. 2020 Oct 13;28:102464. doi: 10.1016/j.nicl.2020.102464

Table 9.

Comparison between the proposed deep learning two-stage approach against the latest studies in the literature on CMBs Detection.

Reference	Method	Data Subjects /CMBs	In-plane resolution( ${mm}^{2}$ )	Performance
Reference	Method	Data Subjects /CMBs	In-plane resolution( ${mm}^{2}$ )	Sen.^*(%)	${F P}_{a v g}$	Prec.^*(%)	Test Time/subject(sec)	Overall Sen. (%)
(Dou et al., 2016)	1st stage: 3D-FCN	3.0 T320/1149	0.45 × 0.45	98.29	282.8	$-$	64.35	91.45^♦
(Dou et al., 2016)	2nd stage: 3D-CNN	3.0 T320/1149	0.45 × 0.45	93.16	2.74	44.31	$-$	91.45^♦
(Liu et al., 2019)	1st stage: 3D-FRST	1.5 T and 3.0 T220/1641	0.45 × 0.57and 0.50 × 0.50	99.40	276.8	$-$	39	95.24^♦
(Liu et al., 2019)	2nd stage: 3D-ResNet	1.5 T and 3.0 T220/1641	0.45 × 0.57and 0.50 × 0.50	95.80	1.60	70.90	9	95.24^♦
(Chen et al., 2019)	1st stage: 2D-FRST	7.0 T73/2835	0.50 × 0.50	86.50	231.88	$-$	$-$	$-$
(Chen et al., 2019)	2nd stage: 3D-ResNet	7.0 T73/2835	0.50 × 0.50	94.69	11.58	71.98	$-$	$-$
(Kuijf et al., 2012)	3D-FRST	7.0 T18/66	0.35 × 0.35	71.20	17.17	13.20	900	71.20
(Wang et al., 2019)	2D-DenseNet	20/68847	$-$	97.78	11.8^♦	97.65	$-$	97.78
(Hong et al., 2019)	2D-ResNet-50	10/4287	$-$	95.71	3.4^♦	99.18^♦	$-$	95.71
Proposed Work	Implemented 3D-FRST	3.0 T107/572	LR 0.80 × 0.80	80.06	946.1	$-$	9.24	62.59
	3D-CNN on 3D-FRST candidates			78.17	4.69	41.63	0.961	62.59
	1st stage: YOLO			78.85	155.5	$-$	0.69	72.78
	2nd stage: 3D-CNN			91.80	1.89	67.21	0.159	72.78
	Implemented 3D-FRST	3.0 T72/188	HR 0.50 × 0.50	97.34	497.4	$-$	33.03	81.38
	3D-CNN on 3D-FRST candidates			83.61	3.74	36.26	0.505	81.38
	1st stage: YOLO			93.62	52.18	$-$	0.69	88.30
	2nd stage: 3D-CNN			94.32	1.42	61.94	0.053	88.30

Sen. and Prec. refer to the sensitivity and precision indices, respectively.

^♦

These values are not provided in the related articles, but were computed from other results.