Gesture Recognition with non-contact sensor for Natural User Interface in the COVID-19 era

Abstract

COBID-19, an infectious disease transmitted by droplet and contact, is prevalent. In order to reduce the risk of contact infection, various operations should be performed in silence and non-contact. A user interface using non-contact sensors is effective in such an environment. Among them, Natural User Interface based on Gesture Recognition using non-contact sensors are useful, we think. We have developed our NUI system in which the user instructs the computer in a full-body gesture. In this paper, we discuss several methods available for gesture recognition based on skeleton recognition. And, for some of the gesture recognition systems we have implemented with the combination of such methods, the design policy and experimental results of each are presented.