CN118097674B - Deep learning-based intelligent automatic auditing method and device for enterprise text information - Google Patents
Deep learning-based intelligent automatic auditing method and device for enterprise text information Download PDFInfo
- Publication number
- CN118097674B CN118097674B CN202311529112.1A CN202311529112A CN118097674B CN 118097674 B CN118097674 B CN 118097674B CN 202311529112 A CN202311529112 A CN 202311529112A CN 118097674 B CN118097674 B CN 118097674B
- Authority
- CN
- China
- Prior art keywords
- enterprise
- field
- image
- identification
- information
- 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.)
- Active
Links
- 238000013135 deep learning Methods 0.000 title claims abstract description 41
- 238000000034 method Methods 0.000 title claims abstract description 31
- 238000012795 verification Methods 0.000 claims description 48
- 238000012545 processing Methods 0.000 claims description 17
- 230000004044 response Effects 0.000 claims description 13
- 238000004590 computer program Methods 0.000 claims description 11
- 238000012550 audit Methods 0.000 claims description 6
- 230000004927 fusion Effects 0.000 description 46
- 238000003062 neural network model Methods 0.000 description 22
- 238000000605 extraction Methods 0.000 description 19
- 230000006870 function Effects 0.000 description 10
- 238000010586 diagram Methods 0.000 description 9
- 239000003086 colorant Substances 0.000 description 6
- 238000004891 communication Methods 0.000 description 6
- 230000007246 mechanism Effects 0.000 description 6
- 238000012015 optical character recognition Methods 0.000 description 4
- 238000013528 artificial neural network Methods 0.000 description 3
- 238000002372 labelling Methods 0.000 description 3
- 230000003287 optical effect Effects 0.000 description 3
- 239000013598 vector Substances 0.000 description 3
- 238000013527 convolutional neural network Methods 0.000 description 2
- 230000000694 effects Effects 0.000 description 2
- 230000008569 process Effects 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000005540 biological transmission Effects 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 239000000835 fiber Substances 0.000 description 1
- 239000004973 liquid crystal related substance Substances 0.000 description 1
- 239000013307 optical fiber Substances 0.000 description 1
- 230000000644 propagated effect Effects 0.000 description 1
- 239000004065 semiconductor Substances 0.000 description 1
- 238000012549 training Methods 0.000 description 1
- 238000012546 transfer Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V30/00—Character recognition; Recognising digital ink; Document-oriented image-based pattern recognition
- G06V30/10—Character recognition
- G06V30/14—Image acquisition
- G06V30/148—Segmentation of character regions
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V30/00—Character recognition; Recognising digital ink; Document-oriented image-based pattern recognition
- G06V30/10—Character recognition
- G06V30/19—Recognition using electronic means
- G06V30/191—Design or setup of recognition systems or techniques; Extraction of features in feature space; Clustering techniques; Blind source separation
- G06V30/19147—Obtaining sets of training patterns; Bootstrap methods, e.g. bagging or boosting
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02P—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN THE PRODUCTION OR PROCESSING OF GOODS
- Y02P90/00—Enabling technologies with a potential contribution to greenhouse gas [GHG] emissions mitigation
- Y02P90/30—Computing systems specially adapted for manufacturing
Landscapes
- Engineering & Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Multimedia (AREA)
- Theoretical Computer Science (AREA)
- Character Discrimination (AREA)
Abstract
The embodiment of the invention discloses an enterprise text information intelligent automatic auditing method and device based on deep learning. One embodiment of the method comprises the following steps: acquiring text information of an enterprise to be checked; inputting the first enterprise identification file image and the second enterprise identification file image into a pre-trained main field identification model to obtain a first image field identification result and a second image field identification result; inputting the enterprise text information form into a main field recognition model to obtain a form field recognition result; acquiring enterprise information corresponding to the enterprise name from the associated enterprise information query system as target enterprise information according to the enterprise name; and according to the target enterprise information, checking and checking the first image field identification result, the second image field identification result and the form field identification result to obtain an enterprise text information checking and checking result. According to the method and the device, the identification efficiency is improved, and the auditing period of enterprise information is shortened.
Description
Technical Field
The embodiment of the disclosure relates to the technical field of computers, in particular to an enterprise text information intelligent automatic auditing method and device based on deep learning.
Background
At present, the OCR recognition business license information only supports the structural recognition of key fields such as certificate numbers, social credit codes, unit names, addresses, legal persons, types, established dates, effective dates, operation ranges and the like of different formats of business licenses.
However, the current way of identifying business license information generally has the following technical problems:
Firstly, unstructured identification of the file cannot be performed, so that the identification efficiency of enterprise text information (business license information) is low, and the auditing period is long;
Second, when the business license image is recognized, the degree of association of the image with text information is low, which makes it difficult to accurately recognize the text information in the business license.
The above information disclosed in this background section is only for enhancement of understanding of the background of the inventive concept and, therefore, may contain information that does not form the prior art that is already known to those of ordinary skill in the art in this country.
Disclosure of Invention
The disclosure is in part intended to introduce concepts in a simplified form that are further described below in the detailed description. The disclosure is not intended to identify key features or essential features of the claimed subject matter, nor is it intended to be used to limit the scope of the claimed subject matter.
Some embodiments of the present disclosure provide a deep learning-based method, apparatus, electronic device, and computer-readable medium for intelligent automatic auditing of enterprise text information, to address one or more of the technical problems mentioned in the background section above.
