Research map / Video / Video-language understanding
SE-DETR: Explicit Semantic Exploration for Generalizability and Distinguishability in Video Temporal Grounding
Chengyang Hu, Guanshuo Wang, Fufu Yu, Qiong Jia, Shouhong Ding, Lizhuang Ma
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 Semantics-Proxied Alignment (SPA) (ablation) | DETR baseline · QVHighlights val | VMR R1@0.5: 62.19 -> 67.74 | Improved |
| add Temporal Sparsity Modulation (TSM) (ablation) | DETR baseline · QVHighlights val | VMR R1@0.5: 62.19 -> 64.97 | Improved |
| add Adaptive Feature Fusion (AFF) multi-layer fusion (ablation) | DETR baseline · QVHighlights val | VMR R1@0.5: 62.19 -> 63.47 | Improved |
| add TSM on top of SPA (ablation) | SPA · QVHighlights val | VMR R1@0.5: 67.74 -> 68.80 | Improved |
| add AFF on top of SPA+TSM (ablation) | SPA+TSM · QVHighlights val | VMR R1@0.5: 68.80 -> 69.51 | Improved |
| increase semantic proxies from 100 to 1000 (ablation) | SPA · QVHighlights val | VMR R1@0.7: 51.16 -> 53.61 | Improved |
| increase semantic proxies to 2000 (ablation) | SPA · QVHighlights val | VMR R1@0.5: 69.51 -> 66.97 | Hurt |
| use multi-layer CLIP features (SE-DETR+, K=4) vs single layer | SE-DETR · Charades-STA | mIoU: 48.6 -> 51.2 | Improved |
| use SE-DETR (K=1) vs R2-Tuning (K=4) on Charades-STA | video grounding · Charades-STA | mIoU: 50.9 -> 48.6 | Hurt |
Closest papers by meaning
- Training-free Video Temporal Grounding using Large-scale Pre-trained Models (ECCV 2024 · similarity 0.92)
- Empower Words: DualGround for Structured Phrase and Sentence-Level Temporal Grounding (NeurIPS 2025 · similarity 0.91)
- Decoupling Moment from Event for Video Temporal Grounding (ECCV 2026 · similarity 0.91)
- HiTeA: Hierarchical Temporal Alignment for Training-Free Long-Video Temporal Grounding (ICLR 2026 · similarity 0.91)
- DART: Difficulty-Adaptive Routing for Zero-Shot Video Temporal Grounding (ECCV 2026 · similarity 0.91)
- Universal Video Temporal Grounding with Generative Multi-modal Large Language Models (NeurIPS 2025 · similarity 0.90)
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.
