CN114067335A - Electronic archive text recognition method, system, computer equipment and storage medium - Google Patents
Electronic archive text recognition method, system, computer equipment and storage medium Download PDFInfo
- Publication number
- CN114067335A CN114067335A CN202111318684.6A CN202111318684A CN114067335A CN 114067335 A CN114067335 A CN 114067335A CN 202111318684 A CN202111318684 A CN 202111318684A CN 114067335 A CN114067335 A CN 114067335A
- Authority
- CN
- China
- Prior art keywords
- text
- recognition
- file
- electronic
- inputting
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
Abstract
The invention relates to the technical field of electronic archive text recognition, and discloses an electronic archive text recognition method, a system, computer equipment and a storage medium. The method comprises the following steps: acquiring a scanning image of a paper file, and establishing a data set; carrying out target detection on the scanned image to obtain a plurality of local image blocks to be identified; performing character recognition on the local image block to be recognized to obtain a corresponding recognition text; inputting the recognition text into a neural network for training to obtain a label corresponding to the recognition text; and inputting the identification text into a preset standard format of the electronic file according to the label to generate the electronic file corresponding to the paper file. The invention can automatically convert the paper file into the electronic file with the standard format, thereby improving the filing efficiency of the electronic file and reducing the consumption of manpower and material resources.
Description
Technical Field
The present invention relates to the field of text recognition technology, and in particular, to a method, a system, a computer device, and a storage medium for text recognition of an electronic archive.
Background
At present, with the development of informatization, more and more industries enter a paperless office stage, and file management of a campus is bound to go through the paperless office stage. At present, the number of files grows enormously with the lengthening of the time span, and the increase of personnel. The past files are usually stored in a paper form, the files with the older age are more difficult to store, and the past paper files are converted into corresponding electronic files.
Disclosure of Invention
In order to solve the above technical problems, an object of the present invention is to provide an electronic archive text recognition method, system, computer device and storage medium, which can scan paper archives and automatically recognize and generate corresponding electronic archives, and can improve the filing efficiency and the standardization process of archives.
In a first aspect, the present invention provides a method for recognizing a text of an electronic archive, the method comprising:
acquiring a scanning image of a paper file, and establishing a data set;
carrying out target detection on the scanned image to obtain a plurality of local image blocks to be identified;
performing character recognition on the local image block to be recognized to obtain a corresponding recognition text;
inputting the recognition text into a neural network for training to obtain a label corresponding to the recognition text;
and inputting the identification text into a preset standard format of the electronic file according to the label to generate the electronic file corresponding to the paper file.
Further, the step of obtaining a scanned image of the paper archive and creating the data set comprises:
generating a label corresponding to each title in all formats according to a preset standard format of the electronic file and a preset format of the paper file;
and scanning the paper file to generate a file scanning image, labeling the label, and generating a training set and a testing set.
Further, the step of performing target detection on the scanned image to obtain a plurality of local image blocks to be identified includes:
inputting the scanned image into a YOLO model for target detection to obtain a plurality of local image boundary frames of the scanned image;
and taking the local image in the boundary frame as a local image block to be identified.
Further, the step of performing character recognition on the local image block to be recognized to obtain a corresponding recognition text includes: and inputting the local image block to be recognized into a CRNN model for character recognition to obtain a corresponding recognition text.
Further, the step of inputting the recognition text into a neural network for training to obtain a label corresponding to the recognition text includes:
training the CNN convolutional neural network by using a training set;
inputting the recognition text into the trained CNN convolutional neural network model for classification prediction to obtain a classification label corresponding to the recognition text.
Further, the step of inputting the text to be recognized into a preset standard format of an electronic file according to the tag, and generating the electronic file corresponding to the paper file includes:
comparing the label with each title in the standard format of the electronic file;
and inputting the identification text corresponding to the tags which are consistent in comparison into the title position corresponding to the electronic file to generate the electronic file corresponding to the paper file.
