At ACIC, we use Artificial Intelligence to validate the detection of any type of subject

In the image analytics sector, the use of AI is becoming essential. Customer requests are increasing and more and more complex both in the analytics scenarios as in the type of objects to be detected.

On a visible camera, the use of AI is therefore a real advantage for the validation of detection in the VCA. At ACIC, we integrate the Deep Neuronal Network (DNN) into many solutions to meet the most complex demands.

With AI it is now possible to detect and classify all types of objects and subjects in many scenarios. We just need to train our model to detect certain objects such as certain types of vehicles, bicycles, helmets, etc. AI requires the use of a GPU which can lead to additional hardware costs.

Counting vehicles

The use of AI is not systematic. We recommend using it only in the most complex scenarios and on visible camera. In most situations, classic image analytics (foreground / background comparison) works fine. Adding a module of AI only makes it possible to validate the detection before raising an alarm in the VMS. This will therefore increase the reliability of the analysis. AI is most often recommended in counting (link to the AI counting page) and traffic management (link to the TM page) solutions.

Advantages:

  • Detection of any type of object
  • Strengthen the analysis
  • Classification
  • Complex scenarios
  • Expand possibilities.

 

Disadvantages:

  • GPU needed
  • Server only
  • More expensive hardware
  • Need to train the model for specific needs.

To access our commercial brochure please click on the button below

#ACIC #VideoAnalytics