CN104217203B - Complex background card face information identifying method and system - Google Patents
Complex background card face information identifying method and system Download PDFInfo
- Publication number
- CN104217203B CN104217203B CN201310216817.8A CN201310216817A CN104217203B CN 104217203 B CN104217203 B CN 104217203B CN 201310216817 A CN201310216817 A CN 201310216817A CN 104217203 B CN104217203 B CN 104217203B
- Authority
- CN
- China
- Prior art keywords
- character information
- image
- character
- block
- pixels
- 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
Landscapes
- Character Discrimination (AREA)
- Character Input (AREA)
Abstract
The application provides a kind of complex background card face information identifying method and system.This method comprises: the character information region in the image in the positioning card face to be identified, to obtain one or more character information images；The extraction of complex characteristic vector is carried out to the character information image of acquisition；According to the complex characteristic vector, the character information in the character information image is identified.The problem of information of (such as bank card) accurately identifies under the background of complicated card face, it is based on complicated image feature, overcome that traditional method based on Threshold segmentation faced can not be to the defect that complex background card face information is effectively accurately identified, the information recognition accuracy of card face is improved, realizes the identification of efficient complex background card face information.
Description
Technical field
This application involves field of image processing more particularly to a kind of complex background card face information identifying method and systems.
Background technique
With the development of internet, the quick payments such as on-line payment, mobile-phone payment are more and more common.During this period,
The case where needing bank card user to directly input card information is more and more.Bank card card face is obtained using image-capturing apparatus to believe
Breath, auxiliary information Rapid input also have become trend trend.
Currently, optical character recognition technology can be used for decoding using including including but not limited to scanner and digital camera
The image of the character of such as horizontal line of text that various types of image-capturing apparatus obtain etc, for reading bank card card
The information that piece surface is printed, such as card number, validity period, holder name and bank card hair fastener tissue etc..Optical character identification side
Method, basic calculating process are display foreground background segment, binaryzation, refinement, coding and identification.Due to having dull background false
If in traditional Optical Character Recognition system, text prospect is easy compared with the segmentation of background；The text prospect extracted can
Carry out binaryzation, refinement, coding are then identified that calculate on the whole easier, calculation amount is few.However, such method master
To be suitable for the text identification of dull background, can not identify the character in complex background.Thus, known based on traditional optical character
The identification technology of other principle has a significant limitation, and it is relatively low that limitation is mainly based upon the recognition result degree of reliability, secondly
Traditional optical character identifying method can not be efficiently applied to usually have complicated back mainly for dull background text document design
The card face information of the bank card of scape pattern identifies that user needs to spend more energy inspection, correcting identification result, to be protected
Stored correctness.
Therefore, demand one kind can accurately identify the technology of card release face information from the image of complex background card face.
Summary of the invention
The main purpose of the application is to provide a kind of complex background card face information identifying method and its system, to solve reality
The problem of information of (such as bank card) accurately identifies under the background of existing complexity card face.
According to the embodiment of the one aspect of the application, a kind of complex background card face information identifying method is provided, comprising: fixed
Character information region in the image in the position card face to be identified, to obtain one or more character information images；It is right
The character information image obtained carries out the extraction of complex characteristic vector；According to the complex characteristic vector, the character is identified
Character information in information image.
According to the embodiment of the another aspect of the application, a kind of complex background card face information identification system is provided, comprising: fixed
Bit location positions the character information region in the image in the card face to be identified, to obtain one or more character letters
Cease image；Fisrt feature extraction unit carries out the extraction of complex characteristic vector to the character information image of acquisition；Character information
Recognition unit identifies the character information in the character information image according to the complex characteristic vector.
Compared with prior art, the application provides one kind and is based on complicated image feature such as Harris, SIFT feature, is suitable for
The card face information identifying method and its system of complex background, overcome that traditional method based on Threshold segmentation faced can not be right
The defect that complex background card face information is effectively accurately identified improves the information recognition accuracy of card face, realizes efficient
The identification of complex background card face information.
Detailed description of the invention
The drawings described herein are used to provide a further understanding of the present application, constitutes part of this application, this Shen
Illustrative embodiments and their description please are not constituted an undue limitation on the present application for explaining the application.In the accompanying drawings:
Fig. 1 is the broad flow diagram of the complex background card face information identifying method of the application one embodiment；
Fig. 2 is the flow chart of preferred preprocess method in the recognition methods of the application one embodiment；
Fig. 3 is feature extraction and calculation schematic diagram in the preprocess method of the application one embodiment；
Fig. 4 is the flow chart of preferred character locating method in the recognition methods of the application one embodiment；
Fig. 5 is images match glide direction precedence diagram in the character locating method of the application one embodiment；
Fig. 6 is that recognition methods flow chart is preferably judged in the recognition methods of the application one embodiment；
Fig. 7 is the structural block diagram of the complex background card face information identification system of the application one embodiment.
Specific embodiment
The main idea of the present application lies in that based on complicated image feature, the identification of card face information suitable for complex background,
It applies complicated image feature such as Harris and SIFT feature, by acquiring a large amount of labeled data sample training classifiers as propped up
Hold after vector machine, artificial neural network etc. are trained have can be carried out identification series of parameters data classifier (with
It is known as identifier down), it is corresponding to obtain which be used to fit through the character picture region that above-mentioned identifier is partitioned into
Character information, so that solve that traditional method based on Threshold segmentation faced can not identify asking for complex background information
Topic.
