CN110400560A - Data processing method and device, storage medium, electronic device - Google Patents
Data processing method and device, storage medium, electronic device Download PDFInfo
- Publication number
- CN110400560A CN110400560A CN201910673507.6A CN201910673507A CN110400560A CN 110400560 A CN110400560 A CN 110400560A CN 201910673507 A CN201910673507 A CN 201910673507A CN 110400560 A CN110400560 A CN 110400560A
- Authority
- CN
- China
- Prior art keywords
- model
- testing data
- tone testing
- data
- input
- 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.)
- Granted
Links
- 238000003672 processing method Methods 0.000 title claims abstract description 10
- 238000012360 testing method Methods 0.000 claims abstract description 187
- 238000000034 method Methods 0.000 claims abstract description 22
- 238000012549 training Methods 0.000 claims description 27
- 238000006243 chemical reaction Methods 0.000 claims description 12
- 238000004590 computer program Methods 0.000 claims description 12
- 238000012545 processing Methods 0.000 claims description 11
- 241001672694 Citrus reticulata Species 0.000 claims description 7
- 230000001537 neural effect Effects 0.000 claims description 3
- 238000012812 general test Methods 0.000 claims 1
- 238000010586 diagram Methods 0.000 description 4
- 210000004218 nerve net Anatomy 0.000 description 4
- 238000013528 artificial neural network Methods 0.000 description 3
- 230000000694 effects Effects 0.000 description 3
- 230000006870 function Effects 0.000 description 3
- 230000005540 biological transmission Effects 0.000 description 2
- 238000004364 calculation method Methods 0.000 description 2
- 230000005611 electricity Effects 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000003014 reinforcing effect Effects 0.000 description 1
- 230000003252 repetitive effect Effects 0.000 description 1
Classifications
-
- 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
- G10L15/00—Speech recognition
- G10L15/06—Creation of reference templates; Training of speech recognition systems, e.g. adaptation to the characteristics of the speaker's voice
- G10L15/063—Training
-
- 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
- G10L15/00—Speech recognition
- G10L15/08—Speech classification or search
- G10L15/16—Speech classification or search using artificial neural networks
Landscapes
- Engineering & Computer Science (AREA)
- Artificial Intelligence (AREA)
- Computational Linguistics (AREA)
- Health & Medical Sciences (AREA)
- Audiology, Speech & Language Pathology (AREA)
- Human Computer Interaction (AREA)
- Physics & Mathematics (AREA)
- Acoustics & Sound (AREA)
- Multimedia (AREA)
- Evolutionary Computation (AREA)
- Telephonic Communication Services (AREA)
- Electrically Operated Instructional Devices (AREA)
Abstract
The present invention provides a kind of data processing method and device, storage medium, electronic equipments, wherein, the above method includes: that the first tone testing data that will acquire are input in the first model, wherein first model is used to the first tone testing data being converted to the second tone testing data；Obtain the second tone testing data of the first model output；The second tone testing data are input in the second model, to indicate that second model is trained according to parameter of the second tone testing data to second model, wherein, second model is for identifying voice messaging, the voice messaging includes: the first tone testing data, the second tone testing data.
Description
Technical field
The present invention relates to computer fields, in particular to a kind of data processing method and device, storage medium, electricity
Sub-device.
Background technique
In the related technology, the mark of speech recognition training data is at high cost, but it is relatively easy to acquire data.Assuming that existing
The mandarin pronunciation data that marks of a batch, but using the trained speech recognition system of the voice data for the language that has an accent
Sound data discrimination is not high.However acquire the data having an accent and it is labeled, training one is directed to the language of this kind of accent
Sound identifying system cost is relatively high.
For in the related technology, during training speech model, for two kinds of tone testing data, model can not effectively into
The problems such as row identification, not yet there is effective solution at present.
Summary of the invention
The embodiment of the invention provides a kind of data processing method and device, storage medium, electronic devices, to solve voice
In identifying system, training is directed to the problems such as speech recognition system having an accent is prohibitively expensive.According to one embodiment of present invention,
Provide a kind of data processing method, comprising: the first tone testing data that will acquire are input in the first model, wherein
First model is used to the first tone testing data being converted to the second tone testing data；Obtain the first model output
The second tone testing data；The second tone testing data are input in the second model, to indicate second model
Be trained according to parameter of the second tone testing data to second model, wherein second model for pair
Voice messaging is identified that the voice messaging includes: the first tone testing data, the second tone testing data.
