Table 5 Efficiency analysis of comparative models on the M3FD dataset.

From: A dual-stream feature decomposition network with weight transformation for multi-modality image fusion

Backbone

Methods

Size(M)

Time(s)

CNN

IFCNN

0.084

0.104±0.058

U2Fusion

0.659

0.128±0.034

SeAFusion

0.167

0.062±0.013

SuperFusion

1.962

0.735±0.134

TarDAL

0.297

0.102±0.018

SDDGAN

3.186

0.738±0.296

Transformer

SwinFusion

0.974

7.087±1.328

CoCoNet

9.131

7.191±1.473

Ours

1.095

2.361±0.440

Diffusion Model

Diff-IF

23.74

5.062±0.535

Back to article page