In a first aspect, some embodiments of the present disclosure provide an enterprise text information intelligent automatic auditing method based on deep learning, the method including: acquiring enterprise text information to be checked, wherein the enterprise text information to be checked comprises a first enterprise identification file image, a second enterprise identification file image and an enterprise text information table; inputting the first enterprise identification file image and the second enterprise identification file image into a pre-trained main field identification model to obtain a first image field identification result and a second image field identification result; inputting the enterprise text information form into the main field recognition model to obtain a form field recognition result, wherein the form field recognition result comprises an enterprise name; acquiring enterprise information corresponding to the enterprise name from an associated enterprise information query system as target enterprise information according to the enterprise name; according to the target enterprise information, checking and checking the first image field identification result, the second image field identification result and the form field identification result to obtain an enterprise text information checking and checking result; and sending the enterprise text information auditing and checking result to the associated enterprise information auditing terminal so that the enterprise information auditing terminal selects a target enterprise supply end.
In a second aspect, some embodiments of the present disclosure provide an enterprise text information intelligent automatic auditing apparatus based on deep learning, the apparatus including: the system comprises a first acquisition unit, a second acquisition unit and a third acquisition unit, wherein the first acquisition unit is configured to acquire enterprise text information to be checked, and the enterprise text information to be checked comprises a first enterprise identification file image, a second enterprise identification file image and an enterprise text information table; the first input unit is configured to input the first enterprise identification file image and the second enterprise identification file image into a pre-trained main field identification model to obtain a first image field identification result and a second image field identification result; a second input unit configured to input the enterprise text information table into the main field recognition model to obtain a table field recognition result, wherein the table field recognition result includes an enterprise name; a second acquiring unit configured to acquire, from an associated enterprise information inquiry system, enterprise information corresponding to the enterprise name as target enterprise information, according to the enterprise name; the verification unit is configured to carry out verification and verification on the first image field identification result, the second image field identification result and the form field identification result according to the target enterprise information to obtain an enterprise text information verification and verification result; and the sending unit is configured to send the enterprise text information auditing and checking result to the associated enterprise information auditing terminal so as to select a target enterprise supply end by the enterprise information auditing terminal.
In a third aspect, some embodiments of the present disclosure provide an electronic device comprising: one or more processors; a storage device having one or more programs stored thereon, which when executed by one or more processors causes the one or more processors to implement the method described in any of the implementations of the first aspect above.
In a fourth aspect, some embodiments of the present disclosure provide a computer readable medium having a computer program stored thereon, wherein the computer program, when executed by a processor, implements the method described in any of the implementations of the first aspect above.
The above embodiments of the present disclosure have the following advantageous effects: according to the deep learning-based enterprise text information intelligent automatic auditing method, the identification efficiency is improved, and the auditing period of enterprise information is shortened. Specifically, the reason for the longer audit period is that: the unstructured recognition of the document cannot be performed, resulting in a low recognition efficiency for the enterprise text information (business license information). Based on this, in some embodiments of the present disclosure, an intelligent automatic auditing method for enterprise text information based on deep learning, first, enterprise text information to be audited is obtained. The enterprise text information to be checked comprises a first enterprise identification file image, a second enterprise identification file image and an enterprise text information table. Thus, auditing of the enterprise-related text information is facilitated. And secondly, inputting the first enterprise identification file image and the second enterprise identification file image into a pre-trained main field identification model to obtain a first image field identification result and a second image field identification result. Thus, unstructured files of the enterprise file image class may be identified. And inputting the enterprise text information form into the main field recognition model to obtain a form field recognition result. Wherein, the table field identification result includes the name of the enterprise. And acquiring enterprise information corresponding to the enterprise name from the associated enterprise information query system as target enterprise information according to the enterprise name. Therefore, standard enterprise information can be queried, and verification of the recognized text result is facilitated. And according to the target enterprise information, checking and checking the first image field identification result, the second image field identification result and the form field identification result to obtain an enterprise text information checking and checking result. Therefore, the structured file and the unstructured file can be checked at the same time, and the checking efficiency of enterprise text information is improved. Thus, the auditing period is shortened. And finally, sending the enterprise text information auditing and checking result to the associated enterprise information auditing terminal so that the enterprise information auditing terminal selects a target enterprise supply end. Therefore, business personnel can conveniently select the matched enterprises according to the enterprise text information auditing and checking results.
Drawings
The above and other features, advantages, and aspects of embodiments of the present disclosure will become more apparent by reference to the following detailed description when taken in conjunction with the accompanying drawings. The same or similar reference numbers will be used throughout the drawings to refer to the same or like elements. It should be understood that the figures are schematic and that elements and components are not necessarily drawn to scale.
FIG. 1 is a flow diagram of some embodiments of a deep learning based business text information intelligent automatic auditing method according to the present disclosure;
FIG. 2 is a schematic structural diagram of some embodiments of a deep learning based enterprise text information intelligent automatic auditing device according to the present disclosure;
fig. 3 is a schematic structural diagram of an electronic device suitable for use in implementing some embodiments of the present disclosure.
Detailed Description
Embodiments of the present disclosure will be described in more detail below with reference to the accompanying drawings. While certain embodiments of the present disclosure are shown in the drawings, it should be understood that the present disclosure may be embodied in various forms and should not be construed as limited to the embodiments set forth herein. Rather, these embodiments are provided so that this disclosure will be thorough and complete. It should be understood that the drawings and embodiments of the present disclosure are for illustration purposes only and are not intended to limit the scope of the present disclosure.
It should be noted that, for convenience of description, only the portions related to the present invention are shown in the drawings. Embodiments of the present disclosure and features of embodiments may be combined with each other without conflict.
