Image Processing: A Substitute for Human Eyes and Brains

I work on “image processing” research, which is a field that tries to use cameras and computers to process—or perhaps to surpass—what human beings do using their eyes and brains. Some familiar examples where this research is being used would include: devices for reading automobile license plates, assembly robots in factories, sorting systems for perishable foods, digital cameras with facial recognition, and automobile driving support systems that sense people or obstacles and avoid collisions.

Image processing technology has evolved rapidly and is now able to process changes down to 1/1000th of a second. One area of my research that is making full use of that technology to move forward is RoboCup* (an autonomous robot soccer competition). During the game, overhead cameras (visual) confirm the positions of each robot and artificial intelligence sends out commands regarding formation and game tactics, while the robots use sensors to recognize the ball at their feet, coordinate their movements around the field, and kick the ball. What’s difficult is “How to approximate human movements?” and “How to predict the opposing team’s movements?” In order to make computers do what human beings do normally, we must work on writing it into a program, that is, transposing it into an objective description. Our ultimate goal is to make it possible to write into a program what human beings see and think and what they predict.

* RoboCup
“RoboCup,” which started in 1997, is a soccer competition using autonomous robots. The Aichi Prefectural University team “RoboDragons,” led by the Murakami Research Lab and the Naruse Research Lab, has been the reigning champion in Japan for the past four years (as of 2010). They won second place in the Small Size Robot League at the 2009 world championships.

Applying Media Research to a Variety of Fields

Human beings can infer the whole from partial information, but even in the field of communication, progress is being made on technology that infers the whole from the part. By researching various levels of media, such as image-based media and sound-based media, it will likely become possible in the future to recognize such things as words, intentions, and expressions of emotion from the slightest word or facial movement.

At our university, we have installed a state-of-the-art computer seminar room that is capable of handling various media such as computer graphics and stereoscopic images. Furnished with a continuous large-screen display of twenty-four 46-inch panels and other equipment compatible with three-dimensional images, students can edit images and quickly distribute them to the outside world via the Internet. We also have prepared a debate support system, by which students can engage in arguments while searching for information using PCs arranged in a square, and a digital signage system that automatically and systematically shows research results or technical explanations on displays outside the room.

In our department, we have done many things until now by making the most of image processing technology, including producing teaching materials for the university and conducting joint research with companies. The range of applications for image processing is wide, so there are possibilities for joint research with all kinds of fields, such as medicine, economics, and the fine arts. Moreover, since this is a topic that involves “vision” and “eyes,” which most people possess, I believe that everyone can speak to each other on the same playing field. Our department welcomes your proposals and consultations.