CN103973453A - Vocal print secret key generating method and device and logging-in method and system based on vocal print secret key - Google Patents
Vocal print secret key generating method and device and logging-in method and system based on vocal print secret key Download PDFInfo
- Publication number
- CN103973453A CN103973453A CN201410211288.7A CN201410211288A CN103973453A CN 103973453 A CN103973453 A CN 103973453A CN 201410211288 A CN201410211288 A CN 201410211288A CN 103973453 A CN103973453 A CN 103973453A
- Authority
- CN
- China
- Prior art keywords
- vocal print
- acoustic information
- print key
- vector set
- coupling
- 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
Links
- 230000001755 vocal effect Effects 0.000 title claims abstract description 208
- 238000000034 method Methods 0.000 title claims abstract description 83
- 238000013523 data management Methods 0.000 claims abstract description 41
- 239000013598 vector Substances 0.000 claims description 100
- 230000008878 coupling Effects 0.000 claims description 65
- 238000010168 coupling process Methods 0.000 claims description 65
- 238000005859 coupling reaction Methods 0.000 claims description 65
- 238000000605 extraction Methods 0.000 claims description 18
- 238000007726 management method Methods 0.000 claims description 15
- 238000012545 processing Methods 0.000 claims description 9
- 230000003595 spectral effect Effects 0.000 claims description 9
- 238000005516 engineering process Methods 0.000 description 5
- 238000012544 monitoring process Methods 0.000 description 5
- 238000013459 approach Methods 0.000 description 1
- 230000005540 biological transmission Effects 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 230000005611 electricity Effects 0.000 description 1
- 238000002474 experimental method Methods 0.000 description 1
- 230000001815 facial effect Effects 0.000 description 1
- 230000004927 fusion Effects 0.000 description 1
- 238000012423 maintenance Methods 0.000 description 1
Abstract
The invention discloses a vocal print secret key generating method and device and a logging-in method and system based on a vocal print secret key. The logging-in method comprises the steps that when a mobile client side logs into a data management system for the first time, first voice information is input, and a first vocal print secret key is generated according to the vocal print secret key generating method and stored in an authentication server; when the mobile client side logs into the data management system not for the first time, second voice information is input, a second vocal print secret key is generated, and authentication matching is carried out on the second vocal print secret key and the first vocal print secret key through the authentication server, and when the authentication matching is finished, the authentication confirming information is sent to the mobile client side to carry out logging in. By means of the vocal print secret key method, the safety of having access to the data management system through the mobile client side is guaranteed, and a database can be remotely controlled in real time.
Description
Technical field
The present invention relates to computer security technique, espespecially a kind of vocal print key generation method, device and based on vocal print key login method, system.
Background technology
Along with the continuous introduction of computer technology, a large amount of application data platforms are applied, the data management platform of the cloud data such as electricity business, data service service provider, operator or large data is developed, as the Amazon web service (AWS of Amazon, Amazon Web Services) platform, domestic Ali's cloud, the platforms such as fertile cloud.The powerful computing capability of these data management platforms has been widely used in national product field, as Taobao's platform of 12306 train tickets ticket bookings websites, Alibaba etc.Realize storage and the supervision of mass users data, DBA administer and maintain burden heavier.In addition, the geographical position of data management platform place database data center (IDC) often has certain physical distance with DBA's Office Area.For more convenient ground maintenance and management data resource, generally data management system is mapped on public network, login data management system and be mapped in the address of public network by public network IP, carry out management and the O&M of data management platform database.
At present, this data base management system is mapped on public network, login data management system by public network IP, there is following defect in the management and the O&M that carry out data management platform database: because the database of data management platform is carrying a large amount of data resources, need data management platform moment focused data storehouse situation, as data management platform staff, not in Administrative Area, (computer in Administrative Area has all been strengthened the defence setting of fire compartment wall, in security domain), cannot be by the real-time sign-on access data base management system in computer terminal outside Administrative Area, database is carried out to real-time servicing control.
In order to realize the real-time monitoring to data management system, can adopt mobile client logon data management system to carry out the supervision of data management system.But, in the time of mobile client logon data base management system, if adopt traditional password to carry out safety certification, there is following problem:
1, in the time that user's setting code is too simple, be easily subject to network attack, password is easily cracked; When too complicated, inconvenience memory, by other recording modes, easily causes password stolen.
2, the certain means of the easy victim utilization of password that traditional approach is set crack, as by the relevant information of user identity (ID card No., birthday etc.); Or dictionary attack method (dictionaryattack);
In addition, if account information is common source, system cannot be distinguished the worker information of login behavior, easily causes legal dispute.
If adopt biometrics identification technology (fingerprint, face recognition) to carry out safety certification, have following problem:
1, the privacy concerns of biometric sample; Such as article theft user fingerprints such as, the cups that assailant can touch from user, the facial characteristics information that hiding image collecting device may pirates shoot user;
2, because biological characteristic belongs to the feature based on statistical property, so there is ambiguity (fuzziness) in class.Ambiguity refers to due to many factors such as position, light, angle, backgrounds, and what transducer gathered same user repeatedly can exist certain difference between biometric sample, easily cause system erroneous judgement (especially face recognition exists these problems);
3, most of biological templates are not discrete, but the real number vector that utilizes certain signal processing technology to obtain in continuous domain, the deviation that easily causes appearance to be judged by accident.
To sum up, current existing data management system, if staff not in Office Area, cannot adopt the computer in Office Area to realize in real time effectively monitoring and management; Adopt existing mobile client to carry out the supervision of data management system, there is safety certification problem, therefore, the method of a kind of safety certification of exigence, for in the time leaving Administrative Area, can realize by mobile client the safety certification of data management system, realize the real-time monitoring of data management platform and management.
Summary of the invention
In order to solve the problems of the technologies described above, the invention provides a kind of vocal print key generation method, device and based on vocal print key login method, system, can utilize acoustic information to generate vocal print key, ensure to use the safety of mobile client logon data management system, realize the real-time monitoring to data management system.
In order to reach foregoing invention object, the invention discloses a kind of vocal print key generation method, comprising:
Utilize two kinds of vocal print feature extracting methods respectively acoustic information to be carried out to vocal print feature extraction, obtain the set of two vocal print characteristic vectors;
Calculate the set of each vocal print characteristic vector, obtain respectively the coupling mark vector set of the first preset value number;
Corresponding element in the set of each coupling mark vector is built and is fused to two tuples;
To build two tuples after merging as coordinate points, carry out the matching of curve analytic expression, using the coefficient of curve analytic expression sort and discretization processing after as vocal print key.
Further, two kinds of vocal print feature extracting methods comprise following two kinds of methods arbitrarily: Mel frequency cepstral coefficient MFCC, Mel linear spectral frequency MFSL, linear prediction cepstrum coefficient coefficient LPCC, residual error phase angle residual phase.
Further, calculate the set of each vocal print characteristic vector, the coupling mark vector set that obtains respectively the first preset value number comprises: by the vector of each vocal print characteristic vector set is inputed to respectively to gauss hybrid models-universal background model GMM-UBM or GMM model, to obtain respectively the coupling mark vector set of the first preset value number.
On the other hand, it is a kind of based on vocal print key login method that the application also provides, and comprising:
Mobile client is first when logon data management system, and input the first acoustic information, to generate the first vocal print key and to be stored on certificate server; When the non-management system of logon data first, input the second acoustic information, to generate the second vocal print key, certificate server authenticates the second vocal print key to mate with the first vocal print key, in the time completing certification coupling, send the definite information of certification to mobile client, login;
The generation method of described the first vocal print key/the second vocal print key is:
Adopt two kinds of vocal print feature extracting methods to extract respectively the vocal print characteristic vector set of the first acoustic information/the second acoustic information;
Calculate each vocal print characteristic vector set of the first acoustic information/the second acoustic information, obtain respectively the coupling mark vector set of the first preset value number of the first acoustic information/the second acoustic information;
Corresponding element in each coupling mark vector set of the first acoustic information/the second acoustic information is built and is fused to two tuples;
Two tuples after merging using the structure of the first acoustic information/the second acoustic information are as coordinate points, carry out the matching of the curve analytic expression of the first acoustic information/the second acoustic information, using the coefficient of the curve analytic expression of the first acoustic information/the second acoustic information sort and discretization process after as vocal print key.
Further, two kinds of vocal print feature extracting methods comprise following two kinds of methods arbitrarily: Mel frequency cepstral coefficient MFCC, Mel linear spectral frequency MFSL, linear prediction cepstrum coefficient coefficient LPCC, residual error phase angle residual phase.
Further, calculate each vocal print characteristic vector set of the first acoustic information/the second acoustic information, the coupling mark vector set that obtains respectively the first preset value number of the first acoustic information/the second acoustic information comprises: by the vector of each vocal print characteristic vector set of the first acoustic information/the second acoustic information is inputed to GMM-UBM or GMM model, to obtain respectively the coupling mark vector set of the first acoustic information/the second acoustic information of the first preset value number.
Further, the content of the second acoustic information when certificate server sets in advance each certification coupling, before the certification coupling that completes vocal print key is logined, the method also comprises: certificate server is determined when the content of the second acoustic information is the content setting in advance, logined; Otherwise, refusal login.
On the one hand, the application also provides a kind of vocal print key generating device, comprising: extraction unit, computing unit, structure integrated unit and matching generation unit again; Wherein,
Extraction unit, for utilizing two kinds of vocal print feature extracting methods respectively acoustic information to be carried out to vocal print feature extraction, obtains the set of two vocal print characteristic vectors;
Computing unit, for calculating the set of each vocal print characteristic vector, obtains respectively the coupling mark vector set of the first preset value number;
Build integrated unit, for the corresponding element of each coupling mark vector set is built and is fused to two tuples;
Matching generation unit, for to build two tuples after merging as coordinate points, carries out the matching of curve analytic expression, using the coefficient of curve analytic expression sort and discretization processing after as vocal print key.
Further, two kinds of vocal print feature extracting methods comprise following two kinds of methods arbitrarily: Mel frequency cepstral coefficient MFCC, Mel linear spectral frequency MFSL, linear prediction cepstrum coefficient coefficient LPCC, residual error phase angle residual phase.
Further, computing unit, specifically for by the vector of each vocal print characteristic vector set is inputed to GMM-UBM or GMM model, gathers with the coupling mark vector that obtains respectively the first preset value number.
Again on the one hand, it is a kind of based on vocal print key login system that the application also provides, and comprising: mobile client, certificate server and above-mentioned vocal print key generating device; Wherein,
Mobile client, during for logon data management system first, input the first acoustic information, sends to vocal print key generating device, to generate the first vocal print key information; When non-data management system first, input the second acoustic information, sends to vocal print key generating device, to generate the second vocal print key; Receive the login of certificate server and determine that information logins;
Certificate server, when receiving first logon data management system, the first vocal print key being generated by vocal print key generating device is stored; While receiving the non-management system of logon data first, the the second vocal print key being generated by vocal print key generating device, the second vocal print key is authenticated and mated with the first vocal print key, in the time completing certification coupling, send authenticate-acknowledge information to mobile client to mobile client;
Vocal print key generating device, generates the first vocal print key/the second vocal print key for receiving the first acoustic information/the second acoustic information, mails to certificate server service and stores/authenticate coupling.
Further, certificate server also comprises and sets in advance unit and content matching unit; Wherein,
Set in advance unit, the content of the second acoustic information while coupling for setting in advance each certification, mails to client and carries out the second Speech input;
Content matching unit, for before the certification coupling that completes vocal print key is logined, when the content of determining the second acoustic information is the content setting in advance, logins; Otherwise, refusal login.
Present techniques scheme comprises: mobile client is first when logon data management system, and input the first acoustic information, to generate the first vocal print key and to be stored on certificate server; When the non-management system of logon data first, input the second acoustic information, to generate the second vocal print key, certificate server authenticates the second vocal print key to mate with the first vocal print key, in the time completing certification coupling, send the definite information of certification to mobile client, login; The generation method of the first vocal print key/the second vocal print key is: adopt two kinds of vocal print feature extracting methods to extract respectively the vocal print characteristic vector set of the first acoustic information/the second acoustic information; Calculate each vocal print characteristic vector set of the first acoustic information/the second acoustic information, obtain respectively the coupling mark vector set of the first preset value number of the first acoustic information/the second acoustic information; Corresponding element in each coupling mark vector set of the first acoustic information/the second acoustic information is built and is fused to two tuples; Two tuples after merging using the structure of the first acoustic information/the second acoustic information are as coordinate points, carry out the matching of the curve analytic expression of the first acoustic information/the second acoustic information, using the coefficient of the curve analytic expression of the first acoustic information/the second acoustic information sort and discretization process after as vocal print key.The present invention, by adopting vocal print encryption key method, has ensured the safety of mobile client visit data management system, is conducive to real-time monitoring data management system.
Brief description of the drawings
Accompanying drawing described herein is used to provide a further understanding of the present invention, forms the application's a part, and schematic description and description of the present invention is used for explaining the present invention, does not form inappropriate limitation of the present invention.In the accompanying drawings:
Fig. 1 is the flow chart that the present invention realizes vocal print key generation method;
Fig. 2 is the structured flowchart that the present invention realizes vocal print key generating device;
Fig. 3 is that the present invention realizes the structured flowchart based on vocal print key login system.
Embodiment
Fig. 1 is the flow chart that the present invention realizes vocal print key generation method, as shown in Figure 1, comprising:
Step 100, utilize two kinds of vocal print feature extracting methods respectively acoustic information to be carried out to vocal print feature extraction, obtain the set of two vocal print characteristic vectors.
In this step, the vocal print characteristic vector set of extracting acoustic information comprises: utilize following any two kinds of vocal print feature extracting methods: Mel frequency cepstral coefficient (MFCC), Mel linear spectral frequency (MFSL), linear prediction cepstrum coefficient coefficient (LPCC), residual error phase angle (residual phase).
It should be noted that, the present invention carries out the extraction of vocal print feature and mainly carries out spectral frequency extracting method, general, just can apply and the present invention as long as can realize spectral frequency extraction.
Step 101, calculate the set of each vocal print characteristic vector, obtain respectively the coupling mark vector set of the first preset value number.
Here according to conventional experiment, the number of the coupling mark vector set of acquisition is greatly about about 120-150, and the first preset value number is main relevant with the interference of acoustic information.
In this step, calculate the set of each vocal print characteristic vector, the coupling mark vector set that obtains respectively the first preset value number comprises: by the vector of each vocal print characteristic vector set being inputed to respectively to gauss hybrid models-universal background model (GMM-UBM) or (GMM) model, to obtain respectively the coupling mark vector set of the first preset value number.Here the vector that refers to independent vocal print characteristic vector set is input to gauss hybrid models-universal background model (GMM-UBM) or (GMM) model.
It should be noted that, GMM-UBM or GMM model are computation model well known to those skilled in the art.
Step 102, the corresponding element in the set of each coupling mark vector is built and is fused to two tuples.
Here, suppose two vocal print characteristic vectors set of calculating in step 101, the coupling mark vector set that obtains the first preset value and be 6 is respectively A=(a, b, c, d, e, f), B=(g, h, i, j, k, l), building the binary composition merging Wei <a, g>, <b, h>, <c, i>, <d, j>, <e, k>, <f, l>,
Utilize first value of above two tuples as the corresponding x value of abscissa, utilize second value of two tuples as the corresponding y value of ordinate, build two-dimensional coordinate system.
Step 103, to build two tuples after fusion as coordinate points, carry out the matching of curve analytic expression, using the coefficient of curve analytic expression sort and discretization process after as vocal print key.
It should be noted that, the matching of curve analytic expression can be carried out matching by methods such as existing lagrange-interpolation, least square curve fittings, and the laggard line ordering of process of fitting treatment adopts intermediate value to carry out obtaining vocal print key after discretization processing.
Here, two tuples after merging, as coordinate points, suppose that the polynomial curve analytic expression of matching is:
p(x)＝a
0+a
1x+…+a
nk-1x
kn-1+a
nkx
kn
Will after the coefficient sequence of curve analytic expression, obtain A={a
0, a
1... a
kn}
Coefficient A sorts by size and then takes out intermediate value am, and the S after discretization comprises:
Here, if S is applied, need to utilize the basic handling method of biological characteristic, arrange as binaryzation character string forms, because this part is common practise in field, not repeat them here.
A kind of based on vocal print key login method, comprising: mobile client is first when logon data management system, and input the first acoustic information, to generate the first vocal print key and to be stored on certificate server; When the non-management system of logon data first, input the second acoustic information, to generate the second vocal print key, certificate server authenticates the second vocal print key to mate with the first vocal print key, in the time completing certification coupling, send the definite information of certification to mobile client, login;
The generation method of the first vocal print key/the second vocal print key is:
Adopt two kinds of vocal print feature extracting methods to extract respectively the vocal print characteristic vector set of the first acoustic information/the second acoustic information.Here, two kinds of vocal print feature extracting methods comprise following two kinds of method: MFCC, MFSL arbitrarily, residual phase, LPCC.
Calculate each vocal print characteristic vector set of the first acoustic information/the second acoustic information, obtain respectively the coupling mark vector set of the first preset value number of the first acoustic information/the second acoustic information;
Calculate each vocal print characteristic vector set of the first acoustic information/the second acoustic information, the coupling mark vector set that obtains respectively the first preset value number of the first acoustic information/the second acoustic information comprises: by the vector of each vocal print characteristic vector set of the first acoustic information/the second acoustic information is inputed to GMM-UBM or GMM model, to obtain respectively the coupling mark vector set of the first acoustic information/the second acoustic information of the first preset value number.
Corresponding element in each coupling mark vector set of the first acoustic information/the second acoustic information is built and is fused to two tuples.
Two tuples after merging using the structure of the first acoustic information/the second acoustic information are as coordinate points, carry out the matching of the curve analytic expression of the first acoustic information/the second acoustic information, using the coefficient of the curve analytic expression of the first acoustic information/the second acoustic information sort and discretization process after as vocal print key.
The content of the second acoustic information when certificate server sets in advance each certification coupling, before the certification coupling that completes vocal print key is logined, the inventive method also comprises: certificate server is determined when the content of the second acoustic information is the content setting in advance, logined; Otherwise, refusal login.
It should be noted that, by setting in advance the content of the second acoustic information of each login, while authenticating login, staff must complete the acoustic information consistent with setting in advance content, just can realize login.Determining of the content of the second acoustic information can realize by existing audio recognition method, the content of the second acoustic information of determining is mated with the content setting in advance, when identical, realize login, avoid other people to usurp staff's acoustic information and carry out system login.
Fig. 2 is the structured flowchart that the present invention realizes vocal print key generating device; As shown in Figure 2, comprising: extraction unit, computing unit, structure integrated unit and matching generation unit; Wherein,
Extraction unit, for utilizing two kinds of vocal print feature extracting methods respectively acoustic information to be carried out to vocal print feature extraction, obtains the set of two vocal print characteristic vectors.
Here, two kinds of vocal print feature extracting methods comprise following two kinds of method: MFCC, MFSL arbitrarily, residual phase, LPCC.
Computing unit, for calculating the set of each vocal print characteristic vector, obtains respectively the coupling mark vector set of the first preset value number.
Computing unit, specifically for by the vector of each vocal print characteristic vector set is inputed to GMM-UBM or GMM model, gathers with the coupling mark vector that obtains respectively the first preset value number.
Build integrated unit, for the corresponding element of each coupling mark vector set is built and is fused to two tuples.
Matching generation unit, for to build two tuples after merging as coordinate points, carries out the matching of curve analytic expression, using the coefficient of curve analytic expression sort and discretization processing after as vocal print key.
Fig. 3 is that the present invention realizes the structured flowchart based on vocal print key login system, as shown in Figure 3, comprising: mobile client, certificate server and vocal print key generating device; Wherein,
Mobile client, during for logon data management system first, input the first acoustic information, sends to vocal print key generating device, to generate the first vocal print key information; When non-data management system first, input the second acoustic information, sends to vocal print key generating device, to generate the second vocal print key; Receive the login of certificate server and determine that information logins;
Certificate server, when receiving first logon data management system, the first vocal print key being generated by vocal print key generating device is stored; While receiving the non-management system of logon data first, the the second vocal print key being generated by vocal print key generating device, the second vocal print key is authenticated and mated with the first vocal print key, in the time completing certification coupling, send authenticate-acknowledge information to mobile client to mobile client;
Vocal print key generating device, comprises extraction unit, computing unit, structure integrated unit and matching generation unit; Wherein,
Extraction unit, for receiving the first acoustic information, adopts two kinds of vocal print feature extracting methods respectively the first acoustic information to be carried out to vocal print feature extraction, obtains the set of two vocal print characteristic vectors; Receive the second acoustic information, adopt two kinds of vocal print feature extracting methods respectively the second acoustic information to be carried out to vocal print feature extraction, obtain the set of two vocal print characteristic vectors.
Two kinds of vocal print feature extracting methods comprise following two kinds of method: MFCC, MFSL arbitrarily, residual phase, LPCC.
Computing unit, for calculating each vocal print characteristic vector set of the first acoustic information, obtains respectively the coupling mark vector set of the first preset value number of the first acoustic information; Calculate each vocal print characteristic vector set of the second acoustic information, obtain respectively the coupling mark vector set of the first preset value number of the second acoustic information.
Computing unit, specifically for by the vector of each vocal print characteristic vector set of the first acoustic information is inputed to GMM-UBM or GMM model, to obtain respectively the coupling mark vector set of the first acoustic information of the first preset value number; By the vector of each vocal print characteristic vector set of the second acoustic information is inputed to GMM-UBM or GMM model, to obtain respectively the coupling mark vector set of the second acoustic information of the first preset value number.
Build integrated unit, for the corresponding element of each coupling mark vector set of the first acoustic information is built and is fused to two tuples; Corresponding element in each coupling mark vector set of the second acoustic information is built and is fused to two tuples;
Matching generation unit, for building two tuples after merging using the first acoustic information as coordinate points, carries out the matching of curve analytic expression, using the coefficient of curve analytic expression sort and discretization processing after as the first vocal print key; Build two tuples after merging as coordinate points using the second acoustic information, carry out the matching of curve analytic expression, using the coefficient of curve analytic expression sort and discretization processing after as the second vocal print key.
Certificate server also comprises and sets in advance unit and content matching unit; Wherein,
Set in advance unit, the content of the second acoustic information while coupling for setting in advance each certification, mails to client and carries out the second Speech input;
Content matching unit, for before the certification coupling that completes vocal print key is logined, when the content of determining the second acoustic information is the content setting in advance, logins; Otherwise, refusal login.
Below in conjunction with specific embodiment, the inventive method is known to detailed description, embodiment only the present invention is not limited to protection scope of the present invention for more clearly stating.
Embodiment 1
The present embodiment is applied as implementation environment using the login of cloud platform data management system.
For guaranteeing to realize in mobile client the secure log of cloud platform data management system, the present embodiment is described based on cloud platform data management system user vocal print key authentication scheme, comprising:
First in the time of mobile client login cloud platform access database, need to carry out the generation of vocal print key, detailed process is as follows:
Staff inputs the voice messaging of user profile and certain time length and (thinks setting, unsuitable too short, here be made as temporarily 10 seconds), mobile client submits to the acoustic information of registration to cloud platform registrar by secure socket layer protocol layer (SSL) passage.
It should be noted that, the language message here can be the voice of identical content, can be also the language of different content.Here, stolen in order to prevent staff's acoustic information, can set in advance the content of the second acoustic information of login each time, carry out Speech input according to the content of the second acoustic information, in the time that input voice content is identical with the content setting in advance, just can realize system login.
When cloud platform registrar receives after the acoustic information of registration, user profile and transmission of speech information, to vocal print key generating device, are produced to the first vocal print key.And by user profile and the first vocal print key in cloud platform storage server, storage in remoteuser table in registered user's sample database.
In the time that staff logins again, by mobile client input voice information, according to the method that generates the first vocal print key, generate the second vocal print key, the second vocal print key is mated with the first vocal print key in cloud platform storage server, in the time that coupling is passed through, login system.
The generation method of vocal print key is identical with the method that above execution mode vocal print key generates.
Although the disclosed execution mode of the application as above, the execution mode that described content only adopts for ease of understanding the application, not in order to limit the application.Those of skill in the art under any the application; do not departing under the prerequisite of the disclosed spirit and scope of the application; can in the form of implementing and details, carry out any amendment and variation; but the application's scope of patent protection, still must be as the criterion with the scope that appending claims was defined.
Claims (12)
1. a vocal print key generation method, is characterized in that, comprising:
Utilize two kinds of vocal print feature extracting methods respectively acoustic information to be carried out to vocal print feature extraction, obtain the set of two vocal print characteristic vectors;
Calculate the set of each vocal print characteristic vector, obtain respectively the coupling mark vector set of the first preset value number;
Corresponding element in the set of each coupling mark vector is built and is fused to two tuples;
To build two tuples after merging as coordinate points, carry out the matching of curve analytic expression, using the coefficient of curve analytic expression sort and discretization processing after as vocal print key.
2. vocal print key generation method according to claim 1, it is characterized in that, described two kinds of vocal print feature extracting methods comprise following two kinds of methods arbitrarily: Mel frequency cepstral coefficient MFCC, Mel linear spectral frequency MFSL, linear prediction cepstrum coefficient coefficient LPCC, residual error phase angle residual phase.
3. vocal print key generation method according to claim 1, it is characterized in that, the set of the each vocal print characteristic vector of described calculating, the coupling mark vector set that obtains respectively the first preset value number comprises: by the vector of each vocal print characteristic vector set is inputed to respectively to gauss hybrid models-universal background model GMM-UBM or GMM model, to obtain respectively the coupling mark vector set of the first preset value number.
4. based on a vocal print key login method, it is characterized in that, comprising:
Mobile client is first when logon data management system, and input the first acoustic information, to generate the first vocal print key and to be stored on certificate server; When the non-management system of logon data first, input the second acoustic information, to generate the second vocal print key, certificate server authenticates the second vocal print key to mate with the first vocal print key, in the time completing certification coupling, send the definite information of certification to mobile client, login;
The generation method of described the first vocal print key/the second vocal print key is:
Adopt two kinds of vocal print feature extracting methods to extract respectively the vocal print characteristic vector set of the first acoustic information/the second acoustic information;
Calculate each vocal print characteristic vector set of the first acoustic information/the second acoustic information, obtain respectively the coupling mark vector set of the first preset value number of the first acoustic information/the second acoustic information;
Corresponding element in each coupling mark vector set of the first acoustic information/the second acoustic information is built and is fused to two tuples;
Two tuples after merging using the structure of the first acoustic information/the second acoustic information are as coordinate points, carry out the matching of the curve analytic expression of the first acoustic information/the second acoustic information, using the coefficient of the curve analytic expression of the first acoustic information/the second acoustic information sort and discretization process after as vocal print key.
5. according to claim 4 based on vocal print key login method, it is characterized in that, described two kinds of vocal print feature extracting methods comprise following two kinds of methods arbitrarily: Mel frequency cepstral coefficient MFCC, Mel linear spectral frequency MFSL, linear prediction cepstrum coefficient coefficient LPCC, residual error phase angle residual phase.
6. according to claim 4 based on vocal print key login method, it is characterized in that, calculate each vocal print characteristic vector set of the first acoustic information/the second acoustic information, the coupling mark vector set that obtains respectively the first preset value number of the first acoustic information/the second acoustic information comprises: by the vector of each vocal print characteristic vector set of the first acoustic information/the second acoustic information is inputed to GMM-UBM or GMM model, to obtain respectively the coupling mark vector set of the first acoustic information/the second acoustic information of the first preset value number.
7. method according to claim 4, it is characterized in that, the content of the second acoustic information when certificate server sets in advance each certification coupling, before the certification coupling that completes vocal print key is logined, the method also comprises: certificate server is determined when the content of the second acoustic information is the content setting in advance, logined; Otherwise, refusal login.
8. a vocal print key generating device, is characterized in that, comprising: extraction unit, computing unit, structure integrated unit and matching generation unit; Wherein,
Extraction unit, for utilizing two kinds of vocal print feature extracting methods respectively acoustic information to be carried out to vocal print feature extraction, obtains the set of two vocal print characteristic vectors;
Computing unit, for calculating the set of each vocal print characteristic vector, obtains respectively the coupling mark vector set of the first preset value number;
Build integrated unit, for the corresponding element of each coupling mark vector set is built and is fused to two tuples;
Matching generation unit, for to build two tuples after merging as coordinate points, carries out the matching of curve analytic expression, using the coefficient of curve analytic expression sort and discretization processing after as vocal print key.
9. vocal print key generating device according to claim 8, it is characterized in that, described two kinds of vocal print feature extracting methods comprise following two kinds of methods arbitrarily: Mel frequency cepstral coefficient MFCC, Mel linear spectral frequency MFSL, linear prediction cepstrum coefficient coefficient LPCC, residual error phase angle residual phase.
10. vocal print key generating device according to claim 8, it is characterized in that, described computing unit, specifically for by the vector of each vocal print characteristic vector set is inputed to GMM-UBM or GMM model, gathers with the coupling mark vector that obtains respectively the first preset value number.
11. 1 kinds based on vocal print key login system, it is characterized in that, comprising: the vocal print key generating device of mobile client, certificate server and claim 8～10; Wherein,
Mobile client, during for logon data management system first, input the first acoustic information, sends to vocal print key generating device, to generate the first vocal print key information; When non-data management system first, input the second acoustic information, sends to vocal print key generating device, to generate the second vocal print key; Receive the login of certificate server and determine that information logins;
Certificate server, when receiving first logon data management system, the first vocal print key being generated by vocal print key generating device is stored; While receiving the non-management system of logon data first, the the second vocal print key being generated by vocal print key generating device, the second vocal print key is authenticated and mated with the first vocal print key, in the time completing certification coupling, send authenticate-acknowledge information to mobile client to mobile client;
Vocal print key generating device, generates the first vocal print key/the second vocal print key for receiving the first acoustic information/the second acoustic information, mails to certificate server service and stores/authenticate coupling.
12. systems according to claim 11, is characterized in that, described certificate server also comprises and sets in advance unit and content matching unit; Wherein,
Set in advance unit, the content of the second acoustic information while coupling for setting in advance each certification, mails to client and carries out the second Speech input;
Content matching unit, for before the certification coupling that completes vocal print key is logined, when the content of determining the second acoustic information is the content setting in advance, logins; Otherwise, refusal login.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201410211288.7A CN103973453A (en) | 2014-05-19 | 2014-05-19 | Vocal print secret key generating method and device and logging-in method and system based on vocal print secret key |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201410211288.7A CN103973453A (en) | 2014-05-19 | 2014-05-19 | Vocal print secret key generating method and device and logging-in method and system based on vocal print secret key |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN103973453A true CN103973453A (en) | 2014-08-06 |
Family
ID=51242524
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201410211288.7A Pending CN103973453A (en) | 2014-05-19 | 2014-05-19 | Vocal print secret key generating method and device and logging-in method and system based on vocal print secret key |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN103973453A (en) |
Cited By (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104639528A (en) * | 2014-11-19 | 2015-05-20 | 中国联合网络通信集团有限公司 | DBA (database administrator) mobile client counterattack method and DBA mobile client counterattack device |
| CN105096121A (en) * | 2015-06-25 | 2015-11-25 | 百度在线网络技术（北京）有限公司 | Voiceprint authentication method and device |
| CN106228990A (en) * | 2016-07-15 | 2016-12-14 | 北京光年无限科技有限公司 | Login method and operating system towards intelligent robot |
| CN107464111A (en) * | 2017-07-10 | 2017-12-12 | 北京云知科技有限公司 | Block chain wallet address generating method, method of payment, device and electronic equipment based on audio |
| CN111741460A (en) * | 2019-03-25 | 2020-10-02 | 南宁富桂精密工业有限公司 | System and method for automatically connecting Wi-Fi |
| CN113179157A (en) * | 2021-03-31 | 2021-07-27 | 杭州电子科技大学 | Text-related voiceprint biological key generation method based on deep learning |
| CN113726528A (en) * | 2021-11-02 | 2021-11-30 | 深圳奥联信息安全技术有限公司 | Bone voiceprint-based key protection method and system |
-
2014
- 2014-05-19 CN CN201410211288.7A patent/CN103973453A/en active Pending
Cited By (11)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104639528A (en) * | 2014-11-19 | 2015-05-20 | 中国联合网络通信集团有限公司 | DBA (database administrator) mobile client counterattack method and DBA mobile client counterattack device |
| CN105096121A (en) * | 2015-06-25 | 2015-11-25 | 百度在线网络技术（北京）有限公司 | Voiceprint authentication method and device |
| US9792913B2 (en) | 2015-06-25 | 2017-10-17 | Baidu Online Network Technology (Beijing) Co., Ltd. | Voiceprint authentication method and apparatus |
| CN106228990A (en) * | 2016-07-15 | 2016-12-14 | 北京光年无限科技有限公司 | Login method and operating system towards intelligent robot |
| CN107464111A (en) * | 2017-07-10 | 2017-12-12 | 北京云知科技有限公司 | Block chain wallet address generating method, method of payment, device and electronic equipment based on audio |
| CN107464111B (en) * | 2017-07-10 | 2021-01-29 | 北京云知科技有限公司 | Block chain wallet address generation method based on audio, payment method and device and electronic equipment |
| CN111741460A (en) * | 2019-03-25 | 2020-10-02 | 南宁富桂精密工业有限公司 | System and method for automatically connecting Wi-Fi |
| CN113179157A (en) * | 2021-03-31 | 2021-07-27 | 杭州电子科技大学 | Text-related voiceprint biological key generation method based on deep learning |
| CN113179157B (en) * | 2021-03-31 | 2022-05-17 | 杭州电子科技大学 | Text-related voiceprint biological key generation method based on deep learning |
| CN113726528A (en) * | 2021-11-02 | 2021-11-30 | 深圳奥联信息安全技术有限公司 | Bone voiceprint-based key protection method and system |
| CN113726528B (en) * | 2021-11-02 | 2022-02-18 | 深圳奥联信息安全技术有限公司 | Bone voiceprint-based key protection method and system |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN103973453A (en) | Vocal print secret key generating method and device and logging-in method and system based on vocal print secret key | |
| US11528268B2 (en) | System and method for verifying liveliness | |
| TWI530150B (en) | Identity authentication device and method thereof | |
| US7519558B2 (en) | Biometrically enabled private secure information repository | |
| Ibrokhimov et al. | Multi-factor authentication in cyber physical system: A state of art survey | |
| CN110457878A (en) | A kind of identity identifying method based on block chain, apparatus and system | |
| US10015171B1 (en) | Authentication using metadata from posts made to social networking websites | |
| CN100576797C (en) | Network identity validation method based on fingerprint | |
| CN103986584A (en) | Double-factor identity verification method based on intelligent equipment | |
| CN107113315A (en) | Identity authentication method, terminal and server | |
| CN104158664A (en) | Identity authentication method and system | |
| CN108989346A (en) | The effective identity trustship agility of third party based on account concealment authenticates access module | |
| CN103607387A (en) | A network login authentication cloud service system based on face identification and a method | |
| CN103259800A (en) | Internet login system based on face recognition and internet login method | |
| CN103024706A (en) | Short message based device and short message based method for bidirectional multiple-factor dynamic identity authentication | |
| CN109802942A (en) | A kind of voiceprint authentication method and system, mobile terminal of secret protection | |
| CN104079413A (en) | Enhancement type one-time dynamic password authentication method and system | |
| CN104639528A (en) | DBA (database administrator) mobile client counterattack method and DBA mobile client counterattack device | |
| CN104240348A (en) | Admittance identity authentication method based on image identification | |
| CN104703180A (en) | Implicit multiple authentication method based on mobile Internet and intelligent terminal | |
| CN109639623A (en) | Verification System and server unit | |
| JPH10161979A (en) | User authentication by fingerprint at time of log-in to server and converted password | |
| CN110505199A (en) | Email safe login method based on the asymmetric identity of lightweight | |
| CN102299805A (en) | Method for realizing single sign-on based on LDAP (Lightweight Directory Access Protocol) service | |
| CN102769606B (en) | A kind of network digital identity identifying method based on gene certificate |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20140806 |
|
| RJ01 | Rejection of invention patent application after publication |