Table 3 Comparison of model testing performance metrics

From: Automatic defect detection in infrared thermal images of ancient polyptychs based on numerical simulation and a new efficient channel attention mechanism aided Faster R-CNN model

Model

\({\text{AP}}_{50}\)

\({\text{AP}}_{75}\)

\({\text{AP}}_{\text{S}}\)

\({\text{AP}}_{\text{L}}\)

\({\text{AR}}_{1}\)

\({\text{AR}}_{10}\)

\({\text{AR}}_{100}\)

Detection time

Memory consumption

Faster R-CNN

81.6

70.2

46.3

64.7

50.1

54.8

56.4

0.31

3200–7000

Improved modelling

87.3

72.2

54.8

66.5

52.2

57.7

57.8

0.33

3400–7500

  1. AP is in percent; time is in seconds; memory consumption is in MB