Lane Detection On Tusimple

Metrics

Accuracy

F1 score

Results

Performance results of various models on this benchmark

			Paper Title
SCNN_UNet_Attention_PL*	98.38	-	Robust Lane Detection through Self Pre-training with Masked Sequential Autoencoders and Fine-tuning with Customized PolyLoss
PE-RESA	96.93	-	Lane detection with Position Embedding
FOLOLane(ERFNet)	96.92	-	Focus on Local: Detecting Lane Marker from Bottom Up via Key Point
CLRNet(ResNet-34)	96.9%	97.82	CLRNet: Cross Layer Refinement Network for Lane Detection
CLLD	96.82	-	Contrastive Learning for Lane Detection via cross-similarity
CLRNet(ResNet-18)	96.82%	97.89	CLRNet: Cross Layer Refinement Network for Lane Detection
RESA	96.82	96.93	RESA: Recurrent Feature-Shift Aggregator for Lane Detection
CANet-L(ResNet101)	96.76%	97.77	CANet: Curved Guide Line Network with Adaptive Decoder for Lane Detection
CANet-M	96.66%	97.44	CANet: Curved Guide Line Network with Adaptive Decoder for Lane Detection
ENet-SAD	96.64%	95.92	Learning Lightweight Lane Detection CNNs by Self Attention Distillation
HarD-SP	96.58%	96.38	Towards Lightweight Lane Detection by Optimizing Spatial Embedding
CANet-S	96.56%	97.51	CANet: Curved Guide Line Network with Adaptive Decoder for Lane Detection
CondLaneNet-L(ResNet-101)	96.54%	97.24	CondLaneNet: a Top-to-down Lane Detection Framework Based on Conditional Convolution
Pairwise pixel supervision + FCN	96.50%	94.31	Learning to Cluster for Proposal-Free Instance Segmentation
Oblique Convolution	96.50%	97.42	-
EL-GAN	96.40%	96.26	EL-GAN: Embedding Loss Driven Generative Adversarial Networks for Lane Detection
LaneNet	96.4%	94.80	Towards End-to-End Lane Detection: an Instance Segmentation Approach
Discriminative loss function	96.40%	-	Semantic Instance Segmentation with a Discriminative Loss Function
ENet-Label	96.29%	95.23	Agnostic Lane Detection
R-34-E2E	96.22%	96.58	End-to-End Lane Marker Detection via Row-wise Classification

0 of 41 row(s) selected.

Command Palette

Lane Detection On Tusimple

Metrics

Results