Research map / Video / Video-language understanding
SpaceVLLM: Endowing Multimodal Large Language Model with Spatio-Temporal Video Grounding Capability
Jiankang Wang, Zhihan Zhang, Zhihang Liu, Yang Li, Jiannan Ge, Hongtao Xie, Yongdong Zhang
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 |
|---|---|---|---|
| replace DETR-like decoder with lightweight Query-Guided Space Head (ablation) | SpaceVLLM · HCSTVG-v1 test | m_vIoU: 35.5 -> 39.5 | Improved |
| replace Historical-Assisted Decoder with simple space head (ablation) | SpaceVLLM · HCSTVG-v1 test | m_vIoU: 38.5 -> 39.5 | Improved |
| add historical-frame fusion via gating (ablation) | SpaceVLLM · HCSTVG-v1 test | m_vIoU: 38.5 vs 39.5 | No clear effect |
| add synthetic V-STG data (ablation) | SpaceVLLM · HCSTVG-v1 test | m_vIoU: 34.4 -> 39.5 | Improved |
| add synthetic V-STG data (ablation) | SpaceVLLM · HCSTVG-v1 test | m_tIoU: 54.3 -> 57.0 | Improved |
| decoupled space head with STA queries vs text coordinate generation (LLaVA-Video-SFT) | LLaVA-Video-7B · HCSTVG-v1 test | m_vIoU: 27.7 -> 39.5 | Improved |
| batched space head vs sequential coordinate text generation | SpaceVLLM vs LLaVA-ST · VidSTG declarative test | avg time per sample (s): 17.63 -> 2.89 | Improved |
| add STA queries and STVG training | LLaVA-Video-SFT · MVBench | accuracy: 61.6 -> 65.5 | Improved |
| add STA queries | LLaVA-Video-SFT · Charades-STA | R1@0.7: 38.4 -> 41.9 | Improved |
| add STA queries and space head | LLaVA-Video-7B · EgoSchema | accuracy: 57.3 -> 57.4 | No clear effect |
| SpaceVLLM vs task-specific CG-STVG | SpaceVLLM · HCSTVG-v2 val | m_vIoU: 39.5 -> 34.3 | Hurt |
Later papers that compared with this method
2 later accepted papers report SpaceVLLM-7B as a baseline in their results tables; 1 of them report a better number on most of the results they share with it.
- STVG-R1: Incentivizing Instance-Level Reasoning and Grounding in Videos via Reinforcement Learning (ICLR 2026 · better on 4 of 5 shared results)
- OmniGround: A Comprehensive Spatio-Temporal Grounding Benchmark for Real-World Complex Scenarios (CVPR 2026 · better on 0 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
- Unleashing the Potential of Multimodal LLMs for Zero-Shot Spatio-Temporal Video Grounding (NeurIPS 2025 · similarity 0.93)
- LLaVA-ST: A Multimodal Large Language Model for Fine-Grained Spatial-Temporal Understanding (CVPR 2025 · similarity 0.91)
- Grounding Everything in Tokens for Multimodal Large Language Models (CVPR 2026 · similarity 0.90)
- VideoGLaMM : A Large Multimodal Model for Pixel-Level Visual Grounding in Videos (CVPR 2025 · similarity 0.89)
- GroundingME: Exposing the Visual Grounding Gap in MLLMs through Multi-Dimensional Evaluation (CVPR 2026 · similarity 0.89)
- Weakly Supervised Temporal Action Localization via Dual-Prior Collaborative Learning Guided by Multimodal Large Language Models (CVPR 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.