In embodiments of the present invention, it obtains first model and is directed to the third that the first tone testing data are exported
Tone testing data；Using the third tone testing data as the input of first model, so that first model is defeated
To the specified content in the second tone testing data of the second model, accounting is more than pre- in the second tone testing data out
If threshold value.
In embodiments of the present invention, the second tone testing data are input in the second model, to indicate described
After two models are trained according to parameter of the second tone testing data to second model, the method is also wrapped
It includes: the second model corresponding to the parameter after determining training；The second model is to language according to corresponding to the parameter after the training
Message breath is identified, recognition result is obtained；Show the recognition result.
In embodiments of the present invention, the first model includes: Feature Conversion network；Second model includes: two classification nerve nets
Network.
In embodiments of the present invention, the first tone testing data include: test data corresponding to standard mandarin voice,
Second tone testing data include: test data corresponding to non-standard mandarin.
According to another embodiment of the invention, a kind of data processing equipment is additionally provided, comprising: the first input module,
The first tone testing data for will acquire are input in the first model, wherein first model is used for the first language
Sound test data conversion is the second tone testing data；First obtains module, for obtaining the second of the first model output
Tone testing data；Second input module, for the second tone testing data to be input in the second model, to indicate
It states the second model to be trained according to parameter of the second tone testing data to second model, wherein described second
For model for identifying to voice messaging, the voice messaging includes: the first tone testing data, second voice
Test data.
In embodiments of the present invention, described device further include: second obtains module, is directed to for obtaining first model
The third tone testing data that the first tone testing data are exported；Processing module is used for the third tone testing
Input of the data as first model, so that first model is output in the second tone testing data of the second model
Specified content in the second tone testing data accounting be more than preset threshold.
In embodiments of the present invention, described device further include: determining module, for determining corresponding to the parameter after training
Second model；Identification module identifies voice messaging for the second model according to corresponding to the parameter after the training,
Obtain recognition result；Display module, for showing the recognition result.
According to still another embodiment of the invention, a kind of storage medium is additionally provided, meter is stored in the storage medium
Calculation machine program, wherein the computer program is arranged to execute the step in any of the above-described embodiment of the method when operation.
According to another embodiment of the invention, a kind of electronic device, including memory and processor are additionally provided, it is special
Sign is, computer program is stored in the memory, and the processor is arranged to run the computer program to hold
Element processing method described in row any of the above item.
Through the invention, the first tone testing data that will acquire are input in the first model, wherein first mould
Type is used to the first tone testing data being converted to the second tone testing data；Obtain the second voice of the first model output
Test data；The second tone testing data are input in the second model, to indicate second model according to described
Two tone testing data are trained the parameter of second model, wherein second model be used for voice messaging into
Row identification, the voice messaging includes: the first tone testing data, the second tone testing data, using above-mentioned skill
Art scheme solves in the related technology, and during training speech model, for two kinds of tone testing data, model can not be effective
First tone testing data can be converted to the second tone testing data by the first model by the problems such as being identified, make
One tone testing data and the second tone testing data have similitude, then using the second tone testing data after conversion into
Row training can avoid being labeled the second tone testing data by adopting the above technical scheme, not only reduce to the second voice
The cost that test data is labeled, and may be implemented to have the first tone testing data and the second tone testing data
Effect identification.
Detailed description of the invention
The drawings described herein are used to provide a further understanding of the present invention, constitutes part of this application, this hair
Bright illustrative embodiments and their description are used to explain the present invention, and are not constituted improper limitations of the present invention.In the accompanying drawings:
Fig. 1 is a kind of flow chart of optional data processing method according to an embodiment of the present invention；
Fig. 2 is a kind of flow chart of optional speech recognition system training method according to an embodiment of the present invention；
Fig. 3 is a kind of structural block diagram of optional data processing equipment according to an embodiment of the present invention；
Fig. 4 is a kind of another structural block diagram of optional data processing equipment according to an embodiment of the present invention.
Specific embodiment
Hereinafter, the present invention will be described in detail with reference to the accompanying drawings and in combination with Examples.It should be noted that not conflicting
In the case of, the features in the embodiments and the embodiments of the present application can be combined with each other.
