CN105096945A - Voice recognition method and voice recognition device for terminal - Google Patents
Voice recognition method and voice recognition device for terminal Download PDFInfo
- Publication number
- CN105096945A CN105096945A CN201510548212.8A CN201510548212A CN105096945A CN 105096945 A CN105096945 A CN 105096945A CN 201510548212 A CN201510548212 A CN 201510548212A CN 105096945 A CN105096945 A CN 105096945A
- Authority
- CN
- China
- Prior art keywords
- word
- voice
- user
- alphabetic literal
- alphabetical
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Abstract
The embodiments of the present invention disclose a voice recognition method and a voice recognition device for a terminal. The voice recognition method comprises the steps of acquiring alphabetic voices input by a user; recognizing the alphabetic voices to determine alphabetic characters corresponding to the alphabetic voices; and constituting a word by at least one recognized alphabetic character as a recognition result. When the voice recognition device on the terminal is in use, words are constituted in the letter-by-letter voice acquisition and alphabetic character recognition manner. Therefore, the recognition accuracy is increased, and the voice input experience is also improved for users. In this way, the word inquiring operation of users on the terminal is more efficient and more convenient.
Description
Technical field
The embodiment of the present invention relates to speech recognition technology, particularly relates to a kind of audio recognition method and device of terminal.
Background technology
Along with the development of speech recognition technology, carry out by voice a kind of important way that input characters has become PC end and cell phone end text event detection.But because the accuracy rate of speech recognition is on the low side, the experience of phonetic entry is unsatisfactory.For a long time, phonetic entry all fails to become main input mode, user relies on for counsel always or input through keyboard.
For the software of verification certificate word.There is the software of some verification certificate words in the market, wherein most software support phonetic entry word, its speech recognition process is that user faces toward PC end or this whole-word read aloud by the microphone of mobile phone terminal, software is identified the word pronunciation that user inputs by signer-independent sign language recognition (SpeakerIndependentAutomaticSpeechRecognition) technology subsequently, the content that recognition result inputs as user.But such recognition accuracy is general lower, main cause has: the pronunciation of different user is widely different, user may not know the orthoepy of word and misreads word, the word that same pronunciation possibility is corresponding different, and domestic speech recognition technology mainly carries out accuracy rate optimization for Chinese, be optimized foreign language is then less.
It can thus be appreciated that, the accuracy rate that existing phonetic entry word technology carries out speech recognition is low, when user does not know the pronunciation of word, then can not input word by voice, or, when user know its pronunciation may inaccurate time, be just more prone to use input through keyboard, and avoid use phonetic entry.Therefore the input of user is experienced also poor.
Summary of the invention
The embodiment of the present invention provides a kind of audio recognition method and device of terminal, and to solve, word pronunciation recognition accuracy in prior art is low, user inputs and experiences poor problem.
First aspect, embodiments provides a kind of audio recognition method of terminal, comprising:
Gather the alphabetical voice of user's input;
Described alphabetical voice are identified, to determine the alphabetic literal corresponding with described alphabetical voice;
By recognize at least one described in alphabetic literal composition word, as recognition result.
Second aspect, the embodiment of the present invention additionally provides a kind of speech recognition equipment of terminal, comprising:
Gather voice module, for gathering the alphabetical voice of user's input;
Identify voice module, for identifying described alphabetical voice, to determine the alphabetic literal corresponding with described alphabetical voice;
Composition word module, for by recognize at least one described in alphabetic literal composition word, as recognition result.
The audio recognition method of a kind of terminal provided by the invention and device, by gathering the alphabetical voice of user's input, alphabetical voice are identified to the alphabetic literal determining that voice alphabetical with this are corresponding, then at least one alphabetic literal recognized is formed word as recognition result.Owing to usually there will not be the problem of cacoepy when user reads aloud single letter, also can will not to be familiar with or uncertain pronunciation or the word with same pronunciation split into letter, therefore the present invention adopt letter-by-letter voice collecting and identification letter word to form the method for word, improve the phonetic entry experience that recognition accuracy also improves user, make user in terminal, carry out the operations such as looking up words more efficient, convenient.
