CN103699530A

CN103699530A - Method and equipment for inputting texts in target application according to voice input information

Info

Publication number: CN103699530A
Application number: CN201210364155.4A
Authority: CN
Inventors: 孙培犖
Original assignee: Beijing Baidu Netcom Science and Technology Co Ltd
Current assignee: Beijing Baidu Netcom Science and Technology Co Ltd
Priority date: 2012-09-27
Filing date: 2012-09-27
Publication date: 2014-04-02

Abstract

The invention aims to provide a method and equipment for inputting texts in a target application according to voice input information in user equipment. The method comprises the following steps: acquiring the voice input information which is input by a user through a voice acquisition device of the user equipment; performing voice identification processing on the voice input information to obtain a language character string corresponding to the voice input information; according to the language character string, performing matching query on an input character library corresponding to an input method which corresponds to the target application to obtain one or more input candidate texts matched with the language character string; providing at least one of the input candidate texts for the target application. Compared with the prior art, the method and the equipment have the advantages that the matching efficiency of voice identified at the user equipment end and the input candidate texts is improved, the utilization rate of local resources is improved, the matching accuracy of the input candidate texts is improved in a mode combined with the input method, and the using experience of the user is improved.

Description

According to the method and apparatus of speech input information input text in intended application

Technical field

The present invention relates to phonetic entry technical field, relate in particular to a kind of for carry out the technology of text input in intended application according to speech input information at subscriber equipment.

Background technology

Along with the development of input technology, people are increasing for the demand of convenient abundant text input mode.Phonetic entry is the input mode of current main-stream, in future, also can not wither away in a short time.Yet existing phonetic entry technology all needs the support of server conventionally, for example, by client recording sound, be uploaded to after server, server carries out speech recognition, and coupling obtains corresponding input candidate text and returns to client, by client, is responsible for subsequent treatment.The speech recognition of this kind of mode mainly realizes in network-side with input candidate text matches, depends on network and connects, and need huge data model.For limited users equipment all such as processing speed, bandwidth, this kind of mode is unfavorable for user's experience.

Therefore, how, in subscriber equipment, according to the speech input information of user's input, coupling obtains corresponding input candidate text, and in intended application input text, become those skilled in the art and need one of problem of solution badly.

Summary of the invention

The object of this invention is to provide a kind of for carry out the method and apparatus of text input in intended application according to speech input information at subscriber equipment.

According to an aspect of the present invention, provide a kind of for carry out the method for text input in intended application according to speech input information at subscriber equipment, wherein, the method comprises the following steps:

A obtains the speech input information that user inputs by the voice acquisition device of subscriber equipment, and wherein, described speech input information is intended the text of inputting in the intended application of described subscriber equipment corresponding to described user;

B carries out voice recognition processing to described speech input information, to obtain the language character string corresponding with described speech input information;

C, according to described language character string, carries out matching inquiry in the input dictionary corresponding with the corresponding input method of described intended application, to obtain the one or more input candidate texts that match with described language character string;

D is by described one or more input candidate texts, at least one offers described intended application.

According to another aspect of the present invention, also provide a kind of subscriber equipment of text input that carries out according to speech input information in intended application, wherein, this equipment comprises:

Voice acquisition device, the speech input information of inputting by the voice acquisition device of subscriber equipment for obtaining user, wherein, described speech input information is intended the text of inputting in the intended application of described subscriber equipment corresponding to described user;

Speech recognition equipment, for described speech input information is carried out to voice recognition processing, to obtain the language character string corresponding with described speech input information;

Text matches device for according to described language character string, carries out matching inquiry in the input dictionary corresponding with the corresponding input method of described intended application, to obtain the one or more input candidate texts that match with described language character string;

Text generator, for by described one or more input candidate texts, at least one offers described intended application.

Compared with prior art, the speech input information that the present invention inputs in intended application user carries out voice recognition processing, obtain corresponding language character string, again in conjunction with the corresponding input dictionary of input method in subscriber equipment, coupling obtains corresponding input candidate text, improved the efficiency with input candidate text matches in the speech recognition of subscriber equipment end, improved Utilization of local resources rate, having reduced other stores such as this locality, the occupancy of the resources such as network service, the mode combining with input method has also improved the accuracy rate of input candidate text matches, promoted user's experience.

Accompanying drawing explanation

By reading the detailed description that non-limiting example is done of doing with reference to the following drawings, it is more obvious that other features, objects and advantages of the present invention will become:

Fig. 1 illustrate according to one aspect of the invention for carry out the equipment schematic diagram of text input in intended application according to speech input information at subscriber equipment;

Fig. 2 illustrate in accordance with a preferred embodiment of the present invention for carry out the equipment schematic diagram of text input in intended application according to speech input information at subscriber equipment;

Fig. 3 illustrate according to a further aspect of the present invention for carry out the method flow diagram of text input in intended application according to speech input information at subscriber equipment;

Fig. 4 illustrate in accordance with a preferred embodiment of the present invention for carry out the method flow diagram of text input in intended application according to speech input information at subscriber equipment.

In accompanying drawing, same or analogous Reference numeral represents same or analogous parts.

Embodiment

Below in conjunction with accompanying drawing, the present invention is described in further detail.

Fig. 1 illustrate according to one aspect of the invention for carry out the equipment schematic diagram of text input in intended application according to speech input information at subscriber equipment.This subscriber equipment 1 comprises voice acquisition device 101, speech recognition equipment 102, text matches device 103 and text generator 104.

At this; this subscriber equipment includes but not limited to anyly can carry out mutual equipment by voice acquisition device and user; as mobile devices such as mobile phone, notebook, panel computers; or intelligentized Furniture such as intelligent refrigerator, Intelligent microwave oven; or the map shows device in building etc.; those skilled in the art will be understood that above-mentioned subscriber equipment is only for giving an example; other subscriber equipmenies existing or that may occur are from now on as applicable to the present invention; also should be included in protection domain of the present invention, and with way of reference, be contained in this at this.

Wherein, voice acquisition device 101 obtains the speech input information that user inputs by the voice acquisition device of subscriber equipment, and wherein, described speech input information is intended the text of inputting in the intended application of described subscriber equipment corresponding to described user.Particularly, mutual by with subscriber equipment 1 of user, by the voice acquisition device input speech input information on it, as inputted speech input information by microphone, this speech input information is intended the text of inputting in the intended application of this subscriber equipment corresponding to this user, as this user intends input " iPhone " in the search box of intended application " search ", it inputs speech input information " iPhone " by voice acquisition device; This voice acquisition device 101 for example, by calling application programming interfaces (API) that this subscriber equipment provides or the communication mode of other agreements, obtains the speech input information " iPhone " that this user inputs.

Those skilled in the art will be understood that the above-mentioned mode of obtaining speech input information is only for giving an example; other existing or modes of obtaining speech input information that may occur are from now on as applicable to the present invention; also should be included in protection domain of the present invention, and with way of reference, be contained in this at this.

102 pairs of described speech input informations of speech recognition equipment carry out voice recognition processing, to obtain the language character string corresponding with described speech input information.Particularly, the speech input information that speech recognition equipment 102 obtains according to this voice acquisition device 101, by this speech input information is carried out to voice recognition processing, as by carry out matching inquiry in the speech database of this subscriber equipment, obtain the language character string corresponding with this speech input information; Or, the speech input information that this speech recognition equipment 102 obtains according to this voice acquisition device 101, this speech input information is carried out to voice recognition processing, as the voice data to this speech input information is resolved, acquisition can represent the phonetic symbol character string of the corresponding pronunciation of this speech input information, and then, according to this phonetic symbol character string, obtain the language character string matching with described phonetic symbol character string.

Connect precedent, this voice acquisition device 101 obtains the speech input information " iPhone " that this user inputs; 102 pairs of these speech input informations of this speech recognition equipment " iPhone " carry out voice recognition processing, as according to this speech input information " iPhone ", in the speech database of this subscriber equipment, carry out matching inquiry, obtain the language character string corresponding with this speech input information " iPhone ", as " iphone ", " AiFen " etc.

