CN101740031B - Network dynamic load balancing-based voiceprint recognition system and recognition method thereof - Google Patents
Network dynamic load balancing-based voiceprint recognition system and recognition method thereof Download PDFInfo
- Publication number
- CN101740031B CN101740031B CN201010046573XA CN201010046573A CN101740031B CN 101740031 B CN101740031 B CN 101740031B CN 201010046573X A CN201010046573X A CN 201010046573XA CN 201010046573 A CN201010046573 A CN 201010046573A CN 101740031 B CN101740031 B CN 101740031B
- Authority
- CN
- China
- Prior art keywords
- voiceprint recognition
- server
- load
- voice
- user
- 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 14
- 230000001755 vocal effect Effects 0.000 claims description 38
- 238000004891 communication Methods 0.000 claims description 12
- 238000012545 processing Methods 0.000 claims description 11
- 230000006870 function Effects 0.000 claims description 9
- 238000000605 extraction Methods 0.000 claims description 7
- 238000012217 deletion Methods 0.000 claims description 4
- 230000037430 deletion Effects 0.000 claims description 4
- 238000006243 chemical reaction Methods 0.000 claims description 3
- 230000004044 response Effects 0.000 abstract description 4
- 238000010586 diagram Methods 0.000 description 4
- 238000005516 engineering process Methods 0.000 description 4
- 238000004364 calculation method Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000002045 lasting effect Effects 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
- 238000004321 preservation Methods 0.000 description 1
- 239000004575 stone Substances 0.000 description 1
- 238000012795 verification Methods 0.000 description 1
Images
Abstract
The invention relates to a network dynamic load balancing-based voiceprint recognition system, which comprises a voiceprint client used for acquiring the voice generated by a user, wherein the voiceprint client is communicated with a load balancing server; the load balancing server used for searching an optimized voiceprint recognition server is communicated with a voiceprint recognition server group, wherein the voiceprint recognition server group consists of at least two voiceprint recognition servers; and the optimized voiceprint recognition server feeds back a recognition result to the voiceprint client. The invention also discloses a recognition method for the network dynamic load balancing-based voiceprint recognition system. The optimized voiceprint recognition server is searched by the load balancing server, dynamic allocation is carried out on the traffic and voiceprint recognition server cluster automatically, the problems of real-time response of the identity recognition of a speaker and the system safety under the large-scale concurrency condition are solved, and the requirement on the voiceprint recognition with various scales and various concurrency values can be completely met.
Description
Technical field
The present invention relates to the Application on Voiceprint Recognition field, especially a kind of Voiceprint Recognition System of dynamic load leveling Network Based.
Background technology
Along with the continuous progress of all kinds of science and technology, safety issue also more and more comes into one's own.Because vocal print is different from the other biological recognition technology, such as fingerprint recognition, palmmprint identification, iris recognition etc., vocal print has and is difficult to lose or forgets, need not memory, economic dispatch advantage easy to use, safe, the safety verification that is particularly useful for remote service, as seen these advantages all can become the foundation stone of Application on Voiceprint Recognition widespread use.Recent years, the judiciary of countries in the world all begins to accept the vocal print evidence.In the U.S., existing 7/8ths court of appealing, state admits that vocal print can be used as court evidence, and the martial court of the U.S. also adopts vocal print as evidence.Vocal print is identified in national security community, judiciary and financial institution at present urgent demand.Vocal print is also brought into use in the production environment of reality as important identity recognizing technology means.
A typical Application on Voiceprint Recognition flow process comprises the function of three aspects: the first, and the vocal print registration, the sound characteristic that is about to a people is preserved, as the reference data of follow-up identification; The second, vocal print confirms, speaker's voice extracted feature compare with the sound-groove model that registration is good before, confirms whether its identity that shows of speaker; The 3rd, vocal print is differentiated, with speaker's voice extraction feature, afterwards a registered good vocal print set search, to confirm which sound-groove model that belongs to of speaker.