Accompanying drawing explanation
In order to be illustrated more clearly in the technical scheme in the embodiment of the present invention, the accompanying drawing used required in describing embodiment is done one below to introduce simply, apparently, accompanying drawing in the following describes is some embodiments of the present invention, for those of ordinary skill in the art, under the prerequisite not paying creative work, other accompanying drawing can also be obtained according to these accompanying drawings.
Fig. 1 is the process flow diagram of the audio recognition method of a kind of terminal provided in the embodiment of the present invention one;
Fig. 2 is the process flow diagram of the audio recognition method of a kind of terminal provided in the embodiment of the present invention two;
Fig. 3 is the schematic diagram of the speech recognition equipment of a kind of terminal provided in the embodiment of the present invention three.
Embodiment
For making the object, technical solutions and advantages of the present invention clearly, hereinafter with reference to the accompanying drawing in the embodiment of the present invention, by embodiment, technical scheme of the present invention is described clearly and completely, obviously, described embodiment is the present invention's part embodiment, instead of whole embodiments.Based on the embodiment in the present invention, those of ordinary skill in the art, not making the every other embodiment obtained under creative work prerequisite, belong to the scope of protection of the invention.
Embodiment one
As shown in Figure 1, a kind of process flow diagram of audio recognition method of terminal for providing in the embodiment of the present invention one.The technical scheme of the present embodiment to be applicable in terminal by phonetic entry with situations such as looking up words.The electric terminal equipment that terminal can be the input of any support voice and identify, typical in mobile phone, panel computer etc.The method can be performed by speech recognition equipment, and this device can adopt the mode of software and/or hardware to realize, and configuration performs in the terminal.
The audio recognition method of a kind of terminal that the present embodiment provides comprises as follows:
The alphabetical voice of S110, collection user input.
For the language (as English, French, Thai etc.) by letter spelling, the number of letters comprised in its word is general and few, user also there will not be the problem of cacoepy usually when reading aloud letter, and user is for not to be familiar with or the word of uncertain pronunciation also can input by being split as letter.Therefore in this step, user carrys out the speech recognition equipment input voice to terminal by letter-by-letter mode of reading aloud, then there will not be the situation collecting user's input error word pronunciation, also there will not be user to input situation that word pronunciation is same pronunciation word.
In aforesaid operations, the speech recognition equipment of terminal obtains input voice by the mode of alphabetical voice collecting one by one, instead of the mode of whole-word voice collecting in prior art, be convenient to the word that user inputs uncertain pronunciation word and has identical pronunciation, the compatible articulation problems of different user, also be convenient to that speech recognition equipment is follow-up carries out speech recognition, make the mode that the voice collecting accuracy rate of speech recognition equipment gathers higher than word pronunciation.
S120, alphabetical voice to be identified, to determine the alphabetic literal corresponding with alphabetical voice.
When the speech recognition equipment of terminal collects alphabetical voice, speech recognition equipment identifies these alphabetical voice immediately, to determine the alphabetic literal that voice alphabetical with this are corresponding.When multiple alphabetical voice of speech recognition equipment acquisition order user input, order identifies multiple alphabetical voice, to determine the alphabetic literal of each alphabetical voice that user inputs.
In aforesaid operations, because alphabetical voice relative overall word pronunciation is simple, therefore the raising gathering accuracy rate makes speech recognition equipment to the recognition accuracy of alphabetical voice higher than the discrimination to whole-word voice, each alphabetic literal in word can be determined to the identification of alphabetical voice, improve word identification accuracy rate.
S130, at least one alphabetic literal composition word that will recognize, as recognition result.
