CN109359305A - A kind of method and apparatus of multilingual intertranslation in unison - Google Patents
A kind of method and apparatus of multilingual intertranslation in unison Download PDFInfo
- Publication number
- CN109359305A CN109359305A CN201811029882.9A CN201811029882A CN109359305A CN 109359305 A CN109359305 A CN 109359305A CN 201811029882 A CN201811029882 A CN 201811029882A CN 109359305 A CN109359305 A CN 109359305A
- Authority
- CN
- China
- Prior art keywords
- voice
- translation
- institute
- processor
- voice document
- 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
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F40/00—Handling natural language data
- G06F40/40—Processing or translation of natural language
- G06F40/55—Rule-based translation
- G06F40/56—Natural language generation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F40/00—Handling natural language data
- G06F40/40—Processing or translation of natural language
- G06F40/58—Use of machine translation, e.g. for multi-lingual retrieval, for server-side translation for client devices or for real-time translation
-
- G—PHYSICS
- G10—MUSICAL INSTRUMENTS; ACOUSTICS
- G10L—SPEECH ANALYSIS OR SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING; SPEECH OR AUDIO CODING OR DECODING
- G10L15/00—Speech recognition
- G10L15/26—Speech to text systems
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Audiology, Speech & Language Pathology (AREA)
- Health & Medical Sciences (AREA)
- Computational Linguistics (AREA)
- Physics & Mathematics (AREA)
- General Health & Medical Sciences (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Artificial Intelligence (AREA)
- Human Computer Interaction (AREA)
- Acoustics & Sound (AREA)
- Multimedia (AREA)
- Machine Translation (AREA)
Abstract
The present invention provides a kind of methods and apparatus of multilingual intertranslation in unison, this method comprises: carrying out the identification of languages context to the voice document that sound pick up equipment generates;It will identify that the voice document after languages context carries out textual processing, and treated that voice document is cut into multiple voice document units by textual, to identify calculating and translation respectively;Multiple institute's voice file units are distributed to multiple processors, Distributed identification is executed to the institute's voice file unit assigned in a parallel fashion by multiple processors and calculates and translates;And by identified and translation institute's voice file unit weight group be the text file after translating, and the text file after the translation is broadcasted in the form of voice document.The invention also discloses the equipment of multilingual intertranslation in unison.The present invention takes full advantage of the advantage that multiple processor parallel distributeds calculate, and takes full advantage of the computing capability of each processor, identification and translation accuracy rate are high, and speed is fast, pronunciation standard.
Description
Technical field
The present invention relates to multilingual speech recognition and text voice process field, especially a kind of the more language is mutual in unison
Translate method and apparatus.
Background technique
The current multilingual technology of intertranslation in unison has developed into the multilingual intertranslation side in unison using multiprocessor cluster
Formula, wherein each processor possibly also with multiple central processing unit (CPU) carried out by the way of multithreading it is multilingual in unison
Intertranslation.Wherein, each processor carries out multilingual intertranslation in unison as unit of a voice document.It is corresponding with the processor
Although multiple CPU carry out data processing in the way of multithreading, for a voice document, data processing
It is serially to be carried out by multiple CPU.Therefore, recognition efficiency is limited to the multiple threads speed of multiple CPU of processor.
The existing multilingual equipment of intertranslation in unison recognition speed is slow, low efficiency, development difficulty are big, and needs chip manufacturer
Support energetically could complete.
Summary of the invention
In view of the above problems, it proposes the present invention and at least is partially solved the multilingual same of the above problem in order to provide a kind of
The method and apparatus of sound intertranslation.
According to one aspect of the invention, it provides a kind of multilingual inter-translation methods in unison, wherein includes:
The identification of languages context is carried out to the voice document that sound pick up equipment generates;
It will identify that the voice document after languages context carries out textual processing, and treated that voice document is cut by textual
It is divided into multiple voice document units, to identify calculating and translation respectively;
Multiple institute's voice file units are distributed to multiple processors, it is right in a parallel fashion by the multiple processor
The institute's voice file unit assigned to executes Distributed identification and calculates and translate;And
By the text file that identified and translation institute's voice file unit weight group is after translation, and will be after the translation
Text file broadcasted in the form of voice document.
Above-mentioned method, wherein the identification for carrying out languages context to the voice document that sound pick up equipment generates further comprises:
The voice selecting of the voice document generated according to the sound pick up equipment talks with languages;
Discourse context is selected according to the dialogue languages and first institute's voice file;And
Automatically the dialogue languages and discourse context of the voice document that the sound pick up equipment generates are confirmed repeatedly.
