Now, it is most generally to exchange the imformation of characters with their codes. But there is a problem in that, the characters both of foreign language and national language can't be displayed by the terminal if it doesn't have their pattern in its ROM, that is to say, there are no way to display the unknown characters. And it is irrational that many patterns of characters are used in China, Japan and other countries common, but their codes are different in these countries. By the way, if the common components of characters can be segmented and the imformation of their structure can be coded, it will be possible to describe the fonts of the unknown characters. From this viewpoint, we suggest a structural character font database system which consideres the strokes to be the components of characters. With this system, we will be able to communicate with the codebook which will be created from the database. And we propose an algorithm of the segmentation of characters due to the frequency of the appearance of strokes, this makes the segmentation of characters easier. Finaly, we propose the structural codes from the peculiar strokes of character, in order to make the fast reference be possible.
抄録全体を表示