MEMBERS

About Project Members Project Members

The University of Electro-Communications
Professor
Masahide Kaneko
Masahide Kaneko

The face and Nigaoe

We can obtain various information from the face. The information contained in the face can be broadly divided into two kinds. First, there is information to differentiate an individual from other people. This information corresponds to who someone is, gender, race, age group and age, and so on. The other type of information is personal internal information that is expressed on the face such as feelings, physical condition, intent, pleasure or discomfort, and so on. The basic structure of the face is the same for everyone. From top to bottom, it consists of the forehead, eyebrows, eyes, nose, and mouth. However, the shapes of these parts and their arrangement are particular for each person, and even in the same person the the face is changing every moment. Communication with others becomes smooth and productive due to this information. The Nigaoe is a straightforward expression of the face as a “picture”. The Nigaoe contains three characteristics that make its expressiveness unique and rich. These characteristics are: the distinct characteristics of the face of the subject whose Nigaoe is being depicted; the characteristics of the artist drawing the Nigaoe (drawing style, interpretation of the face of the subject, etc.); and the characteristics of the materials by which the Nigaoe is drawn (various tools such as differences in paper and paint). Moreover, rather than simply drawing a single portrait, this computer technology will make it possible to add a fourth characteristic by developing facial application technology to change the expression or impression, add movement, implement as an anthropomorphic agent, or conduct a search of specific facial images. Through FaceMaker, we hope to broaden the scope of applications for the Nigaoe.

Background to research into Nigaoe

I became involved with digital image processing technology when I was a graduate student, and thereafter have been pursuing image-related research and development such as image processing, image coding, and computer graphics for around 40 years. I became involved with facial images in the 1980s when I began research into intelligent coding of facial image sequences. I started research in Nigaoe from 1998. Initially, I began with the simple thought that it would be fun if a computer could be used to generate a Nigaoe. As my research progressed, I discovered that generating a Nigaoe was an extremely in-depth process involving facial recognition by humans and a lot of technologies, such as extracting features from a facial image, the analysis, description, and recognition of extracted features, exaggeration of a person’s distinct features, its drawing, the interpretation and manipulation of expressions and impressions, retrieval of face from database, and application to agents. In fact, it would not be an overstatement to say that “research into Nigaoe” is “research into the face itself.”

uec.ac.jp