Summary of the project
The goal of this project is to develop a means of generating collections of high-quality, diverse images of the various structures and objects found in volume data sets. Volume data is generally complicated to explore and analyze, and finding the right combinations of viewer parameters and visual properties to best represent underlying details of interest is generally a tedious, time consuming process. Consequently, we plan on combining volume rendering with evolutionary optimization in order to explore the space of viewer properties and color/opacity mappings that can generate visual representations of the volume. We will define quality and diversity criteria based on combinations of data, surface and geometry-based features of the volume structures, and establish optimization objectives and population discriminators based on these criteria. The resulting software prototype will allow the generation of quality, diverse images, illustrating different aspects of the objects contained within the volume dataset. To this extent, we will analyze the potential of various combinations of evolutionary algorithms, feature extraction methods and volume rendering engines, in order to find the ones that best achieve our goals. While, traditionally, the state-of-the-art in this field has mostly targeted a relatively narrow, specialized user base, we plan on addressing much broader categories of users, including those involved in media, visual arts and 3D modelling.
The team:
|
|
|
|
|
|
