A Guide to Hikvision DeepInMind Technology

Hikvision's DeepInMind revolutionises what CCTV products can do and how they can help provide additional features and functionality to your security infrastructure. Using deep learning technology that trains the system to perform human-like tasks such as speech recognition, defining certain images and objects including vehicles, licence plates and humans and even the ability to describe content.

This deep learning has moved into the security industry too which enables certain supported systems to offer enhanced site security and operational efficiency. These attributes include determining the difference between moving tree branches vs animals' vs humans to help prevent false alarms, vehicle and licence plate recognition for automating security operations when entering a secured location (i.e. opening gates for employees, etc.), attendance tracking using facial recognition plus much more.

What DeepInMind offers

Analysing videos and images captured from a CCTV camera on an NVR (Network Video Recorder) in a way that can make informed decisions on your behalf based on captured content is now easily available thanks to Hikvision's DeepInMind NVR systems. Using incredibly powerful computer components that stand above the competition, the Hikvision DeepInMind NVR series claims up to 90 percent accuracy in false alarm reduction and intelligent facial recognition technology, and the more data that gets fed to the system, the more accurate it will become over time.

Intelligent features also include perimeter protection which can smartly analyse when borders are crossed and search for intruding targets using the built-in smart analysis functionality. Models such as the 6700, 7700 and 9600 series can do either perimeter protector or facial recognition, whereas the 96000 series can do both.

That's great, but what else can it do?

People

Using artificial intelligence algorithms, DeepInMind can be used to analyse face shape data for more accurate facial recognition, to whether or not a person is wearing a mask and can recommend a mask is equipped on the premises to forcing a user to wear a mask before granting entry if that is a requirement of the entered location.

This can also be used for construction sites to detect whether a person is wearing a hard hat or not, and for their safety prompt a notification to ensure one is worn.

Another use for this technology is crowd control, detecting when there are too many people within a gathering (predetermined) and alert that people move or the number within the crowd is reduced.

Objects

Vehicle detection is a commonly desired feature of camera systems, and with DeepInMind this can be used not only for the previously mentioned points regarding vehicle and licence plate recognition, but can also branch out to car park management. Often during peak hours, car parks can become full of spaces all the more difficult to find, stressing drivers and potentially creating unwanted panic situations.