Monocular Depth Estimation on KITTI Improved Ground Truth 40
0.068Abs Rel ErrorDynamicDepth
Evaluation results
| Method | Links | |||||||
|---|---|---|---|---|---|---|---|---|
| DynamicDepthTraining=Monocular, WxH=640 x 1922022.03 | 0.068 | 0.362 | 3.454 | 0.111 | 94.3 | 99.1 | 99.8 | |
| ManyDepthTraining=Monocular, WxH=640 x 1922022.03 | 0.07 | 0.399 | 3.455 | 0.113 | 94.1 | 98.9 | 99.7 | |
| DORNTraining=Supervised, WxH=513 x 3852022.03 | 0.072 | 0.307 | 2.727 | 0.12 | 93.2 | 98.4 | 99.5 | |
| Packnet-SFMTraining=Monocular, WxH=640 x 1922022.03 | 0.078 | 0.42 | 3.485 | 0.121 | 93.1 | 98.6 | 99.6 | |
| Johnston et al.Training=Monocular, WxH=640 x 1922022.03 | 0.081 | 0.484 | 3.716 | 0.126 | 92.7 | 98.5 | 99.6 | |
| Wang et al.Training=Monocular, WxH=640 x 1922022.03 | 0.082 | 0.462 | 3.739 | 0.127 | 92.3 | 98.4 | 99.6 | |
| Monodepth2Training=Stereo, WxH=640 x 1922022.03 | 0.085 | 0.537 | 3.868 | 0.139 | 91.2 | 97.9 | 99.3 | |
| Patil et al.Training=Monocular, WxH=640 x 1922022.03 | 0.087 | 0.495 | 3.775 | 0.133 | 91.7 | 98.3 | 99.5 | |
| Kuznietsov et al.Training=Supervised, WxH=621 x 1872022.03 | 0.089 | 0.478 | 3.61 | 0.138 | 90.6 | 98 | 99.5 | |
| SuperDepthTraining=Stereo, WxH=1024 x 3842022.03 | 0.09 | 0.542 | 3.967 | 0.144 | 90.1 | 97.6 | 99.3 | |
| Monodepth2Training=Monocular, WxH=640 x 1922022.03 | 0.09 | 0.545 | 3.942 | 0.137 | 91.4 | 98.3 | 99.5 | |
| 3netTraining=Stereo, WxH=512 x 256, Backbone=ResNet 502022.03 | 0.102 | 0.675 | 4.293 | 0.159 | 88.1 | 96.9 | 99.1 | |
| MonodepthTraining=Stereo, WxH=512 x 2562022.03 | 0.109 | 0.841 | 4.568 | 0.166 | 87.7 | 96.7 | 98.8 | |
| 3netTraining=Stereo, WxH=512 x 256, Backbone=VGG2022.03 | 0.119 | 0.92 | 4.824 | 0.182 | 85.6 | 95.7 | 98.5 | |
| EPC++Training=Monocular, WxH=832 x 2562022.03 | 0.12 | 0.789 | 4.755 | 0.177 | 85.6 | 96.1 | 98.7 | |
| EPC++Training=Stereo, WxH=832 x 2562022.03 | 0.123 | 0.754 | 4.453 | 0.172 | 86.3 | 96.4 | 98.9 | |
| RanjanTraining=Monocular, WxH=832 x 2562022.03 | 0.123 | 0.881 | 4.834 | 0.181 | 86 | 95.9 | 98.5 | |
| DDVOTraining=Monocular, WxH=416 x 1282022.03 | 0.126 | 0.866 | 4.932 | 0.185 | 85.1 | 95.8 | 98.6 | |
| Zhan FullNYUTraining=Supervised, WxH=608 x 1602022.03 | 0.13 | 1.52 | 5.184 | 0.205 | 82.9 | 95.5 | 98.1 | |
| GeoNetTraining=Monocular, WxH=416 x 1282022.03 | 0.132 | 0.994 | 5.24 | 0.193 | 83.3 | 95.3 | 98.5 | |
| Vid2DepthTraining=Monocular, WxH=416 x 1282022.03 | 0.134 | 0.983 | 5.501 | 0.203 | 82.7 | 94.4 | 98.1 | |
| SfMLearnerTraining=Monocular, WxH=416 x 1282022.03 | 0.176 | 1.532 | 6.129 | 0.244 | 75.8 | 92.1 | 97.1 |