CN109583387A - Identity identifying method and device - Google Patents

Identity identifying method and device Download PDF

Info

Publication number
CN109583387A
CN109583387A CN201811458913.2A CN201811458913A CN109583387A CN 109583387 A CN109583387 A CN 109583387A CN 201811458913 A CN201811458913 A CN 201811458913A CN 109583387 A CN109583387 A CN 109583387A
Authority
CN
China
Prior art keywords
biological information
network model
certified
predetermined
authentication
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Pending
Application number
CN201811458913.2A
Other languages
Chinese (zh)
Inventor
聂镭
沙露露
郑权
张峰
聂颖
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Dragon Horse Zhixin (zhuhai Hengqin) Technology Co Ltd
Original Assignee
Dragon Horse Zhixin (zhuhai Hengqin) Technology Co Ltd
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Dragon Horse Zhixin (zhuhai Hengqin) Technology Co Ltd filed Critical Dragon Horse Zhixin (zhuhai Hengqin) Technology Co Ltd
Priority to CN201811458913.2A priority Critical patent/CN109583387A/en
Publication of CN109583387A publication Critical patent/CN109583387A/en
Pending legal-status Critical Current

Links

Classifications

    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06VIMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V40/00Recognition of biometric, human-related or animal-related patterns in image or video data
    • G06V40/70Multimodal biometrics, e.g. combining information from different biometric modalities
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06FELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00Pattern recognition
    • G06F18/20Analysing
    • G06F18/25Fusion techniques
    • G06F18/253Fusion techniques of extracted features

Abstract

The invention discloses a kind of identity identifying method and devices.Wherein, this method comprises: obtaining biological information to be certified;Convert biological information to be certified to the input of authentication network model, wherein, authentication network model is obtained using multiple groups training data by machine learning training, and every group of training data in multiple groups training data includes: predetermined object belonging to biological information and biological information;Obtain the output result of authentication network model, wherein export the predetermined object that result includes: predetermined quantity and the probability that biological information to be certified is each predetermined object in predetermined quantity;Target object belonging to biological information to be certified is determined according to output result.The present invention solves in the related technology for carrying out the lower technical problem of the reliability of the relatively simple caused identification of the mode of identification.