Yet, for the parallel Application on Voiceprint Recognition of continuous large-scale, existing system does not have the real-time of the next abundant assurance of good method system and the dynamic load of server, this problem can be more general in the large-scale pooled applications such as similar bank of reality, the user can run into waits for too long, and system is difficult to the problem of dilatation.Generally solve the problems referred to above by third-party load plug-in unit at present, this load plug-in unit is not to formulate for the characteristic of Application on Voiceprint Recognition, can't well deal with problems, simultaneously can't good cross-platform operation.
Summary of the invention
Primary and foremost purpose of the present invention is to provide that a kind of extendability is strong, the Voiceprint Recognition System of dynamic load leveling Network Based that can equally loaded.
For achieving the above object, the present invention has adopted following technical scheme: a kind of Voiceprint Recognition System of dynamic load leveling Network Based, comprise the vocal print client that gathers for the sound that the user is sent, vocal print client and load-balanced server communication, for load-balanced server and the communication of Application on Voiceprint Recognition server group of searching optimum Application on Voiceprint Recognition server, Application on Voiceprint Recognition server group is comprised of at least two Application on Voiceprint Recognition servers, and optimum Application on Voiceprint Recognition server feeds back to the vocal print client with recognition result.
Another object of the present invention is to provide a kind of recognition methods of Voiceprint Recognition System of dynamic load leveling Network Based, the method comprises the step of following order:
(1) voice that send of vocal print client user, and search optimum Application on Voiceprint Recognition server by load-balanced server, load-balanced server sends user voice data to this optimum Application on Voiceprint Recognition server;
(2) optimum Application on Voiceprint Recognition server is processed user voice data, and recognition result is fed back to the vocal print client.
As shown from the above technical solution, the present invention not only provides the vocal print registration of core, confirm and identification function, support simultaneously the load of parallel and a plurality of services of these functions, utilize load-balanced server to search optimum Application on Voiceprint Recognition server, automatically flow harmony line identified server cluster is carried out dynamic assignment, solve real-time response and the security of system of speaker ' s identity identification under the large-scale concurrent condition, can satisfy the Application on Voiceprint Recognition demand of various scales and various concurrency fully, so that system can be very easily online dilatation and multi-platform deployment, thereby provide strong technical support for the large-scale application of Application on Voiceprint Recognition.
Description of drawings
Fig. 1 is circuit structure block diagram of the present invention;
Fig. 2 is workflow diagram of the present invention;
Fig. 3 is network architecture synoptic diagram of the present invention;
The workflow diagram that Fig. 4,5,6 is respectively voice registration among the present invention, voice are confirmed and voice are differentiated.
Embodiment
A kind of Voiceprint Recognition System of dynamic load leveling Network Based, comprise the vocal print client 10 that gathers for the sound that the user is sent, vocal print client 10 and load-balanced server 20 communications, be used for searching the load-balanced server 20 and 30 communications of Application on Voiceprint Recognition server group of optimum Application on Voiceprint Recognition server 32, Application on Voiceprint Recognition server group 30 is comprised of at least two Application on Voiceprint Recognition servers 31, reciprocity between each vocal print server 31, and can expand, unified saving the data on the same voice print database server 50, each vocal print server 31 visits voice print database server 50 by the unified access interface of voice print database server 50, optimum Application on Voiceprint Recognition server 32 feeds back to vocal print client 10 with recognition result, such as Fig. 1, shown in 3.
In conjunction with Fig. 1,3, the interface module 31a that described Application on Voiceprint Recognition server 31 comprises directly and load-balanced server 20 carries out communication, interface module 31a is connected with the memory module 31c that is connected voice data is stored with the speech processing module 31b that is used for the sound that the user sends is processed respectively, interface module 31a calls speech processing module 31b and memory module 31c's, by calculating, preservation, inquiry and the identification of calling to finish vocal print.
