A Hand Motion Recognition System Combining Wearable Sensing Gloves and Template Matching Algorithm

Hand motion recognition systems, especially those these using wearable electronic glove, have received growing attention in the field of human-computer interaction and medical rehabilitation. However, two problems still exist: (1) the sensors are usually single, which fails to acquire the multi-sensor information; and (2) Pattern recognition based on deep learning algorithms tend to require a large number of samples to ensure pattern recognition accuracy, whereas obtaining a large amount of hand movement information from a patient is difficult. To address the above problems, this study develops a hand motion recognition system through combining a wearable sensing glove and a new template matching algorithm. The wearable sensing glove employs five Flex sensors and one Inertial Measurement Unit (IMU) to obtain the gesture information, followed by realization of the human-computer interaction through the Unity3D software. Meanwhile, a simple template matching algorithm is constructed for the hand gesture recognition, which can recognize 10 numbers and 20 letters in American Sign Language (ASL) with an accuracy of 90.11% and 87.13%, respectively. The proposed hand motion recognition system can basically meet the needs of patients for hand movement recognition.