Table 5 Model evaluation on external dataset on CT images containing COVID-19 lesions and on whole CT volumes.
From: COVID-rate: an automated framework for segmentation of COVID-19 lesions from chest CT images
Method | Input data | Validation type | DSC | SPC | SEN | MAE |
|---|---|---|---|---|---|---|
Ref.40 | CTs with infection regions | Cross-validation | 0.831 | 0.993 | 0.867 | – |
\(\text {COVID-Rate}\) | CTs with infection regions | External validation | 0.794 | 0.9931 | 0.901 | 0.0087 |
Enhanced \(\text {COVID-Rate}\) | CTs with infection regions | External validation | 0.806 | 0.9933 | 0.914 | 0.0082 |
Ref.41 | Whole lung volumes | External validation | 0.597 | 0.977 | 0.865 | 0.033 |
\(\text {COVID-Rate}\) | Whole lung volumes | External validation | 0.787 | 0.9961 | 0.901 | 0.0049 |
Enhanced \(\text {COVID-Rate}\) | Whole lung volumes | External validation | 0.7998 | 0.9962 | 0.914 | 0.0046 |