And for example, the speech input information " iPhone " that 102 pairs of voice acquisition device 101 of this speech recognition equipment obtain carries out voice recognition processing, as this speech input information " iPhone " is simplified according to the algorithm of formulating, obtain for the more needed characteristic of model, again the audio model data of storing in this characteristic and this subscriber equipment 1 are compared, the phonetic symbol character string of the pronunciation that represents this speech input information " iPhone " that acquisition meets, as " Ai1|Fen4 "; Then, this speech recognition equipment 102, according to this phonetic symbol character string, by modes such as matching inquiries, obtains the language character string matching with described phonetic symbol character string.Preferably, this speech recognition equipment 102 can also be in conjunction with the corresponding applicational language information of this intended application, obtain the language character string matching with this phonetic symbol character string, as carried out the modes such as matching inquiry in the speech database by corresponding with this applicational language information, obtain the language character string matching with this phonetic symbol character string, while being English as the applicational language information when intended application, obtain this corresponding language character string of phonetic symbol character string " Ai1|Fen4 " " iphone "; When the applicational language information of this intended application is Chinese, obtain this corresponding language character string of phonetic symbol character string " Ai1|Fen4 " " AiFen ".

Those skilled in the art will be understood that the mode of above-mentioned acquisition language character string is only for giving an example; the mode of other acquisition language character strings existing or that may occur is from now on as applicable to the present invention; also should be included in protection domain of the present invention, and with way of reference, be contained in this at this.

Text matches device 103, according to described language character string, carries out matching inquiry in the input dictionary corresponding with the corresponding input method of described intended application, to obtain the one or more input candidate texts that match with described language character string.Particularly, the language character string that text matches device 103 obtains according to these speech recognition equipment 102 couplings, by the mode of the matching inquiries such as exact matching, broad match, in the input dictionary corresponding with the corresponding input method of this intended application, carry out matching inquiry, to obtain the one or more input candidate texts that match with this language character string.

Connect example, when the applicational language information of this intended application is English, this speech recognition equipment 102 obtains this corresponding language character string of phonetic symbol character string " Ai1|Fen4 " " iphone "; Subsequently, text matches device 103 is according to this language character string " iphone ", in the input dictionary corresponding with the corresponding English input method of this intended application, carry out matching inquiry, a plurality of input candidate texts " iphone " that acquisition and this language character string match, " ip ", " i " etc.

If the applicational language information of this intended application is Chinese, speech recognition equipment 102 obtains this corresponding language character string of phonetic symbol character string " Ai1|Fen4 " " AiFen "; Subsequently, text matches device 103 is according to this language character string " AiFen ", in the input dictionary corresponding with the corresponding Chinese character coding input method of this intended application, carry out matching inquiry, a plurality of input candidate texts that acquisition and this language character string match " love is fragrant ", " liking to divide ", " Ai Fen " etc.

At this, the mode that text coalignment 103 coupling in input dictionary obtains one or more input candidate texts includes but not limited to the modes such as exact matching, broad match.Exact matching for example this language character string mates completely with this input candidate text, as " iphone " and " iphone ", " AiFen " and " liking fragrant " etc.; Broad match for example this language character string is mated with this input candidate textual portions, as " iphone " and " ip ", " AiFen " and " liking crazy " etc.

Those skilled in the art will be understood that the mode of the one or more input of above-mentioned coupling candidate text is only for giving an example; the mode of the one or more input of other couplings existing or that may occur from now on candidate text is as applicable to the present invention; also should be included in protection domain of the present invention, and with way of reference, be contained in this at this.

Text generator 104 is by described one or more input candidate texts, at least one offers described intended application.Particularly, text generator 104 mates the one or more input candidate texts that obtain according to text matches device 103, such as by calling dynamic web page techniques such as JSP, ASP or PHP, or, by calling modes such as showing control, by in these one or more input candidate texts at least one, offer this intended application with form in certain sequence.For example, text generator 104 is by calling demonstration control, by these one or more input candidate texts at least one, in the entry hurdle of the corresponding input method of this intended application, offer this intended application.At this, this entry hurdle for example only shows a line input candidate text, in this row entry hurdle, the number of shown input candidate text can be default also can being set by the user, by pressed special function keys by user, show lastrow or next line input candidate text, this special function keys can be for example "+" and "-"; Or text generator 104 can offer this intended application by least one form with multirow entry hurdle in these one or more input candidate texts according to the screen width of this subscriber equipment 1, further processes for this intended application.

Those skilled in the art will be understood that the above-mentioned mode of input candidate text that provides is only for giving an example; other existing or modes that input candidate text is provided that may occur are from now on as applicable to the present invention; also should be included in protection domain of the present invention, and with way of reference, be contained in this at this.

Preferably, between each device of subscriber equipment 1, be constant work.Particularly, voice acquisition device 101 obtains the speech input information that user inputs by the voice acquisition device of subscriber equipment, and wherein, described speech input information is intended the text of inputting in the intended application of described subscriber equipment corresponding to described user; 102 pairs of described speech input informations of speech recognition equipment carry out voice recognition processing, to obtain the language character string corresponding with described speech input information; Text matches device 103, according to described language character string, carries out matching inquiry in the input dictionary corresponding with the corresponding input method of described intended application, to obtain the one or more input candidate texts that match with described language character string; Text generator 104 is by described one or more input candidate texts, at least one offers described intended application.At this, it will be understood by those skilled in the art that " continuing " refers to that each device of subscriber equipment 1 carries out in real time or constantly respectively the acquisition of the obtaining of speech input information, voice recognition processing, language character string and the coupling of input candidate text and provides, until this subscriber equipment 1 stops obtaining the speech input information that user inputs by the voice acquisition device of subscriber equipment in a long time.

At this, the speech input information that the present invention inputs in intended application user carries out voice recognition processing, obtain corresponding language character string, again in conjunction with the corresponding input dictionary of input method, coupling obtains corresponding input candidate text, improved the efficiency with input candidate text matches in the speech recognition of subscriber equipment end, improved Utilization of local resources rate, reduced the occupancy of other resources such as this locality storage, network service, the mode combining with input method has also improved the accuracy rate of input candidate text matches, has promoted user's experience.

Further, even if this subscriber equipment 1 cannot connection server, the present invention still can realize to the obtaining of speech input information, voice recognition processing, the acquisition of language character string and the coupling of input candidate text with provide etc.

Preferably, this speech recognition equipment 102 also comprises voice recognition unit (not shown) and string matching unit (not shown).

This voice recognition unit carries out voice recognition processing to described speech input information, to obtain the phonetic symbol character string corresponding with described speech input information.Particularly, the speech input information that voice recognition unit obtains according to this voice acquisition device 101, this speech input information is carried out to voice recognition processing, as the voice data to this speech input information is resolved, acquisition can represent the phonetic symbol character string of the corresponding pronunciation of this speech input information.

For example, this voice acquisition device 101 obtains the speech input information " iPhone " that this user inputs; This voice recognition unit carries out voice recognition processing to this speech input information " iPhone ", as this speech input information " iPhone " is simplified according to the algorithm of formulating, obtain for the more needed characteristic of model, again the audio model data of storing in this characteristic and this subscriber equipment 1 are compared, the phonetic symbol character string of the pronunciation that represents this speech input information " iPhone " that acquisition meets, as " Ai1|Fen4 ".

String matching unit, according to described phonetic symbol character string, obtains the language character string matching with described phonetic symbol character string.Particularly, string matching unit, according to the phonetic symbol character string of acquisition that this voice recognition unit is identified, by modes such as matching inquiries, obtains the language character string matching with described phonetic symbol character string.

For example, the speech input information of supposing the user that voice acquisition device 101 obtains is " iPhone ", voice recognition unit determines that the corresponding phonetic symbol character string of this speech input information " iPhone " is for " Ai1|Fen4 ", string matching unit is according to the phonetic symbol character string of acquisition that this voice recognition unit is identified, by modes such as matching inquiries, obtain the language character string matching with described phonetic symbol character string, as " iphone ", " AiFen " etc.

Preferably, this subscriber equipment 1 also comprises adjusts treating apparatus (not shown), this adjusts treating apparatus according to described phonetic symbol character string, and in conjunction with the speech related information of described speech input information, described phonetic symbol character string is adjusted to processing, to obtain through adjusting the phonetic symbol character string after processing; Wherein, this string matching unit through adjusting the phonetic symbol character string after processing, obtains the described language character string matching with described phonetic symbol character string according to described.

Particularly, adjust the speech input information that treating apparatus obtains according to voice acquisition device 101, by this speech input information is carried out to speech analysis, or, by carrying out matching inquiry in the user information database the corresponding user of this speech input information, obtain the speech related information of this speech input information, this speech related information includes but not limited to accent information, element information, voiceprint of the corresponding user of this speech input information etc.; This adjusts treating apparatus according to the determined phonetic symbol character string of voice recognition unit, one or more in conjunction with in the above-mentioned speech related information of speech input information again, this phonetic symbol character string is adjusted to processing, as when the corresponding applicational language information of this intended application being when Chinese, the front and back nasal sound of this speech input information, flat cacuminal etc. are adjusted to processing, to obtain through adjusting the phonetic symbol character string after processing; For example, suppose that voice recognition unit is " Cuan1Yue4 " according to the determined phonetic symbol character string of language input message, adjust treating apparatus in conjunction with the speech related information of this speech input information, this phonetic symbol character string is adjusted to processing, as be adjusted to " Chuan1Yue4 "; Subsequently, this string matching unit through adjusting the phonetic symbol character string after processing, obtains the described language character string matching with this phonetic symbol character string according to this.

More preferably, described speech related information comprises following at least any one:

The accent information of-described speech input information;

The element information of-described speech input information;

The voiceprint of-described speech input information.

At this, the voice accent information of described speech input information is as the speech with a guy, place or native language feature, adjust treating apparatus according to the determined phonetic symbol character string of voice recognition unit, then in conjunction with the accent information of speech input information, this phonetic symbol character string is adjusted to processing.

The phonetic feature information of described speech input information includes but not limited to the element characteristics such as pitch, loudness of a sound, the duration of a sound and tone color of this speech input information, adjust treating apparatus according to the determined phonetic symbol character string of voice recognition unit, in conjunction with the element information of speech input information, this phonetic symbol character string is adjusted to processing again.

The voice voiceprint of described speech input information is the sound wave spectrum that carries language message, adjust treating apparatus according to the determined phonetic symbol character string of voice recognition unit, in conjunction with the voiceprint of speech input information, this phonetic symbol character string is adjusted to processing again.

Those skilled in the art will be understood that above-mentioned speech related information is only for giving an example, and other speech related information existing or that may occur from now on, as applicable to the present invention, also should be included in protection domain of the present invention, and with way of reference, are contained in this at this.

At this, the speech input information that the present invention inputs in intended application user carries out voice recognition processing, obtain phonetic symbol character string, and then, in conjunction with the speech related information of this speech input information, this phonetic symbol character string is adjusted to processing, and based on through adjusting the phonetic symbol character string after processing, mate and obtain language character string, further improved the accuracy rate of language character string coupling, and then improved the accuracy rate of follow-up input candidate text matches, promoted user's experience.

More preferably, this string matching unit is according to described phonetic symbol character string, and in conjunction with the corresponding applicational language information of described intended application, obtains the language character string matching with described phonetic symbol character string.At this, this applicational language information is for characterizing the information such as current languages input state of the input method such as language environment, correspondence of this intended application.

Particularly, string matching unit is according to the phonetic symbol character string of acquisition that this voice recognition unit is identified, and in conjunction with the corresponding applicational language information of this intended application, by modes such as matching inquiries, as carried out the modes such as matching inquiry in the speech database by corresponding with this applicational language information, obtain the language character string matching with described phonetic symbol character string.

For example, the speech input information of supposing the user that voice acquisition device 101 obtains is " iPhone ", voice recognition unit determines that the corresponding phonetic symbol character string of this speech input information " iPhone " is for " Ai1|Fen4 ", when the applicational language information of intended application is English, this string matching unit, by matching algorithm, obtains this corresponding language character string of phonetic symbol character string " Ai1|Fen4 " " iphone "; If the applicational language information of this intended application is Chinese, string matching unit obtains this corresponding language character string of phonetic symbol character string " Ai1|Fen4 " " AiFen "; If the language that the applicational language information of this intended application is other languages, string matching unit, according to this phonetic symbol character string " Ai1|Fen4 ", obtains the language character string of other languages corresponding with this phonetic symbol character string.

Preferably, this string matching unit is according to described phonetic symbol character string, and in conjunction with the corresponding applicational language information of described intended application and described user's user related information, obtains the language character string matching with described phonetic symbol character string; Wherein, described user related information comprises following at least any one:

The preference setting of-described user to described applicational language information;

The history input record of-described user to described applicational language information.

Particularly, string matching unit is such as by carrying out alternately with user information database, carry out the modes such as mutual with user's input journal, obtain user's user related information, this user related information includes but not limited to that this user records to the preference setting of described applicational language information, to the history input of described applicational language information; String matching unit is according to the determined phonetic symbol character string of voice recognition unit, and in conjunction with the corresponding applicational language information of this intended application and described user's user related information, by modes such as matching inquiries, obtain the language character string matching with this phonetic symbol character string.

At this, this user to the preference setting example of this applicational language information as the input language type of this user to this intended application preference, as preference input Chinese, English or other language etc.; This user comprises that to the history input record of this applicational language information this user selects record etc. to the history of the historical input record of the voice of this intended application, the historical input record of word, input candidate text.

At this, in this user information database or user's input journal, store this user to the preference setting of this applicational language information, historical input record, or, the speech related information that this user is corresponding etc.

Preferably, string matching unit is according to the phonetic symbol character string after adjust processing, and in conjunction with the corresponding applicational language information of this intended application and this user's user related information, obtains the language character string matching with this phonetic symbol character string.

Those skilled in the art will be understood that above-mentioned user related information is only for giving an example, and other user related informations existing or that may occur from now on, as applicable to the present invention, also should be included in protection domain of the present invention, and with way of reference, are contained in this at this.

At this, the speech input information that the present invention inputs in intended application user carries out voice recognition processing, obtains phonetic symbol character string, and then in conjunction with this user's user related information, coupling obtains language character string, for user provides personalized experience.

Preferably, this string matching unit is according to described phonetic symbol character string, in the corresponding speech database of the enter ancillary information with described applicational language information and described subscriber equipment, carry out matching inquiry, obtain the language character string matching with described phonetic symbol character string; Wherein, described enter ancillary information comprises following at least any one:

The registered place information of-described subscriber equipment;

The information of roaming place of-described subscriber equipment.

At this, when this subscriber equipment 1 is mobile device, this mobile device also has the enter ancillary informations such as registered place information, information of roaming place.The registered place of this registered place this mobile device of information representation, by this registered place information, can know the speech habits etc. of the registered place of this mobile device; This information of roaming place characterizes the roaming location of this mobile device, by this information of roaming place, can know the speech habits of this mobile device roaming region wherein etc.This string matching unit is according to the determined phonetic symbol character string of voice recognition unit, in the corresponding speech database of the enter ancillary information with this applicational language information and this mobile device, carry out matching inquiry, in speech database as corresponding in the registered place information with this applicational language information and this mobile device, carry out matching inquiry, in the corresponding speech database of the information of roaming place with this applicational language information and this mobile device, carry out matching inquiry etc., obtain the language character string matching with this phonetic symbol character string.

For example, when the registered place of this mobile device is in City of South China, " en " and " eng " in the database that this enter ancillary information is corresponding, pre-nasal sounds such as " in " and " ing " and rear nasal sound have mapping relations, or, " ci " and " chi ", flat cacuminals such as " si " and " shi " has mapping relations, and applicational language information corresponding to this intended application is Chinese, string matching unit is according to the determined phonetic symbol character string of voice recognition unit, as " Ai1|Fen4 ", by carrying out matching inquiry in the corresponding speech database of the registered place information with this applicational language information and this mobile device, obtain this corresponding language character string of phonetic symbol character string " Ai1|Fen4 " " AiFen ", " AiFeng " etc.

Those skilled in the art will be understood that above-mentioned enter ancillary information is only for giving an example, and other enter ancillary informations existing or that may occur from now on, as applicable to the present invention, also should be included in protection domain of the present invention, and with way of reference, are contained in this at this.

At this, the speech input information that the present invention inputs in intended application user carries out voice recognition processing, obtain phonetic symbol character string, and then, in conjunction with the enter ancillary information of this mobile device, coupling obtains language character string, has further improved the accuracy rate of language character string coupling, and then improved the accuracy rate of follow-up input candidate text matches, promoted user's experience.

Preferably, this string matching unit is according to described phonetic symbol character string, and in conjunction with the corresponding applicational language information of described intended application, obtains the one or more candidate language character strings that match with described phonetic symbol character string; According to speech model, described one or more candidate language character strings are carried out to Screening Treatment, to obtain the language character string after Screening Treatment; Wherein, described text matches device 103, according to described language character string after Screening Treatment, carries out matching inquiry in described input dictionary, to obtain described one or more input candidate text.

Particularly, this string matching unit is according to the determined phonetic symbol character string of voice recognition unit, and in conjunction with the corresponding applicational language information of this intended application, by modes such as matching inquiries, as by carry out the modes such as matching inquiry in the speech database corresponding with this applicational language information, obtain the one or more candidate language character strings that match with this phonetic symbol character string.At this, this candidate language character string has a plurality of, for example hypothetical target is applied corresponding applicational language information for Chinese, this string matching unit is according to the determined phonetic symbol character string of voice recognition unit " chuai4 ", obtains a plurality of candidate language character strings " chuai " of matching with this phonetic symbol character string, " cuai " etc.

