Keywords: Biometrics, Electromyogram, Personal Authentication
Biometrics is a technology that recognizes user’s information by using unique physical features of
his or her body such as face, fingerprint, and iris. It also uses behavioral features such as signature,
electrocardiogram (ECG), electromyogram (EMG), and electroencephalogram (EEG). Among them,
the EMG signal is a sign generated when the muscles move, which can be used in various fields
such as motion recognition, personal identification, and disease diagnosis. In this paper, we analyze
EMG-based biometrics and implement a motion recognition and personal identification system. The
system extracted features using non-uniform filter bank and Waveform Length (WL), and reduced
the dimension using Principal Component Analysis (PCA) and Linear Discriminant Analysis (LDA).
Afterward, it classified the features using Euclidean Distance (ED), Support Vector Machine (SVM),
and K Nearest Neighbors (KNN). As a result of the motion recognition experiment, 95% of acquired
EMG data and 84.66% of UCI data were obtained, and as a result of the personal recognition experiment,
85% of acquired EMG data and 88.66% of UCI data were obtained.