JP2004145237A - Interactive doll with translating function - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide an interactive doll with a translating function that has a speech translation system and is used to learn a basic foreign language conversation by converting a speech that a user speaks into a speech of a different language and voicing it. <P>SOLUTION: Disclosed is the interactive doll with the translating function and when a conventional tandem type color image forming apparatus is made small-sized, it becomes difficult to secure positioning precision and driving precision of an image formation unit, so a color image has more color slurring and color unevenness. Further, the image formation unit itself is small-sized, so work space for replacement of it as a consumption article by the user is eliminated to cause a problem of deterioration in operability. A speech of a 1st language by the user is recognized and converted into a speech of a 2nd language and a previously stored speech is voiced corresponding to the speech of the 2nd language or the speech of the 1st language to interact with the user. <P>COPYRIGHT: (C)2004,JPO

Description

[0001]
TECHNICAL FIELD OF THE INVENTION
The present invention relates to an interactive doll, and more particularly, to provide a voice translation system in a doll, convert a voice uttered by a user into a voice based on a different language, and utter the voice to satisfy the playfulness of the user. The present invention also relates to an interactive doll with a translation function that can be used for elementary foreign language learning. That is, the present invention relates to an interactive doll with a translation function that has not existed in the past.
[0002]
[Prior art]
Generally, children tend to learn life education through interesting play and toys, and the intimate contact with the toys performs imitation learning that leads to the real world. The majority of such imitation learning is performed through toy dolls, for example, a child creates his own imitation learning scenario and induces the toy doll to respond appropriately according to the scenario. In other words, the child is immersed in the imitation learning by making the appropriate speech expression and the action act interesting and interactive.
[0003]
Education using toy dolls like this has been handed down to children since ancient times.Recently, research on voice dolls has become active in anticipation of their educational effects, and the production of more progressive toy dolls has been promoted. Have been tried. In most of such toy dolls according to the related art, a touch sensor is provided at a predetermined position. When a child operates the touch sensor, a magnetic recording medium (magnetic tape) or a semiconductor recording medium (IC memory) is used. simple sentences, which were recorded in), for example, "Hello.", "I am a ○○○.", "who are you?", and the voice, such as "do you like what you." is uttered Is done. For example, Patent Literature 1 discloses an utterance doll that can interest children and further interest conversation.
[0004]
[Patent Document 1]
Registered Utility Model No. 25662439
[0005]
Specifically, in order to achieve the above-mentioned purpose, Patent Document 1 described above stores a head detachably attached to the body of a doll, a detecting means for detecting an operation from outside, and a plurality of vocabularies stored therein. A housing containing a storage unit for performing the processing, a selecting unit for selecting an arbitrary vocabulary from a plurality of vocabularies, a voice synthesizing unit for synthesizing a voice from the selected vocabulary, and a voice generating unit for uttering the synthesized voice are arranged in the body. There is disclosed a voice doll in which a battery accommodating portion formed so as to stabilize with a padding and protruding from an upper portion of a housing is projected upward from a body portion, and the head is fitted to the battery accommodating portion.
[0006]
The toy doll described in Patent Document 1 is a sporadic, simple puppet that speaks a simple sentence. The operation of the touch sensor allows the user to listen to a sound recorded with a simple sentence without a scenario. It can be triggered temporarily. However, there is a problem that the child gets tired immediately and the period of playing with such a toy doll is actually shortened, so that the educational effect is low. In addition, the voice sentences spoken by the conventional toy dolls are not interactive scenarios but a series of discontinuous sentences, which is not realistic. From this, there is a problem that the educational effect is low.
