Hand Gesture Recognition Pad Using an Array of Inductive Sensors

Gesture recognition is a field of study within human-computer interaction technology and is used in an increasing number of applications. So far, the systems designed for gesture recognition have been able to sense multiple static and dynamic gestures. However, they suffer from some limitations. This study introduces a novel sensing pad system consisting of an array of inductive sensors which can recognize and differentiate specific static hand gestures through machine learning algorithms (MLAs). It is designed to be a non-contact apparatus where the gestures made by a user can be perceived by the system. It uses five coils, one for each finger, and can sense the fingers that are unfolded while making a particular gesture. Ten volunteer users participated in this study. Ten gestures, numbers 1-10 of the American Sign Language (ASL) are chosen to be tested upon, ten times each for every user. The responses from the sensing coils were measured via a data acquisition board and sent to the PC for processing. A total of 1000 responses were recorded and processed using MLAs which provided an accuracy of 98.7% using fivefold cross-validation (5F-CV) and 97.3% using leave-one-subject-out CV (LOSO-CV) proving that the system can successfully distinguish hand gestures instantly.

Automated summary

Gesture recognition is a growing field in human-computer interaction technology. This study introduces a new sensing pad system that uses machine learning algorithms to recognize specific hand gestures. It achieves high accuracy in distinguishing gestures and is designed to be a non-contact apparatus.