To keep the purposes, technical schemes and advantages of the application clearer, below in conjunction with drawings and the specific embodiments, to this
Application is described in further detail.
The application does not do picture prospect, background segment, directly uses character candidates region entirety to be identified as text
Prospect calculates its feature vector, and using machine learning method training classifier to obtain identifier；Meanwhile it is accurate to obtain
Character candidates region needs repeatedly to be identified in the calculating of the application, and used characteristics of image is larger, calculates relative complex
Requirement thus to operation platform is relatively high.
Fig. 1 shows main flow Figure 100 of the complex background card face information identifying method of the application one embodiment.
Step S110 inputs the card face image to be identified with complex background card.
Usually object (such as bank to be identified is obtained using image capture device such as camera, camera, scanner etc.
Card) card face image.On these card face images with complex background, there are the various information for needing typing, mainly character is believed
Breath, comprising: number, letter, text, other various characters.
Step S120, on the card face image, location of concern information/character information to be identified (the various letters of such as card
Breath: card number, number, date, name etc.) image-region.
It can be by coarse localization to the process for being accurately positioned (or crying accurate positionin).
For example, positioning character (including number, card number, text etc.) the information institute on the card face image of bank card to be identified
Region.
Here standard can be made according to card, such as bank card, to bank card number region, validity date region, held
These include the region of character (number, card number) in people's name region etc., Primary Location are carried out, such as preferred characters information locating method
Coarse localization mentioned by 400.
Then it is accurately positioned, such as is carried out on Primary Location result periphery quickly into one based on card face texture information again
Step is accurately positioned, to obtain the accurate positionin in card face character information region.Preferably, character information localization method can be used
The 400 accurate positionin modes referred to realize that it is corresponding that single character information region even can be obtained during accurate positioning
Character information image.
Step S130, based on the positioning to the character information region on the image of card face, to character information image to be identified
In character information identified.
Card face information is mainly that various character informations include text (such as holder name, validity date), digital (such as bank
Card number, the date number of card) etc..It is main as follows:
First, obtain single character information image (image of each individually character information).
A kind of mode is, for kinds of characters information area, can to carry out character after character information region is accurately positioned
The segmentation of information.Partitioning scheme is for example: can using card number information region segmentation at 21 sub-regions as card number alternative area, can
Validity date region is divided into month and time (month and year information back slash "/" are divided), can be by holder's surname
Name separates its each English alphabet.It can determine that single character information region in this way, and obtain single character information image.
Preferred another way can just obtain character information image in positioning, such as: pass through following character informations
Positioning step 400 orients the accurate location of each character information, and during the accurate positionin, just obtains single word
Information image is accorded with, and a series of single character information image is provided and is used in case identifying.It has been obtained thus, it is possible to receive
Each of individual character information image.
Second, to a series of character information images (a series of images of single character informations) of acquisition, carry out feature
It extracts, it is preferable that can be using the feature extraction mode in following preprocess methods 200, as to 24*24 pixel criterion image
8*8 pixel node-by-node algorithm, obtain as Harris isogonism point feature, SIFT have rotational invariance feature (i.e. complex characteristic to
Amount).
Third, further according to these features (complex characteristic vector), using the identifier that training classifier obtains in advance,
To be judged the character information to identify in correspondence image region.Machine learning algorithm (neural network, support can be passed through
The classifiers such as vector machine SVM) character database is trained in advance, obtain identifier (i.e. the classifier with supplemental characteristic or
Claim character classifier), by extracting the complex characteristics vector such as character notable feature for example above-mentioned Harris, SIFT, with identifier into
The judgement of row Classification and Identification is to identify character information.Preferably, as believed using following preferred preprocess methods 200 and character
The judgement recognition methods 600 of breath.
Step S140 verifies the result identified, is verified using algorithm and judges the word identified through identifier
Accord with information.For bank card card, the Luhn mould ten that bank card number coding generallys use can be used and verify
Algorithm verification for example judges the card number identified in recognition methods 600 to confirm final recognition result.It verifies and successfully then exports most
Terminate fruit.Error message can also be provided when verification failure.
As shown in Fig. 2, for adoptable preferred pretreatment method flow diagram in the information identifying method of complex background card face
200。
Acquisition step S210, acquire a large amount of image data sample in advance and these image data samples are split and
Label.Thus, it is possible to obtain the individual character information image (sample image) largely classified.
It before the information on the image to be identified to input identifies, is pre-processed, to obtain energy
Sample (character information sample) database and identifier identified to information to be identified is (with supplemental characteristic after training
Classifier).
A kind of specific mode, can be and acquire a large amount of card face (such as: credit card) image data sample, usually band in advance
Then the image of character information is split these image data samples, the single character information image that will be obtained after segmentation
(calling sample image in the following text) is marked.
By the segmentation, the character information in image data sample can be separated, form individual character information image
(sample image).And by the label, can individual character information image each in the sample image of acquisition be determined and arrive its institute
In the character information classification of category.
The segmentation, label (i.e. image data sample analysis) can specify acquisition by modes such as artificial or machine algorithms
The classification of character information belonging to sample image (individual character information image) after segmentation.
Here, acquisition mode can provide bank's card graphic by picture catching/capturing apparatus to carry out image data
Sample collection.