Description

Identity identifying method and device
Technical field
The present invention relates to technical field of information processing, in particular to a kind of identity identifying method and device.
Background technique
In the epoch of this current information development, identity identifying technology is one of the important technology to ensure information safety, raw Object identification technology is widely used in authentication, since everyone biological characteristic has the uniqueness of itself and at one section Be able to maintain in time stablize it is constant, compared with traditional authentication, using biological characteristic carry out authentication will more pacify Entirely, conveniently, reliably and accurately.In existing bio-identification mode, recognition of face, fingerprint recognition, Application on Voiceprint Recognition identification method list One.
For above-mentioned in the related technology for carrying out the reliable of the relatively simple caused identification of the mode of identification The lower problem of property, currently no effective solution has been proposed.
Summary of the invention
The embodiment of the invention provides a kind of identity identifying method and devices, at least to solve in the related technology for carrying out The lower technical problem of the reliability of identification caused by the mode of identification is relatively simple.
According to an aspect of an embodiment of the present invention, a kind of identity identifying method is provided, comprising: obtain life to be certified Object characteristic information;Convert the biological information to be certified to the input of authentication network model, wherein the body Part certification network model is obtained using multiple groups training data by machine learning training, every in the multiple groups training data Group training data includes: predetermined object belonging to biological information and the biological information;The identity is obtained to recognize Demonstrate,prove the output result of network model, wherein the output result includes: the predetermined object and the life to be certified of predetermined quantity Object characteristic information is the probability of each predetermined object in the predetermined quantity;It is determined according to the output result described to be certified Biological information belonging to target object.
Optionally, before obtaining biological information to be certified, the identity identifying method further include: obtain to be certified The biological information of object, and the biological information is pre-processed;According to the pretreated biological information and the life The corresponding semaphore of object information obtains the biological information to be certified.
Optionally, pre-process to the biological information includes: that there are storage classes in the biological information to scheme When the biological information of picture, the corresponding image of biological information that storage class is image is converted into the grayscale image to conform to a predetermined condition Picture;In the biological information there are storage class be audio biological information when, by predetermined way to storage class be sound The biological information of frequency is pre-processed, wherein the predetermined way includes at least one of: framing adding window, end-point detection, language Sound noise reduction.
Optionally, obtaining the authentication network model by training data training includes: to the trained number Biological attribute data in is pre-processed;Using the pretreated biological attribute data as the authentication net The input layer of network model;By the processing network layer in the authentication network model to the biological characteristic of input Data are handled, and output node layer is obtained, wherein the processing network layer is in the authentication network model except input The network layer of layer and output layer;It is trained according to the input layer and the output node layer, obtains the identity and recognize Demonstrate,prove network model.
Optionally, the training data is a part for the historical data collected from multiple users, the history number Another part in is as verify data, wherein the verify data is for testing the authentication network model Card.
Optionally, determine that target object belonging to the biological information comprises determining that institute according to the output result State the maximum probability in the probability that biological information is each predetermined object in the predetermined quantity;By the maximum probability Corresponding predetermined object is as target object belonging to the biological information.
Optionally, the full articulamentum of each of multiple full articulamentums in the authentication network model is connected to same A network structure, to be merged to every kind of characteristic information in biological information.
Another aspect according to an embodiment of the present invention, additionally provides a kind of identification authentication system, comprising: first obtains Unit, for obtaining biological information to be certified;Converting unit, for converting the biological information to be certified For the input of authentication network model, wherein the authentication network model is to pass through machine using multiple groups training data What learning training obtained, every group of training data in the multiple groups training data includes: biological information and the biology Predetermined object belonging to characteristic information;Second acquisition unit, for obtain the output of the authentication network model as a result, its In, the output result includes: the predetermined object of predetermined quantity and the biological information to be certified is the predetermined number The probability of each predetermined object in amount;Determination unit, for determining that the biology to be certified is special according to the output result Target object belonging to reference breath.
Optionally, the identification authentication system further include: processing unit, for obtaining the biological characteristic letter to be certified Before breath, the biological information of object to be certified is obtained, and pre-process to the biological information;Third acquiring unit, is used for The biological characteristic to be certified is obtained according to the pretreated biological information and the corresponding semaphore of the biological information Information.
Optionally, the processing unit includes: conversion module, is schemed for there are storage classes in the biological information When the biological information of picture, the corresponding image of biological information that storage class is image is converted into the grayscale image to conform to a predetermined condition Picture;First processing module, for, there are when the biological information that storage class is audio, passing through predetermined party in the biological information Formula pre-processes the biological information that storage class is audio, wherein the predetermined way includes at least one of: framing Adding window, end-point detection, voice de-noising.
Optionally, the converting unit includes: Second processing module, for the biological characteristic number in the training data According to being pre-processed;First determining module, for using the pretreated biological attribute data as the authentication net The input layer of network model;First obtains module, for passing through the processing network layer pair in the authentication network model The biological attribute data of input is handled, and output node layer is obtained, wherein the processing network layer is that the identity is recognized Demonstrate,prove the network layer that input layer and output layer are removed in network model;Second obtains module, for according to the input layer and institute It states output node layer to be trained, obtains the authentication network model.
Optionally, the training data is a part for the historical data collected from multiple users, the history number Another part in is as verify data, wherein the verify data is for testing the authentication network model Card.
Optionally, the determination unit includes: the second determining module, for determining that the biological information is described pre- Maximum probability in the probability of each predetermined object in fixed number amount;Second determining module, for the maximum probability is corresponding Predetermined object as target object belonging to the biological information.
Optionally, the full articulamentum of each of multiple full articulamentums in the authentication network model is connected to same A network structure, to be merged to every kind of characteristic information in biological information.
Another aspect according to an embodiment of the present invention, additionally provides a kind of storage medium, the storage medium includes The program of storage, wherein described program execute it is any one of above-mentioned described in identity identifying method.
Another aspect according to an embodiment of the present invention, additionally provides a kind of processor, the processor is for running Program, wherein described program run when execute it is any one of above-mentioned described in identity identifying method.
In embodiments of the present invention, the biological information to be certified using acquisition;By biological information to be certified It is converted into the input of authentication network model, wherein authentication network model is to pass through machine using multiple groups training data What learning training obtained, every group of training data in multiple groups training data includes: biological information and biological information Affiliated predetermined object;Obtain the output result of authentication network model, wherein output result includes: the pre- of predetermined quantity Determine object and biological information to be certified is the probability of each predetermined object in predetermined quantity;It is determined according to output result Target object belonging to biological information to be certified, the middle identity identifying method provided can be real through the embodiment of the present invention A variety of biological informations are now converted into biological information, and biological information is merged, and according to fused life Object characteristic information carries out the purpose of authentication, has reached the technical effect for improving the reliability of authentication, and then solve In the related technology for carrying out the lower technical problem of the reliability of the relatively simple caused identification of the mode of identification.
Detailed description of the invention
The drawings described herein are used to provide a further understanding of the present invention, constitutes part of this application, this hair Bright illustrative embodiments and their description are used to explain the present invention, and are not constituted improper limitations of the present invention.In the accompanying drawings:
Fig. 1 is the flow chart of identity identifying method according to an embodiment of the present invention;
Fig. 2 is the schematic diagram of identification authentication system according to an embodiment of the present invention.
Specific embodiment
In order to enable those skilled in the art to better understand the solution of the present invention, below in conjunction in the embodiment of the present invention Attached drawing, technical scheme in the embodiment of the invention is clearly and completely described, it is clear that described embodiment is only The embodiment of a part of the invention, instead of all the embodiments.Based on the embodiments of the present invention, ordinary skill people The model that the present invention protects all should belong in member's every other embodiment obtained without making creative work It encloses.
It should be noted that description and claims of this specification and term " first " in above-mentioned attached drawing, " Two " etc. be to be used to distinguish similar objects, without being used to describe a particular order or precedence order.It should be understood that using in this way Data be interchangeable under appropriate circumstances, so as to the embodiment of the present invention described herein can in addition to illustrating herein or Sequence other than those of description is implemented.In addition, term " includes " and " having " and their any deformation, it is intended that cover Cover it is non-exclusive include, for example, the process, method, system, product or equipment for containing a series of steps or units are not necessarily limited to Step or unit those of is clearly listed, but may include be not clearly listed or for these process, methods, product Or other step or units that equipment is intrinsic.
Embodiment one
According to embodiments of the present invention, a kind of embodiment of the method for identity identifying method is provided, it should be noted that attached The step of process of figure illustrates can execute in a computer system such as a set of computer executable instructions, though also, So logical order is shown in flow charts, but in some cases, it can be to be different from shown by sequence execution herein Or the step of description.
Fig. 1 is the flow chart of identity identifying method according to an embodiment of the present invention, as shown in Figure 1, the identity identifying method Include the following steps:
Step S102 obtains biological information to be certified.
Wherein, above-mentioned biological information is converted to by the biological information of the target object acquired, wherein the life Object information can include but is not limited to following several: face, fingerprint and sound etc., i.e., the biological information can for it is all with The variation at age and the biological information changed, also may include venous information.
As a kind of optional embodiment, before obtaining biological information to be certified, which can With further include: the biological information of object to be certified is obtained, and biological information is pre-processed;According to pretreated biology letter It ceases semaphore corresponding with biological information and obtains biological information to be certified.
Wherein, carrying out pretreatment to biological information may include: that there are the lifes that storage class is image in biological information When object information, the corresponding image of biological information that storage class is image is converted into the gray level image to conform to a predetermined condition;? There are when the biological information that storage class is audio in biological information, the biology that storage class is audio is believed by predetermined way Breath is pre-processed, wherein predetermined way includes at least one of: framing adding window, end-point detection, voice de-noising.
For example, the biological information of the object to be certified obtained includes: face, fingerprint, sound.The life that can so will acquire Object information is stored, and storage mode specifically can be stores face and fingerprint in the form of images, sound Then stored in the form of audio.It wherein, can be by storage form after being stored biological information in a predetermined format It is cut by color conversion at gray level image, and by the size of the gray level image after turning for the corresponding image of biological information of image The specification that authentication network model can identify;It deposits in the case of voice, sound can be carried out in biological information simultaneously The processing such as framing adding window, end-point detection and voice de-noising.
It should be noted that in embodiments of the present invention, if biological information includes: face, fingerprint, sound, providing Recognition sequence of the authentication network model to biological information are as follows: face, vocal print, sound.
In addition, obtaining biological characteristic to be certified according to pretreated biological information and the corresponding semaphore of biological information Information may include: the format and the semaphore that the semaphore of the biological information is arranged according to the type of the biological information.
Step S104 converts biological information to the input of authentication network model, wherein authentication network Model is obtained using multiple groups training data by machine learning training, and every group of training data in multiple groups training data wraps It includes: predetermined object belonging to biological information and biological information.
Step S106 obtains the output result of authentication network model, wherein output result includes: predetermined quantity Predetermined object and biological information to be certified are the probability of each predetermined object in predetermined quantity.
Step S108 determines target object belonging to biological information to be certified according to output result.
Through the above steps, available biological information to be certified;Biological information to be certified is converted For the input of authentication network model, wherein authentication network model is to pass through machine learning using multiple groups training data What training obtained, every group of training data in multiple groups training data includes: belonging to biological information and biological information Predetermined object;Obtain the output result of authentication network model, wherein output result includes: predetermined pair of predetermined quantity As the probability with biological information to be certified for each predetermined object in predetermined quantity;It is determined according to output result wait recognize Target object belonging to the biological information of card.Relative in the related technology when carrying out authentication, the authenticating party of use Formula is relatively simple, the lower drawback of the reliability of caused identification, through the embodiment of the present invention the middle authentication provided Method may be implemented a variety of biological informations being converted to biological information, and merge to biological information, and according to Fused biological information carries out the purpose of authentication, has reached the technical effect for improving the reliability of authentication, And then it solves lower for carrying out the reliability of the relatively simple caused identification of the mode of identification in the related technology The technical issues of.
In above-mentioned steps S104, obtaining authentication network model by training data training may include: to training Biological attribute data in data is pre-processed;Using pretreated biological attribute data as authentication network model Input layer;The biological attribute data of input is handled by the processing network layer in authentication network model, is obtained To output node layer, wherein processing network layer is to remove the network layer of input layer and output layer in authentication network model;According to Input layer and output node layer are trained, and obtain authentication network model.
Wherein, before obtaining authentication network model by training data training, which can be with It include: acquisition training data, it is preferred that above-mentioned training data is a part for the historical data collected from multiple users. The historical data is then to carry out biomedical information acquisition to the predetermined object of magnanimity in historical time section, and by the biological information pair The object record answered is got off, and will include biological information and the corresponding object of the biological information as historical data.
After collecting above-mentioned historical data, need to include that face, fingerprint and sound etc. is not of the same race to acquisition The biological information of class is pre-processed.In addition, after handling the biological information of acquisition, further includes: by the life of acquisition Object information is stored according to predetermined format, wherein the information of face and fingerprint pattern is stored in the form of images, will The information of the types such as sound is stored in the form of audio.It then, is the corresponding coloured silk of biological information of image by storage form Chromatic graph picture is converted to gray level image, and the gray level image is cut to image of the same size, and be audio to storage form Image carries out the pretreatment such as framing adding window, end-point detection and voice de-noising.
Then, three semaphores [s1, s2, s3] are set, the value for working as s1, s2 or s3 indicates signal when being 1, s1, s2 or The value of s3 indicates no signal when being 0, then value shares eight kinds of modes, and since [0,0,0] indicates s1, s2, s3 is without signal Pass through, therefore this kind of situation is rejected.Therefore, seven kinds of value modes are shared, [0,0,1] is respectively as follows:, [0,1,0], [0,1,1], [1,0,0], [1,0,1], [1,1,0], [1,1,1].Then, the value in above-mentioned semaphore is randomly selected, after and pre-processing Biological information training data is obtained by semaphore.For example, obtained biological information is [Xy, Xp, Xs], wherein Xy table Show face information, Xp indicates that finger print information, Xs indicate that acoustic information, obtained semaphore are [1,1,0], the then data set generated For [Xy, Xp, zeros ()], wherein zeros () is indicated and the consistent full null matrix of Xs dimension.Wherein, " 1 " in the semaphore Indicate that signal passes through, " 0 " indicates that no signal passes through.
In addition, another part in above-mentioned historical data is as verify data, wherein verify data is used for authentication Network model is verified.
Preferably, after obtaining above-mentioned training data and verify data, using the training data to authentication network Model is trained, wherein the network layer of the authentication network model may include: input layer, volume base, excitation layer, pond Change layer, full articulamentum and output layer.Wherein, input layer is the network layer for input data, it will usually be done at some data Reason;Convolutional layer is for carrying out feature extraction;Pond layer is mainly used for placing over-fitting (that is, guaranteeing the authentication network mould The authentication result that type not only verifies the information in some training datas is reliable, while guaranteeing will be some New biological information is input to same available reliable output result after authentication network model);In addition, A pond layer is all connected with after each volume base;The result of convolutional layer is then mainly done Nonlinear Mapping by excitation layer, wherein is swashed Encouraging layer functions has sigmoid, tanh, Relu, Leaky Relu, ELU, Maxout;Full articulamentum would generally be in convolutional Neural net The tail portion of network, all neurons between adjacent two layers all have the right to reconnect, and output layer is for exporting calculated result.
It should be noted that the number of plies and parameter in network layer need to carry out parameter selection and excellent for training data Change.
