A Study on Improvement of Japanese Cursive Style Character Recognition by using SOM Template with Clustering of Dictionary Samples

Abstract

It is one of the fundamental works to read the Japanese historical documents in Japanese historical studies.Most of these documents are written with "kuzushi-ji" characters and it has been studied well to apply character recognition techniques to "kuzushi-ji" character recognition.Above all, "SOM template" which is used in a recognition of ornament letters can reduce a decline of the recognition accuracy for blurred and also fractured characters, but total accuracy is lower than conventional methods.In this research, we apply clustering to dictionary samples before making each SOM template to improve total recognition accuracy.