Automatic View Selection Using Viewpoint Entropy and its Application to Image‐Based Modelling

Abstract

In the last decade a new family of methods, namely Image-Based Rendering, has appeared. These techniques rely on the use of precomputed images to totally or partially substitute the geometric representation of the scene. This allows to obtain realistic renderings even with modest resources. The main problem is the amount of data needed, mainly due to the high redundancy and the high computational cost of capture. In this paper we present a new method to automatically determine the correct camera placement positions in order to obtain a minimal set of views for Image-Based Rendering. The input is a 3D polyhedral model including textures and the output is a set of views that sample all visible polygons at an appropriate rate. The viewpoints should cover all visible polygons with an adequate quality, so that we sample the polygons at sufficient rate. This permits to avoid the excessive redundancy of the data existing in several other approaches. We also reduce the cost of the capturing process, as the number of actually computed reference views decreases. The localization of interesting viewpoints is performed with the aid of an information theory-based measure, dubbed viewpoint entropy. This measure is used to determine the amount of information seen from a viewpoint. Next we develop a greedy algorithm to minimize the number of images needed to represent a scene. In contrast to other approaches, our system uses a special preprocess for textures to avoid artifacts appearing in partially occluded textured polygons. Therefore no visible detail of these images is lost.

Upright orientation of man-made objects

“…Automatically selecting good viewpoints for 3D models has always been important, especially for applications such as browsing a huge database of 3D models. Previous work either maximizes the visibility of interesting content using metrics like viewpoint entropy [Vázquez et al 2003], view saliency [Lee et al 2005] or shape distinction [Shilane and Funkhouser 2007], or minimizes visible redundant information such as symmetry [Podolak et al 2006] or similarity [Yamauchi et al 2006]. None of these methods considers shape orientation: due to the rotation-invariant metrics employed, they cannot distinguish between different orientations of a shape around a view direction.…”

Viewpoint information channel for illustrative volume rendering

“…Best view selection algorithms have been applied to computer graphics domains, such as scene understanding and virtual exploration [20,25], and volume visualization [2,4,22,26]. Shannon's information measures, such as entropy and mutual information, have been used in these fields to measure the quality of a viewpoint from which a given scene is rendered.…”

An Introduction to Viewpoint Selection in Volume Visualization

“…There are various criteria for evaluating viewpoint, In Vázquez's opinion [17], the Viewpoint Entropy (VE) is the proper criteria for viewpoint selection. They comparing many different criteria, simple ones to Fleishman's [18], which is the number of faces properly captured.…”

“…They decompose the volume into a set of feature Interval Volume components, and use the surface-based view point selection method suggested in [2,17] to find the optimal view for each component. Then they calculate the globally optimal view by a compromise between the local optimal views of all the feature components.…”