Drivable Area Detection On Bdd100K Val
評価指標
Params (M)
mIoU
評価結果
このベンチマークにおける各モデルのパフォーマンス結果
モデル名 | Params (M) | mIoU | Paper Title | Repository |
|---|---|---|---|---|
| TwinLiteNetPlus-Nano | 0.03 | 87.3 | TwinLiteNetPlus: A Stronger Model for Real-time Drivable Area and Lane Segmentation | |
| TwinLiteNetPlus-Large | 1.94 | 92.9 | TwinLiteNetPlus: A Stronger Model for Real-time Drivable Area and Lane Segmentation | |
| HybridNets | 12.8 | 90.5 | HybridNets: End-to-End Perception Network | |
| TwinLiteNetPlus-Medium | 0.48 | 92.0 | TwinLiteNetPlus: A Stronger Model for Real-time Drivable Area and Lane Segmentation | |
| YOLOP | 7.9 | 91.5 | YOLOP: You Only Look Once for Panoptic Driving Perception | |
| TwinLiteNet | 0.43 | 91.3 | TwinLiteNet: An Efficient and Lightweight Model for Driveable Area and Lane Segmentation in Self-Driving Cars | |
| TwinLiteNetPlus-Small | 0.12 | 90.6 | TwinLiteNetPlus: A Stronger Model for Real-time Drivable Area and Lane Segmentation | |
| A-YOLOM(s) | - | 91 | You Only Look at Once for Real-time and Generic Multi-Task | |
| YOLOPv2 | 38.9 | 93.2 | YOLOPv2: Better, Faster, Stronger for Panoptic Driving Perception |
0 of 9 row(s) selected.