Table 3 The comparison of the performance between the model, radiologist and radiologist-model collaboration.

From: Rib fracture detection system based on deep learning

Group

Model

Radiologist A

Radiologist B

Radiologist A-model collaboration

Radiologist B-model collaboration

F1-score

0.890

0.796

0.889

0.925

0.970

Recall

0.913

0.693

0.853

0.920

0.972

Precision

0.869

0.935

0.928

0.930

0.968

NPV

0.969

0.989

0.985

0.985

0.993

Time (seconds)

20 ± 5.8

242.6 ± 83.0*

153.6 ± 34.2*

207.0 ± 47.9a

58.6 ± 31.4a

  1. NPV negative predictive value.
  2. *Indicated the p value of the comparison between model and radiologists was < 0.001.
  3. aIndicated the p value of the comparison before and after using the model was < 0.001.
Back to article page