Further, the step of inputting the text to be recognized into a preset standard format of an electronic archive according to the tag and generating the electronic archive corresponding to the paper archive further includes:
inputting the identification text corresponding to the tags which are inconsistent in comparison into a preset electronic document;
and taking the electronic document as a second electronic file corresponding to the paper file.
In a second aspect, the present invention provides an electronic archive text recognition system, comprising:
the data set establishing module is used for acquiring a scanning image of the paper file and establishing a data set;
the image detection module is used for carrying out target detection on the scanned image to obtain a plurality of local image blocks to be identified;
the text recognition module is used for carrying out character recognition on the local image blocks to be recognized to obtain corresponding recognition texts;
the text classification module is used for inputting the recognition text into a neural network for training to obtain a label corresponding to the recognition text;
and the file generation module is used for inputting the identification text into a preset standard format of the electronic file according to the label to generate the electronic file corresponding to the paper file.
In a third aspect, an embodiment of the present invention further provides a computer device, which includes a memory, a processor, and a computer program stored in the memory and executable on the processor, where the processor implements the steps of the method when executing the computer program.
In a fourth aspect, the present invention further provides a computer-readable storage medium, on which a computer program is stored, where the computer program is executed by a processor to implement the steps of the above method.
The invention provides a method, a system, computer equipment and a storage medium for recognizing electronic archive texts. By the method, the paper file can be automatically identified and converted into the electronic document with the standard format, so that the manpower resource is saved, the economic consumption is reduced, and the conversion efficiency of the electronic file and the file standardization process are improved.
Drawings
FIG. 1 is a flowchart illustrating a text recognition method for an electronic file according to an embodiment of the invention;
FIG. 2 is a schematic flow chart of step S10 in FIG. 1;
FIG. 3 is a schematic flow chart of step S20 in FIG. 1;
FIG. 4 is a schematic flow chart of step S40 in FIG. 1;
FIG. 5 is a schematic flow chart of step S50 in FIG. 1;
FIG. 6 is another schematic flow chart of step S50 in FIG. 1;
FIG. 7 is a block diagram of an electronic archive text recognition system in accordance with an embodiment of the present invention;
fig. 8 is an internal structural diagram of a computer device in the embodiment of the present invention.
Detailed Description
In order to make the objects, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are some, but not all, embodiments of the present invention. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
Referring to FIG. 1, a text recognition method for an electronic file according to a first embodiment of the present invention includes steps S10-S50:
and step S10, acquiring a scanned image of the paper file and establishing a data set.
As shown in fig. 2, the specific steps of creating a data set for a paper archive are as follows:
and S101, generating a label corresponding to each title in all formats according to a preset standard format of the electronic file and a preset format of the paper file.
And S102, scanning the paper archive to generate an archive scanning image, labeling the label, and generating a training set and a test set.
Firstly, the format of the electronic file needs to be set so that the subsequent file management is performed in a standardized manner, so that different standardized formats are set for different types of electronic files in advance, and the header in the standardized format is extracted as the label of the data set.
For the original paper file, the format of the paper file is not uniform due to time, the format of each type of paper file needs to be sorted, each title is extracted, the titles are compared with the titles of the electronic file in a similar manner, the same or similar titles are used as the same label, and different titles are used as labels independently. This has the advantage of preventing the loss of content due to parts of the paper file that do not conform to the standard format. For the paper file, a corresponding scanned image needs to be generated by scanning, and a scanner or scanning software and the like can be used without specific setting. And establishing a corresponding training set and a corresponding test set for the obtained image and the label for use in subsequent steps, wherein the data set establishing process is carried out according to a conventional method, and details are not repeated here.
And step S20, performing target detection on the scanned image to obtain a plurality of local image blocks to be identified.
