A modeling-based OCR recognition result judgment method and device

A recognition result and judgment technology, applied in the field of OCR recognition result judgment based on modeling, can solve the problems of limited use range, inability to OCR recognition result error detection, error correction, and complicated operation, so as to improve the recognition accuracy and reduce The effect of manual intervention workload and large number of features

Active Publication Date: 2021-11-09
厦门商集网络科技有限责任公司
View PDF16 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

This technical solution must determine its allowable set in advance. As the recognized text changes, the allowed set must be updated synchronously. The operation is complicated and the scope of use is limited. It cannot simultaneously detect and correct the OCR recognition results of various texts.

Method used

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View more

Image

Smart Image Click on the blue labels to locate them in the text.
Viewing Examples
Smart Image
  • A modeling-based OCR recognition result judgment method and device
  • A modeling-based OCR recognition result judgment method and device

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0030] see figure 1 , a decision method based on modeling, comprising the steps of: establishing a feature extraction model of a character image; inputting each standard character image into the feature extraction model to obtain a standard feature matrix corresponding to each character, and establishing a feature matrix containing the standard features The standard character feature library of the matrix; the text is recognized by the OCR recognition engine to obtain the OCR recognition image; the single character feature matrix of each character in the OCR recognition image is obtained by the feature extraction model; the single character feature matrix of the character is corresponding to the character one by one Compare the standard character matrix to judge the similarity between the two, and judge the correctness of the OCR recognition result.

[0031] The standard characters refer to clear and accurate characters that are selected as standard characters through manual s...

Embodiment 2

[0042] A model-based OCR recognition result judgment device, including a memory and a processor, the memory stores instructions, and the instructions are suitable for being loaded by the processor and performing the following steps: establishing a feature extraction model of a character image; confirming that it is correct Each character image of each character image is input into the feature extraction model to obtain the corresponding standard feature matrix of each character, and a standard character feature library comprising the standard feature matrix is ​​set up; the text is recognized by the OCR recognition engine to obtain an OCR recognition image; through the feature extraction The model obtains the single-character feature matrix of each character in the OCR recognition image; compares the single-character feature matrix of the character with the standard character matrix corresponding to the character one by one, and judges the correctness of the OCR recognition resu...

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to view more

PUM

No PUM Login to view more

Abstract

The invention relates to a modeling-based OCR recognition result judgment method, comprising the following steps: establishing a feature extraction model of a character image; inputting each character image confirmed to be correct into the feature extraction model to obtain a standard feature matrix corresponding to each character, Set up a standard character feature library comprising the standard feature matrix; recognize the text by the OCR recognition engine to obtain an OCR recognition image; obtain the single-character feature matrix of each character in the OCR recognition image by the feature extraction model; The feature matrix is ​​compared with the standard character matrix corresponding to the character to determine the correctness of the OCR recognition result. The present invention has the advantages of establishing a character image feature extraction model and using the model to convert the image into a feature matrix, comparing the feature matrix of each character image in the recognition result image with the corresponding standard feature matrix, thereby judging the correctness of the recognition result and making the recognition As a result, the accuracy rate is higher than 99.95%, and the work efficiency is improved.

Description

technical field [0001] The invention relates to a modeling-based OCR recognition result judgment method, which belongs to the field of model application. Background technique [0002] Existing texts and images are generally entered into computers through automatic recognition. Since the recognition accuracy rate of the OCR recognition algorithm can reach 99%, automatic entry is generally realized through OCR recognition. However, in fields such as finance and taxation, the recognition accuracy rate of 99% still cannot meet its requirements. Therefore, it is necessary to manually check the recognition results twice or even three times to find out the location of the recognition error, which wastes a lot of human resources. , which increases entry costs. [0003] The invention patent with the publication number CN106650715A "A Method for Error Detection and Correction of Character String OCR Recognition Results Based on Allowed Sets" discloses a technical solution: by making...

Claims

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to view more

Application Information

Patent Timeline
no application Login to view more
Patent Type & Authority Patents(China)
IPC IPC(8): G06K9/46G06K9/34
CPCG06V30/153G06V10/462G06V30/10
Inventor 郝占龙庄国金陈文传杜保发林玉玲吴建杭方恒凯
Owner 厦门商集网络科技有限责任公司
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Try Eureka
PatSnap group products

Browse by: Latest US Patents, China's latest patents, Technical Efficacy Thesaurus, Application Domain, Technology Topic.

© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap