Fire360 introduces a 50-hour 360-degree video benchmark where vision-language models trail humans by over 40 points on recognizing fire-damaged objects and answering safety questions.
The epic- kitchens dataset: Collection, challenges and baselines
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CV 1years
2025 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
Fire360: A Benchmark for Robust Perception and Episodic Memory in Degraded 360-Degree Firefighting Videos
Fire360 introduces a 50-hour 360-degree video benchmark where vision-language models trail humans by over 40 points on recognizing fire-damaged objects and answering safety questions.