Table 11 Efficiency and parameters of different ConvNeXt backbones.

From: Transferring enhanced material knowledge via image quality enhancement and feature distillation for pavement condition identification

Backbone

Accuracy on different datasets (%)

Param(MB)

YouTube-w/o-ALI

YouTube-w-ALI

ConvNeXt-T

96.23

96.08

29

ConvNeXt-S

97.71

97.28

50

ConvNeXt-B

97.82

97.35

89

  1. “Param (MB)” indicates the number of parameters in M (million).
Back to article page