Identification of Manual Alphabets Based Gestures Using s-EMG for Realizing User Authentication

At the present time, since mobile devices such as tablet-type PCs and smart phones have penetrated deeply into our daily lives, an authentication method that prevents shoulder surfing attacks comes to be important. We are investigating a new user authentication method for mobile devices that uses surface electromyogram (s-EMG) signals, not screen touching. The s-EMG signals, which are generated by the electrical activity of muscle fibers during contraction, can be detected over the skin surface, and muscle movement can be differentiated by analyzing the s-EMG signals. Taking advantage of the characteristics, we proposed a method that uses a list of gestures as a password in the previous study. In order to realize this method, we have to prepare a sufficient number of gestures that are used to compose passwords. In this paper, we adopted fingerspelling as candidates of such gestures. We introduced manual kana of the Japanese Sign Language syllabary and compared the identification performance of some candidate sets of feature values with adopting support vector machines.