Table 3 Comparison between our model and other methods based on TuSimple dataset.
From: Efficient spatial and channel net for lane marker detection based on self-attention and row anchor
Model | Accuracy (%) | FP | FN |
|---|---|---|---|
ResNet-184 | 93.78 | 0.1035 | 0.0964 |
ResNet-344 | 94.66 | 0.0804 | 0.0775 |
LaneNet2 | 93.38 | 0.0780 | 0.0224 |
SCNN1 | 95.12 | 0.0610 | 0.0643 |
PolyLaneNet3 | 93.36 | 0.0942 | 0.0933 |
ERFNet6 | 94.34 | 0.0850 | 0.0777 |
ENet5 | 94.68 | 0.0977 | 0.0603 |
ESCN model based on Resnet-34(ours) | 95.49 | 0.0307 | 0.0342 |
Full Matrix Channel and Spatial attention based on ResNet-34(ours) | 95.30 | 0.0322 | 0.0368 |
