JP2007041443A - Device, program, and method for speech conversion - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To protect individual's privacy and to perform conversion into speech data with which the meaning of a conversation can be understood. <P>SOLUTION: A device for speech conversion is equipped with a speech data input section 11 which inputs speech data, a substituted word detection section 15 which detects a word part including a predetermined word to be substituted from the input speech data, and a synchronization section 21 which substitutes alternative speech data corresponding to the word part of the input speech data for the word part. <P>COPYRIGHT: (C)2007,JPO&INPIT

Description

  The present invention relates to a voice conversion device, a voice conversion program, and a voice conversion method, and more particularly, to a voice conversion device, a voice conversion program, and a voice conversion method for converting voice into a voice that protects privacy.

  In recent years, with the enforcement of the Personal Information Protection Law, there is a growing tendency to protect human privacy. This tendency is particularly high in the medical field. On the other hand, in the medical field, attempts are being made to record the conversations of doctors, nurses, and patients in order to reduce medical errors. However, recording the conversation during diagnosis or treatment as it is cannot protect the patient's privacy. This is because the recorded voice includes the name identifying the patient and the name of the patient's disease. For this reason, Japanese Patent Application Laid-Open No. 2004-357294 describes an apparatus that suppresses playback of a restricted keyword if the restricted keyword is included in the voice. Specific examples of the inhibition of reproduction include conversion of reproduction sound by interfering sound or skipping the restriction keyword itself.

However, when the recorded voice is played back, if the playback of the restricted keyword is suppressed, there is a problem that the contents of the conversation cannot be understood. For example, if there is a sentence “Yamada is lung cancer” that includes the patient's name and disease name, and the restriction keywords “Yamada” and “lung cancer” exist, the reproduced sentence is “??? Is that? " “?” Indicates a disturbing sound. For this reason, for example, “san” can recognize that he is a person, but it is unknown whether he is a doctor, a nurse, or a patient. Also, from "I", you can see that you are affirming something, but you are not sure what you are affirming. For this reason, there is a problem that the subject and the predicate are not understood and the meaning of reproducing the sound is lost.
JP 2004-357294 A

  The present invention has been made to solve the above-mentioned problems, and one of the objects of the present invention is to protect the privacy of the individual and to convert the voice data that can be understood into the meaning of the conversation. It is to provide a conversion device.

  Another object of the present invention is to provide a voice conversion program and a voice conversion method capable of converting personal voice into voice data that can understand the meaning of conversation while protecting personal privacy.

  In order to achieve the above-described object, according to one aspect of the present invention, a speech conversion device includes speech data input means for inputting speech data, and a word including a predetermined replacement target word for the input speech data. Replacement means for replacing the portion with alternative speech data corresponding to the word portion.

  According to the present invention, a word portion including a predetermined replacement target word in the audio data is replaced with the corresponding alternative audio data. For this reason, for example, if the replacement target word is a word that can specify an individual such as a name and address, the portion that specifies the individual from the voice data can be replaced with alternative voice data, thus protecting the privacy of the individual. be able to. Further, since the replacement target word is replaced with an alternative word without being deleted at all, the meaning of the conversation can be understood. As a result, it is possible to provide an audio conversion device capable of protecting personal privacy and converting audio data that can understand the meaning of conversation.

  Preferably, a replacement level input means for inputting a replacement level is further provided, and the replacement means includes a replacement target word determining means for determining a replacement target word according to the input replacement level.

  According to the present invention, it is possible to cope with the fact that the range of information that can be notified varies depending on the listener.

  Preferably, the apparatus further comprises an alternative word table storage unit that stores an alternative word table that defines an alternative word for each replacement level corresponding to one word of the replacement target word, and the replacement unit uses the alternative word table as an input. It further includes alternative voice data generation means for generating alternative voice data corresponding to the replaced level.

  According to the present invention, it is possible to cope with the difference in the range of information that can be notified depending on the listener.

  Preferably, the apparatus further comprises level input means for inputting a replacement level, and an alternative word table storage means for storing an alternative word table that defines an alternative word for each replacement level corresponding to one word of the replacement target word, The substitution means includes substitution voice data generation means for generating substitution voice data corresponding to the inputted substitution level using the substitution word table.

  According to the present invention, it is possible to cope with the difference in the range of information that can be notified depending on the listener.

