CN106529436A - Identity consistency authentication method and device, and mobile terminal - Google Patents

Abstract

The invention discloses an identity consistency authentication method based on multi-mode biological recognition, and the method is suitable to be implemented in a terminal comprising composite imaging equipment. The composite imaging equipment is suitable to divide an incident light source into visible light and infrared light through a regional waveband filter module, and obtains an image A in a mode of visible light imaging and an image B in a mode of infrared light imaging. The method comprises the steps: respectively the image A in the mode of visible light imaging and the image B in the mode of infrared light imaging of a user, and an image sequence in a switching process of the two imaging modes; obtaining the movement vector information of the image sequence and/or the similarity information of the image A and the image B; and carrying out the identity consistency authentication of the user in various types of biological recognition according to the movement vector information and/or the similarity information. The invention also discloses a corresponding identity consistency authentication device and a mobile terminal.

Description

A kind of identity coherence authentication method, device and mobile terminal

Technical field

The present invention relates to field of biological recognition, consistent more particularly, to a kind of identity based on multi-mode bio-identification Property authentication method, device and mobile terminal.

Background technology

In recent years, China the Internet financial development is presented gesture like a raging fire, is increasingly becoming the focus of social concerns.But If remotely the problem of opening an account cannot be properly settled, the Internet model finance of the overwhelming majority will have to be walked under loop line again, So which loses internet community.It can be seen that remotely opening an account most important for the Internet financial development.

Biometrics identification technology provides technology guarantee for the identity verification remotely opened an account.Recognition of face, iris identification, The biometrics identification technologies such as fingerprint recognition are widely used in safety-security area, but financial field will relative to safety-security area The safe class asked is higher, and for open financial sector, at one, potential safety hazard may threaten whole financial sector Safety.Long-distance identity-certifying is completed using the biological characteristic of single mode there is safe drawback and application limitation, such as rainbow Film identification is more sensitive to identification distance, and recognition of face performance is easily affected by ambient light change again.Therefore, on mobile terminals Two kinds of (or two or more) biometrics identification technologies of fusion, to improve the Stability and veracity of identification, to mobile Internet Ensure that user information safety is significant using biometrics identification technology under environment.

Due to the special physiological make-up of iris, iris identification is typically using the near infrared light conduct of 760nm-880nm wave bands Light source, is imaged using special near-infrared camera module.Like this, the front panel of mobile terminal (smart mobile phone etc.) needs Two holes are opened, one is used for placing the visible image capturing head mould group autodyned, and one is used for placing near-infrared camera module, and this is just Cause Hardware Design complexity and cost all significantly to rise, and be unfavorable for the popularization applied.Meanwhile, separate iris The complexity and speed of iris identification and face recognition process fusion can also be increased with man face image acquiring equipment, user's body is affected Test.

The content of the invention

For this purpose, the present invention provides a kind of based on the identity coherence authentication method of multi-mode bio-identification, device and movement Terminal, to try hard to solve or at least alleviate at least one problem for existing above.

According to an aspect of the invention, there is provided a kind of identity coherence authenticating party based on multi-mode bio-identification Method, is suitable to perform in the terminal containing compound imaging device, and the compound imaging device is suitable to using subregion subrane Incident light source is divided into visible light wave range and near infrared light wave band by optical filter box, and is respectively obtained under visual light imaging pattern Image B under image A and near infrared light imaging pattern.The method includes：User to be identified is obtained respectively in visual light imaging mould Image A under formula, the image B under near infrared imaging pattern, and the image sequence of two imaging pattern handoff procedures；Analysis Obtain the similarity information of the motion vector information and/or image A and image B of described image sequence；And according to motion vector Information and/or similarity information carry out identity coherence certification to user.

Alternatively, in the method according to the invention, according to motion vector information and/or the similarity information to user The step of carrying out identity coherence certification can be using any one or more of situations below：(1) determine whether continuous many The motion vector of frame image sequence is unsatisfactory for pre-provisioning request, if so, then reminds user identity concordance authentification failure to be identified； (2) judge whether image A is consistent with the user identity in image B according to similarity information；If it is not, then reminding user's body to be identified Part concordance authentification failure；(3) the first analysis result r of motion vector information is calculated respectively₁With the similarity information The second analysis result r₂, and according to r=r₁*x+r₂* y is calculated and treats identifying user and carry out the final of identity full handshake authentication As a result, wherein, x and y are the weights of respective items.

Alternatively, in the method according to the invention, judge whether the motion vector of image sequence meets pre-provisioning request Operation includes：The image sequence for getting is divided into into multiple block images per frame；What is moved in determining image sequence divides The motion vector of block image and block image；And motion vector direction is horizontal and vector magnitude in difference sequence of computed images More than block image number M and block image number N for moving of first threshold, and the ratio of M and N；If the ratio More than Second Threshold, then judge that the motion vector of the frame image sequence is unsatisfactory for pre-provisioning request.

Alternatively, in the method according to the invention, the block image moved in determining image sequence and block diagram The step of motion vector direction of picture, includes：The block image at a certain position is chosen from image sequence；For selected Block image, obtains the same position block image in same position from previous frame image, and this is with the more of position block image Individual partition neighborhood image, wherein, with position block image and partition neighborhood image collectively as selected block image contrast Block image；The dependency of all contrast block images in selected block image and previous frame image is calculated, and is determined The position of the contrast block image of correlation maximum；If the position of the contrast block image of correlation maximum and selected piecemeal Picture position is different, then judge that the contrast block image is moved, and its motion vector direction points to selected block image, Its motion vector size is the distance between two positions.

Alternatively, in the method according to the invention, the step of being calculated the similarity information of image A and image B is wrapped Include：Carry out iris and canthus positioning respectively to image A and B, obtain iris center therein and canthus position；According to the rainbow of eyes Center membrane and canthus position cut out the area-of-interest comprising eyes respectively from image A and B；Respectively by image A and B Area-of-interest is rotated to horizontal direction, obtains image C and D；Respectively centered on canthus and iris center calculate image C and D In multiple dimensioned higher-dimension local binary pattern LBP features, and using LBP features cascade as image characteristic vector；And The similarity of the characteristic vector of image C and D is calculated respectively.

Alternatively, in the method according to the invention, the step of cutting out the area-of-interest comprising eyes from image Including：The rectangular area comprising eyes is cut out from image as the area-of-interest of the image, wherein, the long side of rectangle with The equal first distance in iris center of eyes, minor face and the equal second distance in its immediate iris center.

Alternatively, in the method according to the invention, the step of multiple dimensioned higher-dimension LBP features for calculating image, includes：Point Respective image pyramid is not built based on image C and D；Respectively with eyes in described image is pyramidal per tomographic image Iris center and canthus centered on interested area division；Each area-of-interest for dividing is divided into into the rectangle of m × m Block；The LBP histogram features in each rectangular block are counted, the LBP histogram features of rectangular block level are obtained；By all rectangular block levels LBP histogram features couple together, obtain the LBP histogram features of key point level；By each key point in every tomographic image LBP histogram features be linked in sequence, obtain the LBP histogram features of level；And by the LBP Nogatas of all levels Figure feature is linked in sequence by image level, obtains the characteristic vector of final image.

