SteelBench: Evaluating Vision-Language Models in Real-World Industrial Environments
Existing video benchmarks evaluate action recognition on consumer videos, egocentric recordings, or simulated industrial environments. They do not test vision-language models under the visual and procedural conditions of real industrial CCTV, where workers appear as distant figures amid dust, steam, low light, glare, occlusion, and overlapping activities. We introduce STEELBENCH, a diagnostic benchmark for industrial surveillance that jointly evaluates per-worker activity recognition, safety-rule reasoning, and annotation provenance. SteelBench contains 1,345 densely annotated clips, curated f
Lineage graph
Paper → model → repo connections mined from source citations (Tier-1 exact match).
Why these links exist
Every edge carries a method, confidence, and the source snippet that justified it — so bad links are debuggable.
- PossiblePossibly related (embedding) · 50%VioletVision-3B →
- PossiblePossibly related (embedding) · 48%vlm-starter →
- PossiblePossibly related (embedding) · 47%darkdevil3610/100-AI-Machine-learning-Deep-learning-Computer-vision-NLP →
- PossiblePossibly related (embedding) · 47%Blaizzy/mlx-vlm →
- PossiblePossibly related (embedding) · 46%xlang-ai/OSWorld →
- LinkedLinked via arxiv author · 85%Suryanarayana Reddy Yarrabothula →
“SteelBench: Evaluating Vision-Language Models in Real-World Industrial Environments”
- LinkedLinked via arxiv author · 85%Manisha Chawla →
“SteelBench: Evaluating Vision-Language Models in Real-World Industrial Environments”
- LinkedLinked via arxiv author · 85%Kunal Sinha →
“SteelBench: Evaluating Vision-Language Models in Real-World Industrial Environments”