Ipsotek

Partners

Ipsotek

AI Video Analytics for Mission-Critical Environments

AI video analytics for mission-critical sites since 2001

Ipsotek have specialised in AI video analytics for mission-critical sites since 2001, and now sit within Eviden, an Atos Group business. Their VISuite platform runs on a site’s existing camera network, covering perimeter protection, crowd management, licence plate recognition and traffic monitoring across more than 800 projects worldwide.

Why does ATEC work with Ipsotek?

Context-aware rule engine
reduces false alarms

Ipsotek’s patented rule engine assesses context before confirming an alarm, so a taxi dropping a passenger is treated differently from a driver abandoning a vehicle. False alarm rates fall, and a control room receives a volume of alerts it can act on properly throughout a shift.

Ipsotek's strengths
at a glance

Patented scenario-based rule engine
Automatic scene calibration and rule templates
Runs on infrastrucure already installed on site
Natural language search locates footage
VISense answers open questions about a scene
More than 800 deployments in 44 countries

Ipsotek's strengths in detail

Rule engine that assesses context

The patented Scenario-Based Rule Engine allows 32 rules per camera stream and combines inputs from several sensors before confirming an alarm. It distinguishes a taxi dropping a passenger from a driver abandoning a vehicle outside a building. Alarm volumes remain at a level a control room can realistically act on.

Shorter commissioning with tested rule templates

Scene calibration runs automatically, and Ipsotek supply a library of tested rule templates covering common scenarios. Our engineers configure a large multi-camera deployment in a fraction of the time full hand-tuning would require, applying proven templates across similar camera positions. Commissioning periods shorten, labour costs fall, and a site reaches full protection earlier in the programme.

Supports the existing infrastructure

VISuite is an open platform that runs on the cameras, video management system and PSIM a site already operates. Adding analytics requires no replacement cameras and no dedicated analytics hardware at the edge. A customer with several hundred working cameras improves detection across the estate while retaining the value of the original installation.

Search using natural language in plain English

The AIVA engine accepts everyday language. An operator types ‘find the red car’ or ‘a person carrying a black bag’ and receives results without learning a query syntax first. Investigation times fall sharply, which is valuable during a live incident when information is needed at speed.

Detection without predefined rules

VISense uses vision language models to answer open questions about a scene, such as whether anything unusual is occurring, with no rules or alarms defined beforehand. Large camera networks are brought into coverage quickly. Events falling outside any predefined rule set are still raised to an operator.

More than 800 deployments across 44 countries

More than 800 deployments across 44 countries, including over 160 airports and 200 government and smart city projects. Airports combine very high camera counts, constant movement, mixed lighting and strict detection requirements. That deployment record demonstrates the platform performs reliably at scale in demanding conditions.