CN105895103A - Speech recognition method and device - Google Patents
Speech recognition method and device Download PDFInfo
- Publication number
- CN105895103A CN105895103A CN201510883295.6A CN201510883295A CN105895103A CN 105895103 A CN105895103 A CN 105895103A CN 201510883295 A CN201510883295 A CN 201510883295A CN 105895103 A CN105895103 A CN 105895103A
- Authority
- CN
- China
- Prior art keywords
- user
- information
- participle
- user profile
- terminal
- 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
- 238000000034 method Methods 0.000 title claims abstract description 30
- 238000012545 processing Methods 0.000 claims description 14
- 230000010365 information processing Effects 0.000 claims description 8
- 239000000203 mixture Substances 0.000 claims description 3
- 230000008878 coupling Effects 0.000 claims description 2
- 238000010168 coupling process Methods 0.000 claims description 2
- 238000005859 coupling reaction Methods 0.000 claims description 2
- 230000008569 process Effects 0.000 description 8
- 238000004891 communication Methods 0.000 description 4
- 238000005516 engineering process Methods 0.000 description 4
- 238000001228 spectrum Methods 0.000 description 4
- 238000012549 training Methods 0.000 description 3
- 238000010586 diagram Methods 0.000 description 2
- 230000007613 environmental effect Effects 0.000 description 2
- 238000012216 screening Methods 0.000 description 2
- 208000031481 Pathologic Constriction Diseases 0.000 description 1
- 230000009471 action Effects 0.000 description 1
- 230000008901 benefit Effects 0.000 description 1
- 230000008859 change Effects 0.000 description 1
- 238000004140 cleaning Methods 0.000 description 1
- 238000000605 extraction Methods 0.000 description 1
- 230000006870 function Effects 0.000 description 1
- 238000012905 input function Methods 0.000 description 1
- 230000003993 interaction Effects 0.000 description 1
- 230000002452 interceptive effect Effects 0.000 description 1
- 238000013507 mapping Methods 0.000 description 1
- 235000012054 meals Nutrition 0.000 description 1
- 230000008439 repair process Effects 0.000 description 1
- 238000013519 translation Methods 0.000 description 1
- 210000001215 vagina Anatomy 0.000 description 1
Classifications
-
- G—PHYSICS
- G10—MUSICAL INSTRUMENTS; ACOUSTICS
- G10L—SPEECH ANALYSIS OR SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING; SPEECH OR AUDIO CODING OR DECODING
- G10L15/00—Speech recognition
- G10L15/28—Constructional details of speech recognition systems
- G10L15/30—Distributed recognition, e.g. in client-server systems, for mobile phones or network applications
-
- G—PHYSICS
- G10—MUSICAL INSTRUMENTS; ACOUSTICS
- G10L—SPEECH ANALYSIS OR SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING; SPEECH OR AUDIO CODING OR DECODING
- G10L15/00—Speech recognition
- G10L15/08—Speech classification or search
- G10L15/10—Speech classification or search using distance or distortion measures between unknown speech and reference templates
-
- G—PHYSICS
- G10—MUSICAL INSTRUMENTS; ACOUSTICS
- G10L—SPEECH ANALYSIS OR SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING; SPEECH OR AUDIO CODING OR DECODING
- G10L15/00—Speech recognition
- G10L15/22—Procedures used during a speech recognition process, e.g. man-machine dialogue
Abstract
The invention provides a speech recognition method and device. Speech information transmitted by a terminal is received. The acoustic characteristic information of the speech information is acquired. The acoustic characteristic information is sequentially input into an acoustic model and a language model. The acoustic model and the language model recognize the speech information to acquire initial text information. Based on pre-stored user information, the initial text information is corrected to generate final text information. According to the technical scheme provided by the invention, the recognized initial text information is corrected; errors in the initial text information are corrected; the corrected final text information is sent to the terminal; and the terminal can provide accurate services for a user according to the accurate final text information.
Description
Technical field
The present embodiments relate to speech signal analysis technical field, particularly relate to a kind of audio recognition method
And device.
Background technology
Speech recognition technology is to allow machine convert voice signals into corresponding life by identifying with understanding process
Order or the technology of text.At present, speech recognition technology is widely used in the language such as speech control, voiced translation
Sound interactive product.
At present, multiple terminal possesses speech voice input function, and the various application softwaries being arranged in terminal are both needed to
To perform corresponding operation based on voice identification result, thus to generate the information required for user, present to
User.When the speech recognition of terminal is preferable, it is possible to identify the voice messaging of user's input, ability exactly
The service that enough guarantees are supplied to user is more accurate.Such as, terminal comprises map application software, user
The route between current location to expectation place can be obtained from by this map application software；Such as, when
User thinks in " xx restaurant, Beijing ", and terminal receives the voice messaging of user's input, i.e. to user's input
Voice messaging is identified, and obtains the text message in " xx restaurant, Beijing ", and map application software is on map
The text message in " xx restaurant, Beijing " is scanned for, and according to the current position of user, planning user works as
The route in front position to " xx restaurant, Beijing "；But when Beijing comprises at least two restaurant title, pronunciation
When being the phonetic of " xx restaurant " correspondence, then map application software will present the knowledge of multiple text message
Other result, or, acquiescence is presented " the xx meal that distance users current location is nearest by map application software
Shop ", now, user needs to carry out the Search Results presented manual screening, map application software according to
The result of family manual screening, carries out route planning, or, terminal will present the route of mistake.
As can be seen here, current voice identification result, there is the problem that error rate is high.