Above-mentioned method, wherein by identified institute's voice file unit weight group be translation after text file after,
Further include:
The accuracy rate and/or speech quality evaluation value of text file after checking the translation, the translation to generation
Text file afterwards carries out grammer and speech quality correction.
Above-mentioned method, wherein the speech quality evaluation value includes Y-PSNR and structural similarity index.
Above-mentioned method, wherein before multiple institute's voice file units are distributed to multiple processors, detect described more
The computing capability of a processor, and be distributed to multiple institute's voice file units according to the computing capability of the multiple processor
The multiple processor further comprises:
Detect the number of the multiple processor;
Detect the number of the core in the multiple processor;And/or
Detect the dominant frequency of the core in the multiple processor.
Above-mentioned method, wherein by the multiple processor in a parallel fashion to the institute's voice file list assigned to
After position executes Distributed identification calculating and translation, the method further includes:
Check the result of described identified and translation institute's voice file unit;And
Institute's voice file unit that identification and translation result do not meet predetermined condition is distributed to the multiple place again
Device is managed, to be re-recognized and be translated.
Above-mentioned method, wherein do not meet institute's voice file unit of predetermined condition again with translation result that will identify
It is secondary to be distributed to before the multiple processor, the method further includes:
According to recognition result, the institute's predicate for not meeting predetermined condition in the multiple processor with the recognition result is adjusted
The identification parameter of the corresponding processor of sound file unit;
Institute's voice file unit that the recognition result does not meet predetermined condition is distributed to the multiple processing again
The corresponding processor in device.
Above-mentioned method, wherein after the result for checking described identified and translation institute's voice file unit, institute
The method of stating further comprises:
Discharge the corresponding place of the institute's voice file unit to conform to a predetermined condition in the multiple processor with recognition result
Manage device.
Above-mentioned method, wherein the identification for carrying out languages context to the voice document that sound pick up equipment generates further comprises:
Individual information variation detection is carried out to the voice content of institute's voice file, to identify cutting for the discourse context
It changes.
According to a further aspect of the invention, additionally provide a kind of multilingual intertranslation equipment in unison, wherein including sound pick up equipment,
Cutting device, distribution translating equipment and recombination public address system,
The sound pick up equipment is used for the voice document for being identified and being generated to voice;
The cutting device is used to carry out institute's voice file textual processing, and by textual treated voice text
Part is cut into multiple voice document units;
The distribution translating equipment is used to multiple institute's voice file units being distributed to multiple processors, by the multiple
Processor executes Distributed identification to the institute's voice file unit assigned in a parallel fashion and calculates and translate;And
The recombination public address system is used to be the text after translation by identified and translation institute's voice file unit weight group
This document, and the text file after the translation is broadcasted in the form of voice document.
Above-mentioned multilingual inter-translation method in unison and equipment take full advantage of the advantage of multiple processor distributed computings, effectively
The computing capability of each processor is utilized, recognition speed is fast, and accuracy rate is high.It is truly realized multilingual intertranslation in unison as a result,.
The above description is only an overview of the technical scheme of the present invention, in order to better understand the technical means of the present invention,
And it can be implemented in accordance with the contents of the specification, and in order to allow above and other objects of the present invention, feature and advantage can
It is clearer and more comprehensible, the followings are specific embodiments of the present invention.
Detailed description of the invention
The embodiment of the present invention is described in more detail in conjunction with the accompanying drawings, the above and other purposes of the present invention,
Feature and advantage will be apparent.Attached drawing is used to provide to further understand the embodiment of the present invention, and constitutes explanation
A part of book, is used to explain the present invention together with the embodiment of the present invention, is not construed as limiting the invention.In the accompanying drawings,
Identical reference label typically represents same parts or step.
Fig. 1 shows the flow chart of multilingual inter-translation method in unison according to an embodiment of the invention;
Fig. 2 shows the flow charts of multilingual inter-translation method in unison in accordance with another embodiment of the present invention;
Fig. 3 shows the schematic block diagram of the multilingual equipment of intertranslation in unison according to an embodiment of the invention.
Specific embodiment
In order to enable the object, technical solutions and advantages of the present invention become apparent, root is described in detail below with reference to accompanying drawings
According to example embodiments of the present invention.Obviously, described embodiment is only a part of the embodiments of the present invention, rather than this hair
Bright whole embodiments, it should be appreciated that the present invention is not limited by example embodiment described herein.Based on described in the present invention
Embodiment, those skilled in the art's obtained all other embodiment in the case where not making the creative labor should all be fallen
Enter within protection scope of the present invention.
