Enhancing Virtual Human Realism Through Emotion Systems

Project description

This research is part of a broader project aimed at improving the realism of virtual humans by equipping them with emotionally grounded internal systems. Rather than relying solely on dialogue generation, the project explores how virtual humans can maintain coherent emotional states that evolve over time in response to both internal dynamics and external influences.

By integrating affective models, perception mechanisms, and contextual awareness, the project seeks to create virtual humans that behave in more human-like, believable ways. Ultimately, this work contributes to the development of virtual agents that can engage users more naturally in applications such as human–computer interaction, education, simulation, and digital companionship.

Context

This project is situated in the context of research on virtual humans and human–computer interaction, where increasing emotional realism is essential for creating believable and engaging digital agents.

Results

Integrating psychoacoustic perception into virtual humans enhances perceived emotional realism, sensitivity, and responsiveness compared to agents with only internal mood regulation or no affective modeling. While promising, the study’s scope was limited by a single soundscape, small student sample, and isolated evaluation of the perception system.

Future work should test diverse environments, combine perception with mood regulation, use larger samples, and expand emotional modeling to improve virtual human believability.

Downloads