Table 1 Performance comparison of deep learning models for street-level safety perception assessment

From: Vision AI reveals waste-safety perception relationships in urban environments

Model

Precision

Recall

F1 score

Accuracy

 

Safe

Unsafe

Safe

Unsafe

Safe

Unsafe

 

ResNet-50

0.739

0.757

0.753

0.743

0.746

0.750

0.748

MobileNet-V2

0.719

0.775

0.789

0.702

0.753

0.737

0.745

EffcientNet-B0

0.723

0.754

0.757

0.720

0.740

0.736

0.738

ShuffleNet-V2

0.723

0.754

0.757

0.720

0.740

0.736

0.678

Back to article page