A Sensor-Based Hand Gesture Recognition System for Japanese Sign Language

In this paper, we propose a sensor-based data acquisition glove for Japanese Sign Language (JSL) hand gesture recognition. Five flex sensors, an Inertial Measurement Unit (IMU), and three Force Sensing Resistors (FSRs) are used to detect the bending degree of fingers and hand movement information. The detected data are transmitted to the computer by an Arduino Micro. The average accuracy of the hand gesture recognition for a single subject, using the Support Vector Machine (SVM) based and the Dynamic Time Wrapping (DTW) based algorithm are 96.9% and 94.5%, respectively. Our proposed system also achieves an average recognition accuracy of about 82.5% for the cross-recognition among three subjects. The experimental results indicate that our proposed system has great potential for JSL hand gesture recognition.