Alternatively, in the method according to the invention, identity coherence certification is carried out to user according to similarity information Step includes：The similarity of the characteristic vector of image C and D is calculated, if similarity is more than the 3rd threshold value, process decision chart is as in A and B User identity to be identified it is consistent, otherwise identity is inconsistent.Wherein, the similarity of the characteristic vector of image C and D includes two figures At least one in the degree of association of picture, card side's coefficient, intersecting coefficient and Pasteur's distance.

Alternatively, in the method according to the invention, image A includes the facial information of user to be identified, and image B includes treating The binocular information of identifying user, the method also include：Checked to image A according to the true head portrait of the user stored in data base Verified；If upchecking, under near infrared light imaging pattern, user's adjustment face to be identified and complex imaging is guided to set Standby angle and/or distance, until obtain the image B comprising its eyes；And according to the true of the user stored in data base Real iris image is verified to image B.

According to a further aspect in the invention, there is provided a kind of identity coherence certification based on multi-mode bio-identification is filled Put, be suitable to perform in the terminal containing compound imaging device, the compound imaging device is suitable to the filter using subregion subrane Incident light source is divided into visible light wave range and near infrared light wave band by mating plate component, and respectively obtains the figure under visual light imaging pattern Image under picture and near infrared light imaging pattern, the device include：Man face image acquiring unit, is suitable in visual light imaging pattern Lower image A of the collection comprising user's face information to be identified；Iris image acquiring unit, is suitable under near infrared light imaging pattern Image B of the collection comprising user's binocular information to be identified；Motion vector analysis unit, is suitable to two kinds of imaging patterns of acquisition and switched The image sequence of journey, and analysis obtains the motion vector information of described image sequence；Similarity analysis unit, be suitable to image A and The similarity information of image B；And identity coherence authentication ' unit, it is suitable to according to motion vector information and/or similarity information Identity coherence certification is carried out to user.

Alternatively, in a device in accordance with the invention, identity coherence authentication ' unit is suitable for use with situations below one Plant or various identifying users for the treatment of carry out identity coherence certification：(1) judge whether the motion arrow for continuously having multiple image sequence Amount is unsatisfactory for pre-provisioning request, if so, then reminds user identity concordance authentification failure to be identified；(2) sentenced according to similarity information Whether disconnected image A is consistent with the user identity in image B；If it is not, then reminding authenticating user identification concordance failure to be identified； (3) the first analysis result r of motion vector information is calculated respectively₁With the second analysis result r of the similarity information₂, and According to r=r₁*x+r₂* y is calculated and treats identifying user and carry out the final result of identity coherence certification, and wherein, x and y is right Answer the weights of item.

Alternatively, in a device in accordance with the invention, the image sequence that motion vector analysis unit is suitable to get is every Frame is divided into multiple block images；The block image moved in determining image sequence and the motion vector of block image；With And respectively in sequence of computed images motion vector direction be horizontal and vector magnitude more than first threshold block image number M with Block image number N for moving, and the ratio of M and N；If the ratio is more than Second Threshold, the two field picture sequence is judged The motion vector of row is unsatisfactory for pre-provisioning request.

Alternatively, in a device in accordance with the invention, motion vector analysis unit is suitable to choose a certain from image sequence Block image at position；For selected block image, the same position point in same position is obtained from previous frame image Block image, and this is with multiple partition neighborhood images of position block image, the same position block image and the partition neighborhood figure As the contrast block image collectively as selected block image；Calculate in selected block image and previous frame image The dependency of all contrast block images, and determine the position of the contrast block image of correlation maximum；If correlation maximum The position of contrast block image is different from selected block image position, then judge that the contrast block image is moved, its Motion vector direction points to selected block image.

Alternatively, in a device in accordance with the invention, similarity analysis unit is suitable to carry out iris to image A and B respectively Position with canthus, obtain iris center therein and canthus position；According to the iris center and canthus position of eyes respectively from figure As cutting out the area-of-interest comprising eyes in A and B；Respectively the area-of-interest in image A and B is rotated to level side To obtaining image C and D；Multiple dimensioned higher-dimension local two centered on canthus and iris center in calculating image C and D is entered respectively Molding formula LBP feature, and the LBP features are cascaded into the characteristic vector as image；And the feature of image C and D is calculated respectively The similarity of vector.

Alternatively, in a device in accordance with the invention, similarity analysis unit is suitable to cut out comprising eyes from image Rectangular area as the image area-of-interest, wherein, the long side of rectangle and the iris center of eyes at a distance of first away from From, minor face and the equal second distance in its immediate iris center.

Alternatively, in a device in accordance with the invention, similarity analysis unit is suitable to the structure based on image C and D respectively Build respective image pyramid；Drawn centered on the iris center and canthus of eyes in every tomographic image of image pyramid respectively Divide area-of-interest；Each area-of-interest for dividing is divided into into the rectangular block of m × m；The LBP counted in each rectangular block is straight Square figure feature, obtains the LBP histogram features of rectangular block level；The LBP histogram features of all rectangular block levels are coupled together, is obtained To the LBP histogram features of key point level；The LBP histogram features of each key point in every tomographic image have been linked in sequence Come, obtain the LBP histogram features of level；And the LBP histogram features of all levels have been linked in sequence by image level Come, obtain the characteristic vector of final image.

Alternatively, in a device in accordance with the invention, similarity analysis unit is suitable to calculate the characteristic vector of image C and D Similarity；Identity coherence authentication ' unit be suitable to similarity be more than three threshold values when process decision chart as A and B in use to be identified Family identity is consistent, and otherwise identity is inconsistent.Wherein, image C includes the related of two images to the similarity of the characteristic vector of D At least one in degree, card side's coefficient, intersecting coefficient and Pasteur's distance.Alternatively, in a device in accordance with the invention, image A Including the facial information of user to be identified, image B includes the binocular information of user to be identified, and the device also includes：Recognition of face Unit, is suitable to verify image A according to the true head portrait of the user stored in data base；Position guidance unit, is suitable to After recognition of face is verified, user's adjustment face to be identified and compound imaging device is guided under near infrared light imaging pattern Angle and/or distance, until obtain the image B comprising its eyes；And iris identification unit, it is suitable to deposit according in data base The true iris image of the user of storage is verified to image B.

According to another aspect of the invention, there is provided a kind of mobile terminal, including biological based on multi-mode as above The identity coherence authentication device of identification；And compound imaging device, mutually couple with the identity full handshake authentication device, including： Lens assembly, including the optical lenses of fixed focal length；Optical filter box, including the visible ray for allowing the light of visible light wave range to pass through The near infrared light bandpass region that bandpass filter and permission near infrared light wave band pass through；And imageing sensor, including visible ray Transitional region between imaging region, near infrared light imaging region and the two regions, wherein, it is seen that photoimaging area is can See that, to being imaged by the visible ray of visible band pass filter under photoimaging pattern, near infrared light imaging region is in near-infrared To being imaged by the near infrared light of near infrared light bandpass filter under photoimaging pattern.

The present invention obtains near infrared light image and the visible images of user to be identified using compound imaging device.According to reality The needs of border application, imageing sensor optionally can only export the visible images that visual light imaging region obtains, or The near infrared light image that near infrared light region obtains, for carrying out visible ray recognition of face or the near infrared light iris of single mode Identification.From on hardware cost, this programme does not increase extra CCD camera assembly, will not be to the normal use mobile terminal of user Impact.

