Skip to main content

Table 2 Overview of the radiographic projections in which incorrect network predictions were observed

From: A deep learning approach for projection and body-side classification in musculoskeletal radiographs

True label

Prediction

r

True label

Prediction

r

True label

Prediction

r

True label

Prediction

r

Clavicle AP

Clavicle oblique

0.16

Hip AP

Lauenstein

0.04

Foot AP

Foot oblique

0.06

Dens

Cervical spine AP

0.03

Shoulder AP

0.02

Clavicle oblique

Clavicle AP

0.07

Lauenstein

Hip AP

0.02

Foot oblique

Foot AP

0.01

Lumbar spine AP

Thoracic spine AP

0.01

AC-joint

0.01

Forefoot oblique

0.03

AC-joint

Clavicle oblique

0.02

Knee AP

Knee lateral

0.01

Forefoot AP

Foot AP

0.04

Lumbar spine lateral

Lauenstein

0.01

Shoulder AP

0.02

Shoulder AP

AC-joint

0.04

Knee lateral

Elbow AP

0.01

Forefoot oblique

Foot oblique

0.09

 

Shoulder outlet

0.03

Radial head

Elbow AP

0.17

Ankle lateral

Calcaneus lateral

0.02

Big toe AP

Toe AP

0.05

Elbow lateral

0.03

Hand oblique

Hand AP

0.05

Calcaneus lateral

Ankle lateral

0.07

 

Thumb lateral

Thumb AP

0.05

   
  1. AP Anterior–posterior, r Relative proportion of misclassified radiographs within the test dataset