Estimation of Positive and Negative Emotion Using Visual Saliency and Gaze Measurement

Abstract

The service called "Google Safe Search" has been provided to hide the inappropriate images on web browser. The inappropriate images which cannot be detected by the Google Safe Search are selected by user's report. However, it is difficult to get the information of inappropriate image from inexperienced PC users. If the emotion (pleasant or unpleasant) of users during web browsing can be estimated automatically, it is judged whether the image shown on the web browser is appropriate or not. We thus develop the emotional estimation method which combines the gaze information with the visual saliency. In this paper, we show the relationship between the emotion and the number of visual fixation in the salient region of an image