Research On Off-line Handwritten Chinese Character Recognition Algorithm Based On Chinese Character Recognition Instrument

In recent years,pattern recognition technology has been widely used in production and life,so as a very challenging subject to study scholars pay attention to and research,in which offline handwritten Chinese characters is one of the difficulties of pattern recognition technology.In order to improve the recognition rate of Chinese character recognizer,this paper studies the algorithm of offline handwritten Chinese character recognition based on the research results of offline handwritten Chinese character recognition algorithm at home and abroad.The main contents of this paper are as follows:(1)By analyzing the characteristics of different databases,considering the diversity and representativeness of the samples.It is decided to use the HCL2000 handwritten Chinese character database of Beijing University of Posts and Telecommunications to carry out the feature extraction.The database adopts the way of writing and storing the files in the form of writers,which provides the data base for offline handwritten Chinese character recognition.(2)In order to solve the problem of irregularity in Chinese characters,we need to preprocess the Chinese character image.The processing of handwritten Chinese characters is processed by gray scale,normalization and refinement,so as to solve the problem of non-standard adhesion and writing.It is ready for further feature extraction of handwritten Chinese characters.(3)feature extraction of the image.This paper discusses the methods of Chinese character feature extraction based on traditional statistical decision method and structural analysis method,compares their advantages and disadvantages,and determines the characteristic fusion method which combines various features.We choose a variety of characteristic methods such as high frequency and good use Extraction of handwritten Chinese characters on the image,the left and right extreme points,as well as horizontal,vertical,write,Na and other eigenvalues.(4)handwritten Chinese characters image classification.Traditional statistical classification is time-consuming.BP neural network is a multilayer feedforward network trained by error back propagation algorithm,which is one of the widely used neural network models.Therefore,the BP neural network is used to train the samples to ensure the recognition of Chinese characters more accurately and quickly.