Research map / 3D vision / Gaussian splatting
SLARM: Streaming and Language-Aligned Reconstruction Model for Dynamic Scenes
Zhicheng Qiu, Jiarui Meng, Tong-an Luo, Yican Huang, Xuan Feng, Xuanfu Li, Zhan Xu
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 |
|---|---|---|---|
| use higher-order (Taylor) motion modeling instead of STORM constant-velocity | SLARM-F · Waymo Open Dataset val scene flow | EPE (m): 0.304 -> 0.240 | Improved |
| use higher-order motion modeling instead of constant velocity | SLARM-F · Waymo Open Dataset val scene flow | angular error (rad): 0.667 -> 0.540 | Improved |
| use window attention (online streaming) instead of full attention | SLARM · Waymo Open Dataset val (full image) | PSNR: 27.49 -> 27.30 | Hurt |
| use window attention (online streaming) instead of full attention | SLARM · Waymo Open Dataset val scene flow | EPE (m): 0.240 -> 0.337 | Hurt |
| use window attention (online streaming) instead of full attention | SLARM · Waymo Open Dataset val scene flow | Acc5 (%): 78.15 -> 81.07 | Improved |
| add LSeg distillation loss (Lsem) to base model (ablation) | SLARM · Waymo Open Dataset val (Fig 5b) | PSNR: 27.37 -> 27.43 | Improved |
| switch to classification loss Lcls for final iterations after Lsem (ablation) | SLARM · Waymo Open Dataset val (Fig 5b) | PSNR: 27.43 -> 27.49 | Improved |
| add LSeg distillation loss (Lsem) to base model (ablation) | SLARM · Waymo Open Dataset val (Fig 5b) | Flow EPE: 0.279 -> 0.276 | Improved |
| add Lcls stage after Lsem (ablation) | SLARM · Waymo Open Dataset val (Fig 5b) | Flow EPE: 0.276 -> 0.240 | Improved |
| raise motion expansion order from 1 to 3 (ablation) | SLARM · Waymo Open Dataset val (Fig 5c) | Flow EPE: 0.264 -> 0.240 | Improved |
| raise motion expansion order from 3 to 4 (ablation) | SLARM · Waymo Open Dataset val (Fig 5c) | Flow EPE: 0.240 -> 0.240 | No clear effect |
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.
- GS-LRM* — GS-LRM: Large Reconstruction Model for 3D Gaussian Splatting (ECCV 2024) · 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
- IGGT4D: Streaming 4D Instance-Grounded Geometry Transformer (NeurIPS 2026 · similarity 0.85)
- 4D3R: Motion-Aware Neural Reconstruction and Rendering of Dynamic Scenes from Monocular Videos (NeurIPS 2025 · similarity 0.85)
- ReMatching Dynamic Reconstruction Flow (ICLR 2025 · similarity 0.85)
- HiCoM: Hierarchical Coherent Motion for Dynamic Streamable Scenes with 3D Gaussian Splatting (NeurIPS 2024 · similarity 0.85)
- ReCon-GS: Continuum-Preserved Guassian Streaming for Fast and Compact Reconstruction of Dynamic Scenes (NeurIPS 2025 · similarity 0.85)
- Feed-Forward Bullet-Time Reconstruction of Dynamic Scenes from Monocular Videos (NeurIPS 2025 · similarity 0.85)
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.
