Breast Cancer Identification via Thermography Image Segmentation with a Gradient Vector Flow and a Convolutional Neural Network

Table 2

Quantitative classification results (%).

Technique	External quality indicators
	TPR	PPV	FDR	F1 or HM	SPC	NPV	FPR	ACC	AUC
CNN	100	100	0	100	100	100	0	100	100
TRF	85.71	85.71	14.28	86.95	85.71	85.71	17.85	85.71	85.71
MLP	80	80	20	88.88	100	100	25	88.88	100
NV	82.85	82.85	17.14	82.85	78.57	78.57	21.42	80.95	78.57

ACC: accuracy; AUC: area under the receiver operating characteristic curve; F1: F1 score; FDR: false discovery rate; FPR: fall-out or false-positive rate; HM: harmonic mean; NPV: negative predictive value; PPV: precision or positive predictive value; SPC: specificity or true-negative rate; TPR: sensitivity, recall, or true-positive rate.