Skip to main content

Table 1 Accuracy metrics

From: Classification of chest radiographs using general purpose cloud-based automated machine learning: pilot study

Label

Number of images

Precision

Recall

Normal

305

70.8%

60.7%

Pathology

332

75.6%

75.6%

Artifact

303

55.3%

63.6%

Cardiomegaly

43

100%

0%

Collapse

21

100%

0%

Consolidation

84

100%

0%

Costophrenic_Angle_Blunted

31

100%

0%

Fibrosis

61

100%

0%

Hilar_Prominence

42

100%

0%

Midinspiratory

43

100%

0%

Nodular_Opacities

39

100%

0%

Other_Pathology

37

100%

0%

Pleural_Effusion

56

100%

0%

Prominent_BV_Markings

27

100%

0%

Rotated

165

33.3%

5.9%