[0007]
On the other hand, for example, in Patent Literature 2, in order to solve such a problem, audio output according to a topic is enabled, and a behavior pattern that a child may perform is created and recorded in a scenario, and an arbitrary pattern is set. There is disclosed a toy doll that enables two-way dialogue with the doll depending on the situation. For example, in situations where you interact with a child, in order to lead to various scenarios, after compressing the voice with voice compression software, record it in the memory unit, quickly retrieve it when necessary, and make it possible to select a single topic. It is said that questions and answers can be made immediately in response. Specifically, Patent Document 2 described above records voice compression data obtained by compressing a digital voice signal stream of a large number of sentences at a predetermined compression rate in a doll body formed into a mixture of human and animal forms. A voice-recognition interactive doll toy comprising a first memory unit, and a second memory unit having a calculation area for recognizing a user's voice signal input from the outside is disclosed. .
[0008]
[Patent Document 2]
Patent No. 3164346
[0009]
Certainly, the doll toy described in Patent Document 2 utters a voice in accordance with the conversation of the user. However, there is no mention of the language of the voice uttered by the user and the language of the voice uttered by the voice dolls and doll toys. This is the same in Patent Document 1 described above. In view of the problems and effects of the invention according to Patent Document 2, it is considered that a case is assumed in which a Japanese infant utters in Japanese, and a doll toy that recognizes it utters a voice in Japanese. Can be
[0010]
Here, with the recent internationalization of society, Japan has lagged behind other countries despite the vigorous debate about the importance of foreign language education. This is thought to be due to problems with the living environment in childhood in ordinary households, aside from school education. Then, it is clear that if the above-mentioned speech doll or doll toy has a function as an item for foreign language education, it will also be useful for the advancement and development of foreign language education. In the prior art, a doll toy having such a function is not provided.
[0011]
[Problems to be solved by the invention]
Therefore, the present invention provides a voice translation system in a doll, converts a voice uttered by a user into a voice based on a different language, and utters the voice, thereby satisfying the user's playfulness and learning a foreign language conversation. It is an object of the present invention to provide an interactive doll with a translation function that can be provided to a user. That is, an object of the present invention is to provide an interactive doll with a translation function that has not existed in the past.
[0012]
[Means for Solving the Problems]
In order to solve the above problem, the present invention receives a first voice based on a first language by a user, converts the first voice into a second voice based on a second language, and then converts the converted second voice. There is provided an interactive doll with a translation function, wherein the interactive doll is characterized by uttering a voice or a third voice based on a second language stored in advance in correspondence with the first voice and interacting with the user.
[0013]
The interactive doll with translation function,
An ear portion having audio receiving means for receiving the first audio,
It is preferable to include a portion having voice conversion means for converting the first voice to the second voice, and a mouth portion having voice utterance means for generating the second voice or the third voice.
Preferably, the first voice and the second voice are words.
[0014]
Further, the voice conversion means,
(A) voice recognition means for recognizing the first voice,
(B) voice translation means for translating the recognized first voice into the second language;
(C) It is preferable to include a voice synthesizing unit that synthesizes the voice based on a result of the translation.
[0015]
It is preferable that the voice recognition means (a) recognizes the first voice as a phoneme sequence.
Specifically, the voice recognition means (a)
Audio processor means for receiving an audio signal of the first audio and converting the audio signal into a corresponding electrical signal;
Analog / digital converter means for digitizing the electrical signal at a predetermined sampling rate to form a digitized audio signal;
Means for performing time domain analysis on the subdivided portions of the digitized audio signal to identify a plurality of time domain characteristics of the audio signal, and a plurality of components having predetermined high and low cutoff frequencies; Means for filtering each of the subdivided portions using a filter band and identifying at least one frequency domain characteristic of each of the subdivided portions; and processing the time domain characteristics and frequency domain characteristics. It is preferable to have a voice phoneme identification means including a means for identifying a phoneme included in the voice signal.
[0016]
Further, it is preferable that the voice translation means (b) translates the recognized phoneme sequence into a vocabulary sequence based on a second language.
Further, it is preferable that the voice synthesizing means (c) synthesizes the second voice by computer processing the vocabulary string.
[0017]
Specifically, the voice synthesis means (c) comprises:
A speech conversion subsystem that receives a vocabulary sequence based on the second language and converts the vocabulary sequence into a first phoneme sequence;
Receiving a transformation rule and applying it to the first phoneme sequence to form a second phoneme sequence,
An evaluator that ranks phonemes included in the second phoneme sequence based on a predetermined criterion, and
It is preferable to include a syllable decomposer that receives the second phoneme sequence and decomposes phonemes included in the second phoneme sequence into syllables using the ranking.
