Monochrome Glove: A Robust Real-Time Hand Gesture Recognition Method by using a Fabric Glove with Design of Structured Markers

摘要

This paper presents a method for recognizing human-hand postures in real time, even if environment light cannot be configured appropriately. The key technology is a monochrome glove that is patterned with augmented reality (AR) markers on its palm and is also designed with a structured marker on each finger. As the glove only uses white color for the design of the patterns, it can achieve robust gesture recognition in a natural lighting environment by using a single camera to track a hand wearing the glove. The structured marker facilitates the recognition of a fingertip. Our system can recognize at least 96 hand postures by using the features of the fingers. The AR pattern can be of various configurations. Thus, by utilizing different patterned gloves, we can represent different interactions from the same hand posture. The extensive experiments we conducted demonstrate the accuracy, efficiency, and robustness of our gesture recognition method.