Performance Optimization of Surface Electromyography Based Biometric Sensing System for Both Verification and Identification

Recently, surface electromyography (sEMG) emerged as a novel biometric authentication method. Since EMG system parameters, such the feature extraction methods and number of channels, have been known to affect performances, it is important investigate these effects on performance sEMG-based determine optimal parameters. In this study, three robust methods, Time-domain (TD) feature, Frequency Division Technique (FDT), Autoregressive (AR) their combinations were investigated while channels varying from one eight. For sixteen static wrist hand gestures was systematically in two modes: verification identification. The results 24 participants showed that TD features significantly ( ${p} < 0.05$ ) consistently outperformed FDT AR for all channel numbers. also four-channel setup not different those with higher channels. average equal error rate (EER) sEMG 4% features, 5.3% 10% features. an identification system, Rank-1 (R1E) configuration 3% 12.4% 36.3% electrode position flexor carpi ulnaris (FCU) muscle had critical contribution performance. Thus, combination set electrodes positioned FCU are recommended