Structural Analysis of Human Eye Movement Trajectory

Development of an algorithm for event detection in eye movement trajectory when viewing pictures remains the goal of research by many specialists. The challenge is to develop an algorithm that depends weakly on the user settings and adjusts automatically to the data recorded by the eyetracker. We propose a new principle of eye movement data processing, which uses a two-stage space-time aggregation of gaze points in the coordinate space and velocity space. This approach results in the construction of microevents from which the target events are formed during the further aggregation procedures. All steps of the algorithm are explained in detail and illustrated by examples.