SOTAVerified

Road Segmentation

Road Segmentation is a pixel wise binary classification in order to extract underlying road network. Various Heuristic and data driven models are proposed. Continuity and robustness still remains one of the major challenges in the area.

Papers

Showing 1–10 of 82 papers

TitleStatusHype
RS-Mamba for Large Remote Sensing Image Dense PredictionCode3
Pix2Poly: A Sequence Prediction Method for End-to-end Polygonal Building Footprint Extraction from Remote Sensing ImageryCode2
Semi-Supervised Confidence-Level-based Contrastive Discrimination for Class-Imbalanced Semantic SegmentationCode2
Seeing the roads through the trees: A benchmark for modeling spatial dependencies with aerial imageryCode2
PhilEO Bench: Evaluating Geo-Spatial Foundation ModelsCode2
PREGAN: Pose Randomization and Estimation for Weakly Paired Image Style TranslationCode1
Multi-Modal Multi-Task (3MT) Road SegmentationCode1
Exploiting Low-level Representations for Ultra-Fast Road SegmentationCode1
Lidar Annotation Is All You NeedCode1
Enhancing Boundary Segmentation for Topological Accuracy with Skeleton-based MethodsCode1
Show:102550
← PrevPage 1 of 9Next →

Benchmark Results

#ModelMetricClaimedVerifiedStatus
1U-Net (ResNet-18)DWR46.5—Unverified
2DeepLabV3+ (ResNet-18)DWR46.1—Unverified
3U-Net (ResNet-50)DWR45.7—Unverified
4FCNDWR10.7—Unverified
#ModelMetricClaimedVerifiedStatus
1CoANet + PRNmIoU70.6—Unverified
2SPIN Road Mapper (ours)APLS0.74—Unverified
3D-LinkNetIoU0.64—Unverified
#ModelMetricClaimedVerifiedStatus
1RSM-SSIoU67.35—Unverified
2SPIN Road Mapper (ours)IoU65.24—Unverified