Research map / Video / Video-language understanding
WTS: A Pedestrian-Centric Traffic Video Dataset for Fine-grained Spatial-Temporal Understanding
Quan Kong, Yuki Kawana, Rajat Saini, Ashutosh Kumar, Jingjing Pan, Ta Gu, Yohei Ozao, Balazs Opra, Yoichi Sato, Norimasa Kobori
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 |
|---|---|---|---|
| Use a more detailed prompt with a demonstration sample (P-C) instead of a simple task prompt (P-A) (ablation) | Video-LLaMA / Video-ChatGPT (no fine-tune) · WTS | CIDEr: Video-ChatGPT 0.009 -> 0.282 | Improved |
| Fine-tune a Video-LLaMA-style model on WTS (ablation) | VideoLLM · WTS | CIDEr: 0.210 -> 0.363 | Improved |
| Add instance spatial token from target bounding box mask (ablation) | VideoLLM · WTS | CIDEr: 0.363 -> 0.389 | Improved |
| Add instance spatial token from target bounding box mask (ablation) | VideoLLM · WTS | BLEU-4: 0.101 -> 0.121 | Improved |
| Fine-tune on WTS (semantic score) | Instance-VideoLLM vs Video-LLaMA · WTS | LLMScore semantic: 0.008 -> 0.285 | Improved |
| Use LLMScore instead of n-gram metrics to judge long captions | caption evaluation · WTS | agreement with human score: human 0.243 vs LLMScore 0.242 | Improved |
Closest papers by meaning
- TUMTraf VideoQA: Dataset and Benchmark for Unified Spatio-Temporal Video Understanding in Traffic Scenes (ICML 2025 · similarity 0.89)
- eTraM: Event-based Traffic Monitoring Dataset (CVPR 2024 · similarity 0.88)
- STSBench: A Spatio-temporal Scenario Benchmark for Multi-modal Large Language Models in Autonomous Driving (NeurIPS 2025 · similarity 0.87)
- Traffic Scene Parsing through the TSP6K Dataset (CVPR 2024 · similarity 0.87)
- RoadSocial: A Diverse VideoQA Dataset and Benchmark for Road Event Understanding from Social Video Narratives (CVPR 2025 · similarity 0.87)
- UrbanIng-V2X: A Large-Scale Multi-Vehicle, Multi-Infrastructure Dataset Across Multiple Intersections for Cooperative Perception (NeurIPS 2025 · similarity 0.86)
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.
