CN112699871B - Method, system, device and computer readable storage medium for identifying field content - Google Patents
Method, system, device and computer readable storage medium for identifying field content Download PDFInfo
- Publication number
- CN112699871B CN112699871B CN202011555047.6A CN202011555047A CN112699871B CN 112699871 B CN112699871 B CN 112699871B CN 202011555047 A CN202011555047 A CN 202011555047A CN 112699871 B CN112699871 B CN 112699871B
- Authority
- CN
- China
- Prior art keywords
- content
- field content
- field
- target
- text
- 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
- 238000000034 method Methods 0.000 title claims abstract description 45
- 238000012937 correction Methods 0.000 claims abstract description 32
- 238000004590 computer program Methods 0.000 claims description 10
- 238000000605 extraction Methods 0.000 claims description 2
- 238000012015 optical character recognition Methods 0.000 description 35
- 230000006870 function Effects 0.000 description 5
- 238000010586 diagram Methods 0.000 description 4
- 238000012795 verification Methods 0.000 description 4
- 230000003287 optical effect Effects 0.000 description 3
- 238000012545 processing Methods 0.000 description 3
- 238000004458 analytical method Methods 0.000 description 2
- 238000004891 communication Methods 0.000 description 2
- 238000005516 engineering process Methods 0.000 description 2
- 238000010295 mobile communication Methods 0.000 description 2
- 230000003068 static effect Effects 0.000 description 2
- XLYOFNOQVPJJNP-UHFFFAOYSA-N water Substances O XLYOFNOQVPJJNP-UHFFFAOYSA-N 0.000 description 2
- 230000005540 biological transmission Effects 0.000 description 1
- 238000004364 calculation method Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000010365 information processing Effects 0.000 description 1
- 238000012360 testing method Methods 0.000 description 1
- 238000012549 training Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/20—Image preprocessing
- G06V10/22—Image preprocessing by selection of a specific region containing or referencing a pattern; Locating or processing of specific regions to guide the detection or recognition
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q40/00—Finance; Insurance; Tax strategies; Processing of corporate or income taxes
- G06Q40/03—Credit; Loans; Processing thereof
-
- 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
- G06V30/153—Segmentation of character regions using recognition of characters or words
-
- 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
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Business, Economics & Management (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Evolutionary Biology (AREA)
- General Engineering & Computer Science (AREA)
- Evolutionary Computation (AREA)
- Finance (AREA)
- Bioinformatics & Computational Biology (AREA)
- Multimedia (AREA)
- Artificial Intelligence (AREA)
- Accounting & Taxation (AREA)
- Life Sciences & Earth Sciences (AREA)
- General Business, Economics & Management (AREA)
- Technology Law (AREA)
- Strategic Management (AREA)
- Marketing (AREA)
- Economics (AREA)
- Development Economics (AREA)
- Character Discrimination (AREA)
Abstract
The invention provides a field content identification method, which comprises the steps of obtaining a target image and a service scene corresponding to the target image; invoking a target recognition model based on the service scene to recognize the target image to obtain a plurality of first field contents; according to the field content types of the plurality of first field contents, auxiliary correction is carried out on the plurality of first field contents to obtain target field contents; extracting first digital field content from the first field content when the field content type is digital content type; acquiring a first image associated with a target image; obtaining second digital field content associated with the first digital field content from the first image; and through the association relation between the first digital field content and the second digital field content, the first digital field content is corrected in an auxiliary way, and the target digital field content after auxiliary correction is generated. In the invention, the optimal recognition model (namely the target recognition model) is called according to the service, and the recognition result after recognition is corrected in an auxiliary way, so that the recognition rate of recognizing characters and numbers from the image is effectively improved.
Description
Technical Field
The embodiment of the invention relates to the technical field of image recognition, in particular to a field content recognition method, a field content recognition system, a field content recognition computer device and a field content recognition program.
Background
OCR (Optical Character Recognition ) refers to the process in which an electronic device examines characters printed on paper, determines their shape by detecting dark and light patterns, and then translates the shape into computer text using a character recognition method. The identification card OCR is to take a picture of the identification card through a mobile phone or a terminal device with a camera, and to use OCR character recognition technology to carry out OCR character recognition on the identification card photo, so as to extract the identification card information.
At present, character recognition and error correction are not intelligent and efficient enough, usually an OCR model is combined with a manual auditing process, recognition results are input and audited manually, the error correction accuracy is low, and a large amount of human resources are consumed.
Disclosure of Invention
In view of the above, embodiments of the present invention provide a field content recognition method, system, computer device and computer readable storage medium, which are used for solving the problems of insufficient intelligent and efficient character recognition and error correction and low recognition and error correction accuracy.
The embodiment of the invention solves the technical problems through the following technical scheme:
a field content identification method, comprising:
receiving an identification request sent by a client, wherein the identification request carries a target image and a service scene corresponding to the target image;
invoking a target recognition model in a plurality of candidate recognition models based on the business scenario;
identifying the target image based on the target identification model to obtain a plurality of first field contents;
according to the field content types of the plurality of first field contents, auxiliary correction is carried out on the plurality of first field contents to obtain target field contents; wherein,
when the service scene is a first service scene and the field content type of the first field content is a digital content type, extracting first digital field content from the first field content based on the first service scene;
acquiring a first image associated with the target image;
acquiring second digital field content associated with the first digital field content from the first image;
auxiliary correcting the first digital field content through the association relation between the first digital field content and the second digital field content to generate auxiliary corrected target digital field content; and
And feeding back the target digital field content to the client.
