Konferensartikel

The aim is to develop an Interactive Evolutionary Computation (IEC) system that can automatically generate designs or media content based on the preferences of multiple users by capturing their gaze movements and by using this as a solution to evaluation. In the beginning, experiments conducted have revealed that stimulated images to subjects managed to catch the movements of their gaze. Consequently, it is found that the images with the longest look and the greatest preference by the subjects occurred at a rate of approximately 63%. This suggests that the subjects choose their favourite images. Furthermore, the IEC’s algorithmic performance is simulated based on the results from the experiment. Consequently, it is verified that solutions using the IEC algorithm have partly evolved with a voting accuracy of 63%. From these observations, it is expected that the IEC system is built to catch the movements of the users’ gaze and the evolution in solutions are preferred by multiple users by making users more conscious in advance.

Multiple user preferences, interactive evolutionary computation, gaze movements