Semantic Segmentation on DeepGlobeRoads
75Dice ScoreCEDice
Evaluation results
| Method | Links | |||
|---|---|---|---|---|
| CEDiceModel=Detectron2+DeepLabv3+2026.03 | 75 | 11.77 | 82 | |
| TverskyModel=Detectron2+DeepLabv3+, beta=0.72026.03 | 74 | 9.78 | 82 | |
| clDiceModel=Detectron2+DeepLabv3+, iter=452026.03 | 72 | 6.51 | 82 | |
| CEDice+CEDiceModel=Detectron2+DeepLabv3+2026.03 | 72 | 7.24 | 80 | |
| SkelRecallModel=Detectron2+DeepLabv3+2026.03 | 71 | 15.26 | 79 | |
| SCNP-CEDiceModel=Detectron2+DeepLabv3+, Method Variant=CEDice applied to SCNP2026.03 | 70 | 6.56 | 78 | |
| TopoLossModel=Detectron2+DeepLabv3+2026.03 | 22 | 27.79 | 33 |