Towards improving visual attention models using influencing factors in a video gaming context
Saman Zadtootaghaj, Steven P. Schmidt, Hamed Ahmadi, Sebastian Möller · 2017
Visual attention models (VAMs) have proved useful in various multimedia applications such as video encoding, interface design, and user behavior prediction. However, existing VAMs are not generalizable enough to be utilized in all multimedia applications implying that in each application there exist some unique factors which drive the user attention. Application agnostic VAMs rely on video and image low level features such as motion, intensity, and orientation while for an interactive application such as gaming, other factors also affect the results. There have been few attempts to identify and measure these factors to date. In this paper, we investigate a variety of the influencing factors in three different categories: user, system, and context.