Skip to main content

Table 3 Comparison of Dice coefficient of 3D Res U-Net and state-of-the-art methods

From: Design of lung nodules segmentation and recognition algorithm based on deep learning

Models

Dice

Lung mass

Nodule

Small nodule

Micro nodule

FCN-8 s

0.759

0.327

0.212

0.159

2D PSPNet

0.718

0.593

0.447

0.144

2D Res U-Net

0.829

0.731

0.536

0.208

3D U-Net

0.911

0.698

0.588

0.185

MSS U-Net

0.846

0.685

0.415

0.209

3D Res U-Net

0.910

0.805

0.652

0.466

  1. Bold text represents the highest Dice obtained by different models when segmenting lung nodules of different sizes
Back to article page

BMC Bioinformatics

ISSN: 1471-2105

Contact us