Research map / 3D vision / Face and shape
Probabilistic Implicit Scene Completion
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (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 probabilistic continuous cGCA completion instead of deterministic ConvOcc or IFNet | ShapeNet scene · ShapeNet scene, min. rate 0.2 | min. CD: 3.60 -> 2.80 (ConvOcc); 12.94 -> 2.80 (IFNet) | Helped |
| Use cGCA rather than GCA's discrete voxel completion | Scene completion · ShapeNet scene, min. rate 0.2 | min. CD: 4.97 -> 2.80 | Helped |
| Condition each transition on the encoded initial state (ablation) | cGCA · ShapeNet scene, min. rate 0.2 | min. CD: 2.80 -> 1.75 | Helped |
| Condition each transition on the encoded initial state (ablation) | cGCA · ShapeNet scene, min. rate 0.8 | min. CD: 0.69 -> 0.57 | Helped |
| Use cGCA instead of GCA | 3DFront scene completion · 3DFront, min. rate 0.5 | min. CD: 3.11 -> 2.12 | Helped |
| Condition every transition on the encoded initial state (ablation) | cGCA · 3DFront, min. rate 0.5 | min. CD: 2.12 -> 2.19 | Hurt |
| Condition every transition on the encoded initial state (ablation) | cGCA · 3DFront, min. rate 0.2 | min. CD: 4.07 -> 3.53 | Helped |
| Condition every transition on the encoded initial state (ablation) | cGCA · 3DFront, min. rate 0.8 | min. CD: 1.64 -> 1.47 | Helped |
| Increase cGCA voxel resolution from 32^3 to 64^3 (ablation) | cGCA · ShapeNet chairs/sofas/tables (Avg.) | TMD: 16.47 -> 19.94 | Helped |
| Increase cGCA voxel resolution from 32^3 to 64^3 (ablation) | cGCA · ShapeNet chairs/sofas/tables (Avg.) | MMD: 5.59 -> 5.57 | Helped |
| Condition cGCA transitions on encoded initial state (ablation) | cGCA (64^3) · ShapeNet chairs/sofas/tables (Avg.) | UHD: 8.21 -> 7.07 | Helped |
| Use continuous cGCA rather than voxelized GCA | ShapeNet single-object completion · ShapeNet chairs/sofas/tables (Avg.) | MMD: 5.72 -> 5.57 | Helped |
Closest papers by meaning
- DiffComplete: Diffusion-based Generative 3D Shape Completion (NeurIPS 2023 · similarity 0.91)
- Learning to Generate 3D Shapes with Generative Cellular Automata (ICLR 2021 · similarity 0.90)
- DinoComplete: 3D Shape Completion with Distilled Semantic Priors and State Space Models (NeurIPS 2026 · similarity 0.88)
- PartCrafter: Structured 3D Mesh Generation via Compositional Latent Diffusion Transformers (NeurIPS 2025 · similarity 0.88)
- Context and Geometry Aware Voxel Transformer for Semantic Scene Completion (NeurIPS 2024 · similarity 0.87)
- Outdoor Scene Extrapolation with Hierarchical Generative Cellular Automata (CVPR 2024 · 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.