Electronic Imaging 2000 Symposium
Preparations
It was a gorgeous November morning in Scottsdale. It was the first CIC in the new resort, and after a refreshing early morning swim in the pool (of course, unfailingly Danny Rich had already put several laps behind him), a healthy breakfast on the restaurant terrace, and an intellectually stimulating keynote talk on the physiological mechanisms of primate color vision by David Hubel, the EI 2000 symposium chairs were ready.
John J. McCann (IS&T) and Giordano B. Beretta (SPIE) took the short walk to the Katchina conference room in a stride. After a warm welcome by Calva Leonard and Marshall Weathersby, the two executives from the respective societies, and after seizing the real size of the meeting room (which walls are mirrors?), the two co-chairs noticed the climate had suddenly changed.
On that Wednesday morning in 1998, the perfect -- albeit artificial -- serenity of the resort had given way to a cloudy atmosphere in the Katchina room, with thick black clouds quickly descending on the unaware chairs. The largest conference in the EI symposium would have its periodical move to Australasia, the second largest conference wanted to move to a different symposium, and the main organizer of the third largest conference was retiring. Last but not least, there was serious concern about the Y2K impact, which -- it was feared -- could discourage many traditional attendees from traveling during January 2000.
This contrasted to the chair's expectation to make the first imaging conference of the new millennium a memorable event. McCann's diplomatic skills were instrumental in elegantly solving the defection problems. A focus on education -- fostering education and professional development, promoting communication and interaction in the imaging community, and recognizing contributions in developing imaging scientists & engineers -- was the chair's strategy to make EI 2000 a memorable event to attend. Strong keynote speakers would help the last Y2K fearsome making up their mind.
Planning
To improve the performance and impact of the short courses, the chairs enlisted the help of the symposium organizing committee. Jan P. Allebach -- who was also the 1998 recipient of IS&T's prestigious Raymond C. Bowman Award -- volunteered as the first member of the newly created Education Program Committee. Richard N. Ellson was instrumental in communicating with the Silicon Valley press.
As illustrated in the chart, EI is different from the other IS&T events. Each circle is proportional to the event's number of technical attendees in 1999. The dark sector in each circle denotes the number of papers presented at the event (note that many papers have co-authors). What the chart tells us, is that EI is mostly a scientific conference, where scientists come to interact with their colleagues like in an uplifting bazaar.
In today's scientific world new breakthroughs have to be detected immediately and must be assimilated in a very short time. In this situation the EI symposium fulfills two roles. The first role is to serve as a synchronization event for a community; everybody hears the latest results and can contribute to that emergent property that is the state-of-the-art in electronic imaging. The second role is to interact with others in the community, to seek clarification at the source, to discover new opportunities for synergism, and to learn about the war stories that cannot find a way in scholarly publications. In summary, scientists attend EI to exchange new ideas.
This contrasts to CIC at the other end of the spectrum in the chart. CIC has the goal of fostering color engineers. Companies send their engineers to conferences to learn about the latest technologies. Consequently, attendees enroll in tutorials because they are a valuable source of information about the status quo. While at CIC the emphasis is on learning, at EI it is on exchange.
One could take the attitude, that since short courses are not sufficiently important to EI participants, the chairs should just forget about them and instead nurture community-building activities. However, the latter are already a well working part of EI, in the form of panel discussions, discussion sessions, informal dinners and discussions in the lobbies. Instead, many EI attendees are gifted if not professional teachers. The duration of short courses at EI is double that at some other IS&T conferences, allowing instructors to go more in depth in their subject matter.
With the help of Allebach, the chairs have improved the EI educational program with a more structured curriculum of short courses for engineers attending the symposium and for Silicon Valley engineers who are seeking CEU (Continuing Education Units) credits. We hope that the program can be further improved for EI 2001.
Results
At the time of this writing the detailed attendance numbers have not yet been tabulated. However, the total number and the pre-registration numbers are very encouraging. The total number of technical attendees has increased 6% from 1472 in 1999 to 1556 in 2000. The number of accepted papers has remained constant at 962 (961 in 1999), despite one less conference (24 instead of the 25 in 1999).
The number of registrations at short courses was up by 30% on January 7, the last day of pre-registration. Final numbers are not yet available.
There is anecdotal evidence from cancellations by session chairs in December, that Y2K fears might have had an impact on the attendance. We will know for sure only after all the data is tabulated and we know how the ratio of local vs. non-local pre-registrants has changed from 1999.
Symposium events
The rich assortment of short courses has allowed participants to quickly get in-depth knowledge from the masters of the art. The conferences have been organized in programs that have grouped similar research areas; the system of synchronized presentations has encouraged attendees to hop from conference to conference, maximizing the likelihood of serendipity. Last but not least, the informal atmosphere has encouraged attendees to interact with the speakers; take advantage of discussion sessions, panel sessions, and receptions to network and build strong communities.
In view of the amount of activity in this bustling bazaar, it is not possible to mention all events as this is customary for other IS&T conference reports. We select a few events and apologize to organizers of the other events for their omission -- it shall not be a value judgement.
Evolution of digital photography
The first plenary speaker has been Carver A. Mead, recently emerited from CalTech. To a mesmerized crowd in a hopelessly overfilled auditorium, Mead has unveiled some of the details of his new digital camera. Unlike current digital cameras, where the components of conventional AgX systems are each "digitized," Mead has designed a new digital studio camera from a clean sheet. For example, the user interface is more akin to a laptop computer with an image editor than the viewer of a conventional camera.
