Use Mouse Ballistic Movement for User Authentication Based on Hilbert-Huang Transform

摘要

In order to explore the frequency domain characteristics of mouse operation for user authentication. This paper collected experimental data on mouse ballistic movements of 10 participants on the AML website. Hilbert-Huang transform was used to extract the frequency-domain information of 9 features such as speed and acceleration during mouse movement, and formed a frequency-domain feature matrix. The Bagged-tree algorithm was used to build an authentication model. The method proposed in this paper obtained Precision = 90.25%, Recall = 88.20%. The results show that there are differences in the frequency domain information when different users operate the mouse to complete the same task, which can be used for user authentication.