Research map / Point clouds / Registration and matching
Diff2I2P: Differentiable Image-to-Point Cloud Registration with Diffusion Prior
Juncheng Mu, Chengwei Ren, Weixiang Zhang, Liang Pan, Xiao-Ping Zhang, Yue Gao
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 |
|---|---|---|---|
| add CSD loss (diffusion score distillation) (ablation) | Diff2I2P · 7-Scenes | Registration Recall (%): 81.3 -> 83.0 | Improved |
| use overlap mask in CSD loss (ablation) | Diff2I2P · 7-Scenes | Registration Recall (%): 81.7 -> 83.0 | Improved |
| add augmented circle loss instead of plain circle loss (ablation) | Diff2I2P · 7-Scenes | Registration Recall (%): 82.5 -> 83.0 | Improved |
| add transformation loss (MSE of projected points) (ablation) | Diff2I2P · 7-Scenes | training stability: Gradient Explosion | Hurt |
| use Diff2I2P instead of 2D3D-MATR | 2D3D-MATR · 7-Scenes | Registration Recall (%): 75.8 -> 83.0 | Improved |
| distill diffusion features into backbone instead of using diffusion features at inference (vs FreeReg) | FreeReg · 7-Scenes | Registration Recall (%) / time: 71.2 -> 83.0; 9.634 s -> 0.074 s | Improved |
| perturb input point clouds (shifts, 1% randomized, 10% removed) | Diff2I2P · 7-Scenes | Registration Recall (%): 83.0 -> 82.5 / 82.0 / 81.8 | No clear effect |
| apply Diff2I2P to outdoor KITTI instead of 2D3D-MATR | 2D3D-MATR · KITTI | Registration Recall (%): 75.4 -> 82.2 | Improved |
| Diff2I2P vs 2D3D-MATR on feature matching recall | 2D3D-MATR · 7-Scenes | Feature Matching Recall (%): 92.1 -> 92.2 | No clear effect |
Later papers that compared with this method
2 later accepted papers report Diff2I2P as a baseline in their results tables; 2 of them report a better number on most of the results they share with it.
- FS-I2P: A Hierarchical Focus–Sweep Registration Network with Dynamically Allocated Depth (ICML 2026 · better on 2 of 2 shared results)
- Rethinking 2D-3D Registration: A Novel Network for High-Value Zone Selection and Representation Consistency Alignment (CVPR 2026 · better on 2 of 2 shared results)
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 2 of 2 on most shared results.
- 2D3D-MATR — 2D3D-MATR: 2D-3D Matching Transformer for Detection-Free Registration Between Images and Point Clouds (ICCV 2023) · better on 9 of 9 shared results
- FreeReg — FreeReg: Image-to-Point Cloud Registration Leveraging Pretrained Diffusion Models and Monocular Depth Estimators (ICLR 2024) · better on 4 of 4 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
- FreeReg: Image-to-Point Cloud Registration Leveraging Pretrained Diffusion Models and Monocular Depth Estimators (ICLR 2024 · similarity 0.92)
- Partial Point Cloud Registration with Multi-view 2D Image Learning (AAAI 2025 · similarity 0.89)
- Where Precision Meets Efficiency: Transformation Diffusion Model for Point Cloud Registration (AAAI 2025 · similarity 0.88)
- Cross-PCR: A Robust Cross-Source Point Cloud Registration Framework (AAAI 2025 · similarity 0.87)
- Implicit Correspondence Learning for Image-to-Point Cloud Registration (CVPR 2025 · similarity 0.87)
- DiffCorr: Conditional Diffusion Model with Reliable Pseudo-Label Guidance for Unsupervised Point Cloud Shape Correspondence (AAAI 2025 · similarity 0.87)
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.
