A method and terminal for creating a structured document based on a deep learning model

A structured document and deep learning technology, applied in the field of data processing, can solve problems such as bill displacement, low accuracy, and exceeding the setting range, and achieve the effects of improving accuracy, improving efficiency and accuracy, and saving resources

Active Publication Date: 2019-05-28
上海深杳智能科技有限公司 +1
View PDF8 Cites 16 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

However, this document structure system is only effective when the positions of the fields to be intercepted are fixed in all documents, which limits the scope of use of the system
In actual application, if the invoice printing system sets different key field content printing position settings or the key field content length changes, the content information of these key fields will be offset and exceed the setting range, resulting in errors
For some bill recognition applications, a large number of bills are stored in the computer by scanning or taking pictures with mobile phones, which can easily cause the displacement of bills in the image, and different bills may have different formats, and the position of the same field in the image is not necessarily the same , these characteristics make the above-mentioned document structuring scheme not suitable for businesses such as bill identification
For the above document structuring scheme, the accuracy of converting images into structured document nodes is low for application scenarios prone to position shifting

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 method and terminal for creating a structured document based on a deep learning model
  • A method and terminal for creating a structured document based on a deep learning model
  • A method and terminal for creating a structured document based on a deep learning model

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0062] Such as image 3 As shown, the present invention provides a method for creating a structured document based on a deep learning model, including:

[0063] S1, a preset training sample set; each sample in the training sample set includes a document picture and an annotated document corresponding to the document picture; the annotated document records the position information and location information of each key field in the document picture category information.

[0064] For example, collect 1000 bill pictures, and use them as samples after processing, part of the samples are used as training samples, and part of them are used as test samples. Each ticket includes a certain number of fields, including key fields of interest. Each sample includes an image of the document, and a document annotating the key fields. The annotation document records the position of each key field in the document picture and the category information of the key field. Document annotation can ...

Embodiment 2

[0098] Such as Image 6 As shown, the present invention also provides a terminal for creating structured documents based on a deep learning model, including one or more processors 1 and a memory 2, the memory 2 stores a program, and is configured to be controlled by the one or more Multiple processors 1 perform the following steps:

[0099] S1, a preset training sample set; each sample in the training sample set includes a document picture and an annotated document corresponding to the document picture; the annotated document records the position information and location information of each key field in the document picture category information.

[0100] For example, collect 1000 bill pictures, and use them as samples after processing, part of the samples are used as training samples, and part of them are used as test samples. Each ticket includes a certain number of fields, including key fields of interest. Each sample includes an image of the document, and a document anno...

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 method and terminal for creating a structured document based on a deep learning model, and belongs to the field of data processing. The method is characterized by presettinga training sample set, wherein each sample in the training sample set comprises a document picture and a labeled document corresponding to the document picture, and the labeled document records position information and category information of each key field in the document picture; training a preset first deep learning model by using the training sample set to obtain a second deep learning model;analyzing the first document picture by the second deep learning model to obtain position information and category information of each key field in the first document picture; and creating a structured document corresponding to the first document picture according to the position information and the category information of each key field in the first document picture, so that the accuracy of converting the document picture into the structured document is improved.

Description

technical field [0001] The invention relates to a method and a terminal for creating a structured document based on a deep learning model, belonging to the field of data processing. Background technique [0002] Document structuring is a process of extracting key field information from a large amount of text information in a document, such as the payer, payment date, and payee in the receipt, and saving it according to a certain structure. After a large number of documents are processed through document structure, efficient document retrieval, document analysis and other intelligent services can be provided. The key and main technical difficulty of document structuring is to extract key field information from a large amount of text, including determining the position of the required key field in the document and identifying the positioned text. [0003] For some document structuring applications with large business volume and high precision requirements, such as invoice rei...

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
IPC IPC(8): G06N20/00G06K9/00
Inventor黄征陈凯周曲周异何建华
Owner上海深杳智能科技有限公司