Human attention modeling in a human-machine interface based on the incorporation of contextual features in a Bayesian network

Human attention can only be inferred from certain causal clues. Such an inference process is of high uncertainty. Bayesian network (BN) is often used for modeling such a process; specifically different features that represent human attention can be fused to reach a consistent conclusion. Previous studies on BN have little consideration of so-called contextual features. In this paper, we propose a few contextual features related to human attention. A novel BN model is then formulated which combines both the contextual features and their corresponding observable behavioral features. At the end, an example is used to illustrate the potential use of the new BN model for human-machine interface design.