|
|
Équipe Analyse/Synthèse |
S. Hashimoto classifies information technology in three categories:
He further classifies these three categories according to their differences in evaluation. As an example: for signal processing, evaluations is by means of measurement; for artificial intelligence, recognition; for kansei technology, appreciation.
He considers that a system to create art with human (sic) has to understand not only the user's intentions, but also the environment of performance with multi-modal sensing ability. The system will try to understand the will of a performer while iterating bi-directional communication.
The author considers the most important emotional information in human gestures as the forces applied by the body (that can be computed directly from acceleration of body movements). He opposes this approach to the most general one, consisting of recognizing shapes and positions (of the body). His group at Waseda University has therefore developed new musical environments driven by gestures and employing acceleration sensors and data-gloves as gesture input devices. They also make use of neural networks (NN) on the processing of gesture information stage, since using NN one does not need knowledge on the Kansei processing itself.
Part of this work can be found in [1] and [2].
Finally, he cites some facts about the current Kansei research in Japan. According to the author, in 92, the Japanese Ministry of Education started a three year grant program in ``Kansei information processing'', which was followed by another special grant program on ``Virtual Reality''. Over 50 research groups from different fields (computer science, physiology and psychology) have joined the program and were divided in four groups: Kansei modelling, kansei information in media, kansei in human behavior and kansei in communication.