A Proposal of a Low-Cost Voice Dialogue System Using Live-Action Video Contents

In recent years, there have been technologies that allow us to reunite and communicate with deceased people in the virtual reality world using artificial intelligence (AI) technology and computer graphics (CG). However, the existing technologies require a long production period and high production costs. In this research, we propose a technology that solves this problem. Specifically, we propose a low-cost system that allows us to reunite and interact with a deceased person in a pseudo-simulated manner by using the principle of the CG voice dialogue system that we developed in the past, changing the CG character to a live-action video content of the subject himself, and creating a voice similar to the subject's voice for speech synthesis processing. In this paper, we have developed a prototype of the system and evaluated the effectiveness of the proposed technology by questionnaire evaluation.