Prediction of search targets from fixations based on weighted average map

The eye is the window of perception and cognition. In the visual search task, the prediction of search targets has become more and more important in theoretical and practical applications. How to express the characteristics of fixation in the process of searching target and predict the visual target effectively is still the difficult point of eye movement prediction. In order to verify this hypothesis, we developed the average map with gaze duration to compute the feature representations of targets respectively in color and grayscale images. According to the weighted map to infer features which can effectively express the patches of the fixation and using the weighted map to compute the feature dictionary. The results show that the weighted average map is used to calculate the target corresponding feature dictionary, which improves the prediction accuracy of eye movement data and reduces the computation time.