  Preferably, the replacement means corresponds to the speech recognition means for recognizing the input speech data and outputting the text data, the replacement target word detection means for extracting the replacement target word from the text data, and the replacement target word. An alternative word table storage unit that stores an alternative word table that defines an alternative word, an alternative word determination unit that determines an alternative word corresponding to the extracted word, and an alternative voice data by synthesizing the determined alternative word by voice Voice generation means for generating, and voice data replacement means for replacing the word portion of the input voice data with alternative voice data.

  According to another aspect of the present invention, a speech conversion program includes a step of inputting speech data, and a substitute speech data corresponding to a word portion including a word portion including a predetermined replacement target word of the input speech data. And causing the computer to execute.

  According to the present invention, it is possible to provide a voice conversion program capable of protecting personal privacy and converting voice data that can understand the meaning of conversation.

  According to still another aspect of the present invention, a speech conversion method includes a step of inputting speech data, and an alternative speech corresponding to a word portion including a predetermined replacement target word of the input speech data. And replacing with data.

  According to the present invention, it is possible to provide a voice conversion method capable of protecting personal privacy and converting voice data that can understand the meaning of conversation.

  Hereinafter, embodiments of the present invention will be described with reference to the drawings. In the following description, the same parts are denoted by the same reference numerals. Their names and functions are also the same. Therefore, detailed description thereof will not be repeated.

  FIG. 1 is a block diagram showing a hardware configuration of an audio conversion device according to one embodiment of the present invention. Referring to FIG. 1, the voice conversion device 100 is configured by a general personal computer. The voice conversion device 100 loads a central processing unit (CPU) 101, each of which is connected to the bus 120, a ROM (Read Only Memory) 103 in which a program to be executed by the CPU 101 is recorded, and a program to be executed. A RAM (Random Access Memory) 105 for storing data during program execution, a hard disk drive (HDD) 107 for storing data in a nonvolatile manner, and a card interface on which an IC card 108 is mounted ( I / F) 109, an operation unit 111 and a display unit 113 that serve as an interface with a user, a communication I / F 115 for connecting the audio conversion device 100 to a network 118, and an audio processing circuit 116.

  The CPU 101 loads the voice conversion program recorded in the ROM 103 into the RAM 105 and executes it. The CPU 101 may load the voice conversion program recorded on the IC card 108 attached to the card I / F 109 into the RAM 105 and execute it. Further, a voice conversion program may be downloaded from another computer connected to the network and stored in the HDD 107, and the voice conversion program may be loaded into the RAM 105 and executed. The voice conversion program here includes not only a program directly executable by the CPU 101 but also a program in a source program format, a compressed program, an encrypted program, and the like.

  The operation unit 111 is an input device such as a keyboard or a mouse. The display unit 113 is a liquid crystal display device, a cathode ray tube (CRT), or a plasma display panel. The communication I / F 115 is a communication interface for connecting the audio conversion device 100 to the network 118. As a result, the audio conversion device 100 can communicate with other computers connected to the network 118. The network 118 may be a local area network (LAN) or a wide area network (WAN) such as the Internet. Further, it may be connected directly to another computer via a serial line or parallel line, or may be connected via a telephone line via a modem. The network 118 may be wired or wireless.

  The audio processing circuit 116 is connected to the microphone 117 and the speaker 119. The microphone 117 collects human voice and outputs voice data to the voice processing circuit 116. The audio processing circuit 116 includes an amplification circuit that amplifies audio data of an analog signal, a noise removal circuit that removes noise, an A / D (analog / digital) converter for converting the digital signal into an analog signal, And a D / A (digital / analog) converter. The audio processing circuit 116 amplifies the analog signal audio data input from the microphone 117, removes noise, converts the analog signal into a digital signal, and outputs the digital signal audio data to the CPU 101. In addition, the audio processing circuit 116 converts the digital signal audio data input from the CPU 101 into an analog signal, amplifies it, and then outputs the analog signal audio data to the speaker 119 to generate audio.

  Note that the recording medium for recording the voice conversion program is not limited to the IC card 108, but a flexible disk, a cassette tape, an optical disk (MO (Magnetic Optical Disc / MD (Mini Disc) / DVD (Digital Versatile Disc)), an optical card, It may be a medium that carries a fixed program such as a semiconductor memory such as a mask ROM or EPROM (Erasable Programmable ROM).

