Research map / Video / Video segmentation and tracking
SceneScribe-1M: A Large-Scale Video Dataset with Comprehensive Geometric and Semantic Annotations
Yunnan Wang, Kecheng Zheng, Jianyuan Wang, Minghao Chen, David Novotny, Christian Rupprecht, Yinghao Xu, Xing Zhu, Wenjun Zeng, Xin Jin, Yujun Shen
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 SceneScribe-1M to TartanAir training | MoGe · Monocular depth average (scale-invariant) | Rel: 6.17 -> 6.14 | Improved |
| add SceneScribe-1M to TartanAir training | MoGe · Monocular depth average (affine-invariant depth) | Rel: 4.72 -> 4.68 | Improved |
| add SceneScribe-1M to TartanAir training | MoGe · Monocular depth average (affine-invariant disparity) | Rel: 5.76 -> 5.68 | Improved |
| add SceneScribe-1M to TartanAir training | MoGe · Sintel (scale-invariant depth) | Rel: 19.3 -> 19.6 | Hurt |
| add SceneScribe-1M training data | VGGT · CO3Dv2 and ETH3D pose estimation | AUC30: 89.5 -> 89.9 | Improved |
| add SceneScribe-1M training data | VGGT · CO3Dv2 and ETH3D point map | Overall: 0.677 -> 0.697 | Hurt |
| add SceneScribe-1M training data | MonST3R · Sintel pose estimation | ATE: 0.108 -> 0.099 | Improved |
| add SceneScribe-1M training data | MonST3R · Sintel depth estimation | delta1: 58.5 -> 58.1 | No clear effect |
| add SceneScribe-1M training data | CoTracker3 · TAP-Vid 2D tracking | mean delta_avg^vis: 76.6 -> 77.4 | Improved |
| add SceneScribe-1M training data | SpatialTrackerV2 · TAPVid-3D | average AJ: 23.25 -> 23.5 | Improved |
| add SceneScribe-1M training data | AC3D · RealEstate10K camera control | TransErr: 0.374 -> 0.318 | Improved |
| add SceneScribe-1M training data | AC3D · RealEstate10K generation quality | FVD: 38.20 -> 35.15 | Improved |
Closest papers by meaning
- From an Image to a Scene: Learning to Imagine the World from a Million 360° Videos (NeurIPS 2024 · similarity 0.89)
- Video-3D LLM: Learning Position-Aware Video Representation for 3D Scene Understanding (CVPR 2025 · similarity 0.88)
- Leader360V: A Large-scale, Real-world 360 Video Dataset for Multi-task Learning in Diverse Environment (NeurIPS 2025 · similarity 0.87)
- Stereo4D: Learning How Things Move in 3D from Internet Stereo Videos (CVPR 2025 · similarity 0.87)
- Learning from Videos for 3D World: Enhancing MLLMs with 3D Vision Geometry Priors (NeurIPS 2025 · similarity 0.87)
- Video Depth Anything: Consistent Depth Estimation for Super-Long Videos (CVPR 2025 · 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.
