Research map / 3D vision / 3D generation
Photorealistic Object Insertion with Diffusion-Guided Inverse Rendering
Ruofan Liang, Zan Gojcic, Merlin Nimier-David, David Acuna, Nandita Vijaykumar, Sanja Fidler, Zian Wang
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 DiPIR instead of lighting estimation baselines for object insertion | DiPIR · PolyHaven (11 scenes) | RMSE: 0.056 (StyleLight) -> 0.048 | Improved |
| use DiPIR instead of the baselines on Waymo driving scenes | DiPIR · Waymo (48 scenes) | user study preference: 67.4% to 77.8% preferred across baselines | Improved |
| DiPIR versus NLFE on twilight scenes | DiPIR · Waymo twilight | user study preference: 44.4% | Hurt |
| add LoRA personalization of the diffusion model (vs SDS w/o LoRA) (ablation) | DiPIR · Waymo outdoor driving ablation | user preference for full method: 90.7% preferred | Improved |
| use LDS loss instead of original SDS loss (ablation) | DiPIR · Waymo outdoor driving ablation | user preference for full method: 74.1% preferred | Improved |
| add concept preservation class images to personalization (ablation) | DiPIR · Waymo outdoor driving ablation | user preference for full method: 64.8% preferred | Improved |
| add optimizable tone-mapping curve (ablation) | DiPIR · Waymo outdoor driving ablation | user preference for full method: 68.5% preferred | Improved |
| add two-environment-map fusion (ablation) | DiPIR · Waymo outdoor driving ablation | user preference for full method: 66.7% preferred | Improved |
| replace dataset-update photometric loss on SDEdit edits with LDS guidance (ablation) | DiPIR · Waymo outdoor driving ablation | user preference for full method: 85.2% preferred | Improved |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.
- StyleLight — StyleLight: HDR Panorama Generation for Lighting Estimation and Editing (ECCV 2022) · better on 2 of 2 shared results
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- Diffusion Renderer: Neural Inverse and Forward Rendering with Video Diffusion Models (CVPR 2025 · similarity 0.91)
- Channel-wise Noise Scheduled Diffusion for Inverse Rendering in Indoor Scenes (CVPR 2025 · similarity 0.89)
- Material Anything: Generating Materials for Any 3D Object via Diffusion (CVPR 2025 · similarity 0.88)
- Eclipse: Disambiguating Illumination and Materials using Unintended Shadows (CVPR 2024 · similarity 0.88)
- Inverse Rendering using Multi-Bounce Path Tracing and Reservoir Sampling (ICLR 2025 · similarity 0.88)
- Diffusion Reflectance Map: Single-Image Stochastic Inverse Rendering of Illumination and Reflectance (CVPR 2024 · similarity 0.88)
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.