It should be noted that the terms "first," "second," and the like in this disclosure are merely used to distinguish between different devices, modules, or units and are not used to define an order or interdependence of functions performed by the devices, modules, or units.
It should be noted that references to "one", "a plurality" and "a plurality" in this disclosure are intended to be illustrative rather than limiting, and those of ordinary skill in the art will appreciate that "one or more" is intended to be understood as "one or more" unless the context clearly indicates otherwise.
The names of messages or information interacted between the various devices in the embodiments of the present disclosure are for illustrative purposes only and are not intended to limit the scope of such messages or information.
The present disclosure will be described in detail below with reference to the accompanying drawings in conjunction with embodiments.
Fig. 1 is a flow diagram of some embodiments of a deep learning based business text information intelligent automatic auditing method according to the present disclosure. A flow 100 of some embodiments of a deep learning based business text information intelligent automated auditing method according to the present disclosure is shown. The intelligent automatic auditing method for the enterprise text information based on deep learning comprises the following steps:
and step 101, acquiring text information of an enterprise to be checked.
In some embodiments, an execution body (e.g., a computing device) of the deep learning-based enterprise text information intelligent automatic auditing method may acquire enterprise text information to be audited through a wired connection or a wireless connection. The enterprise text information to be checked comprises a first enterprise identification file image, a second enterprise identification file image and an enterprise text information table. The text information of the enterprise to be checked can represent registered text information of the enterprise to be checked. The first enterprise identification document image may refer to a business license positive document of the enterprise information to be checked. The second enterprise identification file image may refer to a business license copy file of the enterprise information to be reviewed. The business text information table may represent a table made up of various fields included in the business registration information. Enterprise registration information may include, but is not limited to: unified social credit code, business name, legal representative name, unit type, registered funds, business terms, registered locale, etc.
Optionally, a sample of the enterprise text information is obtained.
In some embodiments, the executing entity may obtain a sample of the enterprise text information. Wherein, the enterprise text information sample includes: the system comprises a first enterprise identification file image sample, a second enterprise identification file image sample, an enterprise text information table sample, first enterprise identification file field information corresponding to the first enterprise identification file image sample, second enterprise identification file field information corresponding to the second enterprise identification file image sample and enterprise field information corresponding to the enterprise text information table sample. The first business identification file image sample may represent a first business identification file image. The second business identification file image sample may represent a second business identification file image. The business text information form sample may represent a business text information form. The first enterprise identification file field information may represent a field tag of a first enterprise identification file image sample. The second enterprise identification file field information may represent a field tag of a second enterprise identification file image sample. The enterprise field information may represent enterprise information.
Optionally, feature extraction processing is performed on the first enterprise identification file image sample and the first enterprise identification file field information to obtain a first enterprise identification file image sample feature and a first enterprise identification file field feature.
In some embodiments, the execution body may perform feature extraction processing on the first enterprise identification file image sample and the first enterprise identification file field information to obtain a first enterprise identification file image sample feature and a first enterprise identification file field feature. In practice, the execution body may input the first enterprise identification file image sample and the first enterprise identification file field information into the feature extraction model to obtain the first enterprise identification file image sample feature and the first enterprise identification file field feature. The characteristic extraction model is a neural network which takes an enterprise identification file image sample and enterprise identification file field information as input and takes enterprise identification file image sample characteristics and enterprise identification file field characteristics as output. The feature extraction model comprises an image feature extraction sub-model and a field feature extraction sub-model. The image feature extraction sub-model may be a convolutional neural network that takes an image sample of the enterprise identification file as input and takes a feature of the image sample of the enterprise identification file as output. The image feature extraction sub-model may be a transform coding model that takes the image sample of the enterprise identification document as input and takes the image sample feature of the enterprise identification document as output. The enterprise identification file image sample feature may be an image feature vector. The field feature extraction sub-model may be a convolutional neural network with the enterprise identification file field information as input and the enterprise identification file field feature as output.
Optionally, the image sample feature of the first enterprise identification file and the field feature of the first enterprise identification file are input into a first image-text feature fusion network included in the initial deep learning neural network model, so as to obtain a first image-text fusion feature.
In some embodiments, the execution body may input the image sample feature of the first enterprise identification file and the field feature of the first enterprise identification file into a first image-text feature fusion network included in the initial deep learning neural network model, to obtain a first image-text fusion feature. The first image-text feature fusion network may be an image-text feature fusion network based on a first attention mechanism. The first attention mechanism described above may be a self-attention mechanism. The image-text feature fusion network may be a neural network for fusing image feature vectors and text feature vectors. For example, the above-described image feature fusion network may be a transducer encoding network model.
Optionally, performing feature extraction processing on the second enterprise identification file image sample and the second enterprise identification file field information to obtain a second enterprise identification file image sample feature and a second enterprise identification file field feature.
In some embodiments, the executing body may perform feature extraction processing on the second enterprise identification file image sample and the second enterprise identification file field information to obtain a second enterprise identification file image sample feature and a second enterprise identification file field feature. In practice, the second enterprise identification file image sample and the second enterprise identification file field information may be input into a feature extraction model to obtain a second enterprise identification file image sample feature and a second enterprise identification file field feature.
Optionally, the image sample features of the second enterprise identification file and the field features of the second enterprise identification file are input into a first image-text feature fusion network included in the initial deep learning neural network model, so that second image-text fusion features are obtained.
