Research map / 3D vision / 3D generation
EmbodiedOcc: Embodied 3D Occupancy Prediction for Vision-based Online Scene Understanding
Yuqi Wu, Wenzhao Zheng, Sicheng Zuo, Yuanhui Huang, 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 Gaussian representation instead of voxels (ablation) | EmbodiedOcc · EmbodiedOcc-ScanNet embodied prediction | mIoU: 26.99 -> 42.53 | Improved |
| use Gaussian representation instead of voxels (ablation) | EmbodiedOcc · Occ-ScanNet local prediction | mIoU: 38.12 -> 45.15 | Improved |
| add online Gaussian memory (ablation) | EmbodiedOcc · EmbodiedOcc-ScanNet embodied prediction | mIoU: 40.74 -> 42.53 | Improved |
| add online Gaussian memory (ablation) | EmbodiedOcc · Occ-ScanNet local prediction | mIoU: 45.15 -> 45.15 | No clear effect |
| add depth-aware branch (ablation) | EmbodiedOcc · EmbodiedOcc-ScanNet-mini embodied prediction | mIoU: 30.73 -> 41.45 | Improved |
| add depth-aware branch (ablation) | EmbodiedOcc · Occ-ScanNet local prediction (mini2) | mIoU: 40.07 -> 46.20 | Improved |
| use depth-aware branch instead of naive pseudo point cloud init (ablation) | EmbodiedOcc · Occ-ScanNet-mini2 local | mIoU: 42.73 -> 46.20 | Improved |
| swap DepthAnything-V2 for IndoorDepth (ablation) | EmbodiedOcc · EmbodiedOcc-ScanNet-mini embodied prediction | mIoU: 41.45 -> 37.78 | Hurt |
| use 16200 Gaussians instead of 8100 (ablation) | EmbodiedOcc · EmbodiedOcc-ScanNet-mini embodied prediction | mIoU: 37.99 -> 41.45 | Improved |
| use max scale 0.08 instead of 0.20 (ablation) | EmbodiedOcc · EmbodiedOcc-ScanNet-mini embodied prediction | mIoU: 38.40 -> 41.45 | Improved |
| Look-Back re-processing of earlier frames (K=3) | EmbodiedOcc · EmbodiedOcc-ScanNet K=3 | mIoU: 39.32 -> 40.09 | Improved |
| Look-Back re-processing of earlier frames (K=5) | EmbodiedOcc · EmbodiedOcc-ScanNet K=5 | mIoU: 40.03 -> 40.98 | Improved |
Later papers that compared with this method
5 later accepted papers report EmbodiedOcc as a baseline in their results tables; 5 of them report a better number on most of the results they share with it.
- Generalizing Visual Geometry Priors to Sparse Gaussian Occupancy Prediction (CVPR 2026 · better on 9 of 9 shared results)
- SplatSSC: Decoupled Depth-Guided Gaussian Splatting for Semantic Scene Completion (AAAI 2026 · better on 8 of 8 shared results)
- AdaOcc: Adaptive 3D Occupancy Prediction for Embodied Tasks (NeurIPS 2026 · better on 4 of 4 shared results)
- TGSFormer: Scalable Temporal Gaussian Splatting for Embodied Semantic Scene Completion (CVPR 2026 · better on 3 of 3 shared results)
- Monocular Open Vocabulary Occupancy Prediction for Indoor Scenes (CVPR 2026 · better on 1 of 1 shared result)
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 3 of 3 on most shared results.
- GaussianFormer — GaussianFormer: Scene as Gaussians for Vision-Based 3D Semantic Occupancy Prediction (ECCV 2024) · better on 4 of 4 shared results
- SurroundOcc — SurroundOcc: Multi-camera 3D Occupancy Prediction for Autonomous Driving (ICCV 2023) · better on 4 of 4 shared results
- MonoScene — MonoScene: Monocular 3D Semantic Scene Completion (CVPR 2022) · better on 2 of 2 shared results
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
- AdaOcc: Adaptive 3D Occupancy Prediction for Embodied Tasks (NeurIPS 2026 · similarity 0.93)
- EmbodiedScan: A Holistic Multi-Modal 3D Perception Suite Towards Embodied AI (CVPR 2024 · similarity 0.90)
- OctOcc: High-Resolution 3D Occupancy Prediction with Octree (AAAI 2024 · similarity 0.90)
- ODG: Occupancy Prediction Using Dual Gaussians (NeurIPS 2025 · similarity 0.90)
- VoxelSplat: Dynamic Gaussian Splatting as an Effective Loss for Occupancy and Flow Prediction (CVPR 2025 · similarity 0.89)
- GaussianWorld: Gaussian World Model for Streaming 3D Occupancy Prediction (CVPR 2025 · similarity 0.89)
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.
