kelvin kuo
Dozens of framed movie posters line a hallway of the Technology Square Research Building. Each one is autographed, not by actors but by the Tech students and faculty who worked on the films. The most recent poster in this informal hall of fame is for Disney’s The Princess and the Frog signed by College of Computing professor Jarek Rossignac, now in his 14th year at Tech, and alumnus Brian Whited. With funding from Disney, the two developed software to make the animators’ jobs easier. Whited, CS 03, MS CS 05, PhD CS 09, now works for Disney. But the bright lights of Hollywood have not lured Rossignac from Tech, where he teaches courses in computer graphics and animation and in an office adjoining his MAGIC lab (that is the Modeling, Animation, Graphic, Interaction & Compression Lab) ponders the behaviors and perplexities of various shapes.
The Disney artist’s dilemma: The lead artist produces drawings. That’s where the character’s appearance is born. But they don’t produce all the drawings that you see in the movie. They produce only a few sparse drawings. Since there are 24 frames per second in the movies, somebody has to generate the missing ones. There’s a process of producing what is called in-between frames, the missing frames that need to be there for a smooth transition from one to the next. This is a tedious job.
The solution: The artist draws complicated curves, not just points or little edges but nice curves that 
might represent the shape of the human face, or a mouse that moves, or a hand or some trees that sway. And now what we need to do is understand what portion of this drawing goes to which portion of the other drawing, so that’s called correspondence. And correspondence is easier for humans to understand. … This is a challenge for software to try to put together automatically. Let’s suppose that we only were given this initial drawing and this final drawing. What we do is produce automatically all these frames. And the idea is that if these frames are OK with the artists, then they don’t have to do the job.
Engineering beauty: If the lead artist puts a point here and says 24 frames later, a second later, it has to be there, then a good Georgia Tech engineer would say, ‘Oh, I understand. I have two constraints. I have a constant velocity, zero acceleration — linear path.’ We solve that very nicely, except the artist will not like it at all. They prefer curved, interesting paths. They like beautiful motions. Working with a colleague of mine in Barcelona, Spain, we have developed what we call the ‘equation of beautiful motion.’ Of course, it’s very pretentious to say we have the equation of beauty. But from a mathematical perspective, it’s a very nice formulation of what we call the most beautiful motion. What we’re using with Disney in part is based on this principle of a very steady motion, which has a nice arcing trajectory and no surprises.
On working with artists: It was very interesting because they come from a different perspective than we do. We want everything to be algorithmically correct, and they have all this understanding of the different subjective matters that come to play. … Sometimes what we propose is happily accepted, and sometimes they say, “No, no way. We’re never going to do that.”
Artists he likes: I like Picasso, Matisse. I like the way they play with shapes to simplify them and still convey something.
On coming to Tech: I came to academia from IBM research, where I worked for 11 years, and I love being a professor.
Hometown: I was born in Poland, in Warsaw, and my parents and I moved to Paris when I was a kid. So I grew up in Paris, between the ages of 10 and 18 or so. And then I studied in France, engineering, and came to the U.S. and did my PhD in EE at the University of Rochester and then stayed here.
On his Klein bottle: It’s a mathematical, bizarre contraption that makes our lives more complicated because it goes through itself but it doesn’t have an inside or an outside. Typically when you do 3-D graphics, the surfaces are oriented. This guy doesn’t have an inside, so not all surfaces are orientable. When [students] think all of them are orientable, I show them the Klein bottle and we have to discuss what to do about it. It’s nice to actually have tangible things to show students when trying to talk them into writing an algorithm for it.
Choosing a career: When I was growing up, I was very bored. So I had to invent things to keep busy. My parents bought me one of those first computers, a TRS. So I learned to program with that, and I always wanted to be an engineer.
On TV: We don’t have a television at home, so I haven’t been watching television for a long time. We decided not to have a television when we had children. … And they didn’t miss it too much. We were concerned that they would be diminished somehow, but it gives them more time to do things. And they grew to be happy, smart and successful.
Recently solved problem: I went to a nice workshop that was organized by McGill University in Barbados, and we spent five days with senior faculty talking about ideas. Everybody was proposing a set of problems or projects, so I proposed a few. And we worked on one of them. The problem was when I give you set a numbers, you can compute the average. But what if I give you a set of shapes? What’s the average of a set of shapes? And this is important, for example, in medical studies. … We actually have a nice formulation of how to compute an average.
What he could do without: I can live without a cell phone. … I try to limit the bandwidth between my brain and the rest of the world. I do e-mail twice a day, but I don’t check it all the time. I’m using Skype because I have collaborations with people in different countries. It’s a convenient way of working together. But I do try to be careful and carve some time for me to think and to work on research or talk with students.
Twitter