UCVL converts UCF-Crime and UCF-Crime Annotation into six multimodal QA tasks, showing that MLLMs perform poorly on anomaly perception and that fine-tuning on the benchmark recovers much of the gap.
Negative sample matters: A renaissance of metric learning for temporal grounding,
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
citation-role summary
background 1
citation-polarity summary
fields
cs.CV 1years
2025 1verdicts
CONDITIONAL 1roles
background 1polarities
background 1representative citing papers
citing papers explorer
-
A Benchmark for Crime Surveillance Video Analysis with Large Models
UCVL converts UCF-Crime and UCF-Crime Annotation into six multimodal QA tasks, showing that MLLMs perform poorly on anomaly perception and that fine-tuning on the benchmark recovers much of the gap.