JP2023049059A - WEB CONFERENCE SERVER, WEB CONFERENCE METHOD AND WEB CONFERENCE SYSTEM - Google Patents

Abstract

Kind Code: A1 To improve the speed and accuracy of web conferencing. A web conference server includes a speech recognition server speed determination unit that determines a communication speed between the web conference server and a plurality of speech recognition servers, an information processing speed of the plurality of speech recognition servers, a language attribute, and a plurality of speech recognition servers. Speech recognition server determination that determines a specific speech recognition server to be used for a specific user from a plurality of speech recognition servers based on the communication speed with the speech recognition server and the information processing speed of the plurality of speech recognition servers a translation server speed determination unit that determines a communication speed between a specific speech recognition server used for a specific user and a plurality of translation servers and an information processing speed of the plurality of translation servers; and a plurality of translation servers. a translation server determination unit that determines a specific translation server to be used for a specific user from among a plurality of translation servers based on a communication speed with and an information processing speed of the plurality of translation servers. [Selection drawing] Fig. 2

Description

The present invention relates to a web conference server, web conference method, and web conference system that provide web conference services.

2. Description of the Related Art A web conference system is known in which a plurality of users in remote areas hold online conferences using individual user terminals connected via a network. In recent years, in addition to voice communication while viewing video images, there are web conference systems that simultaneously display text obtained by speech recognition of voice data on the screen. As web conferencing between remote regions becomes more and more common due to the COVID-19 epidemic, there is a demand for a technology that translates speech recognition data into the language used by the user who is the conference partner and simultaneously displays the translated text on the screen.

JP 2019-153099 A JP 2019-061594 A JP 2017-215931 A Japanese Patent No. 6795668

According to Patent Document 1, when a TV conference system is used, a device (site) for text conversion processing is determined based on the communication speed and processing capability of the device. In Patent Document 2, a user arbitrarily designates a translation destination language, and the language is automatically determined according to the attributes of the attendees of the target meeting and the user who has the viewing authority of the minutes. In Patent Document 3, an input language and a translation language are set based on a user's operation. According to Japanese Patent Laid-Open No. 2002-200000, conference participants and the like may set a priority for each translation dictionary when selecting a plurality of translation dictionaries, and translation processing is executed in the order of the set priority.

In view of the circumstances as described above, an object of the present invention is to improve the speed and accuracy of a web conference.

A web conference server according to one aspect of the present invention includes:
a voice acquisition unit that acquires voice data of a specific user included in a plurality of users participating in a web conference from a user terminal of the specific user;
a voice recognition unit that recognizes the voice data and determines the language attribute of the specific user;
a speech recognition server speed determination unit that determines a communication speed between the web conference server and the plurality of speech recognition servers and an information processing speed of the plurality of speech recognition servers;
Based on the language attribute, the communication speed with the plurality of speech recognition servers, and the information processing speed of the plurality of speech recognition servers, a specific a voice recognition server determination unit that determines a voice recognition server;
a translation server speed determination unit that determines a communication speed between the specific speech recognition server used for the specific user and a plurality of translation servers and an information processing speed of the plurality of translation servers;
Translation server determination for determining a specific translation server to be used for the specific user from among the plurality of translation servers based on a communication speed with the plurality of translation servers and an information processing speed of the plurality of translation servers. Department and
The speech data of the specific user is supplied to the specific speech recognition server by supplying the speech data of the specific user and identification information for identifying the specific translation server to the specific speech recognition server. a speech data processing requesting unit for generating speech recognition data, which is text data, by recognizing the voice of the user, and causing the specific translation server to translate the speech recognition data of the specific user to generate translation data, which is text data;
and
A combination of the specific speech recognition server and the specific translation server differs for each of the multiple users participating in the web conference.

According to the present invention, it is possible to improve the speed and accuracy of a web conference.

1 illustrates a web conferencing system according to one embodiment of the present invention; 1 shows a functional configuration of a web conference system; 1 shows a first operational flow of a web conference server; 2 shows a second operational flow of the web conferencing server; 3 shows a third operational flow of the web conference server;

Hereinafter, embodiments of the present invention will be described with reference to the drawings.

1. Web conferencing system overview

FIG. 1 shows a web conference system according to one embodiment of the invention.