Identifying user is treated when authentication is carried out, the face figure of the user is obtained first under visual light imaging pattern Picture, and compare with face database.After recognition of face passes through, use to be identified is obtained under near infrared light imaging pattern The iris image at family, and compare with iris image database.In order to ensure the handoff procedure in face and iris identification The concordance of personnel to be identified, the coloured image that the present invention obtains visual light imaging are similar to the gray level image of near infrared imaging Degree information, and identity coherence certification is carried out to user according to similarity information；Imaging pattern twice can also be obtained to switch The motion vector information of the image sequence of journey, and identity coherence certification is carried out according to the motion vector information；Can also be by one Determining comprehensive both information of weight proportion carries out identity coherence certification.That is, convergence strategy of the present invention by both information, The multi-level authentication level of security that improve user in many aspects.

Description of the drawings

In order to realize above-mentioned and related purpose, some illustrative sides are described herein in conjunction with explained below and accompanying drawing Face, indicate in terms of these can be to put into practice principles disclosed herein various modes, and all aspects and its equivalent aspect It is intended to fall under in the range of theme required for protection.By being read in conjunction with the accompanying detailed description below, the disclosure it is above-mentioned And other purposes, feature and advantage will be apparent from.Throughout the disclosure, identical reference generally refers to identical Part or element.

Fig. 1 shows the structured flowchart of mobile terminal 1 according to an embodiment of the invention 00；

Fig. 2 shows the schematic diagram of compound imaging device according to an embodiment of the invention 200；

Fig. 3 shows the identity coherence authentication method based on multi-mode bio-identification according to an embodiment of the invention 300 flow chart；

Fig. 4 a and 4b respectively illustrate the image A and image B schematic diagrams collected by compound imaging device 200；

Fig. 5 a and 5b respectively illustrate the iris center in image A and image B and canthus schematic diagram；

Fig. 6 a show the schematic diagram for cutting out area-of-interest from image, and Fig. 6 b show and revolve the area-of-interest Go to the schematic diagram after horizontal direction；

Fig. 7 a and 7b respectively illustrate from image A and image B cut out comprising the interested of user's eyes to be identified Area schematic；

Fig. 8 shows the schematic diagram of the LBP features for building image pyramid and extracting iris center；

Fig. 9 shows the identity coherence authentication device based on multi-mode bio-identification according to an embodiment of the invention 900 block diagram.

Specific embodiment

The exemplary embodiment of the disclosure is more fully described below with reference to accompanying drawings.Although showing the disclosure in accompanying drawing Exemplary embodiment, it being understood, however, that may be realized in various forms the disclosure and should not be by embodiments set forth here Limited.On the contrary, there is provided these embodiments are able to be best understood from the disclosure, and can be by the scope of the present disclosure Complete conveys to those skilled in the art.

The present invention proposes a kind of identity coherence authentication device based on multi-mode bio-identification, and which may reside within various In terminal unit, such as mobile phone, flat board or notebook computer etc., it is also possible to reside in computing device.Fig. 1 is shown according to this The structured flowchart of the mobile terminal 1 00 of bright one embodiment.As described in Figure 1, mobile terminal 1 00 includes：Memory interface 102, One or more data processors, image processor and/or CPU 104, and peripheral interface 106.Memorizer connects Mouthfuls 102, one or more processors 104 and/or peripheral interface 106 can both be discrete components, it is also possible to be integrated in one or In multiple integrated circuits.In the mobile terminal 100, various elements can be by one or more communication bus or holding wire come coupling Close.Sensor, equipment and subsystem may be coupled to peripheral interface 106, to help realize several functions.For example, motion is passed Sensor 110, optical sensor 112 and range sensor 114 may be coupled to peripheral interface 106, to facilitate orientation, illumination and find range Etc. function.Other sensors 116 equally can be connected with peripheral interface 106, such as alignment system (such as GPS), temperature Degree sensor, biometric sensor or other sensor devices, it is possible thereby to help implement the function of correlation.

Camera sub-system 120 and optical pickocff 122 can be used for the camera of convenient such as recording photograph and video clipping The realization of function, wherein described camera sub-system and optical pickocff can for example be charge-coupled image sensor (CCD) or complementary gold Category oxide semiconductor (CMOS) optical pickocff.Can help realize by one or more radio communication subsystems 124 Communication function, wherein radio communication subsystem can include radio-frequency transmitter and transmitter and/or light (such as infrared) receiver And transmitter.The particular design of radio communication subsystem 124 and embodiment can depend on that mobile terminal 1 00 is supported Individual or multiple communication networks.For example, mobile terminal 1 00 can include being designed to support GSM network, GPRS network, EDGE nets The communication subsystem 124 of network, Wi-Fi or WiMax network and BlueboothTM networks.Audio subsystem 126 can with raise one's voice Device 128 and mike 130 are coupled, to help implement the function of enabling voice, such as speech recognition, speech reproduction, number Word is recorded and telephony feature.

I/O subsystems 140 can include touch screen controller 142 and/or one or more other input controllers 144. Touch screen controller 142 may be coupled to touch screen 146.For example, the touch screen 146 and touch screen controller 142 can be with The contact and movement or time-out for carrying out therewith is detected using any one of various touch-sensing technologies, wherein senses skill Art includes but is not limited to capacitive character, resistive, infrared and surface acoustic wave technique.One or more other input controllers 144 May be coupled to other input/control devicess 148, for example one or more buttons, rocker switch, thumb wheel, infrared port, The pointer device of USB port, and/or instruction pen etc.One or more of button (not shown)s can be included for controlling The up/down button of 130 volume of speaker 128 and/or mike.

Memory interface 102 can be coupled with memorizer 150.The memorizer 150 can be deposited including high random access Reservoir and/or nonvolatile memory, such as one or more disk storage equipments, one or more optical storage apparatus, and/ Or flash memories (such as NAND, NOR).Memorizer 150 can with storage program area 152, such as Android, IOS or The operating system of Windows Phone etc.The operating system 152 can be included for processing basic system services and execution Depend on the instruction of the task of hardware.Memorizer 150 can also be stored and apply 154.These are applied in operation, can be from memorizer 150 are loaded on processor 104, and run on the operating system run via processor 104, and utilize operating system And the interface that bottom hardware is provided realizes the desired function of various users, such as instant messaging, web page browsing, pictures management etc.. It is operating system offer, or that operating system is carried using being independently of.In addition, being installed to shifting using 154 When in dynamic terminal 100, it is also possible to add drive module to operating system.

It is above-mentioned it is various apply in 154, a kind of application therein is related to the present invention based on multi-mode bio-identification Identity coherence authentication device 900.In certain embodiments, mobile terminal 1 00 is configured to perform base of the invention In the identity coherence authentication method 300 of multi-mode bio-identification.

When iris identification is carried out using mobile terminal 1, it usually needs individually near-infrared photographic head carries out iris imaging, Can not be multiplexed with the photographic head of existing visual light imaging (spectral frequency is in 380-760nm) that (such as smart mobile phone is existing The preposition colour imagery shot having), undoubtedly increased the structure complexity of mobile terminal.And propose in this programme a kind of by moving The single camera of dynamic terminal can carry out the compound imaging device 200 of near infrared light and visible ray two waveband, so as to effectively solving The problem.

Usually, compound imaging device 200 is mutually coupled with identity coherence authentication device 900, for performing based on multimode The identity coherence authentication method 300 of formula bio-identification.Further, the equipment 200 is arranged in mobile terminal screen front Side, such as the top of screen or bottom of screen.Near-infrared light source can include one in 780～880nm wavelength bands Or many infrared LEDs, alternatively, near-infrared light source is positioned under the touch pad of mobile terminal, for example, the front of mobile terminal Button bottom (such as Home key bottom), it is not necessary to cause to affect visual appearance in the special perforate of other positions.

Fig. 2 shows the structural frames of the compound imaging device 200 of mobile terminal 1 according to an embodiment of the invention 00 Figure.As shown in Fig. 2 the compound imaging device 200 includes：Successively along the lens assembly 230, filter set of input path arrangement Part 220 and imageing sensor 210.Wherein, optical filter box 220 includes visible band pass filter 221 and near infrared light band logical Optical filter 222 (being illustrated as being filled with backslash), imageing sensor 210 include the region a (figures near infrared light imaging In be shown as being filled with backslash), the region b for visual light imaging and transitional region c between a, b region (illustrate It is to be filled with horizontal line).

Full spectrum light line is incident to be passed through lens assembly 230 and reaches optical filter box 220, wherein visible band pass filter 221 permissions visible ray (for example, wavelength is 380-760nm) pass through, and near infrared light bandpass filter 222 allows near infrared light (for example, wavelength is 780-880nm) passes through.Visible band pass filter 221 and near infrared light bandpass filter 222 can pass through Plated film is realizing.Substantially it is imaged in the region b of imageing sensor by the visible ray of visible band pass filter 221, and is passed through The near infrared light of near infrared light bandpass filter 222 is imaged in the region a of imageing sensor substantially.Can using image processing software To make a distinction the imaging of the region b and a of imageing sensor respectively, wherein the imaging of region b is corresponding to normal visible ray Imaging, such as user carries out imaging during daily auto heterodyne using mobile terminals such as mobile phones, and the imaging of region a is corresponding near The imaging of imaging under infrared mode, such as user when iris identification is carried out using mobile phone.As such, it is possible to easily realization can The switching seen between light and near infrared light imaging, switches optical filter without being equipped with moving component, is greatly improved steady It is qualitative.

Previously mentioned, for open financial sector, recognition of face performance combines iris identification etc., and other are multi-modal Living things feature recognition mode it is especially necessary；Such as advanced row recognition of face, then carry out iris identification to further confirm that.But answering With during but can there are the following problems：As recognition of face and iris identification are to enter in two stages over time and space Capable, if In vivo detection algorithm is broken, lawless person can carry out face knowledge with the photo and identity document using other people Other illegal operation.For example, (credit card is such as done) when remotely opening an account and can require the facial image and iris figure of typing account holder Picture, if In vivo detection algorithm when lawless person breaks through recognition of face, it is possible to using other people face's picture (such as identity License) recognition of face, then the iris image of typing oneself is carried out, open an account so as to be completed with other people names.Although two images All typings, but be not same person actually in the two images.So, many of face and iris are used under this mode There is potential safety hazard in model organism identification, cannot ensure the concordance of personnel to be identified in its handoff procedure.

For this purpose, present invention further proposes a kind of identity coherence certificate scheme based on multi-mode bio-identification, with Ensure the concordance of the personnel to be identified of the handoff procedure of face and iris identification.

Fig. 3 shows the identity coherence authentication method based on multi-mode bio-identification according to an embodiment of the invention 300 flow chart.As shown in figure 3, the method starts from step S310.

In step S310, image A of the user to be identified under visual light imaging pattern is obtained respectively, near infrared imaging Image B under pattern, and the image sequence of two imaging pattern handoff procedures.Wherein, image A generally includes user to be identified Facial information, it typically is coloured image；Image B generally includes the binocular information of user to be identified, such as iris information, and which leads to It is often gray level image.

Under visual light imaging pattern, mobile terminal gathers the face figure of user to be identified by compound imaging device 200 Picture A, as shown in fig. 4 a.Before step S310, account holder can read the information with its ID card information of remote visiting system, client (head portrait, name, identification card number etc.) uploads onto the server, and compares with the information in Ministry of Public Security data base, it is ensured that its certificate is true Reality.After getting face-image A, man-to-man face is carried out with the head portrait in the ID card information submitted to and is matched, it is ensured that The verity of personnel identity to be opened an account.Certainly, either contrast with the data base of the Ministry of Public Security, or with the body that uploads in server Part information contrast, is construed as comparing with the true head portrait prestored in data base.It should be noted that of the invention Face recognition algorithms are not restricted, any face recognition algorithms can be combined with embodiments of the invention, to complete people Face identification step.

After recognition of face passes through, under near infrared light imaging pattern can be included by the collection of compound imaging device 200 and treated The image B of identifying user eyes, as shown in Figure 4 b.Generally, when gathering facial image and iris image using compound imaging device, It is to people and the angle of collection terminal or required distance and different, so before collection iris image, user to be identified can be guided The angle and/or distance between its face and compound imaging device is adjusted, until collecting the image B comprising its eyes.Certainly, Obtain iris image after, it is also desirable to carry out the identification of iris image, specifically, and with data base in store the user it is true Real iris image is compared, and further determines that the identity of user to be identified.It should be noted that the present invention is to Algorithm of Iris Recognition It is not restricted, any Algorithm of Iris Recognition can be combined with embodiments of the invention, to complete iris identification step.

Although can significantly ensure the verity of user to be identified by recognition of face and iris identification, due to What face and iris identification were carried out in two stages over time and space, so need to ensure cutting for face and iris capturing The concordance of personnel to be identified during changing, in order to avoid cause potential safety hazard.The present invention first obtains the colored human face figure under visible ray Picture, reminds user to enter near-infrared iris capturing pattern, analyzes the motion vector of the handoff procedure image sequence, afterwards using fortune The combined strategy of dynamic vector information and similarity information carries out identity full handshake authentication to user.

Specifically, in step s 320, analysis obtains the motion vector information and/or image A and figure of described image sequence As the similarity information of B；And in step S330, according to the motion vector information and/or the similarity information to user Carry out the identity coherence certification under various bio-identification patterns.Wherein, motion vector information can include described image sequence Motion vector direction and motion vector magnitude.

Further, adopt the combined strategy of motion vector information and similarity information can be for：If motion vector is believed The pre-provisioning request that breath meets then continues to gather near infrared eyes gray level image, and calculates the similarity of image A and image B as most Whole basis for estimation.Certainly, image handoff failure is represented if pre-provisioning request is unsatisfactory for, now can both select no longer to carry out closely Infrared hybrid optical system, i.e., directly terminate authentication procedures；The process can also be terminated, continue to obtain near-infrared gray-scale maps Picture, and the similarity of image A and image B is calculated, and fusional movement vector analyses result and similarity-rough set result are obtained most Whole judged result.

It is, identity coherence certification can be carried out using any one or more of situations below to user：

(1) directly judged according to motion vector information：

Determine whether that the motion vector of continuous multiple frames image sequence is unsatisfactory for pre-provisioning request, if so, then remind to be identified User identity concordance authentification failure.It should be appreciated that the collection of image sequence is before iris image acquiring, it is possible to Collection image sequence, analyzes judgement to which；If continuously multiple are all unsatisfactory for pre-provisioning request, image switching is directly assert Process is wrong, it is understood that there may be the switching of different personnel.Afterwards, no longer carry out the gray level image collection under near infrared light.

(2) directly judged according to similarity information：

Judge according to similarity information whether the user in image A and image B is same people；If it is not, then reminding to be identified User identity concordance authentification failure.Here, regardless of the result of motion vector analysis, can assert that identity coherence is recognized Card failure.

(3) fusional movement Vector Message and similarity information are judged：

In this case, if motion vector analysis result thinks that image handoff procedure is wrong, gray-scale maps are also still gathered Picture, simply finally calculates two analysis results according to its weights.Specifically, the motion vector is calculated respectively First analysis result r of information₁With the second analysis result r of the similarity information₂, and according to r=r₁*x+r₂* y is calculated The final result of identity full handshake authentication is carried out to identifying user is treated, wherein, x and y is the weights of respective items.Such as, if according to Motion vector analysis its be probably same people probability be 30%, according to the probability of similarity analysis be 40%, then according to items Self-defining weights, obtain it is final be whether same people probability.Certainly, this is an exemplary explanation, can use which His mode represents the calculating process of the analysis result and final result, the invention is not limited in this regard.

According to one embodiment, motion vector be unsatisfactory for pre-provisioning request generally refer to image sequence motion vector direction it is total Compare deflection from the point of view of body laterally and its change amplitude is more than certain threshold value, avoid user's imprudence from rocking caused by mobile phone with this Calculation error.Analyze motion vector information when, under mobile terminal records two image acquisition handoff procedure image sequence Row, and analyze motion vector direction and the motion vector magnitude of described image sequence；Often gather an image sequence all first to judge Whether the motion vector of the image meets pre-provisioning request, if continuous m frame image sequences are all unsatisfactory for pre-provisioning request, then it is assumed that deposit In the switching of different people, user identity full handshake authentication failure to be identified is reminded.

Further, whether the motion vector direction of analysis of the image sequence meets pre-provisioning request and can include：For acquisition The every frame image sequence for arriving, is all divided into multiple block images, the block image moved in determining the image sequence, And the motion vector of the block image of each motion, and block image number N, the motion vector side for moving is counted respectively To for laterally and vector magnitude more than first threshold block image number M, and the ratio of M and N.Afterwards, judge M and N's Whether ratio is more than Second Threshold；If so, then judge that the frame image sequence is unsatisfactory for pre-provisioning request；If it is not, being then considered normal Operation, can proceed follow-up image acquisition process.Wherein, Second Threshold can be 0.6.

It should be noted that when continuous m two field pictures are calculated, if all block images of a certain frame image sequence are not sent out Raw motion, then give up the calculating process that the image sequence is not counted in continuous m.For example, if existing continuous 2 image sequences are discontented with Sufficient pre-provisioning request, and the 3rd image is not moved, if then the 4th image is still unsatisfactory for pre-provisioning request, it is believed that having Continuous 3 image sequences are unsatisfactory for pre-provisioning request.

Further, it is determined that the method for the block image moved in image sequence and its motion vector can include： By taking P two field pictures as an example, the previous frame image of the P two field pictures, i.e. P-1 two field pictures are first determined；Choose from P two field pictures Block image at a certain position；For selected block image, the acquisition from P-1 two field pictures is in that of same position Individual same position block image, and obtain by this with position piecemeal centered on multiple partition neighborhood images, such as 3 × 3 partition neighborhood images. Wherein, with position block image and its partition neighborhood image, i.e., comprising this with position piecemeal totally 9 block image ImgSub_i (i =0,1 ..., 8), collectively as the contrast block image of selected block image.

Afterwards, calculate owning for the correspondence position in the selected block image and P-1 two field pictures in P two field pictures Contrast block image ImgSub_i (i=0,1 ..., dependency 8), and determine the position of the contrast block image of correlation maximum Put.Wherein, if the position of the contrast block image of correlation maximum is different from the position of selected block image, judging should Contrast block image is moved, and its motion vector direction points to selected block image, and its motion vector size is two Distance between position.Wherein, image correlation algorithm can adopt existing any particular algorithms, the invention is not limited in this regard.

According to another embodiment, when being calculated the similarity information of image A and image B, can first respectively to image A Iris and canthus positioning are carried out with image B, iris center therein and canthus position is obtained, the key point on human eye for obtaining point Not as shown in the black circle in Fig. 5 a and 5b, including the left and right summit at iris center, the left and right summit of left eye and right eye.Tool Body ground, can first carry out eye location；Then iris center is determined by the iris circle of iris segmentation, is determined by Corner Detection Eyes canthus.Wherein eye location can be examined using hough conversion using the methods such as adaboost graders, iris segmentation The method for surveying circle, can also adopt other conventional methods certainly, and the present invention is without limitation.

Subsequently, the sense comprising eyes is cut out respectively from image A and B according to the iris center and canthus position of eyes emerging Interesting region (Region Of Intrest, ROI).Area-of-interest is a rectangular area, and the size in the region is by eye spacing What nDist wire lengths d and its position determined.According to one embodiment, eye spacing nDist can by iris of both eyes center away from From expression.Cutting method is：The rectangular area for cutting out the iris center comprising eyes from image is emerging as the sense of the image Interesting region, two long sides of the rectangle and equal first distance d in iris center of eyes₁, minor face is immediate with the minor face Equal second distance d in iris center₂。

Fig. 6 a show the schematic diagram for cutting out area-of-interest from image.If the coordinate difference at left and right eye iris center For (x₀,y₀)(x₁,y₁), then iris centre distanceIris center abscissa difference DELTA x =| x₁-x₀|, the angle theta=arcos (Δ x/d) of iris of both eyes line and horizontal direction.

The rectangle P of area-of-interest is solved using information above₁P₂P₃P₄The coordinate on each summit, specifically,

If y₀≤y₁, then the transverse and longitudinal coordinate of each point be respectively：

x(P₁)=x₀-d₁×cosθ–d₂× sin θ, y (P₁)=y₀-d₁×sinθ+d₂×cosθ；

x(P₂)=x₀-d₁×cosθ+d₂× sin θ, y (P₂)=y₀-d₁×sinθ-d₂×cosθ；

x(P₃)=x₁+d₁×cosθ+d₂× sin θ, y (P₃)=y₁+d₁×sinθ-d₂×cosθ；

x(P₄)=x₁+d₁×cosθ–d₂× sin θ, y (P₄)=y₁+d₁×sinθ+d₂×cosθ。

If y₀>y₁, then the transverse and longitudinal coordinate of each point be respectively：

x(P₁)=x₀-d₁×cosθ+d₂× sin θ, y (P₁)=y₀+d₁×sinθ+d₂×cosθ；

x(P₂)=x₀-d₁×cosθ-d₂× sin θ, y (P₂)=y₀+d₁×sinθ-d₂×cosθ；

x(P₃)=x₁+d₁×cosθ-d₂× sin θ, y (P₃)=y₁-d₁×sinθ-d₂×cosθ；

x(P₄)=x₁+d₁×cosθ+d₂× sin θ, y (P₄)=y₁-d₁×sinθ+d₂×cosθ。

According to P₁、P₂、P₃、P₄The coordinate of point can determine that area-of-interest, as shown in the rectangle frame in Fig. 6 a.According to one Individual embodiment, can choose first apart from d₁=d/6, second distance d₂=d/2.Fig. 6 b are shown the region of interest in Fig. 6 a Domain rotates the schematic diagram to horizontal direction, and its long side is located at horizontal direction.The detailed process of its rotation includes：Obtain image Iris centre coordinate (the x of middle eyes₀,y₀), (x₁,y₁)；Iris centre distance d of calculating eyes, and iris of both eyes center Abscissa difference DELTA x；Calculate the angle theta of the iris of both eyes line of centres and horizontal direction；By area-of-interest rotation θ angles, make Its long side is located at horizontal direction.

Subsequently, respectively the area-of-interest in image A and B is rotated to horizontal direction, obtains image C and D, such as Fig. 7 a and Shown in 7b.Afterwards, multiple dimensioned higher-dimension partial binary mould respectively centered on canthus and iris center in calculating image C and D Formula LBP (Local Binary Patterns) feature, and the LBP features are cascaded into the characteristic vector as image.Finally, divide Not Ji Suan image C and D characteristic vector similarity, and judge user to be identified in image A and B is whether according to the similarity For same people.

According to one embodiment, whether identity is consistent with the user to be identified in B to judge image A, can calculate image C and The similarity of the characteristic vector of D, if similarity be more than the 3rd threshold value, process decision chart as A and B in user identity to be identified Cause, otherwise identity is inconsistent.Wherein, the similarity of characteristic vector can be chosen the degree of association of two images, card side's coefficient, intersect The common similarity parameter such as coefficient or Pasteur's distance, the invention is not limited in this regard, the 3rd threshold value can take 0.2.

Further, the step of multiple dimensioned higher-dimension LBP features for calculating image, can include：After image C or D fuzzy Down sample, obtains the image of different resolution, at the same the new image for obtaining every time it is wide be the 1/2 of original image with height, i.e., Respective image pyramid is built based on image C and D.Wherein, the most common fuzzy post-sampling for being namely based on Gauss, obtains To a series of images be referred to as gaussian pyramid.Afterwards, to per layer of image pyramid centered on image key points, such as in iris The heart and canthus, obtain area-of-interest, and the area-of-interest are divided into the rectangular image block of m × m.Count in each rectangular block LBP histogram features, obtain the LBP histogram features of rectangular block level.The LBP histogram features of all rectangular block levels are connected Get up, obtain the LBP histogram features of key point level.By the LBP histogram features of each key point in every tomographic image by suitable Sequence is coupled together, and obtains the LBP histogram features of level.Finally, the LBP histogram features of all levels are pressed into image level suitable Sequence is coupled together and is progressively extracted the LBP histogram features → key of each rectangular block as final characteristic vector, the i.e. present invention Point level LBP histogram features → stratal diagram picture LBP histogram features, and using multiple dimensioned higher-dimension LBP features cascade as Final characteristic vector.Fig. 8 shows the schematic diagram of the LBP features for building image pyramid and extracting iris center, wherein, can To take m=4.

Fig. 9 shows the identity coherence authentication device based on multi-mode bio-identification according to an embodiment of the invention 900 block diagram, which is mutually coupled with compound imaging device 200, and the device includes：Man face image acquiring unit 910, iris image is adopted Collection unit 920, motion vector analysis unit 930, similarity analysis unit 940 and identity coherence authentication ' unit 950.

Man face image acquiring unit 910 gathers the image A of user to be identified, wherein image A under visual light imaging pattern Generally include the facial information of the user.

Iris image acquiring unit 920 gathers the image B comprising user's eyes to be identified under near infrared light imaging pattern, Wherein image A generally includes the binocular information of the user.

Motion vector analysis unit 930 is suitable to obtain the image sequence of two kinds of imaging pattern handoff procedures, and analysis is obtained The motion vector information of described image sequence.Specifically, the image sequence for getting is divided by motion vector analysis unit 930 per frame It is segmented into multiple block images；The block image moved in determining image sequence and the motion vector of block image；And point In other sequence of computed images motion vector direction be horizontal and vector magnitude more than first threshold block image number M and generation Block image number N of motion, and the ratio of M and N；If the ratio is more than Second Threshold, the frame image sequence is judged Motion vector direction is unsatisfactory for pre-provisioning request.

Further, motion vector analysis unit 930 chooses the block image at a certain position from described image sequence； For selected block image, the same position block image in same position, and the same position is obtained from previous frame image Multiple partition neighborhood images of block image, the same position block image is with the partition neighborhood image collectively as selected The contrast block image of block image；The all contrast block images in block image and previous frame image selected by calculating Dependency, and determine the position of the contrast block image of correlation maximum；And if the contrast block diagram of the correlation maximum The position of picture is different from selected block image position, then judge that the contrast block image is moved, its motion vector side To the selected block image of sensing.

Similarity analysis unit 940 is suitable to analyze the similarity information for obtaining image A and image B.Specifically, respectively to figure As A and B carries out iris and canthus positioning, iris center therein and canthus position is obtained；According to iris center and the eye of eyes Angle Position cuts out the area-of-interest comprising eyes respectively from image A and B.Afterwards, respectively will be the sense in image A and B emerging Interesting region is rotated to horizontal direction, obtains image C and D.Wherein, the angle of rotation is the iris of both eyes line of centres and horizontal direction Angle.Afterwards, similarity analysis unit 940 calculates the multiple dimensioned higher-dimension in image C and D centered on canthus and iris center The LBP features are cascaded the characteristic vector as image, and calculate the similarity of the characteristic vector of image C and D by LBP features.Its Middle similarity mainly by calculating degree of association, card side's coefficient, intersecting coefficient or Pasteur's distance of two images etc., can set phase Be the 3rd threshold value like the marginal value spent, meet phase when the user in two images being can be generally thought more than the 3rd threshold value (such as 0.2) Require like degree.

Identity coherence authentication ' unit 950 is suitable to carry out body according to motion vector information and/or similarity information to user Part full handshake authentication.Which can take various strategies to carry out authentication, and such as 1) motion vector direction is directly sentenced after being unsatisfactory for requiring Determine identity coherence authentification failure and no longer gather gray level image；2) direction of motion continues collection gray level image after meeting requirement, Judged according to image similarity, similarity then certification success up to standard, otherwise then failed；3) after motion vector direction is unsatisfactory for requiring Continue collection gray level image, and the similarity result according to image and movement direction decision result, carry out with reference to its weight proportioning Fusion judges, specifically can be found in the introduction in step S330, will not be described here.

According to one embodiment, device 900 can also include：Face identification unit, is suitable to according to storing in data base The true head portrait of the user is verified to image A；Position guidance unit, is suitable to after recognition of face is verified, near red The angle and/or distance of user's adjustment face to be identified and compound imaging device is guided under outer photoimaging pattern, until being wrapped Image B containing its eyes；And iris identification unit, it is suitable to the true iris image pair according to the user stored in data base Image B is verified.

Identity coherence authentication device 900 based on multi-mode bio-identification of the invention, its detail exist Detailed disclosure in the description of Fig. 1-Fig. 8, will not be described here.

Technology according to the present invention scheme, using compound imaging device respectively visual light imaging pattern and near infrared light into As gathering its facial image and binocular images under pattern, and recognition of face and iris identification is carried out respectively, take full advantage of dynamic Two kinds of information of image sequence and still image.Specifically, advanced Mobile state image sequence motion vector direction analysis, excludes different The probability of people's switching, then carries out still image signature verification, further confirms that visible ray is same near infrared imaging identity One property, has stronger security reliability.Certainly, after the dynamic image direction of motion has been analyzed, even if finding which is not up to standard, also may be used To continue collection iris image, and the result of determination with reference to the direction of motion after image similarity has been calculated carries out comprehensively sentencing to which It is fixed, so as to carry out double authentication, improve whole security protection rank.

