New techniques are presented for rendering complex hierarchical skeletal implicit models in several pen-and-ink styles. A particle system is employed to find interesting areas on the surface and perform stroke stylization guided by local shape features. Interesting areas include silhouette strokes and lines following local shape features, such as those caused by CSG junctions. Hidden line removal is performed either by applying a surfel technique for rapid prototyping or, more accurately, by using ray tracing. Examples drawn from simple to complex models illustrate the capabilities of our system.

Laser scanning is an innovative method of measuring tissue volume that combines precision and reproducibility and may have clinical utility for measuring lymphedema.

A method is proposed for photo-realistic modeling and visualization of a growing tree. Recent visualization methods have focused on producing smoothly blending branching structures, however, these methods fail to account for the inclusion of non-smooth features such as branch bark ridges and bud scale scars. These features constitute an important visual aspect of a living tree, and are also observed to vary over time. The proposed method incorporates these features by using an hierarchical implicit modeling system, which provides a variety of tools for combining surface components in both smooth and non smooth configurations. A procedural interface to this system supports the use of inverse modeling, which is a global-to-local methodology, where the local properties of plant organs are described in terms of their global position within the tree architecture. Inverse modeling is used to define both the tree structure at any time, and a continuous developmental sequence for the tree from a seedling. These techniques provide an intuitive paradigm for the definition of complex tree growth sequences and their subsequent visualization, based solely on observed phenomena. Thus, a key advantage is that they do not require any knowledge of, or simulation of, the underlying biological processes.

One of the major applications of implicit suface modeling systems has been the generation of cartoon-like characters. Recently, additional modeling methods have been combined with implicit sufaces to create much more complex models. These methods include constructive solid geometry (CSG), warping, and two-dimensional texture mapping (among others). The BlobTree has been introduced to organize all of these elements into a single structure which allows both local and global applications of each of these techniques in a general and intuitive fashion. The BlobTree lends itself well to rapid and direct specijication of complex models, however current implementations of the BlobTree have not been engineered for eficiency, and peform poorly when attempting to render large models. In this work we apply established techniques, such as spatial subdivision and tree optimization, to the BlobTree. The objective is to increase efJiciency during rendering without restricting the functionality of the BlobTree as a modeling tool.

No abstract