[0018]
Further, in the interactive doll with a translation function,
The voice receiving means receives a keyword based on a first language by the user, the voice recognizing means (a) recognizes the keyword, and the voice uttering means stores in advance a second keyword corresponding to the keyword. Ask a question based on your language,
Then, it is preferable that the first voice by the user is converted into the second voice, and the second voice or the third voice is uttered to interact with the user.
[0019]
In such an interactive doll with a translation function,
The voice recognition means (a) recognizes a specific portion of the first voice,
The voice translation means (b) translates the specific part into voice based on a second language,
The speech synthesis means (c) may synthesize the second speech by applying the translation result to a speech answer pattern based on a second language stored in advance corresponding to the question.
[0020]
In this case, the voice translation means (b)
(B-1) storage means for storing the voice response pattern and a plurality of vocabularies based on a second language expected as vocabulary constituting the specific part in response to the question;
(B-2) selecting means for selecting a vocabulary based on a second language, corresponding to the vocabulary constituting the specific part recognized by the voice recognition means (a),
Preferably, the voice synthesizing means (c) synthesizes a vocabulary voice based on the selected second language, applies the vocabulary to the voice answer pattern, and synthesizes the second voice.
[0021]
Further, it is preferable that the interactive doll with a translation function further includes language specifying means (d) for specifying the first language and the second language.
Further, the interactive doll with a translation function detects an operation from the outside and detects the voice recognition means (a), the voice conversion means (b), the voice utterance means (c), and the language identification means (d). )), It is preferable to have control means (e) for controlling at least one means selected from the group consisting of:
[0022]
In the interactive doll with a translation function, it is preferable that the control means (e) has a function of causing the voice utterance means (c) to utter the second voice or the third voice a plurality of times.
The control means (e) specifies a plurality of languages as the second language in the language specifying means (d), and the second voice or the third voice in the voice utterance means (c). It is preferable to have a function of uttering continuously based on a plurality of languages.
[0023]
BEST MODE FOR CARRYING OUT THE INVENTION
The present invention receives a first voice based on a first language by a user and converts it into a second voice based on a second language, and then converts the converted second voice, or the first voice The present invention relates to an interactive doll with a translation function, characterized by uttering a third voice based on a second language stored in advance corresponding to the voice and interacting with the user. Hereinafter, an interactive doll with a translation function according to the present invention will be described with reference to the drawings.
[0024]
FIG. 1 is a diagram for schematically explaining the functions of the interactive doll with a translation function according to the present invention. As shown in FIG. 1, the interactive doll 1 with a translation function according to the present invention hears a voice based on the first language spoken by the user, for example, a voice saying "My name is cherry blossoms" in Japanese. The voice is received by the voice receiving means provided in the unit 1a, and the voice is based on the second language, for example, in English, by the translation system module including the voice converting means 1b which is incorporated in a part of the doll and which realizes the present invention. It is converted into "My name is Sakura." Or "Main name in Sakura." In German, and this second voice is uttered from the voice uttering means provided in the mouth part 1c.
[0025]
Also, although not shown in FIG. 1, when the first voice is, for example, “song.”, The “happy bass” stored in advance corresponding to the first voice is used. Song may be uttered as the third voice.
However, it is preferable that the first voice and the second voice are words. Because a child in childhood does not speak a sentence but only utters words in a row, on the other hand, if the interactive doll with a translation function according to the present invention can translate in a word to word manner, This is because it greatly contributes to traditional foreign language education, that is, introduction to foreign language education.
