Research map / Video / Video-language understanding
Towards Long-Form Spatio-Temporal Video Grounding
Xin Gu, Bing Fan, Jiali Yao, Zhipeng Zhang, Yan Huang, Cheng Han, Heng Fan, Libo 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 |
|---|---|---|---|
| add all temporal memories without selection (ablation) | ART-STVG · LF-STVG-3min | m_tIoU: 16.7 -> 9.6 | Hurt |
| add temporal memory selection on top of temporal memory (ablation) | ART-STVG · LF-STVG-3min | m_tIoU: 9.6 -> 23.0 | Improved |
| add selective temporal memory vs none (ablation) | ART-STVG · LF-STVG-3min | m_tIoU: 16.7 -> 23.0 | Improved |
| add all spatial memories without selection (ablation) | ART-STVG · LF-STVG-3min | m_tIoU: 21.3 -> 22.1 | Improved |
| add spatial memory selection (ablation) | ART-STVG · LF-STVG-3min | m_tIoU: 22.1 -> 23.0 | Improved |
| use cascaded instead of parallel spatial-temporal decoders (ablation) | ART-STVG · LF-STVG-3min | m_tIoU: 21.5 -> 23.0 | Improved |
| set N_s to 32 instead of 16 (ablation) | ART-STVG · LF-STVG-3min | m_tIoU: 22.7 -> 23.0 | Improved |
| increase N_s from 32 to 48 (ablation) | ART-STVG · LF-STVG-3min | m_tIoU: 23.0 -> 22.5 | Hurt |
| raise temporal similarity threshold from 0.8 to 0.9 (ablation) | ART-STVG · LF-STVG-5min | m_tIoU: 15.0 -> 14.4 | Hurt |
| lower temporal similarity threshold from 0.8 to 0.7 (ablation) | ART-STVG · LF-STVG-1min | m_tIoU: 39.1 -> 38.7 | Hurt |
| add memory modules to baseline | baseline · LF-STVG-3min | m_tIoU: 16.2 -> 23.0 | Improved |
| add memory modules to baseline | baseline · LF-STVG-5min | m_vIoU: 5.3 -> 10.0 | 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.
- TubeDETR — TubeDETR: Spatio-Temporal Video Grounding With Transformers (CVPR 2022) · better on 3 of 3 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
- Knowing Your Target: Target-Aware Transformer Makes Better Spatio-Temporal Video Grounding (ICLR 2025 · similarity 0.92)
- Context-Guided Spatio-Temporal Video Grounding (CVPR 2024 · similarity 0.92)
- STPro: Spatial and Temporal Progressive Learning for Weakly Supervised Spatio-Temporal Grounding (CVPR 2025 · similarity 0.92)
- Unleashing the Potential of Multimodal LLMs for Zero-Shot Spatio-Temporal Video Grounding (NeurIPS 2025 · similarity 0.92)
- Contextual Self-paced Learning for Weakly Supervised Spatio-Temporal Video Grounding (ICLR 2025 · similarity 0.89)
- VideoGrounding-DINO: Towards Open-Vocabulary Spatio-Temporal Video Grounding (CVPR 2024 · 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.
