Table 2. The performance of the LDA model on the test folds.

Flavour	Balanced accuracy	F1 score	ROC AUC	mAP
Vanilla	0.67±0.08	0.79±0.08	0.73±0.12	0.88±0.09
TVC	0.77±0.06	0.85±0.07	0.83±0.09	0.92±0.08
Bins 1-to-100	0.73±0.08	0.82±0.09	0.78±0.06	0.92±0.04
V	0.68±0.08	0.79±0.09	0.65±0.14	0.84±0.10
VC	0.69±0.09	0.87±0.02	0.82±0.06	0.93±0.03

Data are presented as mean ± standard deviation. ROC, receiver operator characteristic; LDA, linear discriminant analysis; AUC, area under the curve; mAP, mean average precision; TVC, combination of image translation (T), segmentation volume adaptation (V) and contour randomization (C).