CN105245497B - A kind of identity identifying method and device - Google Patents
A kind of identity identifying method and device Download PDFInfo
- Publication number
- CN105245497B CN105245497B CN201510542515.9A CN201510542515A CN105245497B CN 105245497 B CN105245497 B CN 105245497B CN 201510542515 A CN201510542515 A CN 201510542515A CN 105245497 B CN105245497 B CN 105245497B
- Authority
- CN
- China
- Prior art keywords
- voice
- signal
- speech
- authentication
- reliable
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
- 238000000034 method Methods 0.000 title claims abstract description 30
- 239000000284 extract Substances 0.000 claims abstract description 12
- 238000000605 extraction Methods 0.000 claims description 16
- 241000283073 Equus caballus Species 0.000 claims description 6
- 230000006870 function Effects 0.000 description 17
- 230000005236 sound signal Effects 0.000 description 16
- 230000008569 process Effects 0.000 description 11
- 238000012545 processing Methods 0.000 description 11
- 108091006146 Channels Proteins 0.000 description 8
- 230000011218 segmentation Effects 0.000 description 8
- 238000005516 engineering process Methods 0.000 description 7
- 238000004590 computer program Methods 0.000 description 6
- 238000010586 diagram Methods 0.000 description 6
- 230000004048 modification Effects 0.000 description 5
- 238000012986 modification Methods 0.000 description 5
- 238000004422 calculation algorithm Methods 0.000 description 4
- 238000004364 calculation method Methods 0.000 description 4
- 239000011159 matrix material Substances 0.000 description 4
- 230000008859 change Effects 0.000 description 3
- NAWXUBYGYWOOIX-SFHVURJKSA-N (2s)-2-[[4-[2-(2,4-diaminoquinazolin-6-yl)ethyl]benzoyl]amino]-4-methylidenepentanedioic acid Chemical compound C1=CC2=NC(N)=NC(N)=C2C=C1CCC1=CC=C(C(=O)N[C@@H](CC(=C)C(O)=O)C(O)=O)C=C1 NAWXUBYGYWOOIX-SFHVURJKSA-N 0.000 description 2
- 238000013459 approach Methods 0.000 description 2
- 238000011161 development Methods 0.000 description 2
- 230000018109 developmental process Effects 0.000 description 2
- 230000007613 environmental effect Effects 0.000 description 2
- 238000009432 framing Methods 0.000 description 2
- 238000009499 grossing Methods 0.000 description 2
- 239000000203 mixture Substances 0.000 description 2
- 238000012544 monitoring process Methods 0.000 description 2
- 230000000717 retained effect Effects 0.000 description 2
- 241000894007 species Species 0.000 description 2
- 238000001228 spectrum Methods 0.000 description 2
- 238000012549 training Methods 0.000 description 2
- 230000009466 transformation Effects 0.000 description 2
- 238000004458 analytical method Methods 0.000 description 1
- 230000003542 behavioural effect Effects 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 210000000056 organ Anatomy 0.000 description 1
- 230000008560 physiological behavior Effects 0.000 description 1
- 230000001766 physiological effect Effects 0.000 description 1
- 230000009467 reduction Effects 0.000 description 1
- 210000001525 retina Anatomy 0.000 description 1
- 238000005070 sampling Methods 0.000 description 1
- 230000007474 system interaction Effects 0.000 description 1
Classifications
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L63/00—Network architectures or network communication protocols for network security
- H04L63/08—Network architectures or network communication protocols for network security for authentication of entities
- H04L63/0861—Network architectures or network communication protocols for network security for authentication of entities using biometrical features, e.g. fingerprint, retina-scan
-
- G—PHYSICS
- G10—MUSICAL INSTRUMENTS; ACOUSTICS
- G10L—SPEECH ANALYSIS TECHNIQUES OR SPEECH SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING TECHNIQUES; SPEECH OR AUDIO CODING OR DECODING
- G10L17/00—Speaker identification or verification techniques
- G10L17/02—Preprocessing operations, e.g. segment selection; Pattern representation or modelling, e.g. based on linear discriminant analysis [LDA] or principal components; Feature selection or extraction
-
- G—PHYSICS
- G10—MUSICAL INSTRUMENTS; ACOUSTICS
- G10L—SPEECH ANALYSIS TECHNIQUES OR SPEECH SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING TECHNIQUES; SPEECH OR AUDIO CODING OR DECODING
- G10L17/00—Speaker identification or verification techniques
- G10L17/16—Hidden Markov models [HMM]
Landscapes
- Engineering & Computer Science (AREA)
- Health & Medical Sciences (AREA)
- Multimedia (AREA)
- Audiology, Speech & Language Pathology (AREA)
- Human Computer Interaction (AREA)
- Physics & Mathematics (AREA)
- Acoustics & Sound (AREA)
- Biomedical Technology (AREA)
- Computer Vision & Pattern Recognition (AREA)
- General Health & Medical Sciences (AREA)
- Computer Hardware Design (AREA)
- Computer Security & Cryptography (AREA)
- Computing Systems (AREA)
- General Engineering & Computer Science (AREA)
- Computer Networks & Wireless Communication (AREA)
- Signal Processing (AREA)
- Telephonic Communication Services (AREA)
Abstract
In order to accurately carry out authentication, the application provides a kind of identity identifying method and device, the method includes the steps: step 1 establishes voice authentication database；Step 2 receives and stores voice document to be certified；Step 3 identifies the true and false of voice document to be certified, if true, then enter step four, if there is vacation, terminates direct authentication output failed message；Step 4 extracts phonic signal character and rebuilds the feature in voice signal；Step 5 authenticates the reliable characteristic vector of voice and exports identity authentication result.
Description
Technical field
This application involves technical field of biometric identification more particularly to a kind of identity identifying methods and device.
Background technique
With the rapid development of information technology, the especially development of Internet, data information deepens continuously.It is more next
More affairs, the process of handling require to carry out authentication, such as: intelligent entrance guard, intelligence are used in public safety field
It can video monitoring, public security are deployed to ensure effective monitoring and control of illegal activities, customs's authentication, practical driving license are verified etc. intelligent robot；In civil and economic field
To all kinds of bank cards, fiscard, credit card, save card holder carry out authentication intelligent robot.For information peace
Entirely, it is usually required before transacting business by after verifying personnel identity.
Biological identification technology is exactly a kind of technology that authentication is carried out using human body biological characteristics.Biological characteristic is unique
, can measure or can automatic identification and verifying physiological property or behavior.Biological recognition system carries out biological characteristic
Sampling, extracts its unique feature and is converted into digital code, and these codes are further formed feature templates, Ren Mentong
When identifying system interaction carries out authentication, identifying system obtains its feature and compares in feature templates, to determine whether
Matching.
Biological identification technology is presently the most convenient and safely identifies technology, does not need to remember complicated password, also not
Need to carry the thing of key, smart card etc.The feature of common bio-identification has fingerprint, iris, retina, DNA
With voice etc..The voice of natural quality one of of the language as the mankind, speaker has respective biological characteristic, everyone hair
Sound organ not only has inborn differences of Physiological, also has posteriori behavioral difference, this makes more and more identity identifying technologies
It is middle that speaker's identity identification is carried out using speech analysis.
But the voice document of people is also easy to be replicated or steal, if the voice document of people to be verified is by fraudulent copying
And be used to handle the business agreed to without party, thus bring very big risk.
In addition, identify the identity of personnel by voice, when receiving voice input, ambient noise will affect for
The accuracy that voice identifies.How the environmental noise of influence to(for) speech recognition be also urgently to be resolved one is solved the problems, such as.
Summary of the invention
In view of this, the application provides one kind for identity identifying method and device, it is avoided that fraudulent copying and reduction
The problem of recognition accuracy, realization with high-accuracy carry out personnel identity identification.
The application provides a kind of identity identifying method, which comprises
Step 1 establishes voice authentication database；
Step 2 receives and stores voice document to be certified；
Step 3 identifies the true and false of voice document to be certified, if true, then enters step four, if there is vacation, terminates straight
Connect authentication output failed message；
Step 4 extracts phonic signal character and rebuilds the reliable speech feature in voice signal；
Step 5 authenticates the reliable characteristic vector of voice and exports identity authentication result.
In one specific embodiment of the application, the step 1 includes: the proprietary reliable speech of acquisition, reliable speech
Feature extraction and voice characteristics information is recorded in the voice authentication database.
In one specific embodiment of the application, the step 3 are as follows: since duplication bring is specific in identification voice document
Feature whether be more than setting first threshold.
In one specific embodiment of the application, calculating the duplication bring special characteristic includes:
If X={ X1[n],…,XT[n] } indicate that frame number is the voice signal of T, then q (1≤q≤T) frame signal Xq[n](0
≤ n≤N-1) discrete Fourier transform are as follows:
Wherein, N is the status number of equine husband's chain in voice signal；
The then expression formula of mean value frame are as follows:
Original real speech and the copying voice difference existing for frequency-portions is extracted by formula (3-3):
Wherein, filter () is arbitrary filter function in the prior art.
In one specific embodiment of the application, the step 4 includes:
The speech feature extraction algorithm used when being established using speech recognition library extracts the voice letter inputted in daily life
The feature vector, X of number λ；
The feature vector, X of the voice signal λ inputted in daily life is separated, speech vector X is divided intorWith noise to
Measure Xu；
Feature vector, X is divided into speech vector X according to the mean value and variance of its Gaussian functionrWith noise vector Xu, and calculate
V-th of given speech vector XrPrior probability P (v | Xr):
By the speech vector X of voice signalrReliable characteristic vector as input speech signal.
Disclosed herein as well is a kind of identification authentication system, described device includes:
Voice authentication database 1, for storing the voice characteristics information of all personnel；
Acquisition module 2 acquires the voice messaging of personnel to be certified；
Authenticity module 3, for identifying the true and false of voice document to be certified；
Characteristic extracting module 4 extracts phonic signal character and rebuilds the feature in voice signal；
Authentication module 5 authenticates the reliable characteristic vector of voice and exports identity authentication result.
In one specific embodiment of the application, the voice authentication database 1 for acquire proprietary reliable speech, can
Voice characteristics information is recorded by the feature extraction of voice and in the voice authentication database.
In one specific embodiment of the application, the authenticity module 3 is specifically used in identification voice document due to multiple
Bring special characteristic processed whether be more than setting first threshold.
In one specific embodiment of the application, calculating the duplication bring special characteristic includes:
If X={ X1[n],…,XT[n] } indicate that frame number is the voice signal of T, then q (1≤q≤T) frame signal Xq[n](0
≤ n≤N-1) discrete Fourier transform are as follows:
Wherein, N is the status number of equine husband's chain in voice signal；
The then expression formula of mean value frame are as follows:
Original real speech and the copying voice difference existing for frequency-portions is extracted by formula (3-3):
Wherein, filter () is arbitrary filter function in the prior art.
In one specific embodiment of the application, the authentication module 5 includes:
The speech feature extraction algorithm used when being established using speech recognition library extracts the voice letter inputted in daily life
The feature vector, X of number λ；
The feature vector, X of the voice signal λ inputted in daily life is separated, speech vector X is divided intorWith noise to
Measure Xu；
Feature vector, X is divided into speech vector X according to the mean value and variance of its Gaussian functionrWith noise vector Xu, and calculate
V-th of given speech vector XrPrior probability P (v | Xr):
By the speech vector X of voice signalrReliable characteristic vector as input speech signal.
Detailed description of the invention
In order to illustrate the technical solutions in the embodiments of the present application or in the prior art more clearly, to embodiment or will show below
There is attached drawing needed in technical description to be briefly described, it should be apparent that, the accompanying drawings in the following description is only this
The some embodiments recorded in application can also be obtained according to these attached drawings other for those of ordinary skill in the art
Attached drawing.
Fig. 1 is the flow chart that voice authentication database is established in the application；
Identity identifying method flow chart is in Fig. 2 the application；
Fig. 3 is identification authentication system structure chart in the application.
Specific embodiment
In order to make those skilled in the art more fully understand the technical solution in the application, below in conjunction with the embodiment of the present application
In attached drawing, the technical scheme in the embodiment of the application is clearly and completely described, it is clear that described embodiment is only
It is some embodiments of the present application, instead of all the embodiments.Based on the embodiment in the application, ordinary skill people
Member's every other embodiment obtained, shall fall within the protection scope of the present application.
In order to solve existing in the prior art use in voice progress authentication due to fraudulent copying and environmental noise
So that the problem of speech recognition accuracy declines, this application discloses a kind of identity identifying method and devices.
Further illustrate that the application implements below with reference to illustrations.
As shown in Figure 1, firstly the need of establishing voice authentication database, the foundation of speech database before carrying out voice authentication
Journey can use any existing speech feature extraction technology, it is preferable that following set-up process can be used:
As needed, all n voice messagings for needing to authenticate personnel are acquired.In order to guarantee the accuracy rate of speech recognition,
It can be continuously increased the voice messaging of all personnel in varied situations, to improve discrimination.
Audio signal x (i) is corresponded to for the voice messaging of collected personnel i and stores the original into speech recognition database
Beginning phonetic storage area, i=1 ... ..., n (i is positive integer).
The characteristic extraction procedure of voice the following steps are included:
The each audio signal x (i) stored in raw tone memory block is carried out the following processing:
(1) audio signal x (i) is divided into a series of continuous frames, Fourier transformation is done to every frame signal.
(2) audio signal is handled using filter, to reduce the mutual leakage of spectrum energy between nearby frequency bands；
Filter function used in filter are as follows:
Filter (t)=Bntn-1e-2πBtcos(2πf0T+ θ) u (t)) formula (1)
Wherein:
Parameter θ is the initial phase of filter, and n is the order of filter；
As t<0, u (t)=0, as t>0, u (t)=1；
B=1.019*ERB (f0), ERB (f0) be filter Equivalent Rectangular Bandwidth, it is the same as filter centre frequency f0's
Relationship are as follows:
ERB(f0)=24.7+0.108f0Formula (2).
(3) the intermediate deviation of audio signal is removed.
After audio signal framing, a certain number of frames are formed a segmentation, in the present invention preferably by 7 frame compositions one
A segmentation, this can be configured according to the processing capacity of system.
The frame length that most of speech recognition systems use be 20ms-30ms, present invention preferably uses 26.5ms as sea
Bright window, a length of 10ms of overlapping frame, the intermediate quantity Q (i, j) of every frame are obtained by calculating the average value of frame energy P (i, j) in section:
Since the present invention preferably 7 frames form a segmentation, thus M=3 in formula (3).I is channel number, and j is required
The sequence of frame, j ' are the sequence of each frame in required segmentation.
In noise energy removal process, it can be indicated using the ratio (AM/GM) of arithmetic mean of instantaneous value and geometrical mean
The degree that voice signal is corroded obtains after seeking logarithm to above-mentioned ratio:
Z is floor coefficient in formula (4), guarantees that the deviation of calculated result is allowing with to avoid bearing infinitesimal valuation
In range；J is the sequence sum of frame.
Assuming that B (i) is the deviation as caused by ambient noise, i indicates channel sequence, and by conditional probability, that thing is obtained, and removes
Intermediate quantity Q ' (i, j | B (i)) after deviation are as follows:
Q'(i, j | B (i))=max (Q (i, j)-B (i), 10-3Q (i, j)) formula (5)
It is available:
It, can be in the hope of when AM/GM value of the ratio of AM/GM under noise situations closest to acoustic signal for formula (6)
The estimated value of B (i) are as follows:
B'(i)=min B (i) | and G'(i | B (i)) >=Gc(i) } formula (7)
Wherein, Gc(i) G (i) respective value in acoustic signal is indicated, it is right to being obtained after each channel calculation formula (7)
In each time-frequency BIN signal (i, j), the ratio of noise removal are as follows:
For smoothing computation, the noise removal ratio of channel i-N to i+N is averaged, final function after adjustment are as follows:
Audio signals all in filter are handled using formula (10), as filter after the intermediate deviation of removal
Output.
(4) non-linear power-function arithmetic, used power function are done to the audio signal data of all filters output are as follows:
Y=X0.1Formula (11).
(5) speech characteristic parameter is obtained after further doing discrete cosine transform to the output of (4) step.
As discrete cosine transform (DCT) be speech processes field well known to processing mode, details are not described herein.
In the database by the phonetic feature being calculated storage.
As shown in Fig. 2, this application discloses a kind of identity identifying methods comprising following steps:
Step1: voice authentication database is established.
Arbitrary speech feature extraction technology in the prior art can be used and establish voice authentication database, can also lead to
It crosses preferred embodiment described above and establishes voice authentication database.
Step2: voice document to be certified is received and stored.
Speech prompt information can be set in voice capture device, the personnel of identity to be identified is prompted to input voice text
Part.For example, passing through the voice of microphone collector.Other voice capture devices can also be used.
Step3: identifying the true and false of voice document to be certified, if true, then enters Step4, if there is vacation, terminates directly
Authentication output failed message.
In general, what the approach that false voice document is usually agreed to without party was obtained by the method replicated, but
It is the copying voice file by way of multiple copies, will necessarily changes the characteristic information in voice document, and this change
Usually equably exist along with the signal in entire voice document.Thus, by identifying in voice document due to dubs
Whether the special characteristic come is more than the first threshold of setting to carry out authenticity.
Firstly, setting X={ X1[n],…,XT[n] } indicate that frame number is the voice signal of T, then q (1≤q≤T) frame signal Xq
The discrete Fourier transform of [n] (0≤n≤N-1) (status number that N is equine husband's chain in voice signal) are as follows:
The then expression formula of mean value frame are as follows:
In general, original real speech and copying voice has differences in frequency-portions, can be mentioned by formula (3-3)
Take this species diversity:
Wherein, filter () is the filter function in arbitrary filter function, such as formula (1) in the prior art.
In this application, through overtesting, preferably first threshold=0.53.
Step4: it extracts phonic signal character and rebuilds the feature in voice signal.
The voice signal usually used when constructing voice authentication database is usually that profession is carried out under quiet environment
Acquisition, and be usually may exist various during actual authentication, when voice inputs in daily living environment
Noise, if directly carrying out feature extraction to the voice signal inputted under noise conditions will receive the influence of noise speech information,
And then influence the accuracy rate of authentication.
