Public speaking words (Indian sign language) recognition using EMG

Abstract

Hearing and speech impaired people uses Sign Language (hand gesture language) instead as a tool for communication instead of verbal communication. There are different sign languages used by different countries. Indian sign language (ISL) is the hand gesture recognition language used by Indians. Electromyography (EMG) detects the electric potential generated by muscle cells when these cells are electrically or neurologically activated. EMG data from the subjects who are performing ISL were collected using BIOPAC MP-45. From the collected EMG signals features like mean absolute value (MAV), simple square integral (SSI), standard deviation (STD), root mean square (RMS), average amplitude change (AAC), maximum (MAX), minimum (MIN), average power in the channel (P) are extracted using MATLAB algorithm. Support vector machine classifier is used to classify the different gestures. The study shows that real-time ISL recognition system using EMG has an accuracy of 90%.