The web conference system 1 has multiple web conference servers 10 , multiple speech recognition servers 20 , and multiple translation servers 30 . A plurality of web conference servers 10, a plurality of speech recognition servers 20, and a plurality of translation servers 30 are interconnected via a network N such as the Internet.

A plurality of web conference servers 10 are installed in a plurality of different countries and regions. A plurality of web conference servers 10 communicate with a plurality of user terminals 40 (personal computers, smart phones, tablet computers, wearable devices, etc.) used by a plurality of users participating in the web conference via a network N, A computer that provides web conferencing services. The web conference server 10 accessed by a plurality of user terminals 40 is different for each user terminal 40 even for a plurality of users participating in the same web conference. Each web conference server 10 determines, for each user, a specific speech recognition server 20 and a specific translation server 30 to be used for that user. For this reason, the combination of the specific speech recognition server 20 and the specific translation server 30 differs for each of the multiple users participating in the web conference, even for multiple users participating in the same web conference. Each web conference server 10 acquires voice data of a specific user participating in the web conference from the user terminal 40 used by the user, and supplies the voice data to one of the voice recognition servers 20 .

A plurality of speech recognition servers 20 are installed in a plurality of different countries and regions. The multiple speech recognition servers 20 are typically computers provided by different providers and running different speech recognition software. Each speech recognition server 20 acquires speech data of a specific user participating in the web conference from any of the web conference servers 10, recognizes the speech data to generate speech recognition data which is text data, The recognition data is supplied to the web conferencing server 10 and the specific translation server 30 from which the voice data is supplied.

A plurality of translation servers 30 are installed in a plurality of different countries and regions. The multiple translation servers 30 are typically computers provided by different providers and running different translation software. Each translation server 30 acquires speech recognition data of a specific user participating in the web conference from any of the speech recognition servers 20, translates the speech recognition data to generate translation data which is text data, and translates the translation data. is supplied to the web conference server 10 that supplies the audio data.

The web conference server 10 transmits speech recognition data and translation data obtained from voice data of multiple users participating in the web conference in real time during the web conference to multiple users participating in the web conference. Output to terminal 40 . In addition, after the web conference ends, the web conference server 10 creates minutes data of the speech recognition data and the translation data, and outputs the minutes data to the plurality of user terminals 40 of the users participating in the web conference.

2. Functional configuration of web conferencing system

FIG. 2 shows the functional configuration of the web conference system.

The web conference server 10 loads the information processing program recorded in the ROM into the RAM by the CPU and executes it. 104, translation server speed determination unit 105, content determination unit 106, emotion determination unit 107, translation server determination unit 108, voice data processing request unit 109, processing data acquisition unit 110, context check unit 111, real-time output unit 112, and meeting minutes It operates as the creation unit 113 . Web conferencing server 10 has non-volatile or volatile storage 120 .

The user terminal 40 operates as a voice input unit 401 , a real-time input unit 402 and a minutes acquisition unit 403 by having the CPU load an information processing program recorded in the ROM into the RAM and execute it. The user terminal 40 has an external or built-in microphone 411 , an external or built-in display 412 , and a nonvolatile or volatile storage device 413 .

3. Web conferencing system operation flow

FIG. 3 shows the first operation flow of the web conference server (from start to end of web conference).

First, at the start of the web conference, each of the user terminals 40 of the users participating in the web conference receives a specific (one ), accesses the selected web conferencing server 10, and requests sign-in to the web conference. Typically, each user terminal 40 selects the web conference server 10 having an IP address that identifies the country or region closest to the country or region identified by the IP address of the user terminal 40 . Therefore, the specific web conference server 10 accessed by each user terminal 40 differs for each user terminal 40 of a plurality of users participating in the web conference. As a result, the optimum web conference server 10 in terms of speed can be used for each of the plurality of user terminals 40 of the plurality of users, and processing can be distributed in the web conference as a whole, enabling stable communication and information processing.

