. Author manuscript; available in PMC: 2020 Aug 20.

Published in final edited form as: Mach Learn Med Imaging. 2018 Sep 15;11046:337–345. doi: 10.1007/978-3-030-00919-9_39

Table 2:

Ablation tests: testing performance comparison of different models (last row is our model). The comparison includes the Accuracy (Acc), F₂ score, Precision (Pre), and Recall (Rec) of all methods (Reg: Regularization, D/O: Drop-Out, Aug: Augmentation).

Model	Simple				Complex
Model	Acc%	F₂	Pre	Rec	Acc%	F₂	Pre	Rec
3D-CNN	68.7	0.71	0.68	0.72	66.5	0.69	0.67	0.70
3D-CNN+Reg	77.6	0.77	0.74	0.78	77.4	0.75	0.72	0.76
3D-CNN+Reg+D/O	83.1	0.811	0.78	0.82	79.7	0.82	0.79	0.84
3D-CNN+Reg+D/O+Aug (Ours)	94.1	0.93	0.92	0.94	88.3	0.89	0.88	0.91