Sign Language Interpreter: Classification of Forearm EMG and IMU Signals for Signing Exact English
Soo Pei Yi Jane, Sangit Sasidhar · 2018
The deaf often find themselves unable to converse with hearing people due to the presence of a severe language barrier. In this project, a system that can bridge the communication gap between the hearing and the deaf was designed. Using a single Myo armband with accelerometer, gyroscope, magnetometer and surface electromyography (sEMG) sensors, spatial information on the hand signs are collected and sent to MATLAB for processing. These raw data are filtered using wavelet denoising techniques and segmented using Teager-Kaiser energy operator (TKEO) thresholds. Various time and frequency-domain features are extracted from the processed signal. The tested artificial neural network classifier achieved a average classification rate of 97.12% for a 48 word Signing Exact English (SEE-II) lexicon.