After user has inputted a word, the alphabetic literal composition word that order identifies by speech recognition equipment, this word is the word that speech recognition equipment goes out according to the alphabetical speech recognition that user inputs, and the word of this composition is carried out subsequent operation as recognition result for user by speech recognition equipment.
The audio recognition method of a kind of terminal that the present embodiment provides, by gathering the alphabetical voice of user's input, alphabetical voice are identified to the alphabetic literal determining that voice alphabetical with this are corresponding, then at least one alphabetic literal recognized is formed word as recognition result.Owing to usually there will not be the problem of cacoepy when user reads aloud single letter, also can will not to be familiar with or uncertain pronunciation or the word with same pronunciation split into letter, therefore the present embodiment adopt letter-by-letter voice collecting and identification letter word to form word, improve the phonetic entry experience that recognition accuracy also improves user, make user in terminal, carry out the operations such as looking up words more efficient, convenient.
Embodiment two
As shown in Figure 2, a kind of process flow diagram of audio recognition method of terminal for providing in the embodiment of the present invention two.The technical scheme of the present embodiment to be applicable in terminal by phonetic entry with situations such as looking up words.The method can be performed by speech recognition equipment, and this device can adopt the mode of software and/or hardware to realize, and configuration performs in the terminal.
Speech recognition technology comprises particular person speech recognition technology and signer-independent sign language recognition technology.The articulation type of different user is different, and therefore the recognition accuracy that identifies the alphabetical voice that user inputs of apparent employing particular person speech recognition technology is higher than the recognition accuracy of signer-independent sign language recognition technology.In the present embodiment, the audio recognition method of terminal preferably adopts particular person speech recognition technology to carry out speech recognition, can increase substantially recognition accuracy.
The audio recognition method of a kind of terminal that the present embodiment provides comprises as follows:
The base word mother pronunciation that in the required language of S210, collection user input, basic alphabetic literal is corresponding, using the identification coupling master meter of the base word mother pronunciation of user letter in identifiable language.
The quantity of the basic alphabetic literal that most of language pack contains is also few, and the basic alphabetic literal as English is 26, and Spanish basic alphabetic literal is 29.The present embodiment preferably adopts particular person speech recognition technology to carry out speech recognition, therefore preferential collection active user is needed to input base word mother pronunciation corresponding to basic alphabetic literal in required language, to input the identification coupling master meter of required language word mother pronunciation as this user.
26 basic alphabetic literals of English are read aloud for this user, user inputs corresponding english foundation letter voice, speech recognition equipment gathers speech data, and according to the base word mother pronunciation of user's input and the base word mother pronunciation of user's input and the corresponding relation of english foundation alphabetic literal, set up the identification coupling master meter of this user's English, for uniquely identifying all English letter voice that this user inputs.Therefore the base word mother pronunciation of required language of preserving user's input in coupling master meter is identified, and the corresponding relation of base word mother pronunciation that basic alphabetic literal and user input.
By setting up the identification coupling master meter of language needed for user in this operation, so that the alphabetical voice of language needed for follow-up this user of employing particular person speech recognition technology identification.
The alphabetical voice of S220, collection user input.
In aforesaid operations, the speech recognition equipment of terminal obtains input voice by the mode of alphabetical voice collecting one by one, instead of the mode of whole-word voice collecting in prior art, be convenient to the word that user inputs uncertain pronunciation word and has identical pronunciation, also be convenient to that speech recognition equipment is follow-up carries out speech recognition, make the mode that the voice collecting accuracy rate of speech recognition equipment gathers higher than word pronunciation.
S230, by alphabetical voice identification coupling master meter in mate.
In aforesaid operations, the alphabetical voice that the user collected inputs are mated master meter with the identification of required language and mate by speech recognition equipment.Particularly, speech recognition equipment needs the alphabetical voice collected to mate one by one with the identification base word mother pronunciation of mating in master meter, to determine the base word mother pronunciation the highest with alphabetical voice match degree.
