Abstract
As the computers are ubiquitous, the linguistic communication by the digital media (e.g. e-mail) has become very popular. However, the simple text doesn't express our emotion, so it is importanto establish the emotional font generation support system that can estimate the user's emotion from sentences and generate the suitable fonts dynamically. Firstly we pointed out that the font shape, the emotion andKanseihad the close relation and influenced each other. Secondly we have deduced the basic elements of the font shape, the emotion andKanseiby categorizing the large number of words that appear in the articles of the font design. As the result we had 32 elements for the emotion, 34 forKanseiKansei and 56 for the font shape. Finally we conducted the experiment to know the relation among them. The result showed that the emotion andKanseigave the influence on the font shape independently in some cases and dependently in the other cases. We have successfully deduced several rules to generate the font shape form the specified emotion andKansei.
