icetana AI Aggressive Behaviour Detection Software

Aggressive behaviour detection uses your existing camera network to analyse live video for the movement patterns that indicate violence. icetana AI evaluates every stream continuously, identifies fights, assaults and vandalism as they occur, and sends an event to the security control room and an alert through the relay agent within seconds.

icetana AI detects around 3 in 4 aggressive behaviour incidents in live deployment. We publish that figure rather than a benchmark accuracy score because it reflects what the system does in a real venue, with real camera angles, crowding and lighting. Detection rate improves as the anomaly engine learns your specific site.

No. The aggression model looks for physical altercation, but icetana's anomaly engine runs alongside it and flags the behaviour around an incident: people running, scattering, changing direction or gathering in numbers that do not fit the normal pattern for that camera. That means chases, standoffs and crowd reactions can generate an alert even where contact is brief or partly obscured.

Yes. icetana AI integrates with Genetec Security Center and Milestone XProtect, and alerts can be routed to email, mobile or your existing control room workflow through the relay agent.

Antara Core, icetana's private AI engine. It runs entirely inside your own infrastructure, so video is analysed on site and no footage is sent to an external provider. That makes aggressive behaviour detection workable in restricted, regulated and air-gapped networks where cloud AI cannot be used at all, and it keeps your video under your own retention policy. Antara Core also runs on a fixed annual fee instead of per-use cloud pricing, so your costs do not climb as you add cameras or as incident volume rises.

No. Detection runs on Antara Core inside your environment, so footage never leaves your infrastructure. Cloud-based detection tools cannot make that commitment, which rules them out for many health, government and correctional sites.