Introduction

With 400 MHz processors on 100 MHz system busses and 128M bytes of RAM, today's personal computers have bandwidth to spare after the user's primary requirements have been fulfilled. This leaves considerable performance available for improved color imaging; in this conference we see some good examples.

After surpassing the mega-pixel resolution hurdle, digital cameras are now a viable tool both for professionals and serious amateurs. While there is still room for innovative research in image reconstruction, a new technology on the horizon is spectral imaging, as is illustrated in the session on Spectral Imaging and Cameras.

This increased computing power allows us to run extensive and sophisticated printer models. As the papers in the session on Modeling for Hardcopy witness, it is now possible to either deploy detailed physical models or apply neural network methods in the printer guerrilla.

Yet, mapping the various device gamuts one into an other remains a hard problem. The blue hue line in the CIELAB color space is still a big practical hurdle, but the concluding paper in the session on Gamut Mapping Algorithms puts powerful Internet tools in the hands of color scientists to help them building an intuition for gamuts. The more general Color Mapping Algorithms problem is be the focus of a separate session starting the next day.

As McLuhan noted a generation ago, instead of saving work, labor-saving devices permit everybody to do their own work. Nowhere is this more evident as in the pre-press industry, a large number of small businesses which are disappearing faster than the horseshoe blacksmithes upon the advent of the automobile. The session on Workflow will shed light on this phenomenon. The Panel Discussion promises to be an animated discussion of the technological roadblocks still remaining to unleash the electronic publishing revolution.

Instrumentation for colorimetry is becoming more and more widespread as cost and convenience have improved dramatically over the past decade. The two sessions on Colorimetry are devoted on the instruments and the interpretation of the obtained results.

Although personal computers now ship with a 9G byte harddisk, moving large color images is still cumbersome. Also, there are still many personal computers that have only 8-bit color. These problems are addresses in the session on Color Quantization and Compression.

The medium where file size is most important is the Internet. We continue the traditional series of Internet-related sessions with a session on Internet Imaging. This series started in 1997 with Internet Printing and continued in 1998 with Electronic Publishing. While many papers thoughout the conference mention Internet applications, the papers in this session are more specific.

In the end, how well the viewer will like an image produced with a good color reproduction system will depend on its Tone Reproduction and Image Quality. This is a very difficult field, because an intuitive aesthetic judgement has to be made explicit by expressing it in a semantically relevant language. Only after this knowledge transformation has been carried out, it is possible to design good algorithms.

The conference concludes with two sessions on Halftoning. The first communicates the latest developments in error diffusion algorithms, such as vector error diffusion and the treatment on boundary artifacts. The second session comprises other methods, like stochastic screening and blue noise dithering, as well as the inverse problems of descreening and inverse dithering.

Giordano Beretta
Reiner Eschbach

Conference Program