Csuri - Interviews csuri work

Over the years you've been heavily involved in research. What's the most important project you've got going right now?

I started the Ohio Supercomputer Graphics Project, but I decided to divest myself of much of my research responsibilities. I've got bright young people around me, and I'm letting them run with the ball, as they should; they've got to make their mark. It is really their project.

Right now I want to play the artist and see what I can do. I'm getting personal satisfaction out of doing my own creative work, and I want to test software as a user. I'm fortunate to be in a situation where I have people writing custom software for scientific visualization, but I can apply it to art.

If it doesn't work, I can ask them to fix it. By putting on my user hat again and allowing my mind to play with notions about reality, and perhaps developing a sense of the ridiculous as I look at reality, I am able to ask questions that in a curious way may drive the technology.

Can you give us some examples?

I was working with a radiosity algorithm developed by Steve Spencer, and I loved the quality of light and the illumination model, but I was frustrated by the computational cost. Then Scott Dyer, who wrote our scan-line algorithm, got together with Steve to see if one could have the best of both worlds-illumination from radiosity and speed from scan line.

Now I can set up a lighting model with simplified oblects at low resolution, and this information is passed on to the scan-line algorithm for final calculation at high resolution. I also discovered that radiosity does not handle bump mapping or mirrors, but that's a difficult problem to solve.

Another time I worked with Jeff Light, who is developing an image processing package and was trying-with edge detection and playing games with colors-to get the effect of brush strokes. He had scanned in a photograph and thought he had made it look like it was created with brush strokes. I said, "That's not like an oil painting. Let me go get some art books: I'll show you what brush strokes look like." I went out and spent about a hundred dollars on art books.

I said, "Let's look at Cezanne and Seurat, let's look at van Gogh, let's look at Turner." We talked about optical color, the mixing of color that makes colors look alive. We talked about how to achieve the appearance of brush strokes. I said I wanted things that had the look of heavy thick paint or impasto where the brush marks trail off.

We now produce images from photographs that absolutely look like they were done by a painter. Jeff did a wonderful job.