Applying hidden markov models to visual activity analysis for simple digital control panel operations

Abstract

The paper presents an application of Hidden Markov Models (HMM) to fixations’ sequences analysis. The examination concerns eye tracking data gathered during performing simple comparison and decision tasks for four versions of plain control panels. The panels displayed the target and current velocity either on a digital or analog (clock-face) speedometers. Subjects were to decide whether increase or decrease the current speed by pressing the appropriate button. The obtained results suggest that females, generally exhibit different covert attention patterns than men. Moreover, the article demonstrates the estimated four HMM with three hidden states for every examined control panels variant and provides discussion of the outcomes.