首页>
外国专利>
METHOD AND DEVICE FOR GENERATING A FUZZY RULE BASE FOR CLASSIFYING LOGICAL STRUCTURE FEATURES OF PRINTED DOCUMENTS
METHOD AND DEVICE FOR GENERATING A FUZZY RULE BASE FOR CLASSIFYING LOGICAL STRUCTURE FEATURES OF PRINTED DOCUMENTS
展开▼
机译:用于生成打印文档逻辑结构特征分类的模糊规则基础的方法和设备
展开▼
页面导航
摘要
著录项
相似文献
摘要
In a first step, character recognition features are provided from a certain printed document. In a second step, a number of physical structure features is determined on the basis of the provided character recognition features. This second step is done for each line of the certain printed document. In a third step, training data including an input-output sample are provided, wherein the input is represented by the number of physical structure features and the output is represented by a manually labelled logical structure feature. This third step is done for each line of the certain printed document. In a fourth step, a distribution for each physical structure feature in the certain printed document is determined. In a fifth step, a fuzzy set having linguistic variables and corresponding membership degrees is provided on the basis of the respective calculated distribution; This fifth step is done for each line and for each physical structure feature. In a sixth step, for each line and for each physical structure feature, selecting the linguistic variable with the maximum membership degree. This sixth step is done for each line and for each physical structure feature. In a seventh step, a fuzzy rule for the fuzzy rule base is generated on the basis of the input-output frame, wherein the respective physical structure feature of the input is represented by the corresponding selected linguistic variable with its membership degree and the output is represented by the manually labelled logical structure feature. This seventh step is done for each line.
展开▼