Research map / Video / Video segmentation and tracking
SpatialVID: A Large-Scale Video Dataset with Spatial Annotations
Jiahao Wang, Yufeng Yuan, Rujie Zheng, Youtian Lin, Jian Gao, Lin-Zhuo Chen, Yajie Bao, Chang Zeng, Yanxi Zhou, Xiao-Xiao Long, Hao Zhu, Zhaoxiang Zhang and 2 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 |
|---|---|---|---|
| train camera-controlled video model on SpatialVID-HQ instead of RealEstate10K | Wan2.2-5B camera-control model · SpatialVID test | TransErr: 5.16 -> 4.33 | Improved |
| train on SpatialVID-HQ instead of Sekai-Real | Wan2.2-5B camera-control model · Sekai-Real test | TransErr: 6.49 -> 6.04 | Improved |
| train on SpatialVID-HQ instead of RealEstate10K | Wan2.2-5B camera-control model · RealEstate10K test | RotErr: 1.15 -> 0.99 | Improved |
| train on SpatialVID-HQ instead of RealEstate10K | Wan2.2-5B camera-control model · RealEstate10K test | Imaging Quality: 72.71 -> 75.68 | Improved |
| train on SpatialVID-HQ instead of RealEstate10K | Wan2.2-5B camera-control model · RealEstate10K test | Motion Smoothness: 99.05 -> 98.77 | Hurt |
| train GS-LRM on SpatialVID-HQ instead of RealEstate10K | GS-LRM · DL3DV | PSNR: 27.01 -> 27.80 | Improved |
| train GS-LRM on SpatialVID-HQ instead of RealEstate10K | GS-LRM · SpatialVID test | LPIPS: 0.222 -> 0.203 | Improved |
| fine-tune CUT3R on SpatialVID | CUT3R · TUM-dynamics | ATE: 0.049 -> 0.040 | Improved |
| add SpatialVID-HQ to VGGT fine-tuning | VGGT · TUM-dynamics | ATE: 0.015 -> 0.013 | Improved |
| add SpatialVID-HQ to VGGT fine-tuning | VGGT · Sintel | ATE: 0.134 -> 0.148 | Hurt |
| fine-tune VGGT with SpatialVID-HQ added | VGGT · Dycheck | RPE rot: 1.087 -> 1.090 | No clear effect |
| fine-tune CUT3R on SpatialVID | CUT3R · Dycheck | ATE: 0.020 -> 0.019 | No clear effect |
Closest papers by meaning
- Video World Models with Long-term Spatial Memory (NeurIPS 2025 · similarity 0.89)
- SpaceVista: All-Scale Visual Spatial Reasoning from mm to km (ICML 2026 · similarity 0.89)
- SpatialBench: Is Your Spatial Foundation Model an All-Round Player (NeurIPS 2026 · similarity 0.88)
- From an Image to a Scene: Learning to Imagine the World from a Million 360° Videos (NeurIPS 2024 · similarity 0.88)
- Leader360V: A Large-scale, Real-world 360 Video Dataset for Multi-task Learning in Diverse Environment (NeurIPS 2025 · similarity 0.87)
- Implicit Zoo: A Large-Scale Dataset of Neural Implicit Functions for 2D Images and 3D Scenes (NeurIPS 2024 · similarity 0.87)
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.
