Bird's Eye View Detection on KITTI (val)
89.96APBEV (IoU=0.7, Easy)SECOND
Evaluation results
| Method | Links | ||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| SECONDSensor=LiDAR2021.08 | 89.96 | — | — | — | — | — | — | 87.07 | 79.66 | — | — | — | |
| SECOND (our teacher)Sensor=LiDAR2021.08 | 89.93 | — | — | — | 98.16 | 90.2 | 89.71 | 87.75 | 86.67 | — | — | — | |
| LIGA-StereoSensor=Stereo2021.08 | 89.35 | — | — | — | 97.22 | 90.27 | 88.36 | 77.26 | 69.05 | — | — | — | |
| PL++: P-RCNN + SL*Sensor=LiDAR, Note=utilizes 4-beam LiDAR to refine stereo depth estimation2021.08 | 88.2 | — | — | — | — | — | — | 76.9 | 73.4 | — | — | — | |
| CG-StereoSensor=Stereo2021.08 | 87.31 | — | — | — | 97.04 | 88.58 | 80.34 | 68.69 | 65.8 | — | — | — | |
| MV3D (LiDAR)Sensor=LiDAR2021.08 | 86.55 | — | — | — | — | — | — | 78.1 | 76.67 | — | — | — | |
| PLUME-LargeSensor=Stereo2021.08 | 84.7 | — | — | — | 91.3 | 86.6 | 81.6 | 71.1 | 65.1 | — | — | — | |
| DSGNSensor=Stereo2021.08 | 83.24 | — | — | — | — | — | — | 63.91 | 57.83 | — | — | — | |
| PL++: P-RCNNSensor=Stereo2021.08 | 82 | — | — | — | 89.8 | 83.8 | 77.5 | 64 | 57.3 | — | — | — | |
| OC-StereoSensor=Stereo2021.08 | 77.66 | — | — | — | 90.01 | 80.63 | 71.06 | 65.95 | 51.2 | — | — | — | |
| Disp-RCNNSensor=Stereo2021.08 | 77.63 | — | — | — | 90.67 | 80.45 | 71.03 | 64.38 | 50.68 | — | — | — | |
| PL: F-PointNetSensor=Stereo2021.08 | 72.8 | — | — | — | 89.8 | 77.6 | 68.2 | 51.8 | 44.3 | — | — | — | |
| Stereo-RCNNSensor=Stereo2021.08 | 68.5 | — | — | — | 87.13 | 74.11 | 58.93 | 48.3 | 41.47 | — | — | — | |
| AM3DTraining with depth modality=true2021.06 | 43.75 | — | — | — | — | — | — | 28.39 | 23.87 | — | — | — | |
| D4LCNTraining with depth modality=true2021.06 | 34.82 | — | — | — | — | — | — | 25.83 | 23.53 | — | — | — | |
| ImVoxelNetTraining with depth modality=false2021.06 | 31.67 | — | — | — | — | — | — | 23.68 | 19.73 | — | — | — | |
| MonoFENetTraining with depth modality=true2021.06 | 30.21 | — | — | — | — | — | — | 20.47 | 17.58 | — | — | — | |
| TLNetSensor=Stereo2021.08 | 29.22 | — | — | — | 62.46 | 45.99 | 41.92 | 21.88 | 18.83 | — | — | — | |
| M3D-RPNModality=Mono2021.03 | 26.86 | — | — | — | — | — | — | 21.15 | 17.14 | — | — | — | |
| M3D-RPNTraining with depth modality=false2021.06 | 25.94 | — | — | — | — | — | — | 21.18 | 17.9 | — | — | — | |
| RTM3DTraining with depth modality=false2021.06 | 25.56 | — | — | — | — | — | — | 22.12 | 20.91 | — | — | — | |
| MonoGRNetType=Mono, Inference Time (s)=0.062018.11 | 24.97 | 73.1 | 60.66 | 46.86 | 54.21 | 39.69 | 33.06 | 19.44 | 16.3 | — | — | — | |
| MonoDISModality=Mono2021.03 | 24.26 | — | — | — | — | — | — | 18.43 | 16.95 | — | — | — | |
| MonoDISTraining with depth modality=false2021.06 | 24.26 | — | — | — | — | — | — | 18.43 | 16.95 | — | — | — | |
| IAFAModality=Mono2021.03 | 22.75 | — | — | — | — | — | — | 19.6 | 19.21 | — | — | — | |
| MF3DType=Mono, Inference Time (s)=0.122018.11 | 22.03 | — | — | — | 55.02 | 36.73 | 31.27 | 13.63 | 11.6 | — | — | — | |
| BaselineModality=Mono2021.03 | 19.99 | — | — | — | — | — | — | 15.61 | 15.28 | — | — | — | |
| SMOKETraining with depth modality=false2021.06 | 19.99 | — | — | — | — | — | — | 15.61 | 15.28 | — | — | — | |
| ROI-10DModality=Mono2021.03 | 14.5 | — | — | — | — | — | — | 9.91 | 8.73 | — | — | — | |
| 3DOPType=Stereo, Inference Time (s)=4.22018.11 | 12.63 | 71.41 | 57.78 | 51.91 | 55.04 | 41.25 | 34.55 | 9.49 | 7.59 | — | — | — | |
| 3DOPSensor=Stereo2021.08 | 12.63 | — | — | — | 55.04 | 41.25 | 34.55 | 9.49 | 7.59 | — | — | — | |
| OFTNetModality=Mono2021.03 | 11.06 | — | — | — | — | — | — | 8.79 | 8.91 | — | — | — | |
| OFTNetTraining with depth modality=false2021.06 | 11.06 | — | — | — | — | — | — | 8.79 | 8.91 | — | — | — | |
| Mono3DType=Mono, Inference Time (s)=32018.11 | 5.22 | 32.76 | 25.15 | 23.65 | 30.5 | 22.39 | 19.16 | 5.19 | 4.13 | — | — | — | |
| Mono3DModality=Mono2021.03 | 5.22 | — | — | — | — | — | — | 5.19 | 4.13 | — | — | — | |
| CenterNetModality=Mono2021.03 | 4.46 | — | — | — | — | — | — | 3.23 | 3.53 | — | — | — | |
| F-ConvNetModality=R+L2020.06 | — | — | — | — | — | — | — | — | — | 90.23 | 88.79 | 86.84 | |
| F-PointnetModality=R+L2020.06 | — | — | — | — | — | — | — | — | — | 88.16 | 84.02 | 76.44 | |
| Fast PointRCNNModality=L2020.06 | — | — | — | — | — | — | — | — | — | 90.12 | 88.1 | 86.24 | |
| MV3DModality=R+L2020.06 | — | — | — | — | — | — | — | — | — | 86.55 | 78.1 | 76.67 | |
| Point-GNNModality=L2020.06 | — | — | — | — | — | — | — | — | — | 89.82 | 88.31 | 87.16 | |
| SECONDModality=L2020.06 | — | — | — | — | — | — | — | — | — | 89.96 | 87.07 | 79.66 | |
| STDModality=L2020.06 | — | — | — | — | — | — | — | — | — | 90.5 | 88.5 | 88.1 | |
| SVGA-NetModality=L2020.06 | — | — | — | — | — | — | — | — | — | 90.27 | 89.16 | 88.11 | |
| VoxelnetModality=L2020.06 | — | — | — | — | — | — | — | — | — | 89.6 | 84.81 | 78.57 |