CN105989464A - Composition method of cash saving and withdrawing system of mobile phone deposit card - Google Patents

Abstract

The invention belongs to a composition method of a cash saving and withdrawing system of a mobile phone deposit card in the image processing field. The composition method comprises the following steps: carrying out the communication of a mobile phone deposit card number, and carrying out the cash saving and withdrawing of the mobile phone deposit card. The composition method is characterized in that a 3D screen code displayed on a screen is dynamic and can not be copied by lawbreakers. A mobile phone and an ATM (Automatic Teller Machine) are communicated through light media so as to be unlikely to be received by the lawbreakers who hide nearby, and safety is high. In addition, the composition method has the advantages of being simple in operation and easy in mastering. No much equipment needs to be added, and the composition method has unique advantages for the popularization of the cash saving and withdrawing system of the mobile phone deposit card.

Description

A kind of constructive method of the cash-access system of mobile phone deposit card

[technical field]

The invention belongs to the constructive method of the cash-access system of a kind of mobile phone deposit card in image processing field.

[background technology]

Recently, the Third-party payment of shopping at network has knocked financial field barrier for many years open, and the formation of a kind of new network-centric financial sector is the most imperative.Here, the core of society's various aspects keen competition is exactly Mobile-Payment Technology.In China, the technology being applied to mobile-phone payment at first is exactly the Versatile two-dimension bar code being disclosed.(non-patent literature 1)

There is the delivering of iPhone6 of NFC (Near Field Communication near field communication (NFC)) function, again the attention of people has been focused on the mobile phone electronic payment system of RFID technique.In some such as buy the convenience store of ticket, magazine or snacks widely used.The most also there is delivering of substantial amounts of patent documentation.Representational patent application has " payment system and payment terminal " patent (patent documentation 1) that Japan electric coating company applies for.

[patent documentation]

[patent documentation 1] (JP 2014-78074 bulletin)

[non-patent literature 1] (wechat Quick Response Code payment function using method)

(http://news.mydrivers.com/1/198/198121.htm)

The access time: on January 24th, 2015

The method described in above-mentioned non-patent literature 1, once dangerous and halt with Quick Response Code by national authority financial institution, the solution carrying out mobile-phone payment with Quick Response Code lies on the table.

Hand call payment system described in above-mentioned patent documentation 1 carries out code exchange by wireless telecommunications, owing to safety problem is not the most fully solved, mobile phone legitimate holder particularly can be made when mobile phone is lost bigger loss occur, and the most this hand set paying method is only suitable for small amount payment.

[summary of the invention]

It is an object of the invention to: overcome the deficiency of conventional art, a position by each organ of the face of random distribution and size is provided to carry out the specific of location parameter under probability scale metric space, and by artificial intervention, the fuzzy message of organ position and size is carried out fixed pattern, one position both having considered each organ of face and the probabilistic information of distribution of sizes proposed, simultaneously it is further contemplated that the constructive method of the more stable mobile phone deposit card cash-access brush face payment system of fuzzy message.

In order to solve above-mentioned problem, following technical scheme proposed:

The cash of a kind of mobile phone deposit card is stored in the constructive method of system, is by the communication authentication step of mobile phone savings card number, and the cash of mobile phone deposit card is stored in step composition, and its feature is as follows:

The communication authentication step of mobile phone savings card number, by the bluetooth of mobile phone, WiFi or sound and ATM communication, or the electronic image of the 3D half-tone screen code by distinguishing at the screen of mobile phone deposit card display optics, read this image by ATM and realize certification, complete to be sent in ATM the card number of deposit card；

The cash of mobile phone deposit card is stored in step, and cash is stored in the account of mobile phone deposit card after receiving certified savings card number and the cash that is stored in by ATM automatically.

The constructive method of the cash withdrawal system of a kind of mobile phone deposit card, is the number step being extracted cash by input, mobile phone deposit card authenticating step, enchashment gold step composition, and its feature is as follows:

The number step of input enchashment gold, selects mobile phone deposit card kind, if only one can be skipped, and the number of input extraction cash on ATM, or extract amount in cash in mobile phone terminal input, realize the number of input extraction cash by mobile phone with communicating of ATM；

Mobile phone deposit card authenticating step, the face characteristic information certification of " the probability scale metric space " of the mobile phone brush face blended by local feature and the global feature of the face of holder, or there is the voiceprint of adaptive learning；

Enchashment gold step, ATM receives input savings card number, and the amount of money of the enchashment gold of input, the most just can enter the operation of enchashment gold.

And, described 3D half-tone screen code image is the screen by ATM or mobile phone plane shows has and includes Two-dimensional electron image, with a kind of electronic image that can be displayed on screen in the three-dimensional electronic image of many-valued gray value, and the three-dimensional electronic image with the many-valued gray value of flicker.

nullAnd，The face characteristic information of described probability scale metric space，Refer to the characteristic information of face random distribution，By including normal distribution (Normal distribution)，Exponential (Exponential distribution)，Erlangian distribution (Erlang Distribution)，Weibull distribution (Weibull distribution)，Angular distribution (triangular distribution)，In beta distribution (Beta Distribution), the parameter of the probability attribute of at least one tool row probability distribution is as self-organizing probability scale，The characteristic information with relative certainty obtained eventually through the algorithm of the self-organizing of probability scale.

And, the face face image of the holder that the face recognition of holder is to rely on mobile phone photograph is the coloured image without monochrome information, and realizing being transformed into face image on the basis of not losing face image color information can be with the code of specific holder.

And, the face image of holder is under a certain color, and the density rule of the pixel distribution under a certain gray value, imports probability scale self-organized algorithm, is extracted out by human face five-sense-organ positional information automatically.

And, above-mentioned face characteristic information is obtained the eigenvalue closest to parent by probability self-organized algorithm, import again the theory of fuzzy mathematics simultaneously, for the information of the above-mentioned face obtained by the way of artificial intervention, using based on membership functions a plurality of defined in artificial experience (Membership Function), carrying out between 0 to n numerical value quantizes directly generates the characteristic vector with image code character.

And, described brush face certification or the generation of voiceprint code are the actions of the nictation by including eyeball, open one's mouth the action shut up, muscle minor variations when face smiles, the change of pupil, the small acceleration rocked of face, the color of face is in the identification of the life entity image of the brush face image of at least one reflection life entity characteristics of image interior.

And, the generation of described voiceprint code is can to propose some enquirements relevant with the problem that holder logs in advance by random by voiceprint, allows voiceprint person answer, thus whether person is life entity to identify voiceprint；Whether the life entity identification of voiceprint is in the methods such as identical state to realize the identification of life entity vocal print also by the password of plural number time.

And, described face characteristic information is the local feature information by brush face image, collectively forms with global feature information.

And, described voiceprint is the local feature information by vocal print signal, collectively forms with global feature information.

[explanation of nouns]

[probability scale metric space (Probability Scale Distance)]

If a given ordered series of numbers g with probability distribution₁, g₂... g_ζCollection be combined into G ∈ g_f(f=1,2 ..., ζ), the central value of this set is A (G), central value be the probability scale of A (G) be M [G, A (G)], and by self organization iteration calculated with the central value A (G of (n-1)th time^(n-1)), and the radius M [G on the basis of this central value^(n-1), A (G^(n-1))] in there is the ordered series of numbers g of k probability distribution₁, g₂... g_kCollection be combined into G⁽ⁿ⁾∈g_f(f=1,2 ..., k), then

[formula 1]

A⁽ⁿ⁾=A (G⁽ⁿ⁾)

M⁽ⁿ⁾=M [G⁽ⁿ⁾, A (G⁽ⁿ⁾)]

G⁽ⁿ⁾=G{ [A (G^(n-1)), M [G^(n-1), A (G^(n-1))]]

Here, probability scale M⁽ⁿ⁾It it is the parameter of probability statistics with multiple attributes.For example normal distribution, exponential, Erlangian distribution, Weibull distribution, angular distribution, beta distribution etc..Such as probability scale M⁽ⁿ⁾Can serve as the dispersion value of normal distribution.

By above-mentioned iterative formula 1 through the central value that iteration several times is calculated be the ordered series of numbers g for probability distribution₁, g₂... g_ζThe obtained estimated value closest to parent, and final horizon radius value is a probability scale, on the basis of final central value, the ordered series of numbers g of all of probability distribution in the range of probability scale₁', g₂' ... g_k' can each belong to probability distribution ordered series of numbers g₁, g₂... g_ζTrue value.

