Research map / Point clouds / Shape and surface completion
Cycle4Completion: Unpaired Point Cloud Completion Using Cycle Transformation With Missing Region Coding
Xin Wen, Zhizhong Han, Yan-Pei Cao, Pengfei Wan, Wen Zheng, Yu-Shen Liu
Check 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 bidirectional cycle transformation with missing region coding instead of Pcl2Pcl | Cycle4Completion · ShapeNet (3D-EPN split) | per point Chamfer distance × 10^4: 17.4 -> 14.1 | Improved |
| Add partial matching loss (ablation) | Cycle4Completion · ShapeNet, four categories | per point Chamfer distance × 10^4: 23.7 -> 9.1 | Improved |
| Add adversarial discriminator and loss (ablation) | Cycle4Completion · ShapeNet, four categories | per point Chamfer distance × 10^4: 12.8 -> 9.1 | Improved |
| Add cycle matching loss (ablation) | Cycle4Completion · ShapeNet, four categories | per point Chamfer distance × 10^4: 10.4 -> 9.1 | Improved |
| Add missing region codes and code matching loss (ablation) | Cycle4Completion · ShapeNet, four categories | per point Chamfer distance × 10^4: 9.4 -> 9.1 | Improved |
| Increase input resolution from 256 to 2048 points (ablation) | Cycle4Completion · ShapeNet | CD × 10^4: 14.4 -> 9.1 | Improved |
| Update auto-encoder and transformation parameters together when applying cycle loss (ablation) | Cycle4Completion · ShapeNet, four categories | CD × 10^4: 9.1 -> 10.1 | Hurt |
| Update auto-encoder and transformation parameters together when applying generator loss (ablation) | Cycle4Completion · ShapeNet, four categories | completion outcome | Hurt |
| Update auto-encoder and transformation parameters together when applying partial matching loss (ablation) | Cycle4Completion · ShapeNet, four categories | completion outcome | Hurt |
| Increase partial matching weight from 0.1 to 1.0 with cycle weight 0.01 (ablation) | Cycle4Completion · ShapeNet car | CD × 10^4: 9.8 -> 7.5 | Improved |
| Remove partial matching loss by setting its weight to 0 (ablation) | Cycle4Completion · ShapeNet car | CD × 10^4: 7.5 -> 17.1 | Hurt |
Later papers that compared with this method
3 later accepted papers report Cycle4Completion as a baseline in their results tables; 3 of them report a better number on most of the results they share with it.
- KT-Net: Knowledge Transfer for Unpaired 3D Shape Completion (AAAI 2023 · better on 1 of 1 shared result)
- Optimization over Disentangled Encoding: Unsupervised Cross-Domain Point Cloud Completion via Occlusion Factor Manipulation (ECCV 2022 · better on 8 of 8 shared results)
- Learning a Structured Latent Space for Unsupervised Point Cloud Completion (CVPR 2022 · 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
- Learning Geometric Transformation for Point Cloud Completion (IJCV 2023 · similarity 0.92)
- KT-Net: Knowledge Transfer for Unpaired 3D Shape Completion (AAAI 2023 · similarity 0.91)
- SymmCompletion: High-Fidelity and High-Consistency Point Cloud Completion with Symmetry Guidance (AAAI 2025 · similarity 0.88)
- DC-PCN: Point Cloud Completion Network with Dual-Codebook Guided Quantization (AAAI 2025 · similarity 0.88)
- Orthogonal Dictionary Guided Shape Completion Network for Point Cloud (AAAI 2024 · similarity 0.88)
- CDPNet: Cross-Modal Dual Phases Network for Point Cloud Completion (AAAI 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.