In addition, the present invention is to the order of recognition of face and iris identification and is not construed as limiting, can be to the people that collects first Face image carries out human eye detection, automatically selects current optimal living things feature recognition pattern according to the image space of human eye, and according to Sequence completes the authentication under recognition of face and iris identification both of which.Specifically, if human eye is in visual light imaging region, First carrying out recognition of face carries out iris identification again, if detecting human eye near infrared imaging region, first carries out iris identification again Carry out recognition of face.Certainly, user can also independently select recognition mode according to the custom of itself and demand.Whole identification process It is more intelligent, convenient, while identification accuracy is ensured, also with good Consumer's Experience.

B10, the device as described in B9, the identity coherence authentication ' unit are suitable for use with one kind in situations below or many Kind treats identifying user carries out authentication：(1) judge whether continuously have the motion vector of multiple image sequence to be unsatisfactory for making a reservation for Require, if so, then remind authenticating user identification failure to be identified；(2) image A and image B are judged according to the similarity information In user identity it is whether consistent；If it is not, then reminding user identity concordance authentification failure to be identified；(3) it is calculated respectively First analysis result r of the motion vector information₁With the second analysis result r of the similarity information₂, and according to r=r₁*x +r₂* y is calculated and treats identifying user and carry out the final result of identity coherence certification, and wherein, x and y is the power of respective items Value.

B11, the device as described in B9 or B10, the motion vector analysis unit are suitable to according to following methods judge Whether the motion vector of image sequence meets pre-provisioning request：The image sequence for getting is divided into into multiple block diagrams per frame Picture；The block image moved in determining described image sequence and the motion vector of block image；And calculate described respectively In image sequence, motion vector direction is horizontal and vector magnitude is more than block image number M of first threshold and moves Block image number N, and the ratio of M and N；If the ratio is more than Second Threshold, the motion arrow of the frame image sequence is judged Amount is unsatisfactory for pre-provisioning request.

B12, the device as described in B11, the motion vector analysis unit are suitable to determine described image according to following methods The block image moved in sequence and its motion vector：The block diagram at a certain position is chosen from described image sequence Picture；For selected block image, the same position block image in same position is obtained from previous frame image, and this is same Multiple partition neighborhood images of position block image, the same position block image is with the partition neighborhood image collectively as selected Block image contrast block image；Calculate selected block image and all contrast block images in previous frame image Dependency, and determine correlation maximum contrast block image position；And if the contrast piecemeal of the correlation maximum The position of image is different from selected block image position, then judge that the contrast block image is moved, its motion vector Selected block image is pointed in direction, and its motion vector size is the distance between two positions.

B13, the device as described in B9, the similarity analysis unit are suitable to be calculated image A and B according to following methods Similarity information：Carry out iris and canthus positioning respectively to image A and B, obtain iris center therein and canthus position；Root The area-of-interest comprising eyes is cut out from image A and B respectively according to the iris center and canthus position of eyes；Respectively will figure As the area-of-interest in A and B is rotated to horizontal direction, obtain image C and D and calculated centered on canthus and iris center respectively Multiple dimensioned higher-dimension local binary pattern LBP features in image C and D, and using LBP features cascade as image feature to Amount；And the similarity of the characteristic vector of image C and D is calculated respectively.

B14, the device as described in B13, wherein described similarity analysis unit are suitable to calculate image according to following methods Multiple dimensioned higher-dimension LBP features：Respective image pyramid is built based on image C and D respectively；It is pyramidal in described image Per the interested area division centered on the iris center and canthus of eyes respectively in tomographic image；The region of interest that each is divided Rectangular block of the regional partition for m × m；The LBP histogram features in each rectangular block are counted, the LBP rectangular histograms of rectangular block level are obtained Feature；The LBP histogram features of all rectangular block levels are coupled together, the LBP histogram features of key point level are obtained；Will per layer The LBP histogram features of each key point in image are linked in sequence, and obtain the LBP histogram features of level；And The LBP histogram features of all levels are linked in sequence by image level, the characteristic vector of final image is obtained.

B15, the device as described in B13, wherein, the similarity analysis unit is suitable to calculate the characteristic vector of image C and D Similarity；The identity coherence authentication ' unit be suitable to similarity be more than three threshold values when, process decision chart as A and B in treat Identifying user identity is consistent, and otherwise identity is inconsistent；The similarity of the wherein characteristic vector of described image C and D includes two figures At least one of the degree of association of picture, card side's coefficient, intersecting coefficient and Pasteur's distance.

B16, the device as described in B9, described image A include the facial information of user to be identified, and described image B includes treating The binocular information of identifying user, described device also include：Face identification unit, is suitable to according to the user's stored in data base True head portrait is verified to described image A；Position guidance unit, is suitable to after recognition of face is verified, near infrared light The angle and/or distance of user's adjustment face to be identified and compound imaging device is guided under imaging pattern, until obtaining comprising which The image B of eyes；And iris identification unit, it is suitable to according to the true iris image of the user stored in data base to described Image B is verified.

It should be appreciated that in order to simplify the disclosure and help understand one or more in each inventive aspect, it is right above The present invention exemplary embodiment description in, the present invention each feature be grouped together into sometimes single embodiment, figure or In person's descriptions thereof.However, should the method for the disclosure be construed to reflect following intention：I.e. required for protection is sent out The bright feature more features required than being expressly recited in each claim.More precisely, as the following claims As book is reflected, inventive aspect is less than all features of single embodiment disclosed above.Therefore, it then follows concrete real Thus the claims for applying mode are expressly incorporated in the specific embodiment, and wherein each claim itself is used as this Bright separate embodiments.

Those skilled in the art should be understood the module of the equipment in example disclosed herein or unit or group Part can be arranged in equipment as depicted in this embodiment, or alternatively can be positioned at and the equipment in the example In one or more different equipment.Module in aforementioned exemplary can be combined as a module or be segmented in addition multiple Submodule.

Those skilled in the art are appreciated that can be carried out adaptively to the module in the equipment in embodiment Change and they are arranged in one or more different from embodiment equipment.Can be the module in embodiment or list Unit or component are combined into a module or unit or component, and can be divided in addition multiple submodule or subelement or Sub-component.In addition at least some in such feature and/or process or unit is excluded each other, can adopt any Combination is disclosed to all features disclosed in this specification (including adjoint claim, summary and accompanying drawing) and so to appoint Where all processes or unit of method or equipment are combined.Unless expressly stated otherwise, this specification (includes adjoint power Profit is required, summary and accompanying drawing) disclosed in each feature can it is identical by offers, be equal to or the alternative features of similar purpose carry out generation Replace.

Although additionally, it will be appreciated by those of skill in the art that some embodiments described herein include other embodiments In some included 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 the following claims, embodiment required for protection appoint One of meaning can in any combination mode using.

Additionally, some heres in the embodiment be described as can be by the processor of computer system or by performing The combination of method or method element that other devices of the function are implemented.Therefore, with for implementing methods described or method The processor of the necessary instruction of element forms the device for implementing the method or method element.Additionally, device embodiment Element described in this is the example of following device：The device is used for implementing by order to performed by implementing the element of the purpose of the invention Function.

As used in this, unless specifically stated so, come using ordinal number " first ", " second ", " the 3rd " etc. Description plain objects are merely representative of the different instances for being related to similar object, and are not intended to imply that the object being so described must There must be the given order that the time is upper, spatially, in terms of sequence or in any other manner.

