A sunrise image generated from an HDR radiance map using Dorian Chan’s algorithm. Original radiance map from Mark Fairchild’s HDR Survey.
Each year, the SIGGRAPH conference hosts the ACM Student Research Competition, a unique opportunity for undergraduate and graduate students to present their original research. This year, Dorian Chan (Carnegie Mellon University) received the first-place prize in the graduate category for the Poster “Bound-constrained Optimized Dynamic Range Compression.” We connected with Dorian, who shares how he developed his research and what advice he has for students interested in submitting to the competition.
SIGGRAPH: What was your reaction to winning first place in the ACM Student Research Competition for SIGGRAPH 2020?
Dorian Chan (DC): I was really happy! I’ve been working on this project for a while, and it was amazing to see all of my hard work pay off. All of the other contestants did amazing work, so it really means a lot for my work to be selected.
SIGGRAPH: Tell us about your winning research. What inspired the new bound-constrained framework for high dynamic range compression that you developed?
DC: In our work, we basically developed a new tone-mapping operator that improves upon existing gradient-domain operators by lifting some of the inherent limitations of these traditional methods. You can reference our Poster for details — it’s available for on-demand viewing through 27 October. This approach came out of a lot of conversations with my project advisor, James F. O’Brien. We reviewed a bunch of previous literature in high dynamic range compression, and we noted a few places where past algorithms could be improved in both speed as well as in the quality of the results. After a bunch of experimentation, we figured out a nice, new framework that addressed those problems.
SIGGRAPH: How do you anticipate your research will be used? What problems does it solve?
DC: I hope our work gets applied to movies and other creative applications. Accurately depicting real-life high dynamic range (HDR) scenes on low dynamic range (LDR) displays and software is a challenging problem that people have worked on for a long time.
SIGGRAPH: Share how you developed your Poster. What steps were involved? What challenges did you face/overcome?
DC: To develop the Poster, I took all of my work, thoughts, and results and compressed them into as small of a space as possible. I think the hardest part was figuring out what to cut and shorten, as there’s so much more I want to say than can fit on a single poster. After that, it was just some fiddling in Illustrator to make everything look nice … and my poster was done. I then made a slide deck for my video, which I reviewed with James. After some practice, I recorded myself giving a presentation, and that was that!
SIGGRAPH: Now that you’ve presented this research at SIGGRAPH 2020, what’s next for your project?
DC: I’m working on a full-length paper version of my poster now, to be submitted hopefully later this year. The Student Research Competition judges gave me a lot of great feedback and advice, which I’m using to improve my work.
SIGGRAPH: Share a bit about your experience attending the virtual SIGGRAPH 2020 conference. How did it go? Any favorite memories or sessions you enjoyed?
DC: Virtual SIGGRAPH was pretty interesting. I’ve never been to an in-person SIGGRAPH, so I can’t speak to the differences very much, but it was a really cool experience this year. A lot of the Technical Papers talks were really interesting to me, and I enjoyed seeing what other people are working on in the field these days. The industry presentations also were nice to see and witness how the research we work on actually gets applied in practice.
SIGGRAPH: How will receiving this SIGGRAPH award help you in your career post-graduation?
DC: I think the recognition will be extremely helpful professionally. A lot of people reached out to me and urged me to keep up the good work, which was really nice to hear especially early on in my graduate career.
SIGGRAPH: What advice would you give to other graduate students thinking of submitting to an ACM SIGGRAPH conference/competition?
DC: You should go ahead and submit, and put your best foot forward. The most important thing is to give it a shot and try your best — I almost didn’t submit my poster to the Student Research Competition, and I’m really happy I did because obviously everything worked out.
Over 250 hours of SIGGRAPH 2020 content is available on-demand through 27 October. Not yet registered? Registration remains open until 19 October — register now.
Dorian Chan is a Ph.D. student at Carnegie Mellon University in the computer science department, where he is advised by Matthew P. O’Toole. He previously collaborated with James F. O’Brien while he was an undergraduate at UC Berkeley. He is broadly interested in problems in computer graphics and imaging.
