Interactive experience of singing learning and perception using gesture robots based on artificial intelligence games

Abstract

The traditional way of learning singing has certain limitations, while gesture robots, as a new auxiliary tool, can provide richer interaction methods. Therefore, this study aims to design a gesture robot system that can recognize and learn singing gestures. A gesture model was designed to accurately recognize different gesture actions. Then, an interactive logic based on artificial intelligence games was proposed, allowing users to learn singing gestures by interacting with gesture robots. In terms of static gesture recognition, gesture segmentation and hand posture estimation were carried out. In terms of dynamic gesture recognition, feature extraction, recognition strategies, and loss functions were carried out, and their recognition performance was analyzed. In the interactive experience experiment, users learn singing gestures by interacting with gesture robots. Interactive perception learning systems can adjust learning strategies in real-time based on user gestures. The experiment and result analysis have demonstrated that gesture robots based on artificial intelligence games have great potential in singing learning and perception. The study also proposed an optimization strategy for singing gesture perception to further improve the interactive experience.