Research and development of IMU sensors-based approach for sign language gesture recognition

Abstract

This paper discusses a few Inertial Measurement Unit (IMU) sensor-based approaches for sign language gesture recognition. Generally, there are three main research areas for the IMU sensor-based approach which consist of the device structure, sensors fusion algorithm and calibration method, and finally gesture recognition and classification method. The device structure includes the number and placement of the sensors to cover the degrees of freedom. Sensors fusion algorithms, such as complementary filter, Kalman filter, and EKF are implemented to combine a variety of sensors used for data acquisition. Several gesture classification and recognition methods are also reviewed in this paper. Some of the limitations related to sensor-based technique such as device structure and classification technique are discussed as a research gap for future references.