Cerberus: Real-Time Video Anomaly Detection via Cascaded Vision-Language Models
It addresses the computational cost and unstable performance of vision-language models for real-time deployment in video analytics, offering a practical solution.
The paper tackles the problem of real-time video anomaly detection by introducing Cerberus, a two-stage cascaded system that achieves 57.68 fps on average with a 151.79× speedup and 97.2% accuracy, comparable to state-of-the-art methods.
Video anomaly detection (VAD) has rapidly advanced by recent development of Vision-Language Models (VLMs). While these models offer superior zero-shot detection capabilities, their immense computational cost and unstable visual grounding performance hinder real-time deployment. To overcome these challenges, we introduce Cerberus, a two-stage cascaded system designed for efficient yet accurate real-time VAD. Cerberus learns normal behavioral rules offline, and combines lightweight filtering with fine-grained VLM reasoning during online inference. The performance gains of Cerberus come from two key innovations: motion mask prompting and rule-based deviation detection. The former directs the VLM's attention to regions relevant to motion, while the latter identifies anomalies as deviations from learned norms rather than enumerating possible anomalies. Extensive evaluations on four datasets show that Cerberus on average achieves 57.68 fps on an NVIDIA L40S GPU, a 151.79$\times$ speedup, and 97.2\% accuracy comparable to the state-of-the-art VLM-based VAD methods, establishing it as a practical solution for real-time video analytics.