The web conference server 10 accepts access and web conference sign-in requests from user terminals 40 used by specific users included in a plurality of users participating in the web conference, and permits sign-in to the web conference. Hereinafter, unless otherwise specified, one web conference server 10 and one user terminal 40 with which communication has been established, and a user using this user terminal 40 will be described. In addition, the uniquely determined web conference server 10, speech recognition server 20, translation server 30 and user terminal 40 are defined as a specific web conference server 10, a specific speech recognition server 20, a specific translation server 30 and a specific user terminal 40. It is sometimes called

When the web conference starts, the voice input unit 401 of the user terminal 40 starts supplying user voice data input by the user via the microphone 411 to one web conference server 10 with which communication has been established.

The voice acquisition unit 101 of the web conference server 10 acquires user voice data from the user terminal 40 (step S101).

The speech recognition unit 102 of the web conference server 10 recognizes the speech data acquired from the user terminal 40 and determines the language attribute of the specific user (step S102). Language attributes include, for example, language (eg, English), dialect (eg, Australian English), accent (eg, Dutch accent), and the like.

The speech recognition server speed determination unit 103 of the web conference server 10 determines the communication speed between the web conference server 10 and the plurality of speech recognition servers 20 and the information processing speed of the plurality of speech recognition servers 20 (step S103). The communication speed and information processing speed each include not only a reference value but also a real-time speed.

The speech recognition server determination unit 104 of the web conference server 10 determines the language attribute determined by the speech recognition unit 102, the communication speed between the web conference server 10 and the plurality of voice recognition servers 20 determined by the speech recognition server speed determination unit 103, and A specific speech recognition server 20 to be used for a specific user is determined from among the plurality of speech recognition servers 20 based on the information processing speed of the plurality of speech recognition servers 20 (step S104).

As an example, the speech recognition server determination unit 104 predefines one speech recognition server 20 that has high language recognition accuracy for a specific language attribute and is close to the web conference server 10 (resulting in high communication speed). are doing. The speech recognition server determination unit 104 may select this speech recognition server 20 if the information processing speed of this speech recognition server 20 is equal to or higher than a predetermined threshold. As another example, the speech recognition server determination unit 104 predefines a plurality of candidates for the speech recognition server 20 with high language recognition accuracy for a specific language attribute. The speech recognition server determination unit 104 selects one speech recognition server 20 having the fastest communication speed with a plurality of speech recognition servers 20 from the candidates, and the selected speech recognition server 20 has an information processing speed equal to or higher than a predetermined threshold. If so, this voice recognition server 20 should be selected.

The translation server speed determination unit 105 of the web conference server 10 determines the communication speed between the specific speech recognition server 20 used for the specific user and the plurality of translation servers 30 determined by the speech recognition server determination unit 104, and the information processing speed of the translation server 30 (step S105). The communication speed and information processing speed each include not only a reference value but also a real-time speed.

The content determination unit 106 of the web conference server 10 determines the content of the web conference based on the result of voice recognition by the voice recognition unit 102 (step S106). The content determination unit 106 may determine the content of the web conference by, for example, inputting the results of voice recognition performed by the voice recognition unit 102 into an AI model created in advance. The content of the web conference may be, for example, the overall theme of the web conference, such as sports or chemistry.

The emotion determination unit 107 of the web conference server 10 determines the user's emotion (irritated, calm, etc.) based on the voice recognition result of the voice recognition unit 102 (step S107). The emotion determination unit 107 may determine the user's emotion by, for example, inputting the result of voice recognition performed by the voice recognition unit 102 into an AI model created in advance.

The translation server determination unit 108 of the web conference server 10 determines the communication speed between the specific speech recognition server 20 and the plurality of translation servers 30 determined by the translation server speed determination unit 105 and the information processing speed of the plurality of translation servers 30. Based on this, a specific translation server 30 to be used for a specific user is determined from a plurality of translation servers 30 . The translation server determination unit 108 selects a specific translation server 30 to be used for a specific user further based on the content of the web conference determined by the content determination unit 106 and/or the specific user's emotion determined by the emotion determination unit 107. may be determined (step S108).

As an example, the translation server determination unit 108 selects one translation server 30 that has high translation accuracy for specific content and/or specific emotions and is close to the web conference server 10 (resulting in high communication speed). stipulated in advance. The translation server determination unit 108 may select this translation server 30 if the information processing speed of this translation server 30 is equal to or higher than a predetermined threshold. As another example, the translation server determination unit 108 predefines a plurality of translation server 30 candidates that have high accuracy in translating specific content and/or specific emotions. The translation server determination unit 108 selects one translation server 30 having the fastest communication speed with a plurality of translation servers 30 from the candidates, and if the information processing speed of the selected translation server 30 is equal to or higher than a predetermined threshold, This translation server 30 can be selected.

