To put up a stronger fight against COVID-19, Aerialtronics, a subsidiary of the Drone Volt group, has developed AI-based software for the detection of the wearing of protective masks. Developed using deep learning through the training of a neural network, this solution can be used in many other applications.
Using an already installed IP surveillance camera or camera network, this software solution allows the real-time verification of whether an individual is or is not wearing a mask.
The technology employed here consists of processing the video stream from each camera. The principle of the algorithm is based on the detection of the face as an object as a first step followed by the classification as “face with mask” or “without mask” as the second step.
This system therefore does not involve biometry or facial recognition. A “Privacy Mask” option allows the blurring of faces in order to guarantee the anonymity of persons.
This solution is available in DESKTOP mode: the software is integrated into the client’s information system in order to receive and process the video stream and extract the statistics. It is also available in Software as a Service (SaaS) mode in the Cloud: the subscription through a Software as a Service (SaaS) offer can be varied depending on the usage and functions desired by the client.
According to the company, this highly flexible system enables the client to avoid IT maintenance and the purchase of expensive IT equipment.
Moreover, results can be displayed and exported in any form (statistical reports, percentages etc.).