One current technological limitation of digital cameras is the sensor's number of pixels. A higher resolution image can be synthesized by interpolating the color planes over all pixel locations, regardless of color sensitivity, but this introduces aliasing artifacts. Another solution is to use a piezo element to shift the sensor to all four photosites for each pixel, but this is only possible for still images. A third method is to use a separate sensor for each color channel, but this can create optical interface problems. Mead has chosen the third approach and has exposed how he has solved the optical problems.
The camera's performance data shown by Mead is by far superior to that of any other AgX or digital camera. Mead is a master in designing hybrid analog and digital circuits, and we suspect that the largest breakthrough might have been in establishing the locus of the interface between the analog world of photon quanta catch statistics and the digital world of counts. We hope to learn about these new insights at a forthcoming EI conference.
Co-evolution of humans and computers
The plenary talk by A. John Michaelis has also drawn a big crowd that has filled the auditorium. In his philosophical presentation, Michaelis has illustrated how, based on Moore's law, the computing power of microprocessors will surpass the human brain's computing power in the early years of the new millennium.
Humans shape tools, but then the tools shape the humans. Michaelis has suggested that every-so-often we must re-examine our focus in the context of broad technology trends, and January 2000 is a good opportunity to do this. Electronic imaging will not only benefit from raw processing power, but also from new interface technologies, which will make more parameters about the surrounding world available to implementors, as well as bring new output dimensions.
The debate has continued in a panel discussion in the late afternoon, were the focus has been on input devices. The panelists have found a consensus that for text entry the WIMP (windows, icons, mice, and pointers) paradigm appears the most efficient input form when used with keyboard and mouse, at least in the immediate future.
In the discussion, panel member Verplank has pointed out, that computer science has gone through three phases concerning the research goal. The first phase has been anthropomorphic, and AI (artificial intelligence) has been the prime technology, delivering the expert systems that drive today's financial markets and pharmacology. The second phase has been focussed on tools, delivering the GUI (graphical user interface) that makes computers easy to use and has made possible today's productivity gains in business administration. The third phase is focussed on communication, delivering the Internet and the Web, and making possible completely new business models based on B2B (business to business).
If the old adage that a picture is worth thousand words, electronic imaging will play a central role in this communication-centered phase of the evolution of computer science.
Multispectral imaging
One way to exploit the increased computing power and I/O devices available today is to gather and process more information in our images. Although multispectral imaging has been an important topic in Japanese and European conferences, there have been only relatively few papers on this subject at American conferences.
We have been lucky that for EI 2000 we have been able to attract Yoichi Miyake, who last October has organized the International Symposium on Multispectral Imaging in Chiba (Japan). Particularly suggesting in his presentation has been the impact of multispectral imaging on medical diagnosis. Miyake has shown examples of the dependence of diagnostic accuracy from the illuminant and the importance of performing accurate colorimetry (endoscopes are complex instruments in which the components age differently and tristimulus colorimetry can be difficult to normalize).
The currently available more powerful technology also allows the capture of geometric appearance in addition to color appearance. This capability opens important new possibilities in the study, preservation, and restauration of our cultural heritage. This makes possible new teaching paradigms and allows us to strengthen our roots while in our activities we become more global.
At the end of the plenary talk, IS&T president John D. Meyer has conferred to Miyake the 2000 Electronic Imaging Honoree of the Year award for his leadership, personal contributions, and most importantly for his vision in the fields of electronic imaging and color technologies. Miyake has been recognized for his outstanding service to the field, by teaching, by introducing many new students to the field, and by organizing conferences for the continuing education of professionals. Meyer has presented Miyake also with a digital camera kindly sponsored by Hewlett-Packard Company.
Conferences
We cannot mention all conferences, so we just list a few highlights. Conference proceedings can be ordered from http://www.spie.org/bookstore/proceed_index.html (vol. 3954 - 3974).
The conference on Human Vision and Electronic Imaging (3959) has been particularly interactive and the discussion sessions have been very fruitful for the attendees. The annual banquet has been an opportunity to meet old acquaintances and make new friends. Larry Stark's personal reminiscences have been very inspiring. An anecdote particularly to the point for this year's EI has been how, when Stark was occupying the office next to Shannon, he would leave his door open so that some of Shannon's wisdom could transfuse over into his office...
In the Color Imaging conference (3963), the management transition from Beretta/Eschbach to Eschbach/Marcu occurred very smoothly and could maintain the momentum this conference has gained over the years (IS&T members might remember the long effort in focussing this conference to differentiate it from NIP -- to where hardware papers where redirected -- and CIC -- to where papers not related to digital publishing were redirected). The conference has started with sessions on spectral imaging and has concluded as usually late Friday afternoon with strong halftoning sessions that have kept the auditorium packed to the very end.
Attendance to the Internet Imaging conference (3964) has been considerably lower than the organizers have expected. The new chairs have been able to solicit 10 times more papers than their predecessors last year, rendering conference viable. Now that the conference is "on the map," the chairs have to strengthen the program committee so it can produce strong sessions. As a first for EI, this conference has had a high-speed Internet connection available to its speakers.
The Optical Security and Counterfeit Deterrence conference (3973) has been the hottest conference and has had to be moved to a larger auditorium to accommodate all its attendees.
Last but not least, Bhaskaran, Hsing, Tescher, and Stevenson must be commended for having put together a successful interim conference on Image Communications and Processing (3974) while VCIP will be held in Australia later this year.