Novosibirsk Scientists Create Face Recognition System for Masked Faces

Researchers at Novosibirsk State Technical University (NSTU) have developed Russia’s first smart home system capable of recognizing faces covered by medical masks using affordable video adapters, according to a press release from the university.

How the System Works

The developers explain that the system can be used at workplaces with access control points. To identify individuals whose faces are partially hidden by masks, a special algorithm was created. This algorithm detects key points in the upper part of the face and uses them to recognize the person. The system scans the visible part of the face and compares it with biometric data stored in its database.

Affordable and Efficient Technology

The system uses standard processors costing from 14,000 rubles—about 20 times cheaper than the video cards typically used in similar systems. The algorithm allows 30-40 people per minute to pass through a turnstile without needing to stop in front of the camera. Additionally, the algorithm cannot be fooled by a photograph or an image on a screen.

Advanced Features and Security

This development not only recognizes faces and objects but also enables intelligent video surveillance and remote identification. The database does not store facial images, but rather digitized values: the distances between key points on the face. The database is created by photographing all employees from different angles and uploading their biometric data. If someone whose data is not in the database tries to enter, the system will flag it.

Easy Implementation and Current Status

According to Ivan Tomilov, Associate Professor at the Department of Automated Control Systems at NSTU, the system can even be run on a regular computer. The project is currently in the pilot testing stage and is planned to last two years.