3D Object Exploration Using Viewpoint and Mesh Saliency Entropies

This paper introduces a technique to inspect a 3D object in a scene with minimal loss of information. We exploit the concept of the viewpoint entropy and introduce a novel view descriptor called mesh saliency entropy to explore the object by finding a minimal set of camera positions that covers the maximum information. Here we present a greedy choice algorithm which tries to detect a sub-optimal N-best views using the combination of mesh saliency entropy and viewpoint entropy to perceive the information communicated by the object. The main contribution is that the object can be examined with minimal loss of information for which can be used in scientific and medical analysis.