This chapter presents the saliency evaluation approach for visual design of video games, where visual saliency is an important factor to evaluate the impact of visual design on user experience of video games. To introduce visual saliency into game design, we carried out an investigation on several state-of-art saliency estimation methods, and studied on three approaches for saliency estimation: color-based, histogram-based, and information theory based methods. In experiments, these approaches were evaluated on a public saliency dataset and compared with the state-of-art technologies, and it was shown that the proposed information theoretic saliency model can attain a better performance in comparison with several state-of-art methods. Then we applied the information theoretic saliency model to visual game design with image and video examples and demonstrated on how to help game designers to evaluate their visual design with respect to the salience awareness of human visual perception systems.
|Title of host publication||Advances in Low-Level Color Image Processing|
|Editors||M. Emre Celebi, Bogdan Smolka|
|Place of Publication||London|
|Publication status||Published - 17 Dec 2013|
|Name||Lecture Notes in Computational Vision and Biomechanics|