Ukrainian company Hlab.AI has developed a neural network that can find people without masks in the crowd by analyzing videos. It is said that using AI and various data, companies, and governments may predict, detect and analyze abnormal behavior of groups of people or single individual, based on the same geolocation. Onwards, this information can be used to prevent further violations according to the laws of the country where this violation is identified.
Hlab.AI is engaged in the development of AI, it has offices in Ukraine and Germany, and one of its founders is Vitaliy Goncharuk (he is also a founder of Augmented Pixels). He asserts that he is not a supporter of “digital control”, but in the case of epidemics “only taking some measures will allow him to manage without a rollback to the Stone Age.”
According to Goncharuk, the service of face recognition without masks is a point solution for the company, which is engaged in wider tasks, doing business AI. For example, such a service can be used by a company that wants to make sure that its employees do not violate the mode of wearing masks at the company.
Hlab.AI is accurate enough to refuse manual video processing by 80% in such tasks, and by 90% in systems where there are photo databases.
Hlab.AI’s Input (Sensors and Data) is:
● Photos, Videos
● GPS, Gyrometer, Accelerometer
● Text data
● Social Media
● 3D Point Cloud
According to the developer, the team was motivated by curiosity. This technology may be used for medical education, surveillance, workplace, and safety control.
More information can be found in the blog by the developers. The site is created on the Tilda platform.
In this video you can see how this technology works in real-time:
Nota bene: HLab.AI wants to add that technology of determining the presence or absence of a mask doesn’t use or store personal data.
Further use of the results of this work and other HLab.AI technologies is possible only with strict observance of international law, the laws of the country in which this technology is applied and ethical standards.