JP2020119043A - Voice translation system and voice translation method - Google Patents

Abstract

To provide a voice translation system and a voice translation method that smoothly translate a conversation without having to specify that a language is different for each language to be used even in situations that it is the first time for the conversation without any information about the other party.SOLUTION: A voice translation system 1 includes: an information terminal 10 capable of inputting a voice of a conversation in which each user uses a different language and outputting the content of the input voice of the user in a language different from the language used by the user; and a translation server 20 connected to the information terminal via a network 30 and translating the content of the voice of the user input to the information terminal into the language different from the language used by the user. The translation server includes: language specifying means 21 for specifying a plurality of languages used in the conversation based on the voice of the conversation; and translation means 22 for translating the content of the voice of the user into a language different from the language used by the user among the plurality of languages specified by the language specifying means.SELECTED DRAWING: Figure 1

Description

The present invention relates to a speech translation system and a speech translation method.

In order to understand the conversation between people who do not understand each other's language, for example, the conversation between a taxi driver and a customer who can only speak in a language other than Japanese (such as tourists from overseas) in their native language. , Many translation systems have been proposed that translate the speaker's voice into the listener's native language.

For example, in Patent Document 1, a voice can be input and output, a voice translation device that translates the content of the input voice into a different language and outputs the voice, and a voice input and output are possible. And a voice input/output device wirelessly connected to the voice translation device, wherein the voice translation device translates the content of the voice input to the voice input/output device into a different language and outputs the voice from the voice translation device. A voice translation system is disclosed that outputs the voice, translates the content of the voice input to the voice translation device into different languages, and outputs the voice from the voice input/output device.

JP, 2008-173910, A

However, the above-described speech translation system has a problem that the language used in the speech translation system must be designated first.

In addition, in order to specify the language to be used, it is necessary to know in advance which language the other party speaks. For example, there is no information about the other party, such as the relationship between the taxi crew and the customer, and the conversation takes place. If this is the first time, there is a problem that it is difficult to use the above speech translation system.

Further, the above-described voice translation system has a problem that it is necessary to specify that the language is different for each conversation, that is, for each speaker.

In view of the above-mentioned circumstances, the present invention is a situation in which it is not necessary to specify that the language is different for each conversation, and there is no information about the other party (information such as which language to speak) at all for the first time. An object of the present invention is to provide a voice translation system and a voice translation method that can smoothly translate a conversation.

The inventor of the present invention, as a result of earnest research and development on the above-mentioned problems, has found the following epoch-making speech translation system and speech translation method.

A first aspect of the present invention for solving the above problem is that each user can input a voice of conversation using a different language, and the content of the input voice of the user is different from the language used by the user. A speech translation system including an information terminal for outputting in a language and a translation server connected to the information terminal via a network and translating the content of the user's voice input to the information terminal into a language different from the language used by the user. That is, the translation server determines the content of the user's voice from the language of the conversation, and the content of the user's voice from among the plurality of languages identified by the language identification means. And a translation means for translating the language into a language different from the language used by the voice translation system.

Here, “output” is a concept including not only outputting as voice but also displaying the content of voice in characters. In addition, “voice” is a concept that includes voice data as well as voice. However, "voice output" or "voice output" means that the voice data is output as voice.

In the first aspect, the conversation can be smoothly translated for each conversation even if the conversation is the first time without specifying the language used for the conversation and without any information about the other party. You can

A second aspect of the present invention is the speech translation system according to the first aspect, wherein the translation means translates using artificial intelligence.

In the second aspect, more accurate translation can be provided.

A third aspect of the present invention is the voice translation system according to the first or second aspect, wherein the output from the information terminal is output as voice.

In the third aspect, since the translation is output as a voice, conversation using different languages can be smoothly conducted.

A fourth aspect of the present invention is characterized in that the information terminal has a language designating means, and the translation server adds the language designated by the language designating means to the language designated by the language specifying means. The speech translation system according to any one of the first to third aspects.

In the fourth aspect, at least one of the languages used can be designated, so that the content of the input voice can be translated more accurately.

A fifth aspect of the present invention is to attach a user's voice that is attached to a user's ear or its vicinity and is connected to at least one of an information terminal and a translation server via a network, and translated into different languages by the translation server. The voice translation system according to any one of the first to fourth aspects, further comprising a voice output terminal that outputs the content as voice.

Here, “in the vicinity of the user's ear” means a place where the user can reliably hear the output voice, and examples thereof include the upper part of the ear and the neck.

In the fifth aspect, the translated voice can be easily heard, so that conversation using different languages can be conducted more smoothly.

A sixth aspect of the present invention is a speech translation method for outputting the content of the input voice of the user in a language different from the language used by the user among the conversations in which the users use different languages. A language specifying step for specifying a plurality of languages used for conversation and a content of a user's voice are translated from a voice into a language different from the language used by the user among the plurality of languages specified by the language specifying step. And a translation step.

In the sixth aspect, the conversation can be smoothly translated without designating the language used for the conversation and without any information about the other party, for example, even in the case of the first conversation. ..

In the present invention, “means” and “system” do not simply mean physical means, but include cases where the functions of the “means” and “system” are realized by software. Further, even if the function of one “means” or “system” is realized by two or more physical means or devices, the functions of two or more “means” or “system” are one physical means or It may be realized by the device.

FIG. 1 is a schematic conceptual diagram of the speech translation system according to the first embodiment. FIG. 2 is a flowchart showing the operation of the speech translation system according to the first embodiment. FIG. 3 is a schematic conceptual diagram of the speech translation system according to the second embodiment.

An embodiment of a speech translation system according to the present invention will be described below with reference to the accompanying drawings. The present invention is not limited to the embodiments below.
(Embodiment 1)

FIG. 1 is a schematic conceptual diagram of a speech translation system according to this embodiment. As shown in this figure, the speech translation system 1 includes an information terminal 10, a translation server 20 connected to the information terminal 10 via a network 30, and a plurality of (three) speech outputs connected to the information terminal 10. It is composed of a terminal 40.

The information terminal 10 is not particularly limited as long as it can input the voice spoken by the users 100a to 100c. The information terminal 10 may have a voice output device such as a speaker so that a translated voice described later can be output. Examples of the information terminal 10 include a mobile phone, a smartphone, and a dedicated terminal having a voice input/output function. Here, the users 100a to 100b are people who speak in different languages.

The translation server 20 identifies the language used by each of the users 100a to 100c from the conversation of the users 100a to 100c and the content of the user's voice in a plurality of languages identified by the language specifying means 21. Among them, it has a translation means 22 for translating it into a language different from the language used by the user.

The language specifying unit 21 is not particularly limited as long as it has a function of specifying the language used by each of the users 100a to 100c, and examples thereof include a program having such a function. Specifically, for example, when the user 100a speaks Japanese, the user 100b speaks English, and the user 100c speaks French, the language specifying means 21 uses Japanese from the voice of the user 100a for conversation. A program that specifies that the user 100b speaks English, and that the user 100c sounds that French is used in the conversation. Here, as a method of specifying the language, for example, a method of detecting a word included in the input voice and determining which language is used from the word, and the like can be mentioned.

In addition, the translation unit 22 has a function of translating into a language different from the language used by the user among the plurality of languages identified by the language identification unit 21 (for example, first, converting voice information into text data, There is no particular limitation as long as it has a function of converting text data or the like into translated text data translated into another language, and further converting the translated text data into translated voice information. For example, such a function And the like. Specifically, for example, when the user 100a speaks Japanese, the user 100b speaks English, and the user 100c speaks French, the translating means 22 serves as the translating means 22 (the content of the voice of the user 100a (expressed in Japanese). (Contents) into English or French.

The translation unit 22 is not particularly limited as long as it can translate the content of the user's voice into a language different from the language used by the user, but it is preferable that it can be translated using artificial intelligence (AI). .. By using artificial intelligence, the learning effect of artificial intelligence enables more accurate translation. The artificial intelligence is not particularly limited, and known techniques that can be used for translation can be used.

The translation server 20 is not particularly limited as long as it has the language specifying unit 21 and the translation unit 22 described above. Examples of the translation server 20 include personal computers and computer systems in which they are connected in parallel.

The network 30 is not particularly limited as long as it can connect the information terminal 10 and the translation server 20 (including wired connection and wireless connection), and examples thereof include the Internet and an intranet.

The voice output terminal 40 can be connected to the information terminal 10 (including wired connection and wireless connection), and outputs the voice content of the users 100a to 100c translated by the translation server 20 into different languages. There is no particular limitation as long as it can be done. Examples of the voice output terminal 40 include earphones and speakers that are connected to the information terminal 10 wirelessly, but earphones are preferable. By using the earphones, it is possible to suppress the influence of noise such as external environmental sound, so that the translated sounds can be heard by the users 100a to 100c in an easily understandable manner.

Next, the operation of the speech translation system 1 will be described. FIG. 2 is a flowchart showing the operation of the speech translation system 1.

As shown in this figure, when the power of the information terminal 10 or the start button for translation is pressed, the speech translation system 1 starts to operate, and the speech of the conversation of the users 100a to 100c is input to the information terminal 10 (S1). ).

Then, the information terminal 10 transmits the input voice to the translation server (S2). Upon receiving the transmitted voice, the translation server 20 identifies the language used for the conversation by the language identifying means 21 (S3). Specifically, it specifies that the transmitted voice uses Japanese, English, and French.