In some embodiments, the execution body may input the image sample feature of the second enterprise identification file and the field feature of the second enterprise identification file into a first image-text feature fusion network included in the initial deep learning neural network model, to obtain a second image-text fusion feature.
Optionally, performing feature extraction processing on the enterprise text information table sample and the enterprise field information to obtain enterprise text information table sample features and enterprise field information features.
In some embodiments, the executing body may perform feature extraction processing on the enterprise text information table sample and the enterprise field information to obtain an enterprise text information table sample feature and an enterprise field information feature. In practice, the enterprise text information table sample and the enterprise field information can be input into a feature extraction model to obtain enterprise text information table sample features and enterprise field information features.
Optionally, the first image-text fusion feature and the sample feature of the enterprise text information form are input into an enterprise image-text feature fusion network included in the initial deep learning neural network model, so as to obtain a first enterprise image-text fusion feature.
In some embodiments, the executing body may input the first image-text fusion feature and the sample feature of the enterprise text information table into an enterprise image-text feature fusion network included in the initial deep learning neural network model, to obtain a first enterprise image-text fusion feature. The enterprise graphic feature fusion network may be a graphic feature fusion network based on a second attention mechanism. The second attention mechanism may be a multi-headed self-attention mechanism. The enterprise graphic and text feature fusion network may be a neural network for fusing the first graphic and text feature and the sample feature of the enterprise text information table. For example, the enterprise teletext feature fusion network may be a Transformer coding network model.
Optionally, the second image-text fusion feature and the sample feature of the enterprise text information form are input into the enterprise image-text feature fusion network to obtain a second enterprise image-text fusion feature.
In some embodiments, the executing entity may input the second image-text fusion feature and the sample feature of the enterprise text information table into the enterprise image-text feature fusion network to obtain a second enterprise image-text fusion feature.
Optionally, a loss value between the first enterprise graphic fusion feature and the enterprise field information feature is determined as a first loss value.
In some embodiments, the execution body may determine a penalty value between the first enterprise graphic fusion feature and the enterprise field information feature as the first penalty value. In practice, the loss value between the first enterprise graphic fusion feature and the enterprise field information feature may be determined by a cross entropy loss function.
Optionally, a loss value between the second enterprise graphic fusion feature and the enterprise field information feature is determined as a second loss value. In practice, a loss value between the second enterprise graphic fusion feature and the enterprise field information feature may be determined by a cross entropy loss function as a second loss value.
In some embodiments, the execution body may determine a penalty value between the second enterprise graphic fusion feature and the enterprise field information feature as the second penalty value.
Optionally, in response to determining that the first loss value and the second loss value are both less than or equal to a preset threshold, determining the initial deep learning neural network model as a trained primary field recognition model.
In some embodiments, the execution body may determine the initial deep learning neural network model as the trained main field recognition model in response to determining that the first loss value and the second loss value are both less than or equal to a preset threshold.
Optionally, in response to determining that the first loss value and/or the second loss value are/is greater than the preset threshold, adjusting network parameters of the initial deep learning neural network model, and using the adjusted initial deep learning neural network model as the initial deep learning neural network model, re-acquiring the enterprise text information sample to retrain the initial deep learning neural network model again.
In some embodiments, the execution entity may adjust network parameters of the initial deep learning neural network model in response to determining that the first loss value and/or the second loss value are greater than the preset threshold, and re-acquire the enterprise text information sample using the adjusted initial deep learning neural network model as the initial deep learning neural network model to re-train the initial deep learning neural network model
The above-mentioned related matters serve as an invention point of the present disclosure, and solve the second technical problem "the text information in the business license is difficult to accurately identify" mentioned in the background art. Factors that lead to difficulty in accurately identifying text information in business licenses tend to be as follows: when the business license image is identified, the association degree of the image and the text information is low. If the above factors are solved, the effect of improving the recognition accuracy of the text information in the business license can be achieved. To achieve this, first, a sample of the business text information is obtained. And secondly, carrying out feature extraction processing on the first enterprise identification file image sample and the first enterprise identification file field information to obtain first enterprise identification file image sample features and first enterprise identification file field features. And then, inputting the image sample characteristics of the first enterprise identification file and the field characteristics of the first enterprise identification file into a first image-text characteristic fusion network included in the initial deep learning neural network model to obtain first image-text fusion characteristics. Thus, an association between the first enterprise identification file image sample feature and the first enterprise identification file field feature may be established. And then, carrying out feature extraction processing on the second enterprise identification file image sample and the second enterprise identification file field information to obtain second enterprise identification file image sample features and second enterprise identification file field features. And inputting the image sample characteristics of the second enterprise identification file and the field characteristics of the second enterprise identification file into a first image-text characteristic fusion network included in the initial deep learning neural network model to obtain second image-text fusion characteristics. Thus, the association relationship between the second enterprise identification file image sample feature and the second enterprise identification file field feature can be established. And then, carrying out feature extraction processing on the enterprise text information table sample and the enterprise field information to obtain enterprise text information table sample features and enterprise field information features. And then, inputting the first image-text fusion characteristic and the sample characteristic of the enterprise text information form into an enterprise image-text characteristic fusion network included in the initial deep learning neural network model to obtain a first enterprise image-text fusion characteristic. And then, inputting the second image-text fusion characteristic and the sample characteristic of the enterprise text information form into the enterprise image-text characteristic fusion network to obtain the second enterprise image-text fusion characteristic. Then, determining a loss value between the first enterprise graphic fusion characteristic and the enterprise field information characteristic as a first loss value; and determining a loss value between the second enterprise graphic fusion characteristic and the enterprise field information characteristic as a second loss value. And finally, determining the initial deep learning neural network model as a main field recognition model after training in response to determining that the first loss value and the second loss value are smaller than or equal to a preset threshold. Therefore, the accuracy of the trained main field recognition model for recognizing the Chinese information in the business license can be improved by establishing the association relation between the image and the text information.
Step 102, inputting the first enterprise identification file image and the second enterprise identification file image into a pre-trained main field identification model to obtain a first image field identification result and a second image field identification result.
In some embodiments, the execution body may input the first enterprise identification document image and the second enterprise identification document image into a pre-trained main field recognition model to obtain a first image field recognition result and a second image field recognition result. The main field recognition model may be a neural network model which is pre-trained and takes an enterprise identification file image or an enterprise text information table as input and takes a field recognition result as output. For example, the primary field recognition model may be a deep neural network model incorporating optical character recognition (Optical Character Recognition, OCR recognition). The enterprise identification document image may refer to either the first enterprise identification document image or the second enterprise identification document image. The field identification result may refer to key fields in the identified enterprise identification file image or enterprise text information table, and may include, but is not limited to: unifying license-like information contents such as social credit codes, enterprise names, legal representative names, unit types, registered funds, business deadlines, registered places, business license national logo positions and the like, and labeling the business license colors (color, black and white) according to the national logo colors.
And step 103, inputting the enterprise text information form into the main field recognition model to obtain a form field recognition result.
In some embodiments, the execution entity may input the enterprise text information table into the main field recognition model to obtain a table field recognition result. Wherein, the table field identification result includes the name of the enterprise. That is, the form field recognition result may be a key field in the business text information form recognized by the main field recognition model.
Step 104, according to the enterprise name, acquiring enterprise information corresponding to the enterprise name from the associated enterprise information query system as target enterprise information.
In some embodiments, the executing entity may obtain, from the associated enterprise information query system, enterprise information corresponding to the enterprise name as target enterprise information according to the enterprise name. That is, the enterprise information query system may refer to a query system that stores a plurality of enterprise information. For example, the enterprise information query system may be an enterprise query software system. In practice, the enterprise information including the same enterprise name as the above enterprise name may be acquired from the enterprise information query system as the target enterprise information.
And 105, checking and checking the first image field identification result, the second image field identification result and the form field identification result according to the target enterprise information to obtain an enterprise text information checking and checking result.
In some embodiments, the execution body may perform audit verification on the first image field identification result, the second image field identification result, and the form field identification result according to the target enterprise information, to obtain an enterprise text information audit verification result. The first image field identification result may include: unifying license-like information contents such as social credit codes, enterprise names, legal representative names, unit types, registered funds, business deadlines, registered places, business license national logo positions and the like, and labeling the business license colors (color, black and white) according to the national logo colors. The second image field recognition result may include: unifying license-like information contents such as social credit codes, enterprise names, legal representative names, unit types, registered funds, business deadlines, registered places, business license national logo positions and the like, and labeling the business license colors (color, black and white) according to the national logo colors.
In practice, the execution subject may perform an audit verification on the first image field identification result, the second image field identification result, and the table field identification result by:
first, for each field included in the target enterprise information, the following processing steps are performed:
First, it is determined whether or not an identification field corresponding to the above-mentioned field exists in each identification field included in the above-mentioned first image field identification result. That is, it is determined whether or not the same identification field as the above-mentioned field exists in each identification field included in the above-mentioned identification result specifying the above-mentioned first image field.
Then, in response to determining that an identification field corresponding to the above-mentioned field exists in each identification field included in the above-mentioned first image field identification result, a first field verification result is generated. The first field check result indicates that the field check passes.
And then, generating a second field verification result in response to determining that the identification field corresponding to the field does not exist in all identification fields included in the first image field identification result. The first field check result indicates that the field check fails.
And step two, merging the generated first field verification results and the second field verification results into a first image field verification result.
Third, for each field included in the target enterprise information, the following processing steps are performed:
First, it is determined whether or not an identification field corresponding to the above-mentioned field exists in each identification field included in the above-mentioned second image field identification result. That is, it is determined whether or not the same identification field as the above-mentioned field exists in each identification field included in the above-mentioned second image field identification result.
Then, in response to determining that the identification field corresponding to the field exists in the identification fields included in the second image field identification result, a first image field verification result is generated. The first image field verification result indicates that the field verification passes.
And then, generating a second image field verification result in response to determining that the identification field corresponding to the field does not exist in all identification fields included in the second image field identification result. The second image field verification result indicates that the field verification fails.
Fourth, merging the generated first image field verification results and the second image field verification results into second image field verification results.
Fifth, for each field included in the target enterprise information, the following processing steps are performed:
First, it is determined whether or not an identification field corresponding to the above-mentioned field exists in each identification field included in the above-mentioned table field identification result. That is, it is determined whether or not the same identification field as the above-described field exists in each identification field included in the above-described table field identification result.
Then, in response to determining that an identification field corresponding to the above-mentioned field exists in each of the identification fields included in the above-mentioned table field identification result, a first table field verification result is generated. The first table field check result indicates that the above field check passes.
And then, generating a second table field verification result in response to determining that the identification field corresponding to the field does not exist in all the identification fields included in the table field identification result. The second table field check result indicates that the above field check fails.
And sixthly, merging the generated first form field verification results and the second form field verification results into form field verification results.
Seventh, merging the first image field verification result, the second image field verification result and the form field verification result into an enterprise text information auditing verification result.
And step 106, transmitting the enterprise text information auditing and checking result to the associated enterprise information auditing terminal so that the enterprise information auditing terminal selects a target enterprise supply end.
In some embodiments, the executing entity may send the enterprise text information auditing result to an associated enterprise information auditing terminal, so that the enterprise information auditing terminal selects a target enterprise supply terminal. The enterprise information auditing terminal can be a display terminal with a display function and is used for displaying enterprise text information auditing and checking results so that service personnel can select an enterprise supply end.
With further reference to fig. 2, as an implementation of the method shown in the foregoing figures, the present disclosure provides some embodiments of a deep learning-based intelligent automatic auditing apparatus for enterprise text information, which correspond to those method embodiments shown in fig. 1, and may be specifically applied to various electronic devices.
As shown in fig. 2, the deep learning-based enterprise text information intelligent automatic auditing apparatus 200 of some embodiments includes: a first acquisition unit 201, a first input unit 202, a second input unit 203, a second acquisition unit 204, a verification unit 205, and a transmission unit 206. The first obtaining unit 201 is configured to obtain to-be-checked enterprise text information, where the to-be-checked enterprise text information includes a first enterprise identification file image, a second enterprise identification file image and an enterprise text information table; a first input unit 202 configured to input the first enterprise identification file image and the second enterprise identification file image into a pre-trained main field recognition model, so as to obtain a first image field recognition result and a second image field recognition result; a second input unit 203 configured to input the enterprise text information table into the main field recognition model to obtain a table field recognition result, where the table field recognition result includes an enterprise name; a second obtaining unit 204 configured to obtain, from the associated enterprise information inquiry system, enterprise information corresponding to the enterprise name as target enterprise information according to the enterprise name; a verification unit 205 configured to perform verification on the first image field identification result, the second image field identification result, and the table field identification result according to the target enterprise information, to obtain an enterprise text information verification result; and a transmitting unit 206 configured to transmit the enterprise text information auditing result to an associated enterprise information auditing terminal, so that the enterprise information auditing terminal selects a target enterprise supply terminal.
It will be appreciated that the elements described in the deep learning-based intelligent automated auditing device 200 for business text information correspond to the steps of the method described with reference to fig. 1. Thus, the operations, features and the beneficial effects described above for the method are also applicable to the enterprise text information intelligent automatic auditing device 200 based on deep learning and the units contained therein, and are not described herein.
Referring now to fig. 3, a schematic diagram of an electronic device (e.g., computing device) 300 suitable for use in implementing some embodiments of the present disclosure is shown. The electronic devices in some embodiments of the present disclosure may include, but are not limited to, mobile terminals such as mobile phones, notebook computers, digital broadcast receivers, PDAs (personal digital assistants), PADs (tablet computers), PMPs (portable multimedia players), car terminals (e.g., car navigation terminals), and the like, as well as stationary terminals such as digital TVs, desktop computers, and the like. The electronic device shown in fig. 3 is merely an example and should not impose any limitations on the functionality and scope of use of embodiments of the present disclosure.
As shown in fig. 3, the electronic device 300 may include a processing means (e.g., a central processing unit, a graphics processor, etc.) 301 that may perform various suitable actions and processes in accordance with a program stored in a Read Only Memory (ROM) 302 or a program loaded from a storage means 308 into a Random Access Memory (RAM) 303. In the RAM303, various programs and task data required for the operation of the electronic device 300 are also stored. The processing device 301, the ROM302, and the RAM303 are connected to each other via a bus 304. An input/output (I/O) interface 305 is also connected to bus 304.
In general, the following devices may be connected to the I/O interface 305: input devices 306 including, for example, a touch screen, touchpad, keyboard, mouse, camera, microphone, accelerometer, gyroscope, etc.; an output device 307 including, for example, a Liquid Crystal Display (LCD), a speaker, a vibrator, and the like; storage 308 including, for example, magnetic tape, hard disk, etc.; and communication means 309. The communication means 309 may allow the electronic device 300 to communicate with other devices wirelessly or by wire to exchange task data. While fig. 3 shows an electronic device 300 having various means, it is to be understood that not all of the illustrated means are required to be implemented or provided. More or fewer devices may be implemented or provided instead. Each block shown in fig. 3 may represent one device or a plurality of devices as needed.
In particular, according to some embodiments of the present disclosure, the processes described above with reference to flowcharts may be implemented as computer software programs. For example, some embodiments of the present disclosure include a computer program product comprising a computer program embodied on a computer readable medium, the computer program comprising program code for performing the method shown in the flow chart. In such embodiments, the computer program may be downloaded and installed from a network via communications device 309, or from storage device 308, or from ROM 302. The above-described functions defined in the methods of some embodiments of the present disclosure are performed when the computer program is executed by the processing means 301.
It should be noted that, the computer readable medium described in some embodiments of the present disclosure may be a computer readable signal medium or a computer readable storage medium, or any combination of the two. The computer readable storage medium can be, for example, but not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or a combination of any of the foregoing. More specific examples of the computer-readable storage medium may include, but are not limited to: an electrical connection having one or more wires, a portable computer diskette, a hard disk, a Random Access Memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disc read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing. In some embodiments of the present disclosure, a computer readable storage medium may be any tangible medium that can contain, or store a program for use by or in connection with an instruction execution system, apparatus, or device. In some embodiments of the present disclosure, however, the computer-readable signal medium may comprise a task data signal that propagates in baseband or as part of a carrier wave, in which computer-readable program code is carried. Such a propagated signal may take any of a variety of forms, including, but not limited to, electro-magnetic, optical, or any suitable combination of the foregoing. A computer readable signal medium may also be any computer readable medium that is not a computer readable storage medium and that can communicate, propagate, or transport a program for use by or in connection with an instruction execution system, apparatus, or device. Program code embodied on a computer readable medium may be transmitted using any appropriate medium, including but not limited to: electrical wires, fiber optic cables, RF (radio frequency), and the like, or any suitable combination of the foregoing.
