• Blogs
• Forums
• Tradeshows
• Partners
• White Papers
• Podcast Show Notes
• Events

Groundbreaking Research

Jun 1, 2004 12:00 PM, by Audrey Doyle

SIGGRAPH 2004 Papers

When SIGGRAPH premiered in 1973, it was a small, informal, papers-only conference that attracted just R&D types. Then, as the technologies introduced in SIGGRAPH papers evolved from research prototypes to full-fledged products, the conference flourished to also include the numerous exciting and groundbreaking presentations, experiences, and services for which SIGGRAPH is known worldwide.

Yet, despite the glitz and glamour of the tradeshow floor, the fun and entertainment of the Computer Animation Festival, and the myriad additional venues comprising the SIGGRAPH experience, today — 31 years after their debut — the Papers sessions at SIGGRAPH still are the essence of the conference for several thousand attendees.

“The Papers sessions have always played a huge role at SIGGRAPH,” says SIGGRAPH 2004 Papers Chair, Joe Marks. “At a SIGGRAPH that attracts, say, 30,000 attendees, for about one-quarter of those people the Papers sessions are the focus of their experience.” The Papers sessions have enjoyed such staying power because year after year they've provided the forum for highlighting provocative and significant new work in computer graphics and interactive techniques. The papers being presented this year promise to continue that role, with topics ranging from new techniques for creating curves and surfaces, simulating smoke and water, and accomplishing photorealistic lighting, to new technologies for data-driven character animation, video-based rendering, and interacting with images.

But it's not just the topics that consistently make the SIGGRAPH papers so interesting and important for so many attendees. It's also the fact that they represent both incremental and seminal advances that promise to open up new avenues of exploration.

“Both types of advances are equally important,” says Marks. “In some areas you must make incremental progress for several years before you amass enough experience to solve problems well. This year's Papers sessions provide a great combination of incremental and seminal advances.”

In addition to the topics and the types of advances they represent, the papers presented at SIGGRAPH also are known for their extremely high quality. “SIGGRAPH is the only premier forum for computer graphics and interactive techniques,” says Marks. “In other fields, the best papers or the most interesting research are spread out over multiple conferences and journals. But computer graphics is unique in that there's one indisputable ‘best place’ to publish, and that's SIGGRAPH. Therefore, the quality of papers presented is always high, and this year is no exception.”

A Rigorous Review Process

To truly appreciate the fact that the papers presented at SIGGRAPH 2004 represent the best research on computer graphics and interactive techniques it helps to understand how they were chosen.

According to Marks, 478 papers were submitted this year — up 10% from last year. This year, for the first time, participants were allowed to submit their papers, supporting videos, and other material electronically.

Once the papers were submitted, they went through a rigorous review process. Roughly 50 senior reviewers comprised the 2004 Papers Committee, and each had to read 20 papers. For 10 of those papers, they acted as the primary reviewer and had to solicit two more reviews from the outside community. For the remaining 10 papers they acted as the secondary reviewer and had to solicit one additional outside review. According to Marks, only a dozen papers were rejected without review. “We didn't want to waste reviewers' time with papers in which the authors hadn't made an effort,” he says.

By the end of the review process, the Papers Committee had conditionally accepted 83 papers, meaning the authors had to address certain requirements before the paper was accepted. Requirements were as benign as a missing reference or as significant as a rewrite of portions of the paper because they were poorly written or misleading. “The dialogue between authors and senior reviewers resulting from conditional acceptances improved the papers substantially,” Marks says. “Journal-quality reviewing and the publishing speed of a conference: it's a great combination.” (For a list of Papers sessions, see the sidebar.)

In addition to these papers, the committee also chose 12 papers for publication in a regular issue of the ACM Transactions on Graphics, the journal that publishes the SIGGRAPH proceedings. “Offering the publication option was new this year,” Marks says. “Although these papers weren't accepted, they were deemed to be of potential SIGGRAPH quality, but needed more time for revision.”

Growing Challenges

To be sure, it's an honor to have your paper accepted for presentation at SIGGRAPH. It's also an honor to be chosen as Papers Chair, to be selected as a member of the Papers Committee, and to be chosen as a reviewer. “For me, being this year's Papers Chair was a huge deal; like what refereeing the World Cup would be to a soccer fan,” Marks enthuses. “I've been a papers chair for four other conferences, but SIGGRAPH was the biggest and most demanding.”

Naturally, the size of the job brought several challenges, some of which could prove to be increasingly difficult to overcome in the future.

One such challenge was reviewer fatigue. “Although submissions have been up 20%, 20%, and 10%, respectively, over the past three years, the number of reviewers hasn't increased at the same rate,” Marks says. “Last year's Papers Chair, Jessica Hodgins, expanded the committee by bringing in lots of new people, many of whom I inherited. Plus I added some more new people, which helped with the fatigue problem. It also addressed the false concern that SIGGRAPH, particularly the Papers portion, is a closed club. It's not; the median number of times a person has served previously on the committee was just one.”

