Table 4.

Tenfold cross‐validation results for the whole pixel‐level bleeding detection dataset.

Color features	Texture features	Classifier	PREC	REC	SPEC	ACC	F1	MCC	ROC	PRC
None	All	RT	0.876	0.875	0.770	0.875	0.876	0.645	0.823	0.836
All	None	RT	0.938	0.938	0.883	0.938	0.938	0.824	0.911	0.912
All	All	RT	0.962	0.962	0.930	0.962	0.962	0.893	0.946	0.946
All	Five top	RT	0.963	0.963	0.931	0.963	0.963	0.894	0.947	0.946
All	Five different	RT	0.963	0.963	0.932	0.963	0.963	0.895	0.947	0.947
All	All	RF	0.976	0.976	0.959	0.976	0.976	0.931	0.997	0.997
All	Five top	RF	0.975	0.974	0.956	0.974	0.975	0.928	0.996	0.996
All	Five different	RF	0.976	0.975	0.958	0.975	0.975	0.930	0.997	0.996
All	All	LMT	0.973	0.973	0.954	0.973	0.973	0.922	0.995	0.994
All	Five top	LMT	0.970	0.970	0.945	0.970	0.970	0.914	0.994	0.993
All	Five different	LMT	0.973	0.972	0.954	0.972	0.972	0.922	0.995	0.994
‐	‐	ZeroR	0.598	0.773	0.227	0.773	0.674	0.000	0.500	0.649

PREC, precision; REC, sensitivity or recall; SPEC, specificity; F1, F‐Measure; ACC, accuracy; MCC, Matthews correlation coefficient; ROC, receiver operator characteristic curve; PRC, precision‐recall curve; RT, Random tree; LMT, logistic model tree.