Research map / 3D vision / Depth and stereo
PAGE-4D: Disentangled Pose and Geometry Estimation for 4D Perception
Kaichen Zhou, Yuhan Wang, Grace Chen, Gaspard Beaudouin, Fangneng Zhan, Paul Pu Liang, Mengyu Wang
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 dynamics-aware mask attention to middle-layer fine-tuning (ablation) | PAGE-4D (VGGT*) · Sintel video depth (scale align) | Abs Rel: 0.409 -> 0.357 | Improved |
| add dynamics-aware mask attention to middle-layer fine-tuning (ablation) | PAGE-4D (VGGT*) · Sintel video depth (scale align) | delta<1.25: 0.590 -> 0.699 | Improved |
| add dynamics-aware mask attention to middle-layer fine-tuning (ablation) | PAGE-4D (VGGT*) · DyCheck video depth (scale align) | Abs Rel: 0.177 -> 0.170 | Improved |
| fine-tune only middle layers instead of whole model (ablation) | VGGT* · Sintel video depth (scale align) | Abs Rel: 0.405 -> 0.409 | No clear effect |
| fine-tune with dynamic-aware aggregator vs plain VGGT | VGGT · Sintel camera pose | ATE: 0.214 -> 0.178 | Improved |
| fine-tune with dynamic-aware aggregator vs plain VGGT | VGGT · DyCheck point reconstruction | Accuracy error (mean): 1.051 -> 0.403 | Improved |
| fine-tune with dynamic-aware aggregator vs plain VGGT | VGGT · Nerfie novel view synthesis | PSNR: 16.861 -> 17.593 | Improved |
| mask dynamic tokens attention (DD-MSK) on VGGT (ablation) | VGGT · Odyssey | RPE rot / Static-D: RPE rot 0.942 -> 0.869; Static-D 0.085 -> 0.566 | Mixed |
| mask dynamic regions at input (Input-MSK) on VGGT (ablation) | VGGT · Odyssey | RPE rot: 0.942 -> 1.006 | Hurt |
| replace layer 17 global attention output with noise (ablation) | VGGT · Odyssey | Static-D: 0.085 -> 1.663 | 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.
- VGGT — VGGT: Visual Geometry Grounded Transformer (CVPR 2025) · better on 13 of 13 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
- 4D3R: Motion-Aware Neural Reconstruction and Rendering of Dynamic Scenes from Monocular Videos (NeurIPS 2025 · similarity 0.90)
- IGGT4D: Streaming 4D Instance-Grounded Geometry Transformer (NeurIPS 2026 · similarity 0.90)
- Motion4D: Learning 3D-Consistent Motion and Semantics for 4D Scene Understanding (NeurIPS 2025 · similarity 0.89)
- Stereo4D: Learning How Things Move in 3D from Internet Stereo Videos (CVPR 2025 · similarity 0.89)
- Controlling Space and Time with Diffusion Models (ICLR 2025 · similarity 0.89)
- Flux4D: Flow-based Unsupervised 4D Reconstruction (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.