Use above-mentioned formula Isosorbide-5-Nitrae, the computational methods of 5 and 6, can be for the characteristic vector of two probability distribution between, obtain a distance value closest to parent.Under complicated mobile phone shooting environmental, facial image is transformed into one the ID of certification holder's identity can obtain a result the most stable for what the present invention proposed.

[accompanying drawing explanation]

Fig. 1 is wholesale commodity purchasing mobile-phone payment operating process schematic diagram

Fig. 2 is small amount commodity purchasing mobile-phone payment operating process schematic diagram

Fig. 3 is mobile phone brush face Zhi Fuyu vocal print payment schematic diagram of doing shopping under line

Fig. 4 is the schematic diagram that mobile phone credit card deposit card " light " pays

Fig. 5 is the schematic diagram that the cash of mobile phone deposit card is stored in

Fig. 6 is the self-organizing handling process of probability scale distance

Fig. 7 is the establishing method schematic diagram of face recognition face position feature point

Fig. 8 is the definition method schematic diagram of the eigenvalue of face recognition face dimension information

Fig. 9 is the schematic diagram being extracted eigenvalue by the eye position of people and dimension information

Figure 10 is the schematic diagram being extracted eigenvalue by the shape information of face

Figure 11 is the schematic diagram being extracted eigenvalue by the Skin Color Information of face

Figure 12 is the schematic diagram of the information retrieval eigenvalue by face frequency space

Figure 13 is as the processing method schematic diagram of the Global Information of face using the information of the frequency space of face

Figure 14 is the processing method schematic diagram quantizing voiceprint

Figure 15 is the example that the shape information of face is defined as membership function

Figure 16 is brush face or the flow chart of vocal print code-adaptive learning processing method

Figure 17 is the schematic diagram of one of the example of life entity image recognition

Figure 18 is the schematic diagram of the electronic image of the 3D half-tone screen code that optical identification is possible

[detailed description of the invention]

Below in conjunction with accompanying drawing, the embodiment of the present invention is further described, but embodiment of the present invention is illustrative rather than determinate.

The present invention solves it and technical problem is that and take techniques below scheme to realize: illustrate for inventive embodiment according to Figure of description.

Fig. 1 is wholesale commodity purchasing mobile-phone payment operating process schematic diagram.

As shown in Figure 1: first select in commodity step (a) to be bought online, can directly surf the Net, in such as Taobao, the shopping website such as Jingdone district utilize the existing system of these websites, select required article, operate according to the requirement of the website used.

Certain shopping at network platform has the consensus standard of oneself, how to seek unity of standard, and allowing mobile-phone payment generalization is a problem, and this problem does not solve to have a strong impact on mobile-phone payment popularizing without card.

According to current present situation, dock can be docked by the plug-in unit that wechat or Alipay are given with wechat or Alipay, but, so dock with that website, the plug-in unit of that website will be installed, bother very much, make troubles to user, the present invention propose can provide a standard to connector, each website just can be docked with all of website after installing this plug-in unit.

Mobile phone selects in credit card step (b), directly carry out using which credit card, or the selection of bank deposit card at mobile phone terminal.Here, a mobile phone can support the payment of the different credit cards, and settling bank can also be a plurality of.If this picture can be given tacit consent to and need not eject the picture being directly entered brush face certification during the credit card only one of which held.

After automatically into brush face authenticating step (c), mobile phone screen ejects a recognition window, as long as the face of holder is directed at recognition window, so that it may quickly have the mobile phone brush face certification that the local feature of face blends with global characteristics.For preventing illegal person from gaining certification by cheating by photo, the present invention proposes to use life entity to know method for distinguishing and carries out brush face certification.

The learning data that mobile phone brush face pays is placed in oneself mobile phone, need not be put on server, if from mobile phone terminal send a code just can, so can make the speed of process faster, personal information is protected by the mobile phone of oneself, it is also possible to raising safety.

It is sent to bank server (d) step in mobile-phone payment data, mobile-phone payment data are sent to bank server, complete certification and payment operation.

The feature of the theoretical brush face certification of importing probability scale distance:

Directly brush face image can be transformed into code, can be directly as the password of mobile-phone payment, in network transmission, speed is fast, and the code that can be transformed into by brush face image is directly as the password of mobile-phone payment, without making original system make more change on system building, reduce input cost.

Brush face pays or vocal print payment code is stable, also has the function of adaptive learning, systematic function can be made constantly to improve.

Brush face pays has life entity identification function, is possible to prevent illegal person to gain payment by cheating with photo.The concordance certification of the credit card and the holder paid can be solved, hand call payment system can be made to have the highest safety.

Fig. 2 is small amount commodity purchasing mobile-phone payment operating process schematic diagram.

As shown in Figure 2: when carrying out small amount shopping on line, utilize mobile phone vocal print to pay and need also exist for 3 steps,

First select in commodity step (a) to be bought online, identical with the wholesale shopping at network shown in Fig. 1, need on shopping at network platform, select required commodity.With each big website to connect the present invention propose can provide a standard to connector, each website just can be docked with all of website after installing this plug-in unit.

Mobile phone selects in credit card step (b), also identical with the wholesale shopping at network shown in Fig. 1, directly carry out using which credit card, or the selection of bank card at mobile phone terminal.If during the credit card only one of which held, this picture can be given tacit consent to and need not eject, and is directly entered the picture of voiceprint.

After automatically into the step (c) of voiceprint, mobile phone screen ejects a voiceprint identification prompting, this mobile phone microphone to be directed at of holder is said in short, such as " agree to pay " certification achieving that the vocal print mobile-phone payment with adaptive learning, easy to operate, certification is effective.Choose whether as voiceprint, or brush face certification can be selected by client oneself in default menu.

Voiceprint can propose some enquirements relevant with the content that holder logs in advance by random, allows voiceprint person answer, thus whether person is life entity to identify voiceprint.

Whether the life entity identification of voiceprint is in identical state also by the password that plural number is secondary realizes.

It is sent to bank server step (d) in mobile-phone payment data, mobile-phone payment data are sent to bank server, complete certification and payment operation.

The feature of the theoretical voiceprint of importing probability scale distance:

Directly voiceprint can be transformed into code, can be directly as the password of mobile-phone payment, in network transmission, speed is fast, and the code that can be transformed into by voiceprint is directly as the password of mobile-phone payment, without making original system make more change on system building, reduce input cost.

Directly voiceprint can be transformed into the payment system of the mobile phone credit card deposit card of code, be characterized in that code is stable, also there is the function of adaptive learning, the recognition performance of system can be made constantly to improve.

The payment system of the mobile phone credit card deposit card that voiceprint can directly be transformed into code also has life entity identification function, is possible to prevent illegal person to utilize the recording of card holder to gain payment by cheating.The concordance certification of the credit card and the holder paid can be solved, the payment system of mobile phone credit card deposit card can be made to have the highest safety.

The payment system of the mobile phone credit card deposit card that voiceprint can directly be transformed into code is simple to operate, there is not privacy problem, can be generalized to countries in the world.

Fig. 3 is mobile phone brush face Zhi Fuyu vocal print payment schematic diagram of doing shopping under line.

Under carrying out line during shopping, utilize mobile phone brush face to pay or vocal print pay collinear 3 steps of the identical needs of upper shopping, as shown in Figure 3:

In shopping clearing step: after checkout station gone to by the commodity that choice of supermarkets is to be bought, each commodity are inputted by cashier, display of commodity price on the screen of POS, after holder opens the APP of mobile-phone payment, by the communication of sound wave, bluetooth and WiFi, the screen of holder's mobile phone shows inventory and the price buying commodity, holder confirms that commodity price is errorless, can immediately enter the step selecting the credit card to be paid.

In the step selecting the credit card to be paid: if holder's only one of which credit card, the step for of could skipping.

In holder's authenticating step: the mobile phone brush face certification that holder is blended by the local feature of face and global feature in the face of the mobile phone of oneself, or there is the voiceprint of adaptive learning, after certification success, automatically the card number of the credit card held being mail to POS, POS receives the card number of the credit card just can complete the overall process of mobile-phone payment.Here, after certification, mobile phone is directly communicated with the POS of cashier by bluetooth or WiFi, the mobile phone credit card or the card number of deposit card and password is dealt in POS, POS is directly traded with bank after receiving information.Can also allow mobile phone directly with cashier communication, after mobile phone side is sent to cashier the amount of money paid, mobile phone directly be traded with bank.

If what holder selected is that the certification that mobile phone vocal print pays can report out oneself payment password at the mobile phone of oneself, such as, the overall process that the most just can complete mobile-phone payment " can be paid ".

Fig. 4 is the schematic diagram that mobile phone credit card deposit card " light " pays.

As shown in Figure 4: in clearing step；Commodity to be bought for client are entered in cashier by cashier by barcode scanning machine, and the screen of cashier shows the electronic image of the 3D half-tone screen code that optical identification is possible.

In mobile phone payment authentication step, client uses the 3D half-tone screen code image on the screen of the camera gun alignment cashier of the mobile phone credit card or deposit card to distinguish, just commodity settlement information can be received mobile phone terminal, it is achieved the certification that the light of mobile phone credit card deposit card pays；

In mobile phone credit card deposit card payment step, after client can be by the commodity oneself bought that confirm phone, by the handset-selected credit card or deposit card, if holder's only one of which credit card or deposit card, the step for of could skipping.

Mobile phone mails to the card number of the credit card held or deposit card to POS automatically, if need password, can produce password by the brush face of mobile phone or vocal print, and POS receives the card number of the credit card or deposit card just can complete the overall process of mobile-phone payment.Here, after certification, mobile phone is directly communicated with the POS of cashier by bluetooth or WiFi, the mobile phone credit card or the card number of deposit card and password is dealt in POS, POS is directly traded with bank after receiving information.Can also allow mobile phone directly with cashier communication, after mobile phone side is sent to cashier the amount of money paid, mobile phone directly be traded with bank.

If what holder selected is that the certification that mobile phone vocal print pays can report out oneself payment password at the mobile phone of oneself, such as, the overall process that the most just can complete mobile-phone payment " can be paid ".

3D half-tone screen code image is the screen by POS or mobile phone plane shows has and includes Two-dimensional electron image, with a kind of electronic image that can be displayed on screen in the three-dimensional electronic image of many-valued gray value, and the three-dimensional electronic image with the many-valued gray value of flicker.

Here, the information screen coding embedded with 3D that mobile phone screen shows, it is dynamic, and is the information of 3D, replicate thus without by illegal person, compared with common Quick Response Code payment, there is high safety.

Holder, when carrying out the mobile phone of 3D half-tone screen code " light " and paying, is to be carried out by the medium of light between mobile phone and cashier, because mutual information is difficult to be received by neighbouring illegal person under covering, has the feature of system and safety thereof.

Further, holder uses mobile phone " light " payment function of 3D half-tone screen code, can feel simple to operate, be easily mastered.

Mobile phone " light " payment system due to 3D half-tone screen code is in close proximity to the payment system of mobile phone two-dimension code, therefore there is for the shopping website that currently used two-dimensional bar code carries out mobile-phone payment easily transformation, a lot of equipment need not be added, there is vertical rod and see the effect of shadow.

Identify the 3D half-tone screen code universal tool advantageous advantage of row without adding any equipment, to the payment system of mobile phone credit card deposit card.

Fig. 5 is the schematic diagram that the cash of mobile phone deposit card is stored in.

With mobile phone deposit card deposit cash gold time, have a two ways, a kind of bluetooth by mobile phone or WiFi and ATM communication, mobile phone is saved card number and notifies to ATM, then select operation content by ATM key, so that it may realize the transaction of the gold of depositing cash of mobile phone memory card.

Also has a kind of mode, ATM installs a 3D half-tone screen code recognizer, after handset-selected cash transaction function, just can show a 3D half-tone screen code on the screen of mobile phone, mobile phone is placed on the 3D half-tone screen code recognizer of ATM and just can be stored in transaction what the card number of mobile phone deposit card sent that ATM carries out cash to.

When using mobile phone deposit card enchashment gold, initially enter the APP program of mobile phone deposit card, flow process with reference to above-mentioned mobile-phone payment, selecting the kind of deposit card, if deposit card only one of which, this step is skipped and is directly entered mobile phone brush face certification, or the step of voiceprint, the mobile phone brush face certification blended by the local feature of face and global feature, or there is the voiceprint of adaptive learning, just can enter the operation of enchashment gold after certification success.Here, the input enchashment amount of money can be carried out on ATM, it is possible to inputs the enchashment amount of money at mobile phone terminal, and the enchashment being realized mobile phone deposit card by the communication of mobile phone Yu ATM is concluded the business.

Fig. 6 is the self-organizing handling process of probability scale distance.

As shown in Figure 6: set given one and there is the ordered series of numbers g of probability distribution₁, g₂... g_fCollection be combined into G ∈ g_f(f=1,2 ..., 1), then it is made up of following 4 steps based on probability scale self-organized algorithm.

Step 1: pre-treatment step: M⁽⁰⁾As initialization probability yardstick, A⁽⁰⁾As the initial centered value of self-organizing, V is as the convergency value of self-organizing, and MN is as self-organizing maximum tissue time numerical value, and initial n=0 is as the current number of times of self-organizing.

About M⁽⁰⁾As initialization probability yardstick and A⁽⁰⁾Determining method as the initial centered value of self-organizing, it is not necessary to carry out tight setting.By artificial prediction, for final scope, at least part of numerical value is included in initialization probability yardstick M⁽⁰⁾In the range of.Initialization probability yardstick M⁽⁰⁾The biggest, the time of calculating is the longest, otherwise the least, it is possible to can not get correct result.

About V as the establishing method of convergency value, convergency value V is the biggest, it is possible to can not get correct result.Convergency value is the least, calculates the time spent the longest.Correct establishing method is about the 10% of the probability scale of final self-organizing.

About the establishing method of maximum self-organizing number of times MN, it is sufficient to for usually 5-10 time.

Step 2: self-organizing step: carry out n self-organizing and process, A⁽ⁿ⁾As self-organizing central value, probability scale M⁽ⁿ⁾As radius, with central value A⁽ⁿ⁾On the basis of, calculate all numerical value g within radius_f(f=1,2 ..., ζ) meansigma methods V⁽ⁿ⁺¹⁾With dispersion value S⁽ⁿ⁺¹⁾, V⁽ⁿ⁺¹⁾=A⁽ⁿ⁺¹⁾, S⁽ⁿ⁺¹⁾=M⁽ⁿ⁺¹⁾, n=n+1.

[formula 2]

$V^{(n+1)}=\frac{1}{\mathrm{\ζ}}\underset{j=1}{\overset{\mathrm{\ζ}}{\mathrm{\Σ}}}{g}_j$

[formula 3]

$S^{(n+1)}=\frac{1}{\mathrm{\ζ}-1}\underset{j=1}{\overset{\mathrm{\ζ}}{\mathrm{\Σ}}}{(g_j-V^{(n+1)})}^2$

Step 3: self-organizing discriminating step.Self-organizing processes and reaches maximum times (N >=MN) or self-organizing process convergence (M⁽ⁿ⁾-M⁽ⁿ⁺¹⁾≤ V), as YES, just no longer carry out the self-organizing of next time and process, self-organizing terminates to jump to step 4.If NO, just jump to step 2 and proceed self-organizing process.

Step 4: self-organizing process terminates.

Probability scale M⁽ⁿ⁾It it is the parameter of probability statistics with multiple attributes.For example normal distribution, exponential, Erlangian distribution, Weibull distribution, angular distribution, beta distribution etc..Such as probability scale M⁽ⁿ⁾Can serve as the dispersion value of normal distribution.

The present invention proposes to be transformed into facial image the authentication codes of mobile-phone payment, realized by two kinds of methods, a kind of is the local message for face, generate each eigenvalue, again each eigenvalue is constituted a vector quantized, at the vectorial constitutive characteristic vector each quantized by membership function, then the image construction characteristic vector space by different faces.

For the local message of face, the method generating each eigenvalue is as follows:

The present invention is using the positional information of face as the important information of specific holder, and therefore the positional information for face carries out following definition.

Fig. 7 is the establishing method schematic diagram of face recognition face position feature point.

As shown in Figure 7: be may be constructed the characteristic point (Landmarks) of more than 24 by the position of the face of face.Such as, the right and left eyes Angle Position of left eye constitutes l₁With l₂Two characteristic points, the right and left eyes Angle Position of right eye constitutes again l₃With l₄Two characteristic points, left eye constitutes l with the upright position at left eye eyebrow center₅With l₆Two characteristic points, right eye constitutes again l with the upright position at right eye eyebrow center₇With l₈Two characteristic points, nose both sides constitute l₉With l₁₀Two characteristic points, nose both sides constitute again l with the upright position of the connecting line of two₁₁With l₁₂Two characteristic points, the both sides of mouth constitute l₁₃With l₁₄Two characteristic points, the both sides of mouth and the centre of nose constitute again l₁₅With l₁₆Two characteristic points, the left eye angle of left eye and the right eye Angle Position of right eye constitute l₁₇With l₁₈Two characteristic points, the left eye angle of left eye is vertical with the forehead of face with the connecting line at the right eye angle of right eye constitutes again l₁₉With l₂₀Two characteristic points, intersect with the both sides of face after the characteristic point connecting line prolongation on the both sides of mouth and constitute l₂₁With l₂₂Two characteristic points, the vertical line intersecting constituted line and face volume after the characteristic point connecting line prolongation on the both sides of mouth with the both sides of face constitutes again l₂₃With l₂₄Two characteristic points.

Here, the both sides of mouth and the middle l constituted of nose₁₆The l that the upright position of characteristic point and nose both sides and the connecting line of two is constituted₁₁Characteristic point is to overlap, and the left eye angle of left eye and the right eye Angle Position of right eye constitute l₁₇With l₁₈Two characteristic points, be and l₁With l₁Two characteristic point coincidences, the l that the forehead of the left eye angle of left eye and the right eye angle of right eye and face is constituted₁₉With l₂₀The l that the upright position of characteristic point and nose both sides and the connecting line of two is constituted again₁₂Characteristic point, and intersect, with the both sides of face, the l that the vertical line of constituted line and face volume is constituted with after the characteristic point connecting line prolongation on the both sides of mouth₂₄Characteristic point is to overlap, and intersects the l that the vertical line of constituted line and face volume is constituted again after the characteristic point connecting line prolongation on the both sides of mouth with the both sides of face₂₃The l that characteristic point is constituted again with the centre of nose with the both sides of mouth₁₅Characteristic point overlaps.

It is to say, l₁With l₁₇It is the left eye Angle Position of a left eye, l₄With l₁₈It is the right eye Angle Position of a right eye, l₁₂With l₁₉It is the upright position of the connecting line of two, l₁₄With l₁₆It is the center on nose both sides, l₁₅With l₂₃It is the center of the characteristic point on the both sides of mouth, l₂₀With l₂₄Being the position of the characteristic point of face volume, the actual characteristic point describing face's face is 18.

As long as it practice, l₁With l₂Two characteristic points, l₃With l₄Two characteristic points, l₆With l₈Two characteristic points, l₉With l₁₀Two characteristic points, l₁₃With l₁₄Two characteristic points, l₂₁With l₂₂Two characteristic points and l₂₄Characteristic point, 13 characteristic points are extracted out and just may be used altogether.

Can be l₁With l₂The straight line that two characteristic points are connected, l₃With l₄The straight line that two characteristic points are connected, l₅With l₆The straight line that two characteristic points are connected, l₇With l₈The straight line that two characteristic points are connected, l₉With l₁₀The straight line that two characteristic points are connected, l₁₁With l₁₂The straight line that two characteristic points are connected, l₁₃With l₁₄The straight line that two characteristic points are connected, l₁₅With l₁₆The straight line that two characteristic points are connected, l₁₇With l₁₈The straight line that two characteristic points are connected, l₁₉With l₂₀The straight line that two characteristic points are connected, l₂₁With l₂₂The straight line that two characteristic points are connected, l₂₃With l₂₄The straight line that two characteristic points are connected.The present invention is using the length of altogether 12 straight lines as the characteristic information describing face's face position, and the characteristic information i.e. describing face's face positions using these 12 is as the key element of the characteristic vector of 12 certification holders.

As the method utilizing the local message of face to generate eigenvalue, the present invention also proposes the face dimension information of face as the important information of specific holder, and therefore the dimension information for face carries out following definition.

Fig. 8 is the definition method schematic diagram of the eigenvalue of face recognition face dimension information.

Fig. 8 (a) is the scale diagrams of human eye, and the size of human eye is to differentiate that the important information of holder's feature, the particularly size of human eye eyeball are the key characters differentiating holder.As shown in Fig. 8 (a): (8-1) is the eyeball of human eye, generally dark color, the optical fundus of (8-2) human eye, generally light color.V₁It is the width dimensions of eye, V₂It it is the diameter dimension of eyeball.The present invention except shown in above-mentioned by the width dimensions V of eye₁As 1 key element of the characteristic vector of certification holder, the most also by the diameter dimension V of eyeball₂Wherein 1 key element as the characteristic vector of certification holder.

Fig. 8 (b) is the scale diagrams of mouth, and the size of mouth and thickness are the important informations differentiating holder's feature.As shown in Fig. 8 (b): (8-3) is the lip of people, generally redness.V₄It is the width dimensions of mouth, V₃It it is the gauge of mouth.The present invention except shown in above-mentioned by the width dimensions V of mouth₄As 1 key element of the characteristic vector of certification holder, the most also by the gauge V of mouth₃Another key element as the characteristic vector of certification holder.

The face image of the holder that the face that the present invention is directed to holder is identified being to rely on mobile phone photograph is the coloured image without monochrome information, and being transformed into by face image of realizing on the basis of not losing face image color information can be with the code of specific holder.Because the color information of face image is high efficiency to face's face position, the extraction of face dimension information provides important feature.Such as, the color of eyeball, the deepest in face, optical fundus color is the most shallow in face, and lip is red etc..Utilizing the color information without monochrome information quickly and the most above-mentioned face information to be extracted out, simultaneously because do not affected bigger monochrome information by mobile phone shooting environmental, therefore recognition result is little on the impact of environment.

The present invention is directed to the identification of holder face also use face's face position density of the distribution of pixel grey scale under a certain color and carry out the information at face position and extract out.It is maximum that such as eyeball compares its density with other face under black color, and it is maximum that its density is compared with other face under white colour in optical fundus.It is maximum that mouth compares its density with other face under red color.

The present invention utilize above-mentioned human face five-sense-organ image under a certain color, and the density rule of the pixel distribution under a certain gray value, above-mentioned probability scale self-organized algorithm can be imported, automatically human face five-sense-organ positional information is extracted out.Because the result of calculation of probability scale self-organizing, it is exactly to obtain a certain color, and the maximized result of probit of the pixel distribution of a certain gray value, namely can obtain in a certain color, and the maximized result of pixel distribution density value under a certain gray value.

Such as the identification of eye eigenvalue of people, first should start more convenient from eyeball, because the feature of eyeball to be the density of black picture element very big, utilize the algorithm of probability self-organizing can extract the positional information of eye and the dimension information of people very easily near eyeball.

Fig. 9 is the schematic diagram being extracted eigenvalue by the eye position of people and dimension information.

Import the method eye position for people of probability self-organizing and the extraction of size characteristic, it is slightly different with the method shown in above-mentioned Fig. 6, the method of the probability self-organizing shown in Fig. 9 is carried out for one-dimensional data, is that distribution center and the Size calculation of the eyeball for two-dimensional space goes out here.

As shown in Fig. 9 (a): the algorithm provided with reference to Fig. 2 as the method for the probability self-organizing of two-dimensional space data, initially should provide initial center A near eyeball⁽⁰⁾=(x₀, y₀), and initial probability scale M⁽⁰⁾.Finding near the position of eyeball, initial center A⁽⁰⁾=(x₀, y₀), and initial probability scale M⁽⁰⁾Establishing method be that eyeball must be included in A⁽⁰⁾=(x₀, y₀Radius M centered by)⁽⁰⁾In the range of, here, it is not necessary to all parts of eyeball will be included completely, also may be used even if some is included, because the scope that probability self-organized algorithm can will be included automatically, under the calculating of self-organizing repeatedly, move to the place of the highest pixel distribution of density automatically.

In this step, as shown in figure 6 above: the convergency value V of self-organizing to be set, self-organizing maximum tissue time numerical value MN, the current number of times h=0 of self-organizing.Its method can refer to Fig. 2.

As shown in Fig. 9 (b): in the calculating of following probability self-organizing, A^(h)=(x₀, y₀)^(h)Can refer to formula 4 calculate.

