A Novel Chess Interaction System Based on Visual Recognition and Generative Models: Virtual Gestures, Intelligent Assistant, and Multimodal Feedback
Yijin Wang, Jiajun Liang · 2023
With the rapid development of computer vision algorithms and artificial intelligence generative models, interaction methods have become more diverse. This paper proposes a novel international chess interaction system that integrates virtual gesture recognition, large language model (LLM) intelligent assistant, and image-text generation technology to upgrade the traditional chess game experience in three aspects: interaction, decision-making, and feedback layers. We conducted a user survey on the prototype experience system and analyzed it in terms of intelligence and interactivity. In terms of interactivity, our user survey results showed that multimodal generation significantly improves interaction, while virtual interaction technology still needs optimization in usability. In terms of intelligence, the LLM intelligent assistant demonstrated certain usability and potential for higher win-rate improvement.