Method and system for detecting and correcting Chinese characters and computing equipment

Abstract

The invention discloses a method and a system for detecting and correcting a Chinese character and computing equipment. The method comprises the following steps: receiving an image containing a Chinese character to be recognized; analyzing the to-be-recognized Chinese characters in the image based on the preset Chinese character space structure by a trained Chinese character error detection and correction model to obtain a structural layout of the to-be-recognized Chinese characters; and performing error detection and error correction on the to-be-recognized Chinese character based on the structural layout of the to-be-recognized Chinese character. According to the method and system for detecting and correcting the Chinese characters and the computing equipment, error detection and correction are carried out on the Chinese characters to be recognized on the basis of the preset Chinese character space structure, wrongly written characters highly similar to correct characters and unseen wrongly written characters can be recognized, good generalization is achieved, the recognition accuracy is high, and the recognition efficiency is high. And wrongly written characters can be corrected.

Description

technical field [0001] The present invention relates to the field of text recognition, and more specifically, to a method, system and computing device for error detection and correction of Chinese characters. Background technique [0002] There are many types of Chinese characters and their internal structures are complex, so learning Chinese characters is a difficult task. In addition, with the popularization of electronic products (mobile phones, tablets, computers, etc.), people rely more and more on keyboard typing, so that the frequency of writing Chinese characters continues to decrease. This phenomenon causes children who are beginning to learn Chinese characters, including many adults, to often write typos in daily writing. [0003] Chinese characters are highly structured entities, which can be composed of strokes, radicals and spatial structures. One or more strokes can form a radical, and one or more radicals and spatial structures listed in a certain order can ...

AI Technical Summary

Problems solved by technology

To use the existing recognition model to solve the problem of error detection of typos, a large amount of labeling data is required. In addition to labeling correct characters, countless typos must be labeled. It is impossible to include all typos in the labeling method, and the cost of labeling is also very high

In addition, the recognition model modeled on whole characters cannot realize the error correction function for typos

Images