Development of Handwritten Japanese Character Dataset for Auto-Transcription of Modern Japanese Official Documents

摘要

This paper aims at developing a method for creating and correcting a handwritten character dataset for automatic transcription of modern Japanese official documents. The dataset can contain transcription data, images of documents, and handwritten character data. Furthermore, the proposed character database system can verify the data and reduce artificial mistakes. Users can retrieve character images and various other data by employing the proposed system.