JP2017191959A - Multilanguage voice translation system for tv conference system - Google Patents

Abstract

PROBLEM TO BE SOLVED: To translate languages so that communication may be performed in a TV conference using multiple languages.SOLUTION: This TV conference system identifies languages to be used in a conference for each base on the basis of attribute information of attendants, translates a language to the languages set to respective bases by translation apparatus connected to the system, and outputs translated contents to the respective bases as character information or voice information. This invention also includes the following mode of displaying translated contents to one's own base and receiving correction from an attendant. Further, display to the respective bases and correction in parallel is also one mode of the invention.SELECTED DRAWING: Figure 1

Description

  The present invention relates to speech translation technology, and more particularly to translation technology in a TV conference system.

  Currently, there is a conference using a TV conference system. In a video conference, it is possible to omit movement between remote locations and it is effectively used. For example, Patent Document 1 demonstrates the effect of cost reduction and time saving, particularly when attendees of a conference cross a border.

  For example, Patent Document 1 discloses a technique for enabling a connection between multi-site PCs to be established easily and quickly even in a TV conference system using a network line network even if the number of local PCs attending is large. Yes.

JP 2008-187659 Gazette

  Here, paying attention to the fact that a conference can be held between remote locations, which is a characteristic of the TV conference system, a TV conference across borders is assumed. When crossing national borders, each country has a different language. Even if the site does not cross the border, multilinguals may be used by attendees.

  Therefore, in the present invention, based on the attendee attribute information, the language used in the corresponding conference is specified for each site, and is translated into the language set for each site by a translation device connected to the TV conference system. The translated content is output as text information or voice information at each site. The present invention further includes the following aspects. This includes displaying the contents of translation at your site and accepting corrections from participants. Furthermore, it is also one aspect of the invention to display and correct each site in parallel.

  With the above-described configuration of the present invention, a multilingual TV conference can be implemented.

It is a figure which shows the structure and basic concept of one Embodiment of this invention. It is a flowchart for every base of one Embodiment of this invention. It is a system flowchart of one Embodiment of this invention. It is a figure which shows the system configuration | structure figure and database of one Embodiment of this invention. It is a figure which shows the system configuration | structure figure and database transition of one Embodiment of this invention. It is a figure which shows the data transition in S9, S12, and S13 of one Embodiment of this invention.

  Hereinafter, an embodiment of the present invention will be described with reference to the drawings. First, FIG. 1 is a diagram showing a concept of the TV conference system in the present embodiment. In the present embodiment, a function for automatically translating speech into text is installed in the TV conference system. As a result, local language subtitles corresponding to the participants are displayed on the screen of the video conference system, and the use of the video conference system with foreign countries can be more easily promoted. In addition, the text of the voice that you have sent is also reflected on your screen, so you can easily correct the text and change it to other locations. More specifically, personal information is transmitted from each base to the registration management server 4 and registered in the registration management DB 5 in advance. Based on this, the translation language is specified, the character translation server 6 executes the translation, and is displayed on the TV screens 2 and 3 at each site. Hereinafter, the processing contents will be described with reference to FIG.

  FIG. 4 is a system configuration diagram of the present embodiment. Each device is connected via the network 1. The own site TV 2 and the other site TV 3 are TV conference monitors installed in each conference room. Here, the expression is separated from the own base and other bases, but they have the same function. Further, at least one registration management server 4 and registration management DB 5 may be prepared in the video conference system. The registration management server 4 has a registration management DB 5 that stores personal voice and language information for conducting a video conference. As a result, the language used at each site can be specified. Further, the character translation server 6 has a function of converting speech into characters and a function of translating characters into other language characters. The character translation DB 7 stores the characterized result in association with the speech information before conversion. The registration management server 4 and the character translation server 6 described above are realized by so-called computers, and the processing of the flowchart described below is executed by an arithmetic device such as a CPU according to the program that the device itself has.

  Next, the user procedure executed before the meeting will be described with reference to FIG. The processing of FIG. 2 is executed by the registration management server 4 unless otherwise specified. First, in S1, the LDAP of the user who will attend the meeting is registered. In the actual registration, a personal identification number such as LDAP is received via a PC (not shown) possessed by each participant and entered in the item “LDAP” of the registration management DB 5. At this time, the intention of participating in the conference is accepted from the prospective participants, and the result is recorded in the conference participation flag. In addition, with respect to the use base item of the registration management DB 5, when LDAP is received, information for identifying the use base is received and registered.

  Next, in S2, it is determined whether or not language and voice information registration is necessary. In this process, the registration management DB 5 is searched using the LDAP entered in S1 as a key to determine whether it has been registered. This determination may be made based on the presence or absence of LDAP or the presence or absence of information corresponding to the searched LDAP. Here, when judging based on the presence or absence of LDAP, it is judged at the time of registration of S1. As a result of the search, if it has never been registered, the process proceeds to S3.

  In S3 and S4, the information specifying the language accepted in S2 and the voice information are registered in the use language item and the voice code item of the registration management DB5. At this time, the language used, which is information for specifying the language, may be registered by analyzing the voice input from the participant. In S3 and S4, as in S1, information specifying the language used by each participant is received from each participant via the PC. Also, voice input is accepted from the participant via a microphone or the like.

  If it is determined that it has been registered in S2, or if the processing up to S4 is completed, the process proceeds to S5. In S5, it is determined whether there is a person who uses a language other than the registered language among the conference participants. This determination is based on whether a plurality of languages are used at the site, and is determined by a person who plans to participate in the site. If you want to use another language, return to S1 and register for the prospective participants who will use the other language. If there is no person who uses another language, the process proceeds to S6 to register the user information list of the conference. In this registration process, the registration management server 4 transmits the LDAP, use language, use base, and voice code whose conference participation flag is ◯ to the character translation server 6, and the character translation server 6 registers it in the character translation DB 7. The above processing is executed for each base.

  Next, the processing contents during the meeting will be described with reference to FIG. At the time of starting the video conference, as described above, the information on the person whose conference participation flag is marked from the registration management DB 5 is registered in the character translation DB 7 in advance. The contents are as shown in the registration management DB 5 and the character translation DB 7 shown in FIG. That is, at the time of starting the video conference, the voice information, character data, character code, and own site flag of the character translation DB 7 are blank. In this state, when the conference starts, the participant makes a remark and performs the process of S7. In S7, the character translation server 6 receives the voice information of the speaker through its own TV, recognizes the caller from the voice information, and creates character data and character codes. In this example, when the LDAP 111 in the site A says “Thank you”, the voice information is converted from the usage site and the voice code to create the following character data and character code. When character conversion is performed from voice information, a local base flag is set. These pieces of information are stored in the character translation DB7. The stored state is indicated by S7 in FIG.

  In S8, the character data, character code, and own site flag created in S7 are returned to the own site, and the caption is displayed on the upper part of the own site TV screen. At this time, it is determined from the own site flag that the use site A is the own site, and character data is displayed as subtitles at the top of the screen.

  In S9, character data is translated into another language and converted into data. In this example, the character data “Thank you” is translated into character data in accordance with the language used at another base. The translated data is stored in the character translation DB 7 as shown in the table of S9 in FIG. For this storage, the character code created in S7 is also registered in each line. After that, the character data “Thank you”, “Takei”, character code, and own site flag are returned to the sites B and C.

  In S10, other sites are displayed relative to the display in S8. In this display, it is determined that the use base B and the use base C are other bases from the own base flag, and character data is displayed as subtitles at the bottom of the screen. In this embodiment, subtitles at your site are displayed at the top of the screen, and subtitles at other sites are displayed at the bottom of the screen, but it is only necessary to be able to display them in different positional relationships. Other embodiments are also included.

  In S11, an input as to whether the subtitles issued by the user's own site are correctly displayed is accepted. As a result, if an input indicating that there is an error is received, the process proceeds to S12. In addition, it can correct efficiently by performing the flow of S9-S10 and S11-S12 in parallel.

  In S12, the selection of the correction part and the correction contents of the character data from the participant are received, and the correction is performed accordingly. As a result, the character translation DB 7 is corrected to the display content of S12 in FIG. In this example, corrections were made from the TV screen at site A, and the content of the character data was changed from "Thank you" to "I'm sorry".

  In S13, the part corrected in S12 is re-translated and displayed on other sites. As a result, the character translation DB 7 is corrected to the display content of S13 in FIG. In this example, the character data “I'm sorry” is re-translated into the character data shown in the table below according to the language used at the other site. After that, each character data “Sorry”, “vs. no occurrence”, character code, and own site flag are returned to site B and site C. Each base recognizes that it is re-translation because character data with the same character code already exists, and corrects it. The corrected part may be displayed with a color or the like. The above is the content of the processing in this embodiment. S11 and S12 may be performed in parallel with S9 and S10.

DESCRIPTION OF SYMBOLS 1 ... Network, 2 ... Own site TV screen, 3 ... Other site TV screen, 4 ... Registration management server, 5 ... Registration management DB, 6 ... Character translation server, 7 ... Character translation DB

Claims (3)

In a multi-lingual speech translation system in a TV conference system for translating the speech of each speaker in a TV conference system connected via a network,
Based on the attendee attribute information, a means for identifying the language used in the meeting for each location,
A multilingual speech translation system in a TV conference system, comprising: means for translating into a language set in each site and outputting the translated content on the TV conference screen of each site. In the multilingual speech translation system in the video conference system according to claim 1,
A multilingual speech translation system in a TV conference system, further comprising means for displaying a translation content at a local site of the TV conference system and receiving a correction from a participant. The multilingual speech translation system in the video conference system according to claim 2,
The multilingual speech translation system in a TV conference system, wherein the receiving means performs display on the other site and correction reception in parallel.

Images