[formula 4]

${X_0}^{\left(h\right)}={X_0}^{(h-1)}+\frac{\underset{j=1}{\overset{l}{\mathrm{\Σ}}}\underset{i=1}{\overset{k}{\mathrm{\Σ}}}(x_i-{X_0}^{(h-1)})I(x_i,y_j)}{\underset{j=1}{\overset{l}{\mathrm{\Σ}}}\underset{i=1}{\overset{k}{\mathrm{\Σ}}}I(x_i,y_j)}$

${Y_0}^{\left(h\right)}={Y_0}^{(h-1)}+\frac{\underset{j=1}{\overset{l}{\mathrm{\Σ}}}\underset{i=1}{\overset{k}{\mathrm{\Σ}}}(y_i-{Y_0}^{(h-1)})I(x_i,y_j)}{\underset{j=1}{\overset{l}{\mathrm{\Σ}}}\underset{i=1}{\overset{k}{\mathrm{\Σ}}}I(x_i,y_j)}$

Equally, probability scale M^(h)=S^(h)2Can refer to formula 5 calculate.

[formula 5]

$S^{\left(h\right)2}=\frac{\underset{j=1}{\overset{l}{\mathrm{\Σ}}}\underset{i=1}{\overset{k}{\mathrm{\Σ}}}[{(x_i-{X_0}^{\left(h\right)})}^2+{(y_j-{Y_0}^{\left(h\right)})}^2]I(x_i,y_j)}{\underset{j=1}{\overset{l}{\mathrm{\Σ}}}\underset{i=1}{\overset{k}{\mathrm{\Σ}}}I(x_i,y_j)}$

Formula 4 and formula 5 have been merely given as an example, are referred to the two formula and use other similar formula equally to obtain required result, do not enumerate.

Carry out h self-organizing to process, (x₀, y₀)^(h)As self-organizing central value, probability scale M^(h)As radius, the pixel I (x of all eyes within calculating radius_i, y_j) (i=1,2 ..., k, j=1,2 ..., 1) gray-scale intensity dispersion value S^(h+1)。M^(h+1)=S^(h+1), h=h+1.As shown in Fig. 9 (b): probability scale M^(h)=S^(h)2Radius value convergence, center A^(h)=(x₀, y₀)^(h)It is gradually moved to the center of eyeball.

Through the calculating of the probability scale self-organizing of n time, in the step of 9 (c), center A^(h)=(x₀, y₀)^(h)Stop in the center of eyeball, probability scale M^(h)Radius be also parked in around eyeball, the size of eyeball is calculated accurately with position.

The face positional information of the random distribution that brush face image is had, face dimension information, face shape information, face colouring information, face frequency information etc., self-organizing through similar above-mentioned probability scale processes, so that it may obtains the stable brush face information in probability scale metric space, does not enumerates.

Figure 10 is the schematic diagram being extracted eigenvalue by the shape information of face.

The shape information of face is the key character of difference different people, and the difference change depending on shooting environmental is smaller, and will not vary widely at short notice, is not affected by after hair style or cosmetic, is therefore a metastable information.

As shown in Figure 10: the shape information of the face that (a) expression shape of face is thinner, (b) represents that shape of face compares The shape information of common face, the shape information of the face that (c) expression shape of face is more fat.Everyone shape of face has different difference, utilize these different informations can as distinguish different people eigenvalue.

Here, the present invention reintroduces the abstracting method of another kind of face local message, if the kernel function of Two-Dimensional Gabor Wavelets (GWT) is a plane wave by Gaussian envelope function constraint:

[formula 6]

${\mathrm{\ψ}}_{u,v}\left(z\right)=\frac{{\left|\right|k_{u,v}\left|\right|}^2}{{\mathrm{\σ}}^2}{e}^{(-{\left|\right|k_{u,v}\left|\right|}^2{\left|\right|z\left|\right|}^2/{2\mathrm{\σ}}^2)}[e^{i{\stackrel{\→}{k}}_{u,v}z}-e^{-{\mathrm{\σ}}^2/2}]---\left(3\right)$

Wherein,k_v=k_max/f^vRepresent the frequency (yardstick) of kernel function, Represent the direction of kernel function.By arranging different yardsticks and direction, one group of Gabor wavelet kernel function can be obtained.With image, the feature extraction of image is carried out convolution operation respectively by multiple Gabor wavelet kernel functions complete.

In order to preferably extract local message, according to locus, Gabor characteristic being carried out piecemeal, the feature in each block is concatenated into a characteristic vector.So, for a width facial image, we can be obtained by multiple characteristic vector, is referred to as local Gabor characteristic vector (LGFV).

Above-mentioned Fig. 8 to Figure 10 is the local message of extraction face, and obtained result is the eigenvalue at each position, and the eigenvalue at each position constitutes the vector that quantizes of one group of feature that can reflect each position of face,

The Global Information entering face refers to the most one-dimensional information all containing all parts on facial image (the most all pixels) of its characteristic vector, and therefore reflect is the integrity attribute of face.Here the Skin Color Information of face is exactly a concrete eigenvalue.A regional area on the most one-dimensional all only corresponding facial image of local feature, therefore lays particular emphasis on the minutia extracting face.Characteristic vector obtained by face Global Information, the characteristic vector constituted with local message are combined together by the present invention, can reflect the thick spacious information of entirety of face, by the minute information that can reflect face local.The authentication codes of face can be obtained more accurately.

Figure 11 is the schematic diagram being extracted eigenvalue by the Skin Color Information of face.

The Skin Color Information of face is the important information quickly distinguishing different people, as shown in figure 11: (a) represents the face of yellow, and (b) represents white face, and (c) represents the face of Black people.The facial image of RGB color that will be read by mobile phone camera, carry out the conversion of Lab color space, monochrome information L is removed, the image of face is represented with a and b, with above-mentioned probability scale self-organized algorithm, calculate the gray value of the maximum distribution density of color a and b respectively, using the two gray value as the Skin Color Information of face the eigenvalue that obtained face complexion by membership function.

Figure 12 is the schematic diagram of the information retrieval eigenvalue by face frequency space.

As shown in figure 12: 1201 represent the face of face, and 1202 is the wrinkle of face forehead, and 1203 is the eye pouch of face eye, and 1204 is the decree stricture of vagina on mouth both sides.As shown in the 1202 of Figure 12: when the wrinkle of forehead is the most intensive, can be extracted out the feature of wrinkle by the frequency characteristic of the regional area of forehead.

As shown in the 1204 of Figure 12: when wrinkle ratio is more visible, differential calculation can be carried out by the image in the region to wrinkle place, find out the sideline of wrinkle, identify the length of wrinkle.For the differential calculation of wrinkle image, also it is belonging to carry out the calculating of frequency space for wrinkle image.

Figure 13 is as the processing method schematic diagram of the Global Information of face using the information of the frequency space of face.

Using the information of the frequency space of face as the Global Information of face, the degree of roughness of its reflection face, the number of such as wrinkle, cicatrix of face etc., for the white noise making recognition result will not produce when image reading because of the trickle position of face, and affect the stability of identification, the most only take the coefficient of the low frequency part of the fast fourier transform result that face image is carried out, as the vector that quantizes of the frequency space of face.

As shown in figure 13, the information of the frequency space of face is carried out in three steps as the process of the Global Information of face；

First step is to read facial image step, is here read out by the general image of face, as the object of process of information of the frequency space of face.

Second step is fast Fourier transform step, and the above-mentioned facial image read is carried out quick Fourier transform.

Third step is the vectorial composition step that quantizes, and the real number coefficient of the low frequency end of above-mentioned fast fourier transform result constitutes the vector that quantizes of the frequency space of face with imaginary number coefficient.

The present invention utilizes the face positional information of face, the dimension information of human face five-sense-organ, the information of the frequency space of face, the shape information of face and the Skin Color Information of face, it is achieved the certification to holder.But, under different mobile phone shooting environmental, above-mentioned information there will be the problem of random distribution within the specific limits, by above-mentioned probability scale self-organized algorithm, can obtain the expected value closest to probability distribution parent and dispersion value.

For more accurately the face of holder being transformed into more stable code, the present invention is considering the randomness of face information, and take maximally efficient method to calculate the eigenvalue closest to parent, import again the theory of fuzzy mathematics simultaneously, for the information that quantizes of the above-mentioned face obtained by the way of artificial intervention, use based on membership functions a plurality of defined in artificial experience (Membership Function), quantize between 0 to n numerical value, and directly generate the characteristic vector with image code character.

Figure 14 is the processing method schematic diagram quantizing voiceprint.

One-dimensional voiceprint again may be by Fourier transform and generates the information that quantizes.

As shown in figure 14, the process quantized voiceprint is carried out in three steps:

First step is to read vocal print signals step, is here read out by vocal print signal, as the object of the process of voiceprint information.

Second step is fast Fourier transform step, and the above-mentioned vocal print signal read is carried out quick Fourier transform.

Third step is the vectorial composition step that quantizes, and the real number coefficient of the low frequency end of above-mentioned fast fourier transform result and imaginary number coefficient constitute the vector that quantizes of voiceprint.

Figure 15 is the example that the shape information of face is defined as membership function.

As shown in Figure 15 (a): by a₁With a₂The straight line connected into, by a₂With a₃The straight line connected into, and by a₃With a₁The triangle that the straight line connected into is constituted, a ' is this triangle area in addition to the region shared by shape of face, and in general the thinnest when the shape of face of face, the area of a ' is just closer to the area with triangle, when the shape of face of face is the most fat, and the area remained by a ' is the least.Utilize this artificial subjective experience, following membership function can be constructed.If by a₁With a₃The straight line connected into is L₁, by a₂With a₃The straight line connected into is L₂, then the membership function MB that the shape information of face is defined as₁Can be made by formula 7:

[formula 7]

${\mathrm{MB}}_1=\frac{\frac{1}{2}\left(L_1{L}_2\right)-a^{\′}}{\frac{1}{2}\left(L_1{L}_2\right)}100=100[1-\frac{{2a}^{\′}}{L_1{L}_2}]$

When face is the most fat, its area a ' is close to " 0 ", MB₁Close to 100, when face is the thinnest, its area 2a ' is close to " L₁* L₂", MB₁Close to " 0 ", therefore membership function MB₁It is to describe the characteristic function that face is fat or thin.

Again for the example of membership function of a width calculating mouth, if mouth is V by the Breadth Maximum that statistics is obtained_max, minimum widith is V_min, as shown in Fig. 8 (b), set the width of mouth again as V₄, then the membership function MB of the width of mouth₂Can be made by formula 8:

[formula 8]

$\begin{array}{c}{\mathrm{MB}}_2=\frac{(V_{\max }-V_{\min })-(V_4-V_{\min })}{V_{\max }-V_{\min }}100\\ =\frac{V_{\max }-V_4}{V_{\max }-V_{\min }}100\end{array}$

Here setting Vmax ≠ Vmin, formula 8 gives when mouth is close to minimum widith, V₁≈V_minThe membership function MB of the width of mouth₂Close to 100, when entering mouth close to Breadth Maximum, V₄≈V_maxThe membership function MB of the width of mouth₂Close to 0.Therefore membership function MB₂It it is the characteristic function describing mouth width.

The canthus width of other faces, eyeball size, canthus distance, eyeball distance, the distance of nose size, nose and eye, the width of mouth, the thickness of mouth, the position of distance of mouth and the distance of nose, the width of cheek, mouth and forehead etc. reflection face and the information of size, and the wrinkle of face, the colour of skin of face, shape of face etc. information all can refer to the definition method of above-mentioned membership function, and face information is transformed into the characteristic vector of the standard figures of 0 to n, no matter use the membership function of what form, within being all belonging to the scope of the present invention.

Defined feature vector space here, if by the facial image of q different people according to the method described above, obtains q characteristic vector, and the feature of the facial image characteristic vector make-up formula 1 that each characteristic vector has p characteristic element asks quantity space.

It is provided with q individual by p elementary composition vector, can express with the determinant shown in formula 1.

[formula 9]

$V_{\mathrm{ij}}=\left(\begin{array}{cccc}{V}_{11,}& V_{12,}& \·\·\·,& V_{1p}\\ V_{21,}& V_{22,}& \·\·\·,& V_{2p}\\ & \·\·\·& & \\ v_{q1,}& v_{q2,}& \·\·\·,& v_{\mathrm{qp}}\end{array}\right)$

At the vectorial w calculating two probability distribution_i1, w_i2..., w_ipWith vector v_i1, v_i2..., v_ip, during the distance of the probability scale of the probability distribution of each corresponding key element, if vector v_i1, v_i2..., v_ij..., v_ipMiddle jth key element v_ijThe ordered series of numbers of probability distribution be g_j1, g_j2... g_{j ζ}, final central value A that calculated by formula 4_ij, and probability scale M_ij, then

[formula 10]

$|W_{\mathrm{ij}}-V_{\mathrm{ij}}|=\left\{\begin{array}{cc}0,& |W_{\mathrm{ij}}-V_{\mathrm{ij}}|\≤M_{\mathrm{ij}}\\ |W_{\mathrm{ij}}-V_{\mathrm{ij}}|,& |W_{\mathrm{ij}}-V_{\mathrm{ij}}|>M_{\mathrm{ij}}\end{array}\right.$

The result of formula 10 is brought into distance P of the most available probability scale of formula 11_i。

[formula 11]

$P_i={\left(\underset{j=1}{\overset{p}{\mathrm{\Σ}}}{|W_{\mathrm{ij}}-V_{\mathrm{ij}}|}^2\right)}^{\frac{1}{2}}$

Set holder's brush face image or the characteristic vector of vocal print signal againCharacteristic vector V with i-th facial image_i1, V_i2..., V_ip, corresponding Ψ identifies logged learning data through several times, constitutes the matrix L of learning data.

[formula 12]

$L=\left(\begin{array}{cccc}{l}_{11},& l_{12},& ...,& l_{1p}\\ l_{21,}& l_{22},& ...,& l_{2p}\\ & ...& & \\ l_{\mathrm{\ψ}1},& l_{\mathrm{\ψ}2,}& ...,& l_{\mathrm{\ψp}}\end{array}\right)$

According to above-mentioned formula Isosorbide-5-Nitrae and 12, carry out probability scale self-organization computation, central value and the dispersion value L of following learning data matrix can be respectively obtained_AM:

[formula 13]

L_AB=[(A₁, M₁), (A₂, M₂) ..., (A_p, M_p)]

Here is setSecondary when carrying out mobile-phone payment, holder's brush face image or the characteristic vector of vocal print signal With characteristic vector space V_ijIn each characteristic vector between the distance of probability scale according to formula 5,6 can obtain:

[formula 14]

$P=\left\{\begin{array}{c}{p}_1\\ p_2\\ \·\·\·\\ p_q\end{array}\right\}$

Wherein minima p_minCorresponding characteristic vector valueIt is brush face image or the code value of vocal print signal of γ people.

Figure 16 is brush face or the flow chart of vocal print code-adaptive learning processing method.

Facial image captured when holder's brush face or voiceprint or vocal print signal are extracted out by characteristic information, holder's brush face image that the calculating of membership function etc. obtain or the characteristic vector of vocal print signal, carry out the calculating of the distance of probability scale with each characteristic vector of characteristic vector space, obtain the characteristic vector corresponding to minima of probability scale distance of formula 14 as the brush face of holder or vocal print code.For improving brush face or the precision of vocal print payment code and stability, the present invention proposes an adaptive brush face or vocal print code generating method.

As shown in figure 16, brush face or vocal print code-adaptive learning processing method divide 3 steps to carry out.

Step 1: the composition step of adaptive characteristic vector space；In order to using each brush face or voice print database as the data once learnt, by substantial amounts of statistics, ensure that the real learning data that constantly will be closest to probability distribution parent retains, and the false data beyond deviation are rejected, the calculating of probability scale distance is carried out in each learning data of the key element of real characteristic vector, guarantee passes through adaptive learning, brush face or vocal print code is made increasingly to tend towards stability, making code building more accurate, this is the main purpose of adaptive learning.

In order to calculate brush face or the vocal print code of holder more accurately, first by the v in the brush face image of formula 1 or the characteristic vector space of vocal print signal_i1, by learning data Matrix Formula 12, and the distance for the probability scale space of learning data matrix is calculated formula 13, is replaced as the matrix A of the central value of probability scale metric space_ij, can using this matrix as adaptive characteristic vector space,

[formula 15]

$A_{\mathrm{ij}}=\left(\begin{array}{cccc}{a}_{11,}& a_{12,}& \·\·\·,& a_{1p}\\ a_{21,}& a_{22,}& \·\·\·,& a_{2p}\\ & \·\·\·& & \\ l_{q1,}& l_{q2,}& \·\·\·,& l_{\mathrm{qp}}\end{array}\right)$

Again by formula 12,13 can obtain characteristic vector space v_ijIn the dispersion matrix of probability scale metric space, i.e.

[formula 16]

$M_{\mathrm{ij}}=\left(\begin{array}{cccc}{m}_{11,}& m_{12,}& \·\·\·,& m_{1p}\\ m_{21,}& m_{22,}& \·\·\·,& m_{2p}\\ & \·\·\·& & \\ m_{q1,}& m_{q2,}& \·\·\·,& m_{\mathrm{qp}}\end{array}\right)$

And brush face or the matrix D of vocal print payment code_ij, i.e.

[formula 17]

$D_{\mathrm{ij}}=\left(\begin{array}{cccc}{d}_{11,}& d_{12,}& \·\·\·,& d_{1p}\\ d_{21,}& d_{22,}& \·\·\·,& d_{2p}\\ & \·\·\·& & \\ d_{q1,}& d_{q2,}& \·\·\·,& d_{\mathrm{qp}}\end{array}\right)$

The central value of the probability scale metric space obtained by formula 15 and 16, it is changing along with the change of learning data of same the following stated with the dispersion value of probability scale metric space, therefore can have adaptive characteristic, be the data of the parent constantly tending to probability distribution.So process can ensure that the retrieval of the brush face image of holder or the characteristic vector space of vocal print signal code is in optimal state, and the code that can improve brush face or vocal print payment is in the result of calculation of optimum.

Here, as shown in Equation 17: change owing to the central value of probability scale metric space is as the change of learning data, it is thus desirable to introduce one with each key element of the central value matrix of probability scale metric space corresponding to, but its numerical value is constant brush face or the matrix D of vocal print payment code_ij。

If theWhen secondary brush face pays, obtained characteristic vectorAs shown in Equation 15, the characteristic vector of holder's brush face image or vocal print signal and self-adaptive features vector space A_ijDistance P of probability scale_iAs follows:

[formula 18]

[formula 19]

When retrieval theSecondary carry out brush face pay time, holder's brush face image or the characteristic vector of vocal print signal Central value A with probability scale metric space_ijIn each self-adaptive features vector between the minimum self-adaptive features vector of the distance of probability scale, corresponding brush face or the matrix D of vocal print payment code_ijCode value, so that it may asThe code that secondary brush face or vocal print pay.

Step 2: adaptive learning brush face or vocal print payment code acquisition step；Through the process of above-mentioned steps, by holderSecondary carry out brush face image or the characteristic vector of vocal print signal that brush face or vocal print pay Distance through probability scale calculates, retrieve an adaptive learning characteristic vector in the self-adaptive features vector space corresponding to the distance of minimum probability yardstick, find again and a code in the matrix of the brush face corresponding to this adaptive learning characteristic vector or vocal print payment code, using this code as the brush face of adaptive learning or vocal print payment code.

Various method can be had to realize the composition of adaptive learning data matrix according to the thinking of the present invention, now take a single example and illustrate, with reference to formula 14, if holder theSecondary carry out brush face image or the characteristic vector of vocal print signal that brush face or vocal print payMatrix A with the central value of the metric space of probability scale_ijThe minimum range that some characteristic vector in the self-adaptive features vector space constituted also exists is Another characteristic vector also exists ratioThe biggest probability scale minimum range is

Holder's brush face image or the characteristic vector of vocal print signal can be set againThe matrix L of the learning data of the formula 12 constituted through identifying logged learning data several times with Ψ_ijIn each characteristic vector between the distance of probability scale according to formula 5,6 can obtain:

[formula 20]

$P^{\′}=\left\{\begin{array}{c}{p_1}^{\′}\\ {p_2}^{\′}\\ \·\·\·\\ {p_{\mathrm{\ψ}}}^{\′}\end{array}\right\}$

By the P '={ p of formula 20₁', p₂' ..., p_Ψ' distance of probability scale, then carry out probability scale self-organization computation by formula 5,6, i.e. available for holder's brush face image or the characteristic vector of vocal print signalMatrix L with learning data_ijCentral value A ', and dispersion M '.

If meetAndOr () then, brush face image or the characteristic vector of vocal print signalCan be as the matrix L of learning data_ijNew study vector.

Here, as one of condition adding learning data it isIts physical significance is: limit theThe condition that the secondary characteristic vector carrying out brush face or the brush face image of vocal print payment or vocal print signal enters learning data is to try to the distance of the probability scale certain with the holding of other characteristic vectors, to prevent the phenomenon distinguished from occurring by mistake.

Range of conditionPhysical significance be, when differing bigger with the distance of the probability scale of this pattern, will not be because of the characteristic vector of brush face imageIt is not appropriate for entering entering in learning data of learning data and mistake.

About the method how rejecting the inadaptable study vector as learning data, first each key element of the distance of the probability scale in formula 20 is carried out { ω=p_i-A’；I=1,2 ..., Ψ } deviation calculating, by maximum deflection difference value ω_maxCorresponding learning data is rejected and just may be used.

Here, merely provide a characteristic vector data the most adaptive entrance learning data, and how to reject a method being not suitable as learning data, but reality is often for several data simultaneously when system is constituted, or enter step by step in learning data, and reality is often for several data simultaneously when system is constituted, or reject from learning data step by step.As long as with reference to said method with regard to measurable to relevant processing method.

In brush face pays, holder is often due to change hair style, or carry out different cosmetics etc., often occur that the characteristic vector of brush face image has greatly changed, for solving these problems, the present invention also proposes while setting up the matrix of learning data, the characteristic vector of α time currently obtained recently is recorded, and becomes the matrix of CAL data:

[formula 21]

$L^{\′}=\left(\begin{array}{cccc}{l_{11}}^{\′},& {l_{12}}^{\′},& \·\·\·,& {l_{1p}}^{\′}\\ {l_{21}}^{\′},& {l_{22}}^{\′}& \·\·\·,& {l_{2p}}^{\′}\\ & \·\·\·& & \\ {l_{\mathrm{\α}1}}^{\′}& {l_{\mathrm{\α}2}}^{\′}& \·\·\·,& {l_{\mathrm{\αp}}}^{\′}\end{array}\right)$

If central value A calculated in the probability scale metric space of the matrix of CAL data "; and dispersion value M " it is in state that is stable and that restrain, additionally; A " with M " merge the matrix of the new learning data of regeneration with the matrix L of learning data with central value A of the matrix of learning data ' and dispersion value M '; when there is larger distance in probability scale metric space; can be the matrix L of CAL data ', or carry out the matrix of CAL data and displacement of learning data matrix etc..Concrete grammar is referred to above-mentioned example and the mode of thinking that the present invention proposes, and draws inferences about other cases from one instance and can obtain various process means.

Step 3: adaptive learning brush face or vocal print payment code acquisition step: through the process of above-mentioned steps, by holderSecondary carry out brush face image or the characteristic vector of vocal print signal that brush face or vocal print pay Distance through probability scale calculates, retrieve an adaptive learning characteristic vector in the self-adaptive features vector space corresponding to the distance of minimum probability yardstick, find again and a code in the matrix of the brush face corresponding to this adaptive learning characteristic vector or vocal print payment code, using this code as the brush face of adaptive learning or vocal print payment code.

Brush face that the present invention proposes or vocal print pay and are applied not only to commodity and pay, it is also possible to as common credit card when paying, carry out a kind of mode of my certification, it is possible to resolve and the generation of the problem of illegal payment that produce stolen due to the credit card.

During it addition, the brush face of present invention proposition or vocal print method of payment are also used as the cash withdrawal of ordinary bank's cash card, a brush face certification or the step of voiceprint can be increased, the safety of bank cash card can be improved.

The photo utilizing holder in order to prevent illegal person carries out the criminal behavior of brush face payment, needs brush face image is made whether the identification into life entity image, the method that the present invention proposes to carry out life entity image recognition for brush face image as follows.

Figure 17 is the schematic diagram of one of the example of life entity image recognition.