For a scanned image, we use a target detection model YOLO to perform local target detection, and the specific steps are shown in fig. 3:
step S201, inputting the scanned image into a YOLO model for target detection, and obtaining a plurality of local image bounding boxes of the scanned image.
And step S202, taking the local image in the boundary box as a local image block to be identified.
Since the YOLO model is also being updated, we have chosen the YOLO 5 model for object recognition, but other versions of the YOLO model can also perform recognition of this step, and the YOLO 5 model is only used as an example here.
Firstly, a scanned image is preprocessed, the labeling of a data set is processed into a vector form suitable for a Yolov5 model, and the size of the image is uniformly set. Then inputting a YOLOV5 model for local target detection. Since the archive content is usually in a table format or a document format, it is only necessary to use the YOLOV5 model to detect the content frame of each part in the archive file, for example, the whole education experience part or the end-of-term result part is framed, and the content in the whole education experience part or the end-of-term result part is used as the local image blocks to be recognized according to the bounding box, and each paper archive can recognize different numbers of local image blocks to be recognized. The loss function used by the YOLOV5 model is a common intersection-ratio loss function, and therefore is not described more than necessary.
And step S30, performing character recognition on the local image block to be recognized to obtain a corresponding recognition text.
After obtaining a local image block to be recognized, character recognition is carried out on the image block by using a character recognition model CRNN, a convolution circulation neural network CRNN comprises a convolution layer, a circulation layer and a transcription layer, the image block is firstly input into the neural network CRNN, the convolution layer carries out scaling pretreatment on the image, and after convolution, maximum pooling and batch normalization, the characteristics of a sequence are extracted. The loop layer is composed of a bidirectional LSTM loop neural network, the label distribution of each feature vector in the feature sequence is predicted, and the results of the feature sequences predicted by the LSTM network are integrated through the transcription layer and converted into the final output result. The CRNN model is used for text recognition, so that end-to-end training can be performed, character segmentation is not required to be performed on sample data, and a text sequence with any length can be recognized. It should be understood that other models with the same functions may also be applied to the present embodiment, and are not described herein again.
And step S40, inputting the recognition text into a neural network for training to obtain a label corresponding to the recognition text.
After the several sections of recognition texts corresponding to the paper archives are obtained through the above steps, for the training of the recognition texts, refer to the steps shown in fig. 4:
step S401, training the CNN convolutional neural network by using the training set.
Step S402, inputting the recognition text into the trained CNN convolutional neural network model for classification prediction to obtain a classification label corresponding to the recognition text.
We choose to use the CNN convolutional neural network to classify and identify the identification text, first train the CNN convolutional neural network through the data set established in step S10, then input the obtained identification text into the trained CNN neural network for classification and prediction, and the steps after obtaining the corresponding label by prediction are shown in fig. 5:
and step S50, inputting the identification text into a preset standard format of the electronic file according to the label, and generating the electronic file corresponding to the paper file.
Because a plurality of conventional paper files do not have a unified standard format, the predicted label is not completely consistent with the label corresponding to the electronic file with the standard format, and the specific processing steps for the condition that the labels are consistent are as follows:
step S501, comparing the label with each title in the standard format of the electronic file.
Step S502, inputting the identification text corresponding to the tags which are compared to be consistent to the title position corresponding to the electronic file, and generating the electronic file corresponding to the paper file.
The labels obtained after the paper files are classified and predicted are compared with the labels of the electronic files in the standard format, and under the condition that the labels are consistent, the data corresponding to the labels can be directly input to the title content corresponding to the labels of the electronic files. For the case of inconsistent labels, the specific processing steps are shown in fig. 6:
step S503, inputting the identification text corresponding to the inconsistent tag into a preset electronic document.
Step S504, the electronic document is used as a second electronic file corresponding to the paper file.
In order to prevent the loss of the content of the paper file caused by the situation, an electronic document is preset in advance, data with unmatched labels can be input into the electronic document, the electronic document is associated with the electronic file generated by matching the labels, and the electronic document is used as a second electronic file of the paper file, so that the content of the paper file cannot be lost, and the integrity of the file is ensured.
