How it works
Scylla is optimized to work on multiple video streams using a single GPU and provide real-time event tracking.
Once an anomalous event has been recognized based on the series of frames given to the model, Scylla sends alerts to all assigned endpoints.
The module supports real-time multiple stream processing as well as offline analysis of video recordings.
The models are trained on a large amount of anomalous and normal videos. This allows Scylla to operate in versatile environments and scenarios to immediately react and send alerts in case of an anomaly.
The system continuously self-learns, that’s why it can be adjusted specifically for your case, if the current dataset is not fully capturing the peculiarities of the environment at your premises.
What makes Scylla Behavior Recognition & Anomaly Detection System stand out
Real-time detection of fighting and anomalous consumer behavior that may result in shoplifting
Centralized solution capable of handling multiple camera streams simultaneously
Flexible deployment based on the customer’s choice - can be on-site, cloud-hosted or hybrid