Information

Yoshinori Dobashi

Fun with Computer Graphics

Yoshinori Dobashi , Associate Professor

Graduate School of Information Science and Technology (Media Networks, Department of Electronics and Information Engineering, School of Engineering)

High school : Itsukaichi High School (Hiroshima)

Academic background : Graduate School of Engineering, Hiroshima University

Research areas
Computer Graphics
Research keywords
Rendering, Simulation, Sound
Website
http://ime.ist.hokudai.ac.jp/~doba

What is Computer Graphics Research?

When you hear “computer graphics (CG),” a movie or game might come to your mind first. However, CG is an academic field that is studied in universities. Nevertheless, I think there are probably few people who know what CG research is. CG can create images of a virtual world that do not actually exist, but there are various problems. For example, when creating clouds in a virtual sky, how do you determine the shape, color, motion, etc., of the clouds? CG research is finding solutions to these types of problems. Moreover, CG has recently advanced further and deals with things other than images, such as sound simulation, fabricating real objects using 3D printers and developing devices for experiencing or manipulating the virtual world.

Figure 1  Control of cloud simulation. The left is a typical cumulonimbus cloud. The right shows skull-shaped clouds. The pink curved line at the top left shows the target shape.

 

Our Research Project

Recently, various images can be created using CG. However, it does not mean that anyone can easily create high-quality images that we often see in movies or computer games. We are concentrating our efforts on solving this problem. Let us describe our approach by using two examples for simulating clouds.

Clouds can be simulated by calculating the flow and temperature of the atmosphere, the amounts of water vapor and water droplets. However, it is often difficult for an artist to predict the shape of the simulated clouds. This is similar to the fact that a weather forecast is not always accurate. In order to address this problem, we developed a control method for the cloud simulation to form the desired target shape. Figure 1 shows the experimental result computed by our control method. The left image shows an example of forming cumulonimbus cloud shapes that are typical on a hot summer day. The pink curved line shown at the top left is the target shape given by the user. These clouds can be created easily using our method, but generally creating these shapes by adjusting the simulation parameters is quite difficult. The right image in Figure 1 is an example of forming skull-shape clouds that have two holes in the middle. Although these clouds probably would not occur in the real world, a realistic shape can be created since our method simulates the physical process of the cloud formation.

Next, when synthesizing an image of the simulated clouds as realistic as an actual photograph, we have to compute the color of the clouds based the laws of physics. However, it is again difficult to predict the image that is produced by the calculation. Therefore, we developed an automatic cloud color adjustment method by using a photograph of real clouds. Figure 2 shows an experimental result using our method. The images in the upper row show CG clouds, and their colors are automatically adjusted using photographs of real clouds at different times of a day. Moreover, by combining the skull-shaped clouds with this color adjustment method, we can create interesting images such as the ones shown in the lower row of Figure 2. The color of the clouds are adjusted by using a photograph of Sapporo and the skull-shaped clouds gradually forming over the city of Sapporo.


Figure 2  Automatic cloud color adjustment. The upper row shows automatic color adjustment of CG clouds using real images. The lower right image is a combined example of a real image and the skull-shaped clouds.

 

Other Research Projects

Our research projects cover a wide range of topics including sound simulation and fabricating real objects using 3D printers. I would like to explain them but there is not enough space here. Figure 3 shows only images from our research. If you are interested in CG, visit my website noted at the top or Google my name on the Internet.


Figure 3  From the upper left: Calculating wind noise, simulating fumes, editing shade and shadows, controlling an explosion, simulating a flame, displaying a light path through stained glass windows, designing a special lens using a 3D printer, expressing thunder and simulating the feeling when touching water (canoe simulation).