CN106778927A - Update TV semantics recognition dictionary method and device - Google Patents
Update TV semantics recognition dictionary method and device Download PDFInfo
- Publication number
- CN106778927A CN106778927A CN201611270798.7A CN201611270798A CN106778927A CN 106778927 A CN106778927 A CN 106778927A CN 201611270798 A CN201611270798 A CN 201611270798A CN 106778927 A CN106778927 A CN 106778927A
- Authority
- CN
- China
- Prior art keywords
- local
- keyword
- clouds
- dictionary
- television
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V30/00—Character recognition; Recognising digital ink; Document-oriented image-based pattern recognition
- G06V30/10—Character recognition
- G06V30/26—Techniques for post-processing, e.g. correcting the recognition result
- G06V30/262—Techniques for post-processing, e.g. correcting the recognition result using context analysis, e.g. lexical, syntactic or semantic context
- G06V30/268—Lexical context
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/90—Details of database functions independent of the retrieved data types
- G06F16/95—Retrieval from the web
- G06F16/951—Indexing; Web crawling techniques
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F40/00—Handling natural language data
- G06F40/20—Natural language analysis
- G06F40/279—Recognition of textual entities
- G06F40/284—Lexical analysis, e.g. tokenisation or collocates
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F40/00—Handling natural language data
- G06F40/30—Semantic analysis
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V30/00—Character recognition; Recognising digital ink; Document-oriented image-based pattern recognition
- G06V30/10—Character recognition
- G06V30/26—Techniques for post-processing, e.g. correcting the recognition result
- G06V30/262—Techniques for post-processing, e.g. correcting the recognition result using context analysis, e.g. lexical, syntactic or semantic context
- G06V30/274—Syntactic or semantic context, e.g. balancing
Abstract
TV semantics recognition dictionary method and apparatus are updated the invention discloses one kind, methods described includes:When local television receives local voice information, the local keyword that local television is based on the identification of local voice information semantic is obtained, and the local corresponding high in the clouds of television is based on the first high in the clouds keyword of local voice information semantic identification;The local keyword and the first high in the clouds keyword are compared, the difference keyword that local keyword is different from the first high in the clouds keyword is updated to the local dictionary of local television;Local voice information is set up mapping relations and preserved with the difference keyword of local dictionary, the renewal of local dictionary is completed with correspondence.The present invention solves to fail the local voice information local semantics recognition keyword dictionary of correspondence that upgrades in time in the prior art, causes accurately obtain the corresponding every matched data of the input voice information comprehensively, so as to reduce the technical problem of Consumer's Experience.
Description
Technical field
The present invention relates to ntelligent television technolog field, more particularly to a kind of renewal TV semantics recognition dictionary method and dress
Put.
Background technology
At present, semantics recognition technology is widely studied, and TV semantics recognition technology includes being input into user in television
Voice messaging is identified as corresponding semantics recognition keyword, and each of the voice messaging is obtained with by correspondence semantics recognition keyword
Item matched data, to accelerate to user in the recognition speed of television input voice information, especially when network condition is bad, leads to
It is often that user is often locally downloading with the corresponding semantics recognition keyword of voice messaging, know with according to the locally downloading semanteme
Other keyword quick obtaining items matched data, is that every matched data is corresponding based on user input voice messaging
Semantics recognition keyword is acquired, and in the prior art, there is user input voice messaging corresponding semantics recognition beyond the clouds
When keyword changes, failing to upgrade in time corresponds to local semantics recognition keyword dictionary, thus presence can not be comprehensively accurate
The corresponding every matched data of the input voice information is obtained, so as to reduce the technical problem of Consumer's Experience.
The content of the invention
It is a primary object of the present invention to provide a kind of renewal TV semantics recognition dictionary method and device, it is intended to solve existing
Have due to not updating the corresponding local semantics recognition keyword dictionary of user input voice messaging in technology, causing can not be comprehensively accurate
The corresponding every matched data of the user input voice messaging is really obtained, so as to reduce the technical problem of Consumer's Experience.
To achieve the above object, the present invention provides a kind of renewal TV semantics recognition dictionary method, the renewal TV language
Justice identification dictionary method includes:
When local television receives local voice information, obtain local television and known based on local voice information semantic
Other local keyword, and the local corresponding high in the clouds of television is based on the first high in the clouds key of local voice information semantic identification
Word;
The local keyword and the first high in the clouds keyword are compared, local keyword will be different from the first high in the clouds keyword
Difference keyword be updated to the local dictionary of local television;
Local voice information is set up mapping relations and preserved with the difference keyword of local dictionary, completes local with correspondence
The renewal of dictionary.
Preferably, include before the comparison local keyword and the first high in the clouds keyword step:
History identification frequency of the local keyword in the preset duration of past first is obtained, and judges that the history recognizes frequency
Whether the first predeterminated frequency is more than;
When history identification frequency is more than the first predeterminated frequency, performs and compare the local keyword and the first high in the clouds
The step of keyword.
Preferably, it is described that the difference keyword that local keyword is different from first high in the clouds keyword is updated to local electricity
Local dictionary step depending on holding includes:
The difference keyword that local keyword is different from first high in the clouds keyword is updated to the local of local television
Dictionary, and the high in the clouds data message that the difference keyword is mapped is downloaded to local television.