Preferably, the full articulamentum of each of multiple full articulamentums in authentication network model is connected to the same net Network structure, to be merged to every kind of characteristic information in biological information.That is, when every kind of biological information passes through respectively It, can be respectively through pulleying base, excitation layer and pond layer etc. after different input layer inputs, wherein above-mentioned every kind of biology is special Reference breath by full articulamentum processing after, can be aggregated into a network structure realize by every kind of biological information into The purpose of row fusion, specifically can be by being connected to the same network structure for each full articulamentum.Every kind of biological characteristic letter It ceases corresponding full articulamentum to be all connected in the same network structure, may be implemented to merge biological information, And the purpose of parameter training.In addition, after obtaining fused output result, it is also necessary to carry out normalizing to output result Change, gradient occurs to avoid web results in the training process and disappear and gradient explosion phenomenon.Then, it is carried out to result of publishing books After normalization, softmax layers are entered into, identification is carried out to training data.
As a kind of optional embodiment, in step S108, determined belonging to biological information according to output result Target object may include: that determining biological information is most general in the probability of each predetermined object in predetermined quantity Rate;Using the corresponding predetermined object of maximum probability as target object belonging to biological information.
For example, in embodiments of the present invention, the final output of authentication network model can be the matrix of a n*n, Include: in the matrix biological information may belonging to object and the biological information belong to belonging to the possibility pair The probability of elephant.So, it after obtaining the output result of the output layer of authentication network model, can be selected from output result Maximum probability predetermined object is selected as target object belonging to the biological information.
The identity identifying method provided through the embodiment of the present invention is it is possible to prevente effectively from biometric identity authenticates in the related technology Method is relatively simple, and the novel excessive drawback affected by environment of different biological characteristics provides a kind of bio-identification skill of fusion Art comprehensive organism characteristic information and can be applied to multiple scenes, for example, usually regarding to monitoring in terms of public security investigation Frequency and the fingerprint at scene are checked and are compared one by one, if the two is carried out uniformly, investigation range can be reduced, and for It is difficult to obtain the place of video image and fingerprint below, energy function provides new point of penetration to Application on Voiceprint Recognition again.Therefore, the present invention is implemented The identity identifying method that example provides has good robustness and applicability.
Embodiment two
A kind of identification authentication system is additionally provided according to embodiments of the present invention, it should be noted that the embodiment of the present invention Identification authentication system can be used for executing identity identifying method provided by the embodiment of the present invention.The embodiment of the present invention is mentioned below The identification authentication system of confession is introduced.
Fig. 2 is the schematic diagram of identification authentication system according to an embodiment of the present invention, as shown in Fig. 2, the identification authentication system It may include: first acquisition unit 21, converting unit 23, second acquisition unit 25 and determination unit 27.Below to the identity Authentication device is described in detail.
First acquisition unit 21, for obtaining biological information to be certified.
Converting unit 23 is connect with above-mentioned first acquisition unit 21, for converting biological information to be certified to The input of authentication network model, wherein authentication network model is to be instructed using multiple groups training data by machine learning It gets, every group of training data in multiple groups training data includes: belonging to biological information and biological information Predetermined object.
Second acquisition unit 25 is connect with above-mentioned converting unit 23, for obtaining the output knot of authentication network model Fruit, wherein output result includes: the predetermined object of predetermined quantity and biological information to be certified is every in predetermined quantity The probability of a predetermined object.
Determination unit 27 is connect with above-mentioned second acquisition unit 25, for determining biology to be certified according to output result Target object belonging to characteristic information.
In this embodiment it is possible to obtain biological information to be certified first with first acquisition unit 21;Then Biological information to be certified is converted to using converting unit 23 input of authentication network model, wherein identity is recognized Card network model is obtained using multiple groups training data by machine learning training, the trained number of every group in multiple groups training data According to including: predetermined object belonging to biological information and biological information;Second acquisition unit 25 is recycled to obtain body The output result of part certification network model, wherein export the predetermined object and biology spy to be certified that result includes: predetermined quantity Reference breath is the probability of each predetermined object in predetermined quantity;Determination unit 27 is recycled to be determined according to output result to be certified Biological information belonging to target object.Relative in the related technology when carrying out authentication, the authentication mode of use Relatively simple, the lower drawback of the reliability of caused identification, the middle authentication provided fills through the embodiment of the present invention Setting, which may be implemented, is converted to biological information for a variety of biological informations, and merges to biological information, and according to melting Biological information after conjunction carries out the purpose of authentication, has reached the technical effect for improving the reliability of authentication, into And it solves lower for carrying out the reliability of the relatively simple caused identification of the mode of identification in the related technology Technical problem.
As a kind of optional embodiment, which can also include: processing unit, for obtaining wait recognize Before the biological information of card, the biological information of object to be certified is obtained, and pre-process to biological information;Third obtains Unit is believed for obtaining biological characteristic to be certified according to pretreated biological information and the corresponding semaphore of biological information Breath.
In addition, above-mentioned processing unit may include: conversion module, for there are storage class being image in biological information Biological information when, by storage class be image the corresponding image of biological information be converted to the grayscale image to conform to a predetermined condition Picture;First processing module, for, there are when the biological information that storage class is audio, passing through predetermined way pair in biological information Storage class is that the biological information of audio is pre-processed, wherein predetermined way includes at least one of: framing adding window, end Point detection, voice de-noising.
As a kind of optional embodiment, converting unit may include: Second processing module, for in training data Biological attribute data is pre-processed;First determining module, for using pretreated biological attribute data as authentication The input layer of network model;First obtains module, for passing through the processing network layer in authentication network model to defeated The biological attribute data entered is handled, and output node layer is obtained, wherein processing network layer is to remove in authentication network model The network layer of input layer and output layer;Second obtains module, for being trained according to input layer and output node layer, obtains To authentication network model.
As a kind of optional embodiment of the present invention, above-mentioned training data can be the history collected from multiple users A part of data.
In addition, another part in historical data can be used as verify data, wherein verify data is used for authentication Network model is verified.
Also, the full connection of each of multiple full articulamentums in authentication network model in embodiments of the present invention Layer is connected to the same network structure, to merge to every kind of characteristic information in biological information.
As a kind of optional embodiment, which may include: the second determining module, for determining biological characteristic Information is the maximum probability in the probability of each predetermined object in predetermined quantity;Third determining module is used for maximum probability Corresponding predetermined object is as target object belonging to biological information.
Above-mentioned identification authentication system includes first acquisition unit 21, converting unit 23, second acquisition unit 25 and determination Units 27 etc. store in memory as program unit, execute above procedure unit stored in memory by processor To realize corresponding function.
Include kernel in above-mentioned processor, is gone in memory to transfer corresponding program unit by kernel.Kernel can be set One or more determines target object belonging to biological information according to output result by adjusting kernel parameter.
Above-mentioned memory may include the non-volatile memory in computer-readable medium, random access memory (RAM) and/or the forms such as Nonvolatile memory, such as read-only memory (ROM) or flash memory (flash RAM), memory includes extremely A few storage chip.
Another aspect according to an embodiment of the present invention, additionally provides a kind of storage medium, and storage medium includes storage Program, wherein program executes any one of above-mentioned identity identifying method.
Another aspect according to an embodiment of the present invention additionally provides a kind of processor, and processor is used to run program, Wherein, any one of above-mentioned identity identifying method is executed when program is run.
A kind of equipment is additionally provided in embodiments of the present invention, which includes processor, memory and be stored in storage On device and the program that can run on a processor, processor performs the steps of that obtain biology to be certified special when executing program Reference breath;Convert biological information to be certified to the input of authentication network model, wherein authentication network mould Type is obtained using multiple groups training data by machine learning training, and every group of training data in multiple groups training data wraps It includes: predetermined object belonging to biological information and biological information;Obtain authentication network model output as a result, its In, output result includes: the predetermined object of predetermined quantity and biological information to be certified is that each of predetermined quantity is pre- Determine the probability of object;Target object belonging to biological information to be certified is determined according to output result.
A kind of computer program product is additionally provided in embodiments of the present invention, when being executed on data processing equipment, It is adapted for carrying out the program of initialization there are as below methods step: obtaining biological information to be certified;Biology to be certified is special Reference ceases the input for being converted into authentication network model, wherein authentication network model is logical using multiple groups training data Cross what machine learning training obtained, every group of training data in multiple groups training data includes: that biological information and biology are special Predetermined object belonging to reference breath;Obtain the output result of authentication network model, wherein output result includes: predetermined number The predetermined object of amount and biological information to be certified are the probability of each predetermined object in predetermined quantity;It is tied according to output Fruit determines target object belonging to biological information to be certified.
The serial number of the above embodiments of the invention is only for description, does not represent the advantages or disadvantages of the embodiments.
In the above embodiment of the invention, it all emphasizes particularly on different fields to the description of each embodiment, does not have in some embodiment The part of detailed description, reference can be made to the related descriptions of other embodiments.
In several embodiments provided herein, it should be understood that disclosed technology contents can pass through others Mode is realized.Wherein, the apparatus embodiments described above are merely exemplary, such as the division of the unit, Ke Yiwei A kind of logical function partition, there may be another division manner in actual implementation, for example, multiple units or components can combine or Person is desirably integrated into another system, or some features can be ignored or not executed.Another point, shown or discussed is mutual Between coupling, direct-coupling or communication connection can be through some interfaces, the INDIRECT COUPLING or communication link of unit or module It connects, can be electrical or other forms.
The unit as illustrated by the separation member may or may not be physically separated, aobvious as unit The component shown may or may not be physical unit, it can and it is in one place, or may be distributed over multiple On unit.It can some or all of the units may be selected to achieve the purpose of the solution of this embodiment according to the actual needs.
It, can also be in addition, the functional units in various embodiments of the present invention may be integrated into one processing unit It is that each unit physically exists alone, can also be integrated in one unit with two or more units.Above-mentioned integrated list Member both can take the form of hardware realization, can also realize in the form of software functional units.
If the integrated unit is realized in the form of SFU software functional unit and sells or use as independent product When, it can store in a computer readable storage medium.Based on this understanding, technical solution of the present invention is substantially The all or part of the part that contributes to existing technology or the technical solution can be in the form of software products in other words It embodies, which is stored in a storage medium, including some instructions are used so that a computer Equipment (can for personal computer, server or network equipment etc.) execute each embodiment the method for the present invention whole or Part steps.And storage medium above-mentioned includes: that USB flash disk, read-only memory (ROM, Read-Only Memory), arbitrary access are deposited Reservoir (RAM, Random Access Memory), mobile hard disk, magnetic or disk etc. be various to can store program code Medium.
The above is only a preferred embodiment of the present invention, it is noted that for the ordinary skill people of the art For member, various improvements and modifications may be made without departing from the principle of the present invention, these improvements and modifications are also answered It is considered as protection scope of the present invention.

Claims (10)

1. a kind of identity identifying method characterized by comprising
Obtain biological information to be certified;
Convert the biological information to be certified to the input of authentication network model, wherein the authentication Network model is obtained using multiple groups training data by machine learning training, every group of training in the multiple groups training data Data include: predetermined object belonging to biological information and the biological information;
Obtain the output result of the authentication network model, wherein the output result includes: predetermined pair of predetermined quantity As the probability with the biological information to be certified for each predetermined object in the predetermined quantity;
Target object belonging to the biological information to be certified is determined according to the output result.
2. the method according to claim 1, wherein before obtaining the biological information to be certified, Further include:
The biological information of object to be certified is obtained, and the biological information is pre-processed;
The biology to be certified is obtained according to the pretreated biological information and the corresponding semaphore of the biological information Characteristic information.
3. according to the method described in claim 2, it is characterized in that, to the biological information carry out pretreatment include:
In the biological information there are storage class be image biological information when, by storage class be image biological information Corresponding image is converted to the gray level image to conform to a predetermined condition;
In the biological information there are storage class be audio biological information when, by predetermined way to storage class be sound The biological information of frequency is pre-processed, wherein the predetermined way includes at least one of: framing adding window, end-point detection, language Sound noise reduction.
4. the method according to claim 1, wherein obtaining the authentication by training data training Network model includes:
Biological attribute data in the training data is pre-processed;
Using the pretreated biological attribute data as the input layer of the authentication network model;
It is handled, is obtained by the biological attribute data of the processing network layer in the authentication network model to input To output node layer, wherein the processing network layer is to remove the net of input layer and output layer in the authentication network model Network layers;
It is trained according to the input layer and the output node layer, obtains the authentication network model.
5. according to the method described in claim 4, it is characterized in that, the training data is gone through from what multiple users collected A part of history data, another part in the historical data is as verify data, wherein the verify data is used for institute Authentication network model is stated to be verified.
6. the method according to claim 1, wherein determining the biological information according to the output result Affiliated target object includes:
Determine the biological information for the maximum probability in the probability of each predetermined object in the predetermined quantity;
Using the corresponding predetermined object of the maximum probability as target object belonging to the biological information.
7. method according to any one of claim 1 to 6, which is characterized in that in the authentication network model The full articulamentum of each of multiple full articulamentums is connected to the same network structure, to every kind of feature in biological information Information is merged.
8. a kind of identification authentication system characterized by comprising
First acquisition unit, for obtaining biological information to be certified;
Converting unit, for converting the biological information to be certified to the input of authentication network model, wherein The authentication network model is obtained using multiple groups training data by machine learning training, the multiple groups training data In every group of training data include: predetermined object belonging to biological information and the biological information;
Second acquisition unit, for obtaining the output result of the authentication network model, wherein the output result packet Include: the predetermined object of predetermined quantity and the biological information to be certified are each predetermined object in the predetermined quantity Probability;
Determination unit, for determining target object belonging to the biological information according to the output result.
9. a kind of storage medium, which is characterized in that the storage medium includes the program of storage, wherein described program right of execution Benefit require any one of 1 to 7 described in identity identifying method.
10. a kind of processor, which is characterized in that the processor is for running program, wherein right of execution when described program is run Benefit require any one of 1 to 7 described in identity identifying method.
CN201811458913.2A 2018-11-30 2018-11-30 Identity identifying method and device Pending CN109583387A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201811458913.2A CN109583387A (en) 2018-11-30 2018-11-30 Identity identifying method and device