[0026]
Next, the speech conversion means (translation system module) used in the first embodiment of the interactive doll with a translation function according to the present invention will be described. FIG. 2 is a diagram showing a configuration of the translation system module. As shown in FIG. 2, the translation system module according to the present invention includes a voice receiving unit 2, a voice converting unit 3, and a voice uttering unit 4. The voice conversion unit 3 includes a voice recognition unit 3a, a voice translation unit 3b, and a voice synthesis unit 3c. Although only the voice conversion unit 3 may be a translation system module, the translation system module may be a concept including the voice reception unit 2 and the voice utterance unit 4.
[0027]
Explaining with reference to the example shown in FIG. 1, a first voice "My name is cherry blossoms" is first received by the voice receiving means 2 (for example, a microphone, a recorder, a wireless microphone, etc.). You. The received first speech is sent to the speech recognition and conversion means 3 and is recognized by the speech recognition means 3a, and is translated by the speech translation means 3b into "My name is Sakura." By this, it is synthesized into the second voice "My name is Sakura." Then, the second sound is uttered from a sound generating means (for example, a speaker or the like).
[0028]
Here, various researches and developments have been individually made on the speech recognition, the speech translation, and the speech synthesis, respectively. In the present invention, the speech recognition means, the speech translation means, and the speech synthesis means based on the conventional technique are used. They can be used in combination. However, the concept of converting a first speech based on a first language into a second speech based on a second language is newly discovered by the present invention.
[0029]
As an example, the first voice "My name is cherry blossoms" is received as a voice signal by the voice receiving means 2 and transmitted to the voice converting means 3. In the voice converting means 3, the voice recognizing means 3a converts the voice signal into an electric signal and converts it into a text (vocabulary), for example. Next, the text-converted first voice (first text) is transmitted to the voice translating means 3b.
[0030]
Then, as shown in FIG. 3, the speech translation unit 3b is provided with a storage unit 3b-1 and a selection unit 3b-2. The storage unit 3b-1 can be called a memory or a dictionary, and stores a plurality of vocabularies (and / or voices) based on the second language corresponding to the vocabulary configuring the first voice. For example, a plurality of vocabularies (and / or sounds) may be stored for each of a plurality of languages such as English, German, French, Spanish, and Portuguese.
[0031]
Here, using the example shown in FIG. 1, the vocabulary of the first text "My name is cherry blossoms." Corresponding to "sakura" and "is", for example, as English groups, the vocabulary words "My", "name", "is" and "Sakura" and / or "My", "Naim", "Iz" and The voice "Sakura" is stored in the storage unit 3b-1. Further, as the German group, the vocabulary of “Mine”, “name”, “ist” and “Sakura” and / or the voices of “Main”, “name”, “ist” and “Sakura” are stored.
[0032]
Then, the selecting unit 3b-2 selects a vocabulary and / or a voice based on the second language corresponding to the vocabulary constituting the first voice from the plurality of vocabularies and / or voices. In the example of FIG. 1, “My” is selected in correspondence with “my”, and “name” is selected in correspondence with “name”. Then, “is” is selected corresponding to “is”, and “Sakura” is selected corresponding to “sakura”.
[0033]
Next, the speech synthesizer 3c synthesizes the second speech from the selected vocabulary, combines the speech corresponding to each vocabulary to synthesize a second speech, and the synthesized second speech is a speech. It is uttered from the utterance means 4. When the selecting unit 3b-1 selects a voice corresponding to each vocabulary, the voice synthesizing unit 3c connects the individual voices to synthesize a second voice, and outputs the synthesized second voice to the voice uttering unit. You just have to start from 4.
[0034]
A person skilled in the art can improve the conventional speech recognition means, speech conversion means, and speech synthesis means having the above functions in view of the technical significance of the present invention in the present specification. Although possible, a more preferable voice recognition unit will be described below.
[0035]
In the interactive doll with a translation function according to the present invention, it is preferable that the voice recognition means recognizes the first voice as a phoneme sequence. According to the conventional voice recognition means, since a user (speaker) has a variety of habits such as a tone, a way of speaking, and intonation, the degree of voice recognition is influenced by a difference between users, and the accuracy is reduced. There is a problem. Therefore, many of the conventional voice recognition means require training to make the voice recognition means remember the habit of a specific user. On the other hand, if a method of recognizing a voice as a sequence of phonemes is adopted, the voice can be recognized more accurately even if the user is different.