Summary of the invention
The embodiment of the present invention provides a kind of audio recognition method and device, in order to solve current speech recognition knot
Really, there is the problem that error rate is high.
The concrete technical scheme that the embodiment of the present invention provides is as follows:
The embodiment of the present invention provides a kind of audio recognition method, including:
Receive the VoP that terminal sends；Wherein, described VoP comprises voice messaging；
Obtain the acoustic features information of described voice messaging；Wherein, described acoustic features information is for characterizing institute
State the information of the sound property of voice messaging；
Described acoustic features information is sequentially input default acoustic model and language model, obtains described
Voice messaging is identified the original text information obtained；
According to the user profile prestored, it is modified described original text information processing, generates final literary composition
This information；
Described final text message is sent to described terminal.
The embodiment of the present invention provides a kind of speech recognition equipment, including:
Receive unit, for receiving the VoP that terminal sends；Wherein, in described VoP
Comprise voice messaging；
Acoustic features information acquisition unit, for obtaining the acoustic features information of described voice messaging；Its
In, described acoustic features information is the information of the sound property characterizing described voice messaging；
Original text information acquisition unit, for sequentially inputting default acoustics by described acoustic features information
Model and language model, obtain the original text information being identified obtaining to described voice messaging；
Final text message signal generating unit, for according to the user profile prestored, believing described original text
Breath is modified processing, and generates final text message；
Transmitting element, for sending described final text message to described terminal.
The embodiment of the present invention provides a kind of audio recognition method and device, by receiving the voice that terminal sends
Information, obtains the acoustic features information of described voice messaging；Described acoustic features information is sequentially input sound
Learn model and language model, obtain described acoustic model and described voice messaging is carried out by described speech model
Identify the original text information obtained；According to the user profile prestored, described original text information is carried out
Correcting process, generates final text message.Use embodiment of the present invention technical scheme, for the most identified
The original text information obtained is modified processing, to repair the mistake in described original text information
Just, the final text message that will generate after revising sends to described terminal, makes terminal more can be as the criterion to basis
True final text message, provides a user with and services the most accurately.
Accompanying drawing explanation
Fig. 1 is speech recognition system configuration diagram in the embodiment of the present invention；
Fig. 2 is speech recognition flow chart in the embodiment of the present invention one；
Fig. 3 is that the present invention implements two example data base's Establishing process figures；
Fig. 4 is speech recognition equipment structural representation in the embodiment of the present invention three.
Detailed description of the invention
For making the purpose of the embodiment of the present invention, technical scheme and advantage clearer, below in conjunction with this
Accompanying drawing in bright embodiment, is clearly and completely described the technical scheme in the embodiment of the present invention,
Obviously, described embodiment is a part of embodiment of the present invention rather than whole embodiments.Based on
Embodiment in the present invention, those of ordinary skill in the art are obtained under not making creative work premise
The every other embodiment obtained, broadly falls into the scope of protection of the invention.
Below in conjunction with Figure of description, the embodiment of the present invention is described in further detail.
Refering to shown in Fig. 1, in the embodiment of the present invention, speech recognition system configuration diagram, this voice
Identification system comprises terminal and server；Wherein, described terminal is the terminal possessing communication function, and institute
Stating terminal is the terminal possessing human-computer interaction interface, if described terminal is personal computer, and panel computer,
Mobile phones etc., can carry various operating system in described terminal, such as microsoft operation system, Android operation is
System, ios operating system etc., and described terminal can carry various with this terminal in the operating system installed
Compatible application software, such as map application software, chat tool application software etc.；Described service utensil
Standby speech recognition component, speech recognition correcting part, described speech recognition component is for sending out described terminal
The voice messaging sent is identified, and described speech recognition correcting part is for knowing described speech recognition component
Other result is modified；Further, described server also includes voiceprint service parts, TTS (Text
To Speech；From text to language), data, services parts, customer data base etc., wherein, described sound
Stricture of vagina Service Part, is analyzed for the voice messaging sending described terminal, obtains initial user letter
Breath, described TTS, for final text message being converted to voice messaging, described data, services parts,
Initial user information for obtaining described voiceprint service parts is analyzed, and obtains final user's letter
Breath, described data base is for storing the user profile that described data, services Component Analysis obtains and described
The terminal iidentification that user profile is corresponding.
Embodiment one
Refering to shown in Fig. 2, in the embodiment of the present invention, server carries out the process of speech recognition, including:
Step 200: receive the VoP that terminal sends；Wherein, described VoP comprises language
Message ceases.
In the embodiment of the present invention, terminal passes through voice collecting parts, calls SDK (Software
Development Kit；SDK) obtain the voice messaging that user inputs；Described terminal root
According to described voice messaging, generate VoP；And described VoP is sent to described service
Device.
Optionally, comprising cordless communication network between described terminal and described server, described terminal is passed through
The VoP comprising described voice messaging is sent to described server by described cordless communication network.
Further, after server receives the VoP that terminal sends, to the voice gathered
Information is removed noise processed, and to reject the interference factor in described voice messaging, this interference factor is
Such as background music during user input voice information, or background noise etc., thus ensure that acquisition is
The accuracy of whole text message.
Step 210: obtain the acoustic features information of described voice messaging；Wherein, described acoustic features information
For characterizing the information of the sound property of described voice messaging.
In the embodiment of the present invention, described voice messaging is resolved by the speech recognition component in server,
Obtain the acoustic features information comprised in described voice messaging；Wherein, described acoustic features information is one to be
Row spectrum information, reacts due to the pronunciation of each word or word and is acoustically being one section of frequency spectrum, no
The frequency spectrum corresponding with the word of pronunciation is different, and therefore, the sound that this spectrum information can characterize voice messaging is special
The information of property.