It should be noted that description and claims of this specification and term " first " in above-mentioned attached drawing, "
Two " etc. be to be used to distinguish similar objects, without being used to describe a particular order or precedence order.
Fig. 1 is a kind of flow chart of optional data processing method according to an embodiment of the present invention, as shown in Figure 1, the stream
Journey includes the following steps:
Step S102, the first tone testing data that will acquire are input in the first model, wherein first model
For the first tone testing data to be converted to the second tone testing data；
Step S104 obtains the second tone testing data of the first model output；
The second tone testing data are input in the second model by step S106, to indicate the second model root
It is trained according to parameter of the second tone testing data to second model, wherein second model is used for language
Message breath is identified that the voice messaging includes: the first tone testing data, the second tone testing data.
Through the invention, the first tone testing data that will acquire are input in the first model, wherein first mould
Type is used to the first tone testing data being converted to the second tone testing data；Obtain the second voice of the first model output
Test data；The second tone testing data are input in the second model, to indicate second model according to described
Two tone testing data are trained the parameter of second model, wherein second model be used for voice messaging into
Row identification, the voice messaging includes: the first tone testing data, the second tone testing data, using above-mentioned skill
Art scheme solves in the related technology, and during training speech model, for two kinds of tone testing data, model can not be effective
First tone testing data can be converted to the second tone testing data by the first model by the problems such as being identified, make
One tone testing data and the second tone testing data have similitude, then using the second tone testing data after conversion into
Row training can avoid being labeled the second tone testing data by adopting the above technical scheme, not only reduce to the second voice
The cost that test data is labeled, and may be implemented to have the first tone testing data and the second tone testing data
Effect identification.
In embodiments of the present invention, before the second tone testing data being input in the second model, the method
Further include: it obtains first model and is directed to the third tone testing data that the first tone testing data are exported；By institute
Input of the third tone testing data as first model is stated, so that first model is output to the second of the second model
Specified content in the tone testing data accounting in the second tone testing data is more than preset threshold.
Wherein, the first tone testing data (such as can be standard accent data) are input in the first model, can be obtained
To above-mentioned third voice data (such as can be the data that have an accent), until obtained third tone testing data (such as can be
Accent data) it is more than preset threshold or more (such as can be 95% or more), then it represents that and third tone testing data at this time can be with
It is input in the second model as the second tone testing data.
In embodiments of the present invention, the second tone testing data are input in the second model, to indicate described
After two models are trained according to parameter of the second tone testing data to second model, the method is also wrapped
It includes: the second model corresponding to the parameter after determining training；The second model is to language according to corresponding to the parameter after the training
Message breath is identified, recognition result is obtained；Show the recognition result.
In embodiments of the present invention, the first model includes: Feature Conversion network；Second model includes: two classification nerve nets
Network.
In embodiments of the present invention, the first tone testing data include: test data corresponding to standard mandarin voice,
Second tone testing data include: test data corresponding to non-standard mandarin.
Optionally, the first tone testing data that will acquire are input in the first model, comprising: obtain default region
Tone testing data are as the first tone testing data；The the first tone testing data that will acquire are input to the first model
In.
Wherein, the default region can be the region of the mandarins standards of comparison such as Beijing.
Above-mentioned data handling procedure is explained below in conjunction with an example, but is not used in the restriction embodiment of the present invention
Technical solution, the exemplary technical solution of the present invention is as follows:
Fig. 2 is a kind of flow chart of optional speech recognition system training method according to an embodiment of the present invention, such as Fig. 2 institute
Show, which includes:
Step 1, using standard accent data and the data that have an accent, one two Classification Neural of training, the two classification mind
It can be deep neural network (Deep Neural Network, abbreviation DNN) through network.Wherein, standard accent data correspond to
Above-mentioned first tone testing data；The data that have an accent correspond to above-mentioned second tone testing data.
Step 2, using standard accent data one Feature Conversion network of training, using the output of the network as two classification minds
Input through network (such as DNN).Then, the parameter of continuous repetitive exercise this feature switching network inputs standard accent data
This feature switching network after iteration, obtains the data that have an accent, until the probability of the obtained data that have an accent reaches 95% (i.e.