Preferably, the renewal TV semantics recognition dictionary method also includes:
The second high in the clouds keyword that the corresponding high in the clouds of local television is based on strange land voice messaging semantics recognition is obtained, it is described different
Ground voice messaging is the voice messaging that other televisions are received beyond local television;
When history identification frequency of second high in the clouds keyword in the preset duration of past second is more than the second default frequency
During rate, the strange land voice messaging and the second high in the clouds keyword mapping are downloaded to local dictionary.
Preferably, it is described that local voice information is set up mapping relations and preserved with the difference keyword of local dictionary, with
Include after the renewal step of the local dictionary of correspondence completion:
The networking state of local television is detected, when local television is in without under interconnection net state, if local TV
End is when receiving the identical local voice information again, obtained from the updated local dictionary local keyword and
Data message that the local keyword is mapped simultaneously shows.
Additionally, to achieve the above object, the present invention also provides a kind of renewal TV semantics recognition dictionary device, the renewal
TV semantics recognition dictionary device includes:
First acquisition module, is based on for when local television receives local voice information, obtaining local television
Known based on local voice information semantic in the local keyword of local voice information semantic identification, and the local corresponding high in the clouds of television
Other first high in the clouds keyword;
Comparing module, for comparing the local keyword and the first high in the clouds keyword, by area in the first high in the clouds keyword
Difference keyword not in local keyword is updated to the local dictionary of local television;
First download module, for setting up mapping relations and protecting local voice information with the difference keyword of local dictionary
Deposit, the renewal of local dictionary is completed with correspondence.
Preferably, the renewal TV semantics recognition dictionary device also includes:
Judge module, recognizes frequency, and judge for obtaining history of the local keyword in the preset duration of past first
Whether history identification frequency is more than the first predeterminated frequency;
Performing module, for when history identification frequency is more than the first predeterminated frequency, performing and comparing the local pass
The step of keyword and the first high in the clouds keyword.
Preferably, the comparing module includes:
Download unit, for the difference keyword that local keyword is different from the first high in the clouds keyword to be updated to locally
The local dictionary of television, and the high in the clouds data message that the difference keyword is mapped is downloaded to local television.
Preferably, the renewal TV semantics recognition dictionary device also includes:
Second acquisition module, for obtaining the corresponding high in the clouds of local television is based on strange land voice messaging semantics recognition second
High in the clouds keyword, the strange land voice messaging is the voice messaging that other televisions are received beyond local television;
Second download module, frequency is recognized for the history when second high in the clouds keyword in the preset duration of past second
When rate is more than the second predeterminated frequency, the strange land voice messaging and the second high in the clouds keyword mapping are downloaded to local dictionary.
Preferably, the renewal TV semantics recognition dictionary device includes:
3rd acquisition module, the networking state for detecting local television, when local television is in netted without interconnecting
When under state, if local television receives the identical local voice information again, from the updated local dictionary
Obtain data message that local keyword and the local keyword mapped and show.
The present invention is based on local voice by when local television receives local voice information, obtaining local television
The local keyword of information semantic identification, and the local corresponding high in the clouds of television is based on the first of the identification of local voice information semantic
High in the clouds keyword;The local keyword and the first high in the clouds keyword are compared, local pass will be different from the first high in the clouds keyword
The difference keyword of keyword is updated to the local dictionary of local television;Local voice information is crucial with the difference of local dictionary
Word is set up mapping relations and is preserved, and the renewal of local dictionary is completed with correspondence.Due to being monitored according to local television monitoring system
Local television receives corresponding keyword of the local voice information in local television high in the clouds corresponding to local television, and will
Difference keyword is downloaded to the corresponding local dictionary of local television, thus the local semantics recognition of correspondence that can upgrade in time is crucial
Word dictionary, accurately to obtain the corresponding every matched data of the input voice information comprehensively, so as to lift the experience of user.
Brief description of the drawings
Fig. 1 is the schematic flow sheet that the present invention updates TV semantics recognition dictionary method first embodiment;
Fig. 2 compare the local keyword described in TV semantics recognition dictionary method second embodiment for the present invention updates
With the refinement schematic flow sheet before the keyword step of high in the clouds;
Fig. 3 is the high-level schematic functional block diagram that the present invention updates TV semantics recognition dictionary device first embodiment;
Fig. 4 is the high-level schematic functional block diagram that the present invention updates TV semantics recognition dictionary method second embodiment;
Fig. 5 is the schematic diagram of a scenario that the present invention updates TV semantics recognition dictionary method and device.
The realization of the object of the invention, functional characteristics and advantage will be described further referring to the drawings in conjunction with the embodiments.
Specific embodiment
It should be appreciated that the specific embodiments described herein are merely illustrative of the present invention, it is not intended to limit the present invention.
The present invention provides a kind of renewal TV semantics recognition dictionary method, and TV semantics recognition dictionary side is updated in the present invention
In the first embodiment of method, reference picture 1, the renewal TV semantics recognition dictionary method includes:
Step S10, when local television receives local voice information, is obtained local television and is believed based on local voice
The local keyword of semantics recognition is ceased, and the local corresponding high in the clouds of television is based on the first cloud of local voice information semantic identification
End keyword;
When user is detected in local television input voice information, the voice messaging can be the word of phonetic entry
Bar, sentence or isolated word etc., using user local television input voice information as local voice information, based on this
Ground television correspondence monitoring system real-time monitoring obtains language of the local television high in the clouds corresponding to TV to the local voice information
Justice identification keyword, illustrated to specific embodiment, such as user in local TV input phonetic entry " social news ",
Under network kilter, while crucial in the semantics recognition that " social news " are somebody's turn to do in the high in the clouds matching corresponding to TV of local television
Word, semantics recognition process includes matching keywords process, and the matching keywords for being such as somebody's turn to do " social news " can be including social news,
" news " of the items content such as entertainment newses, sports news, then matches social news from the full content of " news ", special
Not, the corresponding subordinate's news of the social news can be matched from the social news, it is bad or without internet in network condition
During state, the semantics recognition keyword of " social news " is preferentially somebody's turn to do in the matching of local television.Wherein by local television to this
Used as local keyword, the corresponding high in the clouds of local television is based on local voice information to the semantics recognition keyword of local voice information
First high in the clouds keyword of semantics recognition.
