Research map / Point clouds / Shape and surface completion
ProtoComp: Diverse Point Cloud Completion with Controllable Prototype
Xumin Yu, Yanbo Wang, Jie Zhou, Jiwen Lu
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 ProtoComp with AdaPoinTr decoder | AdaPoinTr · PCN dataset | Avg CD-l1 (x1000): 6.53 -> 6.47 | Improved |
| wrap decoder with ProtoComp prototype guidance | PoinTr · PCN dataset | Avg CD-l1 (x1000): 8.38 -> 7.47 | Improved |
| complete partial scans with existing completion methods | PCN, SnowFlakeNet, AdaPoinTr · Real-Sensors Benchmark | GD_I (%): 32.8 (partials) vs 27.2 PCN | Hurt |
| add text prompt at inference | ProtoComp · Real-Sensors Benchmark | GD_I (%): 33.4 -> 43.6 | Improved |
| fine-tune vanilla Point-E directly on 800 pairs (ablation) | Model A · Real-Sensors Benchmark (ShapeNet-55 trained) | GD_I (%): 57.1 -> 42.9 | Hurt |
| add partial condition module (ablation) | Model B · Real-Sensors Benchmark (ShapeNet-55 trained) | Fidelity-l1 (x1000): 98.4 -> 16.2 | Improved |
| use side branch instead of full fine-tuning (ablation) | Model C · Real-Sensors Benchmark (ShapeNet-55 trained) | GD_I (%): 44.1 -> 45.4 | Improved |
| initialize side branch with pre-trained weights (copy) (ablation) | Model D · Real-Sensors Benchmark (ShapeNet-55 trained) | GD_I (%): 45.4 -> 46.8 | Improved |
| reduce training categories from 8 to 4 with same total pairs (ablation) | Model E · Real-Sensors Benchmark (ShapeNet-55 trained) | GD_I (%): 46.8 -> 42.3 | Hurt |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 4 of 4 on most shared results.
- AdaPoinTr — AdaPoinTr: Diverse Point Cloud Completion With Adaptive Geometry-Aware Transformers (TPAMI 2023) · better on 3 of 3 shared results
- SnowFlakeNet — SnowflakeNet: Point Cloud Completion by Snowflake Point Deconvolution With Skip-Transformer (ICCV 2021) · better on 2 of 2 shared results
- SeedFormer — SeedFormer: Patch Seeds Based Point Cloud Completion with Upsample Transformer (ECCV 2022) · better on 1 of 1 shared result
- PoinTr — PoinTr: Diverse Point Cloud Completion With Geometry-Aware Transformers (ICCV 2021) · 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
- ComPC: Completing a 3D Point Cloud with 2D Diffusion Priors (ICLR 2025 · similarity 0.87)
- EINet: Point Cloud Completion via Extrapolation and Interpolation (ECCV 2024 · similarity 0.87)
- Point-based Instance Completion with Scene Constraints (ICLR 2025 · similarity 0.87)
- GenPC: Zero-shot Point Cloud Completion via 3D Generative Priors (CVPR 2025 · similarity 0.86)
- PLATYPUS: Progressive Local Surface Estimator for Arbitrary-Scale Point Cloud Upsampling (AAAI 2025 · similarity 0.85)
- Position-Aware Guided Point Cloud Completion with CLIP Model (AAAI 2025 · similarity 0.85)
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.