Above-mentioned preset threshold) more than, then stop iteration.It should be noted that the process only trains the parameter of this feature switching network, no
The parameter of two Classification Neurals of training.Wherein, features described above switching network can be understood as a kind of neural network, may be implemented
First tone testing data are converted to the function of the second tone testing data, it can realization is converted to standard accent data
The function for the data that have an accent.
Step 3, standard accent data are inputted into trained Feature Conversion network, output it as speech recognition system
Feature, which is trained.And the speech recognition system is applied in the scene having an accent and is identified, is identified
As a result.
It can avoid being labeled the data that have an accent by adopting the above technical scheme, by carrying out standard accent data characteristics
Strengthen, makes its feature and the data that have an accent are with high similitude, be trained using the data that have an accent after reinforcing, not only subtracted
Lack and the data that have an accent are marked with prohibitively expensive problem, and has improved speech recognition system to the robust for the data that have an accent
Property.
Through the above description of the embodiments, those skilled in the art can be understood that according to above-mentioned implementation
The method of example can be realized by means of software and necessary general hardware platform, naturally it is also possible to by hardware, but it is very much
In the case of the former be more preferably embodiment.Based on this understanding, technical solution of the present invention is substantially in other words to existing
The part that technology contributes can be embodied in the form of software products, which is stored in a storage
In medium (such as ROM/RAM, magnetic disk, CD), including some instructions are used so that a terminal device (can be mobile phone, calculate
Machine, server or network equipment etc.) execute method described in each embodiment of the present invention.
A kind of data processing equipment is additionally provided in the present embodiment, and the data processing equipment is for realizing above-described embodiment
And preferred embodiment, the descriptions that have already been made will not be repeated.As used below, term " module " may be implemented to make a reservation for
The combination of the software and/or hardware of function.It is hard although device described in following embodiment is preferably realized with software
The realization of the combination of part or software and hardware is also that may and be contemplated.
Fig. 3 is a kind of structural block diagram of optional data processing equipment according to an embodiment of the present invention, as shown in figure 3, should
Device includes:
First input module 30, the first tone testing data for will acquire are input in the first model, wherein institute
The first model is stated for the first tone testing data to be converted to the second tone testing data；
First obtains module 32, for obtaining the second tone testing data of the first model output；
Second input module 34, for the second tone testing data to be input in the second model, described in instruction
Second model is trained according to parameter of the second tone testing data to second model, wherein second mould
For type for identifying to voice messaging, the voice messaging includes: the first tone testing data, and second voice is surveyed
Try data.
Through the invention, the first tone testing data that will acquire are input in the first model, wherein first mould
Type is used to the first tone testing data being converted to the second tone testing data；Obtain the second voice of the first model output
Test data；The second tone testing data are input in the second model, to indicate second model according to described
Two tone testing data are trained the parameter of second model, wherein second model be used for voice messaging into
Row identification, the voice messaging includes: the first tone testing data, the second tone testing data, using above-mentioned skill
Art scheme solves in the related technology, and during training speech model, for two kinds of tone testing data, model can not be effective
First tone testing data can be converted to the second tone testing data by the first model by the problems such as being identified, make
One tone testing data and the second tone testing data have similitude, then using the second tone testing data after conversion into
Row training can avoid being labeled the second tone testing data by adopting the above technical scheme, not only reduce to the second voice
The cost that test data is labeled, and may be implemented to have the first tone testing data and the second tone testing data
Effect identification.
In embodiments of the present invention, Fig. 4 is a kind of the another of optional data processing equipment according to an embodiment of the present invention
Structural block diagram, as shown in figure 4, described device further include:
Second obtains module 36, the exported for obtaining first model for the first tone testing data
Three tone testing data；
Processing module 38, for using the third tone testing data as the input of first model, so that described
First model is output to the specified content in the second tone testing data of the second model in the second tone testing data
Accounting is more than preset threshold.
In embodiments of the present invention, as shown in figure 4, described device further include:
Determining module 40, for determining the second model corresponding to the parameter after training；
Identification module 42 knows voice messaging for the second model according to corresponding to the parameter after the training
Not, recognition result is obtained；
Display module 44, for showing the recognition result.
In embodiments of the present invention, the first model includes: Feature Conversion network；Second model includes: two classification nerve nets
Network.
In embodiments of the present invention, the first model includes: Feature Conversion network；Second model includes: two classification nerve nets
Network.
It should be noted that above-mentioned modules can be realized by software or hardware, for the latter, Ke Yitong
Following manner realization is crossed, but not limited to this: above-mentioned module is respectively positioned in same processor；Alternatively, above-mentioned modules are with any
Combined form is located in different processors.
The embodiments of the present invention also provide a kind of storage medium, computer program is stored in the storage medium, wherein
The computer program is arranged to execute the step in any of the above-described embodiment of the method when operation.
Optionally, in the present embodiment, above-mentioned storage medium can be set to store by executing based on following steps
Calculation machine program:
S1, the first tone testing data that will acquire are input in the first model, wherein first model is used for will
First tone testing data are converted to the second tone testing data；
S2 obtains the second tone testing data of the first model output；
The second tone testing data are input in the second model by S3, to indicate second model according to
Second tone testing data are trained the parameter of second model, wherein second model is used for voice messaging
It is identified, the voice messaging includes: the first tone testing data, the second tone testing data.
Optionally, in the present embodiment, above-mentioned storage medium can include but is not limited to: USB flash disk, read-only memory (Read-
Only Memory, referred to as ROM), it is random access memory (Random Access Memory, referred to as RAM), mobile hard
The various media that can store computer program such as disk, magnetic or disk.
The embodiments of the present invention also provide a kind of electronic device, including memory and processor, stored in the memory
There is computer program, which is arranged to run computer program to execute the step in any of the above-described embodiment of the method
Suddenly.
Optionally, above-mentioned electronic device can also include transmission device and input-output equipment, wherein the transmission device
It is connected with above-mentioned processor, which connects with above-mentioned processor.
Optionally, in the present embodiment, above-mentioned processor can be set to execute following steps by computer program:
S1, the first tone testing data that will acquire are input in the first model, wherein first model is used for will
First tone testing data are converted to the second tone testing data；
S2 obtains the second tone testing data of the first model output；
The second tone testing data are input in the second model by S3, to indicate second model according to
Second tone testing data are trained the parameter of second model, wherein second model is used for voice messaging
It is identified, the voice messaging includes: the first tone testing data, the second tone testing data.
Obviously, those skilled in the art should be understood that each module of the above invention or each step can be with general
Computing device realize that they can be concentrated on a single computing device, or be distributed in multiple computing devices and formed
Network on, optionally, they can be realized with the program code that computing device can perform, it is thus possible to which they are stored
It is performed by computing device in the storage device, and in some cases, it can be to be different from shown in sequence execution herein
Out or description the step of, perhaps they are fabricated to each integrated circuit modules or by them multiple modules or
Step is fabricated to single integrated circuit module to realize.In this way, the present invention is not limited to any specific hardware and softwares to combine.
The foregoing is only a preferred embodiment of the present invention, is not intended to restrict the invention, for the skill of this field
For art personnel, the invention may be variously modified and varied.It is all within principle of the invention, it is made it is any modification, etc.
With replacement, improvement etc., should all be included in the protection scope of the present invention.
Claims (10)
1. a kind of data processing method characterized by comprising
The the first tone testing data that will acquire are input in the first model, wherein first model is used for the first language
Sound test data conversion is the second tone testing data；
Obtain the second tone testing data of the first model output；
The second tone testing data are input in the second model, to indicate second model according to second voice
Test data is trained the parameter of second model, wherein second model for being identified to voice messaging,
The voice messaging includes: the first tone testing data, the second tone testing data.
2. the method according to claim 1, wherein the second tone testing data are input to the second model
In before, the method also includes:
It obtains first model and is directed to the third tone testing data that the first tone testing data are exported；
Using the third tone testing data as the input of first model, so that first model is output to the second mould
Specified content in second tone testing data of the type accounting in the second tone testing data is more than preset threshold.
3. the method according to claim 1, wherein the second tone testing data are input to the second model
In, to indicate that second model is trained it according to parameter of the second tone testing data to second model
Afterwards, the method also includes:
Second model corresponding to parameter after determining training；
The second model according to corresponding to the parameter after the training identifies voice messaging, obtains recognition result；
Show the recognition result.
4. method according to any one of claims 1 to 3, which is characterized in that the first model includes: Feature Conversion network；
Second model includes: two Classification Neurals.
5. method according to any one of claims 1 to 3, which is characterized in that the first tone testing data include: that standard is general
Test data corresponding to call voice, the second tone testing data include: test data corresponding to non-standard mandarin.
6. a kind of data processing equipment characterized by comprising
First input module, the first tone testing data for will acquire are input in the first model, wherein described first
Model is used to the first tone testing data being converted to the second tone testing data；
First obtains module, for obtaining the second tone testing data of the first model output；
Second input module, for the second tone testing data to be input in the second model, to indicate second mould
Type is trained according to parameter of the second tone testing data to second model, wherein second model is used for
Voice messaging is identified, the voice messaging includes: the first tone testing data, the second tone testing number
According to.
7. device according to claim 6, described device further include:
Second obtains module, is directed to the third voice that the first tone testing data are exported for obtaining first model
Test data；
Processing module, for using the third tone testing data as the input of first model, so that first mould
It is super that type is output to the accounting in the second tone testing data of the specified content in the second tone testing data of the second model
Cross preset threshold.
8. device according to claim 6, described device further include:
Determining module, for determining the second model corresponding to the parameter after training；
Identification module identifies voice messaging for the second model according to corresponding to the parameter after the training, obtains
Recognition result；
Display module, for showing the recognition result.
9. a kind of storage medium, which is characterized in that be stored with computer program in the storage medium, wherein the computer
Program is arranged to execute method described in described any one of claims 1 to 5 when operation.
10. a kind of electronic device, including memory and processor, which is characterized in that be stored with computer journey in the memory
Sequence, the processor are arranged to run the computer program to execute side described in any one of claim 1 to 5
Method.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910673507.6A CN110400560B (en) | 2019-07-24 | 2019-07-24 | Data processing method and device, storage medium and electronic device |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910673507.6A CN110400560B (en) | 2019-07-24 | 2019-07-24 | Data processing method and device, storage medium and electronic device |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN110400560A true CN110400560A (en) | 2019-11-01 |
| CN110400560B CN110400560B (en) | 2022-10-18 |
Family
ID=68325913
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910673507.6A Active CN110400560B (en) | 2019-07-24 | 2019-07-24 | Data processing method and device, storage medium and electronic device |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110400560B (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111696551A (en) * | 2020-06-05 | 2020-09-22 | 海尔优家智能科技（北京）有限公司 | Device control method, device, storage medium, and electronic apparatus |
| CN111916105A (en) * | 2020-07-15 | 2020-11-10 | 北京声智科技有限公司 | Voice signal processing method and device, electronic equipment and storage medium |
Citations (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US7457745B2 (en) * | 2002-12-03 | 2008-11-25 | Hrl Laboratories, Llc | Method and apparatus for fast on-line automatic speaker/environment adaptation for speech/speaker recognition in the presence of changing environments |
| CN104732976A (en) * | 2013-12-20 | 2015-06-24 | 上海伯释信息科技有限公司 | Voice recognition method for converting mandarin into dialects |
| CN106023985A (en) * | 2016-05-19 | 2016-10-12 | 北京捷通华声科技股份有限公司 | Linguistic model training method and system and speech recognition system |
| US20170371866A1 (en) * | 2016-06-27 | 2017-12-28 | Facebook, Inc. | Language model using reverse translations |
| CN109359309A (en) * | 2018-12-11 | 2019-02-19 | 成都金山互动娱乐科技有限公司 | A kind of interpretation method and device, the training method of translation model and device |
| CN109616101A (en) * | 2019-02-12 | 2019-04-12 | 百度在线网络技术（北京）有限公司 | Acoustic training model method, apparatus, computer equipment and readable storage medium storing program for executing |
| RU2017140974A3 (en) * | 2017-11-24 | 2019-05-24 | ||
| CN109859737A (en) * | 2019-03-28 | 2019-06-07 | 深圳市升弘创新科技有限公司 | Communication encryption method, system and computer readable storage medium |
| CN109887497A (en) * | 2019-04-12 | 2019-06-14 | 北京百度网讯科技有限公司 | Modeling method, device and the equipment of speech recognition |
