http://vvv.com/interface/ogden.html (PC Press Internet CD, 03/1996)
Remember all those discussions with people who didn't understand you? How many times have you said "Why can't they get what I mean through their thick head?" Now you can relax. The Japanese are working on it -- VIRTUAL PEOPLE -- Robots that recognize body and facial language and react accordingly. It is just one of the many fields, unknown to us in the west, that the Japanese are concentrating on as they move into greater understanding of knowledge about knowledge.Think of those times at a party when you met someone not totally fluent in your language, yet you were able to communicate by charades: pointing, gesturing and facial expressions. In teaching and learning, language alone is often not enough, even among people with a common language.
Such researchers as Yasuhito Suenaga, of the NTT Human Interface Laboratories of the Nippon Telegraph and Telephone Corporation (the world's largest company -- more than twice the size of AT&T, General Motors and IBM combined) are trying to fathom how a robot or "virtual person" can understand and react to a real person and anticipate thoughts!
The basic process at the moment includes merging present computer vision (my own system can recognize 16.8 million different colors and hues) and graphic technologies to create a better human interface through the analysis and synthesis of human images.
This they hope to accomplish by the recognition of head motion, lip motion and facial expressions with graphics research producing 3-D face models, facial expression and realistic hair images.
What will be the end result? No one knows just yet, but we are all aware of the many sensors we keep running into in the course of our daily travels and chores. Elevators, hotel, supermarket, airport and office doors which open as we approach. Back-up sensors on trucks and cars that continually beep while the vehicle is in reverse gear. We unconsciously rely upon such sensors and very quickly they become part of the urban scene. Most work on ultra-red light or ultrasonic signal detection and costs are dropping daily. In many cases they are not detecting humans just sensing the presence of some object in the observation zone. When sensors can act like people and make real decisions, our environment will become much safer.
Once visual sensors and computers with eyes, can interpret human behavior and intentions from facial expressions; gestures and other body movement; the introduction of human interfaces will be much easier to arrange. The same applies in reverse. If computers, in the form of human-like robots, can express themselves, communications between people and virtual people will become dramatically smoother.
The line drawings created by NTT's Yasuhito Suenaga show how the fields of computer vision in the form of image analysis and computer graphics, in the form of image synthesis, can work togetherto produce a "virtual person". Hands, body, lips, face, even hair all make a statement and can be duplicated ... another step on the pathway to virtual reality.
One of the outcomes of similar research, conducted by NASA in the United States, is the creation of "medical examiners", chairs that take vital measurements while you are sitting. Think of the reassurance, especially for the elderly, when they know they can get a doctor's examination every day just by sitting in a chair for a couple of minutes. Especially a chair that is wired to a computer in a physician's office. The reassurance of good health enhances good health.The chair could remember names and verify them from a dozen sensor identifiers and then speak to them as friends.
The next 10 years will be like no other decade past.
More information:
Yasuhito Suenaga,
NTT Human Interface Laboratories,
Nippon Telegraph and Telephone Corporation,
1-6 Uchisaiwai-cho, 1-chome, Chiyoda-ku, Tokyo 100, Japan.
Phone: 03/509-5095 or 03/509-3028. Fax: 03/503-9990.
Return To InterFace Planet Window