Step S20, compares the local keyword and the first high in the clouds keyword, and this will be different from the first high in the clouds keyword
The difference keyword of ground keyword is updated to the local dictionary of local television;
The local keyword and the first high in the clouds keyword are compared, the comparison process includes local keyword and the first high in the clouds
The matching of keyword quantity, and matching for specific key words content, when matching of the corresponding high in the clouds of TV to the local voice information
Keyword be the first high in the clouds keyword it is more more comprehensively when, first high in the clouds keyword is compared one by one with local keyword
It is right, the difference keyword that local keyword is different from the first high in the clouds keyword is downloaded to the corresponding local word of local television
Storehouse simultaneously preserves, and especially, the corresponding high in the clouds matched data of the difference keyword can be downloaded into local television and correspondence preservation,
Illustrated to specific embodiment, when such as user's local voice information is " news ", if the first high in the clouds keyword is social news, body
News, entertainment newses, financial and economic news, world news, animation news, military news, science and technology news etc. are educated, and in local keyword
Only social news, sports news, entertainment newses, during financial and economic news, by the world news, animation news, military news, science and technology
The difference keyword not having in the local keyword such as news is downloaded to the corresponding local dictionary of local television.
Step S30, local voice information is set up mapping relations and preserved with the difference keyword of local dictionary, with correspondence
Complete the renewal of local dictionary.
Because this is downloaded to the difference Keywords matching of local television in local voice information, thus in local electricity
Depending on the local voice information is set up into mapping with the difference keyword for being downloaded to local television in semantics recognition dictionary
Relation is simultaneously preserved, and the renewal of TV semantics recognition dictionary is completed with correspondence.
In the present embodiment, it is based on by when local television receives local voice information, obtaining local television
Known based on local voice information semantic in the local keyword of local voice information semantic identification, and the local corresponding high in the clouds of television
Other first high in the clouds keyword;The local keyword and the first high in the clouds keyword are compared, will be distinguished in the first high in the clouds keyword
The local dictionary of local television is updated in the difference keyword of local keyword;By local voice information and local dictionary
Difference keyword is set up mapping relations and is preserved, and the renewal of local dictionary is completed with correspondence.Due to being monitored according to local television
The local television of system monitoring receives correspondence of the local voice information in local television high in the clouds corresponding to local television and closes
Keyword, and just difference keyword is downloaded to the corresponding local dictionary of local television, thus the local language of correspondence that can upgrade in time
Justice identification keyword dictionary, accurately to obtain the corresponding every matched data of the input voice information comprehensively, so as to lift user
Experience.
Further, on the basis of the first embodiment that the present invention updates TV semantics recognition dictionary method, there is provided more
New TV semantics recognition dictionary method second embodiment, in a second embodiment, reference picture 2 is also wrapped before the step S20
Include:
Step S11, obtains history identification frequency of the local keyword in the preset duration of past first, and judge the history
Whether identification frequency is more than the first predeterminated frequency;
History identification frequency of the local keyword in the preset duration of past first is obtained, the local keyword was in the past the
History identification frequency in one preset duration is corresponding with the local voice information, is the local keyword default in the past first
In duration history identification number of times local voice information corresponding with the keyword or the voice messaging it is identical semanteme letter
Whether positive correlation is ceased into, after history identification frequency is obtained, history identification frequency is judged more than the first predeterminated frequency, it is described
First predeterminated frequency can be average identification frequency of each local keyword in the first preset duration in local dictionary, especially,
First predeterminated frequency can set.
Step S52, when history identification frequency is more than the first predeterminated frequency, perform compare the local keyword with
The step of first high in the clouds keyword.
Because history identification frequency is more than the first predeterminated frequency, history identification frequency is more than the first predeterminated frequency institute
Corresponding local voice information or semantic information identical with local voice information belong to user speech input frequently information,
Thus when history identification frequency is more than the first predeterminated frequency, perform the comparison local keyword and the first high in the clouds is crucial
The step of word.
In the present embodiment, frequency is recognized by obtaining history of the local keyword in the preset duration of past first, and
Judge history identification frequency whether more than the first predeterminated frequency;When history identification frequency is more than the first predeterminated frequency,
The step of comparing the local keyword and the first high in the clouds keyword is performed, due to when local keyword is when presetting in the past first
History identification frequency in length is the corresponding local voice information input search of the local keyword more than the first predeterminated frequency
When frequently, the local dictionary of local voice information is accordingly updated, rather than to all local all keyword frequent updatings of dictionary
In case causing the technical problem of Out of Memory or corresponding server paralysis.
Further, on the basis of the first embodiment that the present invention updates TV semantics recognition dictionary method, there is provided more
New TV semantics recognition dictionary method 3rd embodiment, in the third embodiment, step S20 includes:
The difference keyword that local keyword is different from first high in the clouds keyword is updated to the local of local television
Dictionary, and the high in the clouds data message that the difference keyword is mapped is downloaded to local television.
It is finally to obtain the corresponding high in the clouds data of local voice information due to semantics recognition, the high in the clouds data are to specific
Embodiment illustrates that, when local voice information is " news ", the high in the clouds data are every particular contents of news, and such as society is new
Hear, world news etc., wherein, the high in the clouds data also particular content including subordinate's news corresponding to the social news, is quick
Obtain the corresponding newest most full high in the clouds data of the input voice information, thus the high in the clouds data that the difference keyword is mapped
Information is downloaded to local television, for calling.
In the present embodiment, it is updated to by by the difference keyword that local keyword is different from the first high in the clouds keyword
The local dictionary of local television, and the high in the clouds data message that the difference keyword is mapped is downloaded to local television, by
The high in the clouds data message that the difference keyword is mapped also is downloaded in the difference keyword is not only downloaded, it is thus possible to based on local
Keyword improves Consumer's Experience in the corresponding high in the clouds data message of local side Overall Acquisition local voice information.
Further, on the basis of the first embodiment that the present invention updates TV semantics recognition dictionary method, there is provided more
New TV semantics recognition dictionary method fourth embodiment, in the fourth embodiment, the renewal TV semantics recognition dictionary method
Also include:
The second high in the clouds keyword that the corresponding high in the clouds of local television is based on strange land voice messaging semantics recognition is obtained, it is described different
Ground voice messaging is the voice messaging that other televisions are received beyond local television;
When history identification frequency of second high in the clouds keyword in the preset duration of past second is more than the second default frequency
During rate, the strange land voice messaging and the second high in the clouds keyword mapping are downloaded to local dictionary.
Obtain the strange land voice messaging of other television terminals in addition to local television terminal, when strange land voice messaging or
History identification frequency of the second high in the clouds keyword of strange land voice messaging semantic similarity in the preset duration of past second is more than the
During two predeterminated frequencies, the strange land voice messaging and the second high in the clouds keyword mapping are downloaded to local dictionary, for local dictionary
Renewal, especially, can also by the history in second preset duration identification frequency more than the second predeterminated frequency the pass
The corresponding high in the clouds data message of keyword is downloaded to local television, and is called for local TV end subscriber, concrete scene schematic diagram
As shown in Figure 5.
In the present embodiment, by obtaining the corresponding high in the clouds of local television is based on strange land voice messaging semantics recognition second
High in the clouds keyword, the strange land voice messaging is the voice messaging that other televisions are received beyond local television;When described
When history identification frequency of the second high in the clouds keyword in the preset duration of past second is more than the second predeterminated frequency, by the strange land language
Message ceases and is downloaded to local dictionary with the second high in the clouds keyword mapping.Due to local television correspondence update local television with
The high-frequency of outer other televisions recognizes the second high in the clouds keyword and strange land voice messaging, thus improves Consumer's Experience.
Further, on the basis of the first embodiment that the present invention updates TV semantics recognition dictionary method, there is provided more
The new embodiment of TV semantics recognition dictionary method the 5th, it is described by local voice information and local dictionary in the 5th embodiment
Difference keyword set up mapping relations and preserve, include after the renewal step that local dictionary is completed with correspondence:
The networking state of local television is detected, when local television is in without under interconnection net state, if local TV
End is when receiving the identical local voice information again, obtained from the updated local dictionary local keyword and
Data message that the local keyword is mapped simultaneously shows.
When local television is in without under interconnection net state, local television can not communicate with local TV high in the clouds, if
When local television receives the identical local voice information again, by obtaining local closing in updated local dictionary
Keyword and obtain the data message that the local keyword is matched in local television, and shown on tv display screen.
In the present embodiment, by detecting the networking state of local television, when local television is in netted without interconnecting
When under state, if local television receives the identical local voice information again, from the updated local dictionary
Obtain data message that local keyword and the local keyword mapped and show, it is thus possible under no net state, and
When from local television obtain need data message, improve Consumer's Experience.
Present invention offer is a kind of to update TV semantics recognition dictionary device, in a kind of renewal TV semantics recognition word of the invention
In the first embodiment of storehouse device, reference picture 3, the renewal TV semantics recognition dictionary device includes:
First acquisition module 10, for when local television receives local voice information, obtaining local TV end group
Local voice information semantic is based in the local keyword of local voice information semantic identification, and the local corresponding high in the clouds of television
First high in the clouds keyword of identification;
When user is detected in local television input voice information, the voice messaging can be the word of phonetic entry
Bar, sentence or isolated word etc., using user local television input voice information as local voice information, based on this
Ground television correspondence monitoring system real-time monitoring obtains language of the local television high in the clouds corresponding to TV to the local voice information
Justice identification keyword, illustrated to specific embodiment, such as user in local TV input phonetic entry " social news ",
Under network kilter, while crucial in the semantics recognition that " social news " are somebody's turn to do in the high in the clouds matching corresponding to TV of local television
Word, semantics recognition process includes matching keywords process, and the matching keywords for being such as somebody's turn to do " social news " can be including social news,
" news " of the items content such as entertainment newses, sports news, then matches social news from the full content of " news ", special
Not, the corresponding subordinate's news of the social news can be matched from the social news, it is bad or without internet in network condition
During state, the semantics recognition keyword of " social news " is preferentially somebody's turn to do in the matching of local television.Wherein by local television to this
Used as local keyword, the corresponding high in the clouds of local television is based on local voice information to the semantics recognition keyword of local voice information
First high in the clouds keyword of semantics recognition.
