The co.exist software platform consists of two elements:
THE VIDEO ANALYTICS ENGINE CLIENT INCLUDED
Designed and developed by Sprinx. Based on neural networks and on an innovative combination of two technologies, Deep Learning and 3D Object Tracking, which enhances their performance.
THE WEB INTERFACE FOR DATA VISUALIZATION
Designed and developed by AKQA. The visualisation of data has the main objective of being intuitive and usable both from Desktop from mobile devices.
The engine, the client and the webserver can be installed on standard PCs (Linux OS Ubuntu and Intel Core i7-8559U processor or higher depending on the number of cameras to analyse on the same server), while the front-end can run on any PC and mobile browser. The software platform as well as being equipped with its front-end can be easily integrated into more complex system architectures, public and private, through the activation of standard communication protocols (example: HTTP, TCP in Json format, FTP, ModBus , OPC UA, OPC DA) or custom ones. This functionality represents an added value as the solution offered can be easily integrated with other systems already in place or under development. The software platform can process the standard RTSP / Onvif video streams transmitted by the cameras already installed, drastically reducing the investment necessary for the purchase and installation of dedicated field sensors. The use of neural networks trained for behavior analysis applications, based on Intel® OpenVINO ™, integrated with a 3D object tracking approach, allows being particularly adaptive in pre-existing CCTV systems ensuring good performance even in not optimal installation and environmental conditions.