| Task Name | Dataset Name | SOTA Result | Trend | |
|---|---|---|---|---|
| Novel View Synthesis | RealEstate10K | PSNR32.89 | 116 | |
| Novel View Synthesis | RealEstate10K Hard | PSNR24.18 | 20 | |
| Novel View Synthesis | RealEstate10K Easy | PSNR26.54 | 20 | |
| Few-view 3D Reconstruction | RealEstate10K (test) | PSNR32.2 | 20 | |
| Novel View Synthesis | RealEstate10K v1 (hold-out scenes) | PSNR30.72 | 16 | |
| Novel View Synthesis | RealEstate10K t=5 (test) | LPIPS0.038 | 16 | |
| Scene-level View Synthesis | RealEstate10k (val) | PSNR29.86 | 15 | |
| Multi-view pose regression | RealEstate10K | mAA(30)79.9 | 15 | |
| Novel View Synthesis | RealEstate10K Medium | PSNR19.5346 | 14 | |
| Novel View Synthesis | RealEstate10K (RE10K) t=10 (test) | LPIPS0.049 | 14 | |
| Single-view Novel View Synthesis | RealEstate10K Long-term, 200th frame 84 (test) | PSNR17.13 | 13 | |
| Single-view Novel View Synthesis | RealEstate10K Short-term, 50th frame 84 (test) | PSNR20.32 | 13 | |
| Camera Pose Estimation | RealEstate10K (unseen) | AUC@3093.5 | 12 | |
| Novel View Synthesis | RealEstate10K 80 (test) | PSNR25.845 | 10 | |
| Relative Camera Pose Estimation | RealEstate10K (test) | AUC @ 5 deg69.1 | 10 | |
| 3D Camera-Controlled Video Synthesis | RealEstate10K (unseen camera trajectories) | TransErr0.358 | 9 | |
| Video Generation | RealEstate10K (RE10K) (test) | PSNR23.77 | 8 | |
| Stereo Video Synthesis | RealEstate10K (test) | FVD67.09 | 8 | |
| Novel View Synthesis | RealEstate10K 36 view | PSNR30.25 | 6 | |
| Novel View Synthesis | RealEstate10K 24 view | PSNR29.987 | 6 | |
| Novel View Synthesis | RealEstate10K 12 view | PSNR28.552 | 6 | |
| 3D Scene Generation | RealEstate10K (RE10K) (test) | PSNR27.57 | 6 | |
| Video Generation | RealEstate10K >=256 frames (test) | PSNR13.89 | 6 | |
| Video Generation | RealEstate10K 0~200 frames (test) | PSNR14.57 | 6 | |
| Video Generation | RealEstate10K 0~128 frames (test) | PSNR15.91 | 6 |