CN110263694A - A kind of bank slip recognition method and device - Google Patents
A kind of bank slip recognition method and device Download PDFInfo
- Publication number
- CN110263694A CN110263694A CN201910510860.2A CN201910510860A CN110263694A CN 110263694 A CN110263694 A CN 110263694A CN 201910510860 A CN201910510860 A CN 201910510860A CN 110263694 A CN110263694 A CN 110263694A
- Authority
- CN
- China
- Prior art keywords
- bill
- identified
- image
- images
- item
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Classifications
-
- 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
- G06Q40/12—Accounting
- G06Q40/125—Finance or payroll
-
- 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
- G06V30/40—Document-oriented image-based pattern recognition
- G06V30/41—Analysis of document content
- G06V30/418—Document matching, e.g. of document images
Landscapes
- Engineering & Computer Science (AREA)
- Business, Economics & Management (AREA)
- Accounting & Taxation (AREA)
- Finance (AREA)
- Physics & Mathematics (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Development Economics (AREA)
- Technology Law (AREA)
- General Business, Economics & Management (AREA)
- Strategic Management (AREA)
- Marketing (AREA)
- Artificial Intelligence (AREA)
- Economics (AREA)
- Multimedia (AREA)
- Image Analysis (AREA)
Abstract
This application discloses a kind of bank slip recognition method and devices, the bank slip recognition method orients the note position of bill to be identified from images to be recognized first and identifies the bill type of the bill to be identified, determine the image of bill to be identified, then it is corrected according to image of the default ticket templates corresponding with the bill type to the bill to be identified, obtain correction bill images, the item to be identified in the correction bill images is extracted, the content information of the item to be identified is identified finally by character recognition technology.No matter on the bank slip recognition method and device scanned copy image, there are how many bill images, it can be based on deep learning algorithm and preset ticket templates and parameter identify each of scanned copy bill images respectively, recognition efficiency is high and accurate, can greatly promote the efficiency of finance reimbursement process work.
Description
Technical field
The present invention relates to image recognition technologys, and more specifically, it relates to a kind of bank slip recognition method and devices.
Background technique
Finance reimbursement is the financial process that each Finance Department, incorporated business is frequently necessary to away, this needs financial staff in the process
The bill (for example invoice, receipt) submitted to application reimbursement personnel is audited.In addition, application reimbursement personnel can will need to submit an expense account
Bill it is irregular paste on blank sheet of paper, electronic image is then scanned into, for staying shelves.Since traditional financial staff audits
The features such as there are human inputs for the working method of bill greatly, low efficiency, has developed the phase of some electronic recognition invoices in recent years
Pass technology.
Bill images are done two in the process by the thought for generally using template to the electronic recognition algorithm of bill scanned copy at present
Value pretreatment excludes the pixel interference outside bill paper to facilitate contour detecting, finally carries out text pretreatment and text
Identification.However, the method for above-mentioned electronic recognition bill is usually applicable in just for the image scanning part for only including a bill, to reality
The scene of multiple bills is posted on scanned copy in the work of border and is not suitable for.
Summary of the invention
In view of this, the present invention provides one kind, with overcome the problems, such as in the prior art due to.
To achieve the above object, the invention provides the following technical scheme:
A kind of bank slip recognition method, comprising:
The note position of bill to be identified is oriented from images to be recognized and identifies the bill of the bill to be identified
Type determines the image of bill to be identified；
It is corrected, obtains according to image of the default ticket templates corresponding with the bill type to the bill to be identified
To correction bill images；
Extract the item to be identified in the correction bill images；
The content information of the item to be identified is identified by character recognition technology.
Optionally, it the note position that bill to be identified is oriented from images to be recognized and identifies described to be identified
The bill type of bill determines the image of bill to be identified, comprising:
Document identification technology based on deep learning determines described to be identified in conjunction with bank slip recognition model trained in advance
The location information and bill type of bill to be identified in image, and determine based on the location information image of bill to be identified.
Optionally, it is described according to default ticket templates corresponding with the bill type to the image of the bill to be identified
It is corrected, obtains correction bill images, comprising:
According to default ticket templates corresponding with the bill type, by images match and affine transformation technology to described
The image of bill to be identified is corrected, and obtains correction bill images.
Optionally, the basis default ticket templates corresponding with the bill type, pass through images match and affine change
It changes technology to be corrected the image of the bill to be identified, obtains correction bill images, comprising:
Using SIFT and RANSAC in conjunction with image matching algorithm determined from the image of the bill to be identified in advance
The most like matching area of the ticket templates characteristic image first stored；
The affine transformation matrix for determining the matching area relative to the ticket templates characteristic image is calculated, and according to institute
The correction that affine transformation matrix is rotated, scaled and/or is displaced to the image of the bill to be identified is stated, correction bill is obtained
Image.
Optionally, the item to be identified extracted in the correction bill images, comprising:
According to position coordinates of the identification item predetermined in default ticket templates, in the image of the bill to be identified
Upper extraction identifies item of image, obtains tentatively identifying item；
Line of text detection algorithm based on deep learning positions literal line position in the preliminary identification item, obtains target
Item to be identified.
Optionally, the note position that bill to be identified is oriented from images to be recognized, comprising:
Document identification technology based on deep learning inputs images to be recognized using full convolutional neural networks, described in output
The prediction probability figure of each pixel class in images to be recognized；
Based on the prediction probability figure of each pixel class, the class of each pixel is calculated by OTSU image segmentation algorithm
Other attribute finally extracts foreground pixel profile and obtains note position information.
Optionally, after the content information for identifying the item to be identified by character recognition technology described, the method is also
Include:
According to the Property Name of the item to be identified and the content information identified, to the content of the bill to be identified
Carry out data structured output.
A kind of bank slip recognition device, comprising:
Location type determining module, for oriented from images to be recognized note position to be identified and identify it is described to
The bill type for identifying bill, determines the image of bill to be identified；
Image correction module is used for according to default ticket templates corresponding with the bill type to the bill to be identified
Image be corrected, obtain correction bill images；
Item extraction module is identified, for extracting the item to be identified in the correction bill images；
Content identifier module, for identifying the content information of the item to be identified by character recognition technology.
A kind of computer readable storage medium, is stored thereon with computer program, realization when which is executed by processor
Bank slip recognition method described in any one of the above.
A kind of electronic equipment, comprising:
Processor；And
Memory, for storing the executable instruction of the processor；
Wherein, the processor is configured to execute any one of the above bank slip recognition via the executable instruction is executed
Method.
It can be seen via above technical scheme that compared with prior art, the embodiment of the invention discloses a kind of bank slip recognitions
Method and device, the bank slip recognition method orient note position and the identification of bill to be identified from images to be recognized first
The bill type of the bill to be identified out, determines the image of bill to be identified, then according to corresponding with the bill type
Default ticket templates are corrected the image of the bill to be identified, obtain correction bill images, extract the correction bill
Item to be identified in image identifies the content information of the item to be identified finally by character recognition technology.The bank slip recognition side
No matter there are how many bill images on method and device scanned copy image, deep learning algorithm and preset bill mould can be based on
Plate and parameter identify that recognition efficiency is high and accurate to each of scanned copy bill images respectively, can greatly promote
The efficiency of finance reimbursement process work.
Detailed description of the invention
In order to more clearly explain the embodiment of the invention or the technical proposal in the existing technology, to embodiment or will show below
There is attached drawing needed in technical description to be briefly described, it should be apparent that, the accompanying drawings in the following description is only this
The embodiment of invention for those of ordinary skill in the art without creative efforts, can also basis
The attached drawing of offer obtains other attached drawings.
Fig. 1 is a kind of flow chart of bank slip recognition method disclosed by the embodiments of the present invention；
Fig. 2 is image rectification flow diagram disclosed by the embodiments of the present invention；
Fig. 3 is the flow chart of item to be identified in extraction bill images disclosed by the embodiments of the present invention；
Fig. 4 is the flow chart of another kind bank slip recognition method disclosed in inventive embodiments；
Fig. 5 is a kind of structural schematic diagram of bank slip recognition device disclosed by the embodiments of the present invention；
Fig. 6 is the structural schematic diagram of image correction module disclosed by the embodiments of the present invention；
Fig. 7 is the structural schematic diagram of identification item extraction module disclosed by the embodiments of the present invention；
Fig. 8 is the structural schematic diagram of another kind bank slip recognition device disclosed in inventive embodiments.
Specific embodiment
Following will be combined with the drawings in the embodiments of the present invention, and technical solution in the embodiment of the present invention carries out clear, complete
Site preparation description, it is clear that described embodiments are only a part of the embodiments of the present invention, instead of all the embodiments.It is based on
Embodiment in the present invention, it is obtained by those of ordinary skill in the art without making creative efforts every other
Embodiment shall fall within the protection scope of the present invention.
Fig. 1 is a kind of flow chart of bank slip recognition method disclosed by the embodiments of the present invention, shown in Figure 1, the bill
Recognition methods may include:
Step 101: orienting the note position of bill to be identified from images to be recognized and identify the ticket to be identified
According to bill type, determine the image of bill to be identified.
Wherein, described bill type such as VAT invoice, quota invoice, official receipts etc..
Specifically, the document identification technology based on deep learning, can be used full convolutional neural networks, scan image is inputted,
Each pixel class prediction probability figure is exported, then (OTSU algorithm is to be proposed by Japanese scholars OTSU in 1979 by OTSU
A kind of pair of image carry out the highly effective algorithm of binaryzation) image segmentation algorithm is the category attribute for calculating each pixel, finally mention
It takes foreground pixel profile to can be obtained note position information, the bill in scan image is positioned to realize.Above-mentioned side
Method replaces the bills detection algorithms such as traditional edge detection based on image procossing, lines extraction, can shorten the identification of bill
Time.Document identification technology based on deep learning identifies different types of bill, can be from the scanning of patch ticket complexity
The position coordinates of bill to be identified are positioned in image.
Meanwhile the document identification technology based on deep learning, utilize the thought of transfer learning, it is only necessary to a small amount of bill sample
Training pattern can stablize the bill type in identification scan image using trained model, avoid to a large amount of of the same race
The dependence of the bill training sample of type.
In conjunction with above content, at one in the specific implementation, the ticket for orienting bill to be identified from images to be recognized
It according to position and identifies the bill type of the bill to be identified, determines the image of bill to be identified, may include: based on depth
The document identification technology of study determines the position of bill to be identified in images to be recognized in conjunction with bank slip recognition model trained in advance
Confidence ceases and bill type, and the image of bill to be identified is determined based on the location information.
Wherein, the bank slip recognition model, which can be, pre- first passes through what the training of existing model trainer obtained.In training
In the process, the image of target bill is first intercepted out, then input model training aids, bill images of the model trainer based on input
Sample be trained, obtain bank slip recognition model.It should be noted that in order to guarantee to train obtained bank slip recognition model
Accuracy of identification can input the image of multiple target bills to model trainer, increase the appearance of training sample in the training process
Amount, can be with repetition training, constantly promotion precision, until satisfying the use demand.
It should be noted that in the present embodiment, the case where for including multiple bill images in scan image, the present embodiment
Disclosed bank slip recognition method can identify the different bill images in a scan image respectively, and be identified respectively
Processing.Specifically, in practical applications, for different bills, there is different bank slip recognition models, in a scan image
In when including the identical or different bill images of multiple types, be based on deep learning algorithm, utilize the various bills of training in advance
Identification model, can to scan image, region is identified in whole or in part, therefore can be gone out with Direct Recognition in scan image every
One individual bill images, and be respectively processed.
In addition, identification image illustrates by taking scan image as an example in the present embodiment, but not limiting identification image only includes scanning
Image, such as photo, color photocopying image etc. can be using in the present embodiment, and example interpretation is not intended to limit this
The protection scope of application.
The process of above-mentioned identification bill type and bill positioning passes through training ticket on the whole based on the method for deep learning
According to identification model, note position is directly oriented from scan image, realizes that a step of bill type and note position determines, it is real
Existing process does not need to combine excessive algorithm, is simple and efficient.
After step 101,102 are entered step.
Step 102: according to default ticket templates corresponding with the bill type to the image of the bill to be identified into
Row correction obtains correction bill images.
In the present embodiment, the region containing feature-rich in bill images can be detected by image matching technology, then
Bill images are corrected, to evade the nonstandard problem of bill inclination during artificial patch ticket, reduce ticket contents
Identify difficulty.Wherein, the feature-rich refer in image it is obvious, with significant or identification characteristic area,
Including color characteristic, Gradient Features etc..
Specifically, it is described according to default ticket templates corresponding with the bill type to the image of the bill to be identified
It is corrected, obtains correction bill images, may include: to be passed through according to default ticket templates corresponding with the bill type
Images match and affine transformation technology are corrected the image of the bill to be identified, obtain correction bill images.
The basis default ticket templates corresponding with the bill type, pass through images match and affine transformation technology pair
The image of the bill to be identified is corrected, and the detailed process for obtaining correction bill images may refer to Fig. 2, and Fig. 2 is this hair
Image rectification flow diagram disclosed in bright embodiment, as shown in Fig. 2, image rectification process may include:
Step 201: using SIFT (Scale-invariant feature transform, Scale invariant features transform)
(Random Sample Consensus includes the sample data set of abnormal data according to one group, calculates data with RANSAC
Mathematical model parameter, obtain the algorithm of effective sample data) combine image matching algorithm, from the figure of the bill to be identified
The matching area most like with pre-stored ticket templates characteristic image is determined as in.
Wherein, the ticket templates characteristic image refers to that in ticket templates image containing feature-rich, bill itself carries
, not by printing character influenced and with higher identification region.
Step 202: the affine transformation matrix for determining the matching area relative to the ticket templates characteristic image is calculated,
And the correction for being rotated, being scaled and/or being displaced according to image of the affine transformation matrix to the bill to be identified, it obtains
Correct bill images.
The bill images finished are corrected, it is substantially consistent with ticket templates image.
After step 102,103 are entered step.
Step 103: extracting the item to be identified in the correction bill images.
In the present embodiment, can identification item coarse extraction mode based on predefined coordinate be based on deep learning text detection
The thin extracting mode of identification item combine, be accurately positioned item to be identified.Text detection therein specifically can be literal line detection skill
Art.
In specific implementation, can according to actual finance reimbursement situation and needs, the pre-set region to be identified,
According to the position for wanting identification region above-mentioned in default ticket templates, the corresponding region in bill to be identified is intercepted, then carry out subsequent
Identification.
Fig. 3 is mentioned as shown in connection with fig. 3 for the flow chart of item to be identified in extraction bill images disclosed by the embodiments of the present invention
The item to be identified in bill images is taken to may include:
Step 301: according to position coordinates of the identification item predetermined in default ticket templates, in the ticket to be identified
According to image on extract identification item of image, obtain tentatively identifying item.
Step 302: the line of text detection algorithm based on deep learning positions literal line position in the preliminary identification item,
Obtain target item to be identified.
The above process, according to position coordinates of the identification item predetermined in default ticket templates, in bill to be identified
Image on coarse extraction identify item of image, then can use the line of text detection algorithm CTPN based on deep learning
(Connectionist Text Proposal Network) is accurately positioned the position of each literal line.Wherein, line of text detects
Algorithm uses CTPN (connectionism text generation network) algorithm based on deep learning, is first mentioned with CNN (convolutional neural networks)
Depth characteristic is taken, text box is then detected with the anchor of fixed width (anchor), and the corresponding feature with a line anchor
It is linked to be sequence, is input in RNN (Recognition with Recurrent Neural Network), is finally classified or returned with full articulamentum, and by correct text
Frame is merged into line of text, and this method by CNN and RNN seamless combination improves detection accuracy.
After step 103,104 are entered step.
Step 104: the content information of the item to be identified is identified by character recognition technology.
Specifically, the content information for identifying the item to be identified by character recognition technology, may include: based on deep
The end-to-end text recognition algorithms of degree study identify the content information of the item to be identified.Text recognition algorithms, which use, is based on depth
End-to-end CRNN (convolution loop neural network) algorithm of study, extracts characteristic sequence using convolutional layer from input picture,
Circulation layer is constructed on convolutional network, for predicting the feature that convolutional layer exports, finally uses transcription layer by circulation layer
Every frame prediction is converted into sequence label, and CNN in conjunction with RNN, is carried out joint training by a loss function by CRNN.The algorithm
The input picture of random length can be handled, directly output full line Word-predictor is as a result, better than traditional based on monocase identification
Text recognition algorithms.
In the present embodiment, the bank slip recognition algorithm is succinctly intuitive, small by nominal value influence of noise degree, hardly deposits
In preset empirical value, algorithm robustness is stronger.It, can meanwhile no matter there are how many bill images on scanned copy image
Enough each of scanned copy bill images are known respectively based on deep learning algorithm and preset ticket templates and parameter
Not, recognition efficiency is high and accurate, can greatly promote the efficiency of finance reimbursement process work.
On the basis of the above disclosed embodiments of the present invention, Fig. 4 is another kind bank slip recognition disclosed in inventive embodiments
The flow chart of method, as shown in figure 4, bank slip recognition method may include:
Step 401: orienting the note position of bill to be identified from images to be recognized and identify the ticket to be identified
According to bill type, determine the image of bill to be identified.
Step 402: according to default ticket templates corresponding with the bill type to the image of the bill to be identified into
Row correction obtains correction bill images.
Step 403: extracting the item to be identified in the correction bill images.
Step 404: the content information of the item to be identified is identified by character recognition technology.
Step 405: according to the Property Name of the item to be identified and the content information identified, to the ticket to be identified
According to content carry out data structured output.
Data structured, which refers to, is combined a large amount of bill literal line recognition results according to its semantic information, and by finance
Reimbursement requires classification output.Offset coordinates using predefined each identification item relative to ticket templates image, it is known that identification
The corresponding attribute information of item can reach number to bank slip recognition result disaggregated classification in conjunction with the semantic information that line of text identifies
According to the purpose of structuring.Specifically, can be according to the Property Name of item to be identified and the semantic information pair of the item to be identified
Recognition result carries out structuring output.For example, the Property Name of item to be identified is the amount of money, identification content is " 10,000 thousand yuan of lands
It is whole ", then output is to 16000 yuan of reimbursed sum.
Above-mentioned bank slip recognition method, can according to the semantic information of the Property Name of item to be identified and specific identification content,
Automatic arranging relevant information structuring output, saves artificial statistics, substantially increases the treatment effeciency of reimbursement work.
For the various method embodiments described above, for simple description, therefore, it is stated as a series of action combinations, but
Be those skilled in the art should understand that, the present invention is not limited by the sequence of acts described because according to the present invention, certain
A little steps can be performed in other orders or simultaneously.Secondly, those skilled in the art should also know that, it is retouched in specification
The embodiment stated belongs to preferred embodiment, and related actions and modules are not necessarily necessary for the present invention.
Method is described in detail in aforementioned present invention disclosed embodiment, diversified forms can be used for method of the invention
Device realize that therefore the invention also discloses a kind of devices, and specific embodiment is given below and is described in detail.
Fig. 5 is a kind of structural schematic diagram of bank slip recognition device disclosed by the embodiments of the present invention, shown in Figure 5, bill
Identification device 50 may include:
Location type determining module 501, for orienting note position to be identified from images to be recognized and identifying institute
The bill type for stating bill to be identified determines the image of bill to be identified.
Specifically, the document identification technology based on deep learning, can be used full convolutional neural networks, scan image is inputted,
Each pixel class prediction probability figure is exported, is then the category attribute for calculating each pixel by OTSU image segmentation algorithm,
Finally extracting foreground pixel profile can be obtained note position information, position to realize to the bill in scan image.
The above method replaces the bills detection algorithms such as traditional edge detection based on image procossing, lines extraction, can shorten bill
Recognition time.Document identification technology based on deep learning identifies different types of bill, can be complicated from patch ticket
Scan image in position the position coordinates of bill to be identified.
Meanwhile the document identification technology based on deep learning, utilize the thought of transfer learning, it is only necessary to a small amount of bill sample
Training pattern can stablize the bill type in identification scan image using trained model, avoid to a large amount of of the same race
The dependence of the bill training sample of type.
In conjunction with above content, at one in the specific implementation, the location type determining module 501 is particularly used in: being based on
The document identification technology of deep learning determines bill to be identified in images to be recognized in conjunction with bank slip recognition model trained in advance
Location information and bill type, and determine based on the location information image of bill to be identified.
It should be noted that in the present embodiment, the case where for including multiple bill images in scan image, the present embodiment
Disclosed bank slip recognition method can identify the different bill images in a scan image respectively, and be identified respectively
Processing.
Image correction module 502 is used for according to default ticket templates corresponding with the bill type to described to be identified
The image of bill is corrected, and obtains correction bill images.
In the present embodiment, the region containing feature-rich in bill images can be detected by image matching technology, then
Bill images are corrected, to evade the nonstandard problem of bill inclination during artificial patch ticket, reduce ticket contents
Identify difficulty.
Specifically, described image correction module 502 can be used for: according to default bill mould corresponding with the bill type
Plate is corrected the image of the bill to be identified by images match and affine transformation technology, obtains correction bill images.
The specific structure of described image correction module 502 may refer to Fig. 6, and Fig. 6 is image disclosed by the embodiments of the present invention
The structural schematic diagram of correction module, as shown in fig. 6, image correction module 502 may include:
Images match module 601, the image matching algorithm for being combined using SIFT and RANSAC, from the ticket to be identified
According to image in determine the matching area most like with pre-stored ticket templates characteristic image.
Wherein, the ticket templates characteristic image refers to that in ticket templates image containing feature-rich, bill itself carries
, the region not influenced by printing character.
Affine transformation module 602 determines the matching area relative to the ticket templates characteristic image for calculating
Affine transformation matrix, and rotated according to image of the affine transformation matrix to the bill to be identified, scaled and/or position
The correction of shifting obtains correction bill images.
The bill images finished are corrected, it is substantially consistent with ticket templates image.
Item extraction module 503 is identified, for extracting the item to be identified in the correction bill images.
In the present embodiment, can identification item coarse extraction mode based on predefined coordinate be based on deep learning text detection
The thin extracting mode of identification item combine, be accurately positioned item to be identified.Text detection therein specifically can be literal line detection skill
Art.
In specific implementation, can according to actual finance reimbursement situation and needs, the pre-set region to be identified,
According to the position for wanting identification region above-mentioned in default ticket templates, the corresponding region in bill to be identified is intercepted, then carry out subsequent
Identification.
Fig. 7 is the structural schematic diagram of identification item extraction module disclosed by the embodiments of the present invention, shown in Figure 7, at one
In example, identification item extraction module 503 may include:
Coarse extraction module 701, for the position coordinates according to identification item predetermined in default ticket templates, in institute
It states and extracts identification item of image on the image of bill to be identified, obtain tentatively identifying item.
Thin extraction module 702, positions in the preliminary identification item for the line of text detection algorithm based on deep learning
Literal line position obtains target item to be identified.
The above process, according to position coordinates of the identification item predetermined in default ticket templates, in bill to be identified
Image on coarse extraction identify item of image, then can using based on deep learning line of text detection algorithm CTPN be accurately positioned
The position of each literal line.
Content identifier module 504, for identifying the content information of the item to be identified by character recognition technology.
Specifically, the content identifier module 504 can be used for: the end-to-end text recognition algorithms identification based on deep learning
The content information of the item to be identified.
In the present embodiment, the bank slip recognition device algorithm is succinctly intuitive, small by nominal value influence of noise degree, hardly deposits
In preset empirical value, algorithm robustness is stronger.It, can meanwhile no matter there are how many bill images on scanned copy image
Enough each of scanned copy bill images are known respectively based on deep learning algorithm and preset ticket templates and parameter
Not, recognition efficiency is high and accurate, can greatly promote the efficiency of finance reimbursement process work.
On the basis of the above disclosed embodiments of the present invention, Fig. 8 is another kind bank slip recognition disclosed in inventive embodiments
The structural schematic diagram of device, as shown in figure 8, bank slip recognition device 80 may include:
Location type determining module 501, for orienting note position and the knowledge of bill to be identified from images to be recognized
Not Chu the bill to be identified bill type, determine the image of bill to be identified.
Image correction module 502 is used for according to default ticket templates corresponding with the bill type to described to be identified
The image of bill is corrected, and obtains correction bill images.
Item extraction module 503 is identified, for extracting the item to be identified in the correction bill images.
Content identifier module 504, for identifying the content information of the item to be identified by character recognition technology.
Structuring output module 801, for according to the Property Name of the item to be identified and the content information identified,
Data structured output is carried out to the content of the bill to be identified.
Specifically, can be tied according to the Property Name of item to be identified and the semantic information of the item to be identified to identification
Fruit carries out structuring output.For example, the Property Name of item to be identified is the amount of money, identification content is " 10,000 thousand yuan of lands are whole ", then defeated
Out to 16000 yuan of reimbursed sum.
Above-mentioned bank slip recognition device, can according to the semantic information of the Property Name of item to be identified and specific identification content,
Automatic arranging relevant information structuring output, saves artificial statistics, substantially increases the treatment effeciency of reimbursement work.
Any one described bank slip recognition device in above-described embodiment includes processor and memory, above-described embodiment
In location type determining module, image correction module, identification item extraction module, content identifier module, images match module, imitative
It penetrates conversion module, coarse extraction module, thin extraction module, structuring output module etc. and is stored in memory as program module
In, the above procedure module of storage in the memory is executed by processor to realize corresponding function.
Include kernel in processor, is gone in memory to transfer corresponding program module by kernel.Kernel can be set one
Or it is multiple, the processing of return visit data is realized by adjusting kernel parameter.
Memory may include the non-volatile memory in computer-readable medium, random access memory (RAM) and/
Or the forms such as Nonvolatile memory, such as read-only memory (ROM) or flash memory (flashRAM), memory includes at least one storage
Chip.
The embodiment of the invention provides a kind of storage mediums, are stored thereon with program, real when which is executed by processor
Bank slip recognition method described in existing above-described embodiment.
The embodiment of the invention provides a kind of processor, the processor is for running program, wherein described program operation
Bank slip recognition method described in Shi Zhihang above-described embodiment.
Further, a kind of electronic equipment, including processor and memory are present embodiments provided.Wherein memory is used for
The executable instruction of the processor is stored, the processor is configured to execute above-mentioned reality via the executable instruction is executed
Apply bank slip recognition method described in example.
Each embodiment in this specification is described in a progressive manner, the highlights of each of the examples are with other
The difference of embodiment, the same or similar parts in each embodiment may refer to each other.For device disclosed in embodiment
For, since it is corresponded to the methods disclosed in the examples, so being described relatively simple, related place is said referring to method part
It is bright.
It should also be noted that, herein, relational terms such as first and second and the like are used merely to one
Entity or operation are distinguished with another entity or operation, without necessarily requiring or implying between these entities or operation
There are any actual relationship or orders.Moreover, the terms "include", "comprise" or its any other variant are intended to contain
Lid non-exclusive inclusion, so that the process, method, article or equipment including a series of elements is not only wanted including those
Element, but also including other elements that are not explicitly listed, or further include for this process, method, article or equipment
Intrinsic element.In the absence of more restrictions, the element limited by sentence "including a ...", it is not excluded that
There is also other identical elements in process, method, article or equipment including the element.
The step of method described in conjunction with the examples disclosed in this document or algorithm, can directly be held with hardware, processor
The combination of capable software module or the two is implemented.Software module can be placed in random access memory (RAM), memory, read-only deposit
Reservoir (ROM), electrically programmable ROM, electrically erasable ROM, register, hard disk, moveable magnetic disc, CD-ROM or technology
In any other form of storage medium well known in field.
The foregoing description of the disclosed embodiments enables those skilled in the art to implement or use the present invention.
Various modifications to these embodiments will be readily apparent to those skilled in the art, as defined herein
General Principle can be realized in other embodiments without departing from the spirit or scope of the present invention.Therefore, of the invention
It is not intended to be limited to the embodiments shown herein, and is to fit to and the principles and novel features disclosed herein phase one
The widest scope of cause.
Claims (10)
1. a kind of bank slip recognition method characterized by comprising
The note position of bill to be identified is oriented from images to be recognized and identifies the bill type of the bill to be identified,
Determine the image of bill to be identified；
It is corrected according to image of the default ticket templates corresponding with the bill type to the bill to be identified, obtains school
Positive bill images；
Extract the item to be identified in the correction bill images；
The content information of the item to be identified is identified by character recognition technology.
2. bank slip recognition method according to claim 1, which is characterized in that described to orient from images to be recognized wait know
The note position of other bill and the bill type for identifying the bill to be identified, determine the image of bill to be identified, comprising:
Document identification technology based on deep learning, in conjunction with bank slip recognition model trained in advance, determine in images to be recognized to
It identifies the location information and bill type of bill, and determines the image of bill to be identified based on the location information.
3. bank slip recognition method according to claim 1, which is characterized in that described according to corresponding with the bill type
Default ticket templates are corrected the image of the bill to be identified, obtain correction bill images, comprising:
According to default ticket templates corresponding with the bill type, by images match and affine transformation technology to described wait know
The image of other bill is corrected, and obtains correction bill images.
4. bank slip recognition method according to claim 3, which is characterized in that the basis is corresponding with the bill type
Default ticket templates, are corrected the image of the bill to be identified by images match and affine transformation technology, obtain school
Positive bill images, comprising:
Using SIFT and RANSAC in conjunction with image matching algorithm determine from the image of the bill to be identified and deposit in advance
The most like matching area of the ticket templates characteristic image of storage；
The affine transformation matrix for determining the matching area relative to the ticket templates characteristic image is calculated, and according to described imitative
The correction that transformation matrix is rotated, scaled and/or is displaced to the image of the bill to be identified is penetrated, correction bill is obtained
Picture.
5. bank slip recognition method according to claim 1, which is characterized in that described to extract in the correction bill images
Item to be identified, comprising:
According to position coordinates of the identification item predetermined in default ticket templates, above mentioned in the image of the bill to be identified
Identification item of image is taken, obtains tentatively identifying item；
Line of text detection algorithm based on deep learning positions literal line position in the preliminary identification item, obtains target and waits knowing
Other item.
6. bank slip recognition method according to claim 1, which is characterized in that described to orient from images to be recognized wait know
The note position of other bill, comprising:
Document identification technology based on deep learning inputs the images to be recognized using full convolutional neural networks, described in output
The prediction probability figure of each pixel class in images to be recognized；
Based on the prediction probability figure of each pixel class, the classification category of each pixel is calculated by OTSU image segmentation algorithm
Property, and foreground pixel profile is extracted to obtain note position information.
7. bank slip recognition method according to claim 1, which is characterized in that identify institute by character recognition technology described
After the content information for stating item to be identified, the method also includes:
According to the Property Name of the item to be identified and the content information identified, the content of the bill to be identified is carried out
Data structured output.
8. a kind of bank slip recognition device characterized by comprising
Location type determining module, for orienting note position to be identified from images to be recognized and identifying described to be identified
The bill type of bill determines the image of bill to be identified；
Image correction module, for the figure according to default ticket templates corresponding with the bill type to the bill to be identified
As being corrected, correction bill images are obtained；
Item extraction module is identified, for extracting the item to be identified in the correction bill images；
Content identifier module, for identifying the content information of the item to be identified by character recognition technology.
9. a kind of computer readable storage medium, is stored thereon with computer program, which is characterized in that the program is held by processor
Claim 1-7 described in any item bank slip recognition methods are realized when row.
10. a kind of electronic equipment characterized by comprising
Processor；And
Memory, for storing the executable instruction of the processor；
Wherein, the processor is configured to require 1-7 described in any item via executing the executable instruction and carry out perform claim
Bank slip recognition method.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910510860.2A CN110263694A (en) | 2019-06-13 | 2019-06-13 | A kind of bank slip recognition method and device |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910510860.2A CN110263694A (en) | 2019-06-13 | 2019-06-13 | A kind of bank slip recognition method and device |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN110263694A true CN110263694A (en) | 2019-09-20 |
Family
ID=67918140
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910510860.2A Pending CN110263694A (en) | 2019-06-13 | 2019-06-13 | A kind of bank slip recognition method and device |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110263694A (en) |
Cited By (33)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110674815A (en) * | 2019-09-29 | 2020-01-10 | 四川长虹电器股份有限公司 | Invoice image distortion correction method based on deep learning key point detection |
| CN110751136A (en) * | 2019-11-04 | 2020-02-04 | 北京亿信华辰软件有限责任公司武汉分公司 | Method for extracting value-added tax invoice information |
| CN110956166A (en) * | 2019-12-02 | 2020-04-03 | 中国银行股份有限公司 | Bill marking method and device |
| CN111079531A (en) * | 2019-11-12 | 2020-04-28 | 泰康保险集团股份有限公司 | Data structured output method and device, electronic equipment and storage medium |
| CN111079571A (en) * | 2019-11-29 | 2020-04-28 | 杭州数梦工场科技有限公司 | Identification card information identification and edge detection model training method and device |
| CN111178353A (en) * | 2019-12-16 | 2020-05-19 | 中国建设银行股份有限公司 | Image character positioning method and device |
| CN111209856A (en) * | 2020-01-06 | 2020-05-29 | 泰康保险集团股份有限公司 | Invoice information identification method and device, electronic equipment and storage medium |
| CN111209827A (en) * | 2019-12-31 | 2020-05-29 | 中国南方电网有限责任公司 | OCR (optical character recognition) bill problem recognition method and system based on feature detection |
| CN111241955A (en) * | 2020-01-03 | 2020-06-05 | 北京一览群智数据科技有限责任公司 | Bill information extraction method and system |
| CN111242124A (en) * | 2020-01-13 | 2020-06-05 | 支付宝实验室(新加坡)有限公司 | Certificate classification method, device and equipment |
| CN111241974A (en) * | 2020-01-07 | 2020-06-05 | 深圳追一科技有限公司 | Bill information acquisition method and device, computer equipment and storage medium |
| CN111275037A (en) * | 2020-01-09 | 2020-06-12 | 上海知达教育科技有限公司 | Bill identification method and device |
| CN111291752A (en) * | 2020-01-22 | 2020-06-16 | 山东浪潮通软信息科技有限公司 | Invoice identification method, equipment and medium |
| CN111414917A (en) * | 2020-03-18 | 2020-07-14 | 民生科技有限责任公司 | Identification method of low-pixel-density text |
| CN111428599A (en) * | 2020-03-17 | 2020-07-17 | 北京公瑾科技有限公司 | Bill identification method, device and equipment |
| CN111461099A (en) * | 2020-03-27 | 2020-07-28 | 重庆农村商业银行股份有限公司 | Bill identification method, system, equipment and readable storage medium |
| CN111461100A (en) * | 2020-03-31 | 2020-07-28 | 重庆农村商业银行股份有限公司 | Bill identification method and device, electronic equipment and storage medium |
| CN111553334A (en) * | 2020-04-21 | 2020-08-18 | 招商局金融科技有限公司 | Questionnaire image recognition method, electronic device, and storage medium |
| CN111546804A (en) * | 2020-04-08 | 2020-08-18 | 远光软件股份有限公司 | Automatic original bill pasting method and device |
| CN111639566A (en) * | 2020-05-19 | 2020-09-08 | 浙江大华技术股份有限公司 | Method and device for extracting form information |
| CN111680679A (en) * | 2020-06-03 | 2020-09-18 | 重庆数道科技有限公司 | Automatic document identification method based on OCR |
| CN111832423A (en) * | 2020-06-19 | 2020-10-27 | 北京邮电大学 | Bill information identification method, device and system |
| CN111914835A (en) * | 2020-07-04 | 2020-11-10 | 中信银行股份有限公司 | Bill element extraction method and device, electronic equipment and readable storage medium |
| CN111967391A (en) * | 2020-08-18 | 2020-11-20 | 清华大学 | Text recognition method and computer-readable storage medium for medical laboratory test reports |
| CN111967395A (en) * | 2020-08-18 | 2020-11-20 | 中国银行股份有限公司 | Bank bill identification method and device |
| CN112800848A (en) * | 2020-12-31 | 2021-05-14 | 中电金信软件有限公司 | Structured extraction method, device and equipment of information after bill identification |
| CN112818951A (en) * | 2021-03-11 | 2021-05-18 | 南京大学 | Ticket identification method |
| WO2021123209A1 (en) * | 2019-12-18 | 2021-06-24 | Carrus Gaming | Method for segmenting an input image showing a document containing structured information |
| CN113033269A (en) * | 2019-12-25 | 2021-06-25 | 华为技术服务有限公司 | Data processing method and device |
| CN113052161A (en) * | 2021-04-30 | 2021-06-29 | 中国银行股份有限公司 | Method, device and equipment for identifying bank bill text |
| WO2021151274A1 (en) * | 2020-05-20 | 2021-08-05 | 平安科技（深圳）有限公司 | Image file processing method and apparatus, electronic device, and computer readable storage medium |
| CN113408516A (en) * | 2021-06-25 | 2021-09-17 | 京东数科海益信息科技有限公司 | Bill recognition device and method |
| CN113569650A (en) * | 2021-06-29 | 2021-10-29 | 上海红檀智能科技有限公司 | Unmanned aerial vehicle autonomous inspection positioning method based on electric power tower label identification |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107977665A (en) * | 2017-12-15 | 2018-05-01 | 北京科摩仕捷科技有限公司 | The recognition methods of key message and computing device in a kind of invoice |
| CN108062756A (en) * | 2018-01-29 | 2018-05-22 | 重庆理工大学 | Image, semantic dividing method based on the full convolutional network of depth and condition random field |
| CN108664897A (en) * | 2018-04-18 | 2018-10-16 | 平安科技（深圳）有限公司 | Bank slip recognition method, apparatus and storage medium |
| CN108921166A (en) * | 2018-06-22 | 2018-11-30 | 深源恒际科技有限公司 | Medical bill class text detection recognition method and system based on deep neural network |
| CN109658584A (en) * | 2018-12-14 | 2019-04-19 | 泰康保险集团股份有限公司 | A kind of bill bank slip recognition method and device |
-
2019
- 2019-06-13 CN CN201910510860.2A patent/CN110263694A/en active Pending
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107977665A (en) * | 2017-12-15 | 2018-05-01 | 北京科摩仕捷科技有限公司 | The recognition methods of key message and computing device in a kind of invoice |
| CN108062756A (en) * | 2018-01-29 | 2018-05-22 | 重庆理工大学 | Image, semantic dividing method based on the full convolutional network of depth and condition random field |
| CN108664897A (en) * | 2018-04-18 | 2018-10-16 | 平安科技（深圳）有限公司 | Bank slip recognition method, apparatus and storage medium |
| CN108921166A (en) * | 2018-06-22 | 2018-11-30 | 深源恒际科技有限公司 | Medical bill class text detection recognition method and system based on deep neural network |
| CN109658584A (en) * | 2018-12-14 | 2019-04-19 | 泰康保险集团股份有限公司 | A kind of bill bank slip recognition method and device |
Non-Patent Citations (4)
| Title |
|---|
| ZHI TIAN ET AL.: "Detecting Text in Natural Image with Connectionist Text Proposal Network", 《COMPUTER VISION – ECCV 2016》 * |
| 中国税务学会: "《2016年全国税收理论研讨文集》", 31 December 2017 * |
| 韩九强，杨磊: "《数字图像处理 基于XAVIS组态软件》", 30 September 2018, 西安交通大学出版社 * |
| 黄志文: "基于深度学习的***自动识别***的设计与实现", 《中国优秀硕士学位论文全文数据库信息科技辑》 * |
Cited By (44)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110674815A (en) * | 2019-09-29 | 2020-01-10 | 四川长虹电器股份有限公司 | Invoice image distortion correction method based on deep learning key point detection |
| CN110751136A (en) * | 2019-11-04 | 2020-02-04 | 北京亿信华辰软件有限责任公司武汉分公司 | Method for extracting value-added tax invoice information |
| CN111079531A (en) * | 2019-11-12 | 2020-04-28 | 泰康保险集团股份有限公司 | Data structured output method and device, electronic equipment and storage medium |
| CN111079571A (en) * | 2019-11-29 | 2020-04-28 | 杭州数梦工场科技有限公司 | Identification card information identification and edge detection model training method and device |
| CN110956166A (en) * | 2019-12-02 | 2020-04-03 | 中国银行股份有限公司 | Bill marking method and device |
| CN111178353A (en) * | 2019-12-16 | 2020-05-19 | 中国建设银行股份有限公司 | Image character positioning method and device |
| FR3105529A1 (en) * | 2019-12-18 | 2021-06-25 | Idemia Identity & Security France | A method of segmenting an input image representing a document containing structured information |
| WO2021123209A1 (en) * | 2019-12-18 | 2021-06-24 | Carrus Gaming | Method for segmenting an input image showing a document containing structured information |
| CN113033269B (en) * | 2019-12-25 | 2023-08-25 | 华为技术服务有限公司 | Data processing method and device |
| CN113033269A (en) * | 2019-12-25 | 2021-06-25 | 华为技术服务有限公司 | Data processing method and device |
| CN111209827A (en) * | 2019-12-31 | 2020-05-29 | 中国南方电网有限责任公司 | OCR (optical character recognition) bill problem recognition method and system based on feature detection |
| CN111209827B (en) * | 2019-12-31 | 2023-07-14 | 中国南方电网有限责任公司 | Method and system for OCR (optical character recognition) bill problem based on feature detection |
| CN111241955A (en) * | 2020-01-03 | 2020-06-05 | 北京一览群智数据科技有限责任公司 | Bill information extraction method and system |
| CN111241955B (en) * | 2020-01-03 | 2023-05-16 | 北京一览群智数据科技有限责任公司 | Bill information extraction method and system |
| CN111209856B (en) * | 2020-01-06 | 2023-10-17 | 泰康保险集团股份有限公司 | Invoice information identification method and device, electronic equipment and storage medium |
| CN111209856A (en) * | 2020-01-06 | 2020-05-29 | 泰康保险集团股份有限公司 | Invoice information identification method and device, electronic equipment and storage medium |
| CN111241974A (en) * | 2020-01-07 | 2020-06-05 | 深圳追一科技有限公司 | Bill information acquisition method and device, computer equipment and storage medium |
| CN111241974B (en) * | 2020-01-07 | 2023-10-27 | 深圳追一科技有限公司 | Bill information acquisition method, device, computer equipment and storage medium |
| CN111275037A (en) * | 2020-01-09 | 2020-06-12 | 上海知达教育科技有限公司 | Bill identification method and device |
| CN111275037B (en) * | 2020-01-09 | 2021-06-08 | 上海知达教育科技有限公司 | Bill identification method and device |
| CN111242124A (en) * | 2020-01-13 | 2020-06-05 | 支付宝实验室(新加坡)有限公司 | Certificate classification method, device and equipment |
| CN111242124B (en) * | 2020-01-13 | 2023-10-31 | 支付宝实验室(新加坡)有限公司 | Certificate classification method, device and equipment |
| CN111291752A (en) * | 2020-01-22 | 2020-06-16 | 山东浪潮通软信息科技有限公司 | Invoice identification method, equipment and medium |
| CN111428599B (en) * | 2020-03-17 | 2023-10-20 | 北京子敬科技有限公司 | Bill identification method, device and equipment |
| CN111428599A (en) * | 2020-03-17 | 2020-07-17 | 北京公瑾科技有限公司 | Bill identification method, device and equipment |
| CN111414917A (en) * | 2020-03-18 | 2020-07-14 | 民生科技有限责任公司 | Identification method of low-pixel-density text |
| CN111461099A (en) * | 2020-03-27 | 2020-07-28 | 重庆农村商业银行股份有限公司 | Bill identification method, system, equipment and readable storage medium |
| CN111461100A (en) * | 2020-03-31 | 2020-07-28 | 重庆农村商业银行股份有限公司 | Bill identification method and device, electronic equipment and storage medium |
| CN111546804B (en) * | 2020-04-08 | 2021-03-23 | 远光软件股份有限公司 | Automatic original bill pasting method and device |
| CN111546804A (en) * | 2020-04-08 | 2020-08-18 | 远光软件股份有限公司 | Automatic original bill pasting method and device |
| CN111553334A (en) * | 2020-04-21 | 2020-08-18 | 招商局金融科技有限公司 | Questionnaire image recognition method, electronic device, and storage medium |
| CN111639566A (en) * | 2020-05-19 | 2020-09-08 | 浙江大华技术股份有限公司 | Method and device for extracting form information |
| WO2021151274A1 (en) * | 2020-05-20 | 2021-08-05 | 平安科技（深圳）有限公司 | Image file processing method and apparatus, electronic device, and computer readable storage medium |
| CN111680679A (en) * | 2020-06-03 | 2020-09-18 | 重庆数道科技有限公司 | Automatic document identification method based on OCR |
| CN111832423A (en) * | 2020-06-19 | 2020-10-27 | 北京邮电大学 | Bill information identification method, device and system |
| CN111914835A (en) * | 2020-07-04 | 2020-11-10 | 中信银行股份有限公司 | Bill element extraction method and device, electronic equipment and readable storage medium |
| CN111967395A (en) * | 2020-08-18 | 2020-11-20 | 中国银行股份有限公司 | Bank bill identification method and device |
| CN111967391A (en) * | 2020-08-18 | 2020-11-20 | 清华大学 | Text recognition method and computer-readable storage medium for medical laboratory test reports |
| CN112800848A (en) * | 2020-12-31 | 2021-05-14 | 中电金信软件有限公司 | Structured extraction method, device and equipment of information after bill identification |
| CN112818951A (en) * | 2021-03-11 | 2021-05-18 | 南京大学 | Ticket identification method |
| CN112818951B (en) * | 2021-03-11 | 2023-11-21 | 南京大学 | Ticket identification method |
| CN113052161A (en) * | 2021-04-30 | 2021-06-29 | 中国银行股份有限公司 | Method, device and equipment for identifying bank bill text |
| CN113408516A (en) * | 2021-06-25 | 2021-09-17 | 京东数科海益信息科技有限公司 | Bill recognition device and method |
| CN113569650A (en) * | 2021-06-29 | 2021-10-29 | 上海红檀智能科技有限公司 | Unmanned aerial vehicle autonomous inspection positioning method based on electric power tower label identification |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110263694A (en) | A kind of bank slip recognition method and device | |
| CN109492643B (en) | Certificate identification method and device based on OCR, computer equipment and storage medium | |
| Luo et al. | Moran: A multi-object rectified attention network for scene text recognition | |
| Yuliang et al. | Detecting curve text in the wild: New dataset and new solution | |
| WO2019238063A1 (en) | Text detection and analysis method and apparatus, and device | |
| RU2695489C1 (en) | Identification of fields on an image using artificial intelligence | |
| US8494273B2 (en) | Adaptive optical character recognition on a document with distorted characters | |
| CN108846385B (en) | Image identification and correction method and device based on convolution-deconvolution neural network | |
| CN110874618B (en) | OCR template learning method and device based on small sample, electronic equipment and medium | |
| Bulatovich et al. | MIDV-2020: a comprehensive benchmark dataset for identity document analysis | |
| CN111353491B (en) | Text direction determining method, device, equipment and storage medium | |
| CN109784342A (en) | A kind of OCR recognition methods and terminal based on deep learning model | |
| CN108681735A (en) | Optical character recognition method based on convolutional neural networks deep learning model | |
| Caldeira et al. | Industrial optical character recognition system in printing quality control of hot-rolled coils identification | |
| CN110119460A (en) | Image search method, device and electronic equipment | |
| CN109271980A (en) | A kind of vehicle nameplate full information recognition methods, system, terminal and medium | |
| WO2021232670A1 (en) | Pcb component identification method and device | |
| CN110598566A (en) | Image processing method, device, terminal and computer readable storage medium | |
| CN113158895A (en) | Bill identification method and device, electronic equipment and storage medium | |
| CN112464845A (en) | Bill recognition method, equipment and computer storage medium | |
| CN112241727A (en) | Multi-ticket identification method and system and readable storage medium | |
| CN112115907A (en) | Method, device, equipment and medium for extracting structured information of fixed layout certificate | |
| RU2633182C1 (en) | Determination of text line orientation | |
| CN111462388A (en) | Bill inspection method and device, terminal equipment and storage medium | |
| CN110796145A (en) | Multi-certificate segmentation association method based on intelligent decision and related equipment |
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 | ||
| RJ01 | Rejection of invention patent application after publication | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20190920 |