Referring to Fig. 1 and Fig. 2, Fig. 1 shows the process of multilingual inter-translation method in unison according to an embodiment of the invention
Figure, Fig. 2 shows the flow charts of multilingual inter-translation method in unison in accordance with another embodiment of the present invention.The present invention is to voice text
Part carries out identification and translation processing, to obtain speech recognition and translation.As shown in Figure 1, the multilingual inter-translation method in unison, including
Following steps:
Step S100, the identification that languages context is carried out to the voice document that sound pick up equipment 10 generates, using speech recognition and
Translation technology come to voice document carry out identification and translation handle.The huge voice document of data volume is cut into voice one by one
Identification and translation, with identifying processing respectively.In the next steps, it is carried out as unit of the identification and translation of a voice document more
Language intertranslation in unison;Step S100 can further comprise:
Step S110, the voice selecting of the voice document generated according to the sound pick up equipment 10 talks with languages;
Step S120, discourse context is selected according to the dialogue languages and first institute's voice file;And
Step S130, confirm the dialogue languages and dialogue of the voice document that the sound pick up equipment 10 generates repeatedly automatically
Context.
Wherein, the identification for carrying out languages context to the voice document that sound pick up equipment 10 generates can also further comprise:
Step S140, individual information variation detection is carried out to the voice content of institute's voice file, to identify the dialogue
The switching of context.
Specifically, voice document can be analyzed first, to determine its type.For example, being searched according to token sound
Data in voice document, and multilingual type is determined according to the data found.It is then possible to according to multilingual class
Type reads the data in voice document, and institute's voice file is carried out speech recognition and translation.It is optional according to the more of voice document
Language form carries out identification and translation processing.In voice coding sequence, equipment carries out text to the voice document heard first
Identification, is then translated text file, is finally broadcasted the text file after translation in a manner of voice document.Optional
Switch multilingual type according to session context, identification then is carried out to voice document and translation is handled.
Different from text data, voice data not only includes that temporal information further includes individual information.It can be to voice content
Carry out individual information variation detection.In case of variation, then can think that voice context is switched accordingly.It can root
Switch according to voice context, voice document is identified and is translated.That is, different speech recognitions and translation include difference
Individual information.It can identify that context switches by carrying out feature detection to voice document, such as acquaintance, stranger, men and women
Etc. features.
Processing is identified and translated to voice document according to the switching of the context of voice document to switch particularly suitable for context
Frequent voice document.The switching of context means the beginning for being suitable for new dialogue sequence, ensure that subsequent recognition operation
It goes on smoothly.
Step S200, it will identify that the voice document after languages context carries out textual processing, and treated by textual
Voice document is cut into multiple voice document units, to identify calculating and translation respectively;
Step S300, multiple institute's voice file units are distributed to multiple processors, by the multiple processor with simultaneously
Capable mode executes Distributed identification to the institute's voice file unit assigned to and calculates and translate;
In step S300, multiple institute's voice file units are distributed to multiple processors, by multiple processors pair
The voice document unit assigned to is identified and is translated.Here, it can successively be divided according to the sequence of speech recognition and translation
Issue each processor.Processor may include various computing units, and processor can be multicore.There may be very for processor
It is multiple.The processor for executing the distributed computing of multilingual intertranslation in unison is more, then the calculating energy of the multilingual equipment of intertranslation in unison
Power is stronger, and the recognition speed of voice document is faster.
Multiple processors execute Distributed identification and calculate.What each processor executed is the identification to voice document unit
Operation, for this angle, the calculating that each processor executes is identical.Only object operated by each processor is not
Together, each processor carries out identification operation for different voice document units.What each processor need to only assign to oneself
Voice document unit carries out identifying processing, other speech recognitions and translation without regard to voice document.For what is assigned to
Voice document unit, multiple processors can carry out identifying processing in a parallel fashion.In other words, multiple processors can be simultaneously
The voice document unit respectively assigned to is identified, does not interfere with each other and influences.Each processing is efficiently utilized as a result,
The computing capability of device.
It is appreciated that identifying processing can use any suitable recognizer, which is not limited by the present invention.
Step S400, by the text file that identified and translation institute's voice file unit weight group is after translation, and will
Text file after the translation is broadcasted in the form of voice document, wherein by identified institute's voice file unit weight group
After the text file after translation, further includes:
Step S410, the accuracy rate and/or speech quality evaluation value for checking the text file after the translation, to generation
Text file after the translation carries out grammer and speech quality correction, wherein the speech quality evaluation value includes peak value letter
Make an uproar than with structural similarity index;
Step S420, identification is re-started to the voice document after correction and translation is handled, generate the multilingual correction
Text file afterwards.
