Research map / 3D vision / 3D generation
Panoptic Lifting for 3D Scene Understanding With Neural Fields
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 |
|---|---|---|---|
| Lift machine-generated 2D panoptic masks into a 3D-consistent neural field | Panoptic Lifting · Hypersim | PQ_scene: 44.8 -> 60.1 | Helped |
| Lift machine-generated 2D panoptic masks into a 3D-consistent neural field | Panoptic Lifting · Replica | PQ_scene: 52.5 -> 57.9 | Helped |
| Lift machine-generated 2D panoptic masks into a 3D-consistent neural field | Panoptic Lifting · ScanNet | PQ_scene: 54.3 -> 58.9 | Helped |
| Add segment consistency loss (ablation) | Panoptic Lifting · ScanNet | mIoU: 60.9 -> 65.2 | Helped |
| Add test-time augmentation confidence estimates (ablation) | Panoptic Lifting · ScanNet | mIoU: 63.1 -> 65.2 | Helped |
| Use a bounded probability segmentation field instead of unbounded outputs (ablation) | Panoptic Lifting · ScanNet | PQ_scene: 52.5 -> 58.9 | Helped |
| Block gradients from segmentation branches to geometry (ablation) | Panoptic Lifting · ScanNet | PSNR: 27.2 -> 28.5 | Helped |
| Disable segment consistency, TTA, bounded field, and gradient blocking (ablation) | Panoptic Lifting · ScanNet | PQ_scene: 47.9 -> 58.9 | Hurt |
Closest papers by meaning
- Multi-view Consistent 3D Panoptic Scene Understanding (AAAI 2025 · similarity 0.88)
- Lift3D: Zero-Shot Lifting of Any 2D Vision Model to 3D (CVPR 2024 · similarity 0.88)
- PanoRecon: Real-Time Panoptic 3D Reconstruction from Monocular Video (CVPR 2024 · similarity 0.88)
- Rethinking End-to-End 2D to 3D Scene Segmentation in Gaussian Splatting (CVPR 2025 · similarity 0.87)
- IPFormer: Visual 3D Panoptic Scene Completion with Context-Adaptive Instance Proposals (NeurIPS 2025 · similarity 0.86)
- Scene-Centric Unsupervised Panoptic Segmentation (CVPR 2025 · similarity 0.86)
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.