Perceptual models of viewpoint preference

摘要: The question of what are good views a 3D object has been addressed by numerous researchers in perception, computer vision, and graphics. This led to large variety measures for the goodness as well some special-case viewpoint selection algorithms. In this article, we leverage results user study optimize parameters general model goodness, such that fitted can predict people's preferred broad range objects. Our is represented combination attributes known be important view selection, projected area silhouette length. Moreover, framework easily incorporate new future, based on data from our existing study. We demonstrate combined measure number applications, automatically selecting set representative views, optimizing camera orbits pass through avoid bad trackball controls gently guide viewer towards better views.