We show that accelerometers, touch screens and software keyboards, which are standard components of modern mobile phones, can be used to differentiate different test subjects based on the unique interaction characteristics of each subject. This differentiation ability can be applied to authenticate individuals under a continuous authentication scheme. Based on six 15 minute data sets collected from the test subjects utilizing our data collection platform, we extract multiple features from the data and show an ability to accurately identify individuals at a rate of 83 percent using a simple normal distribution of each feature.
展开▼
