Research map / 3D vision / Gaussian splatting
StreetForward: Perceiving Dynamic Street with Feedforward Causal Dynamics
Zhongrui Yu, Zhao Wang, Yida Wang, Xueyang Zhang, Yifei Zhan, Kun Zhan
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 causal masked attention (ablation) | StreetForward · Waymo (dynamic only) | PSNR: 22.36 -> 24.30 | Improved |
| add causal masked attention (ablation) | StreetForward · Waymo (full image) | PSNR: 24.07 -> 27.01 | Improved |
| use variable causal horizon instead of fixed k=1 (ablation) | StreetForward · Waymo (dynamic only) | PSNR: 22.96 -> 24.30 | Improved |
| add local rigidity loss L_rigid (ablation) | StreetForward · Waymo (dynamic only) | PSNR: 22.64 -> 24.30 | Improved |
| add backward velocity (ablation) | StreetForward · Waymo (dynamic only) | PSNR: 23.26 -> 24.30 | Improved |
| add forward-backward consistency loss L_fb (ablation) | StreetForward · Waymo (dynamic only) | SSIM: 0.721 -> 0.827 | Improved |
| add opacity stabilization loss L_alpha (ablation) | StreetForward · Waymo (dynamic only) | PSNR: 21.02 -> 24.30 | Improved |
| add opacity stabilization loss L_alpha (ablation) | StreetForward · Waymo (full image) | SSIM: 0.683 -> 0.818 | Improved |
| use StreetForward versus DGGT | DGGT · Waymo (dynamic only) | PSNR: 20.99 -> 24.30 | Improved |
| use StreetForward versus DGGT | DGGT · Waymo (full image) | PSNR: 27.41 -> 27.01 | Hurt |
| use StreetForward versus DGGT | DGGT · Waymo sparse-depth (full image) | depth RMSE: 4.08 -> 3.14 | Improved |
| zero-shot transfer to Carla versus DGGT | DGGT · Carla zero-shot (dynamic only) | PSNR: 21.32 -> 22.37 | Improved |
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 1 of 1 shared result
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
- Effective Multi-sensor Conditioning for Street-view Novel-view Synthesis (NeurIPS 2026 · similarity 0.89)
- 3D StreetUnveiler with Semantic-aware 2DGS - a simple baseline (ICLR 2025 · similarity 0.88)
- StreetCrafter: Street View Synthesis with Controllable Video Diffusion Models (CVPR 2025 · similarity 0.88)
- DeSiRe-GS: 4D Street Gaussians for Static-Dynamic Decomposition and Surface Reconstruction for Urban Driving Scenes (CVPR 2025 · similarity 0.88)
- DrivingForward: Feed-forward 3D Gaussian Splatting for Driving Scene Reconstruction from Flexible Surround-view Input (AAAI 2025 · similarity 0.88)
- FRUC: Feedforward Dynamic Scene Reconstruction from Uncalibrated Collaborative Driving Views (NeurIPS 2026 · 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.
