Development of a hand gesture recognition-based system for interaction with video games

Abstract

The topic of this paper is the application of hand gesture recognition for interaction with video games, with particular emphasis on first-person shooter games and the use of light guns. The proposed system employs the MediaPipe software library for detecting hand landmark points, while gesture recognition is performed using deep learning techniques. In the final part of the paper, empirical testing was conducted across various application scenarios, and the results indicate that this approach has potential for future use in video game control systems.

Keywords: Deep Learning, Hand Gestures, MediaPipe, Computer Vision, Light Gun, First-Person Video Games

Published on website: 22.7.2025

Attached files: azivkovic.pdf