Study of the best view selection based on a human observer

Abstract

In this paper we present results from a questionnaire focused on searching for the best view on virtual 3D objects. For our research we collected a database of virtual models from different categories and we tried to find out if the best view selected by respondents varies from category to category. We performed a clustering using the von Mises-Fisher mixture model to find out the means of each group of selected views. Further we would like to determine whether we can predict the best view similar to the one chosen by the questionnaire just by categorizing examined object.