Optionally, the step of calling the target recognition model in the plurality of candidate recognition models based on the service scenario further includes:
determining a plurality of field content areas to be identified based on the service scene;
according to the preset weight of each field content area to be identified and each candidate identification model, calculating the comprehensive weight of each candidate identification model; and
Determining a candidate recognition model with highest comprehensive weight as the target recognition model, and calling the target recognition model; wherein,
when a plurality of target recognition models are determined according to the service scene, acquiring the calling times of the plurality of target recognition models in preset time;
comparing the calling times of the plurality of target recognition models, and determining the target recognition model with the smallest calling times as the target recognition model matched with the service scene.
Optionally, the step of identifying the target image based on the target identification model to obtain a plurality of first field contents further includes:
dividing the target image to obtain a plurality of image blocks;
determining a plurality of field content areas to be identified on a plurality of image blocks based on the format of the target image;
Extracting a plurality of convolution characteristics of a plurality of field content areas to be identified from the target identification model;
inputting the convolution features into a classifier of the target recognition model for recognition;
and generating a plurality of first field contents corresponding to the field content areas to be identified.
Optionally, after the step of obtaining the first digital field content from the plurality of first field contents based on the first service scenario, the method further includes:
judging whether a first designated position in the first digital field content comprises characters or not;
when the first designated position comprises a character, and the character is inconsistent with a preset character, replacing the character with the preset character; the preset character is a preset specific character of the first appointed position; and;
and filling the preset character into the first designated position when the first designated position does not comprise any character.
Optionally, the service scene is a second service scene;
the step of auxiliary correcting the plurality of first field contents to obtain the target field contents according to the field content types of the plurality of first field contents includes:
Acquiring first text field contents from the plurality of first field contents based on the second service scene;
acquiring a second image associated with the target image;
acquiring second text field content associated with the first text field content from the second image;
extracting key text content from the second text field content according to the association relation between the first text field content and the second text field content;
and correcting the first text field content in an auxiliary way through the key text content to obtain the target text field content after auxiliary correction.
Optionally, after the step of obtaining the first text field content from the plurality of first field contents based on the second service scenario, the method further includes:
judging whether a second appointed position in the first text field content comprises text content or not;
when the second appointed position comprises the text content and the text content is inconsistent with the preset text content, replacing the text content with the preset text content; wherein the preset text content is a specific text content associated with the second designated position;
And when the second appointed position does not comprise the text content, filling the preset text content into the second appointed position.
Optionally, the method comprises: and storing the acquired identification request and the corresponding target field content in a blockchain.
In order to achieve the above object, an embodiment of the present invention further provides a field content identification system, including:
the receiving module is used for receiving an identification request sent by a client, wherein the identification request carries a target image and a service scene corresponding to the target image;
the calling module is used for calling a target recognition model in a plurality of candidate recognition models based on the service scene;
the identification module is used for identifying the target image based on the target identification model so as to obtain a plurality of first field contents;
the auxiliary correction module is used for auxiliary correcting the plurality of first field contents according to the field content types of the plurality of first field contents to obtain target field contents; wherein,
the extraction module is used for extracting first digital field content from the first field content based on the service scene when the field content type of the first field content is digital content type;
A first acquisition module for acquiring a first image associated with the target image;
a second acquisition module, configured to acquire, from the first image, second digital field content associated with the first digital field content;
the generation module is used for generating auxiliary corrected target digital field content by auxiliary correcting the first digital field content through the association relation between the first digital field content and the second digital field content;
and the feedback module is used for feeding back the target digital field content to the client.
To achieve the above object, an embodiment of the present invention further provides a computer apparatus including a memory, a processor, and a computer program stored on the memory and executable on the processor, the processor implementing the steps of the field content identification method as described above when executing the computer program.
To achieve the above object, an embodiment of the present invention also provides a computer-readable storage medium having stored therein a computer program executable by at least one processor to cause the at least one processor to perform the steps of the field content identifying method as described above.
According to the field content identification method, the field content identification system, the field content identification computer device and the field content identification computer readable storage medium, the optimal identification model (namely the target identification model) is called according to the service, the auxiliary correction strategy is determined based on a plurality of first field content types and service scenes, and the auxiliary correction is carried out on the identified result through the auxiliary correction strategy, so that characters can be identified and corrected intelligently and efficiently, and the accuracy of identification and correction of the field content in the image is effectively improved.
The invention will now be described in more detail with reference to the drawings and specific examples, which are not intended to limit the invention thereto.
Drawings
FIG. 1 is a flowchart illustrating a field content identification method according to a first embodiment of the present invention;
FIG. 2 is a flowchart illustrating steps for invoking the object recognition model in the field content recognition method according to the first embodiment of the present invention;
FIG. 3 is a flowchart illustrating steps for obtaining a plurality of first field contents in a field content identification method according to a first embodiment of the present invention;
FIG. 4 is a flowchart illustrating steps for verifying specific characters based on a first service scenario in a field content recognition method according to a first embodiment of the present invention;
FIG. 5 is a flowchart illustrating steps of a method for identifying field contents according to a first embodiment of the present invention for assisting in correcting second text field contents based on a second service scenario, thereby obtaining target text field contents after assisting in correction;
FIG. 6 is a flowchart illustrating steps for verifying specific text content based on a second service scenario in a field content recognition method according to a first embodiment of the present invention;
FIG. 7 is a schematic diagram illustrating a field content recognition system according to a second embodiment of the present invention;
fig. 8 is a schematic hardware structure of a computer device according to a third embodiment of the invention.
Detailed Description
The present invention will be described in further detail with reference to the drawings and examples, in order to make the objects, technical solutions and advantages of the present invention more apparent. It should be understood that the specific embodiments described herein are for purposes of illustration only and are not intended to limit the scope of the invention. All other embodiments, which can be made by those skilled in the art based on the embodiments of the invention without making any inventive effort, are intended to be within the scope of the invention.
