The system is intended for human pose estimation in 3D space from a single 2D surveillance camera image.
Application area
- The developed algorithm can be applied to the automatic analysis of video sequences from surveillance cameras to identify dangerous behavioral patterns, as well as to enhance the effectiveness of online training and in the rehabilitation process of people after medical operations.
Main advantages
- The development utilizes an ultra-fast and precise model of human body positioning, which ensures high performance across various hardware configurations. An automatic system calibration process has been implemented, which recognizes the physical parameters of the human body for the algorithm to function correctly.
Contacts:
Nikita Tomashevich
Akim Golubenak
Euphrosyne Polotskaya State University of Polotsk
29 Blokhin Street, Novopolotsk
211440
Belarus
Tel.: +375 29 940 04 04