| CN109949808A (en) * | 2019-03-15 | 2019-06-28 | 上海华镇电子科技有限公司 | The speech recognition appliance control system and method for compatible mandarin and dialect |
-
2019
- 2019-07-24 CN CN201910673507.6A patent/CN110400560B/en active Active
Patent Citations (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US7457745B2 (en) * | 2002-12-03 | 2008-11-25 | Hrl Laboratories, Llc | Method and apparatus for fast on-line automatic speaker/environment adaptation for speech/speaker recognition in the presence of changing environments |
| CN104732976A (en) * | 2013-12-20 | 2015-06-24 | 上海伯释信息科技有限公司 | Voice recognition method for converting mandarin into dialects |
| CN106023985A (en) * | 2016-05-19 | 2016-10-12 | 北京捷通华声科技股份有限公司 | Linguistic model training method and system and speech recognition system |
| US20170371866A1 (en) * | 2016-06-27 | 2017-12-28 | Facebook, Inc. | Language model using reverse translations |
| RU2017140974A3 (en) * | 2017-11-24 | 2019-05-24 | ||
| CN109359309A (en) * | 2018-12-11 | 2019-02-19 | 成都金山互动娱乐科技有限公司 | A kind of interpretation method and device, the training method of translation model and device |
| CN109616101A (en) * | 2019-02-12 | 2019-04-12 | 百度在线网络技术（北京）有限公司 | Acoustic training model method, apparatus, computer equipment and readable storage medium storing program for executing |
| CN109949808A (en) * | 2019-03-15 | 2019-06-28 | 上海华镇电子科技有限公司 | The speech recognition appliance control system and method for compatible mandarin and dialect |
| CN109859737A (en) * | 2019-03-28 | 2019-06-07 | 深圳市升弘创新科技有限公司 | Communication encryption method, system and computer readable storage medium |
| CN109887497A (en) * | 2019-04-12 | 2019-06-14 | 北京百度网讯科技有限公司 | Modeling method, device and the equipment of speech recognition |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111696551A (en) * | 2020-06-05 | 2020-09-22 | 海尔优家智能科技（北京）有限公司 | Device control method, device, storage medium, and electronic apparatus |
| CN111916105A (en) * | 2020-07-15 | 2020-11-10 | 北京声智科技有限公司 | Voice signal processing method and device, electronic equipment and storage medium |
Also Published As
| Publication number | Publication date |
|---|---|
| CN110400560B (en) | 2022-10-18 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108830235A (en) | Method and apparatus for generating information | |
| CN108021806B (en) | Malicious installation package identification method and device | |
| CN107645562A (en) | Data transmission processing method, device, equipment and system | |
| CN109976995B (en) | Method and apparatus for testing | |
| CN110544469B (en) | Training method and device of voice recognition model, storage medium and electronic device | |
| CN110362822A (en) | Text marking method, apparatus, computer equipment and storage medium for model training | |
| CN112860841A (en) | Text emotion analysis method, device and equipment and storage medium | |
| CN110096617B (en) | Video classification method and device, electronic equipment and computer-readable storage medium | |
| CN110046254A (en) | Method and apparatus for generating model | |
| CN111694937A (en) | Interviewing method and device based on artificial intelligence, computer equipment and storage medium | |
| CN111352846B (en) | Method, device, equipment and storage medium for manufacturing number of test system | |
| CN108281158A (en) | Voice biopsy method, server and storage medium based on deep learning | |
| CN108229535A (en) | Relate to yellow image audit method, apparatus, computer equipment and storage medium | |
| CN107729928A (en) | Information acquisition method and device | |
| CN109685068A (en) | A kind of image processing method and system based on generation confrontation neural network | |
| CN110400560A (en) | Data processing method and device, storage medium, electronic device | |
| CN111061867A (en) | Text generation method, equipment, storage medium and device based on quality perception | |
| CN108614987A (en) | The method, apparatus and robot of data processing | |
| CN113343248A (en) | Vulnerability identification method, device, equipment and storage medium | |
| CN115205736A (en) | Video data identification method and device, electronic equipment and storage medium | |
| CN109829431A (en) | Method and apparatus for generating information | |
| CN108985178A (en) | Method and apparatus for generating information | |
| CN109766089B (en) | Code generation method and device based on dynamic diagram, electronic equipment and storage medium | |
| CN116633804A (en) | Modeling method, protection method and related equipment of network flow detection model | |
| CN113555037B (en) | Method and device for detecting tampered area of tampered audio and storage medium |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |