This thesis contains a description of a software process, the result of which is a web application. It records data about the movement of the eyes across the screen and in real time creates temperature maps from them. Data can be exported to an image or by machine readable form. With the help of a secure administration interface, it enables create datasets that can be distributed to other users. It is part of the job analysis of existing solutions, system design, implementation description, evaluation testing and suggestions for future improvements.
GazeCloudAPI is a javascript library that is available on Github. It uses a computer camera and requires calibration. In the initial calibration phase, 17 points are taken. The user must not move his head, only his eyes. It is in the second phase the user is asked to fix the gaze and move only the head. Another 16 points is removed, of which the last 4 are on a black background (all the previous ones were on white). A demo version of the application is available at gazerecorder.com. You can choose a set of images from your computer as input and set the time they will last displayed. As part of the beta test, it is also possible to load the web page. After the application will generate one video once the capture is complete. You can only export a video that shows how the user interacts with the website. Illustrated is also the generation of a temperature map. The application works reliably. Even if the user has glasses in which reflections of the surroundings can be seen through the glass. The disadvantage of the solution is the resulting data presentation. Everything is combined into one video. The application does not differentiate the heatmap for different source files. The resulting data is not machine-readable. Unfortunately, there is no license on the website the author is only a registration form where the reason for use must be stated libraries. The library works in 95 percent of browsers to date [12]. The most sensitive Internet Explorer users may experience problems when deployed. It cannot be used due to the non-support of the Stream API by the browser