Such as: by the image of bank's card of image acquisition device sample, on the card face image of these cards
Include the character informations such as card number, date (image data sample).Next these image data samples can be analyzed
(segmentation and label), such as manual type analysis can be with the specific can be that using image editing tools, such as Photoshop
Numerical character region is found on these collected image data samples, chooses the range of a numerical character, such as the
One 6, then the range for covering 6 this number is taken out, separately save as an image, i.e. segmented image data sample is independent
Character information image (sample image), and its label is set as 6, that is, marks the individual character as sample image
The affiliated character information classification of information image.
Sample image after each label of acquisition is normalized normalization step S220.
Specifically such as, a sample image is amplified, corrected or is contracted to the processing of normal size.Preferably, whole samples
This image is all unified for 24*24 pixel size.
Normalized may include the image processing operations such as correction and the scaling of sample image.It can be incited somebody to action by scaling
Image size normalization.It here, can also be further by the image of acquisition (as shot) in the presence of such as direction of rotation by correction
Situation be rotated back to the situation (the character information direction that such as people get used to) of reference direction.
Here, the sample image after can also further normalizing these is stored in storage equipment appropriate, database
And/or in caching.
Characteristic extraction step S230 carries out feature extraction processing on the sample image after each normalization, carries out to it
Feature vector description.Thus to obtain the complicated vector characteristics for carrying out sample image needed for judgement identifies character information, as spy
Levy sample.
Specifically, on a standard picture (i.e. unified image size criteria), here it is possible to be carried out at normalization
The sample image of reason, such as the image of 24*24 pixel size, to the image node-by-node algorithm of each 8*8 pixel on the image its
Characteristics of image.The characteristics of image of extraction is such as: (Harris Corner Detection Algorithm is that the angle point based on gray level image mentions to Harris feature
Take characteristics of image used in algorithm, feature vector such as related with angle point), SIFT (Scale-invariant feature
Transform scale invariant feature is converted) feature (characteristics of image used in local feature algorithm in image is extracted into SIFT,
Feature vector such as unrelated to scaling, rotation, brightness change),.
One is extracted the side of above-mentioned complex characteristic vector to the image of 8*8 image element each on 24*24 sample image point by point
Formula, such as: it can be and 8*8 block of pixels 8*8 block of pixels adjacent thereto is constituted a big image block such as 16*16's
Big block carries out；The big block can choose such as 36 most strong dimensional feature vector descriptions；Meanwhile the block can be in standard picture
On the upper image such as the 24*24, according to the step-length of 8 pixels, respectively in lateral, longitudinal movement, the 24*24 is finally directed to after combination
The sample image of standard, the character pattern of the entire 24*24 of available high dimensional feature vector description.
Node-by-node algorithm, such as: lateral, longitudinal mobile by 8 pixel step lengths respectively, each block of pixels (block) has 16*16
Pixel, so transverse direction, longitudinal feature vector for thering are 2 positions to calculate 36 dimension respectively.The total length of feature vector is 36*2*2
=144.As shown in Figure 3.
Further, which can also carry out dimensionality reduction, reduce energy after dimension
Promote the recognition speed in the identification processing procedure being described below.Dimension reduction method for example can be using the drop such as pivot analysis PCA
The algorithm of low dimensional.
Here, the characteristics of image in each sample image after extracting normalization, it is preferable that be to extract more complex figure
As feature, and as far as possible it is the constant complex characteristic vector in direction, can be used as the identification processing procedure that the application is described below
The feature samples of middle need each sample image to be used are known otherwise especially in identifying processing by target detection
Optical identification is carried out, needs to use feature samples (such as constant complexity of Harris feature, SIFT feature direction of this kind of complexity
Feature vector), to overcome influence caused by complicated card face background in identification processing procedure.
Here, these features of image are extracted, it will make in judging recognizer when identifying as feature samples
With these feature samples may further be saved in various storage equipment, caching, database etc..
Based on aforesaid way, the application is not necessarily to carry out card face image the segmentation of prospect, background, directly uses word to be identified
The entirety of the candidate region of the card face image of symbol calculates its feature vector as text prospect, that is, has used more more
Complicated characteristics of image all will be above traditional traditional binary segmentation side in recognition accuracy, the recognition result degree of reliability
Method segmentation prospect, background and identify the mode of prospect.
Training step S240 is believed using the feature of the sample image extracted in characteristic extraction step S230 as image
Feature samples in the identifying processing of breath, and with these feature samples training classifier, to obtain corresponding identifying processing (as known
Other algorithm) used in containing parameter data classifier, that is, identifier.
Classifier such as uses support vector machines or artificial neural network.
And classifier training, for example, the classifiers such as Training Support Vector Machines SVM or artificial neural network.
Here, the complex characteristic vector based on extraction, for example, the multiple sample images extracted in step S230
Harris feature can use Harris feature and carry out classifier training, available one classifier comprising supplemental characteristic
That is identifier.The identifier will be used at the judgement identification to character information image to be identified based on the feature samples having
Reason.
As shown in figure 4, for adoptable preferred character locating method flow in the information identifying method of complex background card face
Figure 40 0.
Coarse localization step S410, the character on the card face image in the card face (such as bank card) identified to the needs of input are believed
It ceases region and carries out coarse localization.
Here, the card face image that the needs of input identify, can be provided by picture catching/capturing apparatus has complicated back
The image in the card face (such as bank card) of scape.Furthermore, it is desirable to the card face of identification, is the card face with complex background, it is such as each
Kind bank card (credit card) etc..Traditional Optical Character Recognition system needs to divide text (character) foreground and background, also
Character information identification can only be carried out to the simple card face of background.The cards faces such as the bank card of card face background complexity, before can not dividing
Scape and background, thus, it is that can not be identified using traditional optical character recognition system to information (text, character etc.) thereon
's.