Subsequently, this string matching unit is according to speech model, as according to Chinese speech model, these one or more candidate language character strings are carried out to Screening Treatment, as deleted, obviously do not meet the language character string of logic of language, as above " cuai " in example, to obtain the language character string after Screening Treatment, as obtained the language character string " chuai " after Screening Treatment.

Then, text coalignment 103 is language character string after Screening Treatment according to this, in the corresponding input dictionary of the corresponding input method of this intended application, carry out matching inquiry, to obtain the one or more input candidate texts that match with this language character string.Connect example, text coalignment 103 is according to the language character string " chuai " after Screening Treatment, in the corresponding input in Chinese dictionary of the corresponding Chinese character coding input method of this intended application, carry out matching inquiry, obtain that input candidate text " is kicked ", " carrying " etc.

At this, the speech input information that the present invention inputs in intended application user carries out voice recognition processing, obtain phonetic symbol character string, and then, according to this phonetic symbol string matching, obtain one or more candidate language character strings, according to speech model, these one or more candidate language character strings are carried out to Screening Treatment, further improved the accuracy rate of language character string coupling, and then improved the accuracy rate of follow-up input candidate text matches, promoted user's experience.

Preferably, text coalignment 103 is according to described language character string, and in conjunction with the application knowledge of described intended application, carries out matching inquiry, to obtain described one or more input candidate text in described input dictionary.

Particularly, intended application also has certain application knowledge, and this application knowledge includes but not limited to the information such as the thematic knowledge, context of co-text of this intended application.For example, suppose that this corresponding applicational language information is for Chinese, the thematic knowledge of this intended application is " shopping ", the language character string that string matching units match obtains is " sougou ", text coalignment 103 is according to this language character string " sougou ", and in conjunction with the application knowledge of this intended application, in this input dictionary, carry out matching inquiry, obtain input candidate's text " collection "; If the context of co-text of this intended application is relevant with input method, text coalignment 103 is according to this language character string " sougou ", and in conjunction with the application knowledge of this intended application, in this input dictionary, carry out matching inquiry, obtain input candidate's text " search dog ".

At this, the application knowledge of combining target application of the present invention, coupling obtains input candidate text, has improved the accuracy rate of input candidate text matches, has promoted user's experience.

Fig. 2 illustrate in accordance with a preferred embodiment of the present invention for carry out the equipment schematic diagram of text input in intended application according to speech input information at subscriber equipment.This subscriber equipment 1 also comprises Target Acquisition device 205 and target input media 206.Referring to Fig. 2, the preferred embodiment is described in detail: particularly, voice acquisition device 201 obtains the speech input information that user inputs by the voice acquisition device of subscriber equipment, wherein, described speech input information is intended the text of inputting in the intended application of described subscriber equipment corresponding to described user; 202 pairs of described speech input informations of speech recognition equipment carry out voice recognition processing, to obtain the language character string corresponding with described speech input information; Text matches device 203, according to described language character string, carries out matching inquiry in the input dictionary corresponding with the corresponding input method of described intended application, to obtain the one or more input candidate texts that match with described language character string; Text generator 204 is by described one or more input candidate texts, at least one offers described intended application; Target Acquisition device 205 obtains described user selected target input text in described one or more input candidate texts; Target input media 206 inputs to described intended application by described target input text.Wherein, voice acquisition device 201, speech recognition equipment 202, text matches device 203 and text generator 204 are identical with corresponding intrument shown in Fig. 1 or basic identical, so locate to repeat no more, and mode is by reference contained in this.

Wherein, Target Acquisition device 205 obtains described user selected target input text in described one or more input candidate texts.

Particularly, at text generator 204 by after in one or more input candidate texts, at least one offers intended application, mutual by with this subscriber equipment 1 of this user, from this, at least one input candidate text, selected a target input text, at this, the mode of this user's select target input text includes but not limited to voice selecting, manual selection etc.; This Target Acquisition device 205, by call this subscriber equipment 1 application programming interfaces (API) that provide, voice acquisition device of this subscriber equipment 1 etc., obtains this user selected target input text in these one or more input candidate texts.

Target input media 206 inputs to described intended application by described target input text.Particularly, the target input text that target input media 206 obtains according to this Target Acquisition device 205, such as by calling dynamic web page techniques such as JSP, ASP or PHP, call text input function, calling modes such as showing control, this target input text is inputed to this intended application.

For example, hypothetical target is applied as " search application ", applicational language information corresponding to this intended application is Chinese, input method as corresponding in this intended application is current in input in Chinese state, user intends input " liking crazy " in the search box of this intended application, and it is inputted speech input information by voice acquisition device and " likes crazy "; Voice acquisition device 201 obtains the speech input information that this user inputs; Speech recognition equipment 202 determines that this speech input information " likes " corresponding phonetic symbol character string crazy for " Ai4|Feng1, then, by matching algorithm, obtains this corresponding language character string of phonetic symbol character string " Ai4|Feng1 " " AiFeng "; Subsequently, text matches device 203 is according to this language character string " AiFeng ", in the input dictionary corresponding with this " search application " corresponding Chinese character coding input method, carry out matching inquiry, obtain that a plurality of input candidate texts that match with this language character string " are liked crazy ", " Ai Feng ", " love wind " etc.; Then, text generator 204 offers user by the plurality of input candidate text, as the mode with numbering, the plurality of input candidate text is offered to this user with forms such as " 1) like crazy ", " 2) Ai Feng ", " 3) love wind "; Suppose that this user is by the mode of voice selecting, select " 1) ", selected target candidate text " to like crazy ", Target Acquisition device 205 obtains the selection that this user does; Subsequently, target input media 206 " likes " to input to this intended application " search application " crazy by this target input text, as input in the search box of this " search application ", should " search application " directly according to this target input text, " like " to search for crazy, or, according to user, operation or the phonetic order of search button are searched for.

At this, the present invention combines with application, has further promoted user's experience.

Preferably, this subscriber equipment 1 also comprises model modification device (not shown), and this model modification device, according to described user-selected target input text, is determined the target language character string corresponding with described target input text; According to the mapping relations of described target language character string and the target phonetic symbol character string corresponding with it, set up or upgrade described speech model.

Particularly, mutual by with this subscriber equipment 1 of user selected a target input text from these one or more input candidate texts, and at this, the mode of this user's select target input text includes but not limited to voice selecting, manually selection etc.; Model modification device is by call this subscriber equipment 1 application programming interfaces (API) that provide, voice acquisition device of this subscriber equipment 1 etc., obtain this user selected target input text in these one or more input candidate texts, or, model modification device is mutual by with Target Acquisition device 205 directly, obtains this user-selected target input text; Subsequently, model modification device, according to this target input text, such as by carry out the modes such as reverse find in input dictionary, is determined the target language character string corresponding with this target input text; And then, according to this target language character string, again by carry out the modes such as reverse find in speech database, determine the target phonetic symbol character string corresponding with this target language character string, subsequently, according to the mapping relations of this target language character string and this target phonetic symbol character string, set up or upgrade this speech model.

For example, user has selected target input text " to like crazy ", model modification device is according to this target input text, determine the target language character string " AiFeng " that it is corresponding, and then, according to target language character string " AiFeng ", determine the target phonetic symbol character string " Ai4Feng1 " that it is corresponding, subsequently, this model modification device is according to the mapping relations of this target language character string " AiFeng " and this target phonetic symbol character string " Ai4Feng1 ", sets up or upgrades this speech model.

Preferably, this subscriber equipment 1 also comprises priority determining device (not shown), and this priority determining device, according to predetermined priority rule, is determined the priority of described one or more input candidate texts; Wherein, text generator 204 is according to described priority, by described one or more input candidate texts, at least one offers described intended application; Wherein, described predetermined priority rule based on below at least any one determine:

-described user's text input historical information;

-described applicational language information;

The input state information of-described input method.

For example, suppose that the text of predetermined priority rule based on this user input historical information determines, for the one or more input candidate texts that offer user, the priority of the input candidate text that historical input number of times is more is higher, or the priority of the input candidate text of the last input of this user is higher.

And for example, suppose that predetermined priority rule determines based on this applicational language information, for the one or more input candidate texts that offer user, higher with the priority of the higher input candidate text of the degree of correlation of this intended application.

For another example, suppose that the input state information of predetermined priority rule based on this input method determines, for the one or more input candidate texts that offer user, higher with the priority of the higher input candidate text of the degree of correlation of the input state information of this input method, as under English input state, the priority of English candidate's text is higher.

