3D Object Detection On Dair V2X I
평가 지표
AP|R40(easy)
AP|R40(hard)
AP|R40(moderate)
평가 결과
이 벤치마크에서 각 모델의 성능 결과
모델 이름 | AP|R40(easy) | AP|R40(hard) | AP|R40(moderate) | Paper Title | Repository |
|---|---|---|---|---|---|
| MonoUNI | 90.92 | 87.2 | 87.2 | MonoUNI: A Unified Vehicle and Infrastructure-side Monocular 3D Object Detection Network with Sufficient Depth Clues | |
| PointPillars | 63.1 | 54.0 | 54.0 | PointPillars: Fast Encoders for Object Detection from Point Clouds | - |
| BEVHeight | 77.8 | 65.9 | 65.8 | BEVHeight: A Robust Framework for Vision-based Roadside 3D Object Detection | - |
| MVXNet | 71.0 | 53.8 | 53.7 | MVX-Net: Multimodal VoxelNet for 3D Object Detection | - |
| CBR | 72.0 | 60.1 | 60.1 | Calibration-free BEV Representation for Infrastructure Perception | - |
| CoBEV | 82.0 | 69.7 | 69.6 | CoBEV: Elevating Roadside 3D Object Detection with Depth and Height Complementarity | - |
| BEVFormer | 61.4 | 50.7 | 50.7 | BEVFormer: Learning Bird's-Eye-View Representation from Multi-Camera Images via Spatiotemporal Transformers | - |
| BEVDepth | 75.7 | 63.7 | 63.6 | BEVDepth: Acquisition of Reliable Depth for Multi-view 3D Object Detection | - |
| ImVoxelNet | 44.8 | 37.6 | 37.6 | ImVoxelNet: Image to Voxels Projection for Monocular and Multi-View General-Purpose 3D Object Detection | - |
0 of 9 row(s) selected.