A Dance Training System that Maps Self-Images onto an Instruction Video

摘要

Owing to recent advancements in motion capture technologies, physical exercise systems that use human interaction technologies have been attracting a great deal of attention. There are already various approaches in place that support motion training methods by using motion capture technology and wearable sensors to analyze body motion. In this study, our basic idea was to change the appearance of a dancer in an instruction video to that of the user, who we assume would be interested in seeing what they would look like if they could perform so well. We developed a motion training system that maps the user's image onto an instruction video. Evaluation results demonstrated that our proposed method is effective for motion training in specific situations.