An offline LLM builds a pseudo-scene caption memory; online embedding retrieval against that memory yields zero-shot, real-time, explainable video anomaly detection with SOTA scores on UCF-Crime and XD-Violence.
Mcanet: Multimodal caption aware training-free video anomaly detection via large language model
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
-
Flashback: Memory-Driven Zero-shot, Real-time Video Anomaly Detection
An offline LLM builds a pseudo-scene caption memory; online embedding retrieval against that memory yields zero-shot, real-time, explainable video anomaly detection with SOTA scores on UCF-Crime and XD-Violence.