The voice data processing request unit 109 of the web conference server 10 sends the user's voice data acquired by the voice acquisition unit 101 and the translation determined by the translation server determination unit 108 to the voice recognition server 20 determined by the voice recognition server determination unit 104 . It supplies identification information for identifying the server 30 and requests processing (step S109). A combination of a specific speech recognition server 20 and a specific translation server 30 differs for each of multiple users participating in the web conference. As a result, the optimal speech recognition server 20 and translation server 30 can be used for each of the plurality of user terminals 40 of the plurality of users in terms of both speed and accuracy. communication and information processing.

The speech recognition server 20 receives from the speech data processing request unit 109 of the web conference server 10 the speech data of a specific user and the identification information (IP address, etc.) identifying the translation server 30 determined by the translation server determination unit 108. get. The speech recognition server 20 performs speech recognition to generate speech recognition data, which is text data, and supplies the speech recognition data to the web conference server 10 that supplies the speech data and to the translation server 30 identified by the identification information. .

The translation server 30 acquires, from the speech recognition server 20, speech recognition data and identification information (such as an IP address) that identifies the web conference server 10 that supplies the speech data. The translation server 30 translates the speech recognition data to generate translation data, which is text data, and supplies the translation data to the web conference server 10 identified by the identification information.

The processed data acquisition unit 110 of the web conference server 10 acquires the specific user's speech recognition data from the speech recognition server 20, and acquires the specific user's translation data from the specific translation server 30 (step S110).

The context check unit 111 of the web conference server 10 synchronizes the corresponding speech recognition data and translation data, checks the context of the corresponding speech recognition data and translation data, and checks the corresponding speech recognition data and/or translation data. Correction is made according to the result (step S111). The context check unit 111 may check the context of the speech recognition data and the translation data by, for example, inputting the speech recognition data and the translation data into an AI model created in advance. The context check unit 111 stores the corresponding speech recognition data and translation data corrected according to the check result in the storage device 120 .

The storage device 120 may store the speech recognition data and the translation data at least during and after the web conference, and may be either non-volatile or volatile storage device. The data after correction according to the check result includes data that does not need correction based on the check result and is not corrected.

The real-time output unit 112 of the web conference server 10 outputs the corresponding speech recognition data and translation data corrected according to the check results in real time during the web conference to a plurality of users participating in the web conference. Output to the user terminal 40 (step S112). That is, the real-time output unit 112 outputs not only the user terminal 40 that has input the speech data that is the source of the speech recognition data and the translation data, but also a plurality of user terminals 40 of all users participating in the web conference.

The real-time input units 402 of the plurality of user terminals 40 of all users participating in the web conference are sent from the real-time output unit 112 of the web conference server 10 in real time during the web conference, after correction according to the check results. Acquire corresponding speech recognition data and translation data. The real-time input unit 402 displays speech recognition data and translation data, which are text data, on the display 412 in real time during the web conference.

After the web conference ends, the minutes creation unit 113 of the web conference server 10 reads out the corresponding speech recognition data and translation data after being corrected according to the check result from the storage device 120, and uses the read speech recognition data and translation data as Based on this, create the minutes data of the web conference. Specifically, the minutes creation unit 113 of any one of the web conference servers 10 extracts all user's speech recognition data and translation data are acquired, and the acquired speech recognition data and translation data are arranged in chronological order to create meeting minutes data. The minutes creation unit 113 outputs the created minutes data to a plurality of user terminals 40 of a plurality of users participating in the web conference.

The minutes acquisition unit 403 of each user terminal 40 acquires minutes data from the minutes creation unit 113 of the web conference server 10 and stores it in the storage device 413 . The storage device 413 may store the speech recognition data and the translation data for at least a predetermined period after the end of the web conference, and may be either non-volatile or volatile storage device.

FIG. 4 shows the second operational flow of the web conference server (during the web conference).