Specifically, first carrying out the positioning of the rough character information band of position to the image in the card face of background complexity.It can
To use posterior infromation, such as international standard ISO7810, ISO7816 of bank's card etc., carrys out rough positioning and known
Character position region on the image of other card face, such as: position the transverse direction of the first character (card number), longitudinal positional locations (or
Ultimate character/card number transverse direction, longitudinal positional locations).On the basis of this band of position, then carry out it is following accurate fixed
Position.
It is accurately positioned step S420 (or being accurately positioned step), the character information region based on above-mentioned coarse localization, into
The accurate character information zone location of row.
Character information on these card faces be a series of characters to be identified in the image in card face to be identified (as number,
Card number, text, letter, other symbols etc.) information.
The kinds of characters region position that can be obtained on the image of card face to be identified is accurately positioned, and then navigates to each
Character information position, moreover it is possible to obtain each character information region and obtain a series of corresponding single character informations
Image.So that the character information image of these the single character informations detected can be utilized in identifying processing.
Specifically, near the initial character position in the character information region of above-mentioned coarse localization, sliding identifier is examined
It surveys, using peak response position when being detected using the identifier as the position where final first place character (such as: card number, number)
It sets.
Wherein, which preferably, can be the classifiers such as support vector machines and artificial neural network.
Wherein, detection preferably, can (the first character of such as coarse localization be substantially from the character information region of coarse localization
Region) middle position start, then according to the number in images match shown in Fig. 5 (identifier detection) glide direction precedence diagram
Shown sequence is detected each position using identifier (first slided in the position detection of 1 instruction, again up and down
Move the position detection to 2 instructions, then slide into the position ... ... of 3 instructions until 9 positions indicated), when identifier detects
When some position, output stops detection search when being greater than some threshold value, and exports the position of initial character positioning.It can be reduced in this way
It detects number, improve locating speed.
Same way, it is also possible to by the above-mentioned concrete mode that initial character is accurately positioned, to the last one (end) character
Position is accurately positioned.
Also, it will be evenly distributed between initial character and end character according to 24*24 pixel size, multiple centres can be obtained
Character information region.In this way, sliding along the transverse direction the identifier, can gradually obtain between initial character and end character
Each character information position.The character information of these different locations also can be positioned accurately.
These character information positions have been accurately positioned it, and then by the detection of identifier (matching of image), also energy
Corresponding each character information image of each character information region is obtained, thus, by the one of initial character to end character
Series of characters information image is obtained.
It is similar, on complex background card face such as other character information regions on bank card card face, such as effectively
It date, name, card type mark etc., can be also accurately positioned using aforesaid way, and can go to obtain these based on this
Character information image where different location.
As shown in fig. 6, the judgement for adoptable preferred character information in the information identifying method of complex background card face is known
Other method flow diagram 600.
Receiving step S610 is received and is obtained character information image according to location character information area.
The image in the card face to be identified of input can obtain a series of character information locations after localization process
Domain even gets a series of corresponding character information images.Receive these character information images to be identified.
Preferably, using the accurate positionin character information location of the identifier detection in character information localization method 400
The mode in domain obtains a series of character information image of single character informations.Received is these character information images.
Characteristic extraction step S620 carries out the extraction of characteristics of image based on obtained character information image.
The image feature extraction procedure, reference can be made to the feature extraction mode in preprocess method 200, analyzes character information
A series of each of character information images (image of such as single character information) obtained in localization method 400 extracts
The constant complex characteristic vector in direction (Harris feature, SIFT feature etc.) out.
Judge identification step S630, the complex characteristic vector based on extraction, by these characteristics of image (complex characteristic vector)
It is sent into identifier and is identified.
Preferably, as used the identifier obtained after complex characteristic vector sample training in preprocess method 200, sentenced
Disconnected identification, determines corresponding character information in single character information image to be identified.Identification judgement such as is carried out to card number, then
Judge that some contains whether the image of a numerical character (having an image for the character information of picture digital " 1 ") belongs to this number
(indicating whether digital " 1 ").
Here, due to the complex characteristic vector constant using direction, optical character information is carried out by object detection method
It identifies (i.e. recognizer), because without being influenced by complicated card face background.
Here, the different classifiers of utilization train the classification of obtained identifier to judge number difference.Use two classes point
Class device, such as support vector machines, to digital character information carry out classification judge when, numerical character hum pattern to be identified
It is that (10 subseries could judge which belongs to which number to determine as needing to judge by the identifier of 10 numerical characters
Number).And use multi classifier, such as artificial neural network, then to digital character information carry out classification judge when, this is to be identified
Numerical character information image do a subseries it may determine that come out, i.e., judged by 1 identifier.
According to an embodiment of the present application, the information for corresponding to a kind of complex background card face of the recognition methods is also provided
Identifying system 700.A kind of identifying system structural block diagram of the information in complex background card face as shown in Figure 7.
Preferably, which includes: input unit 710, inputs the card face figure to be identified with complex background card
Picture；Positioning unit 720, input unit 710 input the card face image on, location of concern information/character information to be identified；Word
Symbol information judges recognition unit 730, positions the character information affiliated area on the card face image based on positioning unit 720, treats
Character information in the character information image of identification is identified；Authentication unit 740, to judging what recognition unit 730 identified
As a result it is verified；Pretreatment unit 750, feature samples are obtained ahead of time and using being obtained after feature samples training classifier
Identifier (classifier with series of parameters data), to be applied to judge the character information figure to be identified of recognition unit 730
As feature extraction and the judgement of character information identify.
