School of Computer Science and Software Engineering
Monash University
Australia
Andrei Sherstyuk
School of Computer Science and Software Engineering
Monash University
Australia
Implicit surfaces obtained by the convolution of mixed dimensional primitives g with some potential function f as shown above, are a generalization of popular implicit surface models: blobs, metaballs and soft objects. These models differ in their choice of potential function but agree upon the use of underlying modeling primitives, namely, points, which puts severe limitations on modeling process and restricts the application base of such models.In this dissertation a method is described for creating and rendering convolution surfaces built upon an expanded set of skeletal primitives: points, line segments, polygons, curves and planes. Analytical solutions to the convolution integral are presented for a number of implicit primitives and potential functions. A comparative analysis for a number of convolution kernels is given.
In addition to conventional techniques, commonly used in implicit modeling, this dissertation describes a set of new modeling methods, which offer a better flexibility for modeling with implicit surfaces.
Finally, an efficient ray-tracing algorithm is presented, which is capable of producing high-quality images of objects modeled with convolution surfaces. The algorithm outperforms all ray-tracing algorithms in its class known to date.
Developing, modeling and rendering issues, discussed in this dissertation are illustrated by original images developed and rendered by the author.
The thesis is available as a PDF, in two-column, one-sided formats.
Many pictures from the thesis appear also on my seafood page
Home