Seamless Human–Machine Interaction in Augmented Reality Gaming: A Low-Cost Webcam Approach

This paper presents a human–machine interaction system that utilizes augmented reality (AR) as an immersive gaming environment, enabling users to interact seamlessly with virtual elements through hand gestures. Two powerful platforms (OpenCV and OpenGL) are employed for video processing, image analysis, and real-time 3D rendering. Unlike existing ArUco-based pipelines, the proposed multi-layer filtering strategy ensures reliable marker detection even under reflective surfaces and dynamic illumination, significantly enhancing system robustness. The AR framework further employs an inexpensive webcam instead of costly RGBD sensors, generating depth information mathematically through pose estimation. This approach not only reduces computational complexity but also achieves high precision in marker detection and identification, outperforming the default OpenCV ArUco library in both speed and accuracy. The system runs efficiently on low-end CPUs equipped with GPUs across multiple video resolutions without compromising performance, providing a cost-effective and accessible solution for realistic AR-based gaming and interaction.