Input unit 710, it is preferable that usually using image capture device such as camera, camera, scanner etc. obtain to
The card face image of the object (such as bank card) of identification.On these card face images with complex background, has and various need typing
Information, mainly character information, comprising: number, letter, text, other various characters.
Positioning unit 720 realizes the processing such as step S120.It preferably, may include coarse localization unit 721 and accurate
Positioning unit 722, to realize the processing of character information localization method 400.Preferably, coarse localization unit 721 and accurate positionin
Unit 722 realizes the processing of step S410 and step S420 respectively.
Character information judges recognition unit 730, realizes the processing of step S130.It preferably, may include receiving image letter
Interest statement member 731, fisrt feature extraction unit 732 judge recognition unit 733, realize the processing for judging recognition methods 600.It receives
Image information units 731 obtain and a series of obtained word information symbol images, such as handling using method 400 are accurately positioned
To character information image (i.e. the processing of the step S610 of method 600)；Fisrt feature extraction unit 732 extracts characteristics of image (step
The processing of rapid S620, the processing of the image feature extraction procedure S230 in reference method 200) to extract complex characteristic vector；
Judge recognition unit 733, the spy that the identifier and step S230 obtained using the step S240 training in method 200 is extracted
Sample is levied, the characteristics of image extracted based on feature extraction unit 732 carries out classification judgement, to identify belonging to character picture information
Character information (processing of step S630).
Authentication unit 740, it is preferable that realize the processing of verification step S140.It such as, can be with for bank card card
What ten checking algorithm of the Luhn mould verification generallyd use using bank card number coding for example judged to identify in recognition unit 733
Card number is to confirm final recognition result.It verifies and successfully then exports final result.Error message can also be provided when verification failure.
Pretreatment unit 750, it is preferable that realize the processing of preprocess method 200.It preferably, may include: acquisition unit
751 with realize step S210 processing, normalization unit 752 with realize processing, the second feature extraction unit 753 of step S220 with
Processing and the training unit 754 of step S230 is realized to realize the processing of step S240.
Since the function that the system of the present embodiment is realized essentially corresponds to earlier figures 1 to embodiment of the method shown in fig. 6,
Therefore not detailed place in the description of the present embodiment, it may refer to the related description in previous embodiment, this will not be repeated here.
The application uses the constant complex characteristic vector in direction, and carries out optical character knowledge by the method for target detection
Not, influence caused by the background of complicated card face is overcome, also, the recognition result degree of reliability gets a promotion, in the card of complex background
It is applied in the identification of face information, correctness accuracy is high.
In a typical configuration, calculating equipment includes one or more processors (CPU), input/output interface, net
Network interface and memory.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 (flash RAM).Memory is computer-readable Jie
The example of matter.
Computer-readable medium includes permanent and non-permanent, removable and non-removable media can be by any method
Or technology come realize information store.Information can be computer readable instructions, data structure, the module of program or other data.
The example of the storage medium of computer includes, but are not limited to phase change memory (PRAM), static random access memory (SRAM), moves
State random access memory (DRAM), other kinds of random access memory (RAM), read-only memory (ROM), electric erasable
Programmable read only memory (EEPROM), flash memory or other memory techniques, read-only disc read only memory (CD-ROM) (CD-ROM),
Digital versatile disc (DVD) or other optical storage, magnetic cassettes, tape magnetic disk storage or other magnetic storage devices
Or any other non-transmission medium, can be used for storage can be accessed by a computing device information.As defined in this article, it calculates
Machine readable medium does not include temporary computer readable media (transitory media), such as the data-signal and carrier wave of modulation.
Each embodiment in this specification is generally described in a progressive manner, the highlights of each of the examples are
With the difference of other embodiments, the same or similar parts between the embodiments can be referred to each other.
The application can describe in the general context of computer-executable instructions executed by a computer, such as program
Module or unit.Generally, program module or unit may include executing particular task or realization particular abstract data type
Routine, programs, objects, component, data structure etc..In general, program module or unit can be by softwares, hardware or both
Combination realize.The application can also be practiced in a distributed computing environment, in these distributed computing environments, by passing through
Communication network and connected remote processing devices execute task.In a distributed computing environment, program module or unit can
To be located in the local and remote computer storage media including storage equipment.
Finally, it is to be noted that, the terms "include", "comprise" or its any other variant be intended to it is non-exclusive
Property include so that include a series of elements process, method, commodity or equipment not only include those elements, but also
Further include other elements that are not explicitly listed, or further include for this process, method, commodity or equipment it is intrinsic
Element.In the absence of more restrictions, the element limited by sentence "including a ...", it is not excluded that including described
There is also other identical elements in the process, method of element, commodity or equipment.
It should be understood by those skilled in the art that, embodiments herein can provide as method, system or computer program
Product.Therefore, complete hardware embodiment, complete software embodiment or reality combining software and hardware aspects can be used in the application
Apply the form of example.Moreover, it wherein includes the computer of computer usable program code that the application, which can be used in one or more,
The computer program implemented in usable storage medium (including but not limited to magnetic disk storage, CD-ROM, optical memory etc.) produces
The form of product.
Specific examples are used herein to illustrate the principle and implementation manner of the present application, and above embodiments are said
It is bright to be merely used to help understand the present processes and its main thought；At the same time, for those skilled in the art, foundation
The thought of the application, there will be changes in the specific implementation manner and application range, in conclusion the content of the present specification is not
It is interpreted as the limitation to the application.
