Digifort analytics is state-of-the art and optimised to run on NVidia Graphics Processing Units (GPUs) and purpose designed servers. These use CUDA technology to provide the huge processing power required to run multiple analytics channels quickly, efficiently and cost-effectively.
Nick Bowden, MD of Digifort UK, said, “Digifort’s deep learning analytics can be used to identify object types and their colour profile, alerting operators in real-time – such as an intruder in a blue jacket. In Digifort version 7.4, metadata also allows the user to retrospectively search the recorded video again, but this time looking for different objects and events. So, instead of identifying intruders in blue jackets, this time operators can instead look for vans, bikes, cars, trucks, groups of people, bags, cyclists and much more, including with a specific, new, colour profile. Retrospective searching can be carried out on the same recording as many times as required, each time with different search criteria.
“Retrospective searching only considers the metadata content related to the objects in the camera’s view. This makes it extremely fast, processing 30 days of video in just a few minutes. It is game-changing technology, not only for live systems with an immediate operator response but also for the forensic examination of video by users, post event.”