Implementation of a sign language primitive framework using EMG and motion sensors

This study extracts a sign language primitive (SLP) for sign recognition using surface electromyography (sEMG) sensor and a motion sensor. Considering the characteristics of gesture communication, the proposed SLP is composed of hand-shape vector and motion direction vector. 12 representative hand-shape primitives are recognized as hand-shape vector by sum of eight-channel muscle tension at brachial 6 hand-movement primitives are classified as motion direction vector by using the Euler angle differences. Sign language word is recognized by the combination of the extracted SLP. We confirm that the proposed primitive would classify the sign word in 5 different sentences.