| Model | EPE |
Configuration | Train Input Size |
|---|---|---|---|
| STTR | 0.40(0.43) | sttr_flyingthings3d.yaml | batchrandom |
| PSMNet | 0.95(1.09) | psmnet_sceneflow.yaml | 256x512 |
| CFNet | 0.96(1.04) | cfnet_sceneflow.yaml | 256x512 |
| AANet | 0.80(0.87) | aanet_sceneflow.yaml | 288x576 |
| MobileStereo-2D | 1.10(1.14) | msnet2d_sceneflow.yaml | 256x512 |
| MobileStereo-3D | 0.78(0.80) | msnet3d_sceneflow.yaml | 256x512 |
| GwcNet | 0.74(0.76) | gwcnet_sceneflow.yaml | 256x512 |
| COEX | 0.67(0.68) | coex_sceneflow_amp.yaml | 288x576 |
| FADNet++ | 0.65(0.76) | fadnet_sceneflow.yaml | 384x768 |
| CasStereo | 0.67(0.72) | casnet_psm_sceneflow.yaml | 256x512 |
| IGEV | 0.46(0.47) | igev_sceneflow_amp.yaml | 256x512 |
| StereoBase (Ours) | 0.34 | stereobase_sceneflow.yaml | 320x736 |
The results in parentheses are mentioned in the published papers. Underline refers to evaluation in the non-occluded regions only.
| Model | EPE |
Configuration | Train Input Size |
|---|---|---|---|
| LightStereo-S (Ours) | 0.73 | lightstereo_s_sceneflow.yaml | 256x512 |
| LightStereo-M (Ours) | 0.62 | lightstereo_m_sceneflow.yaml | 256x512 |
| LightStereo-L (Ours) | 0.59 | lightstereo_l_sceneflow.yaml | 256x512 |
| LightStereo-L* (Ours) | 0.51 | lightstereo_lx_sceneflow.yaml | 256x512 |
| StereoBase (Ours) | 0.34 | stereobase_sceneflow.yaml | 320x736 |
Access the checkpoint: googledrive
| Model | D1-bg |
D1-fg |
D1-all |
|---|---|---|---|
| PSMNet | 1.86 | 4.62 | 2.32 |
| PSMNet* | 1.80 | 4.58 | 2.26 |
| IGEV | 1.38 | 2.67 | 1.59 |
| IGEV* | 1.44 | 2.31 | 1.59 |
The model without * indicates the results of the original paper. The model with * means the implementation of OpenStereo.
| Model | D1-bg |
D1-fg |
D1-all |
|---|---|---|---|
| LightStereo-S (Ours) | 2.00 | 3.80 | 2.30 |
| LightStereo-M (Ours) | 1.81 | 3.22 | 2.04 |
| LightStereo-L (Ours) | 1.78 | 2.64 | 1.93 |
| LightStereo-L* (Ours) | 1.60 | 2.92 | 1.82 |
| StereoBase (Ours) | 1.28 | 2.26 | 1.44 |