Stereo Depth Estimation on SUDS Combined
1.99EPEOurs (DEFOM ViT-L)
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| Ours (DEFOM ViT-L)Params. (M)=47.30, Backbone=ViT-L2026.01 | 1.99 | 46.51 | 4.85 | |
| Ours (DEFOM ViT-S)Params. (M)=18.51, Backbone=ViT-S2026.01 | 2.06 | 48.85 | 5.05 | |
| DEFOM (ViT-S)Params. (M)=18.51, Backbone=ViT-S2026.01 | 3.74 | 63.21 | 18.71 | |
| DEFOM (ViT-L)Params. (M)=47.30, Backbone=ViT-L2026.01 | 5.57 | 62.37 | 23.98 | |
| IGEV++Params. (M)=14.53, Backbone=No Vision Transformer2026.01 | 10.54 | 66.03 | 24.09 | |
| FoundationStereo (ViT-S)Params. (M)=37.55, Backbone=ViT-S2026.01 | 12.96 | 71.18 | 24.03 | |
| FoundationStereo (ViT-L)Params. (M)=39.20, Backbone=ViT-L2026.01 | 13.34 | 67.77 | 23.59 | |
| UnderwaterStereoParams. (M)=2.97, Backbone=No Vision Transformer2026.01 | 14.93 | 67.64 | 28.33 |