List your company
ResearchFormerly WizWand

LiDAR Semantic Segmentation on SemanticSTF (val) (Adverse Weather Metrics)

53.2mIoU (D-fog)OmniLiDAR
13.721623.970834.2244.4692Jul 2, 2024Oct 23, 2024Feb 13, 2025Jun 7, 2025Sep 28, 2025Jan 19, 2026May 13, 2026
Updated 5mo ago

Evaluation results

MethodLinks
OmniLiDARBackbone=MinkUNet2026.0553.256.65955.357.9
sup-onlyBackbone=MinkUNet2026.0552.555.158.65456.2
OracleTraining source=SynLiDAR2024.0751.954.657.953.754.7
PointDR*‡Training source=SynLiDAR, Status=reproduced, Validation set for model selection=SynLiDAR sequence 02024.0721.4120.9425.4819.3120.47
PointDR*Training source=SynLiDAR, Status=reproduced2024.0719.0920.2825.2918.9819.78
Baseline SJ+LPDTraining source=SynLiDAR, Augmentation=Selective Jittering + Learnable Point Drop2024.0719.0820.6521.9717.2720.08
Baseline+SJ+LPD‡Training source=SynLiDAR, Augmentation=Selective Jittering + Learnable Point Drop, Validation set for model selection=SynLiDAR sequence 02024.0718.9921.2223.1417.2820.51
PolarMixTraining source=SynLiDAR2024.0716.4718.6919.6315.9818.09
LaserMixTraining source=SynLiDAR2024.0715.3217.9518.5513.816.85
BaselineTraining source=SynLiDAR2024.0715.2415.9716.8312.7615.45