Claims (18)
1. a kind of complex background card face information identifying method characterized by comprising
The character information region in the image in the card face to be identified is positioned, to obtain one or more character information figures
Picture, the corresponding individual character of each character information image, and character information image are as follows: do not carry out prospect and background segment
Card face image-region；
The extraction of complex characteristic vector is carried out to the character information image of acquisition；
According to the complex characteristic vector, the character information in the character information image is identified；
The character information region in the image of card face to be identified is positioned, to obtain one or more character information images, packet
It includes:
Coarse localization step includes the region of character, Primary Location character information in the production standard based on complex background card face
Region；
Step is accurately positioned, according to the character information region of Primary Location, using an identifier in certain direction sequence into
Row detection obtains accurately each character information position, and obtains the character information image of corresponding position.
2. the method according to claim 1, wherein
Coarse localization step further include:
When Primary Location character information region, energy coarse localization first place character information and/or ultimate character information institute are in place
It sets；
Step is accurately positioned further include:
Near the first character information and/or ultimate character information position, using the identifier by up and down
Direction sequence shiding matching to be detected, obtain the accurately position of the first character information and/or ultimate character information；
By unified image size criteria, evenly distribute true by the first character information and/or the accurate location of ultimate character information institute
Fixed character information region, to each character information region for distributing using the identifier by up and down
Direction sequence shiding matching obtains the position of accurately each character information to be detected；
Accurate position based on each character information obtains each character information image in corresponding position region.
3. the method according to claim 1, wherein the character information image to acquisition carries out complex characteristic
Vector extracts
Based on the acquired one or more character information images of positioning, size criteria is unified to the character information image；
A default block of pixels, to the image node-by-node algorithm characteristics of image of the block of pixels each of on the character information image,
To obtain the complex characteristic vector of the character information image.
4. according to the method described in claim 3, it is characterized in that, the node-by-node algorithm characteristics of image includes:
Block of pixels block of pixels adjacent thereto forms block, and the feature vector for choosing 36 dimensions to the block describes；
On the character information image by 8 pixels step-length respectively transversely, longitudinal movement, combination obtain to the character believe
Cease the feature vector description of the higher-dimension of image.
5. the method according to claim 1, wherein further include: pre-treatment step, wherein
Acquisition largely has the card face image data sample of complex background and is split and marks, and classification is marked to obtain
Character information image；
It to the character information image that classification is marked, is normalized, to obtain the character letter of unified size criteria
Cease image；
A default block of pixels, to the image node-by-node algorithm characteristics of image of the block of pixels each of on the character information image,
To obtain the complex characteristic vector of the character information image as feature samples；
Using the feature samples, classifier training is carried out, to obtain the identifier of identifying processing.
6. according to the method described in claim 5, it is characterized in that, identifying the character letter according to the complex characteristic vector
Breath image in character information include:
Based on feature samples, the character information image is sent into the identifier, judges and identifies corresponding character information.
7. according to the method described in claim 5, it is characterized in that, a block of pixels is preset, on the character information image
The image node-by-node algorithm characteristics of image of each block of pixels, using obtain the complex characteristic vector of the character information image as
Feature samples include:
Block of pixels block of pixels adjacent thereto forms block, and the feature vector for choosing 36 dimensions to the block describes；
On the character information image by 8 pixels step-length respectively transversely, longitudinal movement, combination obtain to the character believe
Cease the feature vector description of the higher-dimension of image.
8. the method according to claim 3 or 5, it is characterised in that:
The unified size criteria of character information image is that the character picture is unified for 24*24 pixel criterion image；It is preset
One block of pixels is 8*8 block of pixels；The complex characteristic vector includes the constant feature vector in direction.
9. the method according to claim 1, wherein further include:
Input step inputs the image in the card face to be identified in the way of picture catching；
Verification step verifies to confirm final recognition result the character information identified.
10. a kind of complex background card face information identification system characterized by comprising
Positioning unit positions the character information region in the image in the card face to be identified, to obtain one or more
Character information image, the corresponding individual character of each character information image, and character information image are as follows: do not carry out prospect with
The card face image-region of background segment；
Fisrt feature extraction unit carries out the extraction of complex characteristic vector to the character information image of acquisition；
Character information recognition unit identifies the character information in the character information image according to the complex characteristic vector；
Positioning unit includes:
Coarse localization unit includes the region of character, Primary Location character information in the production standard based on complex background card face
Region；
Unit is accurately positioned, according to the character information region of Primary Location, using an identifier in certain direction sequence into
Row detection obtains accurately each character information position, and obtains the character information image of corresponding position.
11. system according to claim 10, which is characterized in that
Coarse localization unit further include:
It, can coarse localization first place character information and/or ultimate character information position when Primary Location character region；
Unit is accurately positioned further include:
Near the first character information and/or ultimate character information position, using the identifier by up and down
Direction sequence shiding matching to be detected, obtain the accurately position of the first character information and/or ultimate character information；
By unified image size criteria, evenly distribute true by the first character information and/or the accurate location of ultimate character information institute
Fixed character information region, to each character information region for distributing using the identifier by up and down
Direction sequence shiding matching obtains the position of accurately each character information to be detected；
Accurate position based on each character information obtains each character information image in corresponding position region.
12. system according to claim 10, which is characterized in that fisrt feature extraction unit includes:
Based on the acquired one or more character information images of positioning, size criteria is unified to the character information image；
A default block of pixels, to the image node-by-node algorithm characteristics of image of the block of pixels each of on the character information image,
To obtain the complex characteristic vector of the character information image.
13. system according to claim 12, which is characterized in that in fisrt feature extraction unit, the node-by-node algorithm figure
As feature includes:
Block of pixels block of pixels adjacent thereto forms block, and the feature vector for choosing 36 dimensions to the block describes；
On the character information image by 8 pixels step-length respectively transversely, longitudinal movement, combination obtain to the character believe
Cease the feature vector description of the higher-dimension of image.
14. system according to claim 10, which is characterized in that further include: pretreatment unit, the pretreatment unit packet
It includes:
Acquisition unit, acquisition largely has the card face image data sample of complex background and is split and marks, to be marked
It is fired the character information image of class；
The character information image that classification is marked is normalized in normalization unit, to obtain unified big small tenon
Quasi- character information image；
Second feature extraction unit presets a block of pixels, to the image of the block of pixels each of on the character information image
Node-by-node algorithm characteristics of image, to obtain the complex characteristic vector of the character information image as feature samples；
Training unit carries out classifier training, using the feature samples to obtain the identifier of identifying processing.
15. system according to claim 14, which is characterized in that character information recognition unit includes: judgement recognition unit,
It is based on feature samples, and the character information image is sent into the identifier, judges and identifies corresponding character information.
16. system according to claim 14, which is characterized in that second feature extraction unit includes:
Block of pixels block of pixels adjacent thereto forms block, and the feature vector for choosing 36 dimensions to the block describes；
On the character information image by 8 pixels step-length respectively transversely, longitudinal movement, combination obtain to the character believe
Cease the feature vector description of the higher-dimension of image.
17. system described in 2 or 14 according to claim 1, it is characterised in that:
The unified size criteria of character information image is that the character picture is unified for 24*24 pixel criterion image；It is preset
One block of pixels is 8*8 block of pixels；The complex characteristic vector includes the constant feature vector in direction.
18. system according to claim 10, which is characterized in that further include:
Input unit inputs the image in the card face to be identified in the way of picture catching；
Authentication unit verifies to confirm final recognition result the character information identified.
Priority Applications (3)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910468824.4A CN110222687B (en) | 2013-06-03 | 2013-06-03 | Complex background card surface information identification method and system |
| CN201310216817.8A CN104217203B (en) | 2013-06-03 | 2013-06-03 | Complex background card face information identifying method and system |
| HK15103317.3A HK1202954A1 (en) | 2013-06-03 | 2015-04-01 | Method for recognizing information located on surface of card in complex background and system thereof |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201310216817.8A CN104217203B (en) | 2013-06-03 | 2013-06-03 | Complex background card face information identifying method and system |
Related Child Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910468824.4A Division CN110222687B (en) | 2013-06-03 | 2013-06-03 | Complex background card surface information identification method and system |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN104217203A CN104217203A (en) | 2014-12-17 |
| CN104217203B true CN104217203B (en) | 2019-08-23 |
Family
ID=52098668
Family Applications (2)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910468824.4A Active CN110222687B (en) | 2013-06-03 | 2013-06-03 | Complex background card surface information identification method and system |
| CN201310216817.8A Active CN104217203B (en) | 2013-06-03 | 2013-06-03 | Complex background card face information identifying method and system |
Family Applications Before (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910468824.4A Active CN110222687B (en) | 2013-06-03 | 2013-06-03 | Complex background card surface information identification method and system |
Country Status (2)
| Country | Link |
|---|---|
| CN (2) | CN110222687B (en) |
| HK (1) | HK1202954A1 (en) |
Families Citing this family (16)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104809157B (en) * | 2015-03-25 | 2018-05-18 | 小米科技有限责任公司 | Number identification method and device |
| CN106469293A (en) * | 2015-08-21 | 2017-03-01 | 上海羽视澄蓝信息科技有限公司 | The method and system of quick detection target |
| KR101824600B1 (en) * | 2015-09-08 | 2018-03-14 | 에스케이플래닛 주식회사 | A system and method credit card number and expiration date recognition using a terminal device |
| CN106527674A (en) * | 2015-09-14 | 2017-03-22 | 上海羽视澄蓝信息科技有限公司 | Human-computer interaction method, equipment and system for vehicle-mounted monocular camera |
| JP6804292B2 (en) * | 2016-12-28 | 2020-12-23 | オムロンヘルスケア株式会社 | Terminal equipment |
| CN108876716B (en) * | 2017-05-11 | 2020-09-11 | Tcl科技集团股份有限公司 | Super-resolution reconstruction method and device |
| CN107256384A (en) * | 2017-05-22 | 2017-10-17 | 汕头大学 | A kind of card recognition and method of counting based on image and signal transacting |
| 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 |
