Research map / 3D vision / 3D generation
OccAny: Generalized Unconstrained Urban 3D Occupancy
Anh-Quan Cao, Tuan-Hung Vu
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 |
|---|---|---|---|
| add Test-Time View Augmentation via Novel View Rendering (ablation) | OccAny · SemanticKITTI sequence | IoU (%): 19.64 -> 25.91 | Improved |
| add Test-Time View Augmentation (ablation) | OccAny · SemanticKITTI monocular | IoU (%): 11.55 -> 24.03 | Improved |
| add Segmentation Forcing loss (ablation) | OccAny · SemanticKITTI sequence | IoU (%): 24.23 -> 25.91 | Improved |
| add geometry-aware confidence weighting in segmentation forcing (ablation) | OccAny · SemanticKITTI sequence | IoU (%): 24.88 -> 25.91 | Improved |
| add task tokens tg and ts (ablation) | OccAny · SemanticKITTI single-frame | IoU (%): 22.67 -> 24.03 | Improved |
| add encoder distillation loss Lenc in rendering (ablation) | OccAny · SemanticKITTI single-frame | IoU (%): 23.46 -> 24.03 | Improved |
| add Segmentation Forcing for semantic occupancy with Grounded SAM2 (ablation) | OccAny · SemanticKITTI sequence | mIoU (%): 6.83 -> 7.28 | Improved |
| add Segmentation Forcing and NVR over base fine-tuning of MUSt3R (ablation) | OccAny base · Occ3D-NuScenes surround-view | IoU (%): 18.53 -> 34.15 | Improved |
| add rotations, larger steps and lateral moves to forward-only TTVA (ablation) | OccAny · SemanticKITTI sequence | IoU (%): 21.41 -> 25.91 | Improved |
| apply TTVA to baselines (CUT3R, AnySplat) | CUT3R / AnySplat · SemanticKITTI sequence | IoU (%): CUT3R 13.11 -> 15.93; AnySplat 11.67 -> 12.39 | Mixed |
| swap base foundation models to DA3 and SAM3 (OccAny+) | OccAny · SemanticKITTI sequence / Occ3D-NuScenes surround | IoU (%): KITTI 25.91 -> 27.33; NuScenes 34.15 -> 33.49 | Mixed |
| zero-shot OccAny versus in-domain self-supervised methods on nuScenes surround | OccAny · Occ3D-NuScenes surround-view | IoU (%): 45.19 (GaussTR) vs 34.15 | Hurt |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.
- MUSt3R — MUSt3R: Multi-view Network for Stereo 3D Reconstruction (CVPR 2025) · 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
- OctOcc: High-Resolution 3D Occupancy Prediction with Octree (AAAI 2024 · similarity 0.89)
- ODG: Occupancy Prediction Using Dual Gaussians (NeurIPS 2025 · similarity 0.89)
- OctreeOcc: Efficient and Multi-Granularity Occupancy Prediction Using Octree Queries (NeurIPS 2024 · similarity 0.89)
- Learning Occupancy for Monocular 3D Object Detection (CVPR 2024 · similarity 0.88)
- LowRankOcc: Tensor Decomposition and Low-Rank Recovery for Vision-based 3D Semantic Occupancy Prediction (CVPR 2024 · similarity 0.88)
- Achieving Speed-Accuracy Balance in Vision-based 3D Occupancy Prediction via Geometric-Semantic Disentanglement (AAAI 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.