Step 220: described acoustic features information is sequentially input default acoustic model and language model, obtains
Take the original text information that described voice messaging is identified to be obtained.
In the embodiment of the present invention, the speech recognition component in server is by the most defeated for described acoustic features information
Enter default acoustic model and language model, obtain the original text letter that described language model identification obtains
Breath.
Optionally, the input of described acoustic features information is preset by the speech recognition component in described server
Acoustic model, obtains the pronunciation template identification of described acoustic model output；By defeated for described pronunciation template identification
Enter described language model, obtain the original text information of described language model output.Wherein, described acoustics
Model and described speech model adjust principle, Hidden Markov principle, or vector quantity according to dynamic time
Change principle, be trained obtaining to a large amount of training samples.
Concrete, described acoustic features information is comprised in described acoustic model by described acoustic model respectively
Each pronunciation template mate, and obtain described acoustic features information and comprise in described acoustic model
Each pronunciation template between distance, wherein, described acoustics template includes word pronunciation model, half syllable
Model or prime model；Described acoustic model, from all pronunciation templates, obtains and believes with described acoustic features
The pronunciation template of each pronunciation distance minimum comprised in breath；Due to the pronunciation template in acoustic model and institute
There are mapping relations in the text stated in language model, therefore, the mark of described pronunciation template is inputted described
Language model, described language model can obtain the text corresponding with the mark of described pronunciation template；
Optionally, described language model comprises multiple tree, each tree with each word or
Each pronunciation of person is root node, and each child node is the phrase that each word can make up；Due to often
One possible corresponding multiple text of pronunciation, therefore, described language model exports for described acoustic model
Each template identification of pronouncing, is performed both by operating as follows: inquire about corresponding each of this pronunciation template identification
Tree, and according to the pronunciation template identification after this pronunciation template identification, obtain this pronunciation template mark
The mark that pronunciation template identification after knowing corresponding text and this pronunciation template identification is corresponding；With this type of
Push away, obtain all texts that described voice messaging is corresponding, and according to described all texts, generate initial literary composition
This information.Wherein, described language model can export an original text information, it is also possible to exports multiple
Original text information.
Use technique scheme, owing to acoustic model and language model are according to entering a large amount of voice messagings
Row sentific training obtains, and therefore, voice messaging inputs described acoustic model and language model, it is possible to
Obtain original text information more accurately.
Step 230: according to the user profile prestored, is modified described original text information processing, raw
Become final text message.
In the embodiment of the present invention, the speech recognition correcting part in described server is from described customer data base
The user profile that middle extraction prestores；And according to the user profile prestored, described original text information is carried out
Correcting process；Wherein, described user profile is uploaded by terminal by user, and/or, by described server
Obtain according to the voice messaging of a large number of users is identified training.
Optionally, described in the acquisition methods of user profile that prestores, including: server obtains described voice
The mark of the terminal comprised in packet；The mark correspondence of described terminal is searched from user profile set
User profile；Wherein, described user profile includes the position of historical time point user, the year of described user
Age, or the sex of described user；Described user profile set comprises the mark of terminal and user profile
Corresponding relation.
Optionally, according to the user profile prestored, it is modified described original text information processing, raw
Become final text message, specifically include: described original text information is divided, obtains each point
Word；For the position participle in described participle, by described search from described user profile with described currently
The historical time point of time point coupling, and obtain the position of user described in the historical time point found, if
The position of the user of described acquisition is mated unsuccessful with described position participle in whole or in part, and institute's rheme
The position pronunciation similarity of the pronunciation and the user of described acquisition of putting participle reaches predetermined threshold value, then with described
Described position participle is replaced in the position of the user obtained；For the special participle in described participle, according to institute
State the age of user comprised in user profile or user's sex, described special participle is modified place
Reason；Wherein, described special participle is the participle that there is unisonance not synonym.
Optionally, described current point in time and described historical time Point matching, refer to described current point in time
And the time difference between described historical time point is less than Preset Time difference scope；This Preset Time difference scope root
Arrange according to concrete application scenarios.
Such as, when initial text message is " how going Quanjude road conditions ", due to Beijing, to comprise many families complete
Poly-moral, first server obtains the position participle " Quanjude " comprised in described original text information, clothes
It is 18:00 in afternoon that business device obtains current time, and server detects that user once had three times on a 18:10 left side
The right side is positioned at He Ping Men Quanjude shop, and therefore, what server will default to user's search is that " He Ping Men gathers entirely
Moral ", described original text Information revision is " how going He Ping Men Quanjude road conditions " by server.
For another example, when initial text message is " traffic how ", and server is by this original text of acquiescence
Comprising position participle in information, it is 18:00 in afternoon that server obtains current time, and server detects use
Family is respectively positioned on " xx community " about this time point, and therefore, described original text information is repaiied by server
Just for " how going xx cordon traffic situation ".
The most such as, when initial text message is " Yuxi how ", owing to " Yuxi " exists phonetically similar word
" Yue-Sai ", therefore, server obtains age and the sex of described user, when the age of described user is
20-26, when the sex of described user is women, described original text Information revision is by described server
" Yue-Sai how ".
Further, when the number of described process text message is multiple, server can use above-mentioned
Mode, screens original text information the most accurately from described original text information, and chooses described
Original text information be modified.
Further, described server can also be according to the class of the application software sending described VoP
Type, is modified described original text information；Such as, the voice messaging inputted as user is " Yue-Sai
How ", when the application software being currently running due to terminal is map application software, due to " Yue-Sai "
Not being a place name, therefore, described original text Information revision is " Yuxi how " by server.
