Skip to main content

Table 4 Evaluation of pneumothorax segmentation performance with different growth rate k

From: Automated segmentation and diagnosis of pneumothorax on chest X-rays with fully convolutional multi-scale ScSE-DenseNet: a retrospective study

Growth rate mean (standard deviation)

PA

\(k=2\)

\(k=4\)

\(k=6\)

\(k=8\)

\(k=10\)

\(k=12\)

\(k=14\)

\(k=16\)

\({\mathrm {PA}}_1\)

0.81(0.32)*

0.84(0.28)*

0.84(0.28)*

0.85(0.27)

0.858(0.26)

0.859(0.27)

0.857(0.26)

0.85(0.27)

MPA

0.90(0.16)*

0.92(0.14)*

0.92(0.14)*

0.92(0.14)

0.928(0.13)

0.928(0.13)

0.927(0.13)

0.92(0.13)

DSC

\(k=2\)

\(k=4\)

\(k=6\)

\(k=8\)

\(k=10\)

\(k=12\)

\(k=14\)

\(k=16\)

DSC1

0.80(0.32)*

0.81(0.30)*

0.82(0.29)*

0.82(0.29)*

0.832(0.27)

0.838(0.27)

0.833(0.28)

0.83(0.30)

DSC

0.90(0.16)*

0.91(0.15)*

0.91(0.15)*

0.91(0.14)*

0.915(0.14)

0.918(0.14)

0.915 (0.14)

0.91(0.14)

  1. The values are provided in the form of mean (standard deviation)