Research map / Video / Video-language understanding
Spatial-TTT: Streaming Visual-based Spatial Intelligence with Test-Time Training
FANGFU LIU, Diankun Wu, Jiawei Chi, Yimo Cai, Yi-Hsin Hung, Xumin Yu, Hao Li, Han Hu, Yongming Rao, Yueqi Duan
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 spatial-predictive 3D conv to TTT layers (ablation) | Spatial-TTT · VSI-Bench | Avg: 62.1 -> 64.4 | Improved |
| add dense scene-description data (ablation) | Spatial-TTT · VSI-Bench | Avg: 61.3 -> 64.4 | Improved |
| interleave self-attention anchor layers (hybrid) with TTT (ablation) | Spatial-TTT · VSI-Bench | Avg: 53.9 -> 64.4 | Improved |
| add spatial-predictive conv: numerical questions (ablation) | Spatial-TTT · VSI-Bench numerical | MRA: 60.7 -> 64.0 | Improved |
| add hybrid architecture: multiple-choice (ablation) | Spatial-TTT · VSI-Bench multiple-choice | accuracy: 52.4 -> 64.8 | Improved |
| use TTT backbone instead of quadratic attention | Spatial-TTT-2B vs Qwen3-VL-2B · 1024 frames | TFLOPs: 1403.1 -> 799.4 | Improved |
| use TTT backbone instead of quadratic attention: memory | Spatial-TTT-2B vs Qwen3-VL-2B · 1024 frames | peak memory GB: 21.2 -> 11.9 | Improved |
| Spatial-TTT vs strongest proprietary on MindCube | Spatial-TTT-2B · MindCube-Tiny | accuracy: 63.9 -> 76.2 | Improved |
| streaming TTT vs context-limited models on long video counting | Spatial-TTT-2B · VSI-SUPER-Count 120min | MRA: 0.0 -> 38.4 | 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.
- Flash-VStream — Flash-VStream: Efficient Real-Time Understanding for Long Video Streams (ICCV 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
- SpaceVista: All-Scale Visual Spatial Reasoning from mm to km (ICML 2026 · similarity 0.90)
- Spatial-MLLM: Boosting MLLM Capabilities in Visual-based Spatial Intelligence (NeurIPS 2025 · similarity 0.87)
- MLLM-4D: Towards Visual-based Spatial-Temporal Intelligence (ICML 2026 · similarity 0.87)
- SpatialLadder: Progressive Training for Spatial Reasoning in Vision-Language Models (ICLR 2026 · similarity 0.87)
- MindJourney: Test-Time Scaling with World Models for Spatial Reasoning (NeurIPS 2025 · similarity 0.87)
- SpatialRGPT: Grounded Spatial Reasoning in Vision-Language Models (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.