Comparing module 20, for comparing the local keyword and the first high in the clouds keyword, by the first high in the clouds keyword
The difference keyword for being different from local keyword is updated to the local dictionary of local television;
The local keyword and the first high in the clouds keyword are compared, the comparison process includes local keyword and the first high in the clouds
The matching of keyword quantity, and matching for specific key words content, when matching of the corresponding high in the clouds of TV to the local voice information
Keyword be the first high in the clouds keyword it is more more comprehensively when, first high in the clouds keyword is compared one by one with local keyword
It is right, the difference keyword that local keyword is different from the first high in the clouds keyword is downloaded to the corresponding local word of local television
Storehouse simultaneously preserves, and especially, the corresponding high in the clouds matched data of the difference keyword can be downloaded into local television and correspondence preservation,
Illustrated to specific embodiment, when such as user's local voice information is " news ", if the first high in the clouds keyword is social news, body
News, entertainment newses, financial and economic news, world news, animation news, military news, science and technology news etc. are educated, and in local keyword
Only social news, sports news, entertainment newses, during financial and economic news, by the world news, animation news, military news, science and technology
The difference keyword not having in the local keyword such as news is downloaded to the corresponding local dictionary of local television.
First download module 30, for local voice information to be set up into mapping relations simultaneously with the difference keyword of local dictionary
Preserve, the renewal of local dictionary is completed with correspondence.
Because this is downloaded to the difference Keywords matching of local television in local voice information, thus in local electricity
Depending on the local voice information is set up into mapping with the difference keyword for being downloaded to local television in semantics recognition dictionary
Relation is simultaneously preserved, and the renewal of TV semantics recognition dictionary is completed with correspondence.
In the present embodiment, by the first acquisition module 10, for when local television receives local voice information,
The local keyword that local television is based on the identification of local voice information semantic is obtained, and the local corresponding high in the clouds of television is based on
First high in the clouds keyword of local voice information semantic identification;Comparing module 20, for comparing the local keyword and first
High in the clouds keyword, the difference keyword that local keyword is different from the first high in the clouds keyword is updated to the sheet of local television
Ground dictionary;First download module 30, for local voice information to be set up into mapping relations simultaneously with the difference keyword of local dictionary
Preserve, the renewal of local dictionary is completed with correspondence.Received due to monitoring local television according to local television monitoring system
Local voice information local television high in the clouds corresponding to local television corresponding keyword, and just difference keyword download
To the corresponding local dictionary of local television, thus the local semantics recognition keyword dictionary of correspondence that can upgrade in time, with comprehensively accurate
The corresponding every matched data of the input voice information is really obtained, so as to lift the experience of user.
Further, on the basis of the first embodiment that the present invention updates TV semantics recognition dictionary device, there is provided more
New TV semantics recognition dictionary device second embodiment, in a second embodiment, reference picture 4, the renewal TV semantics recognition
Dictionary device also includes:
Judge module 11, recognizes frequency, and sentence for obtaining history of the local keyword in the preset duration of past first
Whether the history of breaking identification frequency is more than the first predeterminated frequency;
History identification frequency of the local keyword in the preset duration of past first is obtained, the local keyword was in the past the
History identification frequency in one preset duration is corresponding with the local voice information, is the local keyword default in the past first
In duration history identification number of times local voice information corresponding with the keyword or the voice messaging it is identical semanteme letter
Whether positive correlation is ceased into, after history identification frequency is obtained, history identification frequency is judged more than the first predeterminated frequency, it is described
First predeterminated frequency can be average identification frequency of each local keyword in the first preset duration in local dictionary, especially,
First predeterminated frequency can set.
Performing module 12, for when history identification frequency is more than the first predeterminated frequency, performing comparison described local
The step of keyword and the first high in the clouds keyword.
Because history identification frequency is more than the first predeterminated frequency, history identification frequency is more than the first predeterminated frequency institute
Corresponding local voice information or semantic information identical with local voice information belong to user speech input frequently information,
Thus when history identification frequency is more than the first predeterminated frequency, perform the comparison local keyword and the first high in the clouds is crucial
The step of word.
In the present embodiment, by judge module 11, for obtaining local keyword in the preset duration of past first
Whether history recognizes frequency, and judge history identification frequency more than the first predeterminated frequency;Performing module 12, for being gone through when described
When history identification frequency is more than the first predeterminated frequency, the step of comparing the local keyword and the first high in the clouds keyword is performed, by
Recognize that frequency is the local pass more than the first predeterminated frequency in the history when local keyword in the preset duration of past first
When the corresponding local voice information input search of keyword is frequent, the local dictionary of local voice information is accordingly updated, rather than right
All all keyword frequent updatings of local dictionary are in case cause the technical problem of Out of Memory or corresponding server paralysis.
Further, on the basis of the first embodiment that the present invention updates TV semantics recognition dictionary device, there is provided more
New TV semantics recognition dictionary device 3rd embodiment, in the third embodiment, walking the comparing module includes:
Download unit, for the difference keyword that local keyword is different from the first high in the clouds keyword to be updated to locally
The local dictionary of television, and the high in the clouds data message that the difference keyword is mapped is downloaded to local television.
It is finally to obtain the corresponding high in the clouds data of local voice information due to semantics recognition, the high in the clouds data are to specific
Embodiment illustrates that, when local voice information is " news ", the high in the clouds data are every particular contents of news, and such as society is new
Hear, world news etc., wherein, the high in the clouds data also particular content including subordinate's news corresponding to the social news, is quick
Obtain the corresponding newest most full high in the clouds data of the input voice information, thus the high in the clouds data that the difference keyword is mapped
Information is downloaded to local television, for calling.
In the present embodiment, by download unit, for the difference by local keyword is different from the first high in the clouds keyword
Different keyword is updated to the local dictionary of local television, and the high in the clouds data message that the difference keyword is mapped is downloaded to
Local television, the high in the clouds data message that the difference keyword is mapped also is downloaded by not only downloading the difference keyword, because
And user can be improved based on local keyword in the corresponding high in the clouds data message of local side Overall Acquisition local voice information
Experience.
Further, on the basis of the first embodiment that the present invention updates TV semantics recognition dictionary device, there is provided more
New TV semantics recognition dictionary device fourth embodiment, in the fourth embodiment, the renewal TV semantics recognition dictionary device
Also include:
Second acquisition module, for obtaining the corresponding high in the clouds of local television is based on strange land voice messaging semantics recognition second
High in the clouds keyword, the strange land voice messaging is the voice messaging that other televisions are received beyond local television;
Second download module, frequency is recognized for the history when second high in the clouds keyword in the preset duration of past second
When rate is more than the second predeterminated frequency, the strange land voice messaging and the second high in the clouds keyword mapping are downloaded to local dictionary.
Obtain the strange land voice messaging of other television terminals in addition to local television terminal, when strange land voice messaging or
History identification frequency of the second high in the clouds keyword of strange land voice messaging semantic similarity in the preset duration of past second is more than the
During two predeterminated frequencies, the strange land voice messaging and the second high in the clouds keyword mapping are downloaded to local dictionary, for local dictionary
Renewal, especially, can also by the history in second preset duration identification frequency more than the second predeterminated frequency the pass
The corresponding high in the clouds data message of keyword is downloaded to local television, and is called for local TV end subscriber, concrete scene schematic diagram
As shown in Figure 5.
In the present embodiment, by the second acquisition module, strange land voice is based on for obtaining the corresponding high in the clouds of local television
Second high in the clouds keyword of information semantic identification, the strange land voice messaging is that other televisions are received beyond local television
Voice messaging;Second download module, knows for the history when second high in the clouds keyword in the preset duration of past second
When other frequency is more than the second predeterminated frequency, the strange land voice messaging and the second high in the clouds keyword mapping are downloaded to local dictionary.
Beyond due to updating local television in local television correspondence the high-frequency of other televisions recognize the second high in the clouds keyword with
Strange land voice messaging, thus improve Consumer's Experience.
Further, on the basis of the first embodiment that the present invention updates TV semantics recognition dictionary device, there is provided more
The new embodiment of TV semantics recognition dictionary device the 5th, in the 5th embodiment, the renewal TV semantics recognition dictionary device
Including:
3rd acquisition module, the networking state for detecting local television, when local television is in netted without interconnecting
When under state, if local television receives the identical local voice information again, from the updated local dictionary
Obtain data message that local keyword and the local keyword mapped and show.
When local television is in without under interconnection net state, local television can not communicate with local TV high in the clouds, if
When local television receives the identical local voice information again, by obtaining local closing in updated local dictionary
Keyword and obtain the data message that the local keyword is matched in local television, and shown on tv display screen.
In the present embodiment, by the 3rd acquisition module, the networking state for detecting local television, when local TV
End is in during without under interconnection net state, if local television receives the identical local voice information again, from it is described
Data message that local keyword and the local keyword mapped is obtained in the local dictionary for updating and is shown, it is thus possible to
Under no net state, the data message for needing is obtained from local television in time, improve Consumer's Experience.
The preferred embodiments of the present invention are these are only, the scope of the claims of the invention is not thereby limited, it is every to utilize this hair
Equivalent structure or equivalent flow conversion that bright specification and accompanying drawing content are made, or directly or indirectly it is used in other related skills
Art field, is included within the scope of the present invention.
Claims (10)
1. it is a kind of to update TV semantics recognition dictionary method, it is characterised in that the renewal TV semantics recognition dictionary method bag
Include:
When local television receives local voice information, obtain local television and be based on the identification of local voice information semantic
Local keyword, and the local corresponding high in the clouds of television is based on the first high in the clouds keyword of local voice information semantic identification;
The local keyword and the first high in the clouds keyword are compared, the difference of local keyword will be different from the first high in the clouds keyword
Different keyword is updated to the local dictionary of local television;
Local voice information is set up mapping relations and preserved with the difference keyword of local dictionary, local dictionary is completed with correspondence
Renewal.
2. it is as claimed in claim 1 to update TV semantics recognition dictionary method, it is characterised in that the comparison local pass
Include before keyword and the first high in the clouds keyword step:
History identification frequency of the local keyword in the preset duration of past first is obtained, and whether judges history identification frequency
More than the first predeterminated frequency;
When history identification frequency is more than the first predeterminated frequency, performs and compare the local keyword and the first high in the clouds key
The step of word.
3. it is as claimed in claim 1 or 2 to update TV semantics recognition dictionary method, it is characterised in that described by the first high in the clouds
The difference keyword that local keyword is different from keyword is updated to the local dictionary step of local television and includes:
The difference keyword that local keyword is different from first high in the clouds keyword is updated to the local dictionary of local television,
And the high in the clouds data message that the difference keyword is mapped is downloaded to local television.
4. it is as claimed in claim 1 or 2 to update TV semantics recognition dictionary method, it is characterised in that the renewal TV language
Justice identification dictionary method also includes:
Obtain the second high in the clouds keyword that the corresponding high in the clouds of local television is based on strange land voice messaging semantics recognition, the strange land language
Message breath is the voice messaging that other televisions are received beyond local television;
When history identification frequency of second high in the clouds keyword in the preset duration of past second is more than the second predeterminated frequency,
The strange land voice messaging and the second high in the clouds keyword mapping are downloaded to local dictionary.
5. it is as claimed in claim 1 to update TV semantics recognition dictionary method, it is characterised in that described by local voice information
Difference keyword with local dictionary is set up mapping relations and is preserved, and is wrapped after the renewal step that local dictionary is completed with correspondence
Include:
The networking state of local television is detected, when local television is in without under interconnection net state, if local television is again
It is secondary when receiving the identical local voice information, obtained from the updated local dictionary local keyword and this
The ground data message that is mapped of keyword simultaneously shows.
6. it is a kind of to update TV semantics recognition dictionary device, it is characterised in that the renewal TV semantics recognition dictionary device bag
Include:
First acquisition module, is based on locally for when local television receives local voice information, obtaining local television
The local keyword of voice messaging semantics recognition, and the local corresponding high in the clouds of television is based on the identification of local voice information semantic
First high in the clouds keyword;
Comparing module, for comparing the local keyword and the first high in the clouds keyword, will be different from the first high in the clouds keyword
The difference keyword of local keyword is updated to the local dictionary of local television;
First download module, for setting up mapping relations and preserving local voice information with the difference keyword of local dictionary,
The renewal of local dictionary is completed with correspondence.
7. it is as claimed in claim 6 to update TV semantics recognition dictionary device, it is characterised in that the renewal TV is semantic to be known
Other dictionary device also includes:
Judge module, recognizes frequency, and judge that this is gone through for obtaining history of the local keyword in the preset duration of past first
Whether history identification frequency is more than the first predeterminated frequency;
Performing module, for when history identification frequency is more than the first predeterminated frequency, performing and comparing the local keyword
The step of with the first high in the clouds keyword.
8. renewal TV semantics recognition dictionary device as claimed in claims 6 or 7, it is characterised in that the comparing module bag
Include:
Download unit, for the difference keyword that local keyword is different from the first high in the clouds keyword to be updated into local TV
The local dictionary at end, and the high in the clouds data message that the difference keyword is mapped is downloaded to local television.
9. renewal TV semantics recognition dictionary device as claimed in claims 6 or 7, it is characterised in that the renewal TV language
Justice identification dictionary device also includes:
Second acquisition module, the second high in the clouds of strange land voice messaging semantics recognition is based on for obtaining the corresponding high in the clouds of local television
Keyword, the strange land voice messaging is the voice messaging that other televisions are received beyond local television;
Second download module is big for the history identification frequency when second high in the clouds keyword in the preset duration of past second
When the second predeterminated frequency, the strange land voice messaging and the second high in the clouds keyword mapping are downloaded to local dictionary.
10. it is as claimed in claim 8 to update TV semantics recognition dictionary device, it is characterised in that the renewal TV is semantic
Identification dictionary device includes:
3rd acquisition module, the networking state for detecting local television, when local television is in without under interconnection net state
When, if local television receives the identical local voice information again, obtained from the updated local dictionary
Data message that local keyword and the local keyword are mapped simultaneously shows.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201611270798.7A CN106778927A (en) | 2016-12-30 | 2016-12-30 | Update TV semantics recognition dictionary method and device |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201611270798.7A CN106778927A (en) | 2016-12-30 | 2016-12-30 | Update TV semantics recognition dictionary method and device |
Publications (1)
Publication Number | Publication Date |
---|---|
CN106778927A true CN106778927A (en) | 2017-05-31 |
Family
ID=58952802
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201611270798.7A Pending CN106778927A (en) | 2016-12-30 | 2016-12-30 | Update TV semantics recognition dictionary method and device |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN106778927A (en) |
Cited By (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN107766482A (en) * | 2017-10-13 | 2018-03-06 | 北京猎户星空科技有限公司 | Information pushes and sending method, device, electronic equipment, storage medium |
CN110097878A (en) * | 2018-01-30 | 2019-08-06 | 阿拉的(深圳)人工智能有限公司 | Polygonal color phonetic prompt method, cloud device, prompt system and storage medium |
CN111312253A (en) * | 2018-12-11 | 2020-06-19 | 青岛海尔洗衣机有限公司 | Voice control method, cloud server and terminal equipment |
Citations (8)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101067858A (en) * | 2006-09-28 | 2007-11-07 | 腾讯科技(深圳)有限公司 | Network advertisment realizing method and device |
US20100074524A1 (en) * | 2008-09-23 | 2010-03-25 | Stollman Jeff | Methods and apparatus related to document processing based on a document type |
CN103247291A (en) * | 2013-05-07 | 2013-08-14 | 华为终端有限公司 | Updating method, device, and system of voice recognition device |
CN103607609A (en) * | 2013-11-27 | 2014-02-26 | Tcl集团股份有限公司 | Voice switching method and device for TV set channels |
CN104183237A (en) * | 2014-09-04 | 2014-12-03 | 百度在线网络技术(北京)有限公司 | Speech processing method and device for portable terminal |
CN105261366A (en) * | 2015-08-31 | 2016-01-20 | 努比亚技术有限公司 | Voice identification method, voice engine and terminal |
CN105791931A (en) * | 2016-02-26 | 2016-07-20 | 深圳Tcl数字技术有限公司 | Smart television and voice control method of the smart television |
CN106210909A (en) * | 2016-08-15 | 2016-12-07 | 深圳Tcl数字技术有限公司 | TV the display processing method of content, Apparatus and system |
-
2016
- 2016-12-30 CN CN201611270798.7A patent/CN106778927A/en active Pending
Patent Citations (8)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101067858A (en) * | 2006-09-28 | 2007-11-07 | 腾讯科技(深圳)有限公司 | Network advertisment realizing method and device |
US20100074524A1 (en) * | 2008-09-23 | 2010-03-25 | Stollman Jeff | Methods and apparatus related to document processing based on a document type |
CN103247291A (en) * | 2013-05-07 | 2013-08-14 | 华为终端有限公司 | Updating method, device, and system of voice recognition device |
CN103607609A (en) * | 2013-11-27 | 2014-02-26 | Tcl集团股份有限公司 | Voice switching method and device for TV set channels |
CN104183237A (en) * | 2014-09-04 | 2014-12-03 | 百度在线网络技术(北京)有限公司 | Speech processing method and device for portable terminal |
CN105261366A (en) * | 2015-08-31 | 2016-01-20 | 努比亚技术有限公司 | Voice identification method, voice engine and terminal |
CN105791931A (en) * | 2016-02-26 | 2016-07-20 | 深圳Tcl数字技术有限公司 | Smart television and voice control method of the smart television |
CN106210909A (en) * | 2016-08-15 | 2016-12-07 | 深圳Tcl数字技术有限公司 | TV the display processing method of content, Apparatus and system |
Non-Patent Citations (3)
Title |
---|
KARTIK AUDHKHASI等: "SEMANTIC WORD EMBEDDING NEURAL NETWORK LANGUAGE MODELS FOR AUTOMATIC SPEECH RECOGNITION", 《2016 IEEE INTERNATIONAL CONFERENCE ON ACOUSTICS, SPEECH AND SIGNAL PROCESSING (ICASSP)》 * |
于金山等: "基于云的语义库设计及机器人语义地图构建", 《机器人》 * |
林茜等: "基于Microsoft Speech SDK 的语音关键词检出系统的设计和实现", 《心智与计算》 * |
Cited By (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN107766482A (en) * | 2017-10-13 | 2018-03-06 | 北京猎户星空科技有限公司 | Information pushes and sending method, device, electronic equipment, storage medium |
CN110097878A (en) * | 2018-01-30 | 2019-08-06 | 阿拉的(深圳)人工智能有限公司 | Polygonal color phonetic prompt method, cloud device, prompt system and storage medium |
CN111312253A (en) * | 2018-12-11 | 2020-06-19 | 青岛海尔洗衣机有限公司 | Voice control method, cloud server and terminal equipment |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
WO2017206661A1 (en) | Voice recognition method and system | |
CN104715752B (en) | Audio recognition method, apparatus and system | |
CN105630876B (en) | Information acquisition method and device across application | |
CN104183237B (en) | Method of speech processing and device for portable terminal | |
CN111831911B (en) | Query information processing method and device, storage medium and electronic device | |
CN108984650B (en) | Computer-readable recording medium and computer device | |
CN110262273A (en) | A kind of home equipment control method, device, storage medium and smart home system | |
CN106778927A (en) | Update TV semantics recognition dictionary method and device | |
CN108780048B (en) | Method for determining detection equipment, detection device and readable storage medium | |
CN107229731B (en) | Method and apparatus for classifying data | |
CN105700897B (en) | A kind of method, apparatus and terminal device starting application program | |
CN104462064A (en) | Method and system for prompting content input in information communication of mobile terminals | |
CN107609047A (en) | Using recommendation method, apparatus, mobile device and storage medium | |
CN105632487A (en) | Voice recognition method and device | |
CN106409285A (en) | Method and apparatus for intelligent terminal device to identify language type according to voice data | |
CN109584887A (en) | A kind of method and apparatus that voiceprint extracts model generation, voiceprint extraction | |
CN108989875A (en) | A kind of method and device generating barrage file | |
CN110929141A (en) | Group mining method, device, equipment and storage medium | |
CN106096520A (en) | A kind of information-pushing method and device | |
CN108170708A (en) | A kind of vehicle entity recognition method, electronic equipment, storage medium, system | |
CN108108391A (en) | For the processing method and device of the information of data visualization | |
CN104239462A (en) | Method and device for displaying search results | |
CN107491464A (en) | Device, method, terminal, server and the system of electronics service manual are provided | |
CN112399013B (en) | Abnormal telephone traffic identification method and device | |
CN105119948B (en) | Method and apparatus for updating desktop browsers shortcut icon in client |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
RJ01 | Rejection of invention patent application after publication |
Application publication date: 20170531 |
|
RJ01 | Rejection of invention patent application after publication |