Research map / 3D vision / Face and shape
Occupancy Networks: Learning 3D Reconstruction in Function Space
Check 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 a continuous occupancy function instead of a fixed-resolution voxel output | ONet · ShapeNet single-image reconstruction (mean) | IoU: 0.493 -> 0.571 | Helped |
| use a continuous occupancy function instead of a fixed-resolution voxel output | ONet · ShapeNet single-image reconstruction (mean) | Chamfer-L1: 0.175 -> 0.215 | Hurt |
| use a continuous occupancy function instead of a fixed-resolution voxel output | ONet · ShapeNet single-image reconstruction (mean) | Normal Consistency: 0.811 -> 0.834 | Helped |
| use a continuous occupancy function for reconstruction from point clouds | ONet · ShapeNet point-cloud reconstruction | IoU: 0.674 -> 0.778 | Helped |
| use a continuous occupancy function for reconstruction from point clouds | ONet · ShapeNet point-cloud reconstruction | Chamfer-L1: 0.117 -> 0.079 | Helped |
| use a continuous occupancy function to super-resolve a coarse voxel input | ONet · ShapeNet voxel super-resolution | IoU: 0.631 -> 0.703 | Helped |
| use uniform sampling in the bounding volume rather than equal inside/outside sampling (ablation) | ONet · ShapeNet single-image reconstruction | IoU: 0.475 -> 0.571 | Helped |
| use uniform sampling in the bounding volume rather than surface sampling (ablation) | ONet · ShapeNet single-image reconstruction | IoU: 0.536 -> 0.571 | Helped |
| increase uniform training samples from 64 to 2048 (ablation) | ONet · ShapeNet single-image reconstruction | IoU: 0.554 -> 0.571 | Helped |
| add ResNet blocks to the decoder (ablation) | ONet · ShapeNet single-image reconstruction | IoU: 0.559 -> 0.571 | Helped |
| add conditional batch normalization instead of a linear input projection (ablation) | ONet · ShapeNet single-image reconstruction | IoU: 0.522 -> 0.571 | Helped |
Closest papers by meaning
- Unsupervised Occupancy Learning from Sparse Point Cloud (CVPR 2024 · similarity 0.89)
- OctOcc: High-Resolution 3D Occupancy Prediction with Octree (AAAI 2024 · similarity 0.89)
- Learning Occupancy for Monocular 3D Object Detection (CVPR 2024 · similarity 0.88)
- 3D Occupancy Prediction with Low-Resolution Queries via Prototype-aware View Transformation (CVPR 2025 · similarity 0.88)
- LowRankOcc: Tensor Decomposition and Low-Rank Recovery for Vision-based 3D Semantic Occupancy Prediction (CVPR 2024 · similarity 0.88)
- ODG: Occupancy Prediction Using Dual Gaussians (NeurIPS 2025 · 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.