Text file in above-mentioned steps, after translation can be reassembled as according to the individual information of voice data file.It can be with
Understand, the serial number of multiple voice document units shows the individual information of multiple voice document units.Optionally, according to multiple languages
The serial number of sound file unit can recombinate identified multiple voice document units forms identified voice text together
Part.
Above-mentioned multilingual inter-translation method in unison takes full advantage of multiple processors and executes distributed computing, is truly realized more
The intertranslation in unison of language.Because processor can efficiently use the multiple voice document unit parallelism recognitions assigned to
The computing capability of each processor, recognition speed is fast, and accuracy rate is high.
In another embodiment, in step S300, multiple institute's voice file units are distributed to multiple
Before processor, further includes:
Step S310, the computing capability of the multiple processor is detected, and according to the computing capability of the multiple processor
Multiple institute's voice file units are distributed to the multiple processor, hardware device self-test can be carried out when hardware powers on,
To detect the computing capability of multiple processors.Thus, it is possible to assess the availability of computing resource.
Optionally, it detects the computing capability of multiple processors and further includes:
Step S311, the number of the multiple processor is detected;
Step S312, the number of the core in the multiple processor is detected;And/or
Step S313, the dominant frequency of the core in the multiple processor is detected.
The number of processor is more, and the core number in processor is more, and the dominant frequency of core is higher, then the calculating energy of processor
Power is stronger.The number of processor significantly affects computing capability.Optionally, voice document unit is distributed according to the number of processor
To multiple processors.If the number of processor is less, and the number of voice document unit is more, then can be each processing
The multiple voice document units of device point.
In some contexts of the invention, multilingual intertranslation in unison can be carried out simultaneously to many voice documents.These
The languages of voice document can be different, such as English, German, French etc..Optionally, by the voice of primary session languages point
The higher processor of computing capability is issued, such as wherein core is more, the higher processor of dominant frequency.And by the voice of other languages
It is distributed to the lower processor of computing capability, such as wherein core is less, the lower processor of dominant frequency.
In other contexts of the invention, a voice document can be translated as to the voice document of different language.Example
An English Phonetics file is such as translated as Chinese, Spanish two kinds of voice documents respectively.In general, the former needs more calculating
Amount, the latter need less calculation amount.The former can be distributed to the higher processor of computing capability, and the latter is distributed to calculating
The lower processor of ability.
The computing capability of measurement processor and the processing task for distributing voice document unit according to the computing capability, can be with
The computing capability for efficiently using each processor allows each processor to make the best use of everything.Also, due to processor can energy person it is more
Labor avoids so respective processing task can be completed at the same time as far as possible because short -board effect causes Respective processors to wait
Situation improves the recognition speed of the multilingual equipment of intertranslation in unison.
Wherein, in step S300, by the multiple processor in a parallel fashion to the institute's voice file list assigned to
After position executes Distributed identification calculating and translation, the method further includes:
Step S320, check it is described identified and translation institute's voice file unit as a result, result may include very
Various aspects, such as the quality and accuracy rate of identified voice document unit correct assessed value.Translation and recognition result can be made
For the foundation of identification operation quality evaluation.If the recognition result for checking speech recognition and translation does not meet predetermined condition, that
The identified voice document that current identification obtains can be abandoned.
Before and after the identification of voice document, it is expected that the speech quality of speech recognition and translation remains unchanged.Assuming that after identification, language
The speech quality and accuracy rate of sound file unit are changed, then it is likely used only to being that mistake has occurred in identification.Check voice
The speech quality and accuracy rate of identification and translation can guarantee the correctness of identification operation to a certain extent.
The speech quality evaluation value of speech recognition and translation is to refer to the parameter of characterization voice quality, such as peak value noise
Than (PSNR) and structural similarity index (SSIM) etc..Speech quality evaluation value is obtained after the completion of effectively characterizing identification operation
The quality of the voice document obtained.The purpose of identification is to correct the syntax gauge of voice document, checks that speech quality evaluation value is effective
It ensure that identification operates the adverse effect not excessive to voice quality, improve user experience.
Step S350, according to recognition result, identification and translation result are not met to institute's voice file list of predetermined condition
Position is distributed to the multiple processor again, to be re-recognized and be translated.
Wherein, it will identify that the institute's voice file unit for not meeting predetermined condition with translation result is divided again in step S350
It is dealt into before the multiple processor, the method further includes:
Step S340, it according to recognition result, adjusts in the multiple processor and does not meet predetermined item with the recognition result
The identification parameter of the corresponding processor of institute's voice file unit of part, to generate and know when executing identification operation again
The ideal speech recognition of other result and translation;
And in step S350, the recognition result original institute's voice file unit of predetermined condition is not met into (not
It is identified) be distributed to corresponding (previously having been distributed) processor in the multiple processor again, with into
Row re-recognizes.Because the identification parameter of processor is adjusted according to last time recognition result, energy when again identifying that
Enough obtain ideal voice document translation.By carrying out detection to recognition result and adjusting processor according to recognition result with again
It identifies speech recognition and translation, ensure that the identification quality of voice document.
Wherein, described after the result that step S320 checks described identified and translation institute's voice file unit
Method further comprises:
Step S330, the institute's voice file unit to conform to a predetermined condition in the multiple processor with recognition result is discharged
Corresponding processor.Release processor in time, avoids meaningless occupancy computing resource.When to be identified there are new voice document,
I.e. using the processor discharged.To make the resource of whole system more effectively be utilized.
It is appreciated that above-mentioned steps S340 it is not necessary to.If step S340 is not present, in step S350,
The speech recognition and translation that recognition result can not met to predetermined condition are distributed to any in multiple processors respectively again
Processor, such as non-corresponding processor.The voice of an example according to the present invention, identification front and back voice document has occurred
Change, then may determine that problem has occurred in identification process, can directly be redistributed to original voice document unit
Random processor in multiple processors re-recognizes.In this way can to avoid because processor reason caused by recognition result not
Ideal can also guarantee the identification quality of voice document to a certain extent.
It will appreciated by the skilled person that above-mentioned multilingual inter-translation method in unison is merely illustrative, and not enough constitute
Limitation of the present invention.Although for example, step S311 is held prior to step S312 in the multilingual step of inter-translation method in unison S310
Row, but its execution sequence is restricted thereto.Step S312 can be executed prior to step S311.Step S311 can also and step
S312 is executed together, to save time cost.
The schematic frame of the multilingual equipment of intertranslation in unison according to an embodiment of the invention is shown referring to Fig. 3, Fig. 3
Figure.Multilingual intertranslation equipment in unison of the invention, including sound pick up equipment 10, cutting device 20, distribution translating equipment 30 and recombination
Public address system 40, wherein the sound pick up equipment 10 is used for the voice document for being identified and being generated to voice, and sound pick up equipment 10 can
Identifying processing is carried out to voice document according to voice document or the switching of voice context;The cutting device 20 is used for institute's predicate
Sound file carries out textual processing, and treated that voice document is cut into multiple voice document units by textual;Described point
Hair translating equipment 30 is used to multiple institute's voice file units being distributed to multiple processors, by the multiple processor with parallel
Mode Distributed identification executed to institute's voice file unit for assigning to calculate and translation;And the recombination public address system 40
For be the text file after translating by identified and translation institute's voice file unit weight group, and by the text after the translation
This document is broadcasted in the form of voice document.The multilingual equipment of intertranslation in unison of the invention can be integrated in such as glasses, cap can
Wearable device or special equipment etc..
Multiple processors can be digital signal processor.Multiple processors can be integrated in respectively on multiple boards.Often
At least one processor is integrated on a board.Board includes network interface, for communicating with the multilingual equipment of intertranslation in unison.
The multilingual equipment of intertranslation in unison can also include processor detection device (not shown).Processor detection device is used
In the computing capability for detecting the multiple processor.Processor detection device be configurable to detect the number of multiple processors,
The dominant frequency of the number of core in multiple processors and/or the core in multiple processors.Distribute translating equipment 30 according to multiple
Multiple voice document units are distributed to multiple processors by the respective computing capability of processor.
The multilingual equipment of intertranslation in unison can also include result check device and retransmission apparatus (not shown).As a result it checks
Device is used to check the recognition result of the voice document through processor identification.Recognition result includes at least one in the following terms
: number of speech frames, size and/or the speech quality evaluation value of identified voice document.Retransmission apparatus is used for recognition result
The original voice file for not meeting predetermined condition is distributed to multiple processors respectively again, to be re-recognized.
The multilingual equipment of intertranslation in unison can also include parameter adjustment controls (not shown).Parameter adjustment controls are used for root
According to result check device recognition result obtained, adjusts in multiple processors and recognition result and do not meet predetermined condition
Speech recognition and the identification parameter for translating corresponding processor.For including the case where that parameter adjustment controls, retransmission apparatus are used for
Recognition result is not met into the speech recognition of predetermined condition and translation is distributed in the multiple processors and voice respectively again
Identify and translate corresponding processor.
The multilingual equipment of intertranslation in unison can also include release device (not shown).Release device is for discharging multiple places
Manage in device, corresponding with speech recognition and translation that recognition result conforms to a predetermined condition processor.
The present invention takes full advantage of the advantage of multiple processor distributed computings, is effectively utilized the calculating of each processor
Ability, recognition speed is fast, and accuracy rate is high.It is truly realized multilingual intertranslation in unison as a result,.
Those of ordinary skill in the art, can by reading the detailed description above for the multilingual inter-translation method in unison
Understand structure, realization and the advantage of the above-mentioned multilingual equipment of intertranslation in unison, therefore which is not described herein again.
Method and apparatus is not inherently related to any particular computer, virtual system, or other device provided herein.
Various common apparatus can also be used together with teachings based herein.As described above, it constructs required by this kind of equipment
Structure be obvious.In addition, the present invention is not also multilingual for any certain programmed.It should be understood that can use each
Kind programs multilingual realization summary of the invention described herein, and is to drape over one's shoulders to specific multilingual done description above
Reveal preferred forms of the invention.
In the instructions provided here, numerous specific details are set forth.It is to be appreciated, however, that implementation of the invention
Example can be practiced without these specific details.In some instances, well known method, structure is not been shown in detail
And technology, so as not to obscure the understanding of this specification.
Similarly, it should be understood that in order to simplify the disclosure and help to understand one or more of the various inventive aspects,
Above in the description of exemplary embodiment of the present invention, each feature of the invention is grouped together into single implementation sometimes
In example, figure or descriptions thereof.However, the disclosed method should not be interpreted as reflecting the following intention: i.e. required to protect
Shield the present invention claims features more more than feature expressly recited in each claim.More precisely, as following
Claims reflect as, inventive aspect is all features less than single embodiment disclosed above.Therefore,
Thus the claims for following specific embodiment are expressly incorporated in the specific embodiment, wherein each claim itself
All as a separate embodiment of the present invention.
It will be understood by those skilled in the art that in addition to such feature and/or at least some of process or unit are
It, can be using any combination to being disclosed in this specification (including the accompanying claims, abstract and drawings) except mutually exclusive
All features and so disclosed any method or all process or units of device be combined.Unless in addition clear
Statement, each feature disclosed in this specification (including adjoint claim, abstract and attached drawing) can by provide it is identical, etc.
With or the alternative features of similar purpose replace.
In addition, it will be appreciated by those of skill in the art that although some embodiments described herein include other embodiments
In included certain features rather than other feature, but the combination of the feature of different embodiments mean it is of the invention
Within the scope of and form different embodiments.For example, in the following claims, embodiment claimed is appointed
Meaning one of can in any combination mode come using.
Various component embodiments of the invention can be implemented in hardware, or to run on one or more processors
Software module realize, or be implemented in a combination thereof.It will be understood by those of skill in the art that can be used in practice
Microprocessor or digital signal processor (DSP) are realized in the multilingual equipment of intertranslation in unison according to an embodiment of the present invention
The some or all functions of some modules.The present invention is also implemented as a part for executing method as described herein
Or whole program of device (for example, computer program and computer program product).It is such to realize that program of the invention
May be stored on the computer-readable medium, or may be in the form of one or more signals.Such signal can be from
Downloading obtains on internet website, is perhaps provided on the carrier signal or is provided in any other form.
It should be noted that the above-mentioned embodiments illustrate rather than limit the invention, and ability
Field technique personnel can be designed alternative embodiment without departing from the scope of the appended claims.In the claims,
Any reference symbol between parentheses should not be configured to limitations on claims.Word "comprising" does not exclude the presence of not
Element or step listed in the claims.Word "a" or "an" located in front of the element does not exclude the presence of multiple such
Element.The present invention can be by means of including the hardware of several different elements and being come by means of properly programmed computer real
It is existing.In the unit claims listing several devices, several in these devices can be through the same hardware branch
To embody.The use of word first, second, and third does not indicate any sequence.These words can be explained and be run after fame
Claim.
The above description is merely a specific embodiment or to the explanation of specific embodiment, protection of the invention
Range is not limited thereto, and anyone skilled in the art in the technical scope disclosed by the present invention, can be easily
Expect change or replacement, should be covered by the protection scope of the present invention.Protection scope of the present invention should be with claim
Subject to protection scope.
Claims (10)
1. a kind of method of multilingual intertranslation in unison characterized by comprising
The identification of languages context is carried out to the voice document that sound pick up equipment generates;
It will identify that the voice document after languages context carries out textual processing, and treated that voice document is cut by textual
Multiple voice document units, to identify calculating and translation respectively;
Multiple institute's voice file units are distributed to multiple processors, by the multiple processor in a parallel fashion to assigning to
Institute's voice file unit execute Distributed identification calculate and translation;And
By the text file that identified and translation institute's voice file unit weight group be after translating, and by the text after the translation
This document is broadcasted in the form of voice document.
2. the method as described in claim 1, which is characterized in that carry out languages context to the voice document that sound pick up equipment generates
Identification further comprises:
The voice selecting of the voice document generated according to the sound pick up equipment talks with languages;
Discourse context is selected according to the dialogue languages and first institute's voice file;And
Automatically the dialogue languages and discourse context of the voice document that the sound pick up equipment generates are confirmed repeatedly.
3. method according to claim 2, which is characterized in that by identified institute's voice file unit weight group be translation after
Text file after, further includes:
The accuracy rate and/or speech quality evaluation value of text file after checking the translation, after the translation of generation
Text file carries out grammer and speech quality correction.
4. method as claimed in claim 3, which is characterized in that the speech quality evaluation value includes Y-PSNR and structure
Similarity indices.
5. method as claimed in claim 3, which is characterized in that multiple institute's voice file units are distributed to multiple processors
Before, the computing capability of the multiple processor is detected, and according to the computing capability of the multiple processor by multiple institute's predicates
Sound file unit is distributed to the multiple processor, further comprises:
Detect the number of the multiple processor;
Detect the number of the core in the multiple processor;And/or
Detect the dominant frequency of the core in the multiple processor.
6. the method as described in any one of claims 1 to 5, which is characterized in that by the multiple processor with parallel
After mode executes Distributed identification calculating and translation to the institute's voice file unit assigned to, the method further includes:
Check the result of described identified and translation institute's voice file unit;And
Institute's voice file unit that identification and translation result do not meet predetermined condition is distributed to the multiple processor again,
To be re-recognized and be translated.
7. method as claimed in claim 6, which is characterized in that do not met described in predetermined condition that will identify with translation result
Voice document unit is distributed to again before the multiple processor, the method further includes:
According to recognition result, the voice text for not meeting predetermined condition in the multiple processor with the recognition result is adjusted
The identification parameter of the corresponding processor of part unit;
Institute's voice file unit that the recognition result does not meet predetermined condition is distributed to again in the multiple processor
The corresponding processor.
8. method as claimed in claim 6, which is characterized in that checking described identified and translation institute's voice file list
After the result of position, the method further includes:
Discharge the corresponding processor of the institute's voice file unit to conform to a predetermined condition in the multiple processor with recognition result.
9. such as the described in any item methods of claim 2 to 5, which is characterized in that carried out to the voice document that sound pick up equipment generates
The identification of languages context further comprises:
Individual information variation detection is carried out to the voice content of institute's voice file, to identify the switching of the discourse context.
10. a kind of equipment of multilingual intertranslation in unison, which is characterized in that including sound pick up equipment, cutting device, distribution translating equipment
With recombination public address system,
The sound pick up equipment is used for the voice document for being identified and being generated to voice;
The cutting device is used to carry out institute's voice file textual processing, and treated that voice document is cut by textual
It is divided into multiple voice document units;
The distribution translating equipment is used to multiple institute's voice file units being distributed to multiple processors, by the multiple processing
Device executes Distributed identification to the institute's voice file unit assigned in a parallel fashion and calculates and translate;And
The recombination public address system is used to be the text text after translation by identified and translation institute's voice file unit weight group
Part, and the text file after the translation is broadcasted in the form of voice document.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201811029882.9A CN109359305A (en) | 2018-09-05 | 2018-09-05 | A kind of method and apparatus of multilingual intertranslation in unison |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201811029882.9A CN109359305A (en) | 2018-09-05 | 2018-09-05 | A kind of method and apparatus of multilingual intertranslation in unison |
Publications (1)
Publication Number | Publication Date |
---|---|
CN109359305A true CN109359305A (en) | 2019-02-19 |
Family
ID=65350262
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201811029882.9A Pending CN109359305A (en) | 2018-09-05 | 2018-09-05 | A kind of method and apparatus of multilingual intertranslation in unison |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN109359305A (en) |
Cited By (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN110516238A (en) * | 2019-08-20 | 2019-11-29 | 广州国音智能科技有限公司 | Voice translation method, device, terminal and computer storage medium |
CN112668346A (en) * | 2020-12-24 | 2021-04-16 | 科大讯飞股份有限公司 | Translation method, device, equipment and storage medium |
CN112818703A (en) * | 2021-01-19 | 2021-05-18 | 传神语联网网络科技股份有限公司 | Multi-language consensus translation system and method based on multi-thread communication |
Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102567312A (en) * | 2011-12-30 | 2012-07-11 | 北京理工大学 | Machine translation method based on distributive parallel computation framework |
KR20160107000A (en) * | 2015-03-03 | 2016-09-13 | 헬로챗 주식회사 | Method for providing multi language learning service |
CN107315742A (en) * | 2017-07-03 | 2017-11-03 | 中国科学院自动化研究所 | The Interpreter's method and system that personalize with good in interactive function |
-
2018
- 2018-09-05 CN CN201811029882.9A patent/CN109359305A/en active Pending
Patent Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102567312A (en) * | 2011-12-30 | 2012-07-11 | 北京理工大学 | Machine translation method based on distributive parallel computation framework |
KR20160107000A (en) * | 2015-03-03 | 2016-09-13 | 헬로챗 주식회사 | Method for providing multi language learning service |
CN107315742A (en) * | 2017-07-03 | 2017-11-03 | 中国科学院自动化研究所 | The Interpreter's method and system that personalize with good in interactive function |
Non-Patent Citations (2)
Title |
---|
姚英彪,易志强主编: "《媒体信号编码》", 31 December 2011, 西安电子科技大学出版社, pages: 0023 * |
王菊丽: "语言 文学 翻译研究专辑2016", 西南交通大学出版社, pages: 0192 * |
Cited By (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN110516238A (en) * | 2019-08-20 | 2019-11-29 | 广州国音智能科技有限公司 | Voice translation method, device, terminal and computer storage medium |
CN110516238B (en) * | 2019-08-20 | 2023-12-19 | 广州国音智能科技有限公司 | Speech translation method, device, terminal and computer storage medium |
CN112668346A (en) * | 2020-12-24 | 2021-04-16 | 科大讯飞股份有限公司 | Translation method, device, equipment and storage medium |
CN112668346B (en) * | 2020-12-24 | 2024-04-30 | 中国科学技术大学 | Translation method, device, equipment and storage medium |
CN112818703A (en) * | 2021-01-19 | 2021-05-18 | 传神语联网网络科技股份有限公司 | Multi-language consensus translation system and method based on multi-thread communication |
CN112818703B (en) * | 2021-01-19 | 2024-02-27 | 传神语联网网络科技股份有限公司 | Multilingual consensus translation system and method based on multithread communication |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
US10777192B2 (en) | Method and apparatus of recognizing field of semantic parsing information, device and readable medium | |
US9665565B2 (en) | Semantic similarity evaluation method, apparatus, and system | |
JP6440732B2 (en) | Automatic task classification based on machine learning | |
KR102163549B1 (en) | Method and apparatus for determining retreat | |
CN108829893A (en) | Determine method, apparatus, storage medium and the terminal device of video tab | |
KR101864361B1 (en) | Method and system for providing translated result | |
CN109359305A (en) | A kind of method and apparatus of multilingual intertranslation in unison | |
JP6655788B2 (en) | Bilingual corpus creation method, apparatus and program, and machine translation system | |
US20150066805A1 (en) | Location-based adaptation of financial management system | |
US20160284344A1 (en) | Speech data recognition method, apparatus, and server for distinguishing regional accent | |
CN106897439A (en) | The emotion identification method of text, device, server and storage medium | |
WO2014117549A1 (en) | Method and device for error correction model training and text error correction | |
CN110019782A (en) | Method and apparatus for exporting text categories | |
CN110516259B (en) | Method and device for identifying technical keywords, computer equipment and storage medium | |
CN109740053A (en) | Sensitive word screen method and device based on NLP technology | |
CN109840883B (en) | Method and device for training object recognition neural network and computing equipment | |
CN109635305A (en) | Voice translation method and device, equipment and storage medium | |
JP2022088602A (en) | Table generation method, device, electronic apparatus, storage medium and program | |
CN114461801A (en) | Patent text classification number identification method and device, electronic equipment and storage medium | |
WO2022178933A1 (en) | Context-based voice sentiment detection method and apparatus, device and storage medium | |
CN109460554A (en) | A kind of method and device of filtering shielding word | |
CN110309513B (en) | Text dependency analysis method and device | |
CN109960752A (en) | Querying method, device, computer equipment and storage medium in application program | |
CN114528851B (en) | Reply sentence determination method, reply sentence determination device, electronic equipment and storage medium | |
CN109657244A (en) | A kind of English long sentence automatic segmentation method and system |
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 |