Motion Primitive Segmentation Based on Cognitive Model in VR-IADL

Recently, many studies have been conducted using virtual reality (VR) technology in the medical field. Our research group has been developing a virtual kitchen challenge (VKC), which is a system to evaluate instrumental activities of daily living (IADL) using VR technology. In the previous study, we focused on motion primitives, which are the smallest unit of motion in VKC. In this study, we focused on the motion of the VKC task when there is no contact with the screen and developed a model that can be segmented by motion primitives. Furthermore, using a two-step process based on time-series data of the subject's fingertip velocity during the VKC task, we developed a model that can be segmented in terms of motion primitives. Therefore, the segmentation accuracy was 83.4%, and the percentage of false positives was as high as 28%.In the future, we plan to revise the feature set.