The Research And Application Of Handwritten Character Recognition

Handwritten character recognition is one of the research topic field of image processing and pattern recognition.In general ,character recognition system is composed of the image acquisition,signal preprocessing,feature extraction,classification identification. The identification mode of recognition system can be divided into online handwritten character recognition,off-line print character recognition and off-line handwritten character recognition etc,in which off-line handwritten character as a result of the written factors that make character images very arbitrary, such as:stroke thickness,font size,the inclination of handwritten characters,distorted partial strokes,differences in gray font that have a direct impact on the correct identification of characters,Therefore,recognition of handwritten characters is the most challenging issues of pattern recognition area.This paper aims to study the processing methods and algorithms to be improved in certain stages of the process of handwritten character recognition.Brief description of the background and development of the handwritten character recognition method,and describes several representative thinning algorithm,handwritten character recognition algorithms and models,discuss the shortcomings of existing algorithms, through further study of these ways, this paper puts forward a practical repair method:The improved algorithm of enhancing handwritten digit thinning effect and the handwritten character recognition method of based on the outline stroke.At present,the basic thinning algorithm method as follows:Hilditch algorithm,Pavlidis thinning algorithm,Rosenfeld thinning algorithm.The number of hand-written,when the spread of ink or the font is too small, sometimes caused lump that much larger than the width of the normal stroke,when is thinning, that will be refined for a line or a point,This will seriously affect the recognition results.To resolve this problem,by the modification of the original Hilditch thinning algorithm,the experimental results have certain enhancement than the original Hilditch thinning algorithm,Rosenfeld algorithm.NLPCA is a kind of neural network algorithm which is a effective method of handwritten character recognition that can identify letters,numbers,but this algorithm has to go through a lot of iteration and training,this paper presents the character recognition method of based on outline stroke,in the letters recognition rate,although this algorithm is not high, but the digital recognition rate is higher than NLPCA,The experimental results have shown that this algorithm has some feasibility.