[0036]
Specifically, the voice recognition means includes:
Audio processor means for receiving an audio signal of the first audio and converting the audio signal into a corresponding electrical signal;
Analog / digital converter means for digitizing the electrical signal at a predetermined sampling rate to form a digitized audio signal;
Means for performing time domain analysis on the subdivided portions of the digitized audio signal to identify a plurality of time domain characteristics of the audio signal, and a plurality of components having predetermined high and low cutoff frequencies; Means for filtering each of the subdivided portions using a filter band and identifying at least one frequency domain characteristic of each of the subdivided portions; and processing the time domain characteristics and frequency domain characteristics. Means for identifying a phoneme included in the audio signal, and preferably includes audio phoneme identification means for identifying the type of the phoneme included in the audio signal.
[0037]
Here, FIG. 4 shows a configuration of a voice recognition means (system) including the voice phoneme identification means.
The voice recognition system 10 shown in FIG. 4 includes an audio processor circuit 14 that converts a voice signal of a first voice received by a voice receiving unit into a corresponding electrical signal. An analog / digital conversion circuit 34 is provided for digitizing the electric signal at a predetermined sampling rate to form a digitized audio signal in order to bring the electric signal into an electric state suitable for digital sampling. The analog / digital conversion circuit 34 receives the electric signal in an analog format, converts the electric signal into a digital format, and transmits the digital signal.
[0038]
The digitized audio signal is then transmitted to the audio identification circuit 16. The voice discrimination circuit 16 programs and analyzes the digitized voice signal and extracts voice characteristics of the voice signal. Then, when necessary voice characteristics are obtained, a specific phoneme included in the voice signal can be identified. The phoneme can be identified without depending on the characteristics of each user (speaker), and can be performed in real time even if the user speaks at a normal conversation speed.
[0039]
The audio identification circuit 16 acquires necessary audio characteristics in two ways. First, a time domain analysis is performed on a plurality of subdivided portions of the digitized audio signal, a plurality of time domain characteristics of the audio signal are identified, and types of phonemes included in the audio signal are identified. The parameter for identifying the type of phoneme included in the audio signal includes, for example, whether the audio is “voiced sound”, “unvoiced sound”, or “silence”.
[0040]
Next, the speech identification circuit 16 filters each of the subdivided portions using a plurality of filter bands having predetermined high and low cutoff frequencies. As a result, a large number of subdivided signals are generated from the audio signal of the first audio having a complicated waveform, and represent a large number of signals representing the waveforms of the individual signals that are components of the audio signal. Then, the voice identification circuit 16 measures each subdivided portion and extracts at least one frequency domain characteristic, for example, various frequency domain data including the frequency and amplitude of the signal.
[0041]
The frequency domain characteristics and the time domain characteristics obtained in this way contain sufficient information to identify phonemes contained in the audio signal. Therefore, the speech identification circuit 16 finally processes the time domain characteristics and the frequency domain characteristics to identify phonemes included in the audio signal.
[0042]
The first speech recognized as described above is then translated by a speech translation unit and a speech synthesis unit incorporated in the speech identification circuit 16 and synthesized into a second speech. In this case, the sequence of phonemes recognized as described above may be translated into a sequence of vocabulary based on the second language. For example, it is possible to perform such translation and speech synthesis by using a language processing program according to the related art.
[0043]
These processes may be controlled by, for example, a control device 22 such as a host computer or a CPU that is connected to the voice identification circuit 16 and that can input, store, and / or control data. As the control device 22, a conventional device can be used, and it is preferable that the control device 22 is built in the voice identification circuit 16. However, it may be omitted depending on the configuration of the voice identification circuit 16.
[0044]
Here, FIG. 5 shows a more detailed configuration of the speech phoneme identification means (system). In the speech recognition system 10 shown in FIG. 5, the first speech received by the speech receiving means 12 is adjusted by the audio processor circuit 14, as in the case of FIG. In the audio processor circuit 14, the audio signal of the first audio is converted into an electric signal and transmitted to the analog / digital converter 34.