Preferably, predetermined priority rule can also in conjunction with above-mentioned any two factors or whole three because usually determining, priority determining device, according to this predetermined priority rule, determines that text coalignment 203 mates the priority of the one or more input candidate texts that obtain.

Subsequently, text generator 204 is according to this priority, by these one or more input candidate texts, at least one offers this intended application.

Those skilled in the art will be understood that the above-mentioned factor that is used for determining predetermined priority rule is only for giving an example; the factor that other existing or being used for of may occurring are from now on determined predetermined priority rule is as applicable to the present invention; also should be included in protection domain of the present invention, and with way of reference, be contained in this at this.

Fig. 3 illustrate according to a further aspect of the present invention for carry out the method flow diagram of text input in intended application according to speech input information at subscriber equipment.

In step S301, subscriber equipment 1 obtains the speech input information that user inputs by the voice acquisition device of subscriber equipment, and wherein, described speech input information is intended the text of inputting in the intended application of described subscriber equipment corresponding to described user.Particularly, mutual by with subscriber equipment 1 of user, by the voice acquisition device input speech input information on it, as inputted speech input information by microphone, this speech input information is intended the text of inputting in the intended application of this subscriber equipment corresponding to this user, as this user intends input " iPhone " in the search box of intended application " search ", it inputs speech input information " iPhone " by voice acquisition device; In step S301, subscriber equipment 1 for example, by calling application programming interfaces (API) that this subscriber equipment provides or the communication mode of other agreements, obtains the speech input information " iPhone " that this user inputs.

In step S302,1 pair of described speech input information of subscriber equipment carries out voice recognition processing, to obtain the language character string corresponding with described speech input information.Particularly, in step S302, subscriber equipment 1 is according to the speech input information obtaining in step S301, by this speech input information is carried out to voice recognition processing, as by carry out matching inquiry in the speech database of this subscriber equipment, obtain the language character string corresponding with this speech input information; Or, in step S302, subscriber equipment 1 is according to the speech input information obtaining in step S301, this speech input information is carried out to voice recognition processing, as the voice data to this speech input information is resolved, acquisition can represent the phonetic symbol character string of the corresponding pronunciation of this speech input information, and then, according to this phonetic symbol character string, obtain the language character string matching with described phonetic symbol character string.

Connect precedent, in step S301, subscriber equipment 1 obtains the speech input information " iPhone " that this user inputs; In step S302,1 pair of this speech input information of subscriber equipment " iPhone " carries out voice recognition processing, as according to this speech input information " iPhone ", in the speech database of this subscriber equipment, carry out matching inquiry, obtain the language character string corresponding with this speech input information " iPhone ", as " iphone ", " AiFen " etc.

And for example, in step S302,1 pair of speech input information obtaining in step S301 of subscriber equipment " iPhone " carries out voice recognition processing, as this speech input information " iPhone " is simplified according to the algorithm of formulating, obtain for the more needed characteristic of model, again the audio model data of storing in this characteristic and this subscriber equipment 1 are compared, the phonetic symbol character string of the pronunciation that represents this speech input information " iPhone " that acquisition meets, as " Ai1|Fen4 "; Then,, in step S302, subscriber equipment 1, according to this phonetic symbol character string, by modes such as matching inquiries, obtains the language character string matching with described phonetic symbol character string.Preferably, in step S302, subscriber equipment 1 can also be in conjunction with the corresponding applicational language information of this intended application, obtain the language character string matching with this phonetic symbol character string, as carried out the modes such as matching inquiry in the speech database by corresponding with this applicational language information, obtain the language character string matching with this phonetic symbol character string, while being English as the applicational language information when intended application, obtain this corresponding language character string of phonetic symbol character string " Ai1|Fen4 " " iphone "; When the applicational language information of this intended application is Chinese, obtain this corresponding language character string of phonetic symbol character string " Ai1|Fen4 " " AiFen ".

In step S303, subscriber equipment 1, according to described language character string, carries out matching inquiry in the input dictionary corresponding with the corresponding input method of described intended application, to obtain the one or more input candidate texts that match with described language character string.Particularly, in step S303, subscriber equipment 1 is according to mate obtained language character string in step S302, by the mode of the matching inquiries such as exact matching, broad match, in the input dictionary corresponding with the corresponding input method of this intended application, carry out matching inquiry, to obtain the one or more input candidate texts that match with this language character string.

Connect example, when the applicational language information of this intended application is English, in step S302, subscriber equipment 1 obtains this corresponding language character string of phonetic symbol character string " Ai1|Fen4 " " iphone "; Subsequently, in step S303, subscriber equipment 1 is according to this language character string " iphone ", in the input dictionary corresponding with the corresponding English input method of this intended application, carry out matching inquiry, a plurality of input candidate texts " iphone " that acquisition and this language character string match, " ip ", " i " etc.

If the applicational language information of this intended application is Chinese,, in step S302, subscriber equipment 1 obtains this corresponding language character string of phonetic symbol character string " Ai1|Fen4 " " AiFen "; Subsequently, in step S303, subscriber equipment 1 is according to this language character string " AiFen ", in the input dictionary corresponding with the corresponding Chinese character coding input method of this intended application, carry out matching inquiry, a plurality of input candidate texts that acquisition and this language character string match " love is fragrant ", " liking to divide ", " Ai Fen " etc.

At this, in step S303, the mode that subscriber equipment 1 coupling in input dictionary obtains one or more input candidate texts includes but not limited to the modes such as exact matching, broad match.Exact matching for example this language character string mates completely with this input candidate text, as " iphone " and " iphone ", " AiFen " and " liking fragrant " etc.; Broad match for example this language character string is mated with this input candidate textual portions, as " iphone " and " ip ", " AiFen " and " liking crazy " etc.

In step S304, subscriber equipment 1 is by described one or more input candidate texts, at least one offers described intended application.Particularly, in step S304, subscriber equipment 1 is according to mate the one or more input candidate texts that obtain in step S303, such as by calling dynamic web page techniques such as JSP, ASP or PHP, or by calling modes such as showing control, by in these one or more input candidate texts at least one, offer this intended application with form in certain sequence.For example, in step S304, subscriber equipment 1 is by calling demonstration control, by these one or more input candidate texts at least one, in the entry hurdle of the corresponding input method of this intended application, offer this intended application.At this, this entry hurdle for example only shows a line input candidate text, in this row entry hurdle, the number of shown input candidate text can be default also can being set by the user, by pressed special function keys by user, show lastrow or next line input candidate text, this special function keys can be for example "+" and "-"; Or in step S304, subscriber equipment 1 can offer this intended application by least one form with multirow entry hurdle in these one or more input candidate texts according to the screen width of this subscriber equipment 1, further processes for this intended application.

Preferably, between each step of subscriber equipment 1, be constant work.Particularly, in step S301, subscriber equipment 1 obtains the speech input information that user inputs by the voice acquisition device of subscriber equipment, and wherein, described speech input information is intended the text of inputting in the intended application of described subscriber equipment corresponding to described user; In step S302,1 pair of described speech input information of subscriber equipment carries out voice recognition processing, to obtain the language character string corresponding with described speech input information; In step S303, subscriber equipment 1, according to described language character string, carries out matching inquiry in the input dictionary corresponding with the corresponding input method of described intended application, to obtain the one or more input candidate texts that match with described language character string; In step S304, subscriber equipment 1 is by described one or more input candidate texts, at least one offers described intended application.At this, it will be understood by those skilled in the art that " continuing " refers to that each step of subscriber equipment 1 carries out in real time or constantly respectively the acquisition of the obtaining of speech input information, voice recognition processing, language character string and the coupling of input candidate text and provide, until this subscriber equipment 1 stops obtaining the speech input information that user inputs by the voice acquisition device of subscriber equipment in a long time.

At this, the speech input information that the present invention inputs in intended application user carries out voice recognition processing, obtain corresponding language character string, again in conjunction with the corresponding input dictionary of input method, coupling obtains corresponding input candidate text, improved the efficiency with input candidate text matches in the speech recognition of subscriber equipment end, improved the utilization rate of local resource, reduced the occupancy of other resources such as this locality storage, network service, the mode combining with input method has also improved the accuracy rate of input candidate text matches, has promoted user's experience.

Preferably, step S302 also comprises sub-step S302a (not shown) and sub-step S302b (not shown).

In sub-step S302a, 1 pair of described speech input information of subscriber equipment carries out voice recognition processing, to obtain the phonetic symbol character string corresponding with described speech input information.Particularly, in sub-step S302a, subscriber equipment 1 is according to the speech input information obtaining in step S301, this speech input information is carried out to voice recognition processing, as the voice data to this speech input information is resolved, acquisition can represent the phonetic symbol character string of the corresponding pronunciation of this speech input information.

