Research map / Video / Video-language understanding
Vid-Group: Temporal Video Grounding Pretraining from Unlabeled Videos in the Wild
Peijun Bao, Chenqi Kong, Siyuan Yang, Zihao Shao, Xinghao Jiang, Boon Poh Ng, Meng Hwa Er, Alex Kot
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 ReCorrect cleaning, adjustment and memory-consensus correction to GPT-4o pretraining (ablation) | GPT-4o Pretraining · Charades-STA (zero-shot) | mIoU: 41.43 -> 45.63 | Improved |
| add ReCorrect cleaning, adjustment and correction in pretraining plus finetuning (ablation) | GPT-4o Pretraining + Finetuning · Charades-STA (unsupervised) | mIoU: 40.91 -> 48.66 | Improved |
| add SPL pseudo-label refinement to GPT-4o pretraining | GPT-4o Pretraining · Charades-STA (zero-shot) | mIoU: 41.43 -> 42.47 | Improved |
| add VTG-GPT boundary refinement to GPT-4o pretraining | GPT-4o Pretraining · ActivityNet Captions (zero-shot) | mIoU: 33.21 -> 34.47 | Improved |
| use ReCorrect instead of SPL refinement in pretraining | GPT-4o Pretraining · Charades-STA (zero-shot) | mIoU: 42.47 -> 45.63 | Improved |
| add pretraining on Vid-Group before unsupervised finetuning with ReCorrect | ReCorrect Finetuning · Charades-STA (unsupervised) | mIoU: 44.48 -> 48.66 | Improved |
| add pretraining on Vid-Group before finetuning with GPT-4o labels | GPT-4o Finetuning · Charades-STA (unsupervised) | mIoU: 40.91 -> 44.22 | Improved |
| increase pretraining data from 6.3K to 200.3K samples (ablation) | ReCorrect · Charades-STA (zero-shot) | Avg R@m: 29.14 -> 48.74 | Improved |
| concatenate frames into a composite image instead of separate images for GPT-4o (ablation) | GPT-4o labeling · Charades-STA (pilot) | mIoU: 27.53 -> 28.11 | No clear effect |
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.
- VTimeLLM-13B — VTimeLLM: Empower LLM to Grasp Video Moments (CVPR 2024) · better on 2 of 2 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
- Training-free Video Temporal Grounding using Large-scale Pre-trained Models (ECCV 2024 · similarity 0.92)
- VideoGrounding-DINO: Towards Open-Vocabulary Spatio-Temporal Video Grounding (CVPR 2024 · similarity 0.90)
- Uncertainty-quantified Rollout Policy Adaptation for Unlabelled Cross-domain Video Temporal Grounding (NeurIPS 2025 · similarity 0.89)
- Time-R1: Post-Training Large Vision Language Model for Temporal Video Grounding (NeurIPS 2025 · similarity 0.89)
- Universal Video Temporal Grounding with Generative Multi-modal Large Language Models (NeurIPS 2025 · similarity 0.89)
- Empower Words: DualGround for Structured Phrase and Sentence-Level Temporal Grounding (NeurIPS 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.