The technical solutions between the embodiments may be combined with each other, but it is necessary to base the implementation on the basis of those skilled in the art that when the combination of technical solutions contradicts or cannot be implemented, it should be considered that the combination of technical solutions does not exist and is not within the scope of protection claimed by the present invention.
Example 1
Referring to fig. 1, a flowchart illustrating steps of a field content identification method according to an embodiment of the present invention is shown. It will be appreciated that the flow charts in the method embodiments are not intended to limit the order in which the steps are performed. The following description is exemplary with a computer device as an execution subject, and specifically follows:
as shown in fig. 1, the field content identification method may include steps S100 to S900, in which:
step S100, receiving an identification request sent by a client, wherein the identification request carries a target image and a service scene corresponding to the target image.
In an exemplary embodiment, different traffic scenarios correspond to different traffic scenarios. According to the different business scenes selected and input by the client, the input target images are different. The target image comprises, but is not limited to, an identification card front image, an identification card back image, a customer hand-held identification card front image, a customer hand-held identification card back image and other credentials.
Specifically, the service scenarios available for the client to select include, but are not limited to: credit card application service, credit card information change service, credit card running water inquiry service, financial product purchasing service, deposit card application service, deposit card information change service, deposit card running water inquiry service and the like.
Step S200, calling a target recognition model in a plurality of candidate recognition models based on the service scene.
Each candidate recognition model can be a recognition model which is trained and optimized in advance according to the corresponding service scene. Each candidate recognition model may be an OCR (Optical Character Recognition, optical field content recognition) model.
The object recognition model may be invoked from an OCR model intelligent selection center. The OCR model intelligent selection center integrates the technical advantages of each large OCR recognition manufacturer, and comprises a plurality of candidate recognition models.
Further, as shown in fig. 2, the step S200 may include steps S201 to S205, where: step S201, determining a plurality of field content areas to be identified based on the service scene; step S202, calculating the comprehensive weight of each candidate recognition model according to the preset weight of each field content area to be recognized and each candidate recognition model; step S203, determining a candidate recognition model with highest comprehensive weight as the target recognition model, and calling the target recognition model; step S204, when a plurality of target recognition models are determined according to the service scene, the calling times of the plurality of target recognition models in a preset time are obtained; and step S205, comparing the calling times of the plurality of target recognition models, and determining the target recognition model with the smallest calling times as the target recognition model matched with the service scene.
In an exemplary embodiment, different business scenarios have different emphasis on the field content area to be identified in the target image. For example, when the service scene is a credit card sponsoring service and the target image is an identity card front image, acquiring fixed point coordinates corresponding to a plurality of standard fields according to the format of the identity card front image so as to determine a plurality of standard field areas; and determining the areas of the contents of the plurality of fields to be identified according to the standard field areas. Wherein, the name, sex, ethnicity, birth, year, month, day, address, citizen identity number and the like are all standard fields with fixed point coordinates; "XXX" corresponding to the name, "woman" corresponding to the sex, "Chinese" corresponding to the ethnicity, "2000" corresponding to the year, "10" corresponding to the month, "10" corresponding to the day, "Shenzhen Futian area Hua Jiang North street XXX" corresponding to the Guangdong city, and "44XXX" corresponding to the citizen identity number are all field content areas to be identified.
For example, according to the fixed point coordinate of the standard field being the name, adding the preset offset coordinate to the fixed point coordinate of the standard field to obtain a new coordinate, and according to the new coordinate, obtaining the content area of the field to be identified corresponding to the standard field.
The content areas of the fields to be identified of different service scenes and the preset weights of the candidate identification models are obtained by analyzing the advantages of all large OCR manufacturers in specific scenes by combining large data. The preset weights of different field content areas to be identified and each candidate identification model of the same service scene can be the same or different.
Specifically, a plurality of OCR manufacturers are obtained, the advantages of each large manufacturer in a specific scene are analyzed, the preset weights of the content area of each field to be identified and the model of each OCR manufacturer in different service scenes are obtained, and then the comprehensive weight of each candidate recognition model corresponding to the service scene is calculated according to the preset weights, so that the API (Application Program Interface ) of the OCR manufacturer with the highest comprehensive weight in the scene is preferentially called.
The sample identity card information and the sample identity card information recognition result are input into the system, then the sample identity card information is automatically uploaded by a program, the API of each manufacturer is called to recognize the sample identity card information by using the OCR model of each manufacturer, the sample recognition result is output to be compared with the sample identity card information recognition result input manually, the recognition passing rate of each OCR model in different field contents is obtained, the preset weights of each OCR model corresponding to different field contents are obtained according to the recognition passing rate of each OCR model in different field contents, and accordingly the comprehensive weights obtained according to the preset weights of the field content area to be recognized and each OCR model required by a specific scene are calculated to determine the OCR model to be called.
For example, each large OCR recognition manufacturer comprises a combination OCR, a scientific OCR and a large data self-grinding OCR, and the large data OCR is found to have higher recognition rate on the content of each field through training of a large amount of sample data. In the process of actually using OCR recognition, big data OCR is usually taken as the dominant, and call is prioritized.
When the number of the obtained target recognition models corresponding to the service types is multiple, the load of each target recognition model, such as the called times of each target recognition model, needs to be considered. The more times a certain target recognition model is called in a preset time, the larger the load of a server bearing the target recognition model is; therefore, the object recognition model with the smallest number of times the object recognition model is called is determined to be the object recognition model matched with the service scene.
The optimal recognition model is called from the OCR model intelligent selection center according to the service, so that the image recognition of different services can be matched with the optimal recognition model for recognition, and the recognition rate of field content in the image is improved correspondingly for different services.
And step S300, identifying the target image based on the target identification model so as to obtain a plurality of first field contents.
For example, the target image may be integrally identified to obtain the plurality of first field contents.
In an exemplary embodiment, referring to fig. 3, the step S300 may include: step S301, dividing the target image to obtain a plurality of image blocks; step S302, determining a plurality of field content areas to be identified on a plurality of image blocks based on the layout of the target image; step S303, extracting a plurality of convolution characteristics of a plurality of field content areas to be identified from the target identification model; step S304, inputting the convolution characteristics into a classifier of the target recognition model for recognition; and step S305, generating a plurality of first field contents corresponding to the plurality of field content areas to be identified. The method for identifying the target image by dividing the image blocks can identify a plurality of image blocks at the same time, so that the time consumption is effectively shortened, and the identification efficiency is greatly improved.
Step S400, according to the field content types of the plurality of first field contents, the plurality of first field contents are corrected in an auxiliary manner to obtain target field contents.
Specifically, the field content types of the plurality of first field contents include text field contents and digital field contents.
Step S500, when the service scenario is a first service scenario and the field content type of the first field content is a digital content type, extracting a first digital field content from the first field content based on the first service scenario.
Step S600, acquiring a first image associated with the target image.
Step S700, obtaining second digital field content associated with the first digital field content from the first image.
Step S800, performing auxiliary correction on the first digital field content through the association relationship between the first digital field content and the second digital field content, to obtain the auxiliary corrected target digital field content.
And step S900, feeding back the target digital field content to the client.
Different field content types and service scenes correspond to different auxiliary correction strategies. For example:
(1) The service scenario may be a first service scenario, and the field content type of the first field content may be a first digital field content.
For example, the target image is an identification card reverse image, the first image associated with the target image is an identification card reverse image, the first digital field content corresponds to the digital field content of the standard field "validity period", and the second digital field content corresponds to the digital field content of the "citizen identification number". The association relation between the first digital field content and the second digital field content is the association relation between the citizen identity number and the validity period in the identity card rule. And when the first digital field content is corrected in an auxiliary way, respectively taking the first 8-bit digital field content and the last 8-bit digital field content of the digital field content corresponding to the effective period, and comparing the last 4 bits of the first 8-bit digital field content with the last 8 bits of the digital field content to obtain the last 4 bits of the digital field content. And correcting the first 4 bits of the first 8-bit digital field content and the last 8 bits of the digital field content based on the second digital field content corresponding to the year and the current year. In the analysis process of the large data error recognition case, the fact that the proportion of the field content recognition errors corresponding to the valid period standard field to the various field content recognition error types of the similar images is 75% is found, so that the field content corresponding to the valid period standard field is corrected in an auxiliary mode, and 75% of recognition failures can be effectively recovered; the recognition rate of recognizing numbers from the identity card images is effectively improved.
In an exemplary embodiment, referring to fig. 4, after the step S500, steps S5011 to S5013 may be further included, where: step S5011, judging whether a first designated position in the first digital field content comprises characters or not; step S5012, when the first designated position comprises a character, and the character is inconsistent with a preset character, replacing the character with the preset character; the preset character is a preset specific character of the first appointed position; and step S5013, when the first designated position does not comprise any character, filling the preset character into the first designated position.
The auxiliary correction operation is exemplarily described below by using the target image as an image on the reverse side of the identification card. Traversing the first digital field content, and positioning a first designated position in the first digital field content, for example, the first designated position is a middle position in a field content area to be identified corresponding to a standard field of 'validity period'; judging whether the first designated position contains characters or not, if the first designated position contains characters, further comparing the characters of the first designated position with preset specific characters of the first designated position, if the comparison result is consistent, indicating that verification is passed, and if the comparison result is inconsistent, replacing the characters with the preset specific characters. It will be understood that if the predetermined specific character is "-", when the first specified position is identified as including the character, the identified character may be "-", "\" or "| -! When the first appointed position is equal to the preset specific character, the characters at the first appointed position are compared, and if the comparison is consistent, the verification is passed; if the comparison is inconsistent, replacing the character identified by the first designated position with the preset specific character '-' so as to carry out auxiliary correction on the character of the first designated position.
(2) The service scenario may be a second service scenario, and the field content type of the first field content may be a first text field content.
Referring to fig. 5, the step S500 may include steps S511 to S515, wherein: step S511, based on the second service scene, acquiring first text field contents from the plurality of first field contents; step S512, acquiring a second image associated with the target image; step S513, obtaining second text field content associated with the first text field content from the second image; step S514, extracting key text content from the second text field content according to the association relation between the first text field content and the second text field content; step S515, the first text field content is corrected in an auxiliary manner by the key text content, so as to obtain the corrected target text field content.
For example, the target image is an identification card reverse image, the second image associated with the target image is an identification card reverse image, the first text field content is text field content corresponding to the "issuing authority" standard field, and the second digital field content is text field content corresponding to the "address" standard field. The association relation between the first text field content and the second text field content is the association relation between an issuing authority and an address in the identity card rule. From the front image of the identification card, it can be known that the key text content includes, but is not limited to, province, autonomous region (inner mongolia, guangxi, ningxia, xinjiang, tibet), city, autonomous state, county, district, etc.
Illustratively, the step S515 may include the following:
when the content of the text field of the key contains county, the standard text combination of the issuing authority is obtained as county name and public security bureau, so as to assist in correcting the content of the text field corresponding to the issuing authority.
When the key word field content does not contain county and district and contains city, the standard word combination of the issuing authority is obtained as city name and public security bureau so as to assist in correcting the word field content corresponding to the issuing authority.
When the key text field content contains 'city' and 'district/new district', the standard text combination of the issuing authority is acquired as 'city+public security bureau+district (minus district word) +branch office', so as to assist in correcting the text field content corresponding to the issuing authority. For example, the "zone" in the address is Pudong zone, and the text field content of the issuing authority is Pudong division in Shanghai city public Annu office.