S240, determine the base word mother pronunciation the highest with alphabetical voice match degree, using basic alphabetic literal corresponding for this base word mother pronunciation as the alphabetic literal corresponding with alphabetical voice.
When determining with the base word mother pronunciation that user's input alphabet voice match degree is the highest from identification coupling master meter, speech recognition equipment mates the corresponding relation of base word mother pronunciation and basic alphabetic literal in master meter according to identifying, using basic alphabetic literal corresponding for base word mother pronunciation as the alphabetic literal corresponding with the alphabetical voice that user inputs.Speech recognition equipment realizes identifying alphabetical voice, to determine the alphabetic literal corresponding with alphabetical voice thus.
Identify to determine that the preferred implementation of the alphabetic literal corresponding with alphabetical voice performs by above-mentioned S230 ~ S240 to alphabetical voice, it will be appreciated by those skilled in the art that the operation that can also perform by other means and identify to determine the alphabetic literal corresponding with alphabetical voice to alphabetical voice.
S250, by least one alphabetic literal recognized composition word show.
When speech recognition equipment recognizes at least one alphabetic literal corresponding to alphabetical voice of user's input, the alphabetic literal recognized composition word shows by the optional display screen by terminal of speech recognition equipment.
The optional alphabetic literal of identification and the alphabetic literal to identify before being combined successively forms word and shows.When reading aloud word " native " as letter-by-letterly in user, be shown as " n " after speech recognition equipment recognizes " n ", be shown as " na " after recognizing " a ", be shown as " nat " after recognizing " t ", be shown as " nati " after recognizing " i ", be shown as " nativ " after recognizing " v ", be shown as " native " after recognizing " e ".
Can be selected in after user reads aloud word " native " letter-by-letterly, the whole alphabetic literals composition words recognized successively show by speech recognition equipment.When reading aloud word " native ", speech recognition equipment recognizes successively " n " as letter-by-letterly in user, " a ", " t ", " i ", " v ", after " e ", be shown as " native ".
If S260 gets the confirmation result of user or the speech pause signal of user, then the word formed by the current alphabetic literal recognized is as recognition result.
Optional when user is bright run through required word after, user can produce speech pause, then speech recognition equipment can not collect again user input alphabetical voice, now speech recognition equipment can obtain this speech pause signal and using instantly recognize alphabetic literal composition word as recognition result.After reading aloud word " native ", when speech recognition equipment is shown as " native ", because user produces speech pause as letter-by-letterly in user, therefore " native " of current display is as recognition result.This optional arrange user speech pause exceed certain hour (as 3s) time, speech pause signal independently triggers, and speech recognition equipment obtains speech pause signal.
Also selectable user can be checked the word of display and confirmed the word of display on the display screen of terminal or cancel or the associative operation such as amendment.When speech recognition equipment receives the confirmation signal of user by configuration display screen in the terminal, then illustrate that user determines that the word that current display screen show is inputted correct word, now the word that formed by the current alphabetic literal that recognizes according to confirmation signal of speech recognition equipment is as recognition result.
By recognize at least one described in alphabetic literal composition word performs using the preferred implementation as recognition result by above-mentioned S250 ~ S260, it will be appreciated by those skilled in the art that can also perform by other means by recognize at least one described in alphabetic literal form word using the operation as recognition result.
The audio recognition method of the terminal that the present embodiment provides, the base word mother pronunciation that in the required language of preferential collection user input, basic alphabetic literal is corresponding is to set up the identification coupling master meter of this user, the alphabetical voice user of collection inputted mate in identification coupling master meter, to be improved the recognition accuracy to user input voice by particular person speech recognition technology.Embodiment improves the Consumer's Experience of phonetic entry, the demand making looking up words such obtains more efficiently, easily meeting.Compared to signer-independent sign language recognition technology, adopt the framework of the audio recognition method of particular person speech recognition technology simple, easily realize, due to complete identification coupling master meter can be set up, recognition accuracy is improved simultaneously, compared with prior art, tool has great advantage.
It should be noted that, it will be appreciated by those skilled in the art that can also perform by other means by recognize at least one described in alphabetic literal composition word using the operation as recognition result, as at least one alphabetic literal recognized is formed word undetermined; Word undetermined is mated with word library, to determine the word that matching degree is the highest, as recognition result.This word undetermined of composition and the word in the word library in terminal carried out mating using by word the highest for matching degree as recognition result.Optional terminal preserves word library and this terminal by networking upgrade word library to mate, or can directly by network mate with the up-to-date word library on network, using by word the highest for matching degree as recognition result.
Embodiment three
As shown in Figure 3, a kind of schematic diagram of speech recognition equipment of terminal for providing in the embodiment of the present invention three.The technical scheme of the present embodiment to be applicable in terminal by phonetic entry with situations such as looking up words.This device can perform the audio recognition method described in any embodiment of the present invention, and this device can adopt the mode of software and/or hardware to realize, and configuration performs in the terminal.
The present embodiment three provides a kind of speech recognition equipment of terminal, and this device comprises: gather voice module 310, identify voice module 320 and composition word module 330.
Wherein, voice module 310 is gathered for gathering the alphabetical voice of user's input; Identify that voice module 320 is for identifying alphabetical voice, to determine the alphabetic literal corresponding with alphabetical voice; At least one alphabetic literal composition word of composition word module 330 for recognizing, as recognition result.
Speech recognition equipment also comprises alternatively: set up base modules 300.
Wherein, set up base modules 300 for before the alphabetical voice gathering user's input, gather the base word mother pronunciation that in the required language of user's input, basic alphabetic literal is corresponding, using the identification coupling master meter of the base word mother pronunciation of user letter in identifiable language.
Identify that voice module 320 comprises alternatively: mate voice unit 321 and determine text unit 322.
Wherein, voice unit 321 is mated for being mated in identification coupling master meter by alphabetical voice; Determine that text unit 322 is for determining the base word mother pronunciation the highest with alphabetical voice match degree, using basic alphabetic literal corresponding for this base word mother pronunciation as the alphabetic literal corresponding with alphabetical voice.
Form word module 330 alternatively to comprise: display sub-word units 331 and determination result unit 332.
Wherein, sub-word units 331 is shown for being shown by least one alphabetic literal recognized composition word; If determination result unit 332 is for the speech pause signal of the confirmation result or user that get user, then using the word of the current alphabetic literal composition recognized as recognition result.
Form word module 330 alternatively to comprise: sub-word units undetermined 333 and determine sub-word units 334.
Wherein, sub-word units 333 undetermined is for forming word undetermined by least one alphabetic literal recognized; Determine that sub-word units 334 is for mating word undetermined with word library, to determine the word that matching degree is the highest, as recognition result.
The speech recognition equipment of a kind of terminal that the present embodiment provides, the alphabetical voice of user's input are gathered by gathering voice module, by identifying that voice module identifies the alphabetic literal determining that voice alphabetical with this are corresponding to alphabetical voice, then by composition word module, at least one alphabetic literal recognized is formed word as recognition result.Owing to usually there will not be the problem of cacoepy when user reads aloud single letter, also can will not to be familiar with or uncertain pronunciation or the word with same pronunciation split into letter, therefore the present embodiment adopt letter-by-letter voice collecting and identification letter word to form word, improve the phonetic entry experience that recognition accuracy also improves user, make user in terminal, carry out the operations such as looking up words more efficient, convenient.
The embodiment of the present invention additionally provides a kind of terminal, and this terminal comprises: speech recognition equipment described above.This terminal is the electric terminal equipment such as computer, mobile phone being configured with microphone, and speech recognition equipment is used for the alphabetical voice being gathered user's input by configuration microphone in the terminal.Particularly, the collection voice module of speech recognition equipment is used for the alphabetical voice being gathered user's input by configuration microphone in the terminal.
