Table 1 Comparison of model performance with different network depths.

From: Edge-guided multi-scale adaptive feature fusion network for liver tumor segmentation

Encoder/Decoder Layers

Dice

VOE

RVD

FLOPs(G)

Param(M)

5 layers

0.6743

0.4319

0.1747

45.64

134.33

4 layers

0.7184

0.3864

0.1238

39.42

34.08

3 layers

0.6974

0.4094

0.1255

31.81

14.31

  1. Significant values are in bold.
Back to article page