In conjunction with Fig. 1,3, described load-balanced server 20 comprises a main load-balanced server 21 and at least one auxiliary load-balanced server 22, main load-balanced server 21 and 30 communications of Application on Voiceprint Recognition server group, and search optimum Application on Voiceprint Recognition server 32, identified by 32 pairs of user speech of optimum Application on Voiceprint Recognition server, auxiliary load-balanced server 22 detects main load-balanced server 21 do not work after, the auxiliary load-balanced server 22 of one of them upgrades to main load-balanced server 21, to guarantee continuous, the stable operation of service.
In conjunction with Fig. 1,3, the information of described load-balanced server 20 all Application on Voiceprint Recognition servers 31 of load poll, as cpu take, the information such as the available way of vocal print mandate and EMS memory occupation, then according to a configurable weight, calculate score and the ordering of each Application on Voiceprint Recognition server 31, then offer vocal print client 10.Wherein, the available way of vocal print mandate refers to carry out simultaneously How many people identification, an important symbol judging that Application on Voiceprint Recognition server 31 can be used simultaneously, can avoid in the situation that a lot of clients access simultaneously identification occuring assembles, for the lasting acquisition of serving provides good guarantee, the available way of vocal print mandate is by software set.Vocal print client 10 need not to know the address of concrete Application on Voiceprint Recognition server 31 on the network, only need to be when connecting query load equalization server 20 at first, load-balanced server 20 can return an optimum Application on Voiceprint Recognition server 32 in real time to vocal print client 10.Simultaneously in order to monitor Application on Voiceprint Recognition servers 31 all on the whole network and the various situations of vocal print client 10, also be provided with a service watch instrument SPC 40, operation service adviser tool SPC 40, the system manager can manage whole vocal print service system, and whole system operates on the environment of gigabit LAN.
Below in conjunction with Fig. 1,2,3,4,5,6 the present invention is described further.
When work, at first vocal print client 10 receives the voice that the user sends, and search optimum Application on Voiceprint Recognition server 32 by load-balanced server 20, optimum Application on Voiceprint Recognition server 32 refers to the most idle Application on Voiceprint Recognition server 31, after finding optimum Application on Voiceprint Recognition server 32, load-balanced server 20 sends a request to this optimum Application on Voiceprint Recognition server 32, and sends user's speech data to this optimum Application on Voiceprint Recognition server 32.Optimum Application on Voiceprint Recognition server 32 response request, accept user's speech data, and calling interface module 31a, interface module 31a is according to the type of user's request, comprise in the present invention three kinds of request types: registration, affirmation and discriminating, read user's related data information by memory module 31c, then call the correlation computations that speech processing module 31b carries out model, and the former road of result of calculation is back to vocal print client 10, finish once typical Application on Voiceprint Recognition funcall.
When carrying out the voice registration, the registering functional of optimum Application on Voiceprint Recognition server 32 calling interface module 31a, speech processing module 31b carries out pre-service to voice, described pre-service comprises quiet section in the deletion voice, white noise in the deletion audio frequency, to format conversion of voice etc., wherein, to the format conversion of voice such as being with VOX, alaw the non-pcm formatted voice such as ulaw convert the pcm form to, pretreated voice are extracted feature according to certain algorithm, feature after the extraction is by calculating user's speech model, memory module 31c stores in model database speech model and user's relevant information, and user's relevant information is user's ID, user's sex etc.
When carrying out the voice affirmation, the affirmation function of optimum Application on Voiceprint Recognition server 32 calling interface module 31a, speech processing module 31b carries out pre-service equally to voice, pretreated voice are extracted feature according to certain algorithm, the speech model that this extraction feature and memory module 31c are stored is compared and is given a mark, if mark is higher than the threshold value of setting, then thinks same people, otherwise, then be not same people.
When carrying out the voice discriminating, the identification function of optimum Application on Voiceprint Recognition server 32 calling interface module 31a, speech processing module 31b carries out pre-service equally to voice, pretreated voice are extracted feature according to certain algorithm, this extraction feature and all users' that will differentiate model are compared one by one, and the ordering that counts the score, if the first place score is higher than the threshold value of setting, then the speaker is exactly the user of first place, otherwise then explanation changes voice and do not belong among the user anyone.
In sum, the invention provides comprehensive Application on Voiceprint Recognition function, in order to satisfy the application of these functions in enterprise, adopted dynamic load leveling to come unified management, greatly promote the stability of whole system, reduced simultaneously the response time under the large-scale concurrent.The invention enables Application on Voiceprint Recognition to obtain larger development space and chance in security fields, can also can allow simultaneously more people be familiar with and approve this easy-to-use new authentication.
Claims (7)
1. the Voiceprint Recognition System of a dynamic load leveling Network Based, it is characterized in that: comprise the vocal print client that gathers for the sound that the user is sent, vocal print client and load-balanced server communication, for load-balanced server and the communication of Application on Voiceprint Recognition server group of searching optimum Application on Voiceprint Recognition server, Application on Voiceprint Recognition server group is comprised of at least two Application on Voiceprint Recognition servers, and optimum Application on Voiceprint Recognition server feeds back to the vocal print client with recognition result;
The interface module that described Application on Voiceprint Recognition server comprises directly and load-balanced server carries out communication, interface module are connected with the memory module that is connected voice data is stored with the speech processing module that is used for the sound that the user sends is processed respectively;
Described load-balanced server comprises a main load-balanced server and at least one auxiliary load-balanced server, main load-balanced server and the communication of Application on Voiceprint Recognition server group.
2. the recognition methods of the Voiceprint Recognition System of dynamic load leveling Network Based according to claim 1, the method comprises the step of following order:
(1) voice that send of vocal print client user, and search optimum Application on Voiceprint Recognition server by load-balanced server, load-balanced server sends user voice data to this optimum Application on Voiceprint Recognition server;
(2) optimum Application on Voiceprint Recognition server is processed user voice data, and recognition result is fed back to the vocal print client;
Described load-balanced server comprises a main load-balanced server and at least one auxiliary load-balanced server, main load-balanced server and the communication of Application on Voiceprint Recognition server group, and search optimum Application on Voiceprint Recognition server, by optimum Application on Voiceprint Recognition server user speech is identified, auxiliary load-balanced server detects main load-balanced server do not work after, the auxiliary load-balanced server of one of them upgrades to main load-balanced server, the interface module that described Application on Voiceprint Recognition server comprises directly and load-balanced server carries out communication, interface module are connected with the memory module that is connected voice data is stored with the speech processing module that is used for the sound that the user sends is processed respectively.
3. the recognition methods of the Voiceprint Recognition System of dynamic load leveling Network Based according to claim 2, it is characterized in that: when carrying out the voice registration, the registering functional of optimum Application on Voiceprint Recognition server calls interface module, speech processing module is carried out pre-service to voice, pretreated voice are extracted feature according to certain algorithm, feature after the extraction is by calculating user's speech model, and memory module is stored speech model and user's relevant information.
4. the recognition methods of the Voiceprint Recognition System of dynamic load leveling Network Based according to claim 3, it is characterized in that: when carrying out the voice affirmation, the affirmation function of optimum Application on Voiceprint Recognition server calls interface module, speech processing module is carried out pre-service to voice, pretreated voice are extracted feature according to certain algorithm, the speech model that this extraction feature and memory module are stored is compared and is given a mark, if mark is higher than the threshold value of setting, then think same people, otherwise, then be not same people.
5. the recognition methods of the Voiceprint Recognition System of dynamic load leveling Network Based according to claim 2, it is characterized in that: when carrying out the voice discriminating, the identification function of optimum Application on Voiceprint Recognition server calls interface module, speech processing module is carried out pre-service to voice, pretreated voice are extracted feature according to certain algorithm, this extraction feature and all users' that will differentiate model are compared one by one, and the ordering that counts the score, if the first place score is higher than the threshold value of setting, then the speaker is exactly the user of first place, otherwise then explanation changes voice and does not belong among the user anyone.
