![]()
Fractal Geometry
In 1984, researchers in the School of Mathematics made a breakthrough in fractal geometry, which can represent shapes that are irregular and broken. That breakthrough has since led to a patented image compression technology that is changing the field of computer graphics. The mathematicians solved "the inverse problem" how the geometry of an actual scene can be used to generate the right algorithm or set of rules to represent that scene.
Courtesy Iterated Systems Inc. Georgia Tech researchers designed a system to automatically analyze the geometry of a given object and regenerate it fractally (such as this image). Fractals allow objects to be represented more efficiently, effectively increasing a computer's capacity to store and access image data.
Using fractal geometry, the researchers were able to determine a natural scene's redundancies, one form of which is self- similarity (shapes of objects remaining the same at different levels of magnification). They then designed a system to automatically analyze the geometry of a given object and regenerate it fractally. Fractals allow objects to be represented more efficiently (that is, with fewer numbers) than other methods, such as pixels. Therefore, fractals effectively increase a computer's capacity to store and access image data.
In 1987, lead researcher and mathematics professor Dr. Michael Barnsley and associate math professor Dr. Alan Sloan co-founded Iterated Systems Inc., an Atlanta company that specializes in digital image science. The next year, the researchers took a leave of absence from Tech to devote their full-time efforts to Iterated Systems. They later resigned from Tech, but Barnsley recently rejoined the faculty as an adjunct professor in the Georgia Tech Center for Dynamical Systems and Nonlinear Studies. He is also a professorial fellow at the University of Melbourne in Australia.
The popularity of Iterated Systems' fractal method of storing and regenerating images grew enormously from 1992 to 1997, when Microsoft used it in the top-selling CD-ROM encyclopedia "Encarta." A more recent technology developed by the company is called StiNGtm. It is a family of core image-handling technologies that includes lossless encoding, patented fractal scaling and automated quality management that translates pixel-based images into resolution-independent images.
Iterated Systems still has a significant Georgia Tech connection among its employees. Two of Barnsley and Sloan's Georgia Tech collaborators, Drs. Stephen Demko and John Elton, are both now employed by Iterated Systems. Other Tech researchers involved in the project were Jeffrey Geronimo and Douglas Hardin.
For more information, contact Dr. Michael Barnsley at barnsley@aol.com.Last updated: October 25, 1999
Contents    Research Horizons    GT Research News    GTRI    Georgia Tech
Send questions and comments regarding these pages to Webmaster@gtri.gatech.edu