Identifying focus points with intelligent self-writing eye tracker

When perceiving complex, saturated images, the task arises not only to track where the user’s gaze is directed, but also to understand how he processes information on the screen, how he switches between objects and where his attention lingers longer. The effectiveness of perception is predetermined by the effectiveness of the interface and the information model built by the developer, the effectiveness of whose work is determined by the identified accents of the human operator’s attention. An eye tracking toolkit has been developed in Python version 3.10 with the connection of the libraries mediapipe, openCV and matplotlib with extended functionality aimed at improving the accuracy of interpretation of gaze behavior and improving the methods of presenting the collected data. The use of the developed toolkit allows us to determine the areas of interest of users, identify accents of attention, which can serve as a basis for constructing attention maps, which can subsequently help in creating an effective user interface.