For example, in step S301, subscriber equipment 1 obtains the speech input information " iPhone " that this user inputs; In sub-step S302a, 1 pair of this speech input information of subscriber equipment " iPhone " carries out voice recognition processing, as this speech input information " iPhone " is simplified according to the algorithm of formulating, obtain for the more needed characteristic of model, again the audio model data of storing in this characteristic and this subscriber equipment 1 are compared, the phonetic symbol character string of the pronunciation that represents this speech input information " iPhone " that acquisition meets, as " Ai1|Fen4 ".

In sub-step S302b, subscriber equipment 1, according to described phonetic symbol character string, obtains the language character string matching with described phonetic symbol character string.Particularly, in sub-step S302b, subscriber equipment 1, according to the phonetic symbol character string of acquisition that this voice recognition unit is identified, by modes such as matching inquiries, obtains the language character string matching with described phonetic symbol character string.

For example, suppose in step S301, the user's that subscriber equipment 1 obtains speech input information is " iPhone ", in sub-step S302a, subscriber equipment 1 determines that the corresponding phonetic symbol character string of this speech input information " iPhone " is " Ai1|Fen4 ", in sub-step S302b, subscriber equipment 1 is according to the phonetic symbol character string of acquisition that this voice recognition unit is identified, by modes such as matching inquiries, obtain the language character string matching with described phonetic symbol character string, as " iphone ", " AiFen " etc.

Preferably, the method also comprises S307 (not shown), in step S307, subscriber equipment 1 is according to described phonetic symbol character string, and in conjunction with the speech related information of described speech input information, described phonetic symbol character string is adjusted to processing, to obtain through adjusting the phonetic symbol character string after processing; Wherein, in sub-step S302b, subscriber equipment 1 through adjusting the phonetic symbol character string after processing, obtains the described language character string matching with described phonetic symbol character string according to described.

Particularly, in step S307, subscriber equipment 1 is according to the speech input information obtaining in step S301, by this speech input information is carried out to speech analysis, or, by carrying out matching inquiry in the user information database the corresponding user of this speech input information, obtain the speech related information of this speech input information, this speech related information includes but not limited to accent information, element information, voiceprint of the corresponding user of this speech input information etc.; In step S307, subscriber equipment 1 is according to determined phonetic symbol character string in sub-step S302a, one or more in conjunction with in the above-mentioned speech related information of speech input information again, this phonetic symbol character string is adjusted to processing, as when the corresponding applicational language information of this intended application being when Chinese, the front and back nasal sound of this speech input information, flat cacuminal etc. are adjusted to processing, to obtain through adjusting the phonetic symbol character string after processing; For example, suppose in sub-step S302a, subscriber equipment 1 is " Cuan1Yue4 " according to the determined phonetic symbol character string of language input message, in step S307 in conjunction with the speech related information of this speech input information, this phonetic symbol character string is adjusted to processing, as be adjusted to " Chuan1Yue4 "; Subsequently, in sub-step S302b, subscriber equipment 1 through adjusting the phonetic symbol character string after processing, obtains the described language character string matching with this phonetic symbol character string according to this.

The accent information of-described speech input information;

The element information of-described speech input information;

The voiceprint of-described speech input information.

At this, the voice accent information of described speech input information is as the speech with a guy, place or native language feature, in step S307, subscriber equipment 1 is according to determined phonetic symbol character string in sub-step S302a, in conjunction with the accent information of speech input information, this phonetic symbol character string is adjusted to processing again.

The phonetic feature information of described speech input information includes but not limited to the element characteristics such as pitch, loudness of a sound, the duration of a sound and tone color of this speech input information, in step S307, subscriber equipment 1 is according to determined phonetic symbol character string in sub-step S302a, in conjunction with the element information of speech input information, this phonetic symbol character string is adjusted to processing again.

The voice voiceprint of described speech input information is the sound wave spectrum that carries language message, in step S307, subscriber equipment 1 is according to determined phonetic symbol character string in sub-step S302a, then in conjunction with the voiceprint of speech input information, this phonetic symbol character string adjusted to processing.

More preferably, in sub-step S302b, subscriber equipment 1 is according to described phonetic symbol character string, and in conjunction with the corresponding applicational language information of described intended application, obtains the language character string matching with described phonetic symbol character string.At this, this applicational language information is for characterizing the information such as current languages input state of the input method such as language environment, correspondence of this intended application.

Particularly, in sub-step S302b, subscriber equipment 1 is according to identify the phonetic symbol character string obtaining in sub-step S302a, and in conjunction with the corresponding applicational language information of this intended application, by modes such as matching inquiries, as carried out the modes such as matching inquiry in the speech database by corresponding with this applicational language information, obtain the language character string matching with described phonetic symbol character string.

For example, suppose in step S301, the user's that subscriber equipment 1 obtains speech input information is " iPhone ", in sub-step S302a, subscriber equipment 1 determines that the corresponding phonetic symbol character string of this speech input information " iPhone " is " Ai1|Fen4 ", when the applicational language information of intended application is English, in sub-step S302b, subscriber equipment 1, by matching algorithm, obtains this corresponding language character string of phonetic symbol character string " Ai1|Fen4 " " iphone "; If the applicational language information of this intended application is Chinese,, in sub-step S302b, subscriber equipment 1 obtains this corresponding language character string of phonetic symbol character string " Ai1|Fen4 " " AiFen "; If the language that the applicational language information of this intended application is other languages,, in sub-step S302b, subscriber equipment 1, according to this phonetic symbol character string " Ai1|Fen4 ", obtains the language character string of other languages corresponding with this phonetic symbol character string.

Preferably, in sub-step S302b, subscriber equipment 1 is according to described phonetic symbol character string, and in conjunction with the corresponding applicational language information of described intended application and described user's user related information, obtains the language character string matching with described phonetic symbol character string; Wherein, described user related information comprises following at least any one:

Particularly, in sub-step S302b, subscriber equipment 1 is such as by carrying out alternately with user information database, carry out the modes such as mutual with user's input journal, obtain user's user related information, this user related information includes but not limited to that this user records to the preference setting of described applicational language information, to the history input of described applicational language information; In sub-step S302b, subscriber equipment 1 is according to determined phonetic symbol character string in sub-step S302a, and in conjunction with the corresponding applicational language information of this intended application and described user's user related information, by modes such as matching inquiries, obtain the language character string matching with this phonetic symbol character string.

Preferably, in sub-step S302b, subscriber equipment 1 is according to the phonetic symbol character string after adjust processing, and in conjunction with the corresponding applicational language information of this intended application and this user's user related information, obtains the language character string matching with this phonetic symbol character string.

Preferably, in sub-step S302b, subscriber equipment 1, according to described phonetic symbol character string, carries out matching inquiry in the corresponding speech database of the enter ancillary information with described applicational language information and described subscriber equipment, obtains the language character string matching with described phonetic symbol character string; Wherein, described enter ancillary information comprises following at least any one:

The registered place information of-described subscriber equipment;

The information of roaming place of-described subscriber equipment.

At this, when this subscriber equipment 1 is mobile device, this mobile device also has the enter ancillary informations such as registered place information, information of roaming place.The registered place of this registered place this mobile device of information representation, by this registered place information, can know the speech habits etc. of the registered place of this mobile device; This information of roaming place characterizes the roaming location of this mobile device, by this information of roaming place, can know the speech habits of this mobile device roaming region wherein etc.In sub-step S302b, mobile device is according to determined phonetic symbol character string in sub-step S302a, in the corresponding speech database of the enter ancillary information with this applicational language information and this mobile device, carry out matching inquiry, in speech database as corresponding in the registered place information with this applicational language information and this mobile device, carry out matching inquiry, in the corresponding speech database of the information of roaming place with this applicational language information and this mobile device, carry out matching inquiry etc., obtain the language character string matching with this phonetic symbol character string.

For example, when the registered place of this mobile device is in City of South China, " en " and " eng " in the database that this enter ancillary information is corresponding, pre-nasal sounds such as " in " and " ing " and rear nasal sound have mapping relations, or, " ci " and " chi ", flat cacuminals such as " si " and " shi " has mapping relations, and applicational language information corresponding to this intended application is Chinese, in sub-step S302b, mobile device is according to determined phonetic symbol character string in sub-step S302a, as " Ai1|Fen4 ", by carrying out matching inquiry in the corresponding speech database of the registered place information with this applicational language information and this mobile device, obtain this corresponding language character string of phonetic symbol character string " Ai1|Fen4 " " AiFen ", " AiFeng " etc.

At this, the speech input information that the present invention inputs in intended application user carries out voice recognition processing, obtain phonetic symbol character string, and then, in conjunction with the enter ancillary information of this subscriber equipment, coupling obtains language character string, has further improved the accuracy rate of language character string coupling, and then improved the accuracy rate of follow-up input candidate text matches, promoted user's experience.

Preferably, in sub-step S302b, subscriber equipment 1 is according to described phonetic symbol character string, and in conjunction with the corresponding applicational language information of described intended application, obtains the one or more candidate language character strings that match with described phonetic symbol character string; Subsequently, subscriber equipment 1, according to speech model, carries out Screening Treatment to described one or more candidate language character strings, to obtain the language character string after Screening Treatment; Then, in step S303, subscriber equipment 1, according to described language character string after Screening Treatment, carries out matching inquiry in described input dictionary, to obtain described one or more input candidate text.

Particularly, in sub-step S302b, subscriber equipment 1 is according to determined phonetic symbol character string in sub-step S302a, and in conjunction with the corresponding applicational language information of this intended application, by modes such as matching inquiries, as by carry out the modes such as matching inquiry in the speech database corresponding with this applicational language information, obtain the one or more candidate language character strings that match with this phonetic symbol character string.At this, this candidate language character string has a plurality of, for example hypothetical target is applied corresponding applicational language information for Chinese, in sub-step S302b, subscriber equipment 1 basis determined phonetic symbol character string " chuai4 " in sub-step S302a, a plurality of candidate language character strings " chuai " that acquisition and this phonetic symbol character string match, " cuai " etc.

Subsequently, in sub-step S302b, subscriber equipment 1 is according to speech model, as according to Chinese speech model, these one or more candidate language character strings are carried out to Screening Treatment, as deleted, obviously do not meet the language character string of logic of language, as above " cuai " in example, to obtain the language character string after Screening Treatment, as obtained the language character string " chuai " after Screening Treatment.

Then, in step S303, subscriber equipment 1 is language character string after Screening Treatment according to this, in the corresponding input dictionary of the corresponding input method of this intended application, carries out matching inquiry, to obtain the one or more input candidate texts that match with this language character string.Connect example, in step S303, subscriber equipment 1, according to the language character string " chuai " after Screening Treatment, carries out matching inquiry in the corresponding input in Chinese dictionary of the corresponding Chinese character coding input method of this intended application, obtains that input candidate text " is kicked ", " carrying " etc.

Preferably, in step S303, subscriber equipment 1 is according to described language character string, and in conjunction with the application knowledge of described intended application, carries out matching inquiry, to obtain described one or more input candidate text in described input dictionary.

Particularly, intended application also has certain application knowledge, and this application knowledge includes but not limited to the information such as the thematic knowledge, context of co-text of this intended application.For example, suppose that this corresponding applicational language information is for Chinese, the thematic knowledge of this intended application is " shopping ", in sub-step S302b, the language character string that subscriber equipment 1 coupling obtains is for " sougou ", in step S303, subscriber equipment 1 is according to this language character string " sougou ", and in conjunction with the application knowledge of this intended application, in this input dictionary, carry out matching inquiry, obtain input candidate's text " collection "; If the context of co-text of this intended application is relevant with input method, in step S303, subscriber equipment 1 is according to this language character string " sougou ", and in conjunction with the application knowledge of this intended application, in this input dictionary, carry out matching inquiry, obtain input candidate's text " search dog ".

Fig. 4 illustrate in accordance with a preferred embodiment of the present invention for carry out the method flow diagram of text input in intended application according to speech input information at subscriber equipment.Referring to Fig. 4, the preferred embodiment is described in detail: particularly, in step S401, subscriber equipment 1 obtains the speech input information that user inputs by the voice acquisition device of subscriber equipment, wherein, described speech input information is intended the text of inputting in the intended application of described subscriber equipment corresponding to described user; In step S402,1 pair of described speech input information of subscriber equipment carries out voice recognition processing, to obtain the language character string corresponding with described speech input information; In step S403, subscriber equipment 1, according to described language character string, carries out matching inquiry in the input dictionary corresponding with the corresponding input method of described intended application, to obtain the one or more input candidate texts that match with described language character string; In step S404, subscriber equipment 1 offers described user by described one or more input candidate texts; In step S405, subscriber equipment 1 obtains described user selected target input text in described one or more input candidate texts; In step S406, subscriber equipment 1 inputs to described intended application by described target input text.Wherein, step S401-S404 is identical or basic identical with corresponding step shown in Fig. 3 respectively, so locate to repeat no more, and mode is by reference contained in this.

Wherein, in step S405, subscriber equipment 1 obtains described user selected target input text in described one or more input candidate texts.

Particularly, in step S404, subscriber equipment 1 is by after in one or more input candidate texts, at least one offers intended application, mutual by with this subscriber equipment 1 of this user, from this, at least one input candidate text, selected a target input text, at this, the mode of this user's select target input text includes but not limited to voice selecting, manual selection etc.; In step S405, subscriber equipment 1, by call this subscriber equipment 1 application programming interfaces (API) that provide, voice acquisition device of this subscriber equipment 1 etc., obtains this user selected target input text in these one or more input candidate texts.

In step S406, subscriber equipment 1 inputs to described intended application by described target input text.Particularly, in step S406, subscriber equipment 1 is according to the target input text obtaining in step S405, such as by calling dynamic web page techniques such as JSP, ASP or PHP, call text input function, calling modes such as showing control, this target input text is inputed to this intended application.

For example, hypothetical target is applied as " search application ", applicational language information corresponding to this intended application is Chinese, input method as corresponding in this intended application is current in input in Chinese state, user intends input " liking crazy " in the search box of this intended application, and it is inputted speech input information by voice acquisition device and " likes crazy "; In step S401, subscriber equipment 1 obtains the speech input information that this user inputs; In step S402, subscriber equipment 1 determine this speech input information " like " corresponding phonetic symbol character string crazy for " Ai4|Feng1; then, subscriber equipment 1, by matching algorithm, obtains this corresponding language character string of phonetic symbol character string " Ai4|Feng1 " " AiFeng "; Subsequently, in step S403, subscriber equipment 1 is according to this language character string " AiFeng ", in the input dictionary corresponding with this " search application " corresponding Chinese character coding input method, carry out matching inquiry, obtain that a plurality of input candidate texts that match with this language character string " are liked crazy ", " Ai Feng ", " love wind " etc.; Then, in step S404, subscriber equipment 1 offers user by the plurality of input candidate text, as the mode with numbering, the plurality of input candidate text is offered to this user with forms such as " 1) like crazy ", " 2) Ai Feng ", " 3) love wind "; Suppose that this user is by the mode of voice selecting, select " 1) ", selected target candidate text " to like crazy ", in step S405, subscriber equipment 1 obtains the selection that this user does; Subsequently, in step S406, subscriber equipment 1 " likes " to input to this intended application " search application " crazy by this target input text, as input in the search box of this " search application ", should " search application " directly according to this target input text, " like " to search for crazy, or, according to user, operation or the phonetic order of search button are searched for.

Preferably, the method also comprises step S408 (not shown), and in step S408, subscriber equipment 1, according to described user-selected target input text, is determined the target language character string corresponding with described target input text; According to the mapping relations of described target language character string and the target phonetic symbol character string corresponding with it, set up or upgrade described speech model.

Particularly, mutual by with this subscriber equipment 1 of user selected a target input text from these one or more input candidate texts, and at this, the mode of this user's select target input text includes but not limited to voice selecting, manually selection etc.; In step S408, subscriber equipment 1 is by call this subscriber equipment 1 application programming interfaces (API) that provide, voice acquisition device of this subscriber equipment 1 etc., obtain this user selected target input text in these one or more input candidate texts, or, in step S408, subscriber equipment 1 directly obtains this user-selected target input text obtaining in step S405; Subsequently, in step S408, subscriber equipment 1, according to this target input text, such as by carry out the modes such as reverse find in input dictionary, is determined the target language character string corresponding with this target input text; And then, according to this target language character string, again by carry out the modes such as reverse find in speech database, determine the target phonetic symbol character string corresponding with this target language character string, subsequently, according to the mapping relations of this target language character string and this target phonetic symbol character string, set up or upgrade this speech model.

For example, user has selected target input text " to like crazy ", in step S408, subscriber equipment 1 is according to this target input text, determine the target language character string " AiFeng " that it is corresponding, and then, according to target language character string " AiFeng ", determine the target phonetic symbol character string " Ai4Feng1 " that it is corresponding, subsequently, in step S408, subscriber equipment 1 is according to the mapping relations of this target language character string " AiFeng " and this target phonetic symbol character string " Ai4Feng1 ", sets up or upgrades this speech model.