In some implementations, the clients, servers may communicate using any currently known or future developed network protocol, such as HTTP (HyperText Transfer Protocol ), and may be interconnected with any form or medium of digital task data communication (e.g., a communication network). Examples of communication networks include a local area network ("LAN"), a wide area network ("WAN"), the internet (e.g., the internet), and peer-to-peer networks (e.g., ad hoc peer-to-peer networks), as well as any currently known or future developed networks.
The computer readable medium may be contained in the electronic device; or may exist alone without being incorporated into the electronic device. The computer readable medium carries one or more programs, which when executed by the electronic device, cause the electronic device to obtain to-be-checked enterprise text information, wherein the to-be-checked enterprise text information includes a first enterprise identification file image, a second enterprise identification file image and an enterprise text information table; inputting the first enterprise identification file image and the second enterprise identification file image into a pre-trained main field identification model to obtain a first image field identification result and a second image field identification result; inputting the enterprise text information form into the main field recognition model to obtain a form field recognition result, wherein the form field recognition result comprises an enterprise name; acquiring enterprise information corresponding to the enterprise name from an associated enterprise information query system as target enterprise information according to the enterprise name; according to the target enterprise information, checking and checking the first image field identification result, the second image field identification result and the form field identification result to obtain an enterprise text information checking and checking result; and sending the enterprise text information auditing and checking result to the associated enterprise information auditing terminal so that the enterprise information auditing terminal selects a target enterprise supply end.
Computer program code for carrying out operations for some embodiments of the present disclosure may be written in one or more programming languages, including a product oriented programming language such as Java, smalltalk, C ++ and conventional procedural programming languages, such as the "C" programming language or similar programming languages. The program code may execute entirely on the user's computer, partly on the user's computer, as a stand-alone software package, partly on the user's computer and partly on a remote computer or entirely on the remote computer or server. In the case of a remote computer, the remote computer may be connected to the user's computer through any kind of network, including a Local Area Network (LAN) or a Wide Area Network (WAN), or may be connected to an external computer (for example, through the Internet using an Internet service provider).
The flowcharts and block diagrams in the figures illustrate the architecture, functionality, and operation of possible implementations of systems, methods and computer program products according to various embodiments of the present disclosure. In this regard, each block in the flowchart or block diagrams may represent a module, segment, or portion of code, which comprises one or more executable instructions for implementing the specified logical function(s). It should also be noted that, in some alternative implementations, the functions noted in the block may occur out of the order noted in the figures. For example, two blocks shown in succession may, in fact, be executed substantially concurrently, or the blocks may sometimes be executed in the reverse order, depending upon the functionality involved. It will also be noted that each block of the block diagrams and/or flowchart illustration, and combinations of blocks in the block diagrams and/or flowchart illustration, can be implemented by special purpose hardware-based systems which perform the specified functions or acts, or combinations of special purpose hardware and computer instructions.
The functions described above herein may be performed, at least in part, by one or more hardware logic components. For example, without limitation, exemplary types of hardware logic components that may be used include: a Field Programmable Gate Array (FPGA), an Application Specific Integrated Circuit (ASIC), an Application Specific Standard Product (ASSP), a system on a chip (SOC), a Complex Programmable Logic Device (CPLD), and the like.
The foregoing description is only of the preferred embodiments of the present disclosure and description of the principles of the technology being employed. It will be appreciated by those skilled in the art that the scope of the invention in the embodiments of the present disclosure is not limited to the specific combination of the above technical features, but encompasses other technical features formed by any combination of the above technical features or their equivalents without departing from the spirit of the invention. Such as the above-described features, are mutually substituted with (but not limited to) the features having similar functions disclosed in the embodiments of the present disclosure.
Claims (5)
1. An enterprise text information intelligent automatic auditing method based on deep learning comprises the following steps:
Acquiring enterprise text information to be checked, wherein the enterprise text information to be checked comprises a first enterprise identification file image, a second enterprise identification file image and an enterprise text information table;
Inputting the first enterprise identification file image and the second enterprise identification file image into a pre-trained main field identification model to obtain a first image field identification result and a second image field identification result;
Inputting the enterprise text information form into the main field recognition model to obtain a form field recognition result, wherein the form field recognition result comprises an enterprise name;
acquiring enterprise information corresponding to the enterprise name from an associated enterprise information query system as target enterprise information according to the enterprise name;
according to the target enterprise information, checking and checking the first image field identification result, the second image field identification result and the form field identification result to obtain an enterprise text information checking and checking result;
and sending the enterprise text information auditing and checking result to the associated enterprise information auditing terminal so that the enterprise information auditing terminal can select a target enterprise supply end.
2. The method of claim 1, wherein the performing audit verification on the first image field identification result, the second image field identification result, and the form field identification result according to the target enterprise information to obtain an enterprise text information audit verification result comprises:
For each field included in the target enterprise information, performing the following processing steps:
Determining whether identification fields corresponding to the fields exist in all identification fields included in the first image field identification result;
Generating a first field verification result in response to determining that an identification field corresponding to the field exists in all identification fields included in the first image field identification result;
Generating a second field verification result in response to determining that an identification field corresponding to the field does not exist in all identification fields included in the first image field identification result;
and merging the generated first field verification results and the second field verification results into a first image field verification result.
