Russian Scientists Train Systems to Recognize Hand Movements
Researchers develop gesture-control technology that lets users operate devices through standard cameras without additional hardware
Scientists at the St. Petersburg Federal Research Center of the Russian Academy of Sciences have developed software that recognizes hand gestures from video streams. The system works with standard cameras and does not require additional equipment, allowing users to control devices by gesturing directly in front of a screen. The program currently recognizes 34 gestures, enabling actions such as making a call, reacting on social media, or selecting objects on a display. The commands are based on familiar, everyday hand movements.
Accuracy Nears Perfection
The technology is built on a neural network model that constructs a three-dimensional representation of the image, separates hands from the background, and accurately tracks motion even in low-light conditions. According to the developers, recognition accuracy exceeds 99.6%.
The system is suited for use in healthcare, the food industry, and other sectors where maintaining clean surfaces is critical, as well as in environments where equipment is frequently handled with dirty hands.
The software can also be integrated into consumer electronics, industrial equipment, and camera-based services, opening the door to new types of interfaces where traditional buttons play a secondary role.
