Identification Key Technology Research Based On Uighur Handwriting Characteristics

Handwriting identification technique is a technique which used to identify thewriter of a given handwriting document based on the writing style.With thecontinuous expansion of the applications, handwriting identification has become avery active research topic in the fields of pattern recognition. Handwritingidentification covers all typical problems in image processing and patternrecognition.Uighur writer identification belongs to the minority language writeridentification research areas, but this kind of research is not mature.Therefore,Uighurhandwriting identification research has great political significance and practicalsignificance.This thesis is to study the offline and text-independent Uighur handwritingidentification.Uighur language uses Arabic letters, its characters are complex andligatures frequently.Based on the characteristics of the Uighur language,this thesisuses improved microstructure features extraction method that based on the probabilitydistribution function and distance measure similarity matching method to makeUighur handwriting identification research.The research result will provide asuccessful experience for other minorities handwriting identification research.The main research contents are as follows:1.Pre-proeessing is the first step of handwriting identifieation.This paper collects120Uygur students handwritings, makes some pre-processing operations:scaningthese handwritings image input, denoising, making gray scale and binaryzation.Thelast step of pre-processing is the handwriting edge extraction.2.The stability and effieieney of feature have crucial influence on theidentifieation system.Based on the characteristics of the Uighur language,this paperuses improved microstructure features extraction method that based on the probabilitydistribution function.Extract grid-window microstructure feature from the edge of thehandwriting image.With the improved microstructure features through the global statistics to describe the frequency of different detailed structure handwriting, thusdemonstrating the writing styles of writers.3.In the aspect of matching handwriting design,this paper uses the distancebetween the characteristic vector to measure the difference of the object,converts thedifferences between the handwritings into comparable numerical expression, formsthe effective identification result through the contrast experiment of normal distancemeasurement method and the weighted distance measurement method.A large number of experimental tests show that the improved microstructurefeatures extraction method that based on the probability distribution function Uighurhandwriting identification research has achieved a good effect. The writeridentification accuracy rate can reach100%. Compared with the existing Uighurhandwriting identification results,this research has greatly improved the accuracy rateof identification and has great application value and practical significance.