Next, the translation server 20 translates the content of the voice into a language different from the language used for the voice among the plurality of languages identified by the language identifying means 21 (S4). Specifically, it translates Japanese voice content into English voice and French voice. That is, the Japanese voice is translated into two languages (English and French).

Then, the translation server 20 transmits the voice translated into a different foreign language to the information terminal 10 (S5). Then, the information terminal 10 transmits the translated voice to the voice output terminal 40 (S6).

The voice output terminal 40 outputs the voice translated as an automatic voice (S7). Specifically, the voice output terminal 40 outputs, for example, an English voice and then a French voice. Then, steps S4 to S7 are repeated until the conversation ends (S8), and when the conversation ends, the speech translation system 1 stops.

As described above, by configuring the speech translation system 1 according to the present embodiment, it is the first time for a conversation without specifying the language used for the conversation and without any information about the other party. However, the conversation can be translated smoothly.
(Embodiment 2)

In the above-described first embodiment, the language used in the conversation is not designated, but when the language to be used is determined, the speech translation system may be configured so that the language to be used can be designated in advance.

For example, the speech translation system 1A according to the present embodiment shown in FIG. 3 includes an information terminal 10A and a translation server 20 connected to the information terminal 10A via a network 30. Unlike the translation system, there is no voice output terminal.

The information terminal 10A can specify a language used by either of the users 100a and 100b, in addition to the function of being able to input a voice spoken by the users 100a and 100b and outputting a translated voice. There is no particular limitation as long as it has a language specification means (not shown) that can be used.

Examples of the language designating means include a program that displays a language name in English on a display unit such as a liquid crystal screen, and can designate the language by touching the language name with a finger.

As the information terminal 10A, for example, a mobile phone, a smartphone, a dedicated terminal having the above-mentioned function, or the like can be given. The translation server 20 and the network 30 are the same as in the first embodiment.

In this embodiment, for example, the language used by the user 100a is known. Therefore, before the user 100a and the user 100b start a conversation, the language to be used can be input to the information terminal 10 in advance by the language designating means. The input information about the language is transmitted to the translation server 20 and added to the language specified by the language specifying means 21. Then, the language used for the conversation (at least one language) will not be erroneously specified. As a result, the probability that the translation server 20 erroneously translates the voice is reduced.

As described above, by configuring the speech translation system 1A according to this embodiment, it is possible to provide a speech translation system that can translate more accurately.
(Other embodiments)

In the above-described embodiment, the case where the number of users who use the voice translation system is two or three has been described, but the present invention is not limited to this, and the voice translation according to the present invention can be applied to a conversation in which more users participate. It goes without saying that the system can be used.

Further, in the above-described embodiment, the voice translation system is configured so as to finally output the voice translated as the automatic voice, but the present invention is not limited to this. For example, a display unit may be provided in the information terminal and the translated text data of the voice may be displayed on the display unit. Even in this case, the same effect as that of the above-described embodiment can be obtained.

Further, in the above-described embodiment, the voice translation system is configured by one information terminal, but the voice translation system may be configured by connecting a plurality of information terminals to the translation server. Even if the voice translation system is configured as described above, the same effect as that of the above-described embodiment can be obtained.

1, 1A voice translation system 10, 10A information terminal 20 translation server 21 language specifying means 22 translation means 30 network 40 voice output terminal

Claims (6)

An information terminal in which each user can input the voice of conversation using a different language, and outputs the content of the input user's voice in a language different from the language used by the user,
A speech translation system comprising a translation server which is connected to the information terminal via a network and translates the content of the user's voice input to the information terminal into a language different from the language used by the user,
The translation server has a language specifying means for specifying a plurality of languages used in the conversation from the voice of the conversation,
Translation means for translating the content of the user's voice into a language different from the language used by the user among the plurality of languages identified by the language identification means.
A voice translation system characterized by the following. The speech translation system according to claim 1, wherein the translation unit translates using artificial intelligence. The speech translation system according to claim 1, wherein the output from the information terminal is output as a voice. The information terminal has a language designating means,
The speech translation system according to any one of claims 1 to 3, wherein the translation server adds the language specified by the language specifying means to the language specified by the language specifying means. It is attached to the user's ear or its vicinity and is connected to at least one of the information terminal and the translation server via a network,
The voice translation system according to any one of claims 1 to 4, further comprising a voice output terminal that outputs, as a voice, the content of the user's voice translated into a different language by the translation server. A speech translation method for outputting, in a conversation in which each user uses a different language, the content of the input user's voice in a language different from the language used by the user,
A language identification step of identifying a plurality of languages used in the conversation from the voice of the conversation,
A translation step of translating the content of the user's voice into a language different from the language used by the user among the plurality of languages identified by the language identification step;
A speech translation method comprising: