方向指数系列の文法的処理による文字認識

概要

論文の詳細を見る
Linguistic or syntactic method is one of the approaches to pattern recognition. Various applications of this method have been made to the patterns which can be transformed into strings of features. This paper describes a general methodology to recognize symbol string patterns and its application to the recognition of handwritten numerals. The patterns are transformed into symbol strings by chain encoding. Then similarity measures between input string and standard strings of each class are calculated by means of dynamic programming. The overall error rate is 1.5% for 1,O00 samples of handwritten numerical characters. This result shows that our method is at least one of the possible ways of handwritten character recognition, though it has many problems to be studied further.
一般社団法人情報処理学会の論文
1978-05-15