Skip to main content

Table 2 Quantitative parameters for lung lobe segmentation

From: A 2D–3D hybrid convolutional neural network for lung lobe auto-segmentation on standard slice thickness computed tomography of patients receiving radiotherapy

N=35

LUL

LLL

RUL

RML

RLL

HD95 (mm)

22.3584±17.2096

20.9913±7.1894

16.9986±7.8134

26.553±13.995

23.4818±11.1656

MSD (mm)

0.9754±0.2355

1.2095±0.3613

1.1752±0.2935

1.9358±0.7122

1.2164±0.5285

DSC

0.9579±0.0125

0.9479±0.0157

0.9507±0.0133

0.9003±0.0331

0.9484±0.0225

Accuracy

99.5715±0.0928

99.5951±0.1209

99.6668±0.0892

99.7161±0.0707

99.5753±0.1421

Sensitivity

98.2261±0.6801

96.1441±1.5422

96.1279±1.7629

92.3785±3.6881

96.0335±2.0398

Specificity

99.6506±0.0652

99.7638±0.0603

99.8104±0.0756

99.8346±0.0762

99.7793±0.0638

  1. LUL left upper lobe, LLL left lower lobe, RUL right upper lobe, RML right middle lobe, RLL right lower lobe.