Access control system
The time from the start of recognition to the opening is 2 seconds.
Recognition accuracy is 99.8%.
A unique module for liveness definition based on the calculations in neural networks.
Realtime face recognition in video stream.

Working time tracking from the video stream at the checkpoint and on workplaces.
Resistance to factors that interfere with recognition (glasses, beard, lighting condition).
Integration with any access control system.

Integration with various types of access controllers (RFID reader imitation, Wiegand-26, Wiegand-34 protocols). The use of currently available access control systems.
Integration with the existing base of company employees.
Scalability and failure capabilities
The inclusion of additional video processing servers, checkpoints and video cameras is performed without changing the current configuration. For the database and the system web-interface a cluster-based project can be implemented to increase fault tolerance.

Only actions in the interface are required to migrate cameras between recognition services.
The use of integrated staff database.
User authentication on workstations via AD.

Viewing the staff from the category "Personnel" with searching and filter options. The system can receive information about employees from the existing customer base.

Integration with third-party systems via RESTful, API, SOAP protocols or other APIs provided by the customer.
Video cameras operation
Integration over TCP/IP with cameras from different manufactures (RTSP protocol). Configuring cameras connection via the web interface, monitoring their status, viewing images from a video camera in the web interface. The number of cameras is unlimited (a video stream processing server is required for every few dozen cameras).

Face recognition in the stream. Usage of video accelerators for face recognition (several times increase in productivity).
Web interface operation
Visiting log.
When a visitor is selected his card is displayed. 2 images can be displayed for the visitor: a high-quality photo from the database (if available) and the last photo from the camera at the checkpoint. Identifying access violators.

Face classification:
  • All recognized faces are classified; the solution maintains the recognition database;
  • The category to which every person was assigned is displayed in the reports generated by the solutions;
  • The number of staff categories is unlimited. The classification rules are determined by the customer business cases.
The company map.
The interactive map of the object with online display of staff location and colour tracking movement. Viewing visiting history of selected areas.
(391) 200 32 20
Techvision LLC
Krasnoyarsk, Kapitanskaya Str., 12, Office 204