Further, according to the user profile prestored, it is modified described original text information processing,
Generate final text message, also include: when this locality does not comprise the user profile of the mark correspondence of described terminal
Time, according to described acoustic features information, determine age and the sex of the user that described voice messaging is provided；
Age according to the user providing described voice messaging determined and sex, enter described original text information
Row correcting process, generates final text message.
Optionally, described acoustic features information, determine age and the property of the user that described voice messaging is provided
Not, specifically include: voiceprint service parts extract the biological attribute data in described acoustic features information, its
In, described biological attribute data comprises tone color, tonequality, tone, word speed etc.；Described voiceprint service parts
According to described biological attribute data, and described acoustic model, obtain age and the sex of described user.
Step 240: described final text message is sent to described terminal.
In the embodiment of the present invention, server by described final text message by cordless communication network send to
Described terminal.
Further, after generating final text message, described server can be by described final text envelope
Breath is converted to voice messaging；And described voice messaging is sent to described terminal, by described in terminal plays
Whole text message.
Further, after generating final text message, described server can be according to described final text
Information, obtains the service of described user request, and generates the data of the service correspondence that described user is asked
Bag sends to terminal.Wherein, described packet can be textual form, it is also possible to for speech form.
Use technique scheme, according to the customized information of user, for the most identified obtain initial
Text message is modified processing, and to be modified the mistake in described original text information, thus carries
The high accuracy of speech recognition；Further, the final text message generated after revising sent to described end
End, makes terminal can provide a user with take the most accurately to according to the most final text message
Business.
Embodiment two
Refering to shown in Fig. 3, in the embodiment of the present invention, the user profile comprised in the data base of server
Generation process, including:
Step 300: receive the VoP that terminal sends；Wherein, described VoP comprises language
Message ceases.
Step 310: obtain the acoustic features information comprised in described voice messaging.
Step 320: according to described acoustic features information, determine the age of the user that described voice messaging is provided
And sex, and final text message；According to determine provide described voice messaging user age and
Sex.
Optionally, server can also be according to described acoustic features information, acquisition environmental data, such as,
Time and user's sphere of action etc..
Step 330: age and the sex to the user determined, and finally text message is analyzed, and
According to analysis result, generate user profile.
Optionally, described server can also generate user profile according to described environmental data.
Step 340: set up the corresponding relation between the mark of described terminal, and the user profile generated, will
Described corresponding relation stores to described user profile set.
Embodiment three
Based on technique scheme, refering to shown in Fig. 4, in the embodiment of the present invention, it is provided that a kind of internal memory is empty
Between cleaning plant, including receive unit 40, acoustic features information acquisition unit 41, original text information obtains
Take unit 42, final text message signal generating unit 43, and transmitting element 44, wherein:
Receive unit 40, for receiving the VoP that terminal sends；Wherein, described VoP
In comprise voice messaging；
Acoustic features information acquisition unit 41, for obtaining the acoustic features information of described voice messaging；Its
In, described acoustic features information is the information of the sound property characterizing described voice messaging；
Original text information acquisition unit 42, for sequentially inputting default sound by described acoustic features information
Learn model and language model, obtain the original text information being identified described voice messaging obtaining；
Final text message signal generating unit 43, for according to the user profile prestored, to described original text
Information is modified processing, and generates final text message；
Transmitting element 44, for sending described final text message to described terminal.
Further, described VoP also comprises terminal iidentification；Also include that prestored information obtains single
Unit 45, is used for: search the user profile that the mark of described terminal is corresponding from user profile set；Its
In, described user profile includes the position of historical time point user, the age of described user, or described
The sex of user；Described user profile set comprises the mark of terminal and the corresponding relation of user profile.
Optionally, described original text information acquisition unit 42, specifically include: described acoustic features is believed
The acoustic model that breath input is preset, obtains the pronunciation template identification of described acoustic model output；By described
Sound template identification inputs described language model, obtains the original text information of described language model output.
Optionally, described final text message signal generating unit 43, specifically for: described original text is believed
Breath divides, and obtains each participle；For the position participle in described participle, by described from described
User profile is searched the historical time point with described current time Point matching, and obtains the history found
The position of user described in time point, if the position of the user of described acquisition and described position participle all or
Part is mated unsuccessful, and the pronunciation of described position participle is similar to the pronunciation of the position of the user of described acquisition
Degree reaches predetermined threshold value, then replace described position participle with the position of the user of described acquisition；For described
Special participle in participle, according to the age of user comprised in described user profile or user's sex, right
Described special participle is modified processing；Wherein, described special participle is to there is dividing of unisonance not synonym
Word.
Further, described final text message signal generating unit 43, it is additionally operable to: when this locality does not comprise described
When identifying corresponding user profile of terminal, according to described acoustic features information, determines the described voice of offer
The age of the user of information and sex；Age according to the user providing described voice messaging determined and property
, it is not modified described original text information processing, generates final text message.
Further, also include processing unit 46, be used for: after generating final text message, to determining
Age of user and sex, and final text message is analyzed, and according to analysis result, generates
User profile；Set up the corresponding relation between the mark of described terminal, and the user profile generated, by institute
State corresponding relation to store to described user profile set.
In sum, in the embodiment of the present invention, by receiving the voice messaging that terminal sends, obtain described
The acoustic features information of voice messaging；Described acoustic features information is sequentially input acoustic model and language mould
Type, obtain that described voice messaging is identified obtaining by described acoustic model and described speech model is initial
Text message；According to the user profile prestored, it is modified described original text information processing, generates
Final text message.Use embodiment of the present invention technical scheme, for the most identified original text obtained
Information is modified processing, to be modified the mistake in described original text information, raw after revising
The final text message become sends to described terminal, makes the terminal can be to according to the most final text envelope
Breath, provides a user with and services the most accurately.