When the key text field content contains 'city' and 'development area', the standard text combination of the issuing authority is acquired as 'city+public security office+area (area word is removed) +branch office', so as to assist in correcting the text field content corresponding to the issuing authority.
Specifically, acquiring key text field content, extracting the first several digits of text field content which is public security from text field content of an issuing authority as text field content to be verified, comparing the key text field content with the text field content to be verified, and if the comparison result is consistent, indicating that verification is passed; and if the comparison results are inconsistent, replacing the text field content to be verified by a plurality of first bits of the district, the county or the city in the key text field content to generate target text field content.
In an exemplary embodiment, as shown in fig. 6, after the step of obtaining the first text field content from the plurality of first field contents based on the second service scenario, steps S5101 to S5103 are further included, where: step S5101, judging whether the second designated location in the first text field content includes text content; step S5102, when the second designated location includes the text content and the text content is inconsistent with the preset text content, replacing the text content with the preset text content; wherein the preset text content is a specific text content associated with the second designated position; step S5103, when the second designated location does not include text content, filling the preset text content into the second designated location.
And continuing to exemplarily illustrate the auxiliary correction operation by taking the target image as the reverse image of the identity card. Traversing the first text field content and locating a second designated position in the first text field content; judging whether the second appointed position contains text contents or not, and if the second appointed position contains characters, further comparing the text contents of the second appointed position with specific text contents associated with the second appointed position to generate a comparison result; if the comparison results are consistent, the verification is passed; and if the comparison results are inconsistent, replacing the text content with the specific text content. For example, the second designated position in the first text field content is a plurality of end positions in the field content area to be identified corresponding to the standard field of the "issuing authority". The specific text associated with the second designated location is "public security office". Illustratively, when the comparison results are inconsistent, the office is complemented in the latter position of the public security. In the analysis process of the large data error recognition case, the content recognition errors corresponding to the standard field of the issuing authority are found to be 95.31 percent in the types of the content recognition errors of the various fields of the similar images, so that the content of the field corresponding to the standard field of the issuing authority is corrected in an auxiliary way, and 95.31 percent of recognition failures can be effectively recovered; the recognition rate of recognizing characters from the identity card image is effectively improved.
The method comprises the following steps: and storing the acquired identification request and the corresponding target field content in a blockchain.
The Blockchain (Blockchain) is essentially a decentralised database, and is a series of data blocks which are generated by correlation with a cryptography method, and each data block contains information of a batch of network transactions and is used for verifying the validity (anti-counterfeiting) of the information and generating a next block. The blockchain may include a blockchain underlying platform, a platform product services layer, an application services layer, and the like.
Under the condition that the OCR technology is relatively closed, the embodiment of the invention can be understood as that in the OCR models provided by a plurality of OCR manufacturers, the optimal selection of the OCR models is carried out to obtain the target recognition model; and the auxiliary correction is carried out on the recognized result through the selected auxiliary correction strategy, so that the characters can be recognized and corrected intelligently and efficiently, the recognition rate of field contents in OCR images is effectively improved, and the recognition and correction accuracy is effectively improved; in the practical test, the OCR recognition rate is improved from 80% to 97%, and the user experience is improved.
Example two
With continued reference to fig. 7, a schematic diagram of the program modules of the field content identification system of the present invention is shown. In this embodiment, the field content identifying system 20 may include or be divided into one or more program modules, which are stored in a storage medium and executed by one or more processors to accomplish the present invention and may implement the field content identifying method described above. Program modules in accordance with embodiments of the present invention are directed to a series of computer program instruction segments capable of performing the specified functions and are more suitable than the programs themselves for describing the execution of the field content identification system 20 in a storage medium. The following description will specifically describe functions of each program module of the present embodiment:
The receiving module 600 is configured to receive an identification request sent by a client, where the identification request carries a target image and a service scenario corresponding to the target image.
And a calling module 610, configured to call a target recognition model in the multiple candidate recognition models based on the service scenario.
The recognition module 620 is configured to recognize the target image based on the target recognition model, so as to obtain a plurality of first field contents.
An auxiliary correction module 630, configured to, according to the field content types of the plurality of first field contents, auxiliary correct the plurality of first field contents to obtain a target field content; wherein,
an extracting module 640, configured to extract, when the field content type of the first field content is a digital content type, a first digital field content from the first field content based on the service scenario;
a first acquisition module 650 for acquiring a first image associated with the target image;
a second obtaining module 660, configured to obtain, from the first image, a second digital field content associated with the first digital field content;
and the generating module 670 is configured to assist in correcting the first digital field content by using the association relationship between the first digital field content and the second digital field content, and generate an object digital field content after the assist correction.
And a feedback module 680, configured to feed back the target digital field content to the client.
In an exemplary embodiment, the calling module 610 is further configured to: determining a plurality of field content areas to be identified based on the service scene; according to the preset weight of each field content area to be identified and each candidate identification model, calculating the comprehensive weight of each candidate identification model; determining a candidate recognition model with highest comprehensive weight as the target recognition model, and calling the target recognition model; when a plurality of target recognition models are determined according to the service scene, acquiring the calling times of the plurality of target recognition models in preset time; comparing the calling times of the plurality of target recognition models, and determining the target recognition model with the smallest calling times as the target recognition model matched with the service scene.
In an exemplary embodiment, the identification module 620 is further configured to: dividing the target image to obtain a plurality of image blocks; determining a plurality of field content areas to be identified on a plurality of image blocks based on the format of the target image; extracting a plurality of convolution characteristics of a plurality of field content areas to be identified from the target identification model; inputting the convolution features into a classifier of the target recognition model for recognition; and generating a plurality of first field contents corresponding to the field content areas to be identified.
