Research On The Mongolian Cyrillic Character Recognition Algorithm Of Handwriting

In the era of cooperation between various countries,China and Mongolia,as good and friendly neighbors,making Mongolian Cyrillic an effective computer input and character recognition can help the economic,social and cultural development between the two countries and regions,as well as promote better communication between the two countries and regional people.In order to meet the basic application requirements,from the perspective of recognition rate and recognition speed,combined with the character and language characteristics of Mongolian Cyrillic,this paper designs a handwritten Mongolian Cyrillic character recognition system.The main contents of this paper are as follows.1.Establishing the Mongolian Cyrillic Character Library and corpus.a Mongolian Cyrillic handwriting character sample set with a certain coverage is completed by writing a character input window and inviting 15 participants to build samples.And the corpus of Mongolian Cyrillic is constructed by transcribing Mongolian Cyrillic articles.2.For the input sample set,combined with the character characteristics andlanguage characteristics of the Mongolian Cyrillic,this paper tries to explore the problems in the recognition of the Mongolian Cyrillic handwriting through the classification experiment of the Mongolian Cyrillic handwriting.Based on the analysis of the experimental results,a Mongolian Cyrillic character recognition system is proposed3.The deep convolution neural network is used as the main classifier,and the neural network framework is determined through a number of experimental tests and comparisons.After fine tuning,the deep convolution neural network model which is most suitable for handwriting Mongolian Cyrillic character recognition is determined.Based on the recognition results,the possibility of further improving the recognition rate is proposed.4.The design of auxiliary classifier and integrated with classifier.Combined with the language characteristics of Mongolian Cyrillic characters,the short-term memory neural network is used to predict the next character,and the prediction results are combined with the output of convolution neural network to determine the classification results.In addition,the GUI interface is made and some sample pre-processing functions are added to the GUI interface.