Although the present invention is described according to the embodiment of limited quantity, benefit from above description, the art It is interior it is clear for the skilled person that in the scope of the present invention for thus describing, it can be envisaged that other embodiments.Additionally, it should be noted that Language used in this specification primarily to the purpose of readable and teaching and select, rather than in order to explain or limit Determine subject of the present invention and select.Therefore, in the case of without departing from the scope of the appended claims and spirit, for this For the those of ordinary skill of technical field, many modifications and changes will be apparent from.For the scope of the present invention, to this The done disclosure of invention is illustrative and be not restrictive, and it is intended that the scope of the present invention be defined by the claims appended hereto.

Claims (10)

1. a kind of identity coherence authentication method based on multi-mode bio-identification, is suitable in the terminal containing compound imaging device Middle execution, the compound imaging device are suitable to incident light source is divided into visible light wave using the optical filter box of subregion subrane Section and near infrared light wave band, and respectively obtain the image under image A and near infrared light imaging pattern under visual light imaging pattern B, the method include：

Image A of the user to be identified under visual light imaging pattern, the image B under near infrared light imaging pattern is obtained respectively, And the image sequence of two imaging pattern handoff procedures；

Analysis obtains the similarity information of the motion vector information and/or described image A and image B of described image sequence；And

The identity under various bio-identification patterns is carried out according to the motion vector information and/or the similarity information to user Concordance certification.

2. the method for claim 1, it is described according to the motion vector information and/or the similarity information to user The step of carrying out identity coherence certification can be using any one or more of situations below：

(1) determine whether that the motion vector of continuous multiple frames image sequence is unsatisfactory for pre-provisioning request, if so, then remind use to be identified Family identity coherence authentification failure；

(2) judge according to the similarity information whether the user in image A and image B is same people；If it is not, then remind to wait to know Other user identity concordance authentification failure；

(3) the first analysis result r of the motion vector information is calculated respectively₁Analyze with the second of the similarity information As a result r₂, and according to r=r₁*x+r₂* y is calculated and treats identifying user and carry out the final result of identity coherence certification, its In, x and y is the weights of respective items.

3. method as claimed in claim 1 or 2, wherein, judges whether the motion vector of described image sequence meets predetermined wanting The operation asked includes：

The image sequence for getting is divided into into multiple block images per frame；

The block image moved in determining described image sequence and the motion vector of block image；And

Respectively calculate described image sequence in motion vector direction be horizontal and vector magnitude more than first threshold block image Number M and block image number N for moving, and the ratio of M and N；

If the ratio is more than Second Threshold, judge that the motion vector of the frame image sequence is unsatisfactory for pre-provisioning request.

4. method as claimed in claim 3, wherein it is described determine image sequence in the block image that moves and block diagram The step of motion vector of picture, includes：

The block image at a certain position is chosen from described image sequence；

For selected block image, the same position block image in same position is obtained from previous frame image, and should With multiple partition neighborhood images of position block image, the same position block image is with the partition neighborhood image collectively as selected The contrast block image of the block image for taking；

The dependency of the multiple contrast block images in selected block image and previous frame image is calculated, and determines dependency The position of maximum contrast block image；And

If the position of the contrast block image of the correlation maximum is different from selected block image position, this pair is judged Than piecemeal image motion, its motion vector direction points to selected block image, and its motion vector size is two positions Distance between putting.

5. the method for claim 1, it is described be calculated described image A and image B similarity information the step of wrap Include：

Carry out iris and canthus positioning respectively to image A and B, obtain iris center therein and canthus position；

Area-of-interest comprising eyes is cut out respectively from image A and B according to the iris center and canthus position of eyes；

Respectively the area-of-interest in image A and B is rotated to horizontal direction, image C and D is obtained；

The multiple dimensioned higher-dimension local binary pattern LBP for being calculated in image C and D centered on canthus and iris center respectively is special Levy, and the LBP features are cascaded into the characteristic vector as image；And

The similarity of the characteristic vector of image C and D is calculated respectively.

6. method as claimed in claim 5, wherein it is described calculate image multiple dimensioned higher-dimension LBP features the step of include：

Respective image pyramid is built based on image C and D respectively；

Region of interest is divided centered on the iris center and canthus of eyes respectively in described image is pyramidal per tomographic image Domain；

Each area-of-interest for dividing is divided into into the rectangular block of m × m；

The LBP histogram features in each rectangular block are counted, the LBP histogram features of rectangular block level are obtained；

The LBP histogram features of all rectangular block levels are coupled together, the LBP histogram features of key point level are obtained；

The LBP histogram features of each key point in every tomographic image are linked in sequence, the LBP rectangular histograms of level are obtained Feature；And

The LBP histogram features of all levels are linked in sequence by image level, the characteristic vector of final image is obtained.

7. method as claimed in claim 4, it is described the step of carry out identity coherence certification according to similarity information to user Including：

Calculate image C and D characteristic vector similarity, if similarity be more than the 3rd threshold value, process decision chart as A and B in treat Identifying user identity is consistent, and otherwise identity is inconsistent；

The similarity of wherein described characteristic vector includes the degree of association of image C and D, card side's coefficient, intersecting coefficient and Pasteur's distance In at least one.

8. the method for claim 1, described image A include the facial information of user to be identified, and described image B includes treating The binocular information of identifying user, methods described also include：

Described image A is verified according to the true head portrait of the user stored in data base；

If upchecking, the angle of to be identified user adjustment face and compound imaging device is guided under near infrared light imaging pattern Degree and/or distance, until the image B under obtaining the near infrared light imaging pattern of the user；And

Described image B is verified according to the true iris image of the user stored in data base.

9. a kind of identity coherence authentication device based on multi-mode bio-identification, is suitable in the terminal containing compound imaging device Middle execution, the compound imaging device are suitable to incident light source is divided into visible light wave using the optical filter box of subregion subrane Section and near infrared light wave band, and the image under image and near infrared light imaging pattern under visual light imaging pattern is respectively obtained, The device includes：

Man face image acquiring unit, is suitable to gather the face-image A of user to be identified under visual light imaging pattern；

Iris image acquiring unit, is suitable to image B of the collection comprising user's eyes to be identified under near infrared light imaging pattern；

Motion vector analysis unit, is suitable to obtain the image sequence of two kinds of imaging pattern handoff procedures, and analysis obtains the figure As the motion vector information of sequence；

Similarity analysis unit, is suitable to analyze the similarity information for obtaining described image A and image B；And

Identity coherence authentication ' unit, is suitable to carry out user according to the motion vector information and/or the similarity information Identity coherence certification.

10. a kind of mobile terminal, fills including identity coherence certification as claimed in claim 9 based on multi-mode bio-identification Put；And

Compound imaging device, is mutually coupled with the identity coherence authentication device, including：

Lens assembly, including the optical lenses of fixed focal length；

Optical filter box, including the visible band pass filter and permission near infrared light wave band that allow the light of visible light wave range to pass through The near infrared light bandpass region for passing through；And

Imageing sensor, including the transition region between visual light imaging region, near infrared light imaging region and the two regions Domain,

Wherein, the visual light imaging region under the visual light imaging pattern to by the visible band pass filter Visible ray is imaged, and the near infrared light imaging region is under the near infrared light imaging pattern to by the near infrared light The near infrared light of bandpass filter is imaged.