| In this dissertation a Multi-Agent System (MAS) for illustrative rendering of implicit surfaces is presented. The system uses techniques from MAS to advance and combine previous particle-based object-space methods and agent-based image-space approaches. Previous approaches to pen-and-ink style renderings of implicit surfaces were based on particle systems, which were slow to achieve a good distribution of particles across a complex surface, and subsequently trace features. The focus was previously on creating an automatic system for feature identification and illustration.;The method proposed in this thesis develops traditional particles into semi-autonomous agents (called smarticles) that use goal directed behaviours to complete tasks. Tasks include identifying and tracing feature lines, place strokes or stipples, and explore a region. Stipples or short or long, curved or straight, strokes can be placed anywhere on the surface relative to smarticle paths. Smarticles also create colour accents or highlights using a low resolution polygonisation of the object that is also used for hidden line removal. Smarticles use attribute properties associated with individual or groups of primitives that create the model. Using this method any subtree in the data structure that represents the model can be rendered in a particular style. |