| CN108596180A (en) * | 2018-04-09 | 2018-09-28 | 深圳市腾讯网络信息技术有限公司 | Parameter identification, the training method of parameter identification model and device in image |
| CN108764230A (en) * | 2018-05-30 | 2018-11-06 | 上海建桥学院 | A kind of bank's card number automatic identifying method based on convolutional neural networks |
| CN110619325B (en) * | 2018-06-20 | 2024-03-08 | 北京搜狗科技发展有限公司 | Text recognition method and device |
| CN110796146A (en) * | 2019-10-11 | 2020-02-14 | 上海上湖信息技术有限公司 | Bank card number identification method, model training method and device |
| CN111242932A (en) * | 2020-01-15 | 2020-06-05 | 中国银联股份有限公司 | Card style checking method and device |
| CN111814789A (en) * | 2020-07-15 | 2020-10-23 | 中国建设银行股份有限公司 | Card number detection method, device, equipment and storage medium |
| CN111860487B (en) * | 2020-07-28 | 2022-08-19 | 天津恒达文博科技股份有限公司 | Inscription marking detection and recognition system based on deep neural network |
| CN114049646A (en) * | 2021-11-29 | 2022-02-15 | 中国平安人寿保险股份有限公司 | Bank card identification method and device, computer equipment and storage medium |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN201054150Y (en) * | 2006-12-15 | 2008-04-30 | 华南理工大学 | An automatic recognition system for post envelope |
| CN101373519A (en) * | 2007-08-20 | 2009-02-25 | 富士通株式会社 | Device and method for recognizing character |
| CN102147858A (en) * | 2011-03-31 | 2011-08-10 | 重庆大学 | License plate character identification method |
| CN102163287A (en) * | 2011-03-28 | 2011-08-24 | 北京邮电大学 | Method for recognizing characters of licence plate based on Haar-like feature and support vector machine |
| CN103020647A (en) * | 2013-01-08 | 2013-04-03 | 西安电子科技大学 | Image classification method based on hierarchical SIFT (scale-invariant feature transform) features and sparse coding |
Family Cites Families (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| KR100601957B1 (en) * | 2004-07-07 | 2006-07-14 | 삼성전자주식회사 | Apparatus for and method for determining image correspondence, apparatus and method for image correction therefor |
| US7570816B2 (en) * | 2005-03-31 | 2009-08-04 | Microsoft Corporation | Systems and methods for detecting text |
| CN101038686B (en) * | 2007-01-10 | 2010-05-19 | 北京航空航天大学 | Method for recognizing machine-readable travel certificate |
| CN102324044A (en) * | 2011-09-09 | 2012-01-18 | 上海合合信息科技发展有限公司 | Card information acquisition method and system |
-
2013
- 2013-06-03 CN CN201910468824.4A patent/CN110222687B/en active Active
- 2013-06-03 CN CN201310216817.8A patent/CN104217203B/en active Active
-
2015
- 2015-04-01 HK HK15103317.3A patent/HK1202954A1/en unknown
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN201054150Y (en) * | 2006-12-15 | 2008-04-30 | 华南理工大学 | An automatic recognition system for post envelope |
| CN101373519A (en) * | 2007-08-20 | 2009-02-25 | 富士通株式会社 | Device and method for recognizing character |
| CN102163287A (en) * | 2011-03-28 | 2011-08-24 | 北京邮电大学 | Method for recognizing characters of licence plate based on Haar-like feature and support vector machine |
| CN102147858A (en) * | 2011-03-31 | 2011-08-10 | 重庆大学 | License plate character identification method |
| CN103020647A (en) * | 2013-01-08 | 2013-04-03 | 西安电子科技大学 | Image classification method based on hierarchical SIFT (scale-invariant feature transform) features and sparse coding |
Also Published As
| Publication number | Publication date |
|---|---|
| CN110222687A (en) | 2019-09-10 |
| CN110222687B (en) | 2021-02-26 |
| HK1202954A1 (en) | 2015-10-09 |
| CN104217203A (en) | 2014-12-17 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN104217203B (en) | Complex background card face information identifying method and system | |
| CN106156766B (en) | Method and device for generating text line classifier | |
| US8744196B2 (en) | Automatic recognition of images | |
| US20190213408A1 (en) | Document authenticity determination | |
| CN108205673B (en) | Method and system for container code identification | |
| WO2019169532A1 (en) | License plate recognition method and cloud system | |
| US8494273B2 (en) | Adaptive optical character recognition on a document with distorted characters | |
| US20190362193A1 (en) | Eyeglass positioning method, apparatus and storage medium | |
| CN108717543B (en) | Invoice identification method and device and computer storage medium | |
| US20140334733A1 (en) | Efficient image matching for large sets of images | |
| CN105512657B (en) | Character identifying method and equipment | |
| CN111079571A (en) | Identification card information identification and edge detection model training method and device | |
| CN109389115B (en) | Text recognition method, device, storage medium and computer equipment | |
| CN106874901B (en) | Driving license identification method and device | |
| CN109766885A (en) | A kind of character detecting method, device, electronic equipment and storage medium | |
| CN111242124A (en) | Certificate classification method, device and equipment | |
| US10423817B2 (en) | Latent fingerprint ridge flow map improvement | |
| CN108154132A (en) | A kind of identity card text extraction method, system and equipment and storage medium | |
| CN112001200A (en) | Identification code identification method, device, equipment, storage medium and system | |
| JP2019079347A (en) | Character estimation system, character estimation method, and character estimation program | |
| CN113011426A (en) | Method and device for identifying certificate | |
| He et al. | Aggregating local context for accurate scene text detection | |
| CN114463767A (en) | Credit card identification method, device, computer equipment and storage medium | |
| CN112001389A (en) | Method and device for identifying text information in multi-scene video and electronic equipment | |
| KR102206604B1 (en) | Apparatus and method for recognizing character |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| REG | Reference to a national code |
Ref country code: HK Ref legal event code: DE Ref document number: 1202954 Country of ref document: HK |
|
| GR01 | Patent grant | ||
| GR01 | Patent grant |