Compared with the traditional method, the method for identifying the text of the electronic file provided by the embodiment cannot automatically convert the paper file with the non-standard format into the electronic file with the standard format, and needs to consume a large amount of manpower and material resources, the method can directly and automatically convert various types of paper files into the electronic file with the standard format, improves the filing efficiency and the standardization process of the electronic file, and ensures the completeness of the electronic file.
Referring to fig. 7, based on the same inventive concept, a second embodiment of the present invention provides an electronic file text recognition system, which includes:
a data set establishing module 10, configured to obtain a scanned image of a paper archive, and establish a data set;
the image detection module 20 is configured to perform target detection on the scanned image to obtain a plurality of local image blocks to be identified;
the text recognition module 30 is configured to perform character recognition on the local image block to be recognized to obtain a corresponding recognition text;
the text classification module 40 is configured to input the recognition text into a neural network for training, so as to obtain a label corresponding to the recognition text;
and the file generation module 50 is configured to input the identification text into a standard format of a preset electronic file according to the tag, and generate an electronic file corresponding to the paper file.
The technical features and technical effects of the electronic file text recognition system provided by the embodiment of the invention are the same as those of the method provided by the embodiment of the invention, and are not repeated herein. The modules in the electronic archive text recognition system can be wholly or partially realized by software, hardware and a combination thereof. The modules can be embedded in a hardware form or independent from a processor in the computer device, and can also be stored in a memory in the computer device in a software form, so that the processor can call and execute operations corresponding to the modules.
Referring to fig. 8, in an embodiment, an internal structure of a computer device may specifically be a terminal or a server. The computer apparatus includes a processor, a memory, a network interface, a display, and an input device connected by a system bus. Wherein the processor of the computer device is configured to provide computing and control capabilities. The memory of the computer device comprises a nonvolatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operation of an operating system and computer programs in the non-volatile storage medium. The network interface of the computer device is used for communicating with an external terminal through a network connection. The computer program is executed by a processor to implement an electronic archive text recognition method. The display screen of the computer equipment can be a liquid crystal display screen or an electronic ink display screen, and the input device of the computer equipment can be a touch layer covered on the display screen, a key, a track ball or a touch pad arranged on the shell of the computer equipment, an external keyboard, a touch pad or a mouse and the like.
It will be appreciated by those of ordinary skill in the art that the architecture shown in FIG. 8 is a block diagram of only a portion of the architecture associated with the subject application, and is not intended to limit the computing devices to which the subject application may be applied, as a particular computing device may include more or less components than those shown in the figures, or may combine certain components, or have the same arrangement of components.
In addition, an embodiment of the present invention further provides a computer device, which includes a memory, a processor, and a computer program stored in the memory and executable on the processor, where the processor implements the steps of the method when executing the computer program.
Furthermore, an embodiment of the present invention further provides a computer-readable storage medium, on which a computer program is stored, and the computer program, when executed by a processor, implements the steps of the above method.
To sum up, the method, the system, the computer device and the storage medium for identifying the text of the electronic archive, provided by the embodiment of the invention, establish a data set by acquiring a scanned image of a paper archive; carrying out target detection on the scanned image to obtain a plurality of local image blocks to be identified; performing character recognition on the local image block to be recognized to obtain a corresponding recognition text; inputting the recognition text into a neural network for training to obtain a label corresponding to the recognition text; and inputting the identification text into a preset standard format of the electronic file according to the label to generate the electronic file corresponding to the paper file. The invention can automatically identify and convert the paper file into the electronic document with the standard format, thereby saving human resources, reducing economic consumption and improving the conversion efficiency of the electronic file and the file standardization process.
