Computer Vision-Driven Human-Robot Interaction for Mobile Devices

Robotic advancements are increasingly becoming integrated into everyday human activities, highlighting the criti-cal importance of smooth human-robot interactions. This paper presents a robust visual perception system designed for real-time application on widely-used mobile devices. Utilizing deep learning techniques, this system not only recognizes but also analyzes human actions based on visual data. This analysis helps robots to make well-informed choices when interacting with humans. Our framework's effectiveness is validated on a mid-range smartphone, the Samsung Galaxy A24, where it delivers notable real-time capabilities. The findings underscore the system's capacity to transform human-robot interactions in various practical settings.