Applying Extreme Gradient Boosting for Surface EMG Based Sign Language Recognition

Abstract

Sign languages are used as a non-verbal form of communication that helps people to exchange information. Various sign language recognition systems have been developed to assist healthy and differently-abled people in understanding and conveying messages using sign languages. These recognition systems are mostly developed using computer vision-based algorithms, which have a limitation that their efficiency depends heavily upon surrounding lighting conditions. On the other hand, surface Electromyography is least affected by surrounding lightning conditions, so can be used as an alternative to computer vision based methods. In this paper, we have focused on surface Electromyography signal’s ability to build a reliable sign language recognition system. We have applied boosting based algorithm to build an accurate and reliable classifier to recognize American Sign Language (ASL) using surface Electromyography signals. For, this a new data-set was made by collecting surface Electromyography signals from ten adult-subjects. The classification model was trained using Extreme gradient boosting (XGBoost) algorithm and we obtain an accuracy of 99.09%. © 2022, The Author(s), under exclusive license to Springer Nature Switzerland AG.