The speech feature extraction algorithm used when being established using speech recognition library extracts the voice letter inputted in daily life
The feature vector, X of number λ.Therefore, it is possible to which the feature vector, X of the voice signal λ inputted in daily life is separated, it is divided into language
Sound vector XrWith noise vector Xu。
The prior probability p (X) of voice signal is established into model, then is obtained by merging, training data:
Wherein, V is the quantity of mixed cell, and v is serial number, and p (v) is that the prior probability P (X | v) an of mixed cell is indicated
V-th of Gaussian Profile, skilled person will appreciate that its Mean Matrix is μv, diagonal covariance matrix σv 2.If one given
The feature of voice signal is just divided into speech vector X according to the mean value of its Gaussian function and variancerWith noise vector Xu, and then calculate
V-th of given speech vector X outrPrior probability P (v | Xr):
In reconstruction process, the speech vector X of voice signalrIt is retained the reliable characteristic as input speech signal
Vector.
Step5: the reliable characteristic vector of voice is authenticated and exports identity authentication result.
The reliable characteristic vector input voice authentication database of voice signal is compared, if in voice authentication database
It finds the voice signal being consistent and then passes through verifying, tested if not passing through if voice authentication database does not find the voice signal being consistent
Card.
As shown in figure 3, present invention also provides a kind of identification authentication systems comprising:
Voice authentication database 1, for storing the voice characteristics information of all personnel.
Arbitrary speech feature extraction technology in the prior art can be used and establish voice authentication database, can also lead to
It crosses preferred embodiment described below and establishes voice authentication database.
As shown in Figure 1, firstly the need of establishing voice authentication database, the foundation of speech database before carrying out voice authentication
Journey can use any existing speech feature extraction technology, it is preferable that following set-up process can be used:
As needed, all n voice messagings for needing to authenticate personnel are acquired.In order to guarantee the accuracy rate of speech recognition,
It can be continuously increased the voice messaging of all personnel in varied situations, to improve discrimination.
Audio signal x (i) is corresponded to for the voice messaging of collected personnel i and stores the original into speech recognition database
Beginning phonetic storage area, i=1 ... ..., n (i is positive integer).
The characteristic extraction procedure of voice the following steps are included:
The each audio signal x (i) stored in raw tone memory block is carried out the following processing:
(1) audio signal x (i) is divided into a series of continuous frames, Fourier transformation is done to every frame signal.
(2) audio signal is handled using filter, to reduce the mutual leakage of spectrum energy between nearby frequency bands；
Filter function used in filter are as follows:
Filter (t)=Bntn-1e-2πBtcos(2πf0T+ θ) u (t)) formula (1)
Wherein:
Parameter θ is the initial phase of filter, and n is the order of filter；
As t<0, u (t)=0, as t>0, u (t)=1；
B=1.019*ERB (f0), ERB (f0) be filter Equivalent Rectangular Bandwidth, it is the same as filter centre frequency f0's
Relationship are as follows:
ERB(f0)=24.7+0.108f0Formula (2).
(3) the intermediate deviation of audio signal is removed.
After audio signal framing, a certain number of frames are formed a segmentation, in the present invention preferably by 7 frame compositions one
A segmentation, this can be configured according to the processing capacity of system.
The frame length that most of speech recognition systems use be 20ms-30ms, present invention preferably uses 26.5ms as sea
Bright window, a length of 10ms of overlapping frame, the intermediate quantity Q (i, j) of every frame are obtained by calculating the average value of frame energy P (i, j) in section:
Since the present invention preferably 7 frames form a segmentation, thus M=3 in formula (3).I is channel number, and j is required
The sequence of frame, j ' are the sequence of each frame in required segmentation.
In noise energy removal process, it can be indicated using the ratio (AM/GM) of arithmetic mean of instantaneous value and geometrical mean
The degree that voice signal is corroded obtains after seeking logarithm to above-mentioned ratio:
Z is floor coefficient in formula (4), guarantees that the deviation of calculated result is allowing with to avoid bearing infinitesimal valuation
In range；J is the sequence sum of frame.
Assuming that B (i) is the deviation as caused by ambient noise, i indicates channel sequence, and by conditional probability, that thing is obtained, and removes
Intermediate quantity Q ' (i, j | B (i)) after deviation are as follows:
Q'(i, j | B (i))=max (Q (i, j)-B (i), 10-3Q (i, j)) formula (5)
It is available:
It, can be in the hope of when AM/GM value of the ratio of AM/GM under noise situations closest to acoustic signal for formula (6)
The estimated value of B (i) are as follows:
B'(i)=min B (i) | and G'(i | B (i)) >=Gc(i) } formula (7)
Wherein, Gc(i) G (i) respective value in acoustic signal is indicated, it is right to being obtained after each channel calculation formula (7)
In each time-frequency BIN signal (i, j), the ratio of noise removal are as follows:
For smoothing computation, the noise removal ratio of channel i-N to i+N is averaged, final function after adjustment are as follows:
Audio signals all in filter are handled using formula (10), as filter after the intermediate deviation of removal
Output.
(4) non-linear power-function arithmetic, used power function are done to the audio signal data of all filters output are as follows:
Y=X0.1Formula (11).
(5) speech characteristic parameter is obtained after further doing discrete cosine transform to the output of (4) step.
As discrete cosine transform (DCT) be speech processes field well known to processing mode, details are not described herein.
In the database by the phonetic feature being calculated storage.
Acquisition module 2 acquires the voice messaging of personnel to be certified.
Speech prompt information can be set in voice capture device, the personnel of identity to be identified is prompted to input voice text
Part.For example, passing through the voice of microphone collector.Other voice capture devices can also be used.
Authenticity module 3, for identifying the true and false of voice document to be certified.
In general, what the approach that false voice document is usually agreed to without party was obtained by the method replicated, but
It is the copying voice file by way of multiple copies, will necessarily changes the characteristic information in voice document, and this change
Usually equably exist along with the signal in entire voice document.Thus, by identifying in voice document due to dubs
Whether the special characteristic come is more than the first threshold of setting to carry out authenticity.
Firstly, setting X={ X1[n],…,XT[n] } indicate that frame number is the voice signal of T, then q (1≤q≤T) frame signal Xq
The discrete Fourier transform of [n] (0≤n≤N-1) (status number that N is equine husband's chain in voice signal) are as follows:
The then expression formula of mean value frame are as follows:
In general, original real speech and copying voice has differences in frequency-portions, can be mentioned by formula (3-3)
Take this species diversity:
Wherein, filter () is the filter function in arbitrary filter function, such as formula (1) in the prior art.
In this application, through overtesting, preferably first threshold=0.53.
Characteristic extracting module 4 extracts phonic signal character and rebuilds the feature in voice signal.
The voice signal usually used when constructing voice authentication database is usually that profession is carried out under quiet environment
Acquisition, and be usually may exist various during actual authentication, when voice inputs in daily living environment
Noise, if directly carrying out feature extraction to the voice signal inputted under noise conditions will receive the influence of noise speech information,
And then influence the accuracy rate of authentication.
The speech feature extraction algorithm used when being established using speech recognition library extracts the voice letter inputted in daily life
The feature vector, X of number λ.Therefore, it is possible to which the feature vector, X of the voice signal λ inputted in daily life is separated, it is divided into language
Sound vector XrWith noise vector Xu。
The prior probability p (X) of voice signal is established into model, then is obtained by merging, training data:
Wherein, V is the quantity of mixed cell, and v is serial number, and p (v) is that the prior probability P (X | v) an of mixed cell is indicated
V-th of Gaussian Profile, skilled person will appreciate that its Mean Matrix is μv, diagonal covariance matrix isIf one given
The feature of voice signal is just divided into speech vector X according to the mean value of its Gaussian function and variancerWith noise vector Xu, and then calculate
V-th of given speech vector X outrPrior probability P (v | Xr):
In reconstruction process, the speech vector X of voice signalrIt is retained the reliable characteristic as input speech signal
Vector.
Authentication module 5 authenticates the reliable characteristic vector of voice and exports identity authentication result.
The reliable characteristic vector input voice authentication database of voice signal is compared, if in voice authentication database
It finds the voice signal being consistent and then passes through verifying, tested if not passing through if voice authentication database does not find the voice signal being consistent
Card.
Certainly, any technical solution for implementing the application must be not necessarily required to reach simultaneously above all advantages.
It will be understood by those skilled in the art that embodiments herein can provide as method, apparatus (equipment) or computer
Program product.Therefore, in terms of the application can be used complete hardware embodiment, complete software embodiment or combine software and hardware
Embodiment form.Moreover, it wherein includes the meter of computer usable program code that the application, which can be used in one or more,
The computer journey implemented in calculation machine usable storage medium (including but not limited to magnetic disk storage, CD-ROM, optical memory etc.)
The form of sequence product.
The application is flow chart of the reference according to method, apparatus (equipment) and computer program product of the embodiment of the present application
And/or block diagram describes.It should be understood that each process in flowchart and/or the block diagram can be realized by computer program instructions
And/or the combination of the process and/or box in box and flowchart and/or the block diagram.It can provide these computer programs to refer to
Enable the processor of general purpose computer, special purpose computer, Embedded Processor or other programmable data processing devices to generate
One machine so that by the instruction that the processor of computer or other programmable data processing devices executes generate for realizing
The device for the function of being specified in one or more flows of the flowchart and/or one or more blocks of the block diagram.
These computer program instructions, which may also be stored in, is able to guide computer or other programmable data processing devices with spy
Determine in the computer-readable memory that mode works, so that it includes referring to that instruction stored in the computer readable memory, which generates,
Enable the manufacture of device, the command device realize in one box of one or more flows of the flowchart and/or block diagram or
The function of being specified in multiple boxes.
These computer program instructions also can be loaded onto a computer or other programmable data processing device, so that counting
Series of operation steps are executed on calculation machine or other programmable devices to generate computer implemented processing, thus in computer or
The instruction executed on other programmable devices is provided for realizing in one or more flows of the flowchart and/or block diagram one
The step of function of being specified in a box or multiple boxes.
Although the preferred embodiment of the application has been described, it is created once a person skilled in the art knows basic
Property concept, then additional changes and modifications may be made to these embodiments.So it includes excellent that the following claims are intended to be interpreted as
It selects embodiment and falls into all change and modification of the application range.Obviously, those skilled in the art can be to the application
Various modification and variations are carried out without departing from spirit and scope.If in this way, these modifications and variations of the application
Belong within the scope of the claim of this application and its equivalent technologies, then the application is also intended to encompass these modification and variations and exists
It is interior.
Claims (4)
1. a kind of identity identifying method, which comprises
Step 1 establishes voice authentication database；
Step 2 receives and stores voice document to be certified；
Step 3 identifies the true and false of voice document to be certified, if true, then enters step four, if there is vacation, terminates directly defeated
Authentification failure message out；
Step 4 extracts phonic signal character and rebuilds the reliable characteristic vector of the voice in voice signal；
Step 5 authenticates the reliable characteristic vector of voice and exports identity authentication result；
The step 3 are as follows: identification voice document in due to duplication bring special characteristic whether be more than setting first threshold；
If being less than the first threshold, voice document be it is true, it is on the contrary then be vacation；
Calculating the duplication bring special characteristic includes:
If X={ X1[n],...,XT[n] } indicate that frame number is the voice signal of T, then q (1≤q≤T) frame signal Xq[n](0≤n
≤ N-1) discrete Fourier transform are as follows:
Wherein, N is the status number of equine husband's chain in voice signal；
The then expression formula of mean value frame are as follows:
Original real speech and the copying voice difference existing for frequency-portions is extracted by formula (3-3):
Wherein, filter () is filter function in the prior art.
2. the method according to claim 1, wherein the step 1 include: the proprietary reliable speech of acquisition,
The feature extraction of reliable speech and voice characteristics information is recorded in the voice authentication database.
3. a kind of identification authentication system, described device include:
Voice authentication database, for storing the voice characteristics information of all personnel；
Acquisition module acquires the voice messaging of personnel to be certified；
Authenticity module, for identifying the true and false of voice document to be certified；
Characteristic extracting module extracts phonic signal character and rebuilds the reliable characteristic vector in voice signal；
Authentication module authenticates the reliable characteristic vector in voice signal and exports identity authentication result；
Whether the authenticity module is specifically used in identification voice document due to duplication bring special characteristic being more than setting
First threshold；If being less than the first threshold, voice document be it is true, it is on the contrary then be vacation；
Calculating the duplication bring special characteristic includes:
If X={ X1[n],...,XT[n] } indicate that frame number is the voice signal of T, then q (1≤q≤T) frame signal Xq[n](0≤n
≤ N-1) discrete Fourier transform are as follows:
Wherein, N is the status number of equine husband's chain in voice signal；
The then expression formula of mean value frame are as follows:
Original real speech and the copying voice difference existing for frequency-portions is extracted by formula (3-3):
Wherein, filter () is filter function in the prior art.
4. device according to claim 3, which is characterized in that the voice authentication database for acquire it is proprietary can
Voice characteristics information is recorded by voice, the feature extraction of reliable speech and in the voice authentication database.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510542515.9A CN105245497B (en) | 2015-08-31 | 2015-08-31 | A kind of identity identifying method and device |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510542515.9A CN105245497B (en) | 2015-08-31 | 2015-08-31 | A kind of identity identifying method and device |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN105245497A CN105245497A (en) | 2016-01-13 |
| CN105245497B true CN105245497B (en) | 2019-01-04 |
Family
ID=55042997
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201510542515.9A Active CN105245497B (en) | 2015-08-31 | 2015-08-31 | A kind of identity identifying method and device |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN105245497B (en) |
Families Citing this family (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109215643B (en) * | 2017-07-05 | 2023-10-24 | 阿里巴巴集团控股有限公司 | Interaction method, electronic equipment and server |
| CN107886956B (en) * | 2017-11-13 | 2020-12-11 | 广州酷狗计算机科技有限公司 | Audio recognition method and device and computer storage medium |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US5054083A (en) * | 1989-05-09 | 1991-10-01 | Texas Instruments Incorporated | Voice verification circuit for validating the identity of an unknown person |
| CN1397929A (en) * | 2002-07-12 | 2003-02-19 | 清华大学 | Speech intensifying-characteristic weighing-logrithmic spectrum addition method for anti-noise speech recognization |
| CN1627365A (en) * | 2003-12-09 | 2005-06-15 | 摩托罗拉公司 | Method and system adaptive to speaker without vetating to speech identifying database |
| CN102324232A (en) * | 2011-09-12 | 2012-01-18 | 辽宁工业大学 | Method for recognizing sound-groove and system based on gauss hybrid models |
| CN103236260A (en) * | 2013-03-29 | 2013-08-07 | 京东方科技集团股份有限公司 | Voice recognition system |
| CN103730114A (en) * | 2013-12-31 | 2014-04-16 | 上海交通大学无锡研究院 | Mobile equipment voiceprint recognition method based on joint factor analysis model |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP6109927B2 (en) * | 2012-05-04 | 2017-04-05 | カオニックス ラブス リミテッド ライアビリティ カンパニー | System and method for source signal separation |
-
2015
- 2015-08-31 CN CN201510542515.9A patent/CN105245497B/en active Active
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US5054083A (en) * | 1989-05-09 | 1991-10-01 | Texas Instruments Incorporated | Voice verification circuit for validating the identity of an unknown person |
| CN1397929A (en) * | 2002-07-12 | 2003-02-19 | 清华大学 | Speech intensifying-characteristic weighing-logrithmic spectrum addition method for anti-noise speech recognization |
| CN1627365A (en) * | 2003-12-09 | 2005-06-15 | 摩托罗拉公司 | Method and system adaptive to speaker without vetating to speech identifying database |
| CN102324232A (en) * | 2011-09-12 | 2012-01-18 | 辽宁工业大学 | Method for recognizing sound-groove and system based on gauss hybrid models |
| CN103236260A (en) * | 2013-03-29 | 2013-08-07 | 京东方科技集团股份有限公司 | Voice recognition system |
| CN103730114A (en) * | 2013-12-31 | 2014-04-16 | 上海交通大学无锡研究院 | Mobile equipment voiceprint recognition method based on joint factor analysis model |
Non-Patent Citations (2)
| Title |
|---|
| 《基于HMM的语音识别技术研究》;魏峰等;《2006通信理论与技术新进展-第十一届全国青年通信学术会议论文集》;20060701;全文 |
| 《基于自适应高斯混合模型与静动态听觉特征融合的说话人识别》;吴迪等;《光学精密工程》;20130615;全文 |
Also Published As
| Publication number | Publication date |
|---|---|
| CN105245497A (en) | 2016-01-13 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Ali et al. | Edge-centric multimodal authentication system using encrypted biometric templates | |
| CN103475490B (en) | A kind of auth method and device | |
| EP3373176B1 (en) | Tamper-resistant element for use in speaker recognition | |
| CN106709402A (en) | Living person identity authentication method based on voice pattern and image features | |
| CN108429619A (en) | Identity identifying method and system | |
| CN104700094B (en) | A kind of face identification method for intelligent robot and system | |
| CN104700018A (en) | Identification method for intelligent robots | |
| Rieger et al. | Speech based emotion recognition using spectral feature extraction and an ensemble of kNN classifiers | |
| CN109920435B (en) | Voiceprint recognition method and voiceprint recognition device | |
| US20210166715A1 (en) | Encoded features and rate-based augmentation based speech authentication | |
| CN107346568A (en) | The authentication method and device of a kind of gate control system | |
| CN105138886B (en) | Robot biometric identification system | |
| Dimaunahan et al. | MFCC and VQ voice recognition based ATM security for the visually disabled | |
| CN105245497B (en) | A kind of identity identifying method and device | |
| Mubeen et al. | Detection of impostor and tampered segments in audio by using an intelligent system | |
| Revathi et al. | Person authentication using speech as a biometric against play back attacks | |
| Wahid et al. | A Gaussian mixture models approach to human heart signal verification using different feature extraction algorithms | |
| US20230401301A1 (en) | Information processing device, information processing method, and recording medium | |
| CN113838469A (en) | Identity recognition method, system and storage medium | |
| US10803873B1 (en) | Systems, devices, software, and methods for identity recognition and verification based on voice spectrum analysis | |
| Shenai et al. | Fast biometric authentication system based on audio-visual fusion | |
| Chakraborty et al. | An improved approach to open set text-independent speaker identification (OSTI-SI) | |
| Al-Hassani et al. | Design a text-prompt speaker recognition system using LPC-derived features | |
| Ndeh de Mbah | Morphed Image Detection using Local Spectrum Analysis | |
| Saranya et al. | An approach to detect replay attack in automatic speaker verification system |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |