inside/out - Fabulous, Fantastic Fractals -- Finally
Images have become so much a part of our digital world that we take them for granted. We give scant thought to the challenge of developing digital images, transmitting them around the digital universe, and reproducing them on a myriad of devices. Yet there are a host of problems associated with imaging that are only recently being addressed. As they are solved, the applications of imaging will permeate more and more of our information technology world, leading to new ways to store and communicate information.
Images are important because the old adage, "a picture is worth a thousand words," is true. Pictures help us communicate facts, opinions, and abstract ideas. But, the digital world throws up a lot of bottlenecks slowing our ability to easily communicate with images -- the most apparent being narrow bandwidths which delay the transfer of images because they are such bit hogs.
Enter fractals, a branch of mathematics that can define structures which posses thematic similarities. That's how John Bacon, president of Iterated Systems, Inc., (Atlanta, Ga., www.iterated.com) describes the technology behind his eleven-year old company. Structures that have similar-looking forms of many different sizes can be expressed mathematically with fractal geometry even though they do not correspond to simple geometric shapes. Examples include clouds, mountains, turbulence, coastlines, roots, branches of trees and blood vessels. By plotting fractal equations, real-world images can be generated.
That is only the beginning, however. It’s the features that make fractals fascinating. The patented technology, a Fractal Transform, developed by Iterated's founders, Michael Barnsley and Alan Sloan, Ph.D.'s from Georgia Institute of Technology, yields the imaging benefits. Using the company's proprietary Imaging Systems Architecture (ISA) announced six months ago, applications built on fractal-based digital imaging gain beneficial features.
First is resolution-on-demand. This is somewhat like Adobe's PostScript page description language that allows page output scaling without loss of quality and creates device-independent printer output files. Fractal encoding eliminates or reduces the need to redigitize data at multiple resolutions. One encoded image can be rendered to many different sizes and resolutions. The same file can display an image of equal quality on a watch or a wall.
Second is bandwidth scalability. This makes it possible to transmit or deliver data with quality, despite bandwidth limitations. It simplifies data transmission over networks, making it easier to move images through the workflow. It also allows transmission over varying types of networks and storage in a wider range of media.
Third is what Iterated calls variable loss encoding. This meets the end-user's demand for high-quality viewing of images. Original images can be reconstructed after decompression while achieving the goal of maximum quality.
In the future, if fractal encoding becomes a standard in image processing, the number and types of digital imaging devices will increase. There will be more flexibility in editing, viewing, printing, encoding and re-encoding images in any order, with ease. Today the process of creating, editing, transmission and viewing is linear. With fractals it can be non-linear. For example, it will not be necessary to rescan an original to gain more resolution so the image can be enlarged.
Fractal images by their very nature make it easier to re-purpose the image. This becomes critical as convergence (the merging of TV, computers and video) takes place. Today, once content is digitized, it is locked into one resolution, one file size and often one purpose. All these features of fractals combine to give better quality. Digital imaging, like digital audio preceding it, brings with it the expectation of better quality. But current technologies used in image processing workflows tend to degrade the image at each step. Fractal Transforms make this annoying problem a thing of the past.
Resolution independence, bandwidth scalability and lossless encoding mean a truly multi-purpose image can be created. Iterated's software tools are now being used by OEMs to develop applications. Today the applications are mostly confined to vertical industry uses particularly in the graphics and pre-press environments. But as this form of digital imaging becomes more common, it will be integrated with many other business-related applications.
It seems only a short time ago that color monitors were thought to be of justifiable use only for design. Now it's hard to find a desktop system with a monochrome monitor. How long will it take for fractals to find their way into software packages for order entry or payroll applications? Stay tuned -- it won't take long.
--After 18 years in marketing and sales at IBM, Bob Diefenbacher founded Denbrook Systems Associates, an IT consulting firm based in Malvern, Pa. firstname.lastname@example.org.