During the web conference, the speech recognition server speed determination unit 103 of the web conference server 10 periodically (Loop) determines the communication speed between the web conference server 10 and a specific (in-communication) speech recognition server 20, The information processing speed of a specific (in communication) speech recognition server 20 is determined (step S201). The communication speed and information processing speed are respectively real-time speeds. The speech recognition server speed determination unit 103 detects that the communication speed between the web conference server 10 and the specific speech recognition server 20 and/or the information processing speed of the specific speech recognition server 20 has changed below a threshold value during the web conference. (step S202). The threshold is, for example, a speed-unacceptable value for conducting a smooth web conference.

If the communication speed between the web conference server 10 and the specific speech recognition server 20 and/or the information processing speed of the specific speech recognition server 20 changes below the threshold during the web conference (step S202, YES), the speech The recognition server speed determination unit 103 determines the communication speed between the web conference server 10 and the plurality of speech recognition servers 20 (the plurality of speech recognition servers 20 other than the speech recognition server 20 currently in communication) and the information of the plurality of speech recognition servers 20. A processing speed is determined (step S203). The communication speed and information processing speed are respectively real-time speeds.

The speech recognition server determination unit 104 of the web conference server 10 determines the language attribute determined by the voice recognition unit 102 (step S102), the web conference server 10 determined by the voice recognition server speed determination unit 103, and the plurality of voice recognition servers 20. and the information processing speed of the plurality of speech recognition servers 20 (step S203), a specific speech recognition server 20 to be used for a specific user is newly determined from among the plurality of speech recognition servers 20. (Step S204). The voice recognition server determination unit 104 may newly determine the specific voice recognition server 20 by the same method as the example described in step S104. The newly determined voice recognition server 20 may not be changed from the voice recognition server 20 currently in communication.

When changing the specific speech recognition server 20 used for a specific user during the web conference (step S205, YES), the translation server speed determination unit 105 of the web conference server 10 uses the newly determined specific The communication speed between the speech recognition server 20 and the plurality of translation servers 30 and the information processing speed of the plurality of translation servers 30 are determined (step S206). The communication speed and information processing speed are respectively real-time speeds.

The communication speed between the newly determined specific speech recognition server 20 and the plurality of translation servers 30 and the information processing speed of the plurality of translation servers 30 are determined (step S206). This is because there is a possibility that the most suitable translation server 30 for step S205, YES) is a translation server 30 other than the translation server 30 currently in communication. By selecting the translation server 30 that is most suitable for cooperating with the new speech recognition server 20, a combination of the speech recognition server 20 and the translation server 30 that is generally optimum in terms of both speed and accuracy can be obtained. , can be selected in real time.

The translation server determination unit 108 of the web conference server 10 selects a plurality of A specific translation server 30 to be used for a specific user is newly determined from the translation servers 30 (step S207). The translation server determination unit 108 may newly determine a specific translation server 30 by the same method as the example described in step S108. It should be noted that the newly determined translation server 30 may not be changed from the translation server 30 currently in communication.

If the specific translation server 30 used for the specific user is not changed during the web conference (step S208, NO), the speech data processing request unit 109 sends the newly determined specific speech recognition server 20 , the voice data of the particular user and the identification information identifying the particular (communicating) translation server 30 (step S209). This allows the specific speech recognition server 20 to be used for a specific user to be changed during the web conference, allowing the optimum speech recognition server 20 to be used in both speed and accuracy in real time, and As a whole, processing is distributed and stable communication and information processing can be performed.

On the other hand, if the specific translation server 30 used for a specific user is to be changed during the web conference (step S208, YES), the voice data processing request unit 109 of the web conference server 10 will The specified speech recognition server 20 is supplied with the specified user's voice data and the identification information for identifying the newly determined specified translation server 30 (step S210). This further allows the specific translation server 30 to be used for a specific user to be changed during the web conference so that the optimal translation server 30 can be used in real time in terms of both speed and accuracy, and the web conference As a whole, processing is distributed and stable communication and information processing can be performed.

FIG. 5 shows the third operation flow of the web conference server (during the web conference).