[0045]
In the audio processor circuit 14, first, the electric signal is amplified to a suitable level by a signal amplifying means such as an amplifier circuit 26, and the output level is limited by a limiting amplifier circuit 28. Then, the high frequency is removed by the filter circuit 30. Various circuits can be used as the amplifier circuit 26, the limiting amplifier circuit 28, and the filter circuit 30. Next, the analog / digital conversion circuit 34 receives the electric signal in an analog form, converts it into a digital form, and transmits it.
[0046]
Next, the speech recognition system 10 shown in FIG. 4 includes a digital speech processor circuit 18 and a host speech processor circuit 20. These are included in the speech identification circuit 16 shown in FIG. 1 and can be configured with equivalent circuits using programmable devices.
[0047]
First, the digital audio processor circuit 18 receives a digitized audio signal, operates based on a program, and extracts various audio characteristics. Specifically, first, a digitalized audio signal in the time domain is analyzed, and at least one type of time-domain audio characteristic is extracted based on the analysis result. This property is useful for determining whether the audio signal contains "voiced", "voiceless" or "quiet" phonemes.
[0048]
The digital audio processor circuit 18 further operates on the digitized audio signal to obtain various frequency domain information related to the audio signal. This can be done by filtering the audio signal with a myriad of filter bands to generate a corresponding myriad of filtered signals. Digital audio processor circuit 18 measures various characteristics exhibited by the individual waveforms and extracts at least one frequency domain audio characteristic. The frequency domain sound characteristics include the frequency, amplitude, gradient, and the like of the signal component obtained by the filtering process. These properties are stored or stored and are used to determine the type of phoneme contained in the audio signal.
[0049]
As shown in FIG. 5, the digital audio processor circuit 18 includes programmable means for analyzing the digitized audio signal under program control, such as a digital audio processor 36. As the digital audio processor circuit 36, a programmable 24-bit general-purpose digital signal processor such as a Motorola DSP56001 can be suitably used. Of course, other commercially available digital signal processors can be used.
[0050]
The digital audio processor 36 is also connected to various components via a bus type standard address, data and control arrangement 38. These components include, for example, a program memory means for storing a series of programs executed by the DSP 36, such as the DSP program memory 40, a data memory means for storing data used by the DSP 36, such as the DSP data memory 42, and an address. And control logic 44 to perform standard time control functions such as data gating and mapping.
[0051]
Next, the host audio processor circuit 20 will be described. The host audio processor circuit 20 is connected to the digital audio processor circuit 18 via a suitable host interface 52. Generally, host audio processor circuit 20 receives various audio signal characteristic information generated by digital audio processor circuit 18 via host interface 52.
[0052]
The host audio processor circuit 20 includes this information in the audio signal by analyzing this information and comparing the signal characteristics to audio standard audio data collected by testing a representative user (speaker). Identify phoneme types. After identifying the phonemes, the host speech processor circuit 20 uses various language processing techniques to translate the phonemes into vocabularies and phrases based on the first and second languages.
[0053]
The host audio processor circuit 20 preferably includes a second programmable means for analyzing the characteristics of the digitized audio signal under program control, such as the host audio processor 54. The host audio processor 54 may be a programmable 32-bit versatile CPU element such as, for example, a Motorola 68EC030.
[0054]
The host audio processor 54 is also connected to various components via a standard address, data and control bus type array 56. These components include, for example, program memory means for storing a series of programs executed by the host audio processor 54, such as the host program memory 58, and data used by the host audio processor 54, such as the host data memory 60. It includes data memory means and control logic 64 which performs standard time control functions such as address and data gating and mapping.
[0055]
The control device 22 is the same as that described in FIG. The control device 22 may be connected to the host audio processor circuit 20 via an interface means 66 such as an RS-232 interface circuit and the cable 24. Of course, depending on the configuration of the digital audio processor circuit 18 and the host audio processor circuit 20, the control device 22 can be omitted. Note that a memory 62 having a dictionary function and a display 68 can be further connected to the host audio processor circuit 20.
