Table 6 Ablation study.

From: FasNet: a hybrid deep learning model with attention mechanisms and uncertainty estimation for liver tumor segmentation on LiTS17

Models

Accuracy

Dice Coefficient

Jaccard Index

Precision

Recall

Specificity

F1 Score

FasNet

0.9919

0.8659

0.8365

0.7931

0.6603

0.9976

0.7171

FasNet with Attention

0.9914

0.8693

0.8413

0.8009

0.7246

0.9794

0.7520

FasNet with Monte Carlo Dropout

0.9902

0.8676

0.8437

0.8298

0.8046

0.9684

0.7910

FasNet with Attention and Monte Carlo Dropout

0.9954

0.8766

0.8487

0.8499

0.8560

0.9695

0.8363

Back to article page