During the web conference, the translation server speed determination unit 105 of the web conference server 10 periodically (loops) the specific (in-communication) speech recognition server 20 and the specific (in-communication) translation server 30 and the information processing speed of a specific (in communication) translation server 30 (step S301). The communication speed and information processing speed are respectively real-time speeds. The translation server speed determination unit 105 determines whether the communication speed between the specific speech recognition server 20 and the specific translation server 30 and/or the information processing speed of the specific translation server 30 has changed below a threshold during the web conference. It is determined whether or not (step S302). The threshold is, for example, a speed-unacceptable value for conducting a smooth web conference.

If the communication speed between the specific speech recognition server 20 and the specific translation server 30 and/or the information processing speed of the specific translation server 30 changes below the threshold during the web conference (step S302, YES), the translation The server speed determination unit 105 determines the communication speed between a specific (in-communication) speech recognition server 20 and a plurality of translation servers 30 (a plurality of translation servers 30 other than the translation server 30 in communication), information processing speed (step S303). The communication speed and information processing speed are respectively real-time speeds.

The translation server determination unit 108 of the web conference server 10 determines the communication speed between the specific speech recognition server 20 determined by the translation server speed determination unit 105 and the plurality of translation servers 30, and the information processing speed of the plurality of translation servers 30 (step S303), a specific translation server 30 to be used for a specific user is newly determined from a plurality of translation servers 30 (step S304). The translation server determination unit 108 may newly determine a specific translation server 30 by the same method as the example described in step S108. It should be noted that the newly determined translation server 30 may not be changed from the translation server 30 currently in communication.

When the specific translation server 30 used for a specific user is changed during the web conference (step S305, YES), the speech recognition server speed determination unit 103 of the web conference server 10 and the web conference server 10 and the information processing speed of a plurality of speech recognition servers 20 (step S306). The communication speed and information processing speed are respectively real-time speeds.

The reason for determining the speed of the plurality of speech recognition servers 20 (step S306) is that the communication speed between the specific speech recognition server 20 and the specific translation server 30 and/or the information processing speed of the specific translation server 30 is (step S302, YES) means that there is a possibility that there is a problem with the specific speech recognition server 20, in which case it may be better to change the speech recognition server 20 to be used. Because there is As a result, by selecting the optimal speech recognition server 20 to cooperate with the new translation server 30, a set of the optimal speech recognition server 20 and translation server 30 in terms of both speed and accuracy can be comprehensively selected. , can be selected in real time.

The speech recognition server determination unit 104 of the web conference server 10 determines the language attribute determined by the speech recognition unit 102 (step S102), the web conference server 10 determined by the speech recognition server speed determination unit 103, and the plurality of voice recognition servers 20. and the information processing speed of the plurality of speech recognition servers 20 (step S306), the specific speech recognition server 20 to be used for the specific user is newly determined from the plurality of speech recognition servers 20 ( step S307). The voice recognition server determination unit 104 may newly determine the specific voice recognition server 20 by the same method as the example described in step S104. The newly determined voice recognition server 20 may not be changed from the voice recognition server 20 currently in communication.

If the specific speech recognition server 20 used for the specific user is not changed during the web conference (step S308, NO), the speech data processing request unit 109 requests the specific (in-communication) speech recognition server 20 , the specific user's voice data and identification information for identifying the newly determined translation server 30 are supplied (step S309). This allows the specific translation server 30 to be used for a specific user to be changed during the web conference, allowing the optimal translation server 30 to be used in terms of both speed and accuracy in real time, as well as the overall web conference. From this point of view, processing is distributed and stable communication and information processing can be performed.

On the other hand, when changing the specific voice recognition server 20 used for a specific user during the web conference (step S308, YES), the voice data processing request unit 109 of the web conference server 10 newly determines The specified speech recognition server 20 is supplied with the specified user's voice data and the identification information for identifying the newly determined specified translation server 30 (step S310). This further allows the particular speech recognition server 20 to be used for a particular user to be changed during the web conference, allowing the optimal speech recognition server 20 to be used in both speed and accuracy in real time, and Processing is decentralized and stable communication and information processing can be performed even when looking at the web conference as a whole.

Although the embodiments and modifications of the present technology have been described above, the present technology is not limited to the above-described embodiments, and various modifications can be made without departing from the gist of the present technology. Of course.

Web conference system 1
Web conference server 10
Voice recognition server 20
translation server 30
User terminal 40

Claims (11)

A web conferencing server,
a voice acquisition unit that acquires voice data of a specific user included in a plurality of users participating in a web conference from a user terminal of the specific user;
a voice recognition unit that recognizes the voice data and determines the language attribute of the specific user;
a speech recognition server speed determination unit that determines a communication speed between the web conference server and the plurality of speech recognition servers and an information processing speed of the plurality of speech recognition servers;
Based on the language attribute, the communication speed with the plurality of speech recognition servers, and the information processing speed of the plurality of speech recognition servers, a specific a voice recognition server determination unit that determines a voice recognition server;
a translation server speed determination unit that determines a communication speed between the specific speech recognition server used for the specific user and a plurality of translation servers and an information processing speed of the plurality of translation servers;
Translation server determination for determining a specific translation server to be used for the specific user from among the plurality of translation servers based on a communication speed with the plurality of translation servers and an information processing speed of the plurality of translation servers. Department and
The speech data of the specific user is supplied to the specific speech recognition server by supplying the speech data of the specific user and identification information for identifying the specific translation server to the specific speech recognition server. a speech data processing requesting unit for generating speech recognition data, which is text data, by recognizing the voice of the user, and causing the specific translation server to translate the speech recognition data of the specific user to generate translation data, which is text data;
and
A web conference server in which a combination of the specific speech recognition server and the specific translation server is different for each of the plurality of users participating in the web conference. The web conferencing server of claim 1, comprising:
The speech recognition server speed determination unit periodically determines a communication speed between the web conference server and the specific speech recognition server and an information processing speed of the specific speech recognition server during the web conference. death,
When the communication speed between the web conference server and the specific speech recognition server and/or the information processing speed of the specific speech recognition server changes below a threshold during the web conference,
The speech recognition server speed determination unit determines a communication speed between the web conference server and the plurality of speech recognition servers and an information processing speed of the plurality of speech recognition servers,
The speech recognition server determination unit newly determines a specific speech recognition server to be used for the specific user from the plurality of speech recognition servers,
The speech data processing request unit supplies the newly determined specific speech recognition server with the speech data of the specific user and identification information for identifying the specific translation server, thereby obtaining the specified translation server. changing the specific speech recognition server used for the user of the web conference server during the web conference. A web conferencing server according to claim 2,
When changing the specific speech recognition server used for the specific user during the web conference,
The translation server speed determination unit determines the communication speed between the newly determined specific speech recognition server and the plurality of translation servers and the information processing speed of the plurality of translation servers,
The translation server determination unit, based on the newly determined communication speed between the specific speech recognition server and the plurality of translation servers and the information processing speed of the plurality of translation servers, from the plurality of translation servers, newly determining a specific translation server to be used for the specific user;
The speech data processing request unit supplies the newly determined specific speech recognition server with the speech data of the specific user and identification information for identifying the newly decided specific translation server. a web conference server that changes the specific translation server used for the specific user during the web conference. A web conference server according to any one of claims 1 to 3,
The translation server speed determination unit periodically determines the communication speed between the specific speech recognition server and the specific translation server and the information processing speed of the specific translation server during the web conference. ,
If the communication speed between the specific speech recognition server and the specific translation server and/or the information processing speed of the specific translation server changes below a threshold value during the web conference,
The translation server speed determination unit determines a communication speed between the specific speech recognition server and the plurality of translation servers and an information processing speed of the plurality of translation servers,
The translation server determination unit newly determines a specific translation server to be used for the specific user from the plurality of translation servers,
The speech data processing request unit supplies the newly determined specific speech recognition server with the speech data of the specific user and identification information for identifying the newly decided specific translation server. a web conference server that changes the specific translation server used for the specific user during the web conference. A web conferencing server according to claim 3,
When changing the specific translation server used for the specific user during the web conference,
The speech recognition server speed determination unit determines the communication speed between the newly determined specific translation server and the plurality of speech recognition servers and the information processing speed of the plurality of speech recognition servers,
The speech recognition server determination unit determines the plurality of speech recognition servers based on the newly determined communication speed between the specific translation server and the plurality of speech recognition servers and the information processing speed of the plurality of speech recognition servers. From the server, newly determine a specific speech recognition server to be used for the specific user,
The speech data processing request unit supplies the newly determined specific speech recognition server with the speech data of the specific user and identification information for identifying the newly decided specific translation server. changing the specific speech recognition server used for the specific user during the web conference. A web conference server according to any one of claims 1 to 5,
a processed data acquisition unit that acquires the speech recognition data of the specific user from the specific speech recognition server and the translation data of the specific user from the specific translation server;
a context check unit that checks the context of the corresponding speech recognition data and the translation data and corrects the corresponding speech recognition data and/or the translation data according to the check result;
Real-time outputting the corresponding speech recognition data and the translation data corrected according to the check result to the plurality of user terminals of the plurality of users participating in the web conference in real time during the web conference. an output unit;
A web conferencing server further comprising: A web conferencing server according to claim 6,
minutes data of the web conference is created based on the corresponding speech recognition data and the translation data after being corrected according to the check result, and the minutes data is sent to multiple users participating in the web conference a minutes creation unit that outputs to the user terminal of
A web conferencing server further comprising: A web conference server according to any one of claims 1 to 7,
a content determination unit that determines content of the web conference based on the result of voice recognition by the voice recognition unit;
an emotion determination unit that determines the emotion of the specific user based on the result of voice recognition by the voice recognition unit;
further comprising
A web conference server, wherein the translation server determination unit determines a specific translation server to be used for the specific user, further based on the content of the web conference and/or emotions of the specific user. A web conference server according to any one of claims 1 to 8,
The user terminal with which the web conference server communicates selects the specific web conference server with the fastest communication speed from the plurality of web conference servers based on the IP address of the user terminal, and selects the specific web conference server. to access
The specific web conference server accessed by the user terminal differs for each user terminal of a plurality of users participating in the web conference. Acquiring voice data of a specific user included in a plurality of users participating in a web conference from a user terminal of the specific user;
recognizing the voice data, determining the language attribute of the specific user,
determining a communication speed between a web conference server and a plurality of speech recognition servers and an information processing speed of the plurality of speech recognition servers;
Based on the language attribute, the communication speed with the plurality of speech recognition servers, and the information processing speed of the plurality of speech recognition servers, a specific determine the speech recognition server,
Determining the communication speed between the specific speech recognition server used for the specific user and the plurality of translation servers and the information processing speed of the plurality of translation servers,
determining a specific translation server to be used for the specific user from among the plurality of translation servers based on a communication speed with the plurality of translation servers and an information processing speed of the plurality of translation servers;
The speech data of the specific user is supplied to the specific speech recognition server by supplying the speech data of the specific user and identification information for identifying the specific translation server to the specific speech recognition server. to generate speech recognition data that is text data, and cause the specific translation server to translate the speech recognition data of the specific user to generate translation data that is text data,
A web conference method in which a combination of the specific speech recognition server and the specific translation server differs for each of the plurality of users participating in the web conference. a web conferencing server interconnected via a network;
a plurality of speech recognition servers;
a plurality of translation servers;
and
The web conferencing server
a voice acquisition unit that acquires voice data of a specific user included in a plurality of users participating in a web conference from a user terminal of the specific user;
a voice recognition unit that recognizes the voice data and determines the language attribute of the specific user;
a speech recognition server speed determination unit that determines a communication speed between the web conference server and the plurality of speech recognition servers and an information processing speed of the plurality of speech recognition servers;
Based on the language attribute, the communication speed with the plurality of speech recognition servers, and the information processing speed of the plurality of speech recognition servers, a specific a voice recognition server determination unit that determines a voice recognition server;
a translation server speed determination unit that determines a communication speed between the specific speech recognition server used for the specific user and the plurality of translation servers and an information processing speed of the plurality of translation servers;
Translation server determination for determining a specific translation server to be used for the specific user from among the plurality of translation servers based on a communication speed with the plurality of translation servers and an information processing speed of the plurality of translation servers. Department and
The speech data of the specific user is supplied to the specific speech recognition server by supplying the speech data of the specific user and identification information for identifying the specific translation server to the specific speech recognition server. a speech data processing requesting unit for generating speech recognition data, which is text data, by recognizing the voice of the user, and causing the specific translation server to translate the speech recognition data of the specific user to generate translation data, which is text data;
has
A web conference system in which a combination of the specific speech recognition server and the specific translation server differs for each of the plurality of users participating in the web conference.

Images