CARLANE: A Lane Detection Benchmark for Unsupervised Domain Adaptation from Simulation to multiple Real-World Domains

Open paper
Benchmark Model Rank Results
domain-adaptation-on-molaneUFLD-SGPCS-ResNet18#1Lane Accuracy (LA): 93.94
domain-adaptation-on-molaneUFLD-SGADA-ResNet18#2Lane Accuracy (LA): 93.82
domain-adaptation-on-molaneUFLD-SGPCS-ResNet32#3Lane Accuracy (LA): 93.53
domain-adaptation-on-molaneUFLD-SGADA-ResNet32#4Lane Accuracy (LA): 93.31
domain-adaptation-on-molaneUFLD-ADDA-ResNet18#5Lane Accuracy (LA): 92.85
domain-adaptation-on-molaneUFLD-ADDA-ResNet32#6Lane Accuracy (LA): 92.39
domain-adaptation-on-molaneUFLD-DANN-ResNet32#7Lane Accuracy (LA): 90.91
domain-adaptation-on-molaneUFLD-DANN-ResNet18#8Lane Accuracy (LA): 87.65
domain-adaptation-on-mulaneUFLD-SGADA-ResNet32#1Lane Accuracy (LA): 91.63
domain-adaptation-on-mulaneUFLD-SGPCS-ResNet18#2Lane Accuracy (LA): 91.57
domain-adaptation-on-mulaneUFLD-SGPCS-ResNet32#3Lane Accuracy (LA): 91.55
domain-adaptation-on-mulaneUFLD-SGADA-ResNet18#4Lane Accuracy (LA): 90.71
domain-adaptation-on-mulaneUFLD-ADDA-ResNet32#5Lane Accuracy (LA): 90.22
domain-adaptation-on-mulaneUFLD-ADDA-ResNet18#6Lane Accuracy (LA): 89.83
domain-adaptation-on-mulaneUFLD-DANN-ResNet32#7Lane Accuracy (LA): 88.76
domain-adaptation-on-mulaneUFLD-DANN-ResNet18#8Lane Accuracy (LA): 86.01
domain-adaptation-on-tulaneUFLD-SGPCS-ResNet32#1Lane Accuracy (LA): 93.29
domain-adaptation-on-tulaneUFLD-SGADA-ResNet32#2Lane Accuracy (LA): 92.04
domain-adaptation-on-tulaneUFLD-SGADA-ResNet18#3Lane Accuracy (LA): 91.70
domain-adaptation-on-tulaneUFLD-SGPCS-ResNet18#4Lane Accuracy (LA): 91.55
domain-adaptation-on-tulaneUFLD-ADDA-ResNet32#5Lane Accuracy (LA): 91.39
domain-adaptation-on-tulaneUFLD-DANN-ResNet32#6Lane Accuracy (LA): 91.06
domain-adaptation-on-tulaneUFLD-ADDA-ResNet18#7Lane Accuracy (LA): 90.72
domain-adaptation-on-tulaneUFLD-DANN-ResNet18#8Lane Accuracy (LA): 88.74

Rank counts only results with a code link.