	0.483	0.823
Standard Deviation	0.014	0.020	0.025	0.033	0.025	0.032

ISSN: 1471-2342