As mentioned above, speech recognition equipment is configurable in the system of the terminals such as the various support voice input PC of word or mobile phone, this terminal gathers the alphabetical voice of user's input by microphone, and the languages of this speech recognition equipment identification can comprise all languages by letter spelling.
Particularly, to adopt the terminal of particular person speech recognition technology, the process of user's phonetic entry word is on the terminal (to input English word):
(1) configuration speech recognition equipment in the terminal guides user to read aloud 26 English alphabets;
(2) this user of system log (SYSLOG) of terminal is to the pronunciation of each letter, analyzes the sound characteristic that this user reads aloud each letter, identifies coupling master meter to set up;
(3) when user needs input English word, user combines out this word into syllables letter-by-letterly.During as needs input " native ", user reads aloud " n ", " a ", " t ", " i ", " v, " e " to terminal microphone;
(4) system of terminal carries out particular person speech recognition, to identify the letter of user's input;
(5) letter that user inputs is combined into English word by terminal system, as final recognition result.
The advantage of the speech recognition equipment configured in this terminal is, can set up complete identification coupling master meter for different user, and then application particular person speech recognition technology, realizes the speech recognition of high-accuracy, improves the experience of phonetic entry word; And only need user can read aloud letter.
Note, above are only preferred embodiment of the present invention and institute's application technology principle.Skilled person in the art will appreciate that and the invention is not restricted to specific embodiment described here, various obvious change can be carried out for a person skilled in the art, readjust and substitute and can not protection scope of the present invention be departed from.Therefore, although be described in further detail invention has been by above embodiment, the present invention is not limited only to above embodiment, when not departing from the present invention's design, can also comprise other Equivalent embodiments more, and scope of the present invention is determined by appended right.
Claims (10)
1. an audio recognition method for terminal, is characterized in that, comprising:
Gather the alphabetical voice of user's input;
Described alphabetical voice are identified, to determine the alphabetic literal corresponding with described alphabetical voice;
By recognize at least one described in alphabetic literal composition word, as recognition result.
2. audio recognition method according to claim 1, is characterized in that, before gathering the alphabetical voice of user's input, also comprises:
Gather the base word mother pronunciation that in the required language of described user input, basic alphabetic literal is corresponding, using the base word mother pronunciation of described user as the identification coupling master meter identifying letter in described language.
3. audio recognition method according to claim 2, is characterized in that, identifies described alphabetical voice, to determine that the alphabetic literal corresponding with described alphabetical voice comprises:
Described alphabetical voice are mated in described identification coupling master meter;
Determine the base word mother pronunciation the highest with described alphabetical voice match degree, using basic alphabetic literal corresponding for this base word mother pronunciation as the alphabetic literal corresponding with described alphabetical voice.
4. audio recognition method according to claim 1, is characterized in that, by recognize at least one described in alphabetic literal composition word, comprise as recognition result:
By recognize at least one described in alphabetic literal composition word show;
If get the confirmation result of user or the speech pause signal of user, then the word formed by the current alphabetic literal recognized is as recognition result.
5. audio recognition method according to claim 1, is characterized in that, by recognize at least one described in alphabetic literal composition word, comprise as recognition result:
By recognize at least one described in alphabetic literal form word undetermined;
Described word undetermined is mated with word library, to determine the word that matching degree is the highest, as recognition result.
6. a speech recognition equipment for terminal, is characterized in that, comprising:
Gather voice module, for gathering the alphabetical voice of user's input;
Identify voice module, for identifying described alphabetical voice, to determine the alphabetic literal corresponding with described alphabetical voice;
Composition word module, for by recognize at least one described in alphabetic literal composition word, as recognition result.
