음성을 통한 감정 인식 기반의 가상 에이전트 감정적 반응 생성의 설계 및 구현

초록

This paper presents the design and implementation of a system that generates emotional responses in virtual agents using speech-based emotion recognition technology. By analyzing users’ speech tone and content, the system estimates emotional states in real-time and provides tailored linguistic and non-linguistic responses, strengthening the sense of emotional connection. The proposed system demonstrated its potential to significantly enhance user experiences in mixed reality (MR) environments and showcased practical applications in various scenarios, such as mental health support, education and training, and social interaction enhancement. However, limitations include the lack of user experiments and diverse dataset evaluations, as well as the restriction of agent responses to predefined behaviors. Despite these challenges, this study introduces new possibilities for enhancing emotional interactions with virtual agents and establishes a foundation for future advancements in interactive systems.

키워드