Device embodiment described above is only schematically, wherein said illustrates as separating component
Unit can be or may not be physically separate, the parts shown as unit can be or
Person may not be physical location, i.e. may be located at a place, or can also be distributed to multiple network
On unit.Some or all of module therein can be selected according to the actual needs to realize the present embodiment
The purpose of scheme.Those of ordinary skill in the art are not in the case of paying performing creative labour, the most permissible
Understand and implement.
Through the above description of the embodiments, those skilled in the art is it can be understood that arrive each reality
The mode of executing can add the mode of required general hardware platform by software and realize, naturally it is also possible to by firmly
Part.Based on such understanding, the portion that prior art is contributed by technique scheme the most in other words
Dividing and can embody with the form of software product, this computer software product can be stored in computer can
Read in storage medium, such as ROM/RAM, magnetic disc, CD etc., including some instructions with so that one
Computer equipment (can be personal computer, server, or the network equipment etc.) performs each to be implemented
The method described in some part of example or embodiment.
Last it is noted that above example is only in order to illustrate the technical scheme of the embodiment of the present invention, and
Non-to its restriction；Although the embodiment of the present invention being described in detail with reference to previous embodiment, ability
The those of ordinary skill in territory is it is understood that it still can be to the technical scheme described in foregoing embodiments
Modify, or wherein portion of techniques feature is carried out equivalent；And these amendments or replacement,
The essence not making appropriate technical solution departs from spirit and the model of the embodiment of the present invention each embodiment technical scheme
Enclose.
Claims (12)
1. an audio recognition method, it is characterised in that including:
Receive the VoP that terminal sends；Wherein, described VoP comprises voice messaging；
Obtain the acoustic features information of described voice messaging；Wherein, described acoustic features information is for characterizing institute
State the information of the sound property of voice messaging；
Described acoustic features information is sequentially input default acoustic model and language model, obtains described
Voice messaging is identified the original text information obtained；
According to the user profile prestored, it is modified described original text information processing, generates final literary composition
This information；
Described final text message is sent to described terminal.
Method the most according to claim 1, it is characterised in that also comprise in described VoP
Terminal iidentification；
The acquisition methods of the described user profile prestored, including:
The user profile that the mark of described terminal is corresponding is searched from user profile set；Wherein, described use
Family information includes the property of the position of historical time point user, the age of described user, or described user
Not；Described user profile set comprises the mark of terminal and the corresponding relation of user profile.
Method the most according to claim 2, it is characterised in that by described acoustic features information successively
Acoustic model that input is preset and language model, obtain be identified obtaining initial to described voice messaging
Text message, specifically includes:
The acoustic model input of described acoustic features information preset, obtains sending out of described acoustic model output
Sound template identification；
Described pronunciation template identification is inputted described language model, obtains the initial of described language model output
Text message.
The most according to the method in claim 2 or 3, it is characterised in that according to the user's letter prestored
Breath, is modified described original text information processing, generates final text message, specifically include:
Described original text information is divided, obtains each participle；For the position in described participle
Put participle, from described user profile, search the historical time with described current time Point matching by described
Point, and obtain the position of user described in the historical time point found, if the position of the user of described acquisition
Mate unsuccessful in whole or in part with described position participle, and the pronunciation of described position participle obtains with described
The position pronunciation similarity of the user taken reaches predetermined threshold value, then replace with the position of the user of described acquisition
Described position participle；For the special participle in described participle, according to the use comprised in described user profile
Family age or user's sex, be modified described special participle processing；Wherein, described special participle
For there is the participle of unisonance not synonym.
Method the most according to claim 4, it is characterised in that according to the user profile prestored, right
Described original text information is modified processing, and generates final text message, also includes:
When this locality does not comprise the user profile of mark correspondence of described terminal, believe according to described acoustic features
Breath, determines age and the sex of the user providing described voice messaging；
Age according to the user providing described voice messaging determined and sex, believe described original text
Breath is modified processing, and generates final text message.
Method the most according to claim 5, it is characterised in that after generating final text message,
Described method also includes:
Age and sex to the user determined, and finally text message is analyzed, and according to analysis
As a result, user profile is generated；
Set up the corresponding relation between the mark of described terminal, and the user profile generated, by described correspondence
Relation stores to described user profile set.
7. a speech recognition equipment, it is characterised in that including:
Receive unit, for receiving the VoP that terminal sends；Wherein, in described VoP
Comprise voice messaging；
Acoustic features information acquisition unit, for obtaining the acoustic features information of described voice messaging；Its
In, described acoustic features information is the information of the sound property characterizing described voice messaging；
Original text information acquisition unit, for sequentially inputting default acoustics by described acoustic features information
Model and language model, obtain the original text information being identified obtaining to described voice messaging；
Final text message signal generating unit, for according to the user profile prestored, believing described original text
Breath is modified processing, and generates final text message；
Transmitting element, for sending described final text message to described terminal.
Device the most according to claim 7, it is characterised in that also comprise in described VoP
Terminal iidentification；
Also include prestored information acquiring unit, be used for:
The user profile that the mark of described terminal is corresponding is searched from user profile set；Wherein, described use
Family information includes the property of the position of historical time point user, the age of described user, or described user
Not；Described user profile set comprises the mark of terminal and the corresponding relation of user profile.
Device the most according to claim 8, it is characterised in that described original text acquisition of information list
Unit, specifically for:
The acoustic model input of described acoustic features information preset, obtains sending out of described acoustic model output
Sound template identification；
Described pronunciation template identification is inputted described language model, obtains the initial of described language model output
Text message.
Device the most according to claim 8 or claim 9, it is characterised in that described final text message
Signal generating unit, specifically for:
Described original text information is divided, obtains each participle；
For the position participle in described participle, by described search from described user profile with described currently
The historical time point of time point coupling, and obtain the position of user described in the historical time point found, if
The position of the user of described acquisition is mated unsuccessful with described position participle in whole or in part, and institute's rheme
The position pronunciation similarity of the pronunciation and the user of described acquisition of putting participle reaches predetermined threshold value, then with described
Described position participle is replaced in the position of the user obtained；
For the special participle in described participle, according to the age of user comprised in described user profile or
User's sex, is modified described special participle processing；Wherein, for there is unisonance in described special participle
The not participle of synonym.
11. devices according to claim 10, it is characterised in that described final text message generates
Unit, is additionally operable to:
When this locality does not comprise the user profile of mark correspondence of described terminal, believe according to described acoustic features
Breath, determines age and the sex of the user providing described voice messaging；
Age according to the user providing described voice messaging determined and sex, believe described original text
Breath is modified processing, and generates final text message.
12. devices according to claim 11, it is characterised in that also include processing unit, use
In:
After generating final text message, age and the sex to the user determined, and final text envelope
Breath is analyzed, and according to analysis result, generates user profile；
Set up the corresponding relation between the mark of described terminal, and the user profile generated, by described correspondence
Relation stores to described user profile set.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510883295.6A CN105895103B (en) | 2015-12-03 | 2015-12-03 | Voice recognition method and device |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510883295.6A CN105895103B (en) | 2015-12-03 | 2015-12-03 | Voice recognition method and device |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN105895103A true CN105895103A (en) | 2016-08-24 |
| CN105895103B CN105895103B (en) | 2020-01-17 |
Family
ID=57002113
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201510883295.6A Active CN105895103B (en) | 2015-12-03 | 2015-12-03 | Voice recognition method and device |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN105895103B (en) |
Cited By (23)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106682642A (en) * | 2017-01-06 | 2017-05-17 | 竹间智能科技（上海）有限公司 | Multi-language-oriented behavior identification method and multi-language-oriented behavior identification system |
| CN107134279A (en) * | 2017-06-30 | 2017-09-05 | 百度在线网络技术（北京）有限公司 | A kind of voice awakening method, device, terminal and storage medium |
| CN107731229A (en) * | 2017-09-29 | 2018-02-23 | 百度在线网络技术（北京）有限公司 | Method and apparatus for identifying voice |
| CN107945806A (en) * | 2017-11-10 | 2018-04-20 | 北京小米移动软件有限公司 | User identification method and device based on sound characteristic |
| CN108122555A (en) * | 2017-12-18 | 2018-06-05 | 北京百度网讯科技有限公司 | The means of communication, speech recognition apparatus and terminal device |
| CN108549628A (en) * | 2018-03-16 | 2018-09-18 | 北京云知声信息技术有限公司 | The punctuate device and method of streaming natural language information |
| CN108597495A (en) * | 2018-03-15 | 2018-09-28 | 维沃移动通信有限公司 | A kind of method and device of processing voice data |
| CN108682421A (en) * | 2018-04-09 | 2018-10-19 | 平安科技（深圳）有限公司 | A kind of audio recognition method, terminal device and computer readable storage medium |
| CN108831479A (en) * | 2018-06-27 | 2018-11-16 | 努比亚技术有限公司 | A kind of audio recognition method, terminal and computer readable storage medium |
| CN109117484A (en) * | 2018-08-13 | 2019-01-01 | 北京帝派智能科技有限公司 | A kind of voice translation method and speech translation apparatus |
| CN109388699A (en) * | 2018-10-24 | 2019-02-26 | 北京小米移动软件有限公司 | Input method, device, equipment and storage medium |
| CN110047467A (en) * | 2019-05-08 | 2019-07-23 | 广州小鹏汽车科技有限公司 | Audio recognition method, device, storage medium and controlling terminal |
| CN110246502A (en) * | 2019-06-26 | 2019-09-17 | 广东小天才科技有限公司 | Voice de-noising method, device and terminal device |
| CN110534098A (en) * | 2019-10-09 | 2019-12-03 | 国家电网有限公司客户服务中心 | A kind of the speech recognition Enhancement Method and device of age enhancing |
| CN110534112A (en) * | 2019-08-23 | 2019-12-03 | 王晓佳 | Distributed speech recongnition error correction device and method based on position and time |
| CN110689881A (en) * | 2018-06-20 | 2020-01-14 | 深圳市北科瑞声科技股份有限公司 | Speech recognition method, speech recognition device, computer equipment and storage medium |