7. speech recognition equipment according to claim 6, is characterized in that, described speech recognition equipment also comprises:
Set up base modules, for before the alphabetical voice gathering user's input, gather the base word mother pronunciation that in the required language of described user input, basic alphabetic literal is corresponding, using the base word mother pronunciation of described user as the identification coupling master meter identifying letter in described language.
8. speech recognition equipment according to claim 7, is characterized in that, described identification voice module comprises:
Coupling voice unit, for mating described alphabetical voice in described identification coupling master meter;
Determine text unit, for determining the base word mother pronunciation the highest with described alphabetical voice match degree, using basic alphabetic literal corresponding for this base word mother pronunciation as the alphabetic literal corresponding with described alphabetical voice.
9. speech recognition equipment according to claim 6, is characterized in that, described composition word module comprises:
Display sub-word units, for by recognize at least one described in alphabetic literal composition word show;
Determination result unit, if for the speech pause signal of the confirmation result or user that get user, then using the word of the current alphabetic literal composition recognized as recognition result.
10. speech recognition equipment according to claim 6, is characterized in that, described composition word module comprises:
Sub-word units undetermined, for by recognize at least one described in alphabetic literal form word undetermined;
Determine sub-word units, for mating, described word undetermined to determine the word that matching degree is the highest, as recognition result with word library.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510548212.8A CN105096945A (en) | 2015-08-31 | 2015-08-31 | Voice recognition method and voice recognition device for terminal |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510548212.8A CN105096945A (en) | 2015-08-31 | 2015-08-31 | Voice recognition method and voice recognition device for terminal |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN105096945A true CN105096945A (en) | 2015-11-25 |
Family
ID=54577231
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201510548212.8A Pending CN105096945A (en) | 2015-08-31 | 2015-08-31 | Voice recognition method and voice recognition device for terminal |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN105096945A (en) |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105913841A (en) * | 2016-06-30 | 2016-08-31 | 北京小米移动软件有限公司 | Voice recognition method, voice recognition device and terminal |
| CN109036410A (en) * | 2018-08-30 | 2018-12-18 | Oppo广东移动通信有限公司 | Audio recognition method, device, storage medium and terminal |
| CN110019667A (en) * | 2017-10-20 | 2019-07-16 | 沪江教育科技（上海）股份有限公司 | It is a kind of that word method and device is looked into based on voice input information |
| CN110647732A (en) * | 2019-09-16 | 2020-01-03 | 广州云从信息科技有限公司 | Voice interaction method, system, medium and device based on biological recognition characteristics |
Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1315809A (en) * | 2000-03-30 | 2001-10-03 | 韦尔博泰克有限公司 | Apparatus and method for spelling speech recognition in mobile communication |
| CN1357821A (en) * | 2000-12-15 | 2002-07-10 | 丽台科技股份有限公司 | Phonetic input method |
| EP1396840A1 (en) * | 2002-08-12 | 2004-03-10 | Siemens Aktiengesellschaft | Spelling speech recognition apparatus |
| CN1667700A (en) * | 2004-03-10 | 2005-09-14 | 微软公司 | New-word pronunciation learning using a pronunciation graph |
| CN1892818A (en) * | 2005-06-29 | 2007-01-10 | 西门子公司 | Method for determining a list of hypotheses from the vocabulary of a speech recognition system |
| CN2893790Y (en) * | 2006-02-20 | 2007-04-25 | 李坤昌 | English alphabet speech inputting translating machine |
| CN101218625A (en) * | 2005-07-07 | 2008-07-09 | 国际商业机器公司 | Dictionary lookup for mobile devices using spelling recognition |
-
2015
- 2015-08-31 CN CN201510548212.8A patent/CN105096945A/en active Pending
Patent Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1315809A (en) * | 2000-03-30 | 2001-10-03 | 韦尔博泰克有限公司 | Apparatus and method for spelling speech recognition in mobile communication |
