Table 3 The test results of the compare study in ETIS-LaribPolypDB and CVC-ColonDB.

From: Dual ensemble system for polyp segmentation with submodels adaptive selection ensemble

Train/Val

Kvasir and CVC-ClinicDB 90%/10%

Test

ETIS-LaribPolypDB

CVC-ColonDB

Metric

Dice

mDice

mIoU

p

Dice

mDice

mIoU

p

UNet

0.668

0.798

0.741

< 0.05

0.704

0.842

0.775

< 0.05

UperNet

0.671

0.804

0.774

< 0.05

0.714

0.833

0.786

< 0.05

Double-UNet

0.633

0.795

0.760

< 0.05

0.631

0.801

0.743

< 0.05

SS-Former

0.802

0.894

0.856

< 0.05

0.789

0.882

0.836

0.18

FCB-Former

0.803

0.899

0.861

0.09

0.792

0.887

0.839

0.42

ESFPNet

0.793

0.891

0.849

< 0.05

0.794

0.892

0.843

0.51

Nanni’s Ens

0.806

0.903

0.863

0.32

0.795

0.891

0.843

0.56

HarDNet-DFUS

0.748

0.862

0.821

< 0.05

0.746

0.861

0.817

< 0.05

DET-Former

0.825

0.910

0.873

_

0.798

0.894

0.851

_

Back to article page