采用混合神经网络高精确度提取机票字符

High Accurate Character Strings Extraction from Coupon Images Us ing Hybrid Neural Network

文章针对机票复杂背景提出了一个进行字符分离的高准确率新算法。该方法采用一个基于主成分分析(Prin-cipalComponentsAnalysis,PCA)和学习向量量化(LearningVectorQuantization,LVQ)混合神经网络作为高效的字符提取器,实际应用证明该字符提取算法准确率高,为准确的字符定位和OCR提供了良好的输入。

A new algorithm is presented in this paper to extract character strings accurately from complex airline coupon images.A hybrid neural network combined with LVQ(Learning Vector Quantization)and PCA(Principal Compo-nents Analysis)is applied as the effective character string extractor to separate coupon strings and background,which satisfies the requirement of accuracy and provides excellent input of OCR.