Music Recommendation System Driven by Interaction between User and Personified Agent Using Speech Recognition, Synthesized Voice and Facial Expression
Ayumi Matsui, Miki Sakurai, Taro Asada, Yasunari Yoshitomi, Masayoshi Tabuse · Proceedings of International Conference on Artificial Life and Robotics · 2021
We propose a music recommendation system that is characterized by an interaction between a user and a personified agent.Speech is recognized using a speech recognition system called Julius, and the facial expression of the agent is then synthesized using preset parameters that depend on each vowel.We add a new function that changes the facial expression of the agent according to the response of the user to music recommended by our previously proposed system.The effectiveness of the proposed system is verified.