In an exemplary embodiment, the auxiliary correction module 630 is further configured to: judging whether a first designated position in the first digital field content comprises characters or not; when the first designated position comprises a character, and the character is inconsistent with a preset character, replacing the character with the preset character; the preset character is a preset specific character of the first appointed position; and; and filling the preset character into the first designated position when the first designated position does not comprise any character.
In an exemplary embodiment, the traffic scenario is a second traffic scenario. The auxiliary correction module 630 is further configured to: acquiring first text field contents from the plurality of first field contents based on the second service scene; acquiring a second image associated with the target image; acquiring second text field content associated with the first text field content from the second image; extracting key text contents from the second text field contents according to the association relation between the first text field contents and the second text field contents; and correcting the first text field content in an auxiliary way through the key text content to obtain the target text field content after auxiliary correction.
In an exemplary embodiment, the auxiliary correction module 630 is further configured to: judging whether a second appointed position in the first text field content comprises text content or not; when the second appointed position comprises the text content and the text content is inconsistent with the preset text content, replacing the text content with the preset text content; wherein the preset text content is a specific text content associated with the second designated position; and when the preset specific position does not contain text contents, filling the preset specific text contents of the specific position into the first text field contents.
Example III
Referring to fig. 8, a hardware architecture diagram of a computer device according to a third embodiment of the present invention is shown. In this embodiment, the computer device 2 is a device capable of automatically performing numerical calculation and/or information processing according to a preset or stored instruction. The computer device 2 may be a rack server, a blade server, a tower server, or a rack server (including a stand-alone server, or a server cluster made up of multiple servers), or the like. As shown in fig. 8, the computer device 2 includes, but is not limited to, at least a memory 21, a processor 22, a network interface 23, and a field content identification system 20, which are communicatively connected to each other via a system bus. Wherein:
In this embodiment, the memory 21 includes at least one type of computer-readable storage medium including flash memory, a hard disk, a multimedia card, a card memory (e.g., SD or DX memory, etc.), a Random Access Memory (RAM), a Static Random Access Memory (SRAM), a read-only memory (ROM), an electrically erasable programmable read-only memory (EEPROM), a programmable read-only memory (PROM), a magnetic memory, a magnetic disk, an optical disk, and the like. In some embodiments, the memory 21 may be an internal storage unit of the computer device 2, such as a hard disk or a memory of the computer device 2. In other embodiments, the memory 21 may also be an external storage device of the computer device 2, such as a plug-in hard disk, a Smart Media Card (SMC), a Secure Digital (SD) Card, a Flash memory Card (Flash Card) or the like, which are provided on the computer device 2. Of course, the memory 21 may also include both internal storage units of the computer device 2 and external storage devices. In the present embodiment, the memory 21 is typically used to store an operating system and various types of application software installed on the computer device 2, such as program codes of the field content identifying system 20 of the above embodiment. Further, the memory 21 may be used to temporarily store various types of data that have been output or are to be output.
The processor 22 may be a central processing unit (Central Processing Unit, CPU), controller, microcontroller, microprocessor, or other data processing chip in some embodiments. The processor 22 is typically used to control the overall operation of the computer device 2. In this embodiment, the processor 22 is configured to execute the program code stored in the memory 21 or process data, for example, execute the field content identifying system 20, to implement the field content identifying method of the above embodiment.
The network interface 23 may comprise a wireless network interface or a wired network interface, which network interface 23 is typically used for establishing a communication connection between the computer apparatus 2 and other electronic devices. For example, the network interface 23 is used to connect the computer device 2 to an external terminal through a network, establish a data transmission channel and a communication connection between the computer device 2 and the external terminal, and the like. The network may be an Intranet (Intranet), the Internet (Internet), a global system for mobile communications (Global System of Mobile communication, GSM), wideband code division multiple access (Wideband Code Division Multiple Access, WCDMA), a 4G network, a 5G network, bluetooth (Bluetooth), wi-Fi, or other wireless or wired network.
It is noted that fig. 8 only shows a computer device 2 having components 20-23, but it is understood that not all of the illustrated components are required to be implemented, and that more or fewer components may alternatively be implemented.
In this embodiment, the field content identifying system 20 stored in the memory 21 may also be divided into one or more program modules, which are stored in the memory 21 and executed by one or more processors (the processor 22 in this embodiment) to complete the present invention.
For example, fig. 7 shows a schematic diagram of a program module for implementing the second embodiment of the field content recognition system 20, where the field content based recognition system 20 may be divided into a receiving module 600, a calling module 610, a recognition module 620, a selecting module 630, an auxiliary correction module 640, and a feedback module 650. Program modules in the present invention are understood to mean a series of computer program instruction segments capable of performing a specific function, more appropriately than a program, for describing the execution of the field content identifying system 20 in the computer device 2. The specific functions of the program modules 600-650 are described in detail in the second embodiment, and are not described herein.
Example IV
The present embodiment also provides a computer-readable storage medium such as a flash memory, a hard disk, a multimedia card, a card-type memory (e.g., SD or DX memory, etc.), a Random Access Memory (RAM), a Static Random Access Memory (SRAM), a read-only memory (ROM), an electrically erasable programmable read-only memory (EEPROM), a programmable read-only memory (PROM), a magnetic memory, a magnetic disk, an optical disk, a server, an App application store, etc., on which a computer program is stored, which when executed by a processor, performs the corresponding functions. The computer readable storage medium of the present embodiment is used to store the field content identifying system 20, and when executed by a processor, implements the field content identifying method of the above embodiment.
The foregoing embodiment numbers of the present invention are merely for the purpose of description, and do not represent the advantages or disadvantages of the embodiments.
From the above description of the embodiments, it will be clear to those skilled in the art that the above-described embodiment method may be implemented by means of software plus a necessary general hardware platform, but of course may also be implemented by means of hardware, but in many cases the former is a preferred embodiment.
The foregoing description is only of the preferred embodiments of the present invention, and is not intended to limit the scope of the invention, but rather is intended to cover any equivalents of the structures or equivalent processes disclosed herein or in the alternative, which may be employed directly or indirectly in other related arts.
Claims (9)
1. A method for identifying field content, comprising:
receiving an identification request sent by a client, wherein the identification request carries a target image and a service scene corresponding to the target image;
invoking a target recognition model in a plurality of candidate recognition models based on the business scenario;
identifying the target image based on the target identification model to obtain a plurality of first field contents;
according to the field content types of the plurality of first field contents, auxiliary correction is carried out on the plurality of first field contents to obtain target field contents; wherein,
when the service scene is a first service scene and the field content type of the first field content is a digital content type, extracting first digital field content from the first field content based on the first service scene;
acquiring a first image associated with the target image;
acquiring second digital field content associated with the first digital field content from the first image;
auxiliary correcting the first digital field content through the association relation between the first digital field content and the second digital field content to generate auxiliary corrected target digital field content; and
Feeding back the target digital field content to the client;
the step of calling the target recognition model in the candidate recognition models based on the service scene further comprises the following steps:
determining a plurality of field content areas to be identified based on the service scene;
according to the preset weight of each field content area to be identified and each candidate identification model, calculating the comprehensive weight of each candidate identification model; and
Determining a candidate recognition model with highest comprehensive weight as the target recognition model, and calling the target recognition model; wherein,
when a plurality of target recognition models are determined according to the service scene, acquiring the calling times of the plurality of target recognition models in preset time;
comparing the calling times of the plurality of target recognition models, and determining the target recognition model with the smallest calling times as the target recognition model matched with the service scene.
2. The field content identifying method according to claim 1, wherein the step of identifying the target image based on the target identification model to obtain a plurality of first field contents further comprises:
dividing the target image to obtain a plurality of image blocks;
Determining a plurality of field content areas to be identified on a plurality of image blocks based on the format of the target image;
extracting a plurality of convolution characteristics of a plurality of field content areas to be identified from the target identification model;
inputting the convolution features into a classifier of the target recognition model for recognition;
and generating a plurality of first field contents corresponding to the field content areas to be identified.
3. The method for identifying field contents according to claim 2, wherein after the step of obtaining the first digital field contents from the plurality of first field contents based on the first service scenario, further comprising:
judging whether a first designated position in the first digital field content comprises characters or not;
when the first designated position comprises a character, and the character is inconsistent with a preset character, replacing the character with the preset character; the preset character is a preset specific character of the first appointed position; and
And filling the preset character into the first designated position when the first designated position does not comprise any character.
4. The field content identifying method according to claim 2, wherein the service scenario is a second service scenario;
The step of auxiliary correcting the plurality of first field contents to obtain the target field contents according to the field content types of the plurality of first field contents includes:
acquiring first text field contents from the plurality of first field contents based on the second service scene;
acquiring a second image associated with the target image;
acquiring second text field content associated with the first text field content from the second image;
extracting key text content from the second text field content according to the association relation between the first text field content and the second text field content;
and correcting the first text field content in an auxiliary way through the key text content to obtain the target text field content after auxiliary correction.
5. The method for identifying field contents according to claim 4, wherein after the step of obtaining the first text field contents from the plurality of first field contents based on the second service scenario, further comprising:
judging whether a second appointed position in the first text field content comprises text content or not;
when the second appointed position comprises the text content and the text content is inconsistent with the preset text content, replacing the text content with the preset text content; the preset text content is a specific text content associated with the second appointed position;
And when the second appointed position does not comprise the text content, filling the preset text content into the second appointed position.
6. A field content identifying method according to claim 1, characterized in that the method comprises: and storing the acquired identification request and the corresponding target field content in a blockchain.
7. A field content identification system, comprising:
the receiving module is used for receiving an identification request sent by a client, wherein the identification request carries a target image and a service scene corresponding to the target image;
the calling module is used for calling a target recognition model in a plurality of candidate recognition models based on the service scene;
the identification module is used for identifying the target image based on the target identification model so as to obtain a plurality of first field contents;
the auxiliary correction module is used for auxiliary correcting the plurality of first field contents according to the field content types of the plurality of first field contents to obtain target field contents; wherein,
the extraction module is used for extracting first digital field content from the first field content based on the service scene when the field content type of the first field content is digital content type;
A first acquisition module for acquiring a first image associated with the target image;
a second acquisition module, configured to acquire, from the first image, second digital field content associated with the first digital field content;
the generation module is used for generating auxiliary corrected target digital field content by auxiliary correcting the first digital field content through the association relation between the first digital field content and the second digital field content; the feedback module is used for feeding back the target digital field content to the client;
wherein, based on the service scenario, invoking a target recognition model in a plurality of candidate recognition models, further comprising:
determining a plurality of field content areas to be identified based on the service scene;
according to the preset weight of each field content area to be identified and each candidate identification model, calculating the comprehensive weight of each candidate identification model; and
Determining a candidate recognition model with highest comprehensive weight as the target recognition model, and calling the target recognition model; wherein,
when a plurality of target recognition models are determined according to the service scene, acquiring the calling times of the plurality of target recognition models in preset time;
Comparing the calling times of the plurality of target recognition models, and determining the target recognition model with the smallest calling times as the target recognition model matched with the service scene.
8. 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 processor implements the steps of the field content identification method according to any of claims 1 to 6 when the computer program is executed.
