Method Of Handwritten Numeral Recognition And Realization

OCR (Optical Character Recognition) has become one of the important method in gathering information and information transformation. Handwritten digit recognition has a promising business feature in many fields in society, for example postcode recognition in post office, the statistics of report forms, financial report forms. So the researching on the handwritten digit recognition is one of the important thing. It will be make huge benefit once the handwritten digit recognition became true. This paper studies the recognition of handwritten digits, including thinning processing, chain code tracing, line diction, character extraction and recognition.This paper studies the different character of the different ten digit, and design the different recognition method for every digit. For every character for recognition ,first, we studies it as a two classes problem .Every recognition method give a result to label the class which the character is belonged to. The last answer is given after the study of the ten different result.We have got a good result on the database of MNIST.