Foxsys is a security and virtual doorman company located in Uruguay. It provides several
services to residential and office buildings. These include monitoring 8 to 10 cameras per
building.
They have over 2.500 cameras that we keep under control every second.
The main goal was to build a Computer Vision engine that could classify people, vehicles, and
provide a second layer of analysis such as pose detection in a scalable and fast way with a
service oriented architecture, in order to optimize the number alarms that their operators
receive.
Artificial intelligence is the simulation of human intelligence processes by machines, especially computer systems. Specific applications of AI include expert systems, natural language processing, speech recognition and machine vision. (source: searchenterpriseai.techtarget.com)
Computer vision is a field of artificial intelligence (AI) that enables computers and systems to derive meaningful information from digital images, videos and other visual inputs — and take actions or make recommendations based on that information. (source: https://www.ibm.com/)
We have a system that detects desired objects in real time using CV. Our client has more than 2.500 real time cameras. At the same time, it detects and fires alarms to alert the operators about movement! All this in 0.5s! Pretty neat, huh?
We created a custom model optimized for the client. It detects and tracks bicycles, motorcycles, people and “people with strange behavior”.
We manage to reduce up to 50% of the alarms that the operators were receiving.
Using a Nvidia T4 we managed to analyze more than 300 cameras. Applying many of the Nvidia TensorRT optimizations.
Some quick screenshots of the APP