Techniques and Applications of Deformable Surfaces

This paper presents a technique for local surface extraction from scalar and vector fields using deformable surfaces. The goal of surface extraction is often to find features, or meaningful patterns in large datasets. To accomplish this, there is a wide spectrum of techniques. On one end of the spectrum, there are purely symbolic techniques, which may use abstract, iconic representations of features. On the other end of the spectrum, there are purely geometric techniques, such as iso-surfaces, which represent features in a naturalistic way, with a wealth of geometric details. Our technique lies in between, since it may represent as many geometric details as the user desires, or give a simplified representation with little computational effort. Using deformable surfaces consists of several steps. First, a region of interest is selected using selection criteria, and an initial surface is placed in the selected region. Then, the surface is gradually deformed by displacing its nodes according to local deformation criteria. Depending on the criteria specified, this deformation process can result in very different types of surfaces. The versatility of the technique is illustrated by two applications: extraction of recirculation zones and vortex tubes from flow fields.