Pulmonary Nodule Detection from 3D CT Image with a Two-Stage Network

Table 5

Comparison of different methods on parameter numbers, FLOPs, and inference.

Methods	#Parameters (M)	FLOPs (T)	Inference time (s)
NoduleNet (2020) [26]	50.19	0.28	0.93
S4ND (2019) [14]	4.58	0.17	0.84
DBResNet (2020) [13]	7.4	0.17	0.85
V-Net (2020) [15]	65.95	4.31	1.73
H-DenseUNet (2018) [22]	40	1.18	1.27
Swin-UNetr (2021) [27]	61.98	0.68	1.02
UNetr (2022) [28]	102.2	4.48	1.74
Ours	93.32	1.09	1.27