Secure authentication on smartphone users is regarded as important. In this research, we aim to perform person authentication using behavioral features of fingers by touch operation on smartphone more safely by applying Support Vector Machine (SVM). As the method of this experiment, examinees perform four kinds of touch operations; Single tap, Double tap, Swipe, Rotation 20 times each, and the results are registered. One or two registered data are extracted and taken as test data and SVM is applied to identify whether or not the person is the principal, and the authentication rate is evaluated. As a result, when single kind of touch operation is used, the authentication rate is as low as 60 to 85%, but when two kinds of operations are used in combination, the maximum authentication rate exceeds 90% and it is enough to be used for personal authentication. The experiments show that as the times of test operations and the number of registered data increase, the authentication rate becomes higher. However, by considering the practicality such as the burden when the user register operations, one or two test operations and at most 20 registered data is considered optimal.

Published in: World Congress on Internet Security (WorldCIS-2017)

  • Date of Conference: 11-14 December 2017
  • DOI: 10.2053/WorldCIS.2017.0014
  • ISBN: 978-1-908320-81-0
  • Conference Location: University of Cambridge, UK