As shown in Figure 17 (a): above-mentioned Fig. 8 gives by the method for probability scale self-organized algorithm identification eyeball size, in order to identify whether as life entity, eyeball according to face is that eyeball can blink in the feature of the lived brush face image of tool, therefore, as shown in Figure 17 (b): use the probability scale self-organized algorithm identifying eyeball size, when brush face image is blinked, the M of Figure 17 (a)⁽ⁿ⁾It is far longer than the M of Figure 17 (b)^{(n) ,}, i.e. M^{(n) ’}＜ M⁽ⁿ⁾State instantaneous possess.

Brush face image is that the identification that above-mentioned eyeball is blinked is not only in the identification of the image of life entity, can also open one's mouth the action shut up according to above-mentioned probability scale self-organized algorithm method identification, muscle minor variations when face smiles, the change of pupil, identify the face small acceleration rocked, identify the method rocked etc. of the color head of face.

Figure 18 is the schematic diagram of the electronic image of the 3D half-tone screen code that optical identification is possible.

As shown in figure 18: (1800) are the screens of screen display, it is possible to be the display screen of mobile phone.The definition of the electronic image of the 3D half-tone screen code that the optical identification of display is possible on so-called screen, as shown in the 1801 of Figure 18: first for the code that optics can be distinguished can be become, first the symbol of its code should be the division of each two dimension on electronic curtain, on screen, i.e. at least it is divided into the zonule symbol as code of several sizes can distinguished more than optics, each symbol is in record information, it it is the different color by including symbol, diverse location, different size, different directions, centralization and decentralization are in interior geometric distribution, or by different modulation systems, different phase contrasts, the different directions of propagation, different intensity profile is in interior physical distribution.

Each symbol of the 3D half-tone screen code that above-mentioned optical identification is possible is by two dimension, the distribution record information of three-dimensional or space-time.Simultaneously as shown in the 1802 of Figure 18: each symbol of the 3D half-tone screen code that optical identification is possible is the most simultaneously by the distribution record information of time.I.e. also can be with red for each symbol, green, blue, in vain, various possible color that is black and that thus combine, also often can be changed in a flash by amplitude modulation 1804 or the mode of frequency modulation 1803, such as by friction speed record information (can be described as frequency modulated information) of this chromatic flicker, record information by the intensity (can be described as AM information) of black or white flicker.

The various information that can show on mobile phone screen, such as: two-dimensional bar code 1805, QR code, Chinese letter co, the mark trade mark etc. of commodity can also record information according to the mode of amplitude modulation 1807 with frequency modulation 1806 in every flashy change.

Described 3D half-tone screen code refers to all QR Quick Response Codes that can launch information on screen, DM Quick Response Code, PDF417 Quick Response Code, Chinese letter co, half-tone screen code, the image that any screen shows.

For not affecting the content that screen shows, and utilize the display space of screen as far as possible, from space field, such as diverse location record information can be taked, so symbol only uses the space of less than 2%, and the space of more than 98% is used for the display of screen picture.

Symbol can also be hidden in certain single order gray scale of display image screen, such as, symbolic information is placed on minimum some positions of bitmap.

From time field, utilize the visual characteristic of the movement perception of human eye, the flicker of symbol can be set less than 0.3 second of picture display times, display image screen occupies the main display time, under being remembered by human eye, just it is not easy to find the state that symbol records information in time field, the effect of Information hiding can be played.

The 3D half-tone screen code that optical identification is possible in a word is the integration visual characteristic that make use of time domain and spatial domain, make screen can normal display screen curtain image, each symbol can be allowed again to imbed information in time domain with spatial domain, can accomplish that again the impact on display image is minimum.

Claims (10)

1. a constructive method for the cash-access system of mobile phone deposit card, including:

The cash of mobile phone deposit card be stored in system constructive method and

The constructive method of the cash withdrawal system of mobile phone deposit card；It is characterized in that, wherein:

It is to be recognized by the communication of mobile phone savings card number that the cash of mobile phone deposit card is stored in the constructive method of system Card step, the cash of mobile phone deposit card is stored in step composition, and its feature is as follows:

The communication authentication step of mobile phone savings card number, by mobile phone and ATM communication, or passes through Screen at mobile phone deposit card shows the 3D half-tone screen code image that can distinguish of optics, by ATM and Read this image and realize certification, complete to be sent in ATM the card number of deposit card；

The cash of mobile phone deposit card is stored in step, and ATM receives certified savings card number and is stored in Cash after automatically cash is stored in the account of mobile phone deposit card；

The constructive method of the cash withdrawal system of a kind of mobile phone deposit card, is extracted cash by input Number step, mobile phone deposit card authenticating step, extract cash step composition, its feature is as follows:

The number step of cash is extracted in input, selects mobile phone deposit card kind, if only one can To skip, on ATM, the number of cash is extracted in input, or extracts amount in cash in mobile phone terminal input, The number of input extraction cash is realized with communicating of ATM by mobile phone；

Mobile phone deposit card authenticating step, by the local feature of the face of holder and global feature phase Merge is completed the face characteristic information of mobile phone brush face by " probability scale metric space " treatment technology Brush face certification, or the sound characteristic being had by oneself by holder processes skill by " probability scale metric space " Art completes the voiceprint with adaptive learning；

Take gold step of withdrawing deposit, ATM receives the savings card number of input, and input extract cash The amount of money, just can enter the operation of enchashment gold after mobile phone brush face or voiceprint success.

The constructive method of the cash-access system of a kind of mobile phone deposit card the most according to claim 1, its It is characterised by: the tool that described 3D half-tone screen code image is the screen by ATM or mobile phone plane shows Have and include Two-dimensional electron image, with the three-dimensional electronic image of many-valued gray value, and with flicker Many-valued gray value three-dimensional electronic image in a kind of electronic image that can be displayed on screen.

The constructive method of the cash-access system of a kind of mobile phone deposit card the most according to claim 1, its It is characterised by: the face characteristic information of described probability scale metric space, refers to face random The characteristic information of distribution, by including normal distribution (Normal distribution), exponential (Exponential distribution), Erlangian distribution (Erlang Distribution), Wei Uncle's distribution (Weibull distribution), angular distribution (triangular distribution), In beta distribution (Beta Distribution), at least one has the probability attribute of probability distribution Parameter as self-organizing probability scale, the tool obtained eventually through the algorithm of the self-organizing of probability scale There is the characteristic information of relative certainty.

The constructive method of the cash-access system of a kind of mobile phone deposit card the most according to claim 1, its It is characterised by: the described face recognition to holder is to rely on the five of the holder of mobile phone photograph Official's face image is the coloured image without monochrome information, is not losing face image color information On the basis of realize being transformed into face image can be with the code of specific holder.

The constructive method of the cash-access system of a kind of mobile phone deposit card the most according to claim 4, its It is characterised by: the face image of described holder is under a certain color, and at a certain gray value Under the density rule of pixel distribution, import probability scale self-organized algorithm, automatically by face five Official's positional information is extracted out.

The constructive method of the cash-access system of a kind of mobile phone deposit card the most according to claim 1, its It is characterised by: described face characteristic information is obtained closest to parent by probability self-organized algorithm Eigenvalue, import again the theory of fuzzy mathematics simultaneously, the information for the above-mentioned face obtained is led to Cross the mode of artificial intervention, use based on membership functions a plurality of defined in artificial experience (Membership Function), carries out quantizing to directly generate having between 0 to n numerical value The characteristic vector of image code character.

The constructive method of the cash-access system of a kind of mobile phone deposit card the most according to claim 1, its It is characterised by: the generation of described brush face authentication codes is the action of the nictation by including eyeball, Opening one's mouth the action shut up, muscle minor variations when face smiles, the change of pupil, face is small The acceleration rocked, the color of face is at the brush face figure of at least one reflection life entity characteristics of image interior The identification of the life entity image of picture.

The constructive method of the cash-access system of a kind of mobile phone deposit card the most according to claim 1, its It is characterised by: the generation of described voiceprint code is can to pass through random proposing by voiceprint Go out some enquirements relevant with the problem that holder logs in advance, allow voiceprint person answer, thus Identify whether voiceprint person is life entity；The life entity identification of voiceprint is also by plural number Password whether be in the methods such as identical state to realize the identification of life entity vocal print.

The constructive method of the cash-access system of a kind of mobile phone deposit card the most according to claim 1, its It is characterised by: described face characteristic information is the local feature information by brush face image, with whole Body characteristics information collectively forms.

The constructive method of the cash-access system of a kind of mobile phone deposit card the most according to claim 1, its It is characterised by: described voiceprint is the local feature information by vocal print signal, special with entirety Reference breath, collectively form.