Evolution of camerawork in automatic comic generation using interactive genetic algorithm with feedbacks from the user model

In our previous systems for automatic comic generation, camerawork parameters were determined by using an interactive genetic algorithm (IGA) with direct user feedbacks or was determined manually through an equipped user interface. The former approach periodically needs direct feedbacks, on multiple comics generated during evolution, from the user, imposing a burden on the the user. The latter approach asks the user to manually perform the camerawork setting for every frame, which lacks diversity because only one comic is produced. In this work, we propose a method that combines the above two: ask the user to manually set the camerawork for each frame, construct the user model based on the determined camerawork parameters, and use the user model, instead of the user themselves, to give feedbacks in IGA. We aim at reducing the user´s burden while achieving highly satisfactory resulting comics.