The embodiments in this specification are described in a progressive manner, and all the same or similar parts of the embodiments are directly referred to each other, and each embodiment is described with emphasis on differences from other embodiments. In particular, for the system embodiment, since it is substantially similar to the method embodiment, the description is simple, and for the relevant points, reference may be made to the partial description of the method embodiment. It should be noted that, the technical features of the embodiments may be arbitrarily combined, and for the sake of brevity, all possible combinations of the technical features in the embodiments are not described, but should be considered as the scope of the present specification as long as there is no contradiction between the combinations of the technical features.
The above-described examples merely represent some of the preferred embodiments of the present application and are illustrative thereof. Are specific and detailed, but are not to be construed as limiting the scope of the invention. It should be noted that, for those skilled in the art, various modifications and substitutions can be made without departing from the technical principle of the present invention, and these should be construed as the protection scope of the present application. Therefore, the protection scope of the present patent shall be subject to the protection scope of the claims.
Claims (10)
1. An electronic archive text recognition method is characterized by comprising the following steps:
acquiring a scanning image of a paper file, and establishing a data set;
carrying out target detection on the scanned image to obtain a plurality of local image blocks to be identified;
performing character recognition on the local image block to be recognized to obtain a corresponding recognition text;
inputting the recognition text into a neural network for training to obtain a label corresponding to the recognition text;
and inputting the identification text into a preset standard format of the electronic file according to the label to generate the electronic file corresponding to the paper file.
2. The method of claim 1, wherein said step of obtaining a scanned image of a paper document and creating a data set comprises:
generating a label corresponding to each title in all formats according to a preset standard format of the electronic file and a preset format of the paper file;
and scanning the paper file to generate a file scanning image, labeling the label, and generating a training set and a testing set.
3. The method of claim 1, wherein the step of performing target detection on the scanned image to obtain a plurality of local image blocks to be recognized comprises:
inputting the scanned image into a YOLO model for target detection to obtain a plurality of local image boundary frames of the scanned image;
and taking the local image in the boundary frame as a local image block to be identified.
4. The method for recognizing the text of the electronic archive according to claim 1, wherein the step of performing character recognition on the local image block to be recognized to obtain the corresponding recognized text comprises: and inputting the local image block to be recognized into a CRNN model for character recognition to obtain a corresponding recognition text.
5. The method of claim 2, wherein the step of inputting the recognition text into a neural network for training to obtain the label corresponding to the recognition text comprises:
training the CNN convolutional neural network by using a training set;
inputting the recognition text into the trained CNN convolutional neural network model for classification prediction to obtain a classification label corresponding to the recognition text.
6. The method for recognizing the text of the electronic file according to claim 1, wherein the step of inputting the text to be recognized into a preset standard format of the electronic file according to the tag and generating the electronic file corresponding to the paper file comprises the steps of:
comparing the label with each title in the standard format of the electronic file;
and inputting the identification text corresponding to the tags which are consistent in comparison into the title position corresponding to the electronic file to generate the electronic file corresponding to the paper file.
7. The method for recognizing the text of the electronic file according to claim 1, wherein the step of inputting the text to be recognized into a preset standard format of the electronic file according to the tag and generating the electronic file corresponding to the paper file further comprises:
inputting the identification text corresponding to the tags which are not consistent in comparison into a preset electronic document;
and taking the electronic document as a second electronic file corresponding to the paper file.
8. An electronic file character recognition system, comprising:
the data set establishing module is used for acquiring a scanning image of the paper file and establishing a data set;
the image detection module is used for carrying out target detection on the scanned image to obtain a plurality of local image blocks to be identified;
the text recognition module is used for carrying out character recognition on the local image blocks to be recognized to obtain corresponding recognition texts;
the text classification module is used for inputting the recognition text into a neural network for training to obtain a label corresponding to the recognition text;
and the file generation module is used for inputting the identification text into a preset standard format of the electronic file according to the label to generate the electronic file corresponding to the paper file.
