Table 7 Model complexity comparison on a 1283 input

From: Structure-aware multi-task learning with domain generalization for robust vertebrae analysis in spinal CT

Method

#Params (M)

FLOPs (G)

Memory (GB)

nnU-Net7

48.0

350.0

12.5

UNETR19

92.1

380.5

15.2

TransBTS46

33.4

333.2

12.0

H2Former22

102.7

410.3

16.8

Scribformer23

88.9

365.7

14.7

Dense-U-Net6

55.6

321.4

6.1

Tao et al.49

36.1

299.8

5.2

VertebraFormer (ours)

85.6

342.4

13.9

  1. #Params and FLOPs are reported in millions (M) and gigaFLOPs (G), respectively43.
Back to article page