| CN1357821A (en) * | 2000-12-15 | 2002-07-10 | 丽台科技股份有限公司 | Phonetic input method |
| EP1396840A1 (en) * | 2002-08-12 | 2004-03-10 | Siemens Aktiengesellschaft | Spelling speech recognition apparatus |
| CN1667700A (en) * | 2004-03-10 | 2005-09-14 | 微软公司 | New-word pronunciation learning using a pronunciation graph |
| CN1892818A (en) * | 2005-06-29 | 2007-01-10 | 西门子公司 | Method for determining a list of hypotheses from the vocabulary of a speech recognition system |
| CN101218625A (en) * | 2005-07-07 | 2008-07-09 | 国际商业机器公司 | Dictionary lookup for mobile devices using spelling recognition |
| CN2893790Y (en) * | 2006-02-20 | 2007-04-25 | 李坤昌 | English alphabet speech inputting translating machine |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105913841A (en) * | 2016-06-30 | 2016-08-31 | 北京小米移动软件有限公司 | Voice recognition method, voice recognition device and terminal |
| CN105913841B (en) * | 2016-06-30 | 2020-04-03 | 北京小米移动软件有限公司 | Voice recognition method, device and terminal |
| CN110019667A (en) * | 2017-10-20 | 2019-07-16 | 沪江教育科技（上海）股份有限公司 | It is a kind of that word method and device is looked into based on voice input information |
| CN109036410A (en) * | 2018-08-30 | 2018-12-18 | Oppo广东移动通信有限公司 | Audio recognition method, device, storage medium and terminal |
| CN110647732A (en) * | 2019-09-16 | 2020-01-03 | 广州云从信息科技有限公司 | Voice interaction method, system, medium and device based on biological recognition characteristics |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US10777192B2 (en) | Method and apparatus of recognizing field of semantic parsing information, device and readable medium | |
| CN113168836B (en) | Computer system, voice recognition method and program product | |
| CN111523306A (en) | Text error correction method, device and system | |
| CN110020422A (en) | The determination method, apparatus and server of Feature Words | |
| WO2021103712A1 (en) | Neural network-based voice keyword detection method and device, and system | |
| CN104731767A (en) | Communication support apparatus, communication support method, and computer program product | |
| CN105096945A (en) | Voice recognition method and voice recognition device for terminal | |
| CN104866308A (en) | Scenario image generation method and apparatus | |
| CN111312209A (en) | Text-to-speech conversion processing method and device and electronic equipment | |
| CN109785829B (en) | Customer service assisting method and system based on voice control | |
| CN110738997B (en) | Information correction method and device, electronic equipment and storage medium | |
| CN113051362A (en) | Data query method and device and server | |
| CN105575402A (en) | Network teaching real time voice analysis method | |
| CN113380238A (en) | Method for processing audio signal, model training method, apparatus, device and medium | |
| CN105335356B (en) | The papery interpretation method and translation pen device of a kind of Semantic-Oriented identification | |
| CN111144118A (en) | Method, system, device and medium for identifying named entities in spoken text | |
| CN111209367A (en) | Information searching method, information searching device, electronic equipment and storage medium | |
| CN104484370A (en) | Transmitting method, receiving method, transmitting device, receiving device and system for answer information on basis of questions and answers | |
| CN113657088A (en) | Interface document analysis method and device, electronic equipment and storage medium | |
| CN113268981B (en) | Information processing method and device and electronic equipment | |
| CN115691503A (en) | Voice recognition method and device, electronic equipment and storage medium | |
| CN114171000A (en) | Audio recognition method based on acoustic model and language model | |
| CN114299930A (en) | End-to-end speech recognition model processing method, speech recognition method and related device | |
| CN114528851A (en) | Reply statement determination method and device, electronic equipment and storage medium | |
| CN110083807B (en) | Contract modification influence automatic prediction method, device, medium and electronic equipment |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20151125 |
|
| RJ01 | Rejection of invention patent application after publication |