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201811458913.2A CN109583387A (en) 2018-11-30 2018-11-30 Identity identifying method and device

Publications (1)

Publication Number Publication Date
CN109583387A true CN109583387A (en) 2019-04-05

Family

ID=65926635

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201811458913.2A Pending CN109583387A (en) 2018-11-30 2018-11-30 Identity identifying method and device

Country Status (1)

Country Link
CN (1) CN109583387A (en)

Citations (11)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN1304114A (en) * 1999-12-13 2001-07-18 中国科学院自动化研究所 Identity identification method based on multiple biological characteristics
WO2015078018A1 (en) * 2013-11-30 2015-06-04 Xiaoou Tang Method and system for face image recognition
CN105224849A (en) * 2015-10-20 2016-01-06 广州广电运通金融电子股份有限公司 A kind of multi-biological characteristic merges authentication identifying method and device
CN106022380A (en) * 2016-05-25 2016-10-12 中国科学院自动化研究所 Individual identity identification method based on deep learning
CN106203333A (en) * 2016-07-08 2016-12-07 乐视控股(北京)有限公司 Face identification method and system
CN106250857A (en) * 2016-08-04 2016-12-21 深圳先进技术研究院 A kind of identity recognition device and method
CN106909905A (en) * 2017-03-02 2017-06-30 中科视拓(北京)科技有限公司 A kind of multi-modal face identification method based on deep learning
CN108021864A (en) * 2017-11-02 2018-05-11 平安科技(深圳)有限公司 Character personality analysis method, device and storage medium
CN108319930A (en) * 2018-03-09 2018-07-24 百度在线网络技术(北京)有限公司 Identity identifying method, system, terminal and computer readable storage medium
CN108376215A (en) * 2018-01-12 2018-08-07 上海大学 A kind of identity identifying method
CN108429619A (en) * 2018-01-18 2018-08-21 北京捷通华声科技股份有限公司 Identity identifying method and system

Patent Citations (12)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN1304114A (en) * 1999-12-13 2001-07-18 中国科学院自动化研究所 Identity identification method based on multiple biological characteristics
WO2015078018A1 (en) * 2013-11-30 2015-06-04 Xiaoou Tang Method and system for face image recognition
CN105849747A (en) * 2013-11-30 2016-08-10 北京市商汤科技开发有限公司 Method and system for face image recognition
CN105224849A (en) * 2015-10-20 2016-01-06 广州广电运通金融电子股份有限公司 A kind of multi-biological characteristic merges authentication identifying method and device
CN106022380A (en) * 2016-05-25 2016-10-12 中国科学院自动化研究所 Individual identity identification method based on deep learning
CN106203333A (en) * 2016-07-08 2016-12-07 乐视控股(北京)有限公司 Face identification method and system
CN106250857A (en) * 2016-08-04 2016-12-21 深圳先进技术研究院 A kind of identity recognition device and method
CN106909905A (en) * 2017-03-02 2017-06-30 中科视拓(北京)科技有限公司 A kind of multi-modal face identification method based on deep learning
CN108021864A (en) * 2017-11-02 2018-05-11 平安科技(深圳)有限公司 Character personality analysis method, device and storage medium
CN108376215A (en) * 2018-01-12 2018-08-07 上海大学 A kind of identity identifying method
CN108429619A (en) * 2018-01-18 2018-08-21 北京捷通华声科技股份有限公司 Identity identifying method and system
CN108319930A (en) * 2018-03-09 2018-07-24 百度在线网络技术(北京)有限公司 Identity identifying method, system, terminal and computer readable storage medium

Similar Documents

Publication Publication Date Title
CN108776788B (en) Brain wave-based identification method
Yuan et al. Fingerprint liveness detection using an improved CNN with image scale equalization
De Souza et al. Deep texture features for robust face spoofing detection
CN107633207B (en) AU characteristic recognition methods, device and storage medium
Regouid et al. Multimodal biometric system for ECG, ear and iris recognition based on local descriptors
Ashby et al. Low-cost electroencephalogram (EEG) based authentication
CN108429619A (en) Identity identifying method and system
WO2019152983A2 (en) System and apparatus for face anti-spoofing via auxiliary supervision
CN106803098A (en) A kind of three mode emotion identification methods based on voice, expression and attitude
CN109190566A (en) A kind of fusion local code and CNN model finger vein identification method
CN106682473A (en) Method and device for identifying identity information of users
Gui et al. A residual feature-based replay attack detection approach for brainprint biometric systems
CN111178130A (en) Face recognition method, system and readable storage medium based on deep learning
CN108664909A (en) A kind of auth method and terminal
Wysoski et al. Adaptive learning procedure for a network of spiking neurons and visual pattern recognition
CN110364163A (en) The identity identifying method that a kind of voice and lip reading blend
CN110363111A (en) Human face in-vivo detection method, device and storage medium based on lens distortions principle
Kassem et al. An enhanced ATM security system using multimodal biometric strategy
Sharma et al. Sign language gesture recognition
Yuan et al. Fingerprint liveness detection using an improved CNN with the spatial pyramid pooling structure
CN109583387A (en) Identity identifying method and device
Agbinya et al. Design and implementation of multimodal digital identity management system using fingerprint matching and face recognition
CN111062345A (en) Training method and device of vein recognition model and vein image recognition device
CN110135229A (en) A kind of driver identity identifying system using neural network
CN109034059A (en) Silent formula human face in-vivo detection method, device, storage medium and processor

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination
RJ01 Rejection of invention patent application after publication

Application publication date: 20190405

RJ01 Rejection of invention patent application after publication