[0056]
As described above, when the speech recognition means (a) recognizes the first speech by phoneme, the speech translation means (b) translates the recognized phoneme string into a vocabulary string based on a second language. It is effective that the speech synthesizer (c) synthesizes the second speech by computer processing the vocabulary string. However, if the speech recognition system as shown in FIGS. 4 and 5 is used, the speech recognition means (a) can have the functions of the speech translation means (b) and the speech synthesis means (c).
[0057]
Here, a conventional voice synthesizing means (c) can be used. However, according to the conventional voice synthesizing means, since the voice is synthesized electrically and mechanically, the interval between vocabulary words is different. And the intonation may not be perfect, and the second voice uttered may be uncomfortable with the voice uttered by a human. Therefore, in the present invention, it is preferable to use the following for the speech synthesis means.
[0058]
That is, the voice synthesis means (c) receives a vocabulary sequence based on the second language, converts the vocabulary sequence into a first phoneme sequence, a voice conversion subsystem, receives a transformation rule, and A speech deformer applied to the phoneme string of the second phoneme string to form a second phoneme string, an evaluator for ranking phonemes included in the second phoneme string based on a predetermined criterion, and the second phoneme string And a syllable decomposer that decomposes phonemes included in the second phoneme sequence into syllables using the ranking.
[0059]
Further, in the interactive doll with a translation function according to the present invention, the voice receiving means receives a keyword based on a first language by the user, the voice recognition means (a) recognizes the keyword, and A voice uttering unit utters a question based on a second language stored in advance corresponding to the keyword, and then converts the first voice by the user into the second voice, Alternatively, it is preferable to have a function of generating the third voice to make the third user interact with the user.
Those skilled in the art can appropriately create a program for such a function and incorporate the program into the speech recognition unit, the speech translation unit, and the speech synthesis unit.
[0060]
Further, the voice recognition means (a) recognizes a specific part of the first voice, and the voice translation means (b) translates the specific part into voice based on a second language. ), It is also effective to apply the result of the translation to a voice answer pattern based on a second language stored in advance corresponding to the question and synthesize the second voice.
[0061]
With this configuration, the voice conversion unit converts the first text constituting the first voice in the first language into the second text in the second language based on a so-called pattern translation method. be able to. The pattern translation method is insufficient for translating a long sentence, but is effective for processing a short sentence. Therefore, it is effective for children in childhood, which is an important period for elementary foreign language education.
[0062]
In this case, the speech translating means (b) calculates (b-1) the speech answer pattern and a plurality of vocabularies based on the second language expected as the vocabulary constituting the specific part corresponding to the question. Storage means for storing; and (b-2) selecting means for selecting a vocabulary based on a second language corresponding to the vocabulary constituting the specific part recognized by the voice recognizing means (a); The synthesizing means (c) may synthesize speech of a vocabulary based on the selected second language, apply the speech to the speech answer pattern, and synthesize the second speech.
[0063]
【The invention's effect】
According to the present invention, a voice translation system is provided in a doll, and a voice uttered by a user is converted into a voice based on a different language and uttered, thereby satisfying the user's playfulness and learning a foreign language conversation. It is possible to provide an interactive doll with a translation function, which can be provided to the user. That is, the present invention can provide an interactive doll with a translation function that has not existed in the past.
[Brief description of the drawings]
FIG. 1 is a diagram for schematically explaining functions of an interactive doll with a translation function according to the present invention.
FIG. 2 is a diagram showing a configuration of a speech conversion means (translation system module) used in the present invention.
FIG. 3 is a diagram showing a configuration of another speech conversion means (translation system module) used in the present invention.
FIG. 4 is a diagram showing a configuration of a voice recognition means (system) including a voice phoneme identification means used in the present invention.
FIG. 5 is a diagram showing in more detail the configuration of a speech recognition means (system) including a speech phoneme identification means used in the present invention.

Claims (12)

A first voice based on the first language by the user is received and converted to a second voice based on the second language, and then the converted second voice, or corresponding to the first voice. An interactive doll with a translation function, characterized by uttering a third voice based on a second language stored in advance to communicate with the user. The interactive doll with a translation function according to claim 1,
An ear portion having audio receiving means for receiving the first audio,
A portion having a voice conversion means for converting the first voice into the second voice, and a mouth portion having a voice uttering means for generating the second voice or the third voice. Interactive doll with translation function. 3. The interactive doll with a translation function according to claim 1, wherein the first voice and the second voice are words. An interactive doll with a translation function according to claim 2,
The voice conversion means,
(A) voice recognition means for recognizing the first voice,
(B) voice translation means for translating the recognized first voice into the second language, and (c) voice synthesis means for synthesizing the voice based on a result of the translation. Interactive doll with translation function. An interactive doll with a translation function according to claim 4,
An interactive doll with a translation function, wherein the voice recognition means (a) recognizes the first voice as a phoneme sequence. An interactive doll with a translation function according to claim 5,
The voice recognition means (a)
Audio processor means for receiving an audio signal of the first audio and converting the audio signal into a corresponding electrical signal;
Analog-to-digital converter means for digitizing the electrical signal at a predetermined sampling rate to form a digitized audio signal; and performing time domain analysis on the subdivided portions of the digitized audio signal; Means for identifying a plurality of time domain characteristics of the audio signal, and a plurality of filter bands having predetermined high and low cutoff frequencies, filtering each of the subdivided portions, And voice phoneme identification means including means for identifying at least one frequency domain characteristic of each portion, and means for processing the time domain characteristics and the frequency domain characteristics to identify phonemes contained in the audio signal. An interactive doll with a translation function. An interactive doll with a translation function according to claim 5 or 6,
An interactive doll with a translation function, wherein the voice translation means (b) translates the recognized phoneme sequence into a vocabulary sequence based on a second language. An interactive doll with a translation function according to claim 7,
An interactive doll with a translation function, wherein the speech synthesis means (c) synthesizes the second speech by computer processing the vocabulary string. An interactive doll with a translation function according to claim 8,
The voice synthesis means (c) comprises:
A speech conversion subsystem that receives a vocabulary sequence based on the second language and converts the vocabulary sequence into a first phoneme sequence;
Receiving a transformation rule and applying it to the first phoneme sequence to form a second phoneme sequence,
An evaluator that ranks phonemes included in the second phoneme string based on a predetermined criterion, and receives the second phoneme string, and is included in the second phoneme string using the ranking. An interactive doll with a translation function, comprising a syllable decomposer for decomposing phonemes into syllables. An interactive doll with a translation function according to any one of claims 1 to 9,
The voice receiving means receives a keyword based on a first language by the user, the voice recognizing means (a) recognizes the keyword, and the voice uttering means stores in advance a second keyword corresponding to the keyword. Ask a question based on your language,
Thereafter, the first voice by the user is converted into the second voice, and the second voice or the third voice is uttered to interact with the user. With interactive doll. The interactive doll with a translation function according to claim 10,
The voice recognition means (a) recognizes a specific portion of the first voice,
The voice translation means (b) translates the specific part into voice based on a second language,
A translation function, wherein the speech synthesis means (c) applies the result of the translation to a speech answer pattern based on a second language stored in advance corresponding to the question and synthesizes the second speech. With interactive doll. The interactive doll with a translation function according to claim 11,
The voice translation means (b)
(B-1) storage means for storing the voice response pattern and a plurality of vocabulary words based on a second language expected as vocabulary words constituting the specific part corresponding to the question; and (b-2). Selecting means for selecting a vocabulary based on a second language corresponding to the vocabulary constituting the specific part recognized by the voice recognition means (a);
The speech synthesis means (c) synthesizes speech of a vocabulary based on the selected second language, applies the speech to the speech answer pattern, and synthesizes the second speech. Interactive doll.

Images