| WO2020024582A1 (en) * | 2018-07-28 | 2020-02-06 | 华为技术有限公司 | Speech synthesis method and related device |
| CN110797014A (en) * | 2018-07-17 | 2020-02-14 | 中兴通讯股份有限公司 | Voice recognition method and device and computer storage medium |
| CN111402870A (en) * | 2019-01-02 | 2020-07-10 | 中国移动通信有限公司研究院 | Voice recognition method, device and equipment |
| CN111475619A (en) * | 2020-03-31 | 2020-07-31 | 北京三快在线科技有限公司 | Text information correction method and device, electronic equipment and storage medium |
| US10964317B2 (en) | 2017-07-05 | 2021-03-30 | Baidu Online Network Technology (Beijing) Co., Ltd. | Voice wakeup method, apparatus and system, cloud server and readable medium |
| CN113766171A (en) * | 2021-09-22 | 2021-12-07 | 广东电网有限责任公司 | Power transformation and defect elimination remote video consultation system and method based on AI voice control |
| US11574632B2 (en) | 2018-04-23 | 2023-02-07 | Baidu Online Network Technology (Beijing) Co., Ltd. | In-cloud wake-up method and system, terminal and computer-readable storage medium |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1758248A (en) * | 2004-10-05 | 2006-04-12 | 微软公司 | Systems, methods, and interfaces for providing personalized search and information access |
| US20120215537A1 (en) * | 2011-02-17 | 2012-08-23 | Yoshihiro Igarashi | Sound Recognition Operation Apparatus and Sound Recognition Operation Method |
| KR20120101855A (en) * | 2011-03-07 | 2012-09-17 | (주)에이치씨아이랩 | Result corrector for dictation speech recognition and result correction method |
| CN102682763A (en) * | 2011-03-10 | 2012-09-19 | 北京三星通信技术研究有限公司 | Method, device and terminal for correcting named entity vocabularies in voice input text |
| CN104508739A (en) * | 2012-06-21 | 2015-04-08 | 谷歌公司 | Dynamic language model |
| CN105095176A (en) * | 2014-04-29 | 2015-11-25 | 华为技术有限公司 | Method for extracting feature information of text information by user equipment and user equipment |
-
2015
- 2015-12-03 CN CN201510883295.6A patent/CN105895103B/en active Active
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1758248A (en) * | 2004-10-05 | 2006-04-12 | 微软公司 | Systems, methods, and interfaces for providing personalized search and information access |
| US20120215537A1 (en) * | 2011-02-17 | 2012-08-23 | Yoshihiro Igarashi | Sound Recognition Operation Apparatus and Sound Recognition Operation Method |
| KR20120101855A (en) * | 2011-03-07 | 2012-09-17 | (주)에이치씨아이랩 | Result corrector for dictation speech recognition and result correction method |
| CN102682763A (en) * | 2011-03-10 | 2012-09-19 | 北京三星通信技术研究有限公司 | Method, device and terminal for correcting named entity vocabularies in voice input text |
| CN104508739A (en) * | 2012-06-21 | 2015-04-08 | 谷歌公司 | Dynamic language model |
| CN105095176A (en) * | 2014-04-29 | 2015-11-25 | 华为技术有限公司 | Method for extracting feature information of text information by user equipment and user equipment |
Cited By (28)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106682642A (en) * | 2017-01-06 | 2017-05-17 | 竹间智能科技（上海）有限公司 | Multi-language-oriented behavior identification method and multi-language-oriented behavior identification system |
| CN107134279A (en) * | 2017-06-30 | 2017-09-05 | 百度在线网络技术（北京）有限公司 | A kind of voice awakening method, device, terminal and storage medium |
| US10964317B2 (en) | 2017-07-05 | 2021-03-30 | Baidu Online Network Technology (Beijing) Co., Ltd. | Voice wakeup method, apparatus and system, cloud server and readable medium |
| CN107731229A (en) * | 2017-09-29 | 2018-02-23 | 百度在线网络技术（北京）有限公司 | Method and apparatus for identifying voice |
| US11011163B2 (en) | 2017-09-29 | 2021-05-18 | Baidu Online Network Technology (Beijing) Co., Ltd. | Method and apparatus for recognizing voice |
| CN107945806A (en) * | 2017-11-10 | 2018-04-20 | 北京小米移动软件有限公司 | User identification method and device based on sound characteristic |
| CN108122555A (en) * | 2017-12-18 | 2018-06-05 | 北京百度网讯科技有限公司 | The means of communication, speech recognition apparatus and terminal device |
| CN108597495A (en) * | 2018-03-15 | 2018-09-28 | 维沃移动通信有限公司 | A kind of method and device of processing voice data |
| CN108549628A (en) * | 2018-03-16 | 2018-09-18 | 北京云知声信息技术有限公司 | The punctuate device and method of streaming natural language information |
| CN108682421A (en) * | 2018-04-09 | 2018-10-19 | 平安科技（深圳）有限公司 | A kind of audio recognition method, terminal device and computer readable storage medium |
| US11574632B2 (en) | 2018-04-23 | 2023-02-07 | Baidu Online Network Technology (Beijing) Co., Ltd. | In-cloud wake-up method and system, terminal and computer-readable storage medium |
| CN110689881A (en) * | 2018-06-20 | 2020-01-14 | 深圳市北科瑞声科技股份有限公司 | Speech recognition method, speech recognition device, computer equipment and storage medium |
| CN108831479A (en) * | 2018-06-27 | 2018-11-16 | 努比亚技术有限公司 | A kind of audio recognition method, terminal and computer readable storage medium |
| CN110797014A (en) * | 2018-07-17 | 2020-02-14 | 中兴通讯股份有限公司 | Voice recognition method and device and computer storage medium |
| WO2020024582A1 (en) * | 2018-07-28 | 2020-02-06 | 华为技术有限公司 | Speech synthesis method and related device |
| CN109117484A (en) * | 2018-08-13 | 2019-01-01 | 北京帝派智能科技有限公司 | A kind of voice translation method and speech translation apparatus |
| CN109117484B (en) * | 2018-08-13 | 2019-08-06 | 北京帝派智能科技有限公司 | A kind of voice translation method and speech translation apparatus |
| US11335348B2 (en) | 2018-10-24 | 2022-05-17 | Beijing Xiaomi Mobile Software Co., Ltd. | Input method, device, apparatus, and storage medium |
| CN109388699A (en) * | 2018-10-24 | 2019-02-26 | 北京小米移动软件有限公司 | Input method, device, equipment and storage medium |
| CN111402870B (en) * | 2019-01-02 | 2023-08-15 | 中国移动通信有限公司研究院 | Voice recognition method, device and equipment |
| CN111402870A (en) * | 2019-01-02 | 2020-07-10 | 中国移动通信有限公司研究院 | Voice recognition method, device and equipment |
| CN110047467B (en) * | 2019-05-08 | 2021-09-03 | 广州小鹏汽车科技有限公司 | Voice recognition method, device, storage medium and control terminal |
| CN110047467A (en) * | 2019-05-08 | 2019-07-23 | 广州小鹏汽车科技有限公司 | Audio recognition method, device, storage medium and controlling terminal |
| CN110246502A (en) * | 2019-06-26 | 2019-09-17 | 广东小天才科技有限公司 | Voice de-noising method, device and terminal device |
| CN110534112A (en) * | 2019-08-23 | 2019-12-03 | 王晓佳 | Distributed speech recongnition error correction device and method based on position and time |
| CN110534098A (en) * | 2019-10-09 | 2019-12-03 | 国家电网有限公司客户服务中心 | A kind of the speech recognition Enhancement Method and device of age enhancing |
| CN111475619A (en) * | 2020-03-31 | 2020-07-31 | 北京三快在线科技有限公司 | Text information correction method and device, electronic equipment and storage medium |
| CN113766171A (en) * | 2021-09-22 | 2021-12-07 | 广东电网有限责任公司 | Power transformation and defect elimination remote video consultation system and method based on AI voice control |
Also Published As
| Publication number | Publication date |
|---|---|
| CN105895103B (en) | 2020-01-17 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN105895103A (en) | Speech recognition method and device | |
| CN108447486B (en) | Voice translation method and device | |
| CN102243871B (en) | Methods and system for grammar fitness evaluation as speech recognition error predictor | |
| CN105374356B (en) | Audio recognition method, speech assessment method, speech recognition system and speech assessment system | |
| CN109410664B (en) | Pronunciation correction method and electronic equipment | |
| CN104185868B (en) | Authentication voice and speech recognition system and method | |
| CN103714048B (en) | Method and system for correcting text | |
| CN105512228A (en) | Bidirectional question-answer data processing method and system based on intelligent robot | |
| CN101567189A (en) | Device, method and system for correcting voice recognition result | |
| CN104008752A (en) | Speech recognition device and method, and semiconductor integrated circuit device | |
| CN108305618B (en) | Voice acquisition and search method, intelligent pen, search terminal and storage medium | |
| CN110120221A (en) | The offline audio recognition method of user individual and its system for vehicle system | |
| CN110021293A (en) | Audio recognition method and device, readable storage medium storing program for executing | |
| CN107240394A (en) | A kind of dynamic self-adapting speech analysis techniques for man-machine SET method and system | |
| CN110019741A (en) | Request-answer system answer matching process, device, equipment and readable storage medium storing program for executing | |
| CN113782026A (en) | Information processing method, device, medium and equipment | |
| CN108364655A (en) | Method of speech processing, medium, device and computing device | |
| Stemmer et al. | Acoustic modeling of foreign words in a German speech recognition system | |
| US11615787B2 (en) | Dialogue system and method of controlling the same | |
| CN110111778B (en) | Voice processing method and device, storage medium and electronic equipment | |
| CN111161718A (en) | Voice recognition method, device, equipment, storage medium and air conditioner | |
| CN110570838A (en) | Voice stream processing method and device | |
| CN111128127A (en) | Voice recognition processing method and device | |
| KR20160138613A (en) | Method for auto interpreting using emoticon and apparatus using the same | |
| CN114783424A (en) | Text corpus screening method, device, equipment and storage medium |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| CB02 | Change of applicant information | ||
| CB02 | Change of applicant information |
Address after: Room 301-1, Room 301-3, Area B2, Animation Building, No. 126 Animation Road, Zhongxin Eco-city, Tianjin Binhai New Area, Tianjin Applicant after: LE SHI ZHI XIN ELECTRONIC TECHNOLOGY (TIANJIN) Ltd. Address before: 300453 Tianjin Binhai New Area, Tianjin Eco-city, No. 126 Animation and Animation Center Road, Area B1, Second Floor 201-427 Applicant before: Xinle Visual Intelligent Electronic Technology (Tianjin) Co.,Ltd. Address after: 300453 Tianjin Binhai New Area, Tianjin Eco-city, No. 126 Animation and Animation Center Road, Area B1, Second Floor 201-427 Applicant after: Xinle Visual Intelligent Electronic Technology (Tianjin) Co.,Ltd. Address before: 300467 Tianjin Binhai New Area, ecological city, animation Middle Road, building, No. two, B1 District, 201-427 Applicant before: LE SHI ZHI XIN ELECTRONIC TECHNOLOGY (TIANJIN) Ltd. |
|
| GR01 | Patent grant | ||
| GR01 | Patent grant | ||
| PP01 | Preservation of patent right | ||
| PP01 | Preservation of patent right |
Effective date of registration: 20210201 Granted publication date: 20200117 |
|
| PD01 | Discharge of preservation of patent | ||
| PD01 | Discharge of preservation of patent |
Date of cancellation: 20240201 Granted publication date: 20200117 |
|
| PP01 | Preservation of patent right | ||
| PP01 | Preservation of patent right |
Effective date of registration: 20240313 Granted publication date: 20200117 |