TABLE 3.

Sensitivity (%) at 2 and 4 false positives per image on the eight subsets of DeepLesion.

	LU		ME		LV		ST		PV		AB		KD		BN
	2	4	2	4	2	4	2	4	2	4	2	4	2	4	2	4
Faster R-CNN	71.83	76.62	72.06	77.13	71.80	77.96	66.67	70.95	71.12	75.88	71.43	76.25	54.83	60.62	46.79	51.72
Cascade R-CNN	72.03	76.77	72.18	78.01	71.97	78.15	67.12	71.24	71.35	75.99	72.11	76.85	55.21	61.37	48.12	52.19
YOLO-v3	70.25	75.69	70.12	76.16	69.83	75.42	65.57	70.18	68.96	74.75	69.36	74.87	54.37	59.30	46.03	50.72
RetinaNet	70.85	76.13	71.56	76.65	70.24	76.39	66.49	70.59	69.52	75.48	70.55	75.37	54.26	59.64	46.77	51.15
3DCE, 3slices	76.03	82.43	77.13	83.34	76.05	83.09	70.18	75.69	76.67	81.38	77.93	83.01	59.57	65.09	50.64	57.90
FPN+3DCE, 3slices	79.29	83.82	78.89	84.12	77.36	83.65	71.38	76.49	78.30	83.71	79.47	84.80	60.64	68.46	50.78	58.43
MVP-Net, 3slices	84.72	89.10	84.62	86.41	83.94	88.99	77.24	79.67	83.07	86.64	85.41	87.14	65.51	69.70	51.52	57.38
Ours, 3slices	87.13	91.66	87.68	91.77	88.59	91.53	77.89	81.38	85.12	87.52	86.53	91.59	65.09	68.46	63.36	67.33