Action Classification of Sewing Operations based on Hands Posture and Fingertip Forces

Abstract

This paper describes a classification method of sewing operation using operator’s hands posture and fingertip force. We construct a measurement system which is mainly consists of an RGBD sensor and fingertip force sensors made of strain gauges. The sensors measure operator’s hands posture and fingertip force. We make sewing operations classification model using these two kinds of data. The proof experiment shows that the proposed method can not only classify sewing actions, but can also extract the operator’s motion data in detail.