3. An enterprise text information intelligent automatic auditing device based on deep learning, comprising:
The system comprises a first acquisition unit, a second acquisition unit and a third acquisition unit, wherein the first acquisition unit is configured to acquire enterprise text information to be checked, and the enterprise text information to be checked comprises a first enterprise identification file image, a second enterprise identification file image and an enterprise text information table;
The first input unit is configured to input the first enterprise identification file image and the second enterprise identification file image into a pre-trained main field identification model to obtain a first image field identification result and a second image field identification result;
the second input unit is configured to input the enterprise text information form into the main field recognition model to obtain a form field recognition result, wherein the form field recognition result comprises an enterprise name;
A second acquisition unit configured to acquire, from an associated enterprise information inquiry system, enterprise information corresponding to the enterprise name as target enterprise information according to the enterprise name;
the verification unit is configured to carry out verification and verification on the first image field identification result, the second image field identification result and the form field identification result according to the target enterprise information to obtain an enterprise text information verification and verification result;
and the sending unit is configured to send the enterprise text information auditing and checking result to the associated enterprise information auditing terminal so that the enterprise information auditing terminal can select a target enterprise supply end.
4. An electronic device, comprising:
One or more processors;
A storage device having one or more programs stored thereon;
When executed by the one or more processors, causes the one or more processors to implement the method of any of claims 1-2.
5. A computer readable medium having stored thereon a computer program, wherein the computer program, when executed by a processor, implements the method of any of claims 1-2.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311529112.1A CN118097674B (en) | 2023-11-16 | 2023-11-16 | Deep learning-based intelligent automatic auditing method and device for enterprise text information |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311529112.1A CN118097674B (en) | 2023-11-16 | 2023-11-16 | Deep learning-based intelligent automatic auditing method and device for enterprise text information |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN118097674A CN118097674A (en) | 2024-05-28 |
| CN118097674B true CN118097674B (en) | 2024-09-06 |
Family
ID=91158076
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202311529112.1A Active CN118097674B (en) | 2023-11-16 | 2023-11-16 | Deep learning-based intelligent automatic auditing method and device for enterprise text information |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN118097674B (en) |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105787028A (en) * | 2016-02-24 | 2016-07-20 | 北京橙鑫数据科技有限公司 | Business card proofreading method and system |
| CN109492981A (en) * | 2018-09-14 | 2019-03-19 | 龙马智芯（珠海横琴）科技有限公司 | The checking method and device of information |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2008204212A (en) * | 2007-02-21 | 2008-09-04 | Promise Co Ltd | Contract information management system |
-
2023
- 2023-11-16 CN CN202311529112.1A patent/CN118097674B/en active Active
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105787028A (en) * | 2016-02-24 | 2016-07-20 | 北京橙鑫数据科技有限公司 | Business card proofreading method and system |
| CN109492981A (en) * | 2018-09-14 | 2019-03-19 | 龙马智芯（珠海横琴）科技有限公司 | The checking method and device of information |
Also Published As
| Publication number | Publication date |
|---|---|
| CN118097674A (en) | 2024-05-28 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111767371B (en) | Intelligent question-answering method, device, equipment and medium | |
| CN109359194B (en) | Method and apparatus for predicting information categories | |
| US20240233334A1 (en) | Multi-modal data retrieval method and apparatus, medium, and electronic device | |
| CN111198945A (en) | Data processing method, device, medium and electronic equipment | |
| CN110826567A (en) | Optical character recognition method, device, equipment and storage medium | |
| CN113507419B (en) | Training method of traffic distribution model, traffic distribution method and device | |
| CN110798567A (en) | Short message classification display method and device, storage medium and electronic equipment | |
| WO2023142914A1 (en) | Date recognition method and apparatus, readable medium and electronic device | |
| US20220391425A1 (en) | Method and apparatus for processing information | |
| CN116129452A (en) | Method, application method, device, equipment and medium for generating document understanding model | |
| CN117633228A (en) | Model training method and device | |
| CN116645211B (en) | Recommended user information generation method, apparatus, device and computer readable medium | |
| CN116720202B (en) | Service information detection method, device, electronic equipment and computer readable medium | |
| CN118097674B (en) | Deep learning-based intelligent automatic auditing method and device for enterprise text information | |
| CN110689285A (en) | Test method, test device, electronic equipment and computer readable storage medium | |
| CN113807056B (en) | Document name sequence error correction method, device and equipment | |
| CN111914535B (en) | Word recognition method and device, computer equipment and storage medium | |
| CN111338621B (en) | Method, apparatus, electronic device and computer readable medium for displaying data | |
| CN111399902B (en) | Client source file processing method and device, readable medium and electronic equipment | |
| CN112669000A (en) | Government affair item processing method and device, electronic equipment and storage medium | |
| CN116974684B (en) | Map page layout method, map page layout device, electronic equipment and computer readable medium | |
| CN113077353B (en) | Method, device, electronic equipment and medium for generating nuclear insurance conclusion | |
| CN118095426B (en) | Click behavior prediction model training method and device, electronic equipment and readable medium | |
| CN112270170B (en) | Implicit expression statement analysis method and device, medium and electronic equipment | |
| CN111933122B (en) | Speech recognition method, apparatus, electronic device, and computer-readable medium |
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 | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |