WO2020052191A1 - 翻译校正方法、装置以及翻译校正设备 - Google Patents

翻译校正方法、装置以及翻译校正设备 Download PDF

Info

Publication number
WO2020052191A1
WO2020052191A1 PCT/CN2019/073762 CN2019073762W WO2020052191A1 WO 2020052191 A1 WO2020052191 A1 WO 2020052191A1 CN 2019073762 W CN2019073762 W CN 2019073762W WO 2020052191 A1 WO2020052191 A1 WO 2020052191A1
Authority
WO
WIPO (PCT)
Prior art keywords
translation
correction
mode
manual
language text
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.)
Ceased
Application number
PCT/CN2019/073762
Other languages
English (en)
French (fr)
Inventor
陈良辉
王文祺
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Shenzhen Water World Co Ltd
Original Assignee
Shenzhen Water World Co Ltd
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by Shenzhen Water World Co Ltd filed Critical Shenzhen Water World Co Ltd
Publication of WO2020052191A1 publication Critical patent/WO2020052191A1/zh
Anticipated expiration legal-status Critical
Ceased legal-status Critical Current

Links

Classifications

    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F40/00—Handling natural language data
    • G06F40/40—Processing or translation of natural language
    • G06F40/51—Translation evaluation
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q20/00—Payment architectures, schemes or protocols
    • G06Q20/08—Payment architectures
    • G06Q20/14—Payment architectures specially adapted for billing systems
    • G06Q20/145—Payments according to the detected use or quantity

Definitions

  • the present application relates to the field of intelligent devices, and in particular, to a method and device for translation correction, and a translation correction device.
  • the main purpose of this application is to provide a translation correction method, which aims to solve the technical problem of low translation accuracy of existing translators or mobile translation apps.
  • This application provides a translation correction method, including:
  • the management server receives a translation mode selection instruction sent by a pre-associated translator.
  • the translation mode includes at least two of an automatic translation mode, a semi-automatic translation mode, and a manual translation mode.
  • a corresponding translation channel is opened according to the translation mode selected by the selection instruction.
  • the translation mode includes an artificial correction option, and the translation channel includes an artificial correction channel corresponding to the artificial correction option; receiving the original speech signal to be translated sent by the translator; and translating the original speech signal to be translated into a target speech signal through the translation channel.
  • This application also provides a translation correction device, which is integrated in the management server and includes:
  • the first receiving module is configured to receive a selection instruction of a translation mode sent by a pre-associated translator.
  • the translation mode includes at least two of an automatic translation mode, a semi-automatic translation mode, and a manual translation mode.
  • the corresponding translation mode opens a corresponding translation channel, wherein the translation mode includes a manual correction option, and the translation channel includes a manual correction channel corresponding to the manual correction option;
  • a second receiving module is configured to receive the original speech signal to be translated sent by the translator; translation A module for translating an original speech signal to be translated into a target speech signal through a translation channel.
  • the present application also provides a translation correction device including a memory, a processor, and an application program.
  • the application program is stored in the memory and configured to be executed by the processor, and the application program is configured to perform the above-mentioned translation correction method.
  • a manual correction option is added on the basis of the original machine translation to improve the accuracy of the translation and meet the user's demand for accurate translation.
  • the translation modes of this application include free automatic translation, paid semi-automatic translation mode and manual translation mode, and the charging methods are flexible and diverse to meet the needs of different users.
  • FIG. 1 is a schematic flowchart of a translation correction method according to an embodiment of the present application.
  • FIG. 2 is a schematic structural diagram of a translation correction device according to an embodiment of the present application.
  • FIG. 3 is a schematic structural diagram of a translation module according to an embodiment of the present application.
  • FIG. 4 is a schematic structural diagram of a second sending unit according to an embodiment of the present application.
  • FIG. 5 is a schematic structural diagram of a second sending unit according to another embodiment of the present application.
  • FIG. 6 is a schematic structural diagram of a translation module according to another embodiment of the present application.
  • FIG. 7 is a schematic structural diagram of a translation correction device according to another embodiment of the present application.
  • FIG. 8 is a schematic structural diagram of a generation module according to an embodiment of the present application.
  • FIG. 9 is a schematic structural diagram of a generating module according to another embodiment of the present application.
  • a translation correction method includes:
  • the management server receives a selection instruction of a translation mode sent by a pre-associated translator.
  • the translation mode includes at least two of an automatic translation mode, a semi-automatic translation mode, and a manual translation mode.
  • the management server in this embodiment connects the translation machine, the servers in the translation process, and the customer service terminals that are manually involved, so that a corresponding translation channel can be set up in time according to the user's selection to meet the user's translation needs.
  • the management server in this embodiment is connected to the translator, each server in the translation process, and the customer service terminal that participates manually.
  • the management server can establish an association relationship by identifying the device number corresponding to each end or the association signal in the handshake protocol.
  • the translation mode of this embodiment includes not only an automatic translation mode, but also a semi-automatic translation mode and a manual translation mode with manual participation to meet the needs of different user groups.
  • manual translation is added to the translation channel, including manual correction options, to improve the accuracy of the translation file.
  • the manual correction options of this embodiment include recognition correction, translation correction, and synthesis correction.
  • Manual translation is used to improve the accuracy of translation, so that the translation content can be closer to the current specific use occasion, such as a specific business meeting or a professional field meeting.
  • a translation mode selected by a user and an original voice signal to be translated are received by a translator, so as to improve the convenience of the user in using the translator.
  • the translator of this embodiment has a mutual translation function.
  • the original voice signal to be translated refers to a voice of a common language entered by a user, so that the voice of the common language is translated into a voice of a current regional communication language (that is, a target voice signal) for real-time communication;
  • the original speech signal to be translated or a language that is not commonly used by the user is received for timely translation into the language that is commonly used by the user (ie, the target speech signal), which is that the user receives the corresponding information in time.
  • the translation mode selected by the selection instruction is a semi-automatic translation mode.
  • the semi-automatic translation mode includes a manual calibration option.
  • the manual correction option includes one or two of recognition correction, translation correction, and synthesis correction.
  • the translation channel includes a correspondence with the manual calibration option. Step S4 of translating the original speech signal to be translated into the target speech signal through the translation channel, including:
  • S40 Send the original voice signal to be translated to the voice server through the translation channel.
  • the management server sends the original speech signal to the speech server, so that after the translation is completed, the management server sends the translated target speech signal to the translator for playback.
  • Other embodiments of the present application may also be directly associated with the voice server through a translator. Only when manual correction access is needed, the manual correction is associated with the management server.
  • Judgment is made by identifying selection information carried by the user-selected translation mode sent by the translator. For example, if it is recognized that the above-mentioned selection information includes the keyword of “recognition correction”, it is determined that the recognition correction is included, otherwise it is not included.
  • the selection information carried in the translation mode of this embodiment includes recognition correction, and then the management server accesses the manual recognition correction after being recognized by the machine, that is, after receiving the first target language text corresponding to the original voice signal fed back by the voice server Sending the above-mentioned first target language text and the original speech signal to be translated to the first customer service terminal corresponding to manual recognition and correction, so as to make the speech-to-text conversion in the voice recognition process more suitable for the current specific use scenario through manual correction To improve translation accuracy.
  • the management server is associated with machine translation and manual correction, and the files completed by each server during the machine translation process are directed to the designated customer service terminal corresponding to the manual correction according to the user's selection.
  • Manually corrected files are re-sent to the server corresponding to the next link in the machine translation process according to the machine translation process, so that the entire translation process is smoother.
  • the machine translation process of this embodiment includes three steps: the process of speech recognition converting speech into text, the process of translating the original language text into the target language text, and the process of synthesizing the target language text into the target speech.
  • Each of the above three links can be connected to the manual correction process, that is, after the previous machine translation link and before the next machine translation link, the manual correction link is connected.
  • speech recognition adds artificial recognition and correction after the process of converting speech to text, and adds manual translation correction after the process of translating the original language text into the target language text, and adds artificial synthesis correction after the process of synthesizing the target language text into the target speech.
  • the original voice signal is used as the reference for correction.
  • the step S45 of translating the corrected first target language text into the target speech signal through a translation channel includes: S450: sending the corrected first target language text to the translation server through the translation channel.
  • S450a Receive the second target language text corresponding to the corrected first target language text sent by the translation server.
  • S451 Determine whether the manual correction option includes translation correction. Judgment is made by identifying selection information carried by the user-selected translation mode sent by the translator. For example, if it is recognized that the above-mentioned selection information includes the "translation correction" keyword, it is determined to include the identification correction, otherwise it is not included.
  • S452 If translation correction is included, the second target language text and the original speech signal to be translated are sent to a second customer service terminal provided by the manual calibration channel for translation correction.
  • S452a Receive the second target language text corrected by the second customer service terminal.
  • the manual translation correction and manual recognition correction in this embodiment are the same customer service terminal to save human resources and simplify the related information of the translation channel.
  • manual recognition correction, manual translation correction, and synthetic synthesis correction may be different customer service terminals, respectively, to improve translation efficiency and translation quality.
  • S453 Send the corrected second target language text to the speech synthesis server through the translation channel for speech synthesis.
  • S453a Receive a target speech signal corresponding to the second target language text sent by the speech synthesis server.
  • This embodiment takes a semi-automatic translation mode in which a manual translation correction and a manual recognition correction are simultaneously added as an example in the machine translation process, and details the implementation process of adding a specified manual correction to the original machine translation channel according to the user's selection information.
  • the semi-automatic translation mode of this embodiment means that the manual correction option includes at least one or two of recognition correction, translation correction, and synthesis correction.
  • the user can freely choose according to his own needs.
  • the management server constructs a translation that meets the user's choice according to the user's choice. aisle. Zh
  • step S45 determining whether the manual correction option includes translation correction, the method includes:
  • S456a Receive a specified target speech signal corresponding to the second target language text sent by the speech synthesis server.
  • Judgment is made by identifying selection information carried by the user-selected translation mode sent by the translator. For example, if it is recognized that the above-mentioned selection information includes the keyword of “composite correction”, it is determined that the identification correction is included, otherwise it is not included.
  • S459 Receive the target voice signal fed back by the third customer service terminal.
  • a semi-automatic translation mode in which artificial synthesis correction and artificial recognition correction are simultaneously added in a machine translation process is taken as an example, and an implementation process of adding a specified manual correction to an original machine translation channel according to user selection information is described in detail.
  • Others include the addition of a single manual correction option, or the addition of both artificial translation correction and artificial synthesis correction, and the formation process and translation process of the translation channels corresponding to the semi-automatic translation modes respectively formed are similar to the specific embodiments described above, and are not repeated here.
  • the translation mode selected by the selection instruction is a manual translation mode.
  • the manual translation mode includes a manual calibration option.
  • the manual correction option includes recognition correction, translation correction, and synthesis correction.
  • the translation channel includes a manual calibration channel corresponding to the manual calibration option. Step S4 of translating the translated original speech signal into the target speech signal through the translation channel includes:
  • S401 Send the original voice signal to be translated to the voice server through the translation channel.
  • S402 Receive a third target language text corresponding to the original voice signal to be translated sent by the voice server.
  • S403 Send the text in the third target language and the original voice signal to be translated to the designated customer service terminal provided by the manual calibration channel for identification and correction.
  • S404 Receive the corrected third target language text fed back by the designated customer service terminal.
  • S405 Send the corrected third target language text to the translation server for translation through the translation channel.
  • S405a Receive the fourth target language text corresponding to the corrected third target language text sent by the translation server.
  • S406 Send the text in the fourth target language and the original voice signal to be translated to the designated customer service terminal provided by the manual calibration channel for translation correction.
  • S407 Receive the corrected fourth target language text returned by the designated customer service terminal.
  • S408 Send the corrected fourth target language text to the speech synthesis server for speech synthesis.
  • S408a Receive the first speech signal corresponding to the corrected fourth target language text sent by the speech synthesis server.
  • S409 Send the first voice signal and the original voice signal to be translated to a designated customer service terminal provided by the manual calibration channel for synthesis correction.
  • S409a Receive the target voice signal from the designated customer service terminal.
  • This embodiment takes an example of an artificial translation mode in which artificial recognition correction, artificial translation correction, and synthetic synthesis correction are added to a machine translation process as an example, and details are provided that a specified artificial correction is added to an original machine translation channel according to user selection information.
  • Implementation process The difference between the manual translation mode and the semi-automatic translation mode in this embodiment is that the number of manual correction option items included is different, and the charges are also different.
  • the implementation process of adding manual correction to the original machine translation process is similar and will not be described in detail.
  • step S4 of translating the original speech signal to be translated into a target speech signal through a translation channel the method includes:
  • the automatic translation of the translator in this embodiment is free, but a fee is charged when manual correction is required, and the semi-automatic translation mode or manual translation mode is selected by the user, and the corresponding translation billing methods are also different, as described above.
  • the preset rules are not only related to the translation mode selected by the user, but also to the specific options of the manual correction option selected by the user. The cost of different manual correction options is different. Selecting a semi-automatic translation mode for manual correction and selecting two manual corrections The billing of the semi-automatic translation model is different.
  • the foregoing preset rule further includes whether the charging method selected by the user is timed charging or charging based on the number of translated words.
  • S6 Send translation billing to a designated terminal to receive translation billing from the user.
  • the translation billing in this embodiment is completed by the management server, it is sent to the translator to facilitate payment by the user.
  • the translator of this embodiment automatically detects whether there is arrears before the user starts the billing mode, and sends a reminder of arrears if there is arrears, and at the same time restricts the continued use of the semi-automatic translation mode or the manual translation mode to control the semi-automatic translation mode. Or manual translation mode is disabled.
  • step S5 of generating a corresponding translation charge according to a preset rule includes:
  • S501 Determine whether the translation mode selected by the selection instruction is a billing mode, where the billing mode includes a semi-automatic translation mode and a manual translation mode, and the semi-automatic translation mode corresponds to a first billing mode and the manual translation mode corresponds to a second billing mode.
  • the first charging method of this embodiment is different according to the manual correction option selected by the user.
  • the cost of selecting one manual correction option is 1 yuan / min (billing unit price)
  • the cost of two manual correction options is 1.2. Yuan / min (billing unit price)
  • Yuan / min that is, after selecting the semi-automatic translation mode, the fee starts to charge from 1 yuan / min (billing unit price), and another manual correction option is added at 1.2 yuan / min (billing unit price). fee.
  • the charge of the second charging method in this embodiment starts from 1.5 yuan / min (billing unit price).
  • the terms “first” and “second” in this embodiment are used only for distinguishing and are not used for limitation. Such terms in other places have the same function and are not repeated.
  • the charging mode of the translator of the present application includes two methods of counting the total cost of timed translation and the cost of translated words. This embodiment uses timed translation as an example.
  • S503 If it is a timed translation, count the charging time from the time when the user starts to record the original speech signals to be translated to the completion of all the original speech signals to be translated.
  • the management server sets a preset time, such as 5 seconds, and automatically receives a timed waiting state after receiving the translation sentence entered by the user to wait for the receiving user. Enter the next sentence to be translated, and determine whether the waiting time exceeds the preset time in real time. If it exceeds the preset time, the user has not received the next sentence to be translated, and the user has not received a click on the mobile translation app or translator.
  • the translation end instruction will automatically pause the translation timing. After automatically suspending the translation timing, start the process of suspending the translation duration statistics, and monitor in real time whether the user continues to enter the sentence to be translated within the specified time.
  • the timer is started to continue to count. If it is not received, it is counted and recorded. Suspend the translation time, determine whether the pause time exceeds the specified time, such as 1 hour, if it exceeds the specified time, determine whether the user has entered the sentence to be translated after the pause time, if not, end the translation time, and record the end time, and Calculate the cost of this time period and send it to the mobile translation APP or translator.
  • the "charging duration end time-start time-pause translation time" in this time period.
  • S504 Calculate the first translation billing corresponding to the semi-automatic translation mode according to the billing duration and the billing standard corresponding to the first billing method; or calculate the manual translation according to the billing duration and the billing standard corresponding to the second billing method.
  • the second translation corresponding to the mode is charged.
  • the first charging method of this embodiment includes a charging type, such as timed charging or charging based on the number of translated words, and further includes a charging standard, which is a charging unit price.
  • step S5 of generating a corresponding translation charge according to a preset rule includes:
  • S505 Determine whether the translation mode selected by the selection instruction is a billing mode, wherein the billing mode includes a semi-automatic translation mode and a manual translation mode, and the semi-automatic translation mode corresponds to a third billing mode and the manual translation mode corresponds to a fourth billing mode.
  • the third charging method of this embodiment is different according to the manual correction option selected by the user.
  • the cost of selecting one manual correction option is 100 yuan / thousand words (billing unit price)
  • the cost of two manual correction options is 150 yuan / thousand words (billing unit price)
  • the fee starts from 100 yuan / thousand words (billing unit price)
  • another manual correction option is added to 150 yuan / thousand words. (Billing unit price).
  • the charge of the fourth charging method in this embodiment starts from 200 yuan / thousand words (charging unit price).
  • the number of translated words refers to the total number of words corresponding to the target speech signal.
  • word count is performed on the text before the target speech signal is synthesized to obtain the total number of words corresponding to the target speech signal.
  • S508 Calculate the third translation charge corresponding to the semi-automatic translation mode according to the total word count and the charging standard corresponding to the third charge method; or calculate the manual translation mode corresponding to the total word count and the fourth charge method.
  • the fourth translation is billed.
  • the third charging method of this embodiment includes a charging type, that is, charging based on the number of translated words, and a charging standard, that is, a charging unit price.
  • a manual correction option is added on the basis of the original machine translation, so as to improve the accuracy of the translation and meet the user's demand for accurate translation.
  • the translation mode in this embodiment includes a free automatic translation, a semi-automatic translation mode with a fee, and a manual translation mode, and the charging methods are flexible and diverse, which can meet the needs of different users.
  • the present application further provides a translation correction device integrated in a management server, including: a first receiving module 1 for receiving a selection instruction of a translation mode sent by a pre-associated translator, and the translation mode includes automatic translation Mode, semi-automatic translation mode and human translation mode.
  • the opening module 2 is configured to open a corresponding translation channel according to a translation mode selected by a selection instruction.
  • the second receiving module 3 is configured to receive the original speech signal to be translated sent by the translator.
  • a translation module 4 is configured to translate an original speech signal to be translated into a target speech signal through a translation channel.
  • the translation mode selected by the selection instruction is a semi-automatic translation mode.
  • the semi-automatic translation mode includes a manual calibration option.
  • the manual correction option includes one or two of recognition correction, translation correction, and synthesis correction.
  • the translation channel includes a manual calibration channel corresponding to the manual calibration option.
  • Translation module 4 includes:
  • the first sending unit 40 is configured to send an original voice signal to be translated to a voice server through a translation channel.
  • the first receiving unit 41 is configured to receive a first target language text corresponding to an original voice signal to be translated sent by a voice server.
  • the first determining unit 42 is configured to determine whether a manual correction option includes recognition correction.
  • the recognition and correction unit 43 is configured to: if the recognition and correction is included, send the first target language text and the original speech signal to be translated to the first customer service terminal provided by the manual calibration channel for recognition and correction, and obtain the corrected first target language text.
  • the second receiving unit 44 is configured to receive the corrected first target language text fed back by the first customer service terminal.
  • the second sending unit 45 is configured to translate the corrected first target language text into a target speech signal through a translation channel.
  • the second sending unit 45 includes:
  • the first sending subunit 450 is configured to send the corrected first target language text to a translation server through a translation channel.
  • the first receiving subunit 450a is configured to receive a second target language text corresponding to the corrected first target language text sent by the translation server.
  • the first judging subunit 451 is configured to judge whether the manual correction option includes translation correction.
  • the second sending subunit 452 is configured to send the second target language text and the original speech signal to be translated to the second customer service terminal provided by the manual calibration channel for translation correction if translation correction is included.
  • the second receiving subunit 452a is configured to receive the corrected second target language text fed back by the second customer service terminal.
  • the third sending subunit 453 is configured to send the corrected second target language text to the speech synthesis server through the translation channel for speech synthesis.
  • the third receiving subunit 453a is configured to receive a target voice signal corresponding to the second target language text sent by the voice synthesis server.
  • the second sending unit 45 includes:
  • the fourth sending subunit 456 is configured to send the second target language text to the speech synthesis server for speech synthesis if translation correction is not included.
  • the fourth receiving subunit 456a is configured to receive a specified target speech signal corresponding to the second target language text sent by the speech synthesis server.
  • the second judging subunit 457 is configured to judge whether the artificial correction option includes synthetic correction.
  • the fifth sending sub-unit 458 is configured to send the designated target voice signal and the original voice signal to be translated to a third customer service terminal provided by the manual calibration channel for synthetic correction if the synthetic correction is included.
  • the fifth receiving subunit 459 is configured to receive a target voice signal fed back by the third customer service terminal.
  • the translation mode selected by the selection instruction is a manual translation mode.
  • the manual translation mode includes a manual calibration option.
  • the manual correction option includes recognition correction, translation correction, and synthesis correction.
  • the manual calibration channel corresponding to the option, the translation module 4, includes: a third sending unit 401, configured to send the original voice signal to be translated to the voice server through the translation channel.
  • the third receiving unit 402 is configured to receive a third target language text corresponding to the original voice signal to be translated sent by the voice server.
  • a fourth sending unit 403 is configured to send the text in the third target language and the original voice signal to be translated to a designated customer service terminal provided by the manual calibration channel for identification and correction.
  • the fourth receiving unit 404 is configured to receive the corrected third target language text fed back by the designated customer service terminal.
  • the fifth sending unit 405 is configured to send the corrected third target language text to a translation server for translation through a translation channel.
  • the fifth receiving unit 405a is configured to receive a fourth target language text corresponding to the corrected third target language text sent by the translation server.
  • a sixth sending unit 406 is configured to send the text in the fourth target language and the original speech signal to be translated to a designated customer service terminal provided by the manual calibration channel for translation correction, and obtain the corrected fourth target language text.
  • the sixth receiving unit 407 is configured to receive the corrected fourth target language text fed back by the designated customer service terminal.
  • the seventh sending unit 408 is configured to send the corrected fourth target language text to the speech synthesis server for speech synthesis.
  • the seventh receiving unit 408a is configured to receive the first speech signal corresponding to the corrected fourth target language text sent by the speech synthesis server.
  • An eighth sending unit 409 is configured to send the first voice signal and the original voice signal to be translated to a designated customer service terminal provided by a manual calibration channel for synthesis correction.
  • An eighth receiving unit 409a is configured to receive a target voice signal from a designated customer service terminal.
  • a translation correction apparatus includes: a generating module 5 configured to generate a corresponding translation charge according to a preset rule.
  • Payment module 6 Send translation billing to a designated terminal to receive translation billing from the user.
  • a generating module 5 includes: a second determining unit 501 configured to determine whether a translation mode selected by a selection instruction is a charging mode, where the charging mode includes a semi-automatic translation mode and a manual translation mode, And the semi-automatic translation mode corresponds to the first charging method and the manual translation mode corresponds to the second charging method.
  • the third determining unit 502 is configured to determine whether the first charging method or the second charging method is a timed translation if the translation mode is a charging mode.
  • the first counting unit 503 is configured to count the charging duration between the time when the user starts recording the original speech signals to be translated and the time when all the original speech signals to be translated are completed, if the timed translation is used.
  • the first calculation unit 504 is configured to calculate a first translation billing corresponding to the semi-automatic translation mode according to the billing duration and a billing standard corresponding to the first billing mode; The fee standard calculates the second translation charge corresponding to the manual translation mode.
  • a generating module 5 includes: a fourth determining unit 505, configured to determine whether a translation mode selected by a selection instruction is a charging mode, where the charging mode includes a semi-automatic translation mode and a manual translation mode , And the semi-automatic translation mode corresponds to the third charging method and the manual translation mode corresponds to the fourth charging method.
  • the fifth determining unit 506 is configured to determine whether the third charging method or the fourth charging method is charging for the number of translated words if the translation mode is a charging mode.
  • the second counting unit 507 is configured to count the total number of words corresponding to the target speech signal if the translation word is charged.
  • the second calculation unit 508 is configured to calculate the third translation charge corresponding to the semi-automatic translation mode according to the total word count and the charge standard corresponding to the third charge method; or according to the total word count and the charge standard corresponding to the fourth charge method , Calculate the fourth translation charge corresponding to the manual translation mode.
  • the present application also provides a translation correction device, including a memory, a processor, and an application program.
  • the application program is stored in the memory and configured to be executed by the processor, and the application program is configured to execute any one of the foregoing embodiments. Translation correction method.
  • the equipment may be specially designed and manufactured for the required purpose, or it may include known equipment in a general purpose computer.
  • the device has computer programs or application programs stored therein that are selectively activated or reconstructed.
  • Such a computer program may be stored in a device (eg, a computer) readable medium or in any type of medium suitable for storing electronic instructions and coupled to a bus, respectively.
  • the computer readable medium includes, but is not limited to, any type of Disk (including floppy disk, hard disk, CD-ROM, CD-ROM, and magneto-optical disk), ROM (Read-Only Memory, read-only memory), RAM (Random Access Memory (random memory), EPROM (Erasable Programmable Read-Only Memory (Erasable Programmable Read-Only Memory), EEPROM (Electrically Erasable Programmable Read-Only Memory), flash memory, magnetic card or optical card. That is, a readable medium includes any medium that stores or transfers information in a readable form by a device (eg, a computer).
  • a device eg, a computer

Landscapes

  • Engineering & Computer Science (AREA)
  • Business, Economics & Management (AREA)
  • Accounting & Taxation (AREA)
  • Theoretical Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • Finance (AREA)
  • Physics & Mathematics (AREA)
  • Computational Linguistics (AREA)
  • General Engineering & Computer Science (AREA)
  • General Health & Medical Sciences (AREA)
  • Health & Medical Sciences (AREA)
  • Audiology, Speech & Language Pathology (AREA)
  • Development Economics (AREA)
  • Economics (AREA)
  • Artificial Intelligence (AREA)
  • Strategic Management (AREA)
  • General Business, Economics & Management (AREA)
  • Machine Translation (AREA)

Abstract

本申请揭示了翻译校正方法，包括：管理服务器接收预关联的翻译机发送的翻译模式的选择指令，翻译模式包括自动翻译模式、半自动翻译模式和人工翻译模式中的至少两种；根据选择指令选择的翻译模式开通对应的翻译通道；接收翻译机发送的待翻译的原语音信号；将待翻译的原语音信号通过翻译通道翻译成目标语音信号。

Description

翻译校正方法、装置以及翻译校正设备 技术领域
本申请涉及到智能设备领域，特别是涉及到翻译校正方法、装置以及翻译校正设备。
 背景技术
过去一年中国出境人口达到 1.3 亿次，语言不通仍然是他们的痛点之一，翻译机或手机翻译APP的需求市场依然在扩大，但现有翻译机或手机翻译APP为机器翻译，准确性低，尤其在重要商务会议使用时，翻译准确性低严重影响商务谈判质量，因此现有翻译机或手机翻译APP的翻译技术还有待改进。
 技术问题
本申请的主要目的为提供一种翻译校正方法，旨在解决现有翻译机或手机翻译APP的翻译准确性低的技术问题。
 技术解决方案
本申请提供了一种翻译校正方法，包括：
管理服务器接收预关联的翻译机发送的翻译模式的选择指令，翻译模式包括自动翻译模式、半自动翻译模式和人工翻译模式中的至少两种；根据选择指令选择的翻译模式开通对应的翻译通道，其中翻译模式包括人工校正选项，翻译通道包括与人工校正选项对应的人工校正通道； 接收翻译机发送的待翻译的原语音信号； 将待翻译的原语音信号通过翻译通道翻译成目标语音信号。
本申请还提供了一种翻译校正装置，集成于管理服务器中，包括：
第一接收模块，用于接收预关联的翻译机发送的翻译模式的选择指令，翻译模式包括自动翻译模式、半自动翻译模式和人工翻译模式中的至少两种；开通模块，用于根据选择指令选择的翻译模式开通对应的翻译通道，其中翻译模式包括人工校正选项，翻译通道包括与人工校正选项对应的人工校正通道； 第二接收模块，用于接收翻译机发送的待翻译的原语音信号；翻译模块，用于将待翻译的原语音信号通过翻译通道翻译成目标语音信号。
本申请还提供了一种翻译校正设备，包括存储器、处理器和应用程序，应用程序被存储在存储器中并被配置为由处理器执行，应用程序被配置为用于执行上述的翻译校正方法。
 有益效果
本申请通过在原机器翻译的基础上增加了人工校正选项，以提高翻译的准确度，满足用户对准确翻译的使用需求。本申请的翻译模式包括免费的自动翻译、收费的半自动翻译模式和人工翻译模式，且收费方式灵活多样，满足不同用户的使用需求。
 附图说明
图1 本申请一实施例的翻译校正方法的流程示意图；
图2 本申请一实施例的翻译校正装置的结构示意图；
图3 本申请一实施例的翻译模块的结构示意图；
图4 本申请一实施例的第二发送单元的结构示意图；
图5 本申请另一实施例的第二发送单元的结构示意图；
图6 本申请另一实施例的翻译模块的结构示意图；
图7 本申请另一实施例的翻译校正装置的结构示意图；
图8 本申请一实施例的生成模块的结构示意图；
图9 本申请另一实施例的生成模块的结构示意图。
 本发明的最佳实施方式
参照图1，本申请一实施例的翻译校正方法，包括：
S1：管理服务器接收预关联的翻译机发送的翻译模式的选择指令，翻译模式包括自动翻译模式、半自动翻译模式和人工翻译模式中的至少两种。
本实施例的管理服务器衔接翻译机、翻译过程中的各服务器以及人工参与的客服终端，以便根据用户的选择及时搭建好对应的翻译通道，满足用户的翻译需求。本实施例的管理服务器衔接翻译机、翻译过程中的各服务器以及人工参与的客服终端，可通过识别各端对应的设备号，或识别握手协议中的关联信号建立关联关系。本实施例的翻译模式不仅包括自动翻译模式，还包括人工参与的半自动翻译模式和人工翻译模式，以满足不同用户群体的使用需求。
S2：根据选择指令选择的翻译模式开通对应的翻译通道。
本实施例通过在翻译通道中增加人工参与的环节，包括人工校正选项，以提高翻译文件的准确性。本实施例的人工校正选项包括识别校正、翻译校正和合成校正，通过人工校正提高翻译的准确度，以便使翻译内容更能贴近当前的具体使用场合，比如特定商务会议、专业领域会议等。
 S3：接收翻译机发送的待翻译的原语音信号。
本实施例通过翻译机接收用户选定的翻译模式以及待翻译的原语音信号，以便提高用户使用翻译机的便利性。
 S4：将待翻译的原语音信号通过翻译通道翻译成目标语音信号。
本实施例的翻译机具有互译功能，上述待翻译的原语音信号指用户录入的常用语种语音，以便将常用语种语音翻译成当前地域交流语种语音（即目标语音信号），以实现实时沟通；上述待翻译的原语音信号或是接收到用户不常用的语种语音，以便及时翻译成用户常用的语种语音（即目标语音信号），是用户及时接收到对应的信息。
进一步地，选择指令选择的翻译模式为半自动翻译模式，半自动翻译模式包括人工校准选项，人工校正选项包括识别校正、翻译校正和合成校正中的一项或两项，翻译通道包括与人工校准选项对应的人工校准通道，将待翻译的原语音信号通过翻译通道翻译成目标语音信号的步骤S4，包括：
S40：通过翻译通道将待翻译的原语音信号发送至语音服务器。
本实施例通过管理服务器将原语音信号发送至语音服务器，以便完成翻译后通过管理服务器将翻译好的目标语音信号发送至翻译机进行播放。本申请其他实施例也可通过翻译机直接与语音服务器关联，仅在需要人工校正接入的时候，通过管理服务器关联人工校正。
S41：接收语音服务器发送的待翻译的原语音信号对应的第一目标语言文本。
本实施例在无人工校正介入的环节，同原机器翻译的实现过程和原理相同，不赘述。
S42：判断人工校正选项是否包括识别校正。
通过识别翻译机发送的用户选择的翻译模式携带的选择信息进行判断。比如识别到上述选择信息中包括“识别校正”关键字，则判定包括识别校正，否则不包括。
S43：若包括识别校正，则将第一目标语言文本和待翻译的原语音信号发送至人工校准通道提供的第一客服终端进行识别校正。
本实施例的翻译模式携带的选择信息中包括了识别校正，则管理服务器则在通过机器识别之后接入人工识别校正，即在接收到语音服务器反馈的原语音信号对应的第一目标语言文本后，将上述第一目标语言文本和待翻译的原语音信号发送至人工识别校正对应的第一客服终端，以通过人工校正使语音识别过程中的语音到文本的转化更贴合当前具体的使用场景，提高翻译准确度。
S44：接收第一客服终端反馈的校正第一目标语言文本。
S45：将校正第一目标语言文本通过上述翻译通道翻译成目标语音信号。
本实施例通过管理服务器关联机器翻译和人工校正，将机器翻译过程中各服务器完成的文件按照用户的选择，指向性地发送到指定的人工校正对应的客服终端，并将客服终端发送的完成了人工校正的文件按照机器翻译流程，重新发送到机器翻译流程中对应的下一环节对应的服务器，以便使整个翻译流程更加顺畅。本实施例的机器翻译流程包括三个环节，语音识别将语音转换为文本的过程、将原语种文本翻译成目标语种文本的过程和将目标语种文本合成目标语音的过程。上述三个环节中，每个环节均可接入人工校正的过程，即在上一机器翻译环节后，进入下一机器翻译环节前，接入人工校正环节。比如语音识别将语音转换为文本的过程后加入人工识别校正、将原语种文本翻译成目标语种文本的过程后加入人工翻译校正、将目标语种文本合成目标语音的过程后加入人工合成校正等，本实施例的人工校正中均以原语音信号为准进行校正。
进一步地，本申请另一实施例中，将校正第一目标语言文本通过翻译通道翻译成目标语音信号的步骤S45，包括：S450：通过翻译通道将校正第一目标语言文本发送至翻译服务器。S450a：接收翻译服务器发送的校正第一目标语言文本对应的第二目标语言文本。S451：判断人工校正选项是否包括翻译校正。通过识别翻译机发送的用户选择的翻译模式携带的选择信息进行判断。比如识别到上述选择信息中包括“翻译校正”关键字，则判定包括识别校正，否则不包括。S452：若包括翻译校正，则将第二目标语言文本和待翻译的原语音信号发送至人工校准通道提供的第二客服终端进行翻译校正。S452a：接收第二客服终端反馈的校正第二目标语言文本。本实施例的人工翻译校正和人工识别校正为同一个客服终端，以节省人力资源、简化翻译通道的关联信息。本申请其他实施例中，人工识别校正、人工翻译校正以及人工合成校正可以分别为不同的客服终端，以提高翻译效率和翻译质量。S453：通过翻译通道将校正第二目标语言文本发送至语音合成服务器进行语音合成。S453a：接收语音合成服务器发送的第二目标语言文本对应的目标语音信号。
本实施例以在机器翻译流程中同时加入了人工翻译校正和人工识别校正的半自动翻译模式为例，详细说明了根据用户的选择信息在原机器翻译通道中加入了指定的人工校正的实现过程。本实施例的半自动翻译模式是指人工校正选项包括识别校正、翻译校正和合成校正中至少一项或两项，用户可根据自身需要自由选择，管理服务器根据用户的选择搭建符合用户选择意向的翻译通道。  
进一步地，判断人工校正选项是否包括翻译校正的步骤S45之后，包括：
S456：若不包括翻译校正，则将第二目标语言文本发送至语音合成服务器进行语音合成；
S456a：接收语音合成服务器发送的第二目标语言文本对应的指定目标语音信号。
S457：判断人工校正选项是否包括合成校正。
通过识别翻译机发送的用户选择的翻译模式携带的选择信息进行判断。比如识别到上述选择信息中包括“合成校正”关键字，则判定包括识别校正，否则不包括。
S458：若包括合成校正，则将指定目标语音信号和待翻译的原语音信号发送至人工校准通道提供的第三客服终端进行合成校正;
S459：接收第三客服终端反馈的目标语音信号。
本实施例以在机器翻译流程中同时加入了人工合成校正和人工识别校正的半自动翻译模式为例，详细说明了根据用户的选择信息在原机器翻译通道中加入了指定的人工校正的实现过程。其他加入某一单一的人工校正选项，或同时加入人工翻译校正和人工合成校正，而分别形成的半自动翻译模式对应的翻译通道的形成过程、翻译过程如上述具体实施例相似，不赘述。
进一步地，选择指令选择的翻译模式为人工翻译模式，人工翻译模式包括人工校准选项，人工校正选项包括识别校正、翻译校正和合成校正，翻译通道包括与人工校准选项对应的人工校准通道，将待翻译的原语音信号通过翻译通道翻译成目标语音信号的步骤S4，包括：
S401：通过翻译通道将待翻译的原语音信号发送至语音服务器。S402：接收语音服务器发送的待翻译的原语音信号对应的第三目标语言文本。S403：将第三目标语言文本和待翻译的原语音信号发送至人工校准通道提供的指定客服终端进行识别校正。S404：接收指定客服终端反馈的校正第三目标语言文本。S405：通过翻译通道将校正第三目标语言文本发送至翻译服务器进行翻译。S405a：接收翻译服务器发送的校正第三目标语言文本对应的第四目标语言文本。S406：将第四目标语言文本和待翻译的原语音信号发送至人工校准通道提供的指定客服终端进行翻译校正。S407：接收指定客服终端反馈的校正第四目标语言文本。S408：将校正第四目标语言文本发送至语音合成服务器进行语音合成。S408a：接收语音合成服务器发送的校正第四目标语言文本对应的第一语音信号。S409：将第一语音信号和待翻译的原语音信号发送至人工校准通道提供的指定客服终端进行合成校正。S409a：接收指定客服终端反馈的目标语音信号。
本实施例以在机器翻译流程中同时加入了人工识别校正、人工翻译校正和人工合成校正的人工翻译模式为例，详细说明了根据用户的选择信息在原机器翻译通道中加入了指定的人工校正的实现过程。本实施例的人工翻译模式与半自动翻译模式，区别在于所包括的人工校正选项项目数不同，收费也不同，但在原机器翻译流程中加入人工校正的实现过程相类似，不赘述。
进一步地，将待翻译的原语音信号通过翻译通道翻译成目标语音信号的步骤S4之后，包括：
S5：按照预设规则生成对应的翻译计费。
本实施例的翻译机的自动翻译是免费的，但需要人工校正时是需要收取费用的，且根据用户选择的是半自动翻译模式或人工翻译模式，所分别对应的翻译计费方法也不同，上述预设规则不仅跟用户选择的翻译模式相关，且跟用户选择的人工校正选项的具体选项有关，不同的人工校正选项对应的费用不同，选择一项人工校正的半自动翻译模式与选择两项人工校正的半自动翻译模式的计费不同。且上述预设规则还包括用户选择的计费方式为计时计费还是按翻译字数收费有关。
S6：将翻译计费发送至指定终端，以接收用户的支付的翻译计费。
本实施例的翻译计费由管理服务器完成后，并发送至翻译机端，以方便用户进行支付。本实施例的翻译机在用户启动计费模式前会自动检测是否存在欠费，若存在欠费则发出欠费提醒，并同时限制继续使用收费的半自动翻译模式或人工翻译模式，控制半自动翻译模式或人工翻译模式处于禁用状态。
进一步地，按照预设规则生成对应的翻译计费的步骤S5，包括：
S501：判断选择指令选择的翻译模式是否为计费模式，其中计费模式包括半自动翻译模式和人工翻译模式，以及半自动翻译模式对应第一计费方式和人工翻译模式对应第二计费方式。
本实施例的第一计费方式根据用户选择的人工校正选项的不同而不同，举例地，选择一项人工校正选项费用为1元/min（计费单价），两项人工校正选项费用为1.2元/min（计费单价），即选择半自动翻译模式后，费用以1元/min（计费单价）起开始计费，增加另一项人工校正选项以1.2元/min（计费单价）计费。本实施例的第二计费方式的费用以1.5元/min（计费单价）起开始计费。本实施例的“第一”、“第二”等用于仅用于区别，不用于限定，其他处此类用语作用相同，不赘述。
S502：若翻译模式为计费模式，则判断第一计费方式或第二计费方式是否为计时翻译。
本申请的翻译机的计费模式中，包括计时翻译和翻译字数收费两种统计总费用方式，本实施例以计时翻译为例。
S503：若为计时翻译，则统计用户从开始录入待翻译的原语音信号起，至完成所有待翻译的原语音信号之间的计费时长。
当用户选择启动了计费模式后，当管理服务器收到用户的第一句待翻译的语音时开始翻译计时，并记录开始时间，直到用户在手机翻译APP或翻译机上点击翻译结束指令时终止翻译计时，并记录结束时间，统计出上述计费时长，即“计费时长=结束时间-开始时间”，并计算上述计费时长对应的翻译费用发送手机翻译APP或翻译机，待用户结算后，下一次才可继续使用半自动翻译模式或人工翻译模式。本实施例为了避免用户忘记在手机翻译APP或翻译机上点击翻译结束指令，管理服务器设定预设时间比如5秒，在接收到用户录入的翻译语句后，自动进入计时等待状态，以等待接收用户录入下一句待翻译的语句，实时判断计时等待状态是否超过预设时间，若超过预设时间，未接收到用户录入下一句待翻译的语句，且未接收到用户在手机翻译APP或翻译机上点击翻译结束指令，则自动暂停翻译计时。自动暂停翻译计时后，启动暂停翻译时长统计过程，在指定时间内实时监听是否接收到用户继续录入待翻译的语句，若接收到，则启动计时器继续累积计时，若未接收到，统计并记录暂停翻译时长，判断暂停翻译时长是否超过指定时间，比如1小时，若超过指定时间则判断暂停翻译计时后是否接收到用户录入的待翻译的语句，若否则结束翻译计时，并记录结束时间，并计算该时间段的费用并发送至手机翻译APP或翻译机，该时间段的“计费时长=结束时间-开始时间-暂停翻译时长”。
S504：根据计费时长和第一计费方式对应的计费标准，计算半自动翻译模式对应的第一翻译计费；或者根据计费时长和第二计费方式对应的计费标准，计算人工翻译模式对应的第二翻译计费。
本实施例的第一计费方式包括计费种类，比如计时计费或按翻译字数收费，还包括计费标准即计费单价。分别根据“翻译计费=计费时长*计费单价”获得半自动翻译模式对应的第一翻译计费，以及人工翻译模式对应的第二翻译计费。
进一步地，按照预设规则生成对应的翻译计费的步骤S5，包括：
S505：判断选择指令选择的翻译模式是否为计费模式，其中计费模式包括半自动翻译模式和人工翻译模式，以及半自动翻译模式对应第三计费方式和人工翻译模式对应第四计费方式。
本实施例的第三计费方式根据用户选择的人工校正选项的不同而不同，举例地，选择一项人工校正选项费用为100元/千字（计费单价），两项人工校正选项费用为150元/千字（计费单价），即选择半自动翻译模式后，费用以100元/千字（计费单价）起开始计费，增加另一项人工校正选项以150元/千字计费（计费单价）。本实施例的第四计费方式的费用以200元/千字（计费单价）起开始计费。
S506：若翻译模式为计费模式，则判断第三计费方式或第四计费方式是否为翻译字数收费。
本实施例按翻译字数收费为例，具体解释实现过程。上述翻译字数指目标语音信号对应的总字数。
S507：若为翻译字数收费，则统计目标语音信号对应的总字数。
本实施例通过对合成目标语音信号之前的文本执行字数统计，获得目标语音信号对应的总字数。
S508：根据总字数和第三计费方式对应的计费标准，计算半自动翻译模式对应的第三翻译计费；或者根据总字数和第四计费方式对应的计费标准，计算人工翻译模式对应的第四翻译计费。
本实施例的第三计费方式包括计费种类，即按翻译字数收费，还包括计费标准即计费单价。分别根据“翻译计费=总字数*计费单价”获得半自动翻译模式对应的第一翻译计费，以及人工翻译模式对应的第二翻译计费。
本实施例通过在原机器翻译的基础上增加了人工校正选项，以提高翻译的准确度，满足用户对准确翻译的使用需求。本实施例的翻译模式包括免费的自动翻译、收费的半自动翻译模式和人工翻译模式，且收费方式灵活多样，满足不同用户的使用需求。
参照图2，本申请还提供了一种翻译校正装置，集成于管理服务器中，包括：第一接收模块1，用于接收预关联的翻译机发送的翻译模式的选择指令，翻译模式包括自动翻译模式、半自动翻译模式和人工翻译模式中的至少两种。开通模块2，用于根据选择指令选择的翻译模式开通对应的翻译通道。第二接收模块3，用于接收翻译机发送的待翻译的原语音信号。翻译模块4，用于将待翻译的原语音信号通过翻译通道翻译成目标语音信号。
本领域技术人员可以理解的是，本实施例的装置和上述实施例所述的方法相辅相成、互相适应，上述方法项中描述的多个细节和说明均可适用于本实施例的装置，为了避免重复，此处不再赘述。
参照图3，本申请一实施例中，选择指令选择的翻译模式为半自动翻译模式，半自动翻译模式包括人工校准选项，人工校正选项包括识别校正、翻译校正和合成校正中的一项或两项，翻译通道包括与人工校准选项对应的人工校准通道，翻译模块4，包括：
第一发送单元40，用于通过翻译通道将待翻译的原语音信号发送至语音服务器。第一接收单元41，用于接收语音服务器发送的待翻译的原语音信号对应的第一目标语言文本。第一判断单元42，用于判断人工校正选项是否包括识别校正。识别校正单元43，用于若包括识别校正，则将第一目标语言文本和待翻译的原语音信号发送至人工校准通道提供的第一客服终端进行识别校正，得到校正第一目标语言文本。第二接收单元44，用于接收第一客服终端反馈的校正第一目标语言文本。第二发送单元45，用于将校正第一目标语言文本通过翻译通道翻译成目标语音信号。
参照图4，本申请一实施例中，第二发送单元45，包括：
第一发送子单元450，用于通过翻译通道将校正第一目标语言文本发送至翻译服务器。第一接收子单元450a，用于接收翻译服务器发送的校正第一目标语言文本对应的第二目标语言文本。第一判断子单元451，用于判断人工校正选项是否包括翻译校正。第二发送子单元452，用于若包括翻译校正，则将第二目标语言文本和待翻译的原语音信号发送至人工校准通道提供的第二客服终端进行翻译校正。第二接收子单元452a，用于接收第二客服终端反馈的校正第二目标语言文本。第三发送子单元453，用于通过翻译通道将校正第二目标语言文本发送至语音合成服务器进行语音合成。第三接收子单元453a，用于接收语音合成服务器发送的第二目标语言文本对应的目标语音信号。  
参照图5，本申请另一实施例中，第二发送单元45，包括：
第四发送子单元456，用于若不包括翻译校正，则将第二目标语言文本发送至语音合成服务器进行语音合成。第四接收子单元456a，用于接收语音合成服务器发送的第二目标语言文本对应的指定目标语音信号。第二判断子单元457，用于判断人工校正选项是否包括合成校正。第五发送子单元458，用于若包括合成校正，则将指定目标语音信号和待翻译的原语音信号发送至人工校准通道提供的第三客服终端进行合成校正。第五接收子单元459，用于接收第三客服终端反馈的目标语音信号。
参照图6，本申请另一实施例中，选择指令选择的翻译模式为人工翻译模式，人工翻译模式包括人工校准选项，人工校正选项包括识别校正、翻译校正和合成校正，翻译通道包括与人工校准选项对应的人工校准通道，翻译模块4，包括：第三发送单元401，用于通过翻译通道将待翻译的原语音信号发送至语音服务器。第三接收单元402，用于接收语音服务器发送的待翻译的原语音信号对应的第三目标语言文本。第四发送单元403，用于将第三目标语言文本和待翻译的原语音信号发送至人工校准通道提供的指定客服终端进行识别校正。第四接收单元404,用于接收指定客服终端反馈的校正第三目标语言文本。第五发送单元405，用于通过翻译通道将校正第三目标语言文本发送至翻译服务器进行翻译。第五接收单元405a，用于接收翻译服务器发送的校正第三目标语言文本对应的第四目标语言文本。第六发送单元406，用于将第四目标语言文本和待翻译的原语音信号发送至人工校准通道提供的指定客服终端进行翻译校正，得到校正第四目标语言文本。第六接收单元407，用于接收指定客服终端反馈的校正第四目标语言文本。第七发送单元408，用于将校正第四目标语言文本发送至语音合成服务器进行语音合成。第七接收单元408a，用于接收语音合成服务器发送的得到校正第四目标语言文本对应的第一语音信号。第八发送单元409，用于将第一语音信号和待翻译的原语音信号发送至人工校准通道提供的指定客服终端进行合成校正。第八接收单元409a，用于接收指定客服终端反馈的得到目标语音信号。
参照图7，本申请又一实施例的翻译校正装置，包括：生成模块5，用于按照预设规则生成对应的翻译计费。支付模块6：将翻译计费发送至指定终端，以接收用户的支付的翻译计费。
参照图8，本申请一实施例的生成模块5，包括：第二判断单元501，用于判断选择指令选择的翻译模式是否为计费模式，其中计费模式包括半自动翻译模式和人工翻译模式，以及半自动翻译模式对应第一计费方式和人工翻译模式对应第二计费方式。第三判断单元502，用于若翻译模式为计费模式，则判断第一计费方式或第二计费方式是否为计时翻译。第一统计单元503，用于若为计时翻译，则统计用户从开始录入待翻译的原语音信号起，至完成所有待翻译的原语音信号之间的计费时长。第一计算单元504，用于根据计费时长和第一计费方式对应的计费标准，计算半自动翻译模式对应的第一翻译计费；或者根据计费时长和第二计费方式对应的计费标准，计算人工翻译模式对应的第二翻译计费。
参照图9，本申请另一实施例的生成模块5，包括：第四判断单元505，用于判断选择指令选择的翻译模式是否为计费模式，其中计费模式包括半自动翻译模式和人工翻译模式，以及半自动翻译模式对应第三计费方式和人工翻译模式对应第四计费方式。第五判断单元506，用于若翻译模式为计费模式，则判断第三计费方式或第四计费方式是否为翻译字数收费。第二统计单元507，用于若为翻译字数收费，则统计目标语音信号对应的总字数。第二计算单元508，用于根据总字数和第三计费方式对应的计费标准，计算半自动翻译模式对应的第三翻译计费；或者根据总字数和第四计费方式对应的计费标准，计算人工翻译模式对应的第四翻译计费。
本申请还提供了一种翻译校正设备，包括存储器、处理器和应用程序，应用程序被存储在存储器中并被配置为由处理器执行，应用程序被配置为用于执行上述任一实施例中的翻译校正方法。
本领域技术人员可以理解，本发明的计步设备和上述所涉及用于执行本申请中方法中的一项或多项的设备。设备可以为所需的目的而专门设计和制造，或者也可以包括通用计算机中的已知设备。设备具有存储在其内的计算机程序或应用程序，这些计算机程序选择性地激活或重构。这样的计算机程序可以被存储在设备（例如，计算机）可读介质中或者存储在适于存储电子指令并分别耦联到总线的任何类型的介质中，计算机可读介质包括但不限于任何类型的盘（包括软盘、硬盘、光盘、CD-ROM、和磁光盘）、ROM（Read-Only Memory，只读存储器）、RAM（Random Access Memory，随机存储器）、EPROM（Erasable Programmable Read-Only Memory，可擦写可编程只读存储器）、EEPROM（Electrically Erasable Programmable Read-Only Memory，电可擦可编程只读存储器）、闪存、磁性卡片或光线卡片。也就是，可读介质包括由设备（例如，计算机）以能够读的形式存储或传输信息的任何介质。

Claims (17)

  1. 一种翻译校正方法，其特征在于，包括：
    管理服务器接收预关联的翻译机发送的翻译模式的选择指令，所述翻译模式包括自动翻译模式、半自动翻译模式和人工翻译模式中的至少两种；
    根据所述选择指令选择的翻译模式开通对应的翻译通道；
    接收所述翻译机发送的待翻译的原语音信号；
    将所述待翻译的原语音信号通过所述翻译通道翻译成目标语音信号。
  2. 根据权利要求1所述的翻译校正方法，其特征在于，所述选择指令选择的翻译模式为半自动翻译模式，所述半自动翻译模式包括人工校准选项，所述人工校正选项包括识别校正、翻译校正和合成校正中的一项或两项，所述翻译通道包括与所述人工校准选项对应的人工校准通道，所述将所述待翻译的原语音信号通过所述翻译通道翻译成目标语音信号的步骤，包括：
    通过所述翻译通道将所述待翻译的原语音信号发送至语音服务器；
    接收所述语音服务器发送的所述待翻译的原语音信号对应的第一目标语言文本；
    判断所述人工校正选项是否包括识别校正；
    若包括识别校正，则将所述第一目标语言文本和所述待翻译的原语音信号发送至所述人工校准通道提供的第一客服终端进行识别校正；
    接收所述第一客服终端反馈的校正第一目标语言文本；
    将所述校正第一目标语言文本通过所述翻译通道翻译成目标语音信号。
  3. 根据权利要求2所述的翻译校正方法，其特征在于，所述将所述校正第一目标语言文本通过所述翻译通道翻译成目标语音信号的步骤，包括：
    通过所述翻译通道将所述校正第一目标语言文本发送至翻译服务器；
    接收所述翻译服务器发送的所述校正第一目标语言文本对应的第二目标语言文本；
    判断所述人工校正选项是否包括翻译校正；
    若包括翻译校正，则将所述第二目标语言文本和所述待翻译的原语音信号发送至所述人工校准通道提供的第二客服终端进行翻译校正；
    接收所述第二客服终端反馈的校正第二目标语言文本；
    通过所述翻译通道将所述校正第二目标语言文本发送至语音合成服务器进行语音合成；
    接收所述语音合成服务器发送的所述第二目标语言文本对应的目标语音信号。
  4. 根据权利要求3所述的翻译校正方法，其特征在于，所述判断所述人工校正选项是否包括翻译校正的步骤之后，包括：
    若不包括翻译校正，则将所述第二目标语言文本发送至语音合成服务器进行语音合成；
    接收所述语音合成服务器发送的所述第二目标语言文本对应的指定目标语音信号；
    判断所述人工校正选项是否包括合成校正；
    若包括合成校正，则将所述指定目标语音信号和所述待翻译的原语音信号发送至所述人工校准通道提供的第三客服终端进行合成校正；
    接收所述第三客服终端反馈的目标语音信号。
  5. 根据权利要求1所述的翻译校正方法，其特征在于，所述选择指令选择的翻译模式为人工翻译模式，所述人工翻译模式包括人工校准选项，所述人工校正选项包括识别校正、翻译校正和合成校正，所述翻译通道包括与所述人工校准选项对应的人工校准通道，所述将所述待翻译的原语音信号通过所述翻译通道翻译成目标语音信号的步骤，包括：
    通过所述翻译通道将所述待翻译的原语音信号发送至语音服务器；
    接收所述语音服务器发送的所述待翻译的原语音信号对应的第三目标语言文本；
    将所述第三目标语言文本和所述待翻译的原语音信号发送至所述人工校准通道提供的指定客服终端进行识别校正；
    接收所述指定客服终端反馈的校正第三目标语言文本；
    通过所述翻译通道将所述校正第三目标语言文本发送至翻译服务器；
    接收所述翻译服务器发送的所述校正第三目标语言文本对应的第四目标语言文本；
    将所述第四目标语言文本和所述待翻译的原语音信号发送至所述人工校准通道提供的指定客服终端进行翻译校正；
    接收所述指定客服终端反馈的校正第四目标语言文本；
    将所述校正第四目标语言文本发送至语音合成服务器进行语音合成；
    接收所述语音合成服务器发送的所述校正第四目标语言文本对应的第一语音信号；
    将所述第一语音信号和所述待翻译的原语音信号发送至所述人工校准通道提供的指定客服终端进行合成校正；
    接收所述指定客服终端反馈的目标语音信号。
  6. 根据权利要求1所述的翻译校正方法，其特征在于，所述将所述待翻译的原语音信号通过所述翻译通道翻译成目标语音信号的步骤之后，包括：
    按照预设规则生成对应的翻译计费；
    将所述翻译计费发送至指定终端，以接收用户支付的所述翻译计费。
  7. 根据权利要求6所述的翻译校正方法，其特征在于，所述按照预设规则生成对应的翻译计费的步骤，包括：
    判断所述选择指令选择的翻译模式是否为计费模式，其中所述计费模式包括所述半自动翻译模式和所述人工翻译模式，以及所述半自动翻译模式对应第一计费方式和人工翻译模式对应第二计费方式；
    若所述翻译模式为计费模式，则判断所述第一计费方式或所述第二计费方式是否为计时翻译；
    若为计时翻译，则统计用户从开始录入所述待翻译的原语音信号起，至完成所有所述待翻译的原语音信号之间的计费时长；
    根据所述计费时长和所述第一计费方式对应的计费标准，计算所述半自动翻译模式对应的第一翻译计费；或者根据所述计费时长和所述第二计费方式对应的计费标准，计算所述人工翻译模式对应的第二翻译计费。
  8. 根据权利要求6所述的翻译校正方法，其特征在于，所述按照预设规则生成对应的翻译计费的步骤，包括：
    判断所述选择指令选择的翻译模式是否为计费模式，其中所述计费模式包括所述半自动翻译模式和所述人工翻译模式，以及所述半自动翻译模式对应的第三计费方式和人工翻译模式对应的第四计费方式；
    若所述翻译模式为计费模式，则判断所述第三计费方式或所述第四计费方式是否为翻译字数收费；
    若为翻译字数收费，则统计所述目标语音信号对应的总字数；
    根据所述总字数和所述第三计费方式对应的计费标准，计算所述半自动翻译模式对应的第三翻译计费；或者根据所述总字数和所述第四计费方式对应的计费标准，计算所述人工翻译模式对应的第四翻译计费。
  9. 一种翻译校正装置，其特征在于，集成于管理服务器中，包括：
    第一接收模块，用于接收预关联的翻译机发送的翻译模式的选择指令，所述翻译模式包括自动翻译模式、半自动翻译模式和人工翻译模式中的至少两种；
    开通模块，用于根据所述选择指令选择的翻译模式开通对应的翻译通道，其中所述翻译模式包括人工校正选项，所述翻译通道包括与所述人工校正选项对应的人工校正通道；
    第二接收模块，用于接收所述翻译机发送的待翻译的原语音信号；
    翻译模块，用于将所述待翻译的原语音信号通过所述翻译通道翻译成目标语音信号。
  10. 根据权利要求9所述的翻译校正装置，其特征在于，所述选择指令选择的翻译模式为半自动翻译模式，所述半自动翻译模式包括人工校准选项，所述人工校正选项包括识别校正、翻译校正和合成校正中的一项或两项，所述翻译通道包括与所述人工校准选项对应的人工校准通道，所述翻译模块，包括：
    第一发送单元，用于通过所述翻译通道将所述待翻译的原语音信号发送至语音服务器；
    第一接收单元，用于接收所述语音服务器发送的所述待翻译的原语音信号对应的第一目标语言文本；
    第一判断单元，用于判断所述人工校正选项是否包括识别校正；
    识别校正单元，用于若所述人工校正选项包括识别校正，则将所述第一目标语言文本和所述待翻译的原语音信号发送至所述人工校准通道提供的第一客服终端进行识别校正；
    第二接收单元，用于接收所述第一客服终端反馈的校正第一目标语言文本；
    第二发送单元，用于将所述校正第一目标语言文本通过所述翻译通道翻译成目标语音信号。
  11. 根据权利要求10所述的翻译校正装置，其特征在于，所述第二发送单元，包括：
    第一发送子单元，用于通过所述翻译通道将所述校正第一目标语言文本发送至翻译服务器；
    第一接收子单元，用于接收所述翻译服务器发送的所述校正第一目标语言文本对应的第二目标语言文本；
    第一判断子单元，用于判断所述人工校正选项是否包括翻译校正；
    第二发送子单元，用于若包括翻译校正，则将所述第二目标语言文本和所述待翻译的原语音信号发送至所述人工校准通道提供的第二客服终端进行翻译校正；
    第二接收子单元，用于接收所述第二客服终端反馈的校正第二目标语言文本；
    第三发送子单元，用于通过所述翻译通道将所述校正第二目标语言文本发送至语音合成服务器进行语音合成；
    第三接收子单元，用于接收所述语音合成服务器发送的所述第二目标语言文本对应的目标语音信号。
  12. 根据权利要求11所述的翻译校正装置，其特征在于，所述第二发送单元，包括：
    第四发送子单元，用于若不包括翻译校正，则将所述第二目标语言文本发送至语音合成服务器进行语音合成；
    第四接收子单元，用于接收所述语音合成服务器发送的所述第二目标语言文本对应的指定目标语音信号；
    第二判断子单元，用于判断所述人工校正选项是否包括合成校正；
    第五发送子单元，用于若包括合成校正，则将所述指定目标语音信号和所述待翻译的原语音信号发送至所述人工校准通道提供的第三客服终端进行合成校正；
    第五接收子单元，用于接收所述第三客服终端反馈的目标语音信号。
  13. 根据权利要求9所述的翻译校正装置，其特征在于，所述选择指令选择的翻译模式为人工翻译模式，所述人工翻译模式包括人工校准选项，所述人工校正选项包括识别校正、翻译校正和合成校正，所述翻译通道包括与所述人工校准选项对应的人工校准通道，所述翻译模块，包括：
    第三发送单元，用于通过所述翻译通道将所述待翻译的原语音信号发送至语音服务器；
    第三接收单元，用于接收所述语音服务器发送的所述待翻译的原语音信号对应的第三目标语言文本；
    第四发送单元，用于将所述第三目标语言文本和所述待翻译的原语音信号发送至所述人工校准通道提供的指定客服终端进行识别校正；
    第四接收单元，用于接收所述指定客服终端反馈的校正第三目标语言文本；
    第五发送单元，用于通过所述翻译通道将所述校正第三目标语言文本发送至翻译服务器；
    第五接收单元，用于接收所述翻译服务器发送的所述校正第三目标语言文本对应的第四目标语言文本；
    第六发送单元，用于将所述第四目标语言文本和所述待翻译的原语音信号发送至所述人工校准通道提供的指定客服终端进行翻译校正；
    第六接收单元，用于接收所述指定客服终端反馈的校正第四目标语言文本；
    第七发送单元，用于将所述校正第四目标语言文本发送至语音合成服务器进行语音合成；
    第七接收单元，用于接收所述语音合成服务器发送的所述校正第四目标语言文本对应的第一语音信号；
    第八发送单元，用于将所述第一语音信号和所述待翻译的原语音信号发送至所述人工校准通道提供的指定客服终端进行合成校正；
    第八接收单元，用于接收所述指定客服终端反馈的目标语音信号。
  14. 根据权利要求9所述的翻译校正装置，其特征在于，包括：
    生成模块，用于按照预设规则生成对应的翻译计费；
    支付模块，用于将所述翻译计费发送至指定终端，以接收用户支付的所述翻译计费。
  15. 根据权利要求14所述的翻译校正装置，其特征在于，所述生成模块，包括：
    第二判断单元，用于判断所述选择指令选择的翻译模式是否为计费模式，其中所述计费模式包括所述半自动翻译模式和所述人工翻译模式，以及所述半自动翻译模式对应第一计费方式和人工翻译模式对应第二计费方式；
    第三判断单元，用于若所述翻译模式为计费模式，则判断所述第一计费方式或所述第二计费方式是否为计时翻译；
    第一统计单元，用于若为计时翻译，则统计用户从开始录入所述待翻译的原语音信号起，至完成所有所述待翻译的原语音信号之间的计费时长；
    第一计算单元，用于根据所述计费时长和所述第一计费方式对应的计费标准，计算所述半自动翻译模式对应的第一翻译计费；或者根据所述计费时长和所述第二计费方式对应的计费标准，计算所述人工翻译模式对应的第二翻译计费。
  16. 根据权利要求14所述的翻译校正装置，其特征在于，所述生成模块，包括：
    第四判断单元，用于判断所述选择指令选择的翻译模式是否为计费模式，其中所述计费模式包括所述半自动翻译模式和所述人工翻译模式，以及所述半自动翻译模式对应的第三计费方式和人工翻译模式对应的第四计费方式；
    第五判断单元，用于若所述翻译模式为计费模式，则判断所述第三计费方式或所述第四计费方式是否为翻译字数收费；
    第二统计单元，用于若为翻译字数收费，则统计所述目标语音信号对应的总字数；
    第二计算单元，用于根据所述总字数和所述第三计费方式对应的计费标准，计算所述半自动翻译模式对应的第三翻译计费；或者根据所述总字数和所述第四计费方式对应的计费标准，计算所述人工翻译模式对应的第四翻译计费。
  17. 一种翻译校正设备，包括存储器、处理器和应用程序，所述应用程序被存储在所述存储器中并被配置为由所述处理器执行，其特征在于，所述应用程序被配置为用于执行权利要求1至8任一项所述的翻译校正方法。
PCT/CN2019/073762 2018-09-11 2019-01-29 翻译校正方法、装置以及翻译校正设备 Ceased WO2020052191A1 (zh)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN201811058673.7A CN109446532A (zh) 2018-09-11 2018-09-11 翻译校正方法、装置以及翻译校正设备
CN201811058673.7 2018-09-11

Publications (1)

Publication Number Publication Date
WO2020052191A1 true WO2020052191A1 (zh) 2020-03-19

Family

ID=65532666

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2019/073762 Ceased WO2020052191A1 (zh) 2018-09-11 2019-01-29 翻译校正方法、装置以及翻译校正设备

Country Status (2)

Country Link
CN (1) CN109446532A (zh)
WO (1) WO2020052191A1 (zh)

Families Citing this family (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN113591491B (zh) * 2020-04-30 2023-12-26 阿里巴巴集团控股有限公司 语音翻译文本校正系统、方法、装置及设备
CN112818704B (zh) * 2021-01-19 2024-04-02 传神语联网网络科技股份有限公司 基于线程间共识反馈的多语种翻译系统与方法
CN116110381A (zh) * 2023-01-09 2023-05-12 中科凡语(武汉)科技有限公司 一种翻译机中语音识别自动纠正系统及方法

Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
GB2323693A (en) * 1997-03-27 1998-09-30 Forum Technology Limited Speech to text conversion
CN1573741A (zh) * 2003-06-20 2005-02-02 微软公司 自适应性机器翻译
CN101309449A (zh) * 2008-07-04 2008-11-19 沈阳格微软件有限责任公司 一种基于手机彩信/短信的远程翻译服务方法
CN105488036A (zh) * 2015-11-23 2016-04-13 百度在线网络技术（北京）有限公司 基于人工智能机器人的翻译提供方法及装置
CN108319592A (zh) * 2018-02-08 2018-07-24 广东小天才科技有限公司 一种翻译的方法、装置及智能终端

Patent Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
GB2323693A (en) * 1997-03-27 1998-09-30 Forum Technology Limited Speech to text conversion
CN1573741A (zh) * 2003-06-20 2005-02-02 微软公司 自适应性机器翻译
CN101309449A (zh) * 2008-07-04 2008-11-19 沈阳格微软件有限责任公司 一种基于手机彩信/短信的远程翻译服务方法
CN105488036A (zh) * 2015-11-23 2016-04-13 百度在线网络技术（北京）有限公司 基于人工智能机器人的翻译提供方法及装置
CN108319592A (zh) * 2018-02-08 2018-07-24 广东小天才科技有限公司 一种翻译的方法、装置及智能终端

Also Published As

Publication number Publication date
CN109446532A (zh) 2019-03-08

Similar Documents

Publication Publication Date Title
US20120330643A1 (en) System and method for translation
JP7108799B2 (ja) ヒューマンマシン対話方法及び電子デバイス
CN109981910A (zh) 业务推荐方法及设备
CA2681473A1 (en) Flexible communication systems and methods
TW202006532A (zh) 播報語音的確定方法、裝置和設備
US10698934B1 (en) Interface between a virtual personal assistant and an interactive voice response system
CN101699840B (zh) 融合通信中智能语音交互系统及其实现方法
US20230282204A1 (en) Text Time Annotation Method and Apparatus, Electronic Device, and Readable Storage Medium
CN103026697B (zh) 业务服务器装置和业务提供方法
WO2019148564A1 (zh) 实现语音翻译的方法、装置和语音翻译设备
CN116319955B (zh) 语音播报方法、装置、存储介质和计算机设备
CN109582760A (zh) 一种应用在智能对话系统中的应答流程引擎系统
US10255914B2 (en) Digital concierge and method
CN107728497B (zh) 用于人机交互的通信方法
CN109202908A (zh) 机器人的控制方法、装置、设备、系统和存储介质
CN109446532A (zh) 翻译校正方法、装置以及翻译校正设备
CN110083694A (zh) 一种基于webhook的电网业务集成方法、装置及存储器
CN109726309A (zh) 音频生成方法、装置及存储介质
WO2024131149A1 (zh) 一种语速调节方法、系统及装置
CN108920377A (zh) 一种日志回放测试方法、系统、装置及可读存储介质
WO2020042878A1 (zh) 智能终端的日历事件提醒交互方法、系统及智能终端
CN103229494B (zh) 业务服务器装置及业务提供方法
CN103544619A (zh) 一种人工视频面对面及时翻译系统及其方法
CN104700830A (zh) 一种语音端点检测方法及装置
CN118410890A (zh) 会议预约方法、装置、存储介质及电子设备

Legal Events

Date Code Title Description
NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 19860903

Country of ref document: EP

Kind code of ref document: A1