Preferably, the method also comprises S409 (not shown), and in step S409, subscriber equipment 1, according to predetermined priority rule, is determined the priority of described one or more input candidate texts; Wherein, in step S404, subscriber equipment 1 is according to described priority, by described one or more input candidate texts, at least one offers described intended application; Wherein, described predetermined priority rule based on below at least any one determine:

-described user's text input historical information;

-described applicational language information;

The input state information of-described input method.

Preferably, predetermined priority rule can also in conjunction with above-mentioned any two factors or whole three because usually determining, in step S409, subscriber equipment 1, according to this predetermined priority rule, is determined the priority of mating the one or more input candidate texts that obtain in step S403.

Subsequently, in step S404, subscriber equipment 1 is according to this priority, by these one or more input candidate texts, at least one offers this intended application.

It should be noted that the present invention can be implemented in the assembly of software and/or software and hardware, for example, can adopt special IC (ASIC), general object computing machine or any other similar hardware device to realize.In one embodiment, software program of the present invention can carry out to realize step mentioned above or function by processor.Similarly, software program of the present invention (comprising relevant data structure) can be stored in computer readable recording medium storing program for performing, for example, and RAM storer, magnetic or CD-ROM driver or flexible plastic disc and similar devices.In addition, steps more of the present invention or function can adopt hardware to realize, for example, thereby as coordinate the circuit of carrying out each step or function with processor.

In addition, a part of the present invention can be applied to computer program, and for example computer program instructions, when it is carried out by computing machine, by the operation of this computing machine, can call or provide the method according to this invention and/or technical scheme.And call the programmed instruction of method of the present invention, may be stored in fixing or movably in recording medium, and/or be transmitted by the data stream in broadcast or other signal bearing medias, and/or be stored in according in the working storage of the computer equipment of described programmed instruction operation.At this, comprise according to one embodiment of present invention a device, this device comprises for storing the storer of computer program instructions and for the processor of execution of program instructions, wherein, when this computer program instructions is carried out by this processor, trigger this device and move based on aforementioned according to the method for a plurality of embodiment of the present invention and/or technical scheme.

To those skilled in the art, obviously the invention is not restricted to the details of above-mentioned one exemplary embodiment, and in the situation that not deviating from spirit of the present invention or essential characteristic, can realize the present invention with other concrete form.Therefore, no matter from which point, all should regard embodiment as exemplary, and be nonrestrictive, scope of the present invention is limited by claims rather than above-mentioned explanation, is therefore intended to be included in the present invention dropping on the implication that is equal to important document of claim and all changes in scope.Any Reference numeral in claim should be considered as limiting related claim.In addition, obviously other unit or step do not got rid of in " comprising " word, and odd number is not got rid of plural number.A plurality of unit of stating in device claim or device also can be realized by software or hardware by a unit or device.The first, the second word such as grade is used for representing title, and does not represent any specific order.

Claims

1. for carry out a method for text input in intended application according to speech input information at subscriber equipment, wherein, the method comprises the following steps:

2. method according to claim 1, wherein, described step b comprises:

B1 carries out voice recognition processing to described speech input information, to obtain the phonetic symbol character string corresponding with described speech input information;

B2, according to described phonetic symbol character string, obtains the language character string matching with described phonetic symbol character string.

3. method according to claim 2, wherein, the method also comprises:

-according to described phonetic symbol character string, and in conjunction with the speech related information of described speech input information, described phonetic symbol character string is adjusted to processing, to obtain through adjusting the phonetic symbol character string after processing;

Wherein, described step b2 comprises:

-according to described phonetic symbol character string after adjusting processing, obtain the described language character string matching with described phonetic symbol character string.

4. method according to claim 3, wherein, described speech related information comprises following at least any one:

The accent information of-described speech input information;

The element information of-described speech input information;

The voiceprint of-described speech input information.

5. method according to claim 2, wherein, described step b2 comprises:

-according to described phonetic symbol character string, and in conjunction with the corresponding applicational language information of described intended application, obtain the language character string matching with described phonetic symbol character string.

6. method according to claim 5, wherein, described step b2 comprises:

-according to described phonetic symbol character string, and in conjunction with the corresponding applicational language information of described intended application and described user's user related information, obtain the language character string matching with described phonetic symbol character string;

Wherein, described user related information comprises following at least any one:

7. method according to claim 5, wherein, described step b2 comprises:

-according to described phonetic symbol character string, in the corresponding speech database of the enter ancillary information with described applicational language information and described subscriber equipment, carry out matching inquiry, obtain the language character string matching with described phonetic symbol character string;

Wherein, described enter ancillary information comprises following at least any one:

The registered place information of-described subscriber equipment;

The information of roaming place of-described subscriber equipment.

8. method according to claim 5, wherein, described step b2 comprises:

-according to described phonetic symbol character string, and in conjunction with the corresponding applicational language information of described intended application, obtain the one or more candidate language character strings that match with described phonetic symbol character string;

-according to speech model, described one or more candidate language character strings are carried out to Screening Treatment, to obtain the language character string after Screening Treatment;

Wherein, described step c comprises:

-according to described language character string after Screening Treatment, in described input dictionary, carry out matching inquiry, to obtain described one or more input candidate text.

9. according to the method described in any one in claim 1 to 8, wherein, the method also comprises:

-obtain described user selected target input text in described one or more input candidate texts;

-described target input text is inputed to described intended application.

10. method according to claim 9, wherein, claim 9 comprises method according to claim 8, wherein, the method also comprises:

-according to described user-selected target input text, determine the target language character string corresponding with described target input text;

-according to the mapping relations of described target language character string and the target phonetic symbol character string corresponding with it, set up or upgrade described speech model.

11. according to the method described in any one in claim 1 to 10, and wherein, described step c comprises:

-according to described language character string, and in conjunction with the application knowledge of described intended application, in described input dictionary, carry out matching inquiry, to obtain described one or more input candidate text.

12. according to the method described in any one in claim 1 to 11, and wherein, the method also comprises:

-according to predetermined priority rule, determine the priority of described one or more input candidate texts;

Wherein, described steps d comprises:

-according to described priority, by described one or more input candidate texts, at least one offers described intended application;

Wherein, described predetermined priority rule based on below at least any one determine:

-described user's text input historical information;

-described applicational language information;

The input state information of-described input method.

13. 1 kinds are carried out the subscriber equipment of text input in intended application according to speech input information, wherein, this equipment comprises:

14. subscriber equipmenies according to claim 13, wherein, described speech recognition equipment comprises:

Voice recognition unit, for described speech input information is carried out to voice recognition processing, to obtain the phonetic symbol character string corresponding with described speech input information;

String matching unit, for according to described phonetic symbol character string, obtains the language character string matching with described phonetic symbol character string.

15. subscriber equipmenies according to claim 14, wherein, this equipment also comprises:

Adjust treating apparatus, for according to described phonetic symbol character string, and in conjunction with the speech related information of described speech input information, described phonetic symbol character string is adjusted to processings, to obtain the phonetic symbol character string after adjustment processing;

Wherein, described string matching unit is used for:

16. subscriber equipmenies according to claim 15, wherein, described speech related information comprises following at least any one:

The accent information of-described speech input information;

The element information of-described speech input information;

The voiceprint of-described speech input information.

17. subscriber equipmenies according to claim 14, wherein, described string matching unit is used for:

18. subscriber equipmenies according to claim 17, wherein, described string matching unit is used for:

19. subscriber equipmenies according to claim 17, wherein, described string matching unit is used for:

The registered place information of-described subscriber equipment;

The information of roaming place of-described subscriber equipment.

20. subscriber equipmenies according to claim 17, wherein, described string matching unit is used for:

Wherein, described text matches device is used for:

21. according to claim 13 to the subscriber equipment described in any one in 20, and wherein, this equipment also comprises:

Target Acquisition device, for obtaining described user from the selected target input text of described one or more input candidate text;

Target input media, for inputing to described intended application by described target input text.

22. subscriber equipmenies according to claim 21, wherein, claim 21 comprises subscriber equipment according to claim 20, wherein, this equipment also comprises model modification device, for:

23. according to claim 13 to the subscriber equipment described in any one in 22, and wherein, described text matches device is used for:

24. according to claim 13 to the subscriber equipment described in any one in 23, and wherein, this equipment also comprises priority determining device, for:

Wherein, described text generator is used for:

-described user's text input historical information;

-described applicational language information;

The input state information of-described input method.