Research map / 3D vision / 3D generation
Holi-Spatial: Evolving Video Streams into Holistic 3D Spatial Intelligence
Yuanyuan Gao, Hao Li, Yifei Liu, Xinhao Ji, Yuning Gong, Yuanjun Liao, Fangfu Liu, Manyuan Zhang, Yuchen Yang, Dan Xu, Xue Yang, Huaxi Huang and 5 more
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 3DGS geometric optimization to Depth-Anything-3 depth (ablation) | Holi-Spatial · ScanNet++ 3D detection | precision / recall @IoU25: P 0.13 -> 0.81, R 0.31 -> 0.89 | Improved |
| add SAM3 confidence filter (ablation) | Holi-Spatial · ScanNet++ 3D detection | precision@IoU25: 0.35 -> 0.67 | Improved |
| add SAM3 confidence filter (ablation) | Holi-Spatial · ScanNet++ 3D detection | recall@IoU25: 0.74 -> 0.69 | Hurt |
| add VLM agent verification of borderline instances (ablation) | Holi-Spatial · ScanNet++ 3D detection | precision / recall @IoU25: P 0.67 -> 0.81, R 0.69 -> 0.89 | Improved |
| fine-tune Qwen3-VL-8B on Holi-Spatial-4M spatial QA | Qwen3-VL-8B · MindCube | accuracy: 29.4 -> 49.1 | Improved |
| fine-tune Qwen3-VL-8B on Holi-Spatial-4M spatial QA | Qwen3-VL-8B · MMSI-Bench | accuracy: 31.1 -> 32.6 | Improved |
| fine-tune Qwen3-VL-8B on Holi-Spatial-4M spatial QA | Qwen3-VL-8B · SparBench-tiny | accuracy: 39.83 -> 46.28 | Improved |
| fine-tune Qwen3-VL-2B on Holi-Spatial-4M spatial QA | Qwen3-VL-2B · MindCube | accuracy: 33.5 -> 44.0 | Improved |
| fine-tune Qwen3-VL-8B on Holi-Spatial-4M 3D grounding data | Qwen3-VL-8B · ScanNet++ 3D grounding | AP50: 13.50 -> 27.98 | Improved |
| fine-tune Qwen3-VL-8B on Holi-Spatial-4M | Qwen3-VL-8B · ViewSpatial | accuracy: 42.35 -> 50.12 | Improved |
Closest papers by meaning
- ReVSI: Rebuilding Visual Spatial Intelligence Evaluation for Accurate Assessment of VLM 3D Reasoning (ICML 2026 · similarity 0.89)
- SpatialLLM: A Compound 3D-Informed Design towards Spatially-Intelligent Large Multimodal Models (CVPR 2025 · similarity 0.89)
- SpaceVista: All-Scale Visual Spatial Reasoning from mm to km (ICML 2026 · similarity 0.89)
- MMSI-Bench: A Benchmark for Multi-Image Spatial Intelligence (ICLR 2026 · similarity 0.89)
- SpatialReasoner: Towards Explicit and Generalizable 3D Spatial Reasoning (NeurIPS 2025 · similarity 0.89)
- From Flatland to Space: Teaching Vision-Language Models to Perceive and Reason in 3D (NeurIPS 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.
