類神經網路辨識手寫中文相似字之研究

Abstract

本論文的目的是應用類神經網路於手寫中文相似字之辨識。相似字辨識一 直是文字辨識的核心問題，其困難的原因是由於相似字之間字形與字形相 同的部份很多，而且手寫字的變異度較大，以致難以正確分類。本研究利 用類神經網路來建構相似字分類器，以完成相似字的辨識。本相似字分類 器包括三個主要的部份，分別是(1)相似字集之建立，(2)特徵擷取， (3)以模組化類神經網路完成辨識。根據所提方法實驗之結果，藉由此相 似字分類器的加入，系統的整體正確率由 86.01% 提昇至 90.12% ，而在 駁回率為 6.7%時，其辨識率可達 94.11%。本論文所能辨識的字為教育部 所選定的 5401 個常用字，並以工研院電通所研發之手寫中文字資料庫為 訓練及測試樣本進行實驗，因此所建立之系統具有廣泛性及一般性。 This thesis presents an application of neural networks on off- line similar handwritten Chinese character recognition. The proposed method consists of three components:(1)confusing character sets construction,(2)feature selection,(3)modular neural network recognition. In order to evaluate the proposed recognition system, we choose 5401 frequently used Chinese characters as our trainning and testing domain. The database of each testing and trainning sample character was created by the Computer and Communication Laboratory of Industrial Technology Research Institute. Because the samples in this database were collected by more than 2600 people, our recognition system could reach a high generality and user-independence. Experimental results show that, the method improves recognition rate from 86.01% to 90.12%.