Calipsa helps eliminate user error in video monitoring and analysis. Accurately identifying objects in a video stream and reporting to aid surveillance or statistical observation.
Applying artificial intelligence to video surveillance
Today there are around 250 million video surveillance cameras (CCTV) worldwide. However, the majority of the video surveillance carried out by the police, military, transport operators and security, is still done by humans, making it challenging, time-consuming, expensive and inefficient. Manually viewing huge quantities of video data for many hours can lead to fatigue, loss of attention and errors at a time when video surveillance has never been more critical.
Key facts and stats
250m – number of video surveillance cameras worldwide
566 petabytes – amount of data generated each day by new video surveillance cameras installed in 2015
1:11 – the UK has one camera for every 11 people
53% - the UK Police Scientific and Development Branch (PSDB) found that human operators viewing one, four, six and nine monitors showed accuracy detection scores of 85%, 74%, 58% and 53% respectively in picking up a person with an umbrella.
Built on state of the art Deep Learning models, Calipsa has developed algorithms that can process and analyse hours of video feeds to provide real time alerts and detailed reports for applications including traffic enforcement, road accidents, public disorder and monitoring of critical infrastructure projects.
How and where Calipsa can be used
Municipal – monitoring for traffic flows/congestion, incident detection, enforcement (speeding, tailgating, lane cutting, etc) and surveys
Law enforcement – prison monitoring, dash-mounted, body-mounted cameras
Mass transit – airports, bus and train stations, freight
Military – aerial surveillance, drones, terrain mapping
Critical infrastructure – monitoring power plants, refineries and docks
Commercial – shopping malls, casinos, offices.
The Calipsa engine uses a feedback loop to continuously evolve and improve over time. Human operators can “teach” the artificial intelligence (AI) using a simple point and click interface thereby automating repetitive parts of their jobs.
Designed to work with any existing camera or video source, the technology can be deployed quickly via the cloud or on-premise, with no retrofit required. It is highly adaptable to all weather and lighting conditions, with high levels of accuracy.
Increased efficiency – highlighting relevant or interesting events in video feeds, operators can focus their attention where needed and be alerted to incidents they may have missed
Increased compliance – improved monitoring leads to better compliance, meaning fewer road accidents and improved road safety
Reduced costs – people costs are reduced, operator fatigue is minimised and overall efficiency increased thanks to AI technology
Auditable – all events and alerts are retained, so it is easy to produce evidence in cases of dispute
Continuously learning engine that is adaptable to new tasks, markets and observations