Exploiting Screen-Touch Trajectory for Passive User Authentication in Industrial Internet of Things Systems

This article exploits the spatial–temporal features of user screen-touch trajectory (STT) to develop a user authentication framework for Industrial Internet of Things (IIoT) systems. We first model the STT as a trajectory image and apply the speeded-up robust features (SURF) algorithm for STT spatial feature characterization. We then model the STT as a time series and employ the hidden Markov model (HMM) for the STT temporal feature extraction. We further design a classifier based on HMM for the above temporal feature and also a classifier based on eXtreme Gradient Boosting for the spatial feature. By combining the two classifiers and assigning each classifier an appropriate weight, we develop a passive user authentication framework. The new framework has the potential to significantly impact the IIoT security practices by offering a flexible and efficient authentication method for IIoT systems, and it also can serve as a complementary solution or an enhancement for the traditional authentication mechanism of such systems.