Data-Glove for Japanese Sign Language Training System with Gyro-Sensor

IT research to support communication between people with normal hearing and the hearing impaired is an emergent and important issue. We propose a data glove for a Japanese sign language training system. Currently, data gloves are used with some video games; however, they are expensive and overly complex for Japanese sign language training. We propose a simple data glove with a gyro sensor that can capture the palm-turning gestures of some Japanese sign language words. In the proposed system, an algorithm determines whether a palm-turning gesture has occurred based on experimentally defined thresholds. We conducted experiments with eight participants without individual enrolling to evaluate the proposed data glove. The results demonstrate that the proposed data glove can capture specific words at greater than 50% accuracy on average. We believe that accuracy can be increased by tuning the thresholds to individuals.