Research map / Video / Video-language understanding
OmniVTG: A Large-Scale Dataset and Training Paradigm for Open-World Video Temporal Grounding
Minghang Zheng, Zihao Yin, Yi Yang, Yuxin Peng, Yang Liu
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 multi-task SFT on OmniVTG (ablation) | Qwen2.5-VL-7B · OmniVTG test set (Full) | R1@0.5: 34.5 -> 62.3 | Improved |
| add multi-task SFT on OmniVTG (ablation) | Qwen2.5-VL-7B · ActivityNet Captions | R1@0.5: 13.6 -> 25.6 | Improved |
| add self-correction CoT finetuning after SFT (ablation) | SFT model · OmniVTG test set (Rare) | R1@0.5: 59.6 -> 61.3 | Improved |
| add self-correction CoT finetuning after SFT (ablation) | SFT model · ActivityNet Captions | R1@0.5: 25.6 -> 32.5 | Improved |
| add self-correction CoT finetuning after SFT (ablation) | SFT model · OmniVTG test set (Full) | R1@0.5: 62.3 -> 62.4 | No clear effect |
| add RL on top of SFT + CoT (ablation) | SFT + CoT model · ActivityNet Captions | R1@0.5: 32.5 -> 39.8 | Improved |
| add RL on top of SFT + CoT (ablation) | SFT + CoT model · OmniVTG test set (Full) | R1@0.5: 62.4 -> 63.3 | Improved |
| add CoT finetuning before RL (ablation) | SFT + RL model · ActivityNet Captions | R1@0.5: 37.2 -> 39.8 | Improved |
| increase SFT data from 10% to 100% (ablation) | SFT model · OmniVTG test set (Full) | R1@0.5: 41.9 -> 62.3 | Improved |
| add content-aware reflection instead of no reasoning (ablation) | SFT model · ActivityNet Captions | R1@0.5: 25.6 -> 39.8 | Improved |
| use content-aware rather than rule-based reflection (ablation) | OmniVTG CoT · ActivityNet Captions | R1@0.5: 37.9 -> 39.8 | Improved |
| use content-aware rather than rule-based reflection (ablation) | OmniVTG CoT · OmniVTG test set (Rare) | R1@0.5: 61.0 -> 62.4 | 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.
- Time-R1 — Time-R1: Post-Training Large Vision Language Model for Temporal Video Grounding (NeurIPS 2025) · better on 9 of 9 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
- OmniGround: A Comprehensive Spatio-Temporal Grounding Benchmark for Real-World Complex Scenarios (CVPR 2026 · similarity 0.93)
- VideoGrounding-DINO: Towards Open-Vocabulary Spatio-Temporal Video Grounding (CVPR 2024 · similarity 0.92)
- Unleashing the Potential of Multimodal LLMs for Zero-Shot Spatio-Temporal Video Grounding (NeurIPS 2025 · similarity 0.92)
- Time-R1: Post-Training Large Vision Language Model for Temporal Video Grounding (NeurIPS 2025 · similarity 0.91)
- Empower Words: DualGround for Structured Phrase and Sentence-Level Temporal Grounding (NeurIPS 2025 · similarity 0.90)
- Uncertainty-quantified Rollout Policy Adaptation for Unlabelled Cross-domain Video Temporal Grounding (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.
