訓点資料を対象とした翻刻支援システムの構築

要旨

Recently ancient sutras have been computerized. However the support for written materials with reading marks such as “okototen” is not sufficient with regard to the transcription of documents and the system for dealing with the contents. This paper aims to establish a transcription support environment for written materials with reading marks. Taking into consideration the fact that the target documents were handwritten and hold a variety of descriptive information, we developed the system so that the descriptive information of reading marks can be stored as coordinates of a rectangular area or a point in the database and the data can be mutually related. Furthermore we attempt to ease the burden of input, by developing an image-based interface using HTML5 Canvas for specifing Chinese characters and reading marks on the document image intuitively, and by implementing automatic acquisition features for character position and reading of marks. Evaluation experiments made sure that there was no problem found in the operation of the system, the average input time was around three seconds, which means the practicability of this system, and the mistakes made in the automatic acquisition were so few that we could correct them manually.

主題