Study On Information Granules And Its Application In Recognition Of Off-line Handwritten Characters

Information Granulesand or Granular Computing is a new subject that developed fastly in recent years and has its special meanings in many aspects.This paper analyses and researches the basic theory and technology of Information Granulesand and presents an information Granulesand system to recognize off-line handwriting digits. we carefully discuss the abstraction of the characteristic information Granulesand, the processing of the characteristic information Granulesand and the construction of classifying device based on information Granulesand,etc.Firstly, this paper provides a neighborhood expanding scheme of the information Granulesand and a scheme of information reduction. Secondly it constructs G_MAXNET network classifying device model based on the technology of information Granulesand. Finally a matching algorithm by using information Granulesand template for decision rules is also proposed. The effectiveness of the approach has been demonstrated by experiments.The last part of the article has analysed some weak aspects of the scheme and has looked forward to the development trend in the field of pattern-recognition based on the Information Granulesand and Granular Computing .