A study of an EMG-based authentication algorithm using an Artificial Neural Network

摘要

The typical method of entering a password for user authentication is vulnerable to hacking; therefore, various security technologies using bio-signals, such as iris scan, electrocardiography, electromyography (EMG), and fingerprint recognition, are being developed. In this research, an authentication algorithm using an EMG signal is proposed to supplement the weakness of personal certification techniques. To improve recognition, an artificial neural network clustering algorithm is employed in this study. It includes pre-processing, feature extraction, and classification. Personal authentication is processed based on five parameters extracted from the EMG signal. The proposed algorithm is verified through experiments, demonstrating that it is able to distinguish 81.6% identities of the subjects.