Research map / 3D vision / 3D generation
OLATverse: A Large-scale Real-world Object Dataset with Precise Lighting Control
Xilong Zhou, Jianchun Chen, Pramod Rao, Timo Teufel, Linjie Lyu, Tigran Minasian, Oleksandr Sotnychenko, Xiao-Xiao Long, Marc Habermann, Christian Theobalt
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| use GS3 instead of Mitsuba with Metashape meshes for inverse rendering | inverse rendering baselines · Inverse rendering (OLATverse validation, 42 objects) | PSNR: 35.906 -> 38.538 | Improved |
| use GS3 instead of Mitsuba+Mshape | inverse rendering baselines · Inverse rendering (OLATverse validation, 42 objects) | LPIPS: 0.0260 -> 0.0263 | No clear effect |
| use GS3 instead of BiGS | inverse rendering baselines · Inverse rendering (OLATverse validation, 42 objects) | PSNR: 32.98 -> 38.538 | Improved |
| use GS3 instead of RNG | inverse rendering baselines · Inverse rendering (OLATverse validation, 42 objects) | SSIM: 0.962 -> 0.982 | Improved |
| use normal-specific models (StableNormal) instead of relighting-oriented RGBX | normal estimation · Normal estimation (OLATverse validation, 4 illuminations) | mean angular error: 51.95 -> 31.85 | Improved |
| use StableNormal instead of DiffusionRender | normal estimation · Normal estimation (OLATverse validation, 4 illuminations) | mean angular error: 34.88 -> 31.85 | Improved |
| use GeoWizard instead of StableNormal | normal estimation · Normal estimation (OLATverse validation, 4 illuminations) | accuracy within 11.25 deg: 8.93 -> 10.98 | Mixed |
| use polarized gradient illumination for normals instead of non-polarized | OLATverse normal extraction · OLATverse normals | normal accuracy | Improved |
| combine bgMatting, SAM and RMBG-2.0 instead of any one alone (ablation) | mask segmentation · OLATverse masks | mask quality: 95% | Improved |
Closest papers by meaning
- OpenIllumination: A Multi-Illumination Dataset for Inverse Rendering Evaluation on Real Objects (NeurIPS 2023 · similarity 0.89)
- Objaverse-XL: A Universe of 10M+ 3D Objects (NeurIPS 2023 · similarity 0.88)
- ICTPolarReal: A Polarized Reflection and Material Dataset of Real World Objects (CVPR 2026 · similarity 0.87)
- LRM-Zero: Training Large Reconstruction Models with Synthesized Data (NeurIPS 2024 · similarity 0.84)
- From an Image to a Scene: Learning to Imagine the World from a Million 360° Videos (NeurIPS 2024 · similarity 0.84)
- IDArb: Intrinsic Decomposition for Arbitrary Number of Input Views and Illuminations (ICLR 2025 · similarity 0.84)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
