A study on textures and their perceptual visual dimensions as application for flexible and effective scientific visualization

Abstract

The use of textures is fundamental in several areas of Computer Graphics, Computer Vision and Image Processing. In this work, we focus on their main relevant attributes, in order to define and design textures as effective visual representations for use in scientific visualization. We concentrate on the problem of visualizing complex multivariate and multi-dimensional datasets as well as in synthesizing multi-fields and temporal evolution of vectorial datasets visualization. Textures features, such as directionality, color and shape are particularly suited for use in a synthesis algorithm, and they serve as effective seed primitives, which can incorporate many visual dimensions for intuitive and flexible data mapping and encoding. As special application, we propose a level-based visualization approach, with a special focus on systematic layering of information for scientific datasets.