A METHOD FOR EVALUATING THE ATTRACTIVENESS OF PEDESTRIANS IN A SHOPPING STREET WITH VISUAL SALIENCY MAP IN REAL AND VIRTUAL ENVIRONMENTS

This research aims to construct a street attractiveness evaluation method that utilizes various eye-tracking, deep learning, and image-processing technologies and to verify its validity in environments with diverse components. Specifically, we will conduct eye-tracking experiments in both real and VR environments, using the attractiveness of a shopping street for pedestrians as an example. Then, we will use the results to train a deep generative model and create a visualization method for general trends in attractiveness distribution for shopping streets. We will also consider individual differences in attractiveness and the influence of environmental differences on visual saliency.