Another challenge concerned the budget. “The SIGGRAPH conference used to make a lot of money, but lately attendance has been down and fewer vendors are exhibiting. It's like this for all ACM conferences,” Marks says.

“The papers review process is the best in computer science. The fact that SIGGRAPH is one-stop shopping for the foremost ideas in the field is an important thing to treasure,” he continues. “Retaining this aspect of SIGGRAPH on a smaller budget was tricky.” To address the issue, the Papers Committee covered half of the roughly $100,000 review process tab (which included admin support, plus travel expenses so reviewers could meet to discuss the papers) through contributions.

“People were very generous: some companies paid expenses for their employees who were reviewers; the University of Washington donated the meeting room and A/V support; and Microsoft paid for a fabulous dinner at the end of the process. Plus, thanks to electronic submission, our FedEx budget dropped from $15,000 to $4500.

A third challenge concerned reviewing the growing number of submissions featuring an experiential aspect. “Text and a video aren't sufficient for judging a paper on a high dynamic range display, or a 3D display that doesn't require glasses, or an interactive technique. To judge these properly, you need to experience them,” Marks says. Because the 2004 Papers Committee is well connected, it was able to enlist many reviewers who had experienced the technology themselves, or had worked on similar systems. “But as we continue to see more submissions like these, this will be an issue SIGGRAPH will have to address,” Marks predicts.

Potential Hits

Despite the challenges they faced, the 2004 Papers Committee was able to select a host of papers that represent the latest and most significant findings in computer graphics and interactive techniques.

A few of Marks's favorites focus on capture and display devices: particularly the sessions “Flash & Color,” “Lightfield Acquisition & Display,” “Capture from Images,” and “HDR & Perception.” In fact, he predicts capture and display devices will be a significant theme in R&D, and then in commercial applications over the coming decade. “In some sense, this is the coming-out party for these devices,” he says. “The work has been going on for a while, but this is the first time such a critical mass is being presented at SIGGRAPH.” Also on Marks's list of favorites are the sessions focusing on interactive techniques: “Interacting with Images,” “Identifying & Sketching the Future,” and “Interactive Modeling.”

But regardless of his favorites, Marks says all the papers being presented this year prove the field of computer graphics and interactive techniques is alive and well. “Some people think computer graphics is a solved problem — the argument being if you give enough money to Pixar, ILM, or PDI they can create any effect you want.

“But it's not a solved problem; computer graphics is not dead,” he concludes. “The next decade will be very exciting. And you can see some of the seeds of that excitement at this year's Papers sessions.”

Monday, August 9

Session: “Graphics is Fun”
8:30-10:15 a.m.

Session: “Curves and Surfaces”
8:30-10:15 a.m.

Session: “Interacting with Images”
10:30 a.m.-12:15 p.m.

Session: “3D Texture”
10:30 a.m.-12:15 p.m.

Session: “Photo & Video Texture”
3:45-5:30 p.m.

Tuesday, August 10

Session: “Dynamics & Modeling”
8:30-10:15 a.m.

Session: “Identifying & Sketching the Future”
10:30 a.m.-12:15 p.m.

Session: “Smoke, Water, & Goop”
10:30 a.m.-12:15 p.m.

Session: “Lighting & Sampling”
1:45-3:15 p.m.

Session: “Data-Driven Character Animation”
3:45-5:30 p.m.

Wednesday, August 11

Session: “Shape & Motion”
8:30-10:15 a.m.

Session: “Video-Based Rendering”
10:30 a.m.-12:15 p.m.

Session: “Shape Analysis”
10:30 a.m.-12:15 p.m.

Session: “Interactive Modeling”
1:45-3:15 p.m.

Session: “Flash & Color”
3:45-5:30 p.m.

Thursday, August 12

Session: “Capture from Images”
8:30-10:15 a.m.

Session: “Reprise of UIST & VRST”
8:30-10:15 a.m.

Session: “HDR & Perception”
10:30 a.m.-12:15 p.m.

Session: “Large Meshes & GPU Programming”
10:30 a.m.-12:15 p.m.

Session: “Lightfield Acquisition & Display”
1:45-3:15 p.m.

Session: “Mesh Parameterization”
1:45-3:15 p.m.

Session: “Fixing Models”
3:45-5:30 p.m.

SIGGRAPH 2004 Papers Advisory Board

CHAIR: Joe Marks
Mitsubishi Electric Research Laboratories (Cambridge, MA)

Michael Cohen
Microsoft Research (Redmond, Washington)

Markus Gross
Swiss Federal Institute of Technology (Zürich, Switzerland)

Hans-Peter Seidel
Max-Planck-Institut für Informatik (Saarbrücken, Germany)

Continue the discussion on “Crosstalk” the Millimeter Forum.
© 2013 NewBay Media, LLC.