CN106599772A - Living body authentication method, identity authentication method and device - Google Patents

Living body authentication method, identity authentication method and device Download PDF

Info

Publication number
CN106599772A
CN106599772A CN201610927708.0A CN201610927708A CN106599772A CN 106599772 A CN106599772 A CN 106599772A CN 201610927708 A CN201610927708 A CN 201610927708A CN 106599772 A CN106599772 A CN 106599772A
Authority
CN
China
Prior art keywords
live body
image
verified
face
identity card
Prior art date
Application number
CN201610927708.0A
Other languages
Chinese (zh)
Inventor
何涛
曹志敏
Original Assignee
北京旷视科技有限公司
北京小孔科技有限公司
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by 北京旷视科技有限公司, 北京小孔科技有限公司 filed Critical 北京旷视科技有限公司
Priority to CN201610927708.0A priority Critical patent/CN106599772A/en
Publication of CN106599772A publication Critical patent/CN106599772A/en

Links

Classifications

    • GPHYSICS
    • G06COMPUTING; CALCULATING; COUNTING
    • G06KRECOGNITION OF DATA; PRESENTATION OF DATA; RECORD CARRIERS; HANDLING RECORD CARRIERS
    • G06K9/00Methods or arrangements for reading or recognising printed or written characters or for recognising patterns, e.g. fingerprints
    • G06K9/00221Acquiring or recognising human faces, facial parts, facial sketches, facial expressions
    • G06K9/00228Detection; Localisation; Normalisation
    • GPHYSICS
    • G06COMPUTING; CALCULATING; COUNTING
    • G06KRECOGNITION OF DATA; PRESENTATION OF DATA; RECORD CARRIERS; HANDLING RECORD CARRIERS
    • G06K9/00Methods or arrangements for reading or recognising printed or written characters or for recognising patterns, e.g. fingerprints
    • G06K9/00885Biometric patterns not provided for under G06K9/00006, G06K9/00154, G06K9/00335, G06K9/00362, G06K9/00597; Biometric specific functions not specific to the kind of biometric
    • G06K9/00899Spoof detection
    • G06K9/00906Detection of body part being alive

Abstract

The invention provides a living body authentication method, an identity authentication method and a device. The living body authentication method comprises the steps of randomly generating a living body action instruction, wherein the living body action instruction is used for indicating a to-be-authenticated object to hold an identity card in hand and execute a corresponding living body action; colleting the image of the to-be-authenticated object when the to-be-authenticated object is executing the living body action so as to obtain a to-be-authenticated image; and based on the to-be-authenticated image, judging whether the to-be-authenticated object passes the living body authentication or not. Based on the above living body authentication method, the device thereof, the identity authentication method and the device thereof, the living body authentication is conducted according to the information of identity cards. Therefore, the living body authentication accuracy is improved.

Description

Live body verification method and device and identity identifying method and device

Technical field

The present invention relates to field of identity authentication, relates more specifically to a kind of live body verification method and device and a kind of identity is recognized Card method and apparatus.

Background technology

In current Above-the-line, the long-range identity for differentiating operator is a kind of common requirement, such as the real name of mobile phone Certification processed, account system of real name certification of finance activities on line etc..Traditional identity identifying method includes:User shines identity card Piece and auto heterodyne photo upload carry out artificial treatment to using backstage;Or user input identification card number, then will shoot the video certainly (or Still image) upload to and carry out artificial treatment using backstage.With the application of the biometric authentication technology such as recognition of face, it is new based on The on-line identification method of recognition of face shortens process time.In order to add reliability and the safety of strong identity authentication, new body Identity authentication method employs live body verification technique.For example, user input identification card number or upload ID Card Image, then in the face of figure As harvester (for example, mobile phone camera) carries out live body checking (doing required movement, say a section etc.), verified by live body After (confirm user be live body), then the body that data and the be previously stored and user such as the image that user is uploaded upload The corresponding actual data of part information is contrasted to verify the identity of user.

Current live body verification technique existing defects.For example, malicious attacker can utilize computer graphics techniques (CG) Software combines video of the photo of the real user stolen required for synthesize the live body checking of deception face, such that it is able to take advantage of Deceive current live body checking system.The defect of above-mentioned live body checking can cause the safety of authentication to be damaged.

The content of the invention

The present invention is proposed in view of the problems referred to above.The invention provides a kind of live body verification method and device and one kind Identity identifying method and device.

According to an aspect of the present invention, there is provided a kind of live body verification method.The live body verification method includes:Random generation is lived Body action command, the live body action command is used to indicate that the hand-held identity card of object to be verified performs corresponding live body action;It is real When gather the image that the object to be verified performs the live body action, to obtain image to be verified;And based on described to be tested Card image determines whether the object to be verified is verified by live body.

Exemplarily, it is described to determine whether the object to be verified is wrapped by live body checking based on the image to be verified Include:Detect the face and identity card in the image to be verified;Perform multiple live bodies and judge operation, wherein, multiple live bodies judge Operation includes that the first live body judges that operation, the second live body judge that operation and the 3rd live body judge operation, wherein, first live body Judge that operation includes:Judge whether the face for detecting belongs to live body based on the image to be verified, second live body judges Operation includes:Judge whether the identity card for detecting belongs to live body based on the image to be verified, the 3rd live body judges behaviour Work includes:Based on the image synthesis to be verified judge described in the face that detects and the identity card that detects be on the whole It is no to belong to live body;And each live body judged in operation according to the plurality of live body is judged described in the judged result determination of operation Whether object to be verified is verified by live body, if arbitrary live body judges that the judged result for operating is no, it is determined that described to be tested Card object otherwise determines that the object to be verified is verified by live body not by live body checking.

Exemplarily, after the face and identity card in the detection image to be verified, the live body authentication Method also includes:The only facial image to be verified comprising the face for detecting is extracted from the image to be verified and is only included The ID Card Image to be verified of the identity card for detecting;Also, first live body judges that operation includes:Treated based on described Whether the face detected described in checking facial image judgement belongs to live body, and second live body judges that operation includes:Based on institute State whether the identity card detected described in ID Card Image judgement to be verified belongs to live body, the 3rd live body judges operation bag Include:Based on the face detected described in the facial image to be verified and the ID Card Image comprehensive descision to be verified and described Whether the identity card for detecting belongs to live body on the whole.

Exemplarily, it is described based on described in the facial image to be verified and the ID Card Image comprehensive descision to be verified Whether the face and the identity card for detecting for detecting belongs on the whole live body includes:By the facial image to be verified and The first convolutional neural networks that the ID Card Image to be verified input is trained, to obtain the face for detecting and described The identity card for detecting belongs on the whole the probability of live body;And the face that detects according to the determine the probability and described Whether the identity card for detecting belongs to live body on the whole.

Exemplarily, the live body verification method also includes:Training data is obtained, the training data includes positive sample figure Picture and negative sample image, the positive sample image includes real human face and real identity card, and the negative sample image includes falseness Face and real identity card;The only positive sample facial image comprising face is extracted from the positive sample image and identity is only included The positive sample ID Card Image of card;The only negative sample facial image comprising face is extracted from the negative sample image and is only included The negative sample ID Card Image of identity card;And with the positive sample facial image and the positive sample ID Card Image as positive sample This, and with the negative sample facial image and the negative sample ID Card Image as negative sample, carry out neural metwork training with Obtain first convolutional neural networks.

Exemplarily, it is described extract from the positive sample image only comprising face positive sample facial image and only wrap Before positive sample ID Card Image containing identity card, the live body verification method also includes:Calculate in the positive sample image Face zoom to positive sample scaling needed for default size, and according to the positive sample scaling to the positive sample Image is zoomed in and out;The only negative sample facial image comprising face is extracted from the negative sample image and only comprising body described Before the negative sample ID Card Image of part card, the live body verification method also includes:Calculate the people in the negative sample image Face zooms to the negative sample scaling needed for the default size, and according to the negative sample scaling to the negative sample Image is zoomed in and out;The only face figure to be verified comprising the face for detecting is extracted from the image to be verified described Before picture and the only ID Card Image to be verified comprising the identity card for detecting, the live body verification method also includes:Meter The face for detecting is zoomed to the image scaling ratio to be verified needed for the default size for calculation, and according to described to be tested Card image scaling ratio is zoomed in and out to the image to be verified.

Exemplarily, it is described that whether live body bag is belonged to based on the face detected described in the facial image judgement to be verified Include:The second convolutional neural networks that the facial image to be verified input is trained, to judge that the face for detecting is It is no to belong to live body.

Exemplarily, it is described whether to belong to living based on the identity card detected described in the ID Card Image judgement to be verified Body includes:The 3rd convolutional neural networks that the ID Card Image to be verified input is trained, to judge described detecting Whether identity card belongs to live body.

Exemplarily, after the face and identity card in the detection image to be verified, the live body authentication Method also includes:If any face is not detected by the image to be verified or any identity card is not detected by, Output re-executes the prompting of live body checking.

Exemplarily, the first live body of the execution judges that operation, the second live body judge that operation and the 3rd live body judge operation Including:First live body is performed according to the order for arranging judge that operation, second live body judge operation and the described 3rd Live body judges operation, if arbitrary live body judges that the result for operating is no, stops the subsequent live body of execution and judges operation.

Exemplarily, the plurality of live body judges that operation also includes that the 4th live body judges operation, wherein the 4th live body Judge that operation includes:Based on the image to be verified judge action that the identity card for detecting performs whether with the live body action Instruction matches.

Exemplarily, the image to be verified is video, and the 4th live body judges that operation is based in the video extremely Few two frames are carried out, and first live body judges that operation, second live body judge that operation and the 3rd live body judge operation base An at least frame at least two frames is carried out.

Exemplarily, the live body action is included in while blocking face with identity card and overturns and/or translate identity card.

According to a further aspect of the invention, there is provided a kind of identity identifying method, including above-mentioned live body verification method, wherein, The identity identifying method also includes:In the case of it is determined that the object to be verified is verified by live body, judge from described to treat Whether the face on identity card detected in authentication image is consistent with the face detected from the image to be verified.

According to a further aspect of the invention, there is provided a kind of live body verifies device, including:Directive generation module, for random Live body action command is generated, the live body action command is used to indicate that the hand-held identity card of object to be verified performs corresponding live body and moves Make;Image to be verified obtains module, for the image that object to be verified described in Real-time Collection performs the live body action, to obtain Image to be verified;And determining module is verified, for whether determining the object to be verified based on the image to be verified Verified by live body.

Exemplarily, the determining module that is verified includes:Detection sub-module, for detecting the image to be verified in Face and identity card;Live body judging submodule, for performing multiple live bodies operation is judged, wherein, the live body judges submodule Block includes:First live body judging unit, for performing the first live body operation is judged, wherein, the first live body judging unit bag Face judgment sub-unit is included, for judging whether the face for detecting belongs to live body based on the image to be verified;Second live body Judging unit, for performing the second live body operation is judged, wherein, the second live body judging unit includes that identity card judges that son is single Unit, for judging whether the identity card for detecting belongs to live body based on the image to be verified;3rd live body judging unit, is used for Perform the 3rd live body and judge operation, wherein, the 3rd live body judging unit includes comprehensive descision subelement, for based on described Whether the face and the identity card for detecting detected described in image synthesis judgement to be verified belongs to live body on the whole;And Determination sub-module is verified, the judged result for each the live body judging unit in the live body judging submodule is true Whether the fixed object to be verified is verified by live body, if the judged result of arbitrary live body judging unit is no, it is determined that institute State object to be verified not verify by live body, otherwise determine that the object to be verified is verified by live body.

Exemplarily, the live body checking device also includes:First image zooming-out module, for from the image to be verified Middle extraction only includes the facial image to be verified of the face for detecting and only includes the to be tested of the identity card for detecting Card ID Card Image;Also, the face judgment sub-unit includes face determination component, for based on the face figure to be verified Whether the face detected as described in judging belongs to live body, and the identity card judgment sub-unit includes identity card determination component, uses In live body whether is belonged to based on the identity card detected described in the ID Card Image judgement to be verified, comprehensive descision is single Unit includes comprehensive descision component, for based on the facial image to be verified and the ID Card Image comprehensive descision institute to be verified State the face that detects and whether the identity card for detecting belongs to live body on the whole.

Exemplarily, the comprehensive descision component includes:First input sub-component, for by the facial image to be verified The first convolutional neural networks trained with the ID Card Image input to be verified, to obtain the face for detecting and institute State the probability that the identity card for detecting belongs on the whole live body;And live body determines sub-component, for according to the determine the probability Whether the face for detecting and the identity card for detecting belong to live body on the whole.

Exemplarily, the live body checking device also includes:Training data acquisition module, for obtaining training data, institute Training data is stated including positive sample image and negative sample image, the positive sample image includes real human face and real identity card, The negative sample image includes false face and real identity card;Second image zooming-out module, for from the positive sample image The middle extraction only positive sample facial image comprising face and only the positive sample ID Card Image comprising identity card;3rd image zooming-out Module, for extracting the only negative sample facial image comprising face and the only negative sample comprising identity card from the negative sample image This ID Card Image;And training module, for the positive sample facial image and the positive sample ID Card Image as just Sample, and with the negative sample facial image and the negative sample ID Card Image as negative sample, carry out neural metwork training To obtain first convolutional neural networks.

Exemplarily, the live body checking device also includes:First Zoom module, in the second image zooming-out mould Block extracts the only positive sample facial image comprising face and only the positive sample identity comprising identity card from the positive sample image Before card image, the face in the positive sample image is zoomed to the positive sample scaling needed for default size for calculating, and The positive sample image is zoomed in and out according to the positive sample scaling;Second Zoom module, in the 3rd figure As extraction module extracts the only negative sample facial image comprising face and bearing only comprising identity card from the negative sample image Before sample identity card image, the face in the negative sample image is zoomed to the negative sample needed for the default size for calculating Scaling, and the negative sample image is zoomed in and out according to the negative sample scaling;And the 3rd Zoom module, use Only include the to be verified of the face for detecting in extracting from the image to be verified in described first image extraction module Before facial image and the only ID Card Image to be verified comprising the identity card for detecting, calculate the people for detecting Face zooms to the image scaling ratio to be verified needed for the default size, and according to the image scaling ratio to be verified to institute State image to be verified to zoom in and out.

Exemplarily, the face determination component includes:Second input sub-component, for by the facial image to be verified The second convolutional neural networks that input is trained, to judge whether the face for detecting belongs to live body.

Exemplarily, the identity card determination component includes:3rd input sub-component, for by the identity card to be verified The 3rd convolutional neural networks that image input is trained, to judge whether the identity card for detecting belongs to live body.

Exemplarily, the live body checking device also includes:Prompting output module, if in the image to be verified In be not detected by any face or be not detected by any identity card, then output re-executes the prompting of live body checking.

Exemplarily, the live body judging submodule also includes the 4th live body judging unit, sentences for performing the 4th live body Disconnected operation, wherein the 4th live body judging unit includes identity card action judgment sub-unit, for based on the figure to be verified As whether the action that the identity card for judging to detect is performed matches with the live body action command.

Exemplarily, the image to be verified is video, and the 4th live body judging unit is based in the video extremely Few two frames perform the 4th live body and judge operation, the first live body judging unit, the second live body judging unit and institute State the 3rd live body judging unit based at least frame at least two frames perform respectively the first live body judge operation, described the Two live bodies judge that operation and the 3rd live body judge operation.Exemplarily, the live body action is included in and is blocked with identity card Identity card is overturn and/or translated while face.

According to a further aspect of the invention, there is provided a kind of identification authentication system, including above-mentioned live body checking device, wherein, The identification authentication system also includes face concordance judge module, for treating in described being verified described in determining module determination In the case that identifying object is by live body checking, judge face on the identity card that detects from the image to be verified and from Whether the face detected in the image to be verified is consistent.

Live body verification method according to embodiments of the present invention and device and identity identifying method and device, it is to be tested due to gathering The hand-held identity card of card object performs the image of live body action and carries out live body checking based on the image of collection, therefore in live body checking During can judge whether object to be verified belongs to live body with reference to the information that identity card brings, can so improve live body and test The accuracy of card.

Description of the drawings

The embodiment of the present invention is described in more detail by combining accompanying drawing, above-mentioned and other purposes of the present invention, Feature and advantage will be apparent from.Accompanying drawing is used for providing further understanding the embodiment of the present invention, and constitutes explanation A part for book, is used to explain the present invention together with the embodiment of the present invention, is not construed as limiting the invention.In the accompanying drawings, Identical reference number typically represents same parts or step.

Fig. 1 illustrates that the exemplary electronic device for realizing live body verification method according to embodiments of the present invention and device is shown Meaning property block diagram;

Fig. 2 illustrates the indicative flowchart of live body verification method according to an embodiment of the invention;

Whether Fig. 3 illustrates and according to an embodiment of the invention determines object to be verified by live body based on image to be verified The indicative flowchart of the step of checking;

Fig. 4 illustrates the schematic network structure of the first convolutional neural networks according to an embodiment of the invention;

Fig. 5 illustrates the schematic flow of the training step of the first convolutional neural networks according to an embodiment of the invention Figure;

Fig. 6 illustrates that live body according to an embodiment of the invention verifies the schematic block diagram of device;And

Fig. 7 illustrates the schematic block diagram of live body checking system according to an embodiment of the invention.

Specific embodiment

In order that the object, technical solutions and advantages of the present invention become apparent from, root is described in detail below with reference to accompanying drawings According to the example embodiment of the present invention.Obviously, described embodiment is only a part of embodiment of the present invention, rather than this Bright whole embodiments, it should be appreciated that the present invention is not limited by example embodiment described herein.Described in the present invention The embodiment of the present invention, those skilled in the art's all other embodiment resulting in the case where creative work is not paid All should fall under the scope of the present invention.

Can be seen from the foregoing, in current authentication application, live body proof procedure and follow-up identity were contrasted Journey is carried out separately.Specifically, live body is verified user plane and makes finger to the image collecting device of such as mobile phone camera Determine interactive action to realize, live body checking using image collecting device gathered comprising user provide face video (or Still image) judge whether face belongs to live body, the ID card information of user is not accounted in this process.Therefore, if attacked The person of hitting may cheat live body checking system using one facial image by attacker of CG software process qualities, while attacker may be used also To steal or using software engineering synthesis by the ID Card Image of attacker, for carrying out subsequent identity contrast.Although identity Card image and facial image are all to forge, but because the two really belongs to same person, as long as therefore having passed through live body and having tested Card, follow-up identity contrast just can be smoothed out so that attacker can pass through authentication.Therefore, in order to ensure The safety of authentication, needs the accuracy for improving live body checking as far as possible, it is to avoid attacker passes through live body using false face Checking.

In order to solve problem as described above, the embodiment of the present invention proposes a kind of live body verification method and device.The method The intersection information that can be based on face and identity card with device carries out live body checking, improves the accuracy of live body checking, additionally, should Method and apparatus can avoid above-mentioned causing because ID card information and face information are used separately when authentication is applied to Problem.It may be noted that present invention could apply in any scene for needing carry out live body checking, including but not limited to finance is led The system of real name certification in domain etc..

First, with reference to Fig. 1 describing the example for realizing live body verification method according to embodiments of the present invention and device Electronic equipment 100.

As shown in figure 1, electronic equipment 100 includes one or more processors 102, one or more storage devices 104, defeated Enter device 106, output device 108 and image collecting device 110, these components are by bus system 112 and/or other forms Bindiny mechanism's (not shown) interconnection.It should be noted that the component and structure of the electronic equipment 100 shown in Fig. 1 are exemplary, and Nonrestrictive, as needed, the electronic equipment can also have other assemblies and structure.

The processor 102 can be CPU (CPU) or perform with data-handling capacity and/or instruction The processing unit of the other forms of ability, and it is desired to perform to control other components in the electronic equipment 100 Function.

The storage device 104 can include one or more computer programs, and the computer program can With including various forms of computer-readable recording mediums, such as volatile memory and/or nonvolatile memory.It is described easy The property lost memorizer can for example include random access memory (RAM) and/or cache memory (cache) etc..It is described non- Volatile memory can for example include read only memory (ROM), hard disk, flash memory etc..In the computer-readable recording medium On can store one or more computer program instructions, processor 102 can run described program instruction, to realize hereafter institute The client functionality (realized by processor) in the embodiment of the present invention stated and/or other desired functions.In the meter Various application programs and various data can also be stored in calculation machine readable storage medium storing program for executing, such as application program use and/or Various data for producing etc..

The input equipment 106 can be device of the user for input instruction, and can include keyboard, mouse, wheat One or more in gram wind and touch screen etc..

The output device 108 can export various information (such as image and/or sound) to outside (such as user), and And can be including one or more in display, speaker etc..

Described image harvester 110 can gather the image for carrying out live body checking, and by acquired image It is stored in the storage device 104 so that other components are used.Image collecting device 110 can be photographic head.It should be appreciated that Image collecting device 110 is only example, and electronic equipment 100 can not include image collecting device 110.In this case, may be used To be used for the image of live body checking using other image acquisition devices, and the image of collection is sent to into electronic equipment 100.

Exemplarily, for realizing that the exemplary electronic device of live body verification method according to embodiments of the present invention and device can To realize on the equipment of personal computer or remote server etc..

Below, live body verification method according to embodiments of the present invention will be described with reference to Fig. 2.Fig. 2 is illustrated according to the present invention one The indicative flowchart of the live body verification method 200 of individual embodiment.As shown in Fig. 2 live body verification method 200 includes following step Suddenly.

Random to generate live body action command in step S210, live body action command is used to indicate the hand-held body of object to be verified Part card performs corresponding live body action.

In step S210, suitable live body action command can be as needed generated.Generation herein, refer to generation and with Certain way (such as voice, word etc.) is notified to object to be verified.Live body action command is used to indicate that object to be verified is performed Corresponding live body action.It is exemplary and without limitation, live body action can be included in while blocking face with identity card and turn over Turn and/or translate identity card.

In live body proof procedure, can require that user's (object i.e. to be verified) performs according to the live body action command for generating Some default live body actions.The live body action that user performs is needed to make identity card that adopting for photographic head is occurred in together with face In the range of collection, so that photographic head can gather the image obtained comprising face and identity card (including video or still image) As image to be verified.One example of live body action command is requirement user (such as live body verify starting stage) by identity The flat act of card flushes with its nose, and user front is towards photographic head, and identity card is parallel with the lens plane of photographic head.Live body action refers to Another example of order is that requirement user overturns identity card, the angle of pitch (pitch angles) and/or yaw angle (yaw of change of status card Angle), while requiring that identity card can block the nose and mouth point of face in switching process.Live body action command it is another Example is requirement user up and down and/or left and right translates identity card, while requiring that identity card can block people in translation motion The nose and mouth of face point.

It should be understood that above live body action command (or the live body action performed by instruction object to be verified) is only exemplary rather than Limit, the present invention can be based on other any suitable live body action commands and (or indicate that the live body performed by object to be verified is moved Make) realize.In addition, the identity card in above-mentioned example blocks the position of face to be only exemplary rather than limiting, identity card is making all Such as overturn or translate etc interactive action when the face position that sheltered from can be any position on face, the present invention is not This is limited.

Require object to be verified make some identity cards interact with face and interaction mode change live body action, can be with Photographic head is enabled to gather identity card and face under distinct interaction state, so as in live body proof procedure, be based on The image to be verified of different situations carries out live body checking, improves the percent of pass of live body checking.

According to embodiments of the present invention, form is played by text importing form and/or audio frequency and exports live body action command. In one example, can will expect that the live body action that user performs shows via output devices such as display screens.Show another In example, can will expect that the live body action that user performs is played back via output devices such as speakers.It is of course also possible to while Live body action command is exported using above two output means.

In step S220, Real-time Collection object to be verified performs the image of live body action, to obtain image to be verified.

Image to be verified can be any suitable image gathered for face and identity card.Image to be verified can be Original image, or the image obtained after pretreatment is carried out to original image that photographic head is collected.Image to be verified Can be still image, or one section of video.

Exemplarily, in system of real name certification occasion, need to carry out live body checking, in such a case, it is possible to be given first Prompting, it is desirable to which user provides its identity card and hand-held identity card performs live body action, and for the hand-held identity card execution of user Live body action gathers image.In one example, user can utilize its mobile phone photographic head gather oneself including identity card With face and the image of identity card live body action, and end of uploading onto the server carries out live body checking.In another example, image Head can be the installation photographic head of such as banking system, and user makes as requested live body action before photographic head, by imaging The identity card and face of head collection user, uploads to and carry out in the processing system of bank rear end live body checking.

Exemplarily, some live body actions are made in order to user using the intersection information of identity card and face, can be required, So that there is interaction between identity card and face, for example, make identity card partial occlusion face etc..Intersection information as herein described is Refer to the information formed on image due to influencing each other between identity card and face (i.e. interactive), such as identity card and face weight The information such as illumination, the focusing of folded part.

In step S230, determine whether object to be verified is verified by live body based on image to be verified.

As described above, during object to be verified performs live body action, the identity card that object to be verified is provided There is interaction and face between, there is intersection information in the image to be verified that certain moment is obtained.Exemplarily, can basis Whether the interactive information checking face of face information, ID card information and identity card and face in image to be verified and identity card Belong to live body.As the interaction mode of the carrying out of live body action, identity card and face can change, obtained in different time Image to be verified in intersection information it is also different, although the image to be verified according to being obtained at certain moment determines to be tested Card object is not verified by live body, it is possible that determining object to be verified according to the image to be verified obtained at other moment Verified by live body.

Exemplarily, in step S230, it can be determined that the face information, ID card information and intersection in image to be verified Whether information is the face obtained in the case where object to be verified performs correct live body action according to live body action command Information, ID card information and intersection information, if it is, determining that object to be verified is verified by live body, otherwise determine to be verified Object is not verified by live body.It is understood that perform correct live body according to live body action command in object to be verified moving In the case of work, the interaction mode of identity card and face substantially determines, ID card information, face information and the friendship for being obtained Fork information should be with certain rule, it may be predetermined that this rule.The rule can be by adopting great amount of samples Image is trained to determine to convolutional neural networks.Treated at authentication image using the convolutional neural networks for training Reason, it is possible to obtain image to be verified is obtained in the case that object to be verified performs correct live body action according to live body action command The probability of the image for obtaining, and then can determine whether object to be verified is verified by live body.The form and instruction of convolutional neural networks Practice process will be described below, do not repeat herein.

In one example, in step S230, except the face information, ID card information and friendship that judge in image to be verified Whether fork information is the people obtained in the case where object to be verified performs correct live body action according to live body action command Outside face information, ID card information and intersection information, whether the action that can also be performed to hand-held identity card is dynamic with the live body Match as instruction and judged, if the judged result of aforementioned four aspects is being, it is determined that object to be verified is by living Experience card, otherwise determines that object to be verified is not verified by live body.

Live body verification method according to embodiments of the present invention, moves due to gathering the hand-held identity card execution live body of object to be verified The image of work simultaneously carries out live body checking based on the image of collection, therefore can bring with reference to identity card in live body proof procedure Information (including ID card information and the intersection information of identity card and face) judging whether object to be verified belongs to live body, so The accuracy of live body checking can be improved.

Exemplarily, live body verification method according to embodiments of the present invention can be in setting with memorizer and processor Realize in standby, device or system.

Live body verification method according to embodiments of the present invention can be deployed at image acquisition end, for example, can be deployed in The image acquisition end of bank's system of real name Verification System.Alternatively, live body verification method according to embodiments of the present invention can also divide It is deployed at server end (or high in the clouds) and client cloth.For example, image to be verified can be gathered in client, client will The image to be verified for collecting sends server end (or high in the clouds) to, and by server end (or high in the clouds) live body checking is carried out.

Whether Fig. 3 illustrates and according to an embodiment of the invention determines object to be verified by live body based on image to be verified The indicative flowchart of the step of checking (step S230).As shown in figure 3, step S230 may comprise steps of.

In step S310, the face and identity card in image to be verified is detected.

Exemplarily, when image to be verified is still image, being directed to the still image carries out face and identity card Detection;When image to be verified is video, for each frame of video in the video face and identity card detection are carried out.

Can be using people any existing or during in the future Face datection algorithm in the cards is to detect image to be verified Face.It is for instance possible to use AdaBoost algorithms, CART (post-class processing) algorithm scheduling algorithms to detect image to be verified in people Face.Deposit in image to be verified in the context of a person's face, the face for detecting can adopt conventional face frame to represent, i.e., with one Individual rectangle frame indicates the position of face.

Likewise it is possible to using it is any existing or in the future identity card detection algorithm in the cards detecting figure to be verified Identity card as in.Similar to Face datection algorithm, it can detect the edge contour of identity card to identity card detection algorithm, and The position of identity card is indicated with an identity card frame (can be rectangle frame).

In the case where image to be verified is video, it is possible to use Face datection algorithm and identity card detection algorithm are examined in real time The face and identity card in each frame of video is surveyed (position) and tracked, can send logical when face is found or identity card loses Know.

In step S320, perform multiple live bodies and judge operation, wherein, multiple live bodies judge that operation includes that the first live body judges Operation, the second live body judge that operation and the 3rd live body judge operation, wherein, the first live body judges that operation includes:Based on to be verified Image judges whether the face for detecting belongs to live body, and the second live body judges that operation includes:Judge to detect based on image to be verified Whether the identity card for going out belongs to live body, and the 3rd live body judges that operation includes:Based on the people that image synthesis to be verified judges to detect Whether face and the identity card for detecting belong to live body on the whole.

First live body judges that operation individually judges whether the face for detecting belongs to live body.For real human face thinks its category In live body, for false face thinks that it is not belonging to live body.Exemplarily, false face can include being obtained for screen reproduction Face, the face using CG Software Creates, by printing the face etc. for obtaining.

Second live body judges that operation individually judges whether the identity card for detecting belongs to live body.For real identity card is thought It belongs to live body, for false identities identification is not belonging to live body for it.Exemplarily, false identities card can be included by being directed to Identity card, Freehandhand-drawing identity card that screen reproduction is obtained etc..

3rd live body judges that operation judges that the face for detecting is whole with the identity card for detecting based on image synthesis to be verified Whether belong to live body on body.The face for detecting be real human face and the identity card that detects be real identity card situation Under, it is believed that the two belongs on the whole live body, in the case of any one is false, all thinks that the two is not belonging on the whole Live body.

Because face surface is up-and-down, therefore real identity card and real human face be not in three dimensions same In one plane, if identity card and face are present interacted, such as if identity card partial occlusion face, real human face and true body Light conditions, circumstance of occlusion, focus condition etc. detailed information (i.e. intersection information) between part card can be with false face and very Real identity card or real human face are demonstrate,proved with false identities or false face is different with during false identities card interaction, therefore according to identity Card may determine that whether the face for detecting and the identity card for detecting belong to live body on the whole with the intersection information of face.Due to By using live body and non-living body in illumination, block, focus on etc. different judging identity card and people in terms of image detail information Whether face belongs on the whole live body, therefore can improve the accuracy that face and identity card verity judge, also just can improve The accuracy of live body checking.

In step S330, judge that each live body in operation judges that the judged result of operation determines according to multiple live bodies to be tested Whether card object is verified by live body, if arbitrary live body judges that the judged result for operating is no, it is determined that object to be verified is not Verified by live body, otherwise determine that object to be verified is verified by live body.

In one example, image to be verified is still image, can only consider face information, body in image to be verified Whether part card information and intersection information meet the requirement of live body checking, you can be to judge whether face belongs to live body, identity card It is no to belong to live body and whether the two belongs on the whole live body, and determine object to be verified whether by living according to judged result Experience card.In this case, the plurality of live body judges that operation can only include that the first live body judges that operation, the second live body are sentenced Disconnected operation and the 3rd live body judge operation.First live body judges that operation is mainly responsible for single face verity and is judged, second lives Body judges that operation is mainly responsible for single identity card verity and judges that the 3rd live body judges operation chief leading cadre's face and identity card Integrated facticity judge.Above-mentioned three live bodies judge that operation each obtains a kind of judged result, indicate the main body being each responsible for It is whether true, that is, indicate whether the main body being each responsible for belongs to live body, if this three live bodies judge arbitrary in operation Live body judges that the judged result of operation indicates that its responsible main body of institute is not belonging to live body, it is determined that provide the to be tested of identity card and face Card object (i.e. user) can not be verified by live body.So, it is overall true due to increased on the basis of the judgement of independent verity Reality judges, therefore can improve the accuracy of live body checking.

In another example, image to be verified is video, except above-mentioned face information, ID card information and intersection information with Outward, can also further consider whether the action that identity card is performed matches with live body action command.In this case, it is described Live body judges that operation may further include the 4th live body and judge operation.Exemplarily, multiple live bodies judge that operation also includes the Four live bodies judge operation, wherein the 4th live body judges that operation includes:Based on the image to be verified judge described in detect Identity card perform action whether match with the live body action command.

Wherein, the 4th live body judges that operation can be realized using various feasible motion detections with method for tracing, herein not Repeated again.This example judges operation by increasing by the 4th live body, can further improve the accuracy of live body checking.

Exemplarily, when the image to be verified is video, the 4th live body judges operation based in the video At least two frames carry out, first live body judges that operation, second live body judge that operation and the 3rd live body judge behaviour Make to be carried out based at least frame at least two frames.Wherein, when the first live body judges operation based on two frames and above video When frame or two and above still image are carried out, only judge for the judged result of each frame of video or each still image When object to be verified is live body, first live body judges that operation just judges object to be verified for live body.Second live body judges operation Judge that the judgment rule of operation is identical with this with the 3rd live body.And in the 4th live body judges operation, if the body of any frame Part card action is mismatched with live body action command, then it is not living for object to be verified that the 4th live body judges to operate the result of output Body.

In a specific example, the image to be verified is video, and the 4th live body judges that operation is regarded based on described All videos frame in frequency is carried out, and first live body judges that operation, second live body judge operation and the 3rd live body Judge that operation is carried out also based on all videos frame, to improve live body judgment accuracy.

In one example, directly above-mentioned live body can be carried out according to original image to be verified and judges operation.Another In one example, facial image to be verified and ID Card Image to be verified can be extracted from image to be verified first and based on to be tested Witness's face image and ID Card Image to be verified carry out above-mentioned live body and judge operation.This latter example is described below.

According to embodiments of the present invention, after step S310, live body verification method 200 can also include:From figure to be verified The facial image to be verified only comprising the face for detecting is extracted as in and the identity to be verified of the identity card for detecting only is included Card image;And the first live body judges that operation can include:Judge whether the face for detecting belongs to based on facial image to be verified In live body, the second live body judges that operation can include:Judge whether the identity card for detecting belongs to based on ID Card Image to be verified In live body, the 3rd live body judges that operation can include:Based on facial image to be verified and ID Card Image comprehensive descision to be verified Whether the face for detecting and the identity card for detecting belong to live body on the whole.

In the case where detection obtains face frame, it is possible to use face frame is treated authentication image and split, by face frame Internal pixel splits acquisition facial image to be verified.Similarly, in the case where detection obtains identity card frame, can be with profit Authentication image is treated with identity card frame to be split, the pixel in identity card inframe portion is split into acquisition identity card figure to be verified Picture.

Subsequently, when the judgement operation of the first live body is performed, the facial image to be verified for extracting can be based on and judges detection Whether the face for going out belongs to live body, can so avoid the interference of the information at other picture positions in addition to face, can be with Improve efficiency and accuracy that face verity judges.When the judgement operation of the second live body is performed, can be based on treating for extracting Checking facial image judges whether the face that detects belongs to live body, similarly, so can avoid in addition to identity card its The interference of the information at his picture position, can improve the efficiency and accuracy of the judgement of identity card verity.Live performing the 3rd When body judges operation, the face that can be detected based on the facial image to be verified and ID Card Image to be verified judgement for extracting Whether belong to live body on the whole with the identity card for detecting, similarly, so can avoid in addition to face and identity card its The interference of the information at his picture position, can improve the efficiency and accuracy of the integrated facticity judgement of face and identity card.

According to embodiments of the present invention, detected based on facial image to be verified and ID Card Image comprehensive descision to be verified Whether face and the identity card for detecting belong on the whole live body (i.e. the 3rd live body judges operation) can include:By witness to be tested The first convolutional neural networks that face image and ID Card Image to be verified input are trained, to obtain the face and detection that detect The identity card for going out belongs on the whole the probability of live body;And the face detected according to determine the probability and the identity card that detects it is whole Whether belong to live body on body.

In the 3rd live body judges operation, it is possible to use the first convolutional neural networks for training are judging.Fig. 4 illustrates root According to the schematic network structure of the first convolutional neural networks of one embodiment of the invention.It should be noted that the network structure shown in Fig. 4 Limitation of the present invention is only exemplary rather than, the first convolutional neural networks in the embodiment of the present invention are not limited to shown in Fig. 4 Network structure, it can have other any suitable network structures, the type of the layer in the first convolutional neural networks, layer it Between connected mode, the number of layer, filter number, wave filter size inside layer etc. can set as needed.

With reference to Fig. 4, the first convolutional neural networks have six convolutional layers, and (convolutional layer, use respectively Conv0, conv1, conv2, conv3, conv4 and conv5 are represented) and a full articulamentum (fully-connected Layer, is represented with fc0), six convolutional layers and a full articulamentum are all divided into upper and lower two-way, are respectively used to receive identity card figure As (positive sample ID Card Image including ID Card Image to be verified and hereinafter described and negative sample ID Card Image) and face Image (positive sample facial image including facial image to be verified and hereinafter described and negative sample facial image).Full articulamentum The characteristic pattern that the two-way of fc0 is exported is combined one articulamentum (concat layer, represented with concat) of input, subsequently One full articulamentum (being represented with fc1) of connection, finally connects output layer (being represented with softmax).

As shown in figure 4, ID Card Image to be verified and facial image to be verified are input into together the first convolutional neural networks, In the processing procedure of the first convolutional neural networks, can comprehensive ID Card Image and facial image information.As described above, exist In the case that identity card interacts (such as identity card partial occlusion face) with face, identity card influences each other with face.The The informix of identity card and face can be got up to process by one convolutional neural networks, you can so that the intersection information of the two to be considered Enter, and then judge whether the two belongs to live body on the whole, if any one is untrue, then it is assumed that the two does not belong on the whole In live body.First convolutional neural networks can export the face for detecting and the identity card for detecting belongs on the whole the general of live body Rate (or claiming confidence level).Exemplarily, if the probability of the first convolutional neural networks output is more than or equal to 0.5, it may be determined that The face for detecting and the identity card for detecting belong on the whole live body, if the probability of the first convolutional neural networks output is less than 0.5, then can determine that the face for detecting and the identity card for detecting are not belonging on the whole live body.First convolutional neural networks can Obtained with advancing with the training of great amount of samples image.

Convolutional neural networks can be with the complicated characteristics of image of autonomic learning, it is possible to achieve high accuracy, high performance image point Class, thus use it for live body judge can obtain accurate judged result, be conducive to improve live body checking accuracy.

According to embodiments of the present invention, live body verification method 200 can also include the training step of the first convolutional neural networks. Fig. 5 illustrates the indicative flowchart of the training step S500 of the first convolutional neural networks according to an embodiment of the invention.

As shown in figure 5, the training step S500 of the first convolutional neural networks is comprised the following steps.

In step S510, obtain training data, the training data includes positive sample image and negative sample image, it is described just Sample image includes real human face and real identity card, and the negative sample image includes false face and real identity card.

A large amount of positive sample images can in advance be gathered.For example can gather comprising real human face and real identity card 10000 videos, wherein some of each video frame of video can be trained as positive sample image.In collection positive sample During image, can require that provide identity card and the object (user) of face overturns while face is blocked with identity card Or translation identity card, the positive sample image different with identity card interaction mode to gather face.

In addition to positive sample image, a large amount of negative sample images can also be in advance gathered.For example can gather comprising false people 10000 videos of face and real identity card, wherein some of each video frame of video can be carried out as negative sample image Training.False face can be face, the face by printing acquisition that the face reproduction for example for playing in screen is obtained Deng.During collection negative sample image, can require that provide identity card and the object (user) of face is being hidden with identity card Identity card, the negative sample image different with identity card interaction mode to gather face are overturn or translated while gear face.

Due to being generally directed to the probability of face fraud than larger, and the difficulty faked simultaneously of face and identity card compared with Height, therefore in the training process of the first convolutional neural networks, can be trained mainly for face, that is, positive sample figure As for real human face collection, for false face collection, the identity card in the case of two kinds can be true to negative sample image Identity card, this can cause to train the first convolutional neural networks for obtaining to be mainly used in judging the true and false of face, further strengthen The judgement precision of face verity.Certainly, in the training process of the first convolutional neural networks, negative sample image can also be included Real human face and false identities card are demonstrate,proved comprising false face and false identities, so that the first convolution god that training is obtained Jing networks are also contemplated for the true and false of identity card.

In step S520, the only positive sample facial image comprising face is extracted from positive sample image and identity card is only included Positive sample ID Card Image.

For each the positive sample image for being gathered, it is possible to use as described above Face datection algorithm detects it In face, and detect identity card therein using identity card detection algorithm as described above.Subsequently, can basis Testing result extracts the only positive sample facial image comprising face and only the positive sample ID Card Image comprising identity card.

In step S530, the only negative sample facial image comprising face is extracted from negative sample image and identity card is only included Negative sample ID Card Image.

Similarly, for each the negative sample image for being gathered, it is possible to use Face datection algorithm as described above Face therein is detected, and identity card therein is detected using identity card detection algorithm as described above.Subsequently, The only negative sample facial image comprising face and only the negative sample identity card figure comprising identity card can be extracted according to testing result Picture.

In step S540, with positive sample facial image and positive sample ID Card Image as positive sample, and with negative sample people Face image and negative sample ID Card Image are negative sample, carry out neural metwork training to obtain the first convolutional neural networks.

Using positive sample facial image and positive sample ID Card Image as positive sample, by negative sample facial image and negative sample ID Card Image has in the first convolutional neural networks of network structure for example as shown in Figure 4 as negative sample, respectively input It is trained.Can be using stochastic gradient descent method training neutral net to restraining, so as to obtain the first required convolutional Neural Network.

The execution sequence of step shown in Fig. 5 is only exemplary rather than limiting, the training step S500 of the first convolutional neural networks There can be other rational execution sequences, such as step S520 can be performed after step S530 or simultaneously.

According to embodiments of the present invention, before step S520, live body verification method 200 can also include:Calculate positive sample Face in this image zooms to the positive sample scaling needed for default size, and according to positive sample scaling to positive sample Image is zoomed in and out;Before step S530, live body verification method 200 can also include:Calculate the people in negative sample image Face zooms to the negative sample scaling needed for default size, and negative sample image is contracted according to negative sample scaling Put;In the facial image to be verified extracted from image to be verified only comprising the face for detecting and only comprising the identity for detecting Before the ID Card Image to be verified of card, live body verification method 200 can also include:Calculating zooms to the face for detecting pre- If the image scaling ratio to be verified needed for size, and treat authentication image according to image scaling ratio to be verified and zoom in and out.

Default size can be any suitable size, for example, can be 150 pixel × 150 pixels.As described above, profit Employment face detection algorithm can detect the face in image, and it can be represented with face frame.For image to be verified, positive sample For image and negative sample image, the mode for detecting wherein face is similar.After face frame is detected, face frame is calculated Zoom to the scaling of default size and the process that image is zoomed in and out can be understood as to face according to scaling The process being normalized.Treat authentication image, positive sample image carries out similar face normalization process with negative sample image, The face size of these images can be adjusted so as to it is basically identical, be convenient for process, reduce live body checking error.

It is to be appreciated that face normalization is not limited to a kind of above-mentioned mode, it can have other rational implementations. For example, for image to be verified, can press after facial image and ID Card Image are extracted from image to be verified Facial image and ID Card Image are zoomed in and out according to image scaling ratio to be verified, in this case, without the need for entirely treating Authentication image is zoomed in and out.For positive sample image and negative sample image, it would however also be possible to employ similar mode carries out face Normalization, repeats no more.It is appreciated that extracting facial image and identity card figure again after zooming in and out to whole image to be verified The mode of picture, due to need not respectively process facial image and ID Card Image, therefore can save certain amount of calculation.

According to embodiments of the present invention, step S220 can include:Gather in the preset period of time after checking start time Object to be verified performs the video of live body action or continuous several still images as image to be verified.

As described above, image to be verified can be still image, or the frame of video in video.In an example In, after live body checking starts, one section of video can be gathered in preset period of time by photographic head, using certain in this section of video A little frame of video carry out live body checking.The mode of frame of video is selected to set as needed from video, the present invention does not enter to this Row is limited.For example, a frame of video can be selected every Fixed Time Interval from video, will be every in selected frame of video One carries out live body checking as image to be verified.Again for example, some frame of video can be randomly choosed from video, will be selected Frame of video in each carry out live body checking as image to be verified.It is of course also possible to select all videos in video Frame is used to carry out live body checking.

Because in different video frame, the interaction mode of face and identity card may be varied from, especially people is being provided In the case that the object of face and identity card makes as requested live body action, the change can be obvious, regards hence with one section Frame of video in frequency carries out live body checking, can account for the image to be verified gathered under different conditions, improves live body and tests The accuracy and percent of pass of card, lifts Consumer's Experience.

According to embodiments of the present invention, live body verification method 200 can also include:If in selected frame of video Arbitrary frame of video determines that object to be verified is verified by live body, it is determined that live body is proved to be successful, and otherwise determines live body authentication failed.

In the present embodiment, after live body checking starts, if according to selected frame of video one in preset period of time Cannot directly determine that object to be verified is verified by live body, then live body authentication failed be can determine, if in authentication application In, then will be unable to proceed subsequent identity coherence authentication operation (i.e. above-mentioned identity contrast operation);, whereas if pre- If the arbitrary frame of video in the period in selected frame of video determines that object to be verified is verified by live body, then can determine Live body is proved to be successful, and can proceed subsequent identity coherence authentication operation.

According to embodiments of the present invention, after step S310, live body verification method 200 can also include:If to be tested Any face is not detected by card image or is not detected by any identity card, then output re-executes carrying for live body checking Show.

As described above, in whole live body proof procedure, with real-time detection (positioning) and image to be verified can be tracked In face and identity card, if being not detected by any face or being not detected by any identity card, illustrate face or Identity card is lost, and prompting can be exported in this case, to inform that the object to be verified for providing identity card and face is re-executed Live body is verified.Aforesaid way can in time find the error in live body proof procedure, help user positively and verified by live body, from And live body verification efficiency can be improved, lift Consumer's Experience.The prompting for re-executing live body checking can be using any suitable Form is exported, for example, can pass through one or more that text importing form, audio frequency are played in form and signal lighties blinking form Output.

According to embodiments of the present invention, judge whether the face that detects belongs to live body (i.e. the based on facial image to be verified One live body judges operation) can include:The second convolutional neural networks that facial image to be verified input is trained, to judge inspection Whether the face measured belongs to live body.

The second convolutional neural networks that facial image to be verified input is trained, by the minutia for analyzing face, May determine that the face for detecting is real human face, or obtain by using CG Software Creates, for modes such as screen reproduction False face.If the face for detecting is real human face, then it is assumed that it belongs to live body, otherwise it is assumed that it is not belonging to live body. Second convolutional neural networks can be obtained in advance using the training of great amount of samples images off-line, and it can be considered as a real human face and sentence Other device.During neural metwork training is carried out to obtain the second convolutional neural networks, the training sample for being adopted can be A large amount of images comprising real human face and the image comprising false face.False face can be obtained including above-mentioned for screen reproduction Face, the face using CG Software Creates, by printing the face etc. for obtaining.With the first convolutional neural networks similarly, Neutral net can be trained to restraining by stochastic gradient descent method, so as to obtain the second required convolutional neural networks.

According to embodiments of the present invention, judge whether the identity card for detecting belongs to live body based on ID Card Image to be verified (the second live body judges operation) can include:The 3rd convolutional neural networks that ID Card Image to be verified input is trained, with Whether the identity card that judgement is detected belongs to live body.

The 3rd convolutional neural networks that ID Card Image to be verified input is trained, it is special by the details for analyzing identity card Levy, it can be determined that whether the identity card for detecting is real identity card.If the identity card for detecting is real identity card, recognize Belong to live body for it, otherwise it is assumed that it is not belonging to live body.3rd convolutional neural networks can in advance using great amount of samples image from Line training is obtained, and it can be considered as a real identity card arbiter.Carrying out neural metwork training to obtain the 3rd convolution god During Jing networks, the training sample for being adopted can include the image of real identity card in a large number and comprising false identities card Image.False identities card can include above-mentioned identity card, Freehandhand-drawing identity card obtained for screen reproduction etc..With the first convolution Neutral net and the second convolutional neural networks are extremely restrained it is likewise possible to pass through stochastic gradient descent method training neutral net, from And the 3rd convolutional neural networks needed for obtaining.

According to embodiments of the present invention, step S320 can include:The first live body is performed according to the order for arranging judge behaviour Make, the second live body judges operation and the 3rd live body judges operation, if arbitrary live body judges that the result for operating is no, stopping is held The subsequent live body of row judges operation.

In one example, can completely perform the first live body and judge that operation, the second live body judge operation and the 3rd live body Judge operation.In another example, can in advance arrange the first live body and judge that operation, the second live body judge operation and the 3rd work Body judges the execution sequence of operation, and performs three live bodies judgement operations successively according to the order for arranging.For example, it is assumed that according to The first live body is first carried out and judges that operation, next second live body of execution judge operation, finally perform what the judgement of the 3rd live body was operated Order is performed, if the first live body judges that the result of operation is that the face for detecting belongs to live body, can continue executing with second Live body judges operation, if the first live body judges that the result of operation is that the face for detecting is not belonging to live body, does not continue to hold Remaining second live body of row judges that operation and the 3rd live body judge operation.When the judgement operation of the second live body is gone to, equally may be used To judge whether that continuing executing with the 3rd live body judges operation, repeats no more according to its result.

In the case of judging the result for operating to be not belonging to live body in certain first live body, it is already possible to determine to be tested Card object is not verified by live body, therefore judges operation without the need for performing subsequent live body again, and this mode can be to a certain degree Upper saving data amount of calculation and live body proving time.

According to a further aspect of the invention, there is provided a kind of identity identifying method.The identity identifying method is tested including above-mentioned live body Card method 200, and also including:In the case of it is determined that object to be verified is verified by live body, judge from image to be verified Whether the face on identity card for detecting is consistent with the face detected from image to be verified.

In authentication application, live body checking can be first carried out, it is determined that providing the to be verified of identity card and face In the case that object is by live body checking, affiliated object actual to identity card can be continued and compareed with object to be verified, be come Whether both certifications are same people.This can be by the identity card that will detect from image to be verified (such as in step S310 Obtained in the identity card for detecting) on face with the face that detects from image to be verified (such as in step S310 Obtained in the face for detecting) compared to realize, if the face on the identity card for detecting with detect Face it is consistent, illustrate that the two belongs to same people, authentication success, otherwise authentication fail.For the identity for detecting Card, can also automatically extract the name and identification card number on identity card, the name and body using optical character recognition (OCR) technology Part card number can be used for carrying out authentication, so as to further lift Consumer's Experience.Specifically, the identity for detecting in acquisition After name and identification card number on card, it is possible to use the biometric authentication technology such as recognition of face or Application on Voiceprint Recognition, by above-mentioned inspection Other data such as the voice signal file that the face measured or object to be verified are uploaded and cut-and-dried and identify The name biological characteristic corresponding with identification card number (the authoritative citizen ID certificate image for for example obtaining from the Ministry of Public Security, or in advance Vocal print signal of each user recorded etc.) contrasted, to confirm whether object to be verified is consistent with the user being previously stored.

Intersection information checking identity card and the face true and false due to combining identity card and face in live body proof procedure, i.e., ID card information is just considered in live body proof procedure, therefore ID card information in conventional authentication procedures can be avoided The low problem of caused safety is used separately with face information, the safety and reliability of authentication can be improved.

According to a further aspect of the invention, there is provided a kind of live body verifies device.Fig. 6 is shown according to one embodiment of the invention Live body verify device 600 schematic block diagram.

As shown in fig. 6, live body checking device 600 according to embodiments of the present invention includes directive generation module 610, to be verified Image obtains module 620 and is verified determining module 630.The modules can respectively be performed and retouched above in conjunction with Fig. 2-5 Each step/function for the live body verification method stated.Hereinafter only the major function of each module of the live body checking device 600 is entered Row description, and omit the detail content having been described above.

Directive generation module 610 is used for random generation live body action command, and the live body action command is to be tested for indicating The hand-held identity card of card object performs corresponding live body action.In the electronic equipment that directive generation module 610 can be as shown in Figure 1 The programmed instruction stored in the Running storage device 104 of processor 102 is realizing.

Image to be verified obtains module 620 is used for the figure that object to be verified described in Real-time Collection performs the live body action Picture, to obtain image to be verified.Image to be verified obtains the processor 102 in the electronic equipment that module 620 can be as shown in Figure 1 The programmed instruction stored in Running storage device 104 is realizing.

Determining module 630 is verified for determining the object to be verified whether by living based on the image to be verified Experience card.The Running storage device 104 of processor 102 being verified in the electronic equipment that determining module 630 can be as shown in Figure 1 The programmed instruction of middle storage is realizing.

According to embodiments of the present invention, the determining module 630 that is verified includes:Detection sub-module, it is described for detecting Face and identity card in image to be verified;Live body judging submodule, for performing multiple live bodies operation is judged, wherein, it is described Live body judging submodule includes:First live body judging unit, for performing the first live body operation is judged, wherein, described first lives Body judging unit includes face judgment sub-unit, for judging whether the face for detecting belongs to living based on the image to be verified Body;Second live body judging unit, for performing the second live body operation is judged, wherein, the second live body judging unit includes body Part card judgment sub-unit, for judging whether the identity card for detecting belongs to live body based on the image to be verified;3rd live body Judging unit, for performing the 3rd live body operation is judged, wherein, the 3rd live body judging unit includes that comprehensive descision is single Unit, be on the whole for the face that detects described in being judged based on the image synthesis to be verified and the identity card that detects It is no to belong to live body;And determination sub-module is verified, judge for each live body in the live body judging submodule The judged result of unit determines whether the object to be verified is verified by live body, if the judgement knot of arbitrary live body judging unit Fruit is no, it is determined that the object to be verified otherwise determines that the object to be verified is verified by live body not by live body checking.

In one example, live body judging submodule also includes the 4th live body judging unit, sentences for performing the 4th live body Disconnected operation, wherein the 4th live body judging unit includes identity card action judgment sub-unit, for based on the figure to be verified Whether the action that the identity card detected as described in judging is performed matches with the live body action command.

Exemplarily, when image to be verified is video, the 4th live body judging unit is based at least two in the video Frame performs the 4th live body and judges operation, the first live body judging unit, the second live body judging unit and described the Three live body judging units perform respectively the first live body and judge that operation, described second live based at least frame at least two frames Body judges that operation and the 3rd live body judge operation.

According to embodiments of the present invention, the live body checking device 600 also includes:First image zooming-out module (not shown), For the only facial image to be verified comprising the face for detecting being extracted from the image to be verified and only being included described The ID Card Image to be verified of the identity card for detecting;Also, the face judgment sub-unit includes face determination component, is used for Whether live body, the identity card judgment sub-unit bag are belonged to based on the face detected described in the facial image judgement to be verified Identity card determination component is included, for whether belonging to living based on the identity card detected described in the ID Card Image judgement to be verified Body, the comprehensive descision subelement includes comprehensive descision component, for based on the facial image to be verified and described to be verified Whether the face and the identity card for detecting detected described in ID Card Image comprehensive descision belongs to live body on the whole.

According to embodiments of the present invention, the comprehensive descision component includes:First input sub-component, for will be described to be verified The first convolutional neural networks that facial image and the ID Card Image to be verified input are trained, to obtain described detecting Face and the identity card for detecting belong on the whole the probability of live body;And live body determines sub-component, for according to described Whether the face and the identity card for detecting detected described in determine the probability belongs to live body on the whole.

According to embodiments of the present invention, the live body checking device 600 also includes:Training data acquisition module, for obtaining Training data, the training data includes positive sample image and negative sample image, the positive sample image comprising real human face and Real identity card, the negative sample image includes false face and real identity card;Second image zooming-out module, for from described The only positive sample facial image comprising face and only the positive sample ID Card Image comprising identity card are extracted in positive sample image;The Three image zooming-out modules, for the only negative sample facial image comprising face to be extracted from the negative sample image and body is only included The negative sample ID Card Image of part card;And training module, for the positive sample facial image and the positive sample identity Card image is positive sample, and with the negative sample facial image and the negative sample ID Card Image as negative sample, carries out god Jing network trainings are obtaining first convolutional neural networks.

According to embodiments of the present invention, the live body checking device 600 also includes:First Zoom module, for described Two image zooming-out modules extract the only positive sample facial image comprising face from the positive sample image and only include identity card Positive sample ID Card Image before, the face in the positive sample image is zoomed to positive sample needed for default size for calculating Scaling, and the positive sample image is zoomed in and out according to the positive sample scaling;Second Zoom module, for The 3rd image zooming-out module is extracted the only negative sample facial image comprising face from the negative sample image and is only included Before the negative sample ID Card Image of identity card, the face in the negative sample image is zoomed to the default size institute by calculating The negative sample scaling for needing, and the negative sample image is zoomed in and out according to the negative sample scaling;And the 3rd Zoom module, only include the people for detecting for extracting from the image to be verified in described first image extraction module Before the facial image to be verified of face and the only ID Card Image to be verified comprising the identity card for detecting, calculating will be described The face for detecting zooms to the image scaling ratio to be verified needed for the default size, and contracts according to the image to be verified Put ratio to zoom in and out the image to be verified.

According to embodiments of the present invention, the face determination component includes:Second input sub-component, for will be described to be verified The second convolutional neural networks that facial image input is trained, to judge whether the face for detecting belongs to live body.

According to embodiments of the present invention, the identity card determination component includes:3rd input sub-component, for will be described to be tested The 3rd convolutional neural networks that card ID Card Image input is trained, to judge whether the identity card for detecting belongs to living Body.

According to embodiments of the present invention, the live body checking device 600 also includes:Prompting output module, if in institute State and any face is not detected by image to be verified or any identity card is not detected by, then output re-executes live body and tests The prompting of card.

According to embodiments of the present invention, the live body action is included in while blocking face with identity card and overturns and/or flat Move identity card.

According to a further aspect of the invention, there is provided a kind of identification authentication system, including above-mentioned live body checking device 600, wherein, The identification authentication system also includes face concordance judge module, for determining institute in the determining module 630 that is verified In the case of object to be verified is stated by live body checking, the face on the identity card that detects from the image to be verified is judged It is whether consistent with the face detected from the image to be verified.Identity according to embodiments of the present invention is hereinbefore described The embodiment of authentication method, those skilled in the art are appreciated that identity is recognized with reference to the above-mentioned description with regard to identity identifying method Implementation and its advantage of card device etc., repeat no more.

Those of ordinary skill in the art are it is to be appreciated that the list of each example with reference to the embodiments described herein description Unit and algorithm steps, being capable of being implemented in combination in electronic hardware or computer software and electronic hardware.These functions are actually Performed with hardware or software mode, depending on the application-specific and design constraint of technical scheme.Professional and technical personnel Each specific application can be used different methods to realize described function, but this realization it is not considered that exceeding The scope of the present invention.

Fig. 7 shows the schematic block diagram of live body checking system 700 according to an embodiment of the invention.Live body checking system System 700 includes image collecting device 710, storage device 720 and processor 730.

Image collecting device 710 is used to gather image to be verified.Image collecting device 710 is optional, live body checking system System 700 can not include image collecting device 710.

The storage device 720 store for realizing live body verification method according to embodiments of the present invention in corresponding steps Program code.

The processor 730 is used to run the program code stored in the storage device 720, to perform according to the present invention The corresponding steps of the live body verification method of embodiment, and for realizing live body checking device 600 according to embodiments of the present invention In directive generation module 610, image to be verified obtains and module 620 and is verified determining module 630.

In one embodiment, the live body checking system 700 is made when described program code is run by the processor 730 Perform following steps:Random to generate live body action command, the live body action command is used to indicate the hand-held identity of object to be verified Card performs corresponding live body action;Object to be verified described in Real-time Collection performs the image of the live body action, to be tested to obtain Card image;And determine whether the object to be verified is verified by live body based on the image to be verified.

In one embodiment, the live body checking system 700 is made when described program code is run by the processor 730 Performed determines that the object to be verified is included the step of whether checking by live body based on the image to be verified:Perform many Individual live body judges operation, wherein, the plurality of live body judges that operation includes that the first live body judges that operation, the second live body judge operation Judge to operate with the 3rd live body, wherein, first live body judges that operation includes:Judge to detect based on the image to be verified Face whether belong to live body, second live body judges that operation includes:Based on the body that the image to be verified judges to detect Whether part card belongs to live body, and the 3rd live body judges that operation includes:The detection is judged based on the image synthesis to be verified Whether the face and the identity card for detecting for going out belongs to live body on the whole;And judged in operation according to the plurality of live body Each live body judge that the judged result for operating determines that whether the object to be verified is verified by live body, if arbitrary live body is sentenced The judged result of disconnected operation is no, it is determined that the object to be verified does not verify that it is described to be verified right otherwise to determine by live body As being verified by live body.

In one embodiment, the live body checking system is made when described program code is run by the processor 730 After the step of face in the detection image to be verified and identity card performed by 700, described program code is by the place Reason device 730 also performs the live body checking system 700 when running:Extract from the image to be verified and only include the detection The facial image to be verified of the face for going out and the only ID Card Image to be verified comprising the identity card for detecting;Also, institute State the first live body and judge that operation includes:Whether belong to living based on the face detected described in the facial image judgement to be verified Body, second live body judges that operation includes:It is based on the identity card detected described in the ID Card Image judgement to be verified No to belong to live body, the 3rd live body judges that operation includes:Based on the facial image to be verified and the identity card to be verified Whether the face and the identity card for detecting detected described in image synthesis judgement belongs to live body on the whole.

In one embodiment, the live body checking system 700 is made when described program code is run by the processor 730 It is performed based on the face detected described in the facial image to be verified and the ID Card Image comprehensive descision to be verified The step of whether belonging to live body on the whole with the identity card for detecting includes:By the facial image to be verified and described treat Checking ID Card Image the first convolutional neural networks for training of input, to obtain the face for detecting and described detect Identity card belong to the probability of live body on the whole;And the face that detects according to the determine the probability and described detect Identity card whether belong to live body on the whole.

In one embodiment, the live body checking system is also made when described program code is run by the processor 730 700 perform:Training data is obtained, the training data includes positive sample image and negative sample image, the positive sample image bag Containing real human face and real identity card, the negative sample image includes false face and real identity card;From the positive sample figure The only positive sample facial image comprising face and only the positive sample ID Card Image comprising identity card are extracted as in;From the negative sample The only negative sample facial image comprising face and only the negative sample ID Card Image comprising identity card are extracted in this image;And with The positive sample facial image and the positive sample ID Card Image are positive sample, and with the negative sample facial image and institute It is negative sample to state negative sample ID Card Image, carries out neural metwork training to obtain first convolutional neural networks.

In one embodiment, the live body checking system is made when described program code is run by the processor 730 Positive sample facial image and only including identity card just that extracting from the positive sample image performed by 700 only includes face Before the step of sample identity card image, the live body checking system is also made when described program code is run by the processor 730 System 700 is performed:Face in the positive sample image is zoomed to the positive sample scaling needed for default size for calculating, and is pressed The positive sample image is zoomed in and out according to the positive sample scaling;Transported by the processor 730 in described program code Make that the only negative sample face comprising face is extracted from the negative sample image performed by the live body checking system 700 during row Before the step of image and only negative sample ID Card Image comprising identity card, described program code is transported by the processor 730 Perform also the live body checking system 700 during row:Calculate and the face in the negative sample image is zoomed to into described presetting greatly Negative sample scaling needed for little, and the negative sample image is zoomed in and out according to the negative sample scaling;Institute State make when program code is run by the processor 730 performed by the live body checking system 700 from the image to be verified Middle extraction only includes the facial image to be verified of the face for detecting and only includes the to be tested of the identity card for detecting Before the step of card ID Card Image, the live body checking system is also made when described program code is run by the processor 730 700 perform:The face for detecting is zoomed to the image scaling ratio to be verified needed for the default size for calculating, and is pressed The image to be verified is zoomed in and out according to the image scaling ratio to be verified.

In one embodiment, the live body checking system is made when described program code is run by the processor 730 After the step of face in the detection image to be verified and identity card performed by 700, described program code is by the place Reason device 730 also performs the live body checking system 700 when running:If be not detected by the image to be verified any Face is not detected by any identity card, then output re-executes the prompting of live body checking.

In one embodiment, the live body action is included in while blocking face with identity card and overturns and/or translate Identity card.

In one embodiment, the live body checking system 700 is made when described program code is run by the processor 730 Performed includes the step of whether belonging to live body based on the face detected described in the facial image judgement to be verified:By institute The second convolutional neural networks that facial image input to be verified is trained are stated, to judge whether the face for detecting belongs to living Body.

In one embodiment, the live body checking system 700 is made when described program code is run by the processor 730 Performed includes the step of whether belonging to live body based on the identity card detected described in the ID Card Image judgement to be verified: The 3rd convolutional neural networks that the ID Card Image to be verified input is trained, to judge that the identity card for detecting is It is no to belong to live body.

In one embodiment, the live body checking system 700 is made when described program code is run by the processor 730 The performed live body of execution first judges that operation, the second live body judge that operation and the 3rd live body include the step of judging operation:Press First live body is performed according to the order for arranging judge that operation, second live body judge that operation and the 3rd live body judge Operation, if arbitrary live body judges that the result for operating is no, stops performing subsequent live body judgement operation.

In one embodiment, the plurality of live body judges that operation also includes that the 4th live body judges operation, wherein described the Four live bodies judge that operation includes:Based on the image to be verified judge the action of the identity card execution for detecting whether with institute State live body action command to match.

In one embodiment, the image to be verified is video, and the 4th live body judges that operation is based on the video In at least two frames carry out, first live body judges that operation, second live body judge that operation and the 3rd live body judge Operation is carried out based at least frame at least two frames.

Additionally, according to embodiments of the present invention, additionally providing a kind of storage medium, program is stored on said storage Instruction, when described program is instructed and run by computer or processor for performing the live body verification method of the embodiment of the present invention Corresponding steps, and for realizing that live body according to embodiments of the present invention verifies the corresponding module in device.The storage medium Storage card, the memory unit of panel computer, the hard disk of personal computer, the read only memory of smart phone can for example be included (ROM), Erasable Programmable Read Only Memory EPROM (EPROM), portable compact disc read only memory (CD-ROM), USB storage, Or the combination in any of above-mentioned storage medium.

In one embodiment, the computer program instructions can cause to calculate by computer or processor when running Machine or processor realize that live body according to embodiments of the present invention verifies each functional module of device, and/or can perform Live body verification method according to embodiments of the present invention.

In one embodiment, the computer program instructions make below the computer execution when being run by computer Step:Random to generate live body action command, the live body action command is used to indicate that the hand-held identity card of object to be verified performs phase The live body action answered;Object to be verified performs the image of the live body action described in Real-time Collection, to obtain image to be verified;With And determine whether the object to be verified is verified by live body based on the image to be verified.

In one embodiment, the computer program instructions make performed by the computer when being run by computer Determine that the object to be verified is included the step of whether checking by live body based on the image to be verified:Detect described to be verified Face and identity card in image;Perform multiple live bodies and judge operation, wherein, the plurality of live body judges that operation includes the first work Body judges that operation, the second live body judge that operation and the 3rd live body judge operation, wherein, first live body judges that operation includes: Judge whether the face for detecting belongs to live body based on the image to be verified, second live body judges that operation includes:It is based on The image to be verified judges whether the identity card for detecting belongs to live body, and the 3rd live body judges that operation includes:Based on institute State image synthesis to be verified judge described in the face that detects and the identity card for detecting whether belong to live body on the whole;With And judge that each live body in operation judges that the judged result of operation determines that described to be verified pair likes according to the plurality of live body It is no to be verified by live body, if arbitrary live body judges that the judged result for operating is no, it is determined that the object to be verified does not pass through Live body checking, otherwise determines that the object to be verified is verified by live body.

In one embodiment, make when being run by computer performed by the computer in the computer program instructions The detection image to be verified in face and the step of identity card after, the computer program instructions are being transported by computer Perform also the computer during row:The only witness to be tested comprising the face for detecting is extracted from the image to be verified Face image and the only ID Card Image to be verified comprising the identity card for detecting;Also, first live body judges operation Including:Whether live body is belonged to based on the face detected described in the facial image judgement to be verified, second live body judges Operation includes:Based on the ID Card Image to be verified judge described in the identity card that detects whether belong to live body, the described 3rd Live body judges that operation includes:Detect based on described in the facial image to be verified and the ID Card Image comprehensive descision to be verified Whether the face and the identity card for detecting for going out belongs to live body on the whole.

In one embodiment, the computer program instructions make performed by the computer when being run by computer Based on the face and the inspection that detect described in the facial image to be verified and the ID Card Image comprehensive descision to be verified The step of whether identity card measured belongs to live body on the whole includes:By the facial image to be verified and the identity to be verified The first convolutional neural networks that card image input is trained, to obtain the face for detecting and the identity card for detecting Belong to the probability of live body on the whole;And the face that detects according to the determine the probability and the identity card for detecting Whether belong to live body on the whole.

In one embodiment, the computer program instructions when being run by computer perform also the computer: Training data is obtained, the training data includes positive sample image and negative sample image, and the positive sample image includes true people Face and real identity card, the negative sample image includes false face and real identity card;Extract from the positive sample image Only the positive sample facial image comprising face and only the positive sample ID Card Image comprising identity card;From the negative sample image Extract the only negative sample facial image comprising face and only the negative sample ID Card Image comprising identity card;And with the positive sample This facial image and the positive sample ID Card Image are positive sample, and with the negative sample facial image and the negative sample ID Card Image is negative sample, carries out neural metwork training to obtain first convolutional neural networks.

In one embodiment, make when being run by computer performed by the computer in the computer program instructions Extract from the positive sample image only comprising face positive sample facial image and only include identity card positive sample identity Before the step of card image, the computer program instructions when being run by computer perform also the computer:Calculating will Face in the positive sample image zooms to the positive sample scaling needed for default size, and scales according to the positive sample Ratio is zoomed in and out to the positive sample image;The computer is made when being run by computer in the computer program instructions Performed extracts the only negative sample facial image comprising face and the only negative sample comprising identity card from the negative sample image Before the step of this ID Card Image, the computer program instructions when being run by computer perform also the computer: Face in the negative sample image is zoomed to the negative sample scaling needed for the default size for calculating, and according to described Negative sample scaling is zoomed in and out to the negative sample image;Make when being run by computer in the computer program instructions The only face figure to be verified comprising the face for detecting is extracted from the image to be verified performed by the computer Before the step of picture and only ID Card Image to be verified comprising the identity card for detecting, the computer program instructions exist Perform also the computer when being run by computer:Calculating zooms to the face for detecting needed for the default size Image scaling ratio to be verified, and the image to be verified is zoomed in and out according to the image scaling ratio to be verified.

In one embodiment, make when being run by computer performed by the computer in the computer program instructions The detection image to be verified in face and the step of identity card after, the computer program instructions are being transported by computer Perform also the computer during row:If any face being not detected by the image to be verified or being not detected by Any identity card, then output re-executes the prompting of live body checking.

In one embodiment, the live body action is included in while blocking face with identity card and overturns and/or translate Identity card.

In one embodiment, the computer program instructions make performed by the computer when being run by computer The step of whether belonging to live body based on the face detected described in the facial image judgement to be verified includes:Will be described to be verified The second convolutional neural networks that facial image input is trained, to judge whether the face for detecting belongs to live body.

In one embodiment, the computer program instructions make performed by the computer when being run by computer The step of whether belonging to live body based on the identity card detected described in the ID Card Image judgement to be verified includes:Treat described The 3rd convolutional neural networks that checking ID Card Image input is trained, to judge whether the identity card for detecting belongs to living Body.

In one embodiment, the computer program instructions make performed by the computer when being run by computer Perform the first live body and judge that operation, the second live body judge that operation and the 3rd live body include the step of judging operation:According to arranging Order perform first live body and judge that operation, second live body judge that operation and the 3rd live body judge operation, such as Really arbitrary live body judges that the result for operating is no, then stop performing subsequent live body judgement operation.

In one embodiment, the plurality of live body judges that operation also includes that the 4th live body judges operation, wherein described the Four live bodies judge that operation includes:Based on the image to be verified judge the action of the identity card execution for detecting whether with institute State live body action command to match.

In one embodiment, the image to be verified is video, and the 4th live body judges that operation is based on the video In at least two frames carry out, first live body judges that operation, second live body judge that operation and the 3rd live body judge Operation is carried out based at least frame at least two frames.

Each module in live body checking system according to embodiments of the present invention can be by reality according to embodiments of the present invention The processor computer program instructions that store in memory of operation of electronic equipment of live body checking are applied realizing, or can be with The computer instruction stored in the computer-readable recording medium of computer program according to embodiments of the present invention is counted Realize when calculation machine runs.

Live body verification method according to embodiments of the present invention and device and identity identifying method and device, it is to be tested due to gathering The hand-held identity card of card object performs the image of live body action and carries out live body checking based on the image of collection, therefore in live body checking During can judge whether object to be verified belongs to live body with reference to the information that identity card brings, can so improve live body and test The accuracy of card.

Although the example embodiment by reference to Description of Drawings here, it should be understood that above-mentioned example embodiment is merely exemplary , and be not intended to limit the scope of the invention to this.Those of ordinary skill in the art can wherein carry out various changes And modification, it is made without departing from the scope of the present invention and spirit.All such changes and modifications are intended to be included in claims Within required the scope of the present invention.

Those of ordinary skill in the art are it is to be appreciated that the list of each example with reference to the embodiments described herein description Unit and algorithm steps, being capable of being implemented in combination in electronic hardware or computer software and electronic hardware.These functions are actually Performed with hardware or software mode, depending on the application-specific and design constraint of technical scheme.Professional and technical personnel Each specific application can be used different methods to realize described function, but this realization it is not considered that exceeding The scope of the present invention.

In several embodiments provided herein, it should be understood that disclosed apparatus and method, it can be passed through Its mode is realized.For example, apparatus embodiments described above are only schematic, for example, the division of the unit, and only Only a kind of division of logic function, can there is other dividing mode when actually realizing, such as multiple units or component can be tied Close or be desirably integrated into another equipment, or some features can be ignored, or do not perform.

In description mentioned herein, a large amount of details are illustrated.It is to be appreciated, however, that the enforcement of the present invention Example can be put into practice in the case of without these details.In some instances, known method, structure is not been shown in detail And technology, so as not to obscure the understanding of this description.

Similarly, it will be appreciated that in order to simplify the present invention and help understand one or more in each inventive aspect, exist To the present invention exemplary embodiment description in, the present invention each feature be grouped together into sometimes single embodiment, figure, Or in descriptions thereof.However, the method for the present invention should be construed to reflect following intention:It is i.e. required for protection The more features of feature that application claims ratio is expressly recited in each claim.More precisely, such as corresponding power As sharp claim reflects, its inventive point is can be with the spy of all features less than certain disclosed single embodiment Levy to solve corresponding technical problem.Therefore, it then follows it is concrete that thus claims of specific embodiment are expressly incorporated in this Separate embodiments of the embodiment, wherein each claim as the present invention itself.

It will be understood to those skilled in the art that in addition to mutually exclusive between feature, any combinations pair can be adopted All features and so disclosed any method disclosed in this specification (including adjoint claim, summary and accompanying drawing) Or all processes or unit of equipment are combined.Unless expressly stated otherwise, this specification (will including adjoint right Ask, make a summary and accompanying drawing) disclosed in each feature can, equivalent identical by offer or similar purpose alternative features replacing.

Although additionally, it will be appreciated by those of skill in the art that some embodiments described herein include other embodiments In included some features rather than further feature, but the combination of the feature of different embodiments means in of the invention Within the scope of and form different embodiments.For example, in detail in the claims, embodiment required for protection one of arbitrarily Can in any combination mode using.

The present invention all parts embodiment can be realized with hardware, or with one or more processor operation Software module realize, or with combinations thereof realization.It will be understood by those of skill in the art that can use in practice Microprocessor or digital signal processor (DSP) are realizing live body checking device according to embodiments of the present invention and authentication The some or all functions of some modules in device.The present invention is also implemented as performing method as described herein Some or all program of device (for example, computer program and computer program).It is such to realize the present invention Program can store on a computer-readable medium, or can have one or more signal form.Such letter Number can download from internet website and to obtain, or provide on carrier signal, or provide in any other form.

It should be noted that above-described embodiment the present invention will be described rather than limits the invention, and ability Field technique personnel can design without departing from the scope of the appended claims alternative embodiment.In the claims, Any reference markss between bracket should not be configured to limitations on claims.Word "comprising" is not excluded the presence of not Element listed in the claims or step.Word "a" or "an" before element does not exclude the presence of multiple such Element.The present invention can come real by means of the hardware for including some different elements and by means of properly programmed computer It is existing.If in the unit claim for listing equipment for drying, several in these devices can be by same hardware branch To embody.The use of word first, second, and third does not indicate that any order.These words can be explained and be run after fame Claim.

The above, the only specific embodiment of the present invention or the explanation to specific embodiment, the protection of the present invention Scope is not limited thereto, any those familiar with the art the invention discloses technical scope in, can be easily Expect change or replacement, all should be included within the scope of the present invention.Protection scope of the present invention should be with claim Protection domain is defined.

Claims (27)

1. a kind of live body verification method, including:
Random to generate live body action command, the live body action command is used to indicate that the hand-held identity card of object to be verified is performed accordingly Live body action;
Object to be verified performs the image of the live body action described in Real-time Collection, to obtain image to be verified;And
Determine whether the object to be verified is verified by live body based on the image to be verified.
2. live body verification method as claimed in claim 1, wherein, it is described determined based on the image to be verified it is described to be verified Whether object is included by live body checking:
Detect the face and identity card in the image to be verified;
Perform multiple live bodies and judge operation, wherein, the plurality of live body judges that operation includes that the first live body judges that operation, second live Body judges that operation and the 3rd live body judge operation, wherein, first live body judges that operation includes:Based on the image to be verified Whether the face that judgement is detected belongs to live body, and second live body judges that operation includes:Judged based on the image to be verified Whether the identity card for detecting belongs to live body, and the 3rd live body judges that operation includes:Sentenced based on the image synthesis to be verified Whether the disconnected face for detecting and the identity card for detecting belong to live body on the whole;And
Judge that each live body in operation judges that the judged result of operation determines the object to be verified according to the plurality of live body Whether verified by live body, if arbitrary live body judges that the judged result for operating is no, it is determined that the object to be verified does not lead to Experience of making a living is demonstrate,proved, and otherwise determines that the object to be verified is verified by live body.
3. live body verification method as claimed in claim 2, wherein, the face and body in the detection image to be verified After part card, the live body verification method also includes:Extract from the image to be verified and only include the face for detecting Facial image to be verified and only comprising the identity card for detecting ID Card Image to be verified;Also,
First live body judges that operation includes:Whether belonged to based on the face detected described in the facial image judgement to be verified In live body, second live body judges that operation includes:Based on the identity detected described in the ID Card Image judgement to be verified Whether card belongs to live body, and the 3rd live body judges that operation includes:Based on the facial image to be verified and the body to be verified Whether the face and the identity card for detecting detected described in part card image synthesis judgement belongs to live body on the whole.
4. live body verification method as claimed in claim 3, wherein, it is described based on the facial image to be verified and described to be tested Whether the face and the identity card for detecting detected described in card ID Card Image comprehensive descision belongs on the whole live body bag Include:
The first convolutional neural networks that the facial image to be verified and the ID Card Image input to be verified are trained, with The face and the identity card for detecting that detect described in obtaining belong on the whole the probability of live body;And
Whether the face and the identity card for detecting detected according to the determine the probability belongs to live body on the whole.
5. live body verification method as claimed in claim 4, wherein, the live body verification method also includes:
Training data is obtained, the training data includes positive sample image and negative sample image, and the positive sample image is comprising true Real face and real identity card, the negative sample image includes false face and real identity card;
The only positive sample facial image comprising face and only the positive sample body comprising identity card are extracted from the positive sample image Part card image;
The only negative sample facial image comprising face and only the negative sample body comprising identity card are extracted from the negative sample image Part card image;And
With the positive sample facial image and the positive sample ID Card Image as positive sample, and with the negative sample face figure Picture and the negative sample ID Card Image are negative sample, carry out neural metwork training to obtain first convolutional neural networks.
6. live body verification method as claimed in claim 5, wherein, only include people in described extraction from the positive sample image Before the positive sample facial image of face and only the positive sample ID Card Image comprising identity card, the live body verification method is also wrapped Include:
Face in the positive sample image is zoomed to the positive sample scaling needed for default size for calculating, and according to described Positive sample scaling is zoomed in and out to the positive sample image;
The only negative sample facial image comprising face and bearing only comprising identity card are extracted from the negative sample image described Before sample identity card image, the live body verification method also includes:
Calculate and the face in the negative sample image is zoomed to into negative sample scaling needed for the default size, and according to The negative sample scaling is zoomed in and out to the negative sample image;
It is described extract from the image to be verified only comprising the face for detecting facial image to be verified and only wrap Before ID Card Image to be verified containing the identity card for detecting, the live body verification method also includes:
The face for detecting is zoomed to the image scaling ratio to be verified needed for the default size for calculating, and according to institute State image scaling ratio to be verified to zoom in and out the image to be verified.
7. live body verification method as claimed in claim 3, wherein, it is described that the inspection is judged based on the facial image to be verified Whether the face measured belongs to live body includes:
The second convolutional neural networks that the facial image to be verified input is trained, to judge that the face for detecting is It is no to belong to live body.
8. live body verification method as claimed in claim 3, wherein, it is described to judge described based on the ID Card Image to be verified Whether the identity card for detecting belongs to live body includes:
The 3rd convolutional neural networks that the ID Card Image input to be verified is trained, to judge the identity for detecting Whether card belongs to live body.
9. live body verification method as claimed in claim 2, wherein, the face and body in the detection image to be verified After part card, the live body verification method also includes:
If any face is not detected by the image to be verified or any identity card is not detected by, output weight The new prompting for performing live body checking.
10. live body verification method as claimed in claim 2, wherein, the first live body of the execution judges that operation, the second live body are sentenced Disconnected operation and the 3rd live body judge that operation includes:
First live body is performed according to the order for arranging judge that operation, second live body judge operation and the 3rd work Body judges operation, if arbitrary live body judges that the result for operating is no, stops the subsequent live body of execution and judges operation.
The 11. live body verification methods as described in any one of claim 2 to 10, wherein, the plurality of live body judges that operation is also wrapped Include the 4th live body and judge operation, wherein the 4th live body judges that operation includes:The inspection is judged based on the image to be verified Whether the action that the identity card measured is performed matches with the live body action command.
12. live body verification methods as claimed in claim 11, wherein, the image to be verified be video, the 4th live body Judge that operation is carried out based at least two frames in the video, first live body judges that operation, second live body judge behaviour Make and the 3rd live body judges that operation is carried out based at least frame at least two frames.
The 13. live body verification methods as described in any one of claim 1 to 12, wherein, the live body action is included in uses identity Card overturns and/or translates identity card while blocking face.
A kind of 14. identity identifying methods, including the live body verification method as described in any one of claim 1 to 13, wherein, it is described Identity identifying method also includes:It is determined that the object to be verified by live body checking in the case of, judge from described to be verified Whether the face on identity card detected in image is consistent with the face detected from the image to be verified.
A kind of 15. live bodies verify device, including:
Directive generation module, for random live body action command is generated, and the live body action command is used to indicate object to be verified Hand-held identity card performs corresponding live body action;
Image to be verified obtains module, for the image that object to be verified described in Real-time Collection performs the live body action, to obtain Obtain image to be verified;And
Determining module is verified, for determining whether the object to be verified is tested by live body based on the image to be verified Card.
16. live bodies as claimed in claim 15 verify device, wherein, the determining module that is verified includes:
Detection sub-module, for detecting the image to be verified in face and identity card;
Live body judging submodule, for performing multiple live bodies operation is judged, wherein, the live body judging submodule includes:
First live body judging unit, for performing the first live body operation is judged, wherein, the first live body judging unit includes people Face judgment sub-unit, for judging whether the face for detecting belongs to live body based on the image to be verified;
Second live body judging unit, for performing the second live body operation is judged, wherein, the second live body judging unit includes body Part card judgment sub-unit, for judging whether the identity card for detecting belongs to live body based on the image to be verified;
3rd live body judging unit, for performing the 3rd live body operation is judged, wherein, the 3rd live body judging unit includes comprehensive Judgment sub-unit is closed, for based on the face and the identity for detecting detected described in the image synthesis judgement to be verified Whether card belongs to live body on the whole;And
Determination sub-module is verified, the judgement for each the live body judging unit in the live body judging submodule is tied Fruit determines whether the object to be verified is verified by live body, if the judged result of arbitrary live body judging unit is no, really The fixed object to be verified otherwise determines that the object to be verified is verified by live body not by live body checking.
17. live bodies as claimed in claim 16 verify device, wherein, the live body checking device also includes:First image is carried Delivery block, for the only facial image to be verified comprising the face for detecting being extracted from the image to be verified and only being wrapped ID Card Image to be verified containing the identity card for detecting;Also,
The face judgment sub-unit includes face determination component, for judging the detection based on the facial image to be verified Whether the face for going out belongs to live body, and the identity card judgment sub-unit includes identity card determination component, for based on described to be tested Whether the identity card detected described in card ID Card Image judgement belongs to live body, and the comprehensive descision subelement includes comprehensive descision Component, for based on the face detected described in the facial image to be verified and the ID Card Image comprehensive descision to be verified Whether belong to live body on the whole with the identity card for detecting.
18. live bodies as claimed in claim 17 verify device, wherein, the comprehensive descision component includes:
First input sub-component, for train the facial image to be verified and the ID Card Image input to be verified First convolutional neural networks, to obtain the face for detecting and the identity card for detecting the general of live body is belonged on the whole Rate;And
Live body determines sub-component, and the face and the identity card for detecting for detecting according to the determine the probability is whole Whether belong to live body on body.
19. live bodies as claimed in claim 18 verify device, wherein, the live body checking device also includes:
Training data acquisition module, for obtaining training data, the training data includes positive sample image and negative sample image, The positive sample image includes real human face and real identity card, and the negative sample image includes false face and true identity Card;
Second image zooming-out module, for extracting the only positive sample facial image and only comprising face from the positive sample image Positive sample ID Card Image comprising identity card;
3rd image zooming-out module, for extracting the only negative sample facial image and only comprising face from the negative sample image Negative sample ID Card Image comprising identity card;And
Training module, for the positive sample facial image and the positive sample ID Card Image as positive sample, and with institute It is negative sample to state negative sample facial image and the negative sample ID Card Image, carries out neural metwork training to obtain described first Convolutional neural networks.
20. live bodies as claimed in claim 19 verify device, wherein, the live body checking device also includes:
First Zoom module, for extracting from the positive sample image only comprising face in the second image zooming-out module Before positive sample facial image and only the positive sample ID Card Image comprising identity card, calculate the people in the positive sample image Face zooms to the positive sample scaling needed for default size, and according to the positive sample scaling to the positive sample image Zoom in and out;
Second Zoom module, for extracting from the negative sample image only comprising face in the 3rd image zooming-out module Before negative sample facial image and only the negative sample ID Card Image comprising identity card, calculate the people in the negative sample image Face zooms to the negative sample scaling needed for the default size, and according to the negative sample scaling to the negative sample Image is zoomed in and out;And
3rd Zoom module, only include the inspection for extracting from the image to be verified in described first image extraction module Before the facial image to be verified of the face measured and the only ID Card Image to be verified comprising the identity card for detecting, meter The face for detecting is zoomed to the image scaling ratio to be verified needed for the default size for calculation, and according to described to be tested Card image scaling ratio is zoomed in and out to the image to be verified.
21. live bodies as claimed in claim 17 verify device, wherein, the face determination component includes:
Second input sub-component, for the second convolutional neural networks for training the facial image input to be verified, to sentence Whether the disconnected face for detecting belongs to live body.
22. live bodies as claimed in claim 17 verify device, wherein, the identity card determination component includes:
3rd input sub-component, for the ID Card Image to be verified to be input into into the 3rd convolutional neural networks for training, with Whether the identity card detected described in judging belongs to live body.
23. live bodies as claimed in claim 16 verify device, wherein, the live body checking device also includes:
Prompting output module, if for any face being not detected by the image to be verified or being not detected by appointing What identity card, then output re-executes the prompting of live body checking.
The 24. live body checking devices as described in any one of claim 16 to 23, wherein, the live body judging submodule also includes 4th live body judging unit, for performing the 4th live body operation is judged, wherein the 4th live body judging unit includes identity card Action judgment sub-unit, for judged based on the image to be verified the action of the identity card execution for detecting whether with institute State live body action command to match.
25. live bodies as claimed in claim 24 verify devices, wherein, the image to be verified is video, the 4th live body Judging unit performs the 4th live body and judges operation based at least two frames in the video, and first live body judges single First, described second live body judging unit and the 3rd live body judging unit are based at least frame difference at least two frames Perform the first live body and judge that operation, second live body judge that operation and the 3rd live body judge operation.
26. live bodies as claimed in claim 15 verify device, wherein, the live body action is included in blocks face with identity card While overturn and/or translate identity card.
A kind of 27. identification authentication systems, including the live body checking device as described in any one of claim 15 to 26, wherein, institute State identification authentication system also include face concordance judge module, for it is described be verified determining module determine it is described to be tested In the case that card object is by live body checking, face on the identity card detected from the image to be verified is judged and from institute Whether consistent state the face detected in image to be verified.
CN201610927708.0A 2016-10-31 2016-10-31 Living body authentication method, identity authentication method and device CN106599772A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201610927708.0A CN106599772A (en) 2016-10-31 2016-10-31 Living body authentication method, identity authentication method and device

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201610927708.0A CN106599772A (en) 2016-10-31 2016-10-31 Living body authentication method, identity authentication method and device

Publications (1)

Publication Number Publication Date
CN106599772A true CN106599772A (en) 2017-04-26

Family

ID=58556164

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201610927708.0A CN106599772A (en) 2016-10-31 2016-10-31 Living body authentication method, identity authentication method and device

Country Status (1)

Country Link
CN (1) CN106599772A (en)

Cited By (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN107273794A (en) * 2017-04-28 2017-10-20 北京建筑大学 Live body discrimination method and device in a kind of face recognition process
CN107316029A (en) * 2017-07-03 2017-11-03 腾讯科技(深圳)有限公司 A kind of live body verification method and equipment
CN107609494A (en) * 2017-08-31 2018-01-19 北京飞搜科技有限公司 A kind of human face in-vivo detection method and system based on silent formula
CN107844748A (en) * 2017-10-17 2018-03-27 平安科技(深圳)有限公司 Auth method, device, storage medium and computer equipment
CN108182409A (en) * 2017-12-29 2018-06-19 北京智慧眼科技股份有限公司 Biopsy method, device, equipment and storage medium
CN108804884A (en) * 2017-05-02 2018-11-13 北京旷视科技有限公司 Identity authentication method, device and computer storage media

Citations (14)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
KR20020069697A (en) * 2001-02-27 2002-09-05 한국전자통신연구원 Apparatus and Method for Controlling Electrical Apparatus by using Bio-signal
WO2009110323A1 (en) * 2008-03-03 2009-09-11 日本電気株式会社 Living body judgment system, method for judging living body and program for judging living body
CN102622588A (en) * 2012-03-08 2012-08-01 无锡数字奥森科技有限公司 Dual-certification face anti-counterfeit method and device
CN103634120A (en) * 2013-12-18 2014-03-12 上海市数字证书认证中心有限公司 Method and system for real-name authentication based on face recognition
CN104361274A (en) * 2014-10-30 2015-02-18 深圳市富途网络科技有限公司 Identity authentication method and system on basis of video identification
CN105447532A (en) * 2015-03-24 2016-03-30 北京天诚盛业科技有限公司 Identity authentication method and device
CN105518713A (en) * 2015-02-15 2016-04-20 北京旷视科技有限公司 Living human face verification method and system, computer program product
CN105512632A (en) * 2015-12-09 2016-04-20 北京旷视科技有限公司 In vivo detection method and device
CN105518708A (en) * 2015-04-29 2016-04-20 北京旷视科技有限公司 Method and equipment for verifying living human face, and computer program product
CN105518711A (en) * 2015-06-29 2016-04-20 北京旷视科技有限公司 In-vivo detection method, in-vivo detection system, and computer program product
CN105612533A (en) * 2015-06-08 2016-05-25 北京旷视科技有限公司 In-vivo detection method, in-vivo detection system and computer programe products
CN105930710A (en) * 2016-04-22 2016-09-07 北京旷视科技有限公司 Living body detection method and device
US20160277397A1 (en) * 2015-03-16 2016-09-22 Ricoh Company, Ltd. Information processing apparatus, information processing method, and information processing system
CN105989263A (en) * 2015-01-30 2016-10-05 阿里巴巴集团控股有限公司 Method for authenticating identities, method for opening accounts, devices and systems

Patent Citations (14)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
KR20020069697A (en) * 2001-02-27 2002-09-05 한국전자통신연구원 Apparatus and Method for Controlling Electrical Apparatus by using Bio-signal
WO2009110323A1 (en) * 2008-03-03 2009-09-11 日本電気株式会社 Living body judgment system, method for judging living body and program for judging living body
CN102622588A (en) * 2012-03-08 2012-08-01 无锡数字奥森科技有限公司 Dual-certification face anti-counterfeit method and device
CN103634120A (en) * 2013-12-18 2014-03-12 上海市数字证书认证中心有限公司 Method and system for real-name authentication based on face recognition
CN104361274A (en) * 2014-10-30 2015-02-18 深圳市富途网络科技有限公司 Identity authentication method and system on basis of video identification
CN105989263A (en) * 2015-01-30 2016-10-05 阿里巴巴集团控股有限公司 Method for authenticating identities, method for opening accounts, devices and systems
CN105518713A (en) * 2015-02-15 2016-04-20 北京旷视科技有限公司 Living human face verification method and system, computer program product
US20160277397A1 (en) * 2015-03-16 2016-09-22 Ricoh Company, Ltd. Information processing apparatus, information processing method, and information processing system
CN105447532A (en) * 2015-03-24 2016-03-30 北京天诚盛业科技有限公司 Identity authentication method and device
CN105518708A (en) * 2015-04-29 2016-04-20 北京旷视科技有限公司 Method and equipment for verifying living human face, and computer program product
CN105612533A (en) * 2015-06-08 2016-05-25 北京旷视科技有限公司 In-vivo detection method, in-vivo detection system and computer programe products
CN105518711A (en) * 2015-06-29 2016-04-20 北京旷视科技有限公司 In-vivo detection method, in-vivo detection system, and computer program product
CN105512632A (en) * 2015-12-09 2016-04-20 北京旷视科技有限公司 In vivo detection method and device
CN105930710A (en) * 2016-04-22 2016-09-07 北京旷视科技有限公司 Living body detection method and device

Cited By (9)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN107273794A (en) * 2017-04-28 2017-10-20 北京建筑大学 Live body discrimination method and device in a kind of face recognition process
CN108804884A (en) * 2017-05-02 2018-11-13 北京旷视科技有限公司 Identity authentication method, device and computer storage media
CN107316029A (en) * 2017-07-03 2017-11-03 腾讯科技(深圳)有限公司 A kind of live body verification method and equipment
CN107316029B (en) * 2017-07-03 2018-11-23 腾讯科技(深圳)有限公司 A kind of living body verification method and equipment
CN107609494A (en) * 2017-08-31 2018-01-19 北京飞搜科技有限公司 A kind of human face in-vivo detection method and system based on silent formula
CN107844748A (en) * 2017-10-17 2018-03-27 平安科技(深圳)有限公司 Auth method, device, storage medium and computer equipment
CN107844748B (en) * 2017-10-17 2019-02-05 平安科技(深圳)有限公司 Auth method, device, storage medium and computer equipment
WO2019075840A1 (en) * 2017-10-17 2019-04-25 平安科技(深圳)有限公司 Identity verification method and apparatus, storage medium and computer device
CN108182409A (en) * 2017-12-29 2018-06-19 北京智慧眼科技股份有限公司 Biopsy method, device, equipment and storage medium

Similar Documents

Publication Publication Date Title
Zhang Automated biometrics: Technologies and systems
Song et al. A finger-vein verification system using mean curvature
CN103383723B (en) Method and system for spoof detection for biometric authentication
US7539330B2 (en) Multispectral liveness determination
CN102414698B (en) To the controlled access of the function of wireless device
US7831072B2 (en) Multispectral imaging biometrics
US8908977B2 (en) System and method for comparing images
De Marsico et al. Mobile iris challenge evaluation (MICHE)-I, biometric iris dataset and protocols
US20050265586A1 (en) Multispectral biometric imaging
EP2397969A1 (en) Multispectral imaging biometrics
CN100361131C (en) Information processing apparatus and information processing method
Bigun et al. Multimodal biometric authentication using quality signals in mobile communications
US8856541B1 (en) Liveness detection
CN102622588B (en) Dual-certification face anti-counterfeit method and device
CN1437161B (en) Personal recognition method, personal recognition apparatus
US20040125994A1 (en) Method for forgery recognition in fingerprint recognition by using a texture classification of gray scale differential images
KR101309889B1 (en) Texture features for biometric authentication
EP2546782B1 (en) Liveness detection
US10095927B2 (en) Quality metrics for biometric authentication
Komulainen et al. Complementary countermeasures for detecting scenic face spoofing attacks
Ferrara et al. The magic passport
CN101669824B (en) Biometrics-based device for detecting indentity of people and identification
CN1849614A (en) Facial liveness assessment system
Revett Behavioral biometrics: a remote access approach
US9298912B2 (en) System and method for distinguishing human swipe input sequence behavior and using a confidence value on a score to detect fraudsters

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
CB02 Change of applicant information
CB02 Change of applicant information

Address after: 100190 Beijing, Haidian District Academy of Sciences, South Road, No. 2, block A, No. 313

Applicant after: MEGVII INC.

Applicant after: Beijing maigewei Technology Co., Ltd.

Address before: 100190 Beijing, Haidian District Academy of Sciences, South Road, No. 2, block A, No. 313

Applicant before: MEGVII INC.

Applicant before: Beijing aperture Science and Technology Ltd.