Looking for breakthrough ideas for innovation challenges? Try Patsnap Eureka!

Data augmentation method and device for OCR, apparatus and storage medium

A data and data set technology, applied in the field of character recognition, can solve the problems of high cost and poor targeting of training samples, and achieve the effect of improving pertinence and reducing costs

Active Publication Date: 2021-09-24
珠海亿智电子科技有限公司
View PDF5 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0004] The purpose of the present invention is to provide a data augmentation method, device, equipment and storage medium for OCR recognition, aiming at solving the problem of obtaining training samples by manual labeling in the prior art and obtaining training samples by data augmentation. The problem of poor sample targeting

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
  • Data augmentation method and device for OCR, apparatus and storage medium
  • Data augmentation method and device for OCR, apparatus and storage medium
  • Data augmentation method and device for OCR, apparatus and storage medium

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0047] figure 1 It shows the implementation process of the data augmentation method for OCR recognition provided by Embodiment 1 of the present invention. For the convenience of description, only the parts related to the embodiment of the present invention are shown, and the details are as follows:

[0048] In step S101, a recognition dictionary is established.

[0049] In the embodiment of the present invention, the recognition dictionary may include Chinese characters, English letters, and Chinese and English punctuation marks. The recognition dictionary can be established according to the Chinese character code standard of our country. The Chinese character code standard GB2312-80 of my country stipulates 3755 first-class Chinese characters and 3008 second-class Chinese characters, totaling 6763 Chinese characters. Compared with the first-level Chinese characters, the word frequency of the second-level Chinese characters is relatively low, but they also appear in daily lif...

Embodiment 2

[0079] figure 2 It shows the implementation process of the OCR model training method based on the data augmentation method described in the first embodiment provided by the second embodiment of the present invention. For the convenience of explanation, only the parts related to the embodiment of the present invention are shown, and the details are as follows:

[0080]Considering that in order to achieve a more ideal text recognition effect, it is not enough to rely on basic data sets (open source data sets and synthetic data sets) to train the OCR model, and some actual data sets are needed to fine-tune the model. Therefore, the following steps can be adopted when training the OCR model:

[0081] In step S201, the OCR model is trained using the basic data set described in Embodiment 1 until the model converges to obtain a pre-trained model;

[0082] In the embodiment of the present invention, the OCR model may be a model adopting a common structure such as CRNN. In the spec...

Embodiment 3

[0088] image 3 The structure of the data augmentation device for OCR recognition provided by Embodiment 3 of the present invention is shown. For the convenience of description, only the parts related to the embodiment of the present invention are shown, including:

[0089] A recognition dictionary building unit 31, configured to build a recognition dictionary;

[0090] The first dictionary establishment unit 32 is used to establish the first word frequency dictionary based on the recognition dictionary and the obtained open source data set;

[0091] A document building unit 33, configured to create a synthetic data set text document based on the first word frequency dictionary; and

[0092] The data augmentation unit 34 is configured to perform data augmentation on the current dataset based on the established attributes of the dataset, the application scenario identified by OCR, and the text document of the synthesized dataset to obtain an augmented basic dataset.

[0093] ...

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 is suitable for the technical field of character recognition, and provides a data augmentation method and device for OCR, an apparatus and a storage medium. The method comprises the steps of building a recognition dictionary, building a first word frequency dictionary based on the recognition dictionary and an obtained open source data set, building a synthetic data set text document based on the first word frequency dictionary, carrying out data augmentation on the current data set based on the established data set attributes, an OCR application scene and the synthesized data set text document, and obtaining the augmented basic data set, so that the cost of obtaining training samples in an OCR depth algorithm is reduced, and the pertinence of data augmentation is improved.

Description

technical field [0001] The invention belongs to the technical field of character recognition, and in particular relates to a data augmentation method, device, equipment and storage medium for OCR recognition. Background technique [0002] OCR (Optical Character Recognition, Optical Character Recognition) refers to the process in which electronic devices check characters printed on paper, determine their shapes by detecting dark and bright patterns, and then use character recognition methods to translate the shapes into computer text. OCR recognition has a wide range of applications, such as document recognition, document recognition, etc. [0003] At present, there are two main methods of OCR: based on the traditional OCR algorithm and the OCR method based on deep learning. In recent years, the application of deep learning network structure has made OCR recognition accuracy and stability much higher than traditional OCR methods. However, deep learning relies on a large num...

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 Applications(China)
IPC IPC(8): G06K9/00G06K9/62G06F40/216G06F40/242
CPCG06F40/216G06F40/242G06F18/214
Inventor 不公告发明人
Owner 珠海亿智电子科技有限公司
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Patsnap Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Patsnap Eureka Blog
Learn More
PatSnap group products

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

© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap|About US| Contact US: help@patsnap.com