フォント

文字認識

個人性

ベジェ曲線

Hand-Written

Font

Character Recognition

Individuality

Bezier Curve

[Abstract] Today, the chances we see hand-written characters are decreasing despite the constant demand for usage. Thus, attempts are being made to invent hand-written-like font. However, making Japanese fonts has much difficulty because of its amount of characters. In order to solve the problem, math-based generation methods are being proposed. The aim of our study is to propose input-based font generation system, which completely reflects users' individualities. Although our method needs input by users, UI that conducts character recognition enables fast mapping to specific character codes and less trouble to users. When vector image is converted from raster image, in general, it needs to traced outlines of image and to be expressed by Bezier Curve. However, our method generates vector images directly from touch panel coordinates. As a result of subjective experiment, we confirmed that our fonts were reflecting users' individuality.