6. according to claim 3 or the recognition methods of the Voiceprint Recognition System of 4 or 5 described dynamic load levelings Network Based, it is characterized in that: described user's relevant information is user's ID, user's sex.
7. it is characterized in that according to claim 3 or the recognition methods of the Voiceprint Recognition System of 4 or 5 described dynamic load levelings Network Based: described pre-service comprises the white noise in format conversion to voice, the deletion voice quiet section, deletion audio frequency.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201010046573XA CN101740031B (en) | 2010-01-21 | 2010-01-21 | Network dynamic load balancing-based voiceprint recognition system and recognition method thereof |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201010046573XA CN101740031B (en) | 2010-01-21 | 2010-01-21 | Network dynamic load balancing-based voiceprint recognition system and recognition method thereof |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN101740031A CN101740031A (en) | 2010-06-16 |
| CN101740031B true CN101740031B (en) | 2013-01-02 |
Family
ID=42463407
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201010046573XA Active CN101740031B (en) | 2010-01-21 | 2010-01-21 | Network dynamic load balancing-based voiceprint recognition system and recognition method thereof |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN101740031B (en) |
Families Citing this family (12)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102347060A (en) * | 2010-08-04 | 2012-02-08 | 鸿富锦精密工业（深圳）有限公司 | Electronic recording device and method |
| CN101950564A (en) * | 2010-10-13 | 2011-01-19 | 镇江华扬信息科技有限公司 | Remote digital voice acquisition, analysis and identification system |
| CN103903622A (en) * | 2012-12-28 | 2014-07-02 | 腾讯科技（深圳）有限公司 | Voice verification method, apparatus and system |
| CN103971687B (en) * | 2013-02-01 | 2016-06-29 | 腾讯科技（深圳）有限公司 | Implementation of load balancing in a kind of speech recognition system and device |
| CN103325371A (en) * | 2013-06-05 | 2013-09-25 | 杭州网豆数字技术有限公司 | Voice recognition system and method based on cloud |
| CN104732972B (en) * | 2015-03-12 | 2018-02-27 | 广东外语外贸大学 | A kind of HMM Application on Voiceprint Recognition based on classified statistics is registered method and system |
| US9940948B2 (en) * | 2015-08-02 | 2018-04-10 | Resonance Software Llc | Systems and methods for enabling information exchanges between devices |
| CN109819057B (en) * | 2019-04-08 | 2020-09-11 | 科大讯飞股份有限公司 | Load balancing method and system |
| CN110691118A (en) * | 2019-08-30 | 2020-01-14 | 许昌许继软件技术有限公司 | Service selection method and device in micro-service cluster |
| CN111048100B (en) * | 2019-11-21 | 2023-09-08 | 深圳市东进银通电子有限公司 | Big data parallelization voiceprint recognition system and method |
| CN112231510B (en) * | 2020-12-17 | 2021-03-16 | 北京远鉴信息技术有限公司 | Voiceprint storage method, voiceprint query method, server and storage medium |
| CN113724711A (en) * | 2021-09-01 | 2021-11-30 | 云从科技集团股份有限公司 | Method, device, system, medium and equipment for realizing voice recognition service |
-
2010
- 2010-01-21 CN CN201010046573XA patent/CN101740031B/en active Active
Also Published As
| Publication number | Publication date |
|---|---|
| CN101740031A (en) | 2010-06-16 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN101740031B (en) | Network dynamic load balancing-based voiceprint recognition system and recognition method thereof | |
| CN104715752B (en) | Audio recognition method, apparatus and system | |
| CN102800008B (en) | The system and method that a kind of accumulated point exchanging and integration are given | |
| CN103136627B (en) | Area real estate intelligent management system | |
| WO2019196303A1 (en) | User identity authentication method, server and storage medium | |
| CN203165005U (en) | Area property intelligent management system | |
| CN104580477A (en) | Voice data processing method and device | |
