Quantitative results from CULane benchmark
Method | Backbone | F1@50 |
---|---|---|
SCNN [10] | VGG16 | 90.60 |
RESA [11] | ResNet50 | 92.10 |
UFLD [15] | ResNet34 | 90.70 |
CondLane [5] | ResNet101 | 93.47 |
CLRNet [6] | DLA34 | 93.73 |
PolyLaneDet | CSPDarkNet53 | 84.29 |
The metric is F1@50 of the Normal category in the CULane benchmark.