Character post processing method and device based on regular expression

Abstract

The invention discloses a character post processing method and a character post processing device based on a regular expression and belongs to the field of character identification. The method and the device are designed by aiming at the defects of poor reusability and expandability and the like of the existing character post processing method. The method provided by the invention comprises the following steps that: a post processing criterion expression of a current identification region is set according to the post processing criterion grammar; the post processing criterion expression is analyzed to obtain an arborescence data structure; identification results are matched; and character post processing results with the highest match value are obtained. According to the method, the grammar elements of the regular expression are utilized for describing the post processing criterion of the character identification results with different post processing requirements, and good universality, expandability and expression capability are realized, so the setting on the post processing criterion is flexible, convenient and fast.

Description

technical field [0001] The invention belongs to the field of character recognition, in particular to a method and device for post-processing characters based on regular expressions. Background technique [0002] Text post-processing refers to the process of screening out the recognition result character string that best meets the criteria requirements according to the preset post-processing criteria after obtaining the recognition result candidates. In previous post-processing methods, the setting of post-processing criteria and the corresponding criteria matching methods are mostly designed according to actual needs. For example, the essence of text post-processing methods based on Optical Character Recognition (OCR) is to convert text images into text recognition results. If the text content has a certain semantic meaning, post-processing criteria can be used to correct it. Recognition result: If the text is an ID number, the post-processing can correct the recognition re...

AI Technical Summary

Problems solved by technology

If it is applied to text post-processing, when there is no exact match among the candidate characters, the recognition result cannot be obtained

[0004] In addition, none of the current text post-processing methods can solve the problem of missing first and last characters in text recognition results caused by wrong text segmentation.

Images