Skip to main content

Table 6 A summary of data employed to perform single-view right/left segmentation model together with information on the training process

From: A fully automated deep learning pipeline for micro-CT-imaging-based densitometry of lung fibrosis murine models

Single-view models summary on multi-class (left lung/right lung)

 

Number of train µCT

Number of test µCT

Image dimension

Number of epochs

Training time [h]

Binarization threshold

Left lung DSC on test set

Right lung DSC on test set

Axial

46 (17,664 2D slices)

6 (820 2D slices)

(192, 120)

6

02:19:12

[0.45, 0.45]

0.952 ± 0.021

0.963 ± 0.043

Sagittal

46 (39,168 2D slices)

6 (1152 2D slices)

(192, 192)

9

02:53:39

[0.40, 0.40]

0.940 ± 0.031

0.950 ± 0.042

Coronal

46 (32,640 2D slices)

6 (960 2D slices)

(192, 120)

11

02:27:39

[0.40, 0.40]

0.957 ± 0.028

0.964 ± 0.027