9. A computer device comprising a memory, a processor and a computer program stored on the memory and executable on the processor, characterized in that the steps of the method of any of claims 1 to 7 are implemented when the computer program is executed by the processor.
10. A computer-readable storage medium, on which a computer program is stored, which, when being executed by a processor, carries out the steps of the method of any one of claims 1 to 7.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111318684.6A CN114067335A (en) | 2021-11-08 | 2021-11-08 | Electronic archive text recognition method, system, computer equipment and storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111318684.6A CN114067335A (en) | 2021-11-08 | 2021-11-08 | Electronic archive text recognition method, system, computer equipment and storage medium |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN114067335A true CN114067335A (en) | 2022-02-18 |
Family
ID=80273724
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202111318684.6A Pending CN114067335A (en) | 2021-11-08 | 2021-11-08 | Electronic archive text recognition method, system, computer equipment and storage medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN114067335A (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115439871A (en) * | 2022-09-13 | 2022-12-06 | 北京航星永志科技有限公司 | Automatic file acquisition method and device and electronic equipment |
| CN116304266A (en) * | 2023-03-03 | 2023-06-23 | 苏州工业园区航星信息技术服务有限公司 | File management system |
-
2021
- 2021-11-08 CN CN202111318684.6A patent/CN114067335A/en active Pending
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115439871A (en) * | 2022-09-13 | 2022-12-06 | 北京航星永志科技有限公司 | Automatic file acquisition method and device and electronic equipment |
| CN116304266A (en) * | 2023-03-03 | 2023-06-23 | 苏州工业园区航星信息技术服务有限公司 | File management system |
| CN116304266B (en) * | 2023-03-03 | 2024-02-27 | 苏州工业园区航星信息技术服务有限公司 | File management system |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US10650231B2 (en) | Method, device and server for recognizing characters of claim document, and storage medium | |
| US10282643B2 (en) | Method and apparatus for obtaining semantic label of digital image | |
| CN108256591B (en) | Method and apparatus for outputting information | |
| CN114067335A (en) | Electronic archive text recognition method, system, computer equipment and storage medium | |
| US20070211942A1 (en) | Multiple image input for optical character recognition processing systems and methods | |
| US7738702B2 (en) | Image processing apparatus and image processing method capable of executing high-performance processing without transmitting a large amount of image data to outside of the image processing apparatus during the processing | |
| CN109271613B (en) | PDF file analysis method | |
| CN111340037A (en) | Text layout analysis method and device, computer equipment and storage medium | |
| US6816277B2 (en) | Systems and methods for converting image data | |
| CN112417899A (en) | Character translation method, device, computer equipment and storage medium | |
| CN114022891A (en) | Method, device and equipment for extracting key information of scanned text and storage medium | |
| CN113850060A (en) | Civil aviation document data identification and entry method and system | |
| CN112839185B (en) | Method, apparatus, device and medium for processing image | |
| CN116860747A (en) | Training sample generation method and device, electronic equipment and storage medium | |
| CN107861931B (en) | Template file processing method and device, computer equipment and storage medium | |
| CN110059559A (en) | The processing method and its electronic equipment of OCR identification file | |
| CN114818627A (en) | Form information extraction method, device, equipment and medium | |
| CN112801086A (en) | Method and device for classifying certificates, electronic equipment and storage medium | |
| CN113822126A (en) | Icon recognition method, device and computer-readable storage medium | |
| CN114365202A (en) | Extensible structure learned via context-free recursive document decomposition | |
| CN113515280A (en) | Page code generation method and device | |
| US20200272688A1 (en) | Information processing apparatus and non-transitory computer readable medium | |
| CN110674091A (en) | File uploading method and system based on artificial intelligence and storage medium | |
| CN115982358B (en) | Document splitting method, device, terminal equipment and computer readable storage medium | |
| US20240135739A1 (en) | Method of classifying a document for a straight-through processing |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |