Machine Learning-Assisted\nGesture Sensor Made with\nGraphene/Carbon Nanotubes for Sign Language Recognition

Abstract

Gesture sensors are essential to collect human movements\nfor human–computer\ninterfaces, but their application is normally hampered by the difficulties\nin achieving high sensitivity and an ultrawide response range simultaneously.\nIn this article, inspired by the spider silk structure in nature,\na novel gesture sensor with a core–shell structure is proposed.\nThe sensor offers a high gauge factor of up to 340 and a wide response\nrange of 60%. Moreover, the sensor combining with a deep learning\ntechnique creates a system for precise gesture recognition. The system\ndemonstrated an impressive 99% accuracy in single gesture recognition\ntests. Meanwhile, by using the sliding window technology and large\nlanguage model, a high performance of 97% accuracy is achieved in\ncontinuous sentence recognition. In summary, the proposed high-performance\nsensor significantly improves the sensitivity and response range of\nthe gesture recognition sensor. Meanwhile, the neural network technology\nis combined to further improve the way of daily communication by\nsign language users.