Utilization of Hōjo Image Archives as Research Materials

Abstract

  In recent years, an increasing number of Hojo images have been made public in web data archives. This paper examines a method of, as described below, and its practicability in processing properly a group of archived image data using presently available techniques, thereby enhancing the efficiency of comparative research of Hojos, and thus contributing to the development of research into the history of calligraphy.

  First, an identifier is given to each of the works listed in Hojos to create source data. The same identifier is given to images of the same work in different materials in order to achieve source control in detecting all the reproduced images of the same work in different materials in a single action.

  Second, a mechanism is devised to recognize the line breaks in the images, line breaks are given to Hojo images semi-automatically, and an identifier is given to the respective lines. By relating these line identifiers to the above-mentioned work identifiers, it is possible to cut out the image of the target work from the overall image data and display it.

  Third, an arbitrary string of letters is selected and the mechanism detects images of identical parts among different materials using image-processing technology. This makes it possible to reference different versions of the same work listed in different Hojos without adding text information to the images.

  Despite the necessity of further technical improvement regarding the referencing of different versions as described in step three, a series of experiments conducted to establish the above-mentioned method suggested the possibility that image processing technology can be developed as a tool to examine more detailed differences found among different lines of Hojos.