Research map / Point clouds / Shape and surface completion
Simba: Towards High-Fidelity and Geometrically-Consistent Point Cloud Completion via Transformation Diffusion
Lirui Zhang, Zhengkai Zhao, Zhi Zuo, Pan Gao, Jie Qin
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 |
|---|---|---|---|
| diffuse the transformation field (Sym-Diffuser) instead of regressing it with a Transformer (ablation) | Simba · PCN | CD-l1 (x10^3): 6.48 -> 6.34 | Improved |
| use three-level gradual upsampling [2x,2x,4x] instead of single 16x stage (ablation) | MBA-Refiner · PCN | CD-l1 (x10^3): 6.70 -> 6.34 | Improved |
| balanced [4x,4x] vs unbalanced [2x,8x] two-level upsampling (ablation) | MBA-Refiner · PCN | CD-l1 (x10^3): 6.56 -> 6.52 | Improved |
| use cross-attention fusion in early blocks instead of MLP (ablation) | MBA-Refiner · PCN | CD-l1 (x10^3): 6.49 -> 6.34 | Improved |
| use Mamba fusion in final block instead of MLP (ablation) | MBA-Refiner · PCN | CD-l1 (x10^3): 6.41 -> 6.34 | Improved |
| use Mamba fusion in final block instead of cross-attention (memory 16.4 GB to 14.7 GB) (ablation) | MBA-Refiner · PCN | CD-l1 (x10^3): 6.35 -> 6.34 | No clear effect |
| use cross-attention early blocks instead of Mamba only (ablation) | MBA-Refiner · PCN | CD-l1 (x10^3): 6.43 -> 6.34 | Improved |
| replace SymmCompletion regression with Simba on KITTI | SymmCompletion · KITTI | MMD (x10^3): 0.970 -> 0.423 | Improved |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 5 of 6 on most shared results.
- AdaPoinTr — AdaPoinTr: Diverse Point Cloud Completion With Adaptive Geometry-Aware Transformers (TPAMI 2023) · better on 3 of 4 shared results
- SymmCompletion — SymmCompletion: High-Fidelity and High-Consistency Point Cloud Completion with Symmetry Guidance (AAAI 2025) · better on 3 of 3 shared results
- CRA-PCN — CRA-PCN: Point Cloud Completion with Intra- and Inter-level Cross-Resolution Transformers (AAAI 2024) · better on 3 of 3 shared results
- PointCFormer — PointCFormer: A Relation-Based Progressive Feature Extraction Network for Point Cloud Completion (AAAI 2025) · better on 1 of 2 shared results
- EINet — EINet: Point Cloud Completion via Extrapolation and Interpolation (ECCV 2024) · better on 1 of 1 shared result
- SeedFormer — SeedFormer: Patch Seeds Based Point Cloud Completion with Upsample Transformer (ECCV 2022) · better on 1 of 1 shared result
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
- SymmCompletion: High-Fidelity and High-Consistency Point Cloud Completion with Symmetry Guidance (AAAI 2025 · similarity 0.88)
- PCDreamer: Point Cloud Completion Through Multi-view Diffusion Priors (CVPR 2025 · similarity 0.88)
- PointMAC: Meta-Learned Adaptation for Robust Test-Time Point Cloud Completion (NeurIPS 2025 · similarity 0.88)
- Where Precision Meets Efficiency: Transformation Diffusion Model for Point Cloud Registration (AAAI 2025 · similarity 0.87)
- Digging into Intrinsic Contextual Information for High-fidelity 3D Point Cloud Completion (AAAI 2025 · similarity 0.87)
- Adaptive 3D Reconstruction via Diffusion Priors and Forward Curvature-Matching Likelihood Updates (NeurIPS 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.
