Research map / Video / Video-language understanding
Spatio-Temporal Video Re-Localization by Warp LSTM
Check your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (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 Warp LSTM instead of ConvLSTM (ablation) | STVR model · AVA | frame-mAP (IoU 0.5): 20.8 -> 21.8 | Helped |
| use Warp LSTM instead of ConvLSTM (ablation) | STVR model · UCF-101-24 | frame-mAP (IoU 0.5): 53.5 -> 59.4 | Helped |
| add temporal information propagation with ConvLSTM rather than predicting independently per clip (ablation) | Clip baseline · AVA | frame-mAP (IoU 0.5): 18.8 -> 20.8 | Helped |
| add temporal information propagation with ConvLSTM rather than predicting independently per clip (ablation) | Clip baseline · UCF-101-24 | frame-mAP (IoU 0.5): 52.9 -> 53.5 | Helped |
| replace ConvLSTM warping with accumulated optical flow (ablation) | STVR model · AVA | frame-mAP (IoU 0.5): 20.8 -> 20.2 | Hurt |
| replace ConvLSTM warping with accumulated optical flow (ablation) | STVR model · UCF-101-24 | frame-mAP (IoU 0.5): 53.5 -> 52.0 | Hurt |
| replace polyharmonic interpolation with TrajGRU-style structure generation (ablation) | STVR model · AVA | frame-mAP (IoU 0.5): 20.8 -> 21.0 | No clear effect |
| replace polyharmonic interpolation with TrajGRU-style structure generation (ablation) | STVR model · UCF-101-24 | frame-mAP (IoU 0.5): 53.5 -> 54.8 | Helped |
Closest papers by meaning
- Transferable Video Moment Localization by Moment-Guided Query Prompting (AAAI 2024 · similarity 0.85)
- STPro: Spatial and Temporal Progressive Learning for Weakly Supervised Spatio-Temporal Grounding (CVPR 2025 · similarity 0.84)
- Re-thinking Temporal Search for Long-Form Video Understanding (CVPR 2025 · similarity 0.84)
- Unleashing the Potential of Multimodal LLMs for Zero-Shot Spatio-Temporal Video Grounding (NeurIPS 2025 · similarity 0.84)
- Context-Guided Spatio-Temporal Video Grounding (CVPR 2024 · similarity 0.84)
- End-to-End Spatio-Temporal Action Localisation with Video Transformers (CVPR 2024 · similarity 0.84)
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.