9. A computer-readable storage medium, in which a computer program is stored, the computer program being executable by at least one processor to cause the at least one processor to perform the steps of the field content identification method according to any one of claims 1 to 6.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011555047.6A CN112699871B (en) | 2020-12-23 | 2020-12-23 | Method, system, device and computer readable storage medium for identifying field content |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011555047.6A CN112699871B (en) | 2020-12-23 | 2020-12-23 | Method, system, device and computer readable storage medium for identifying field content |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN112699871A CN112699871A (en) | 2021-04-23 |
| CN112699871B true CN112699871B (en) | 2023-11-14 |
Family
ID=75510051
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202011555047.6A Active CN112699871B (en) | 2020-12-23 | 2020-12-23 | Method, system, device and computer readable storage medium for identifying field content |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN112699871B (en) |
Families Citing this family (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114339375B (en) * | 2021-08-17 | 2024-04-02 | 腾讯科技（深圳）有限公司 | Video playing method, method for generating video catalogue and related products |
| CN113435993A (en) * | 2021-08-27 | 2021-09-24 | 聆笙(北京)科技有限公司 | Receipt data recognition system and method thereof |
| CN116702024B (en) * | 2023-05-16 | 2024-05-28 | 见知数据科技(上海)有限公司 | Method, device, computer equipment and storage medium for identifying type of stream data |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JPH076200A (en) * | 1992-04-20 | 1995-01-10 | Nec Corp | Optical character reader |
| WO2019071660A1 (en) * | 2017-10-09 | 2019-04-18 | 平安科技（深圳）有限公司 | Bill information identification method, electronic device, and readable storage medium |
| WO2019071662A1 (en) * | 2017-10-09 | 2019-04-18 | 平安科技（深圳）有限公司 | Electronic device, bill information identification method, and computer readable storage medium |
| WO2020010547A1 (en) * | 2018-07-11 | 2020-01-16 | 深圳前海达闼云端智能科技有限公司 | Character identification method and apparatus, and storage medium and electronic device |
| CN111461108A (en) * | 2020-02-21 | 2020-07-28 | 浙江工业大学 | Medical document identification method |
| CN111832382A (en) * | 2019-04-15 | 2020-10-27 | 通用电气公司 | Optical character recognition error correction based on visual and textual content |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US10360703B2 (en) * | 2017-01-13 | 2019-07-23 | International Business Machines Corporation | Automatic data extraction from a digital image |
-
2020
- 2020-12-23 CN CN202011555047.6A patent/CN112699871B/en active Active
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JPH076200A (en) * | 1992-04-20 | 1995-01-10 | Nec Corp | Optical character reader |
| WO2019071660A1 (en) * | 2017-10-09 | 2019-04-18 | 平安科技（深圳）有限公司 | Bill information identification method, electronic device, and readable storage medium |
| WO2019071662A1 (en) * | 2017-10-09 | 2019-04-18 | 平安科技（深圳）有限公司 | Electronic device, bill information identification method, and computer readable storage medium |
| WO2020010547A1 (en) * | 2018-07-11 | 2020-01-16 | 深圳前海达闼云端智能科技有限公司 | Character identification method and apparatus, and storage medium and electronic device |
| CN111832382A (en) * | 2019-04-15 | 2020-10-27 | 通用电气公司 | Optical character recognition error correction based on visual and textual content |
| CN111461108A (en) * | 2020-02-21 | 2020-07-28 | 浙江工业大学 | Medical document identification method |
Also Published As
| Publication number | Publication date |
|---|---|
| CN112699871A (en) | 2021-04-23 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110349038B (en) | Risk assessment model training method and risk assessment method | |
| CN112699871B (en) | Method, system, device and computer readable storage medium for identifying field content | |
| CN110502608B (en) | Man-machine conversation method and man-machine conversation device based on knowledge graph | |
| CN109389723B (en) | Visitor management method and device using face recognition and computer equipment | |
| WO2020098250A1 (en) | Character recognition method, server, and computer readable storage medium | |
| CN111695439A (en) | Image structured data extraction method, electronic device and storage medium | |
| CN110751041A (en) | Certificate authenticity verification method, system, computer equipment and readable storage medium | |
| CN109766072B (en) | Information verification input method and device, computer equipment and storage medium | |
| EP4109332A1 (en) | Certificate authenticity identification method and apparatus, computer-readable medium, and electronic device | |
| CN112650875A (en) | House image verification method and device, computer equipment and storage medium | |
| CN112560964A (en) | Method and system for training Chinese herbal medicine pest and disease identification model based on semi-supervised learning | |
| CN110363222B (en) | Picture labeling method and device for model training, computer equipment and storage medium | |
| CN111222517A (en) | Test sample generation method, system, computer device and storage medium | |
| CN111178147A (en) | Screen crushing and grading method, device, equipment and computer readable storage medium | |
| CN111444226A (en) | Method and system for pushing service reservation network point data | |
| CN112668640A (en) | Text image quality evaluation method, device, equipment and medium | |
| CN112668575B (en) | Key information extraction method and device, electronic equipment and storage medium | |
| CN110992155A (en) | Bidding and enclosing processing method and related product | |
| CN114817340A (en) | Data tracing method and device, computer equipment and storage medium | |
| CN110084467B (en) | Mobile label verification method, mobile label verification device, computer equipment and storage medium | |
| CN111062301A (en) | Identity authentication method and device, electronic equipment and computer readable storage medium | |
| CN110909733A (en) | Template positioning method and device based on OCR picture recognition and computer equipment | |
| CN113642642B (en) | Control identification method and device | |
| CN116681045A (en) | Report generation method, report generation device, computer equipment and storage medium | |
| CN113743129B (en) | Information pushing method, system, equipment and medium based on neural network |
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 |