Learning to Rank Images from Eye movements

Combining multiple information sources can improve the accuracy of search in information retrieval. This paper presents a new image search strategy which combines im- age features together with implicit feedback from users' eye movements, using them to rank images. In order to better deal with larger data sets, we present a perceptron formu- lation of the Ranking Support Vector Machine algorithm. We present initial results on inferring the rank of images presented in a page based on simple image features and implicit feedback of users. The results show that the per- ceptron algorithm improves the results, and that fusing eye movements and image histograms gives better rankings to images than either of these features alone.