Research map / Video / Video-language understanding
EatVid-Bench: A Multimodal Fine-Grained Eating Behavior Video Dataset
Xiangpeng Zheng, Zhenbo Xu, Gong Huang, Zhu Li, Qinghong Yang
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 annotation-guided CoT supervision on top of plain LoRA (ablation) | Qwen2.5-VL-7B · EatVid-QA Hard subset (4 FPS) | Non-SA-WAcc (%): 22.3 -> 36.4 | Improved |
| add plain LoRA fine-tuning on EatVid data (ablation) | Qwen2.5-VL-7B · EatVid-QA Hard subset (4 FPS) | Non-SA-WAcc (%): 14.1 -> 22.3 | Improved |
| fine-tune with AGCoT-LoRA | Qwen2.5-VL-7B · EatVid-QA Hard subset | Acc: 0.2500 -> 0.4069 | Improved |
| fine-tune with AGCoT-LoRA | Qwen2.5-VL-7B · EatVid-QA Hard subset | Non-SA-WAcc: 0.1623 -> 0.3214 | Improved |
| fine-tune with AGCoT-LoRA | Qwen2.5-VL-7B · EatVid-QA Hard subset | WAcc: 0.5798 -> 0.6130 | Improved |
| use larger API model instead of 7B open model | Qwen3.5-35B-A3B · EatVid-QA Hard subset | WAcc: 0.6130 -> 0.7359 | Improved |
| increase frame density from 32 to 64 frames or 1-4 FPS | Video-LLMs · EatVid-QA | accuracy | No clear effect |
Closest papers by meaning
- VideoAutoArena: An Automated Arena for Evaluating Large Multimodal Models in Video Analysis through User Simulation (CVPR 2025 · similarity 0.85)
- LongVideoBench: A Benchmark for Long-context Interleaved Video-Language Understanding (NeurIPS 2024 · similarity 0.85)
- Pic2Prep: A Multimodal Conversational Agent for Cooking Assistance (AAAI 2025 · similarity 0.85)
- HourVideo: 1-Hour Video-Language Understanding (NeurIPS 2024 · similarity 0.84)
- Unleashing Hour-Scale Video Training for Long Video-Language Understanding (NeurIPS 2025 · similarity 0.84)
- Exploring Domain Incremental Video Highlights Detection with the LiveFood Benchmark (AAAI 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.