| CN110782341A (en) | Business collection method, device, equipment and medium | |
| WO2018232887A1 (en) | Mobile person information acquisition feedback method, server, and system | |
| CN110109757B (en) | High-performance computing method based on cloud computing | |
| CN113573029B (en) | Multi-party audio and video interaction method and system based on IOT | |
| CN110688473A (en) | Method for robot to dynamically acquire information | |
| CN111694884B (en) | Intelligent government affair request processing method based on big data | |
| CN111553826B (en) | Smart city data processing method | |
| CN111414364B (en) | User information generation method and device and electronic equipment | |
| CN113936677A (en) | Tone conversion method, device, computer equipment and storage medium | |
| CN114358543A (en) | Information processing method and device | |
| CN113890948A (en) | Resource allocation method based on voice outbound robot dialogue data and related equipment | |
| CN112766822A (en) | Enterprise decision management information service system | |
| CN201259777Y (en) | Chinese speech recognizing system | |
| CN111553600B (en) | Big data-based smart city business distribution system | |
| CN107479959B (en) | Smart city management system | |
| CN106504022A (en) | A kind of network marketing system that is mated based on client | |
| CN107194014B (en) | Data source calling method and device | |
| CN106656906A (en) | Biological identification system based on cloud platform, and method thereof |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| C14 | Grant of patent or utility model | ||
| GR01 | Patent grant | ||
| C56 | Change in the name or address of the patentee |
Owner name: IFLYTEK CO., LTD. Free format text: FORMER NAME: ANHUI USTC IFLYTEK CO., LTD. |
|
| CP03 | Change of name, title or address |
Address after: Wangjiang Road high tech Development Zone Hefei city Anhui province 230088 No. 666 Patentee after: IFLYTEK Co.,Ltd. Address before: 230088 No. 616, Mount Huangshan Road, hi tech Zone, Anhui, Hefei Patentee before: ANHUI USTC IFLYTEK Co.,Ltd. |
|
| TR01 | Transfer of patent right |
Effective date of registration: 20170714 Address after: 230088, Hefei province high tech Zone, 2800 innovation Avenue, 288 innovation industry park, H2 building, room two, Anhui Patentee after: Anhui Puji Information Technology Co.,Ltd. Address before: Wangjiang Road high tech Development Zone Hefei city Anhui province 230088 No. 666 Patentee before: IFLYTEK Co.,Ltd. |
|
| TR01 | Transfer of patent right | ||
| CP01 | Change in the name or title of a patent holder |
Address after: 230088, Hefei province high tech Zone, 2800 innovation Avenue, 288 innovation industry park, H2 building, room two, Anhui Patentee after: ANHUI IFLYTEK MEDICAL INFORMATION TECHNOLOGY CO.,LTD. Address before: 230088, Hefei province high tech Zone, 2800 innovation Avenue, 288 innovation industry park, H2 building, room two, Anhui Patentee before: Anhui Puji Information Technology Co.,Ltd. |
|
| CP01 | Change in the name or title of a patent holder | ||
| CP03 | Change of name, title or address |
Address after: 230088 floor 23-24, building A5, No. 666, Wangjiang West Road, high tech Zone, Hefei, Anhui Province Patentee after: Anhui Xunfei Medical Co.,Ltd. Address before: Room 288, H2 / F, phase II, innovation industrial park, 2800 innovation Avenue, high tech Zone, Hefei City, Anhui Province, 230088 Patentee before: ANHUI IFLYTEK MEDICAL INFORMATION TECHNOLOGY CO.,LTD. |
|
| CP03 | Change of name, title or address | ||
| CP01 | Change in the name or title of a patent holder |
Address after: 230088 floor 23-24, building A5, No. 666, Wangjiang West Road, high tech Zone, Hefei, Anhui Province Patentee after: IFLYTEK Medical Technology Co.,Ltd. Address before: 230088 floor 23-24, building A5, No. 666, Wangjiang West Road, high tech Zone, Hefei, Anhui Province Patentee before: Anhui Xunfei Medical Co.,Ltd. |
|
| CP01 | Change in the name or title of a patent holder |