  FIG. 2 is a functional block diagram showing an outline of the functions of the CPU 101 of the voice conversion device. Referring to FIG. 2, CPU 101 recognizes voice data input from voice input unit 11 connected to voice processing circuit 116 that processes voice from microphone 117, and converts it into text data. A speech recognition unit 13 that performs replacement, a replacement target word detection unit 15 that detects a replacement target word that is a replacement target from text data, and an alternative word determination unit 17 that determines a replacement word for replacing the detected replacement target word. A speech synthesizer 19 that generates speech data of alternative words, a synchronization unit 21 that replaces the replacement target word portion of the speech data with the replacement word speech data, and a replacement word that associates the replacement target word with the replacement word A replacement word list generation unit 25 for generating a list and a replacement level input unit 23 instructed from the operation unit 111 are included.

  Voice data of a person's conversation collected by the microphone 117 is input to the voice processing circuit 116. The audio processing circuit 116 converts the input audio data into audio data of a digital signal and outputs it to the audio input unit 11. The voice input unit 11 outputs the voice data input from the voice processing circuit 116 to the voice recognition unit 13 and the synchronization unit 21. The voice recognition unit 13 is connected to the voice input unit 11 and the replacement target word detection unit 15. The speech recognition unit 13 recognizes the speech data input from the speech input unit 11 and converts it into text data, and outputs the text data to the replacement target word detection unit 15.

  Here, an example in which audio data is input from the audio processing circuit 116 to the audio input unit 11 is shown, but audio data recorded on the IC card 108 is read out by the card I / F 109 and the audio input unit 11 is read out. May be entered. In this case, the recording medium on which the audio data is recorded is not limited to the IC card 108, but may be a magnetic recording device such as the HDD 107, a magneto-optical disk such as a CD-ROM, or an EPROM. The semiconductor memory may be used.

  The replacement level input unit 23 is connected to the operation unit 111, and acquires a replacement level based on information input from the operation unit 111. Here, the replacement level refers to the qualification given to medical personnel such as doctors, nurses, pharmacists, and other non-medical personnel such as patients, family members, acquaintances, etc. The replacement level of the person is “1”, and the replacement level of the other person is “2”. The replacement level input unit 23 has a user authentication function in order to grasp who the user of the voice conversion device 100 is. Specifically, the voice conversion device 100 stores in advance in the HDD 107 user information that associates user identification information for identifying a user, a password, and a replacement level. The qualifications here are doctors, nurses and patients. When the user inputs user identification information and a password to the operation unit 111, the replacement level input unit 23 extracts user information including the input user identification information and password from the HDD 107. If user information including the input user identification information and password is extracted, the replacement level is acquired from the user information. If such user information is not extracted, the replacement level is set to “2”. Although an example in which user authentication is performed using user identification information and a password has been described here, biometric authentication may be used. Alternatively, the user may directly input the replacement level by operating the operation unit 111 without performing user authentication. The replacement level input unit 23 outputs the replacement level to the replacement target word detection unit 15 and the alternative word determination unit 17.

  The replacement target word detection unit 15 receives text data from the speech recognition unit 13 and receives a replacement level from the replacement level input unit 23. The replacement target word detection unit 15 is connected to the ROM 103 and reads an alternative word table 35 stored in the ROM 103.

  Here, the alternative word table 35 will be described. FIG. 3 is a diagram illustrating an example of an alternative word table. The replacement word table 35 defines replacement target words and replacement words for replacing the replacement target words. The replacement target word differs depending on the replacement level, and the alternative word corresponding to one replacement target word differs depending on the replacement level. More specifically, in the alternative word table 35, the replacement target word is defined by a predetermined type of word. In the present embodiment, a predetermined type of word is a person name, a telephone number, a disease name (special) drug name (special), a disease name (general), a drug name (general), a test name, and a treatment name.

  Personal names include doctors, nurses, patients and others. A disease name (special) is a name of an odd disease, and a disease name (general) is a disease name of a disease that generally affects many people such as wind and diabetes. Since there are only a few patients with strange illnesses, the patient name can be identified by identifying the disease name. Therefore, the disease name (special) is classified into a different type from the disease name (general). The drug name (special) is the name of a drug used for the treatment of a strange disease, and the drug name (general) is the name of a drug that is generally used for the treatment of such a disease by many people. Since the patient can be identified by listening to the medicine name (special), the medicine name (special) is classified into a different type from the medicine name (general).

  The alternative word table 35 is defined by changing the replacement target word according to the replacement level. At the replacement level “1”, the replacement target words are a person name, a disease name (special), a drug name (special), and a telephone number, whereas at the replacement level “2”, the replacement target word is a person name, a disease name ( Special), drug name (special), telephone number, disease name (general), drug name (general), examination name and treatment name.

  The alternative word table 35 defines different alternative words corresponding to one replacement target word depending on the replacement level. For example, the replacement word corresponding to the replacement target word “doctor” is a number obtained by adding a number to “Dr” at the replacement level “1”, and only “Dr” at the replacement level “2”. At the replacement level “1”, different alternative words are defined for each doctor. At the replacement level “2”, the same alternative word is defined for all doctors. The replacement word corresponding to the replacement target word “nurse” is “Ns” plus years of experience at the replacement level “1”, and only “Ns” at the replacement level “2”. At the replacement level “1”, different alternative words are defined by nurses with different years of experience, and the years of experience are known from the alternative words. At the replacement level “2”, the same alternative word is defined for all nurses. The substitution word corresponding to the substitution target word “patient” is “Pt” added with a care degree at the substitution level “1”, and only “Pt” at the substitution level “2”. At the replacement level “1”, different alternative words are defined for each patient, and the degree of care is known from the alternative words. At substitution level “2”, the same alternative word is defined for all patients.

  Returning to FIG. 2, the replacement target word detector 15 determines a replacement target word based on the replacement level and the alternative word table. That is, the replacement target word is acquired from the replacement word table and the replacement level, and the replacement target word is extracted by scanning the text data. As described above, the replacement level “1” has a smaller number of replacement target words than the replacement level “2”. Then, the replacement target word detection unit 15 outputs the position of the replacement target word in the text data to the synchronization unit 21 and outputs the replacement target word to the replacement word determination unit 17.

  The replacement word determination unit 17 receives the replacement target word of the text data from the replacement target word detection unit 15 and receives the replacement level from the replacement level input unit 23. The alternative word determination unit 17 is connected to the ROM 103 and reads the alternative word table 35 stored in the ROM 103. The alternative word determination unit 17 determines an alternative word corresponding to the replacement target word based on the replacement level and the replacement table. That is, an alternative word corresponding to the replacement target word is determined from the alternative word table and the replacement level. As described above, in the case of the replacement level “1”, the number of alternative words corresponding to one replacement word is larger than that in the case of the replacement level “2”.

  The voice conversion device 100 according to the present embodiment stores a personal name table that associates personal names and qualifications in the HDD 107 in advance. This personal name table further associates the years of experience in the case of a nurse and the degree of care in the case of a patient. When the replacement target word detection unit 15 receives the name of the replacement target word, the replacement word determination unit 17 determines whether the replacement word determination unit 17 is a doctor, a nurse, or a patient. In the case of a nurse, the years of experience are judged, and in the case of a patient, the degree of care is judged. For example, a case where the qualifications of the person name “Tanaka” and the person name “Yamada” are both “doctors” and the person name “Tanaka” and the person name “Yamada” are determined as replacement target words will be described. When the replacement level is “1”, the alternative word determination unit 17 determines the alternative word for the personal name “Tanaka” as “Dr1” and the alternative word for the personal name “Yamada” as “Dr2”. In the case of the replacement level “2”, the alternative word determination unit 17 determines the alternative word for the personal name “Tanaka” as “Dr” and the alternative word for the personal name “Yamada” as “Dr”. The alternative word determination unit 17 outputs the alternative word to the speech synthesizer 19, and outputs the combination of the replacement target word and the alternative word to the replacement word list generation unit 25.

  An alternative word is input from the alternative word determination unit 17 to the speech synthesizer 19. The voice synthesizer 19 synthesizes an alternative word of text data as voice and converts it into voice data. Then, the voice synthesizer 19 outputs the voice data to the synchronizer 21.

  The synchronizer 21 receives the speech data of the substitute word from the speech synthesizer 19, the speech data from the speech input unit 11, and the position information in the text data of the replacement target word from the replacement target word detector 15. Is done. The synchronization unit 21 replaces a portion corresponding to the replacement target word in the voice data input from the voice input unit with the voice data of the alternative word. The synchronization unit 21 acquires the position of the speech data corresponding to the replacement target word in the speech data from the position information of the text data of the replacement target word, and replaces the speech data at that position with the speech data of the alternative word. Then, the synchronization unit 21 outputs the audio data replaced with the alternative word to the audio processing circuit 116 and the HDD 107. Thereby, the audio data in which the replacement target word is replaced with the alternative word is output from the speaker 119. Also, the HDD 107 stores the audio data 33 in which the replacement target word is replaced with an alternative word.

  The replacement word list generation unit 25 receives a combination of a replacement target word and a replacement word from the replacement word determination unit 17. The replacement word list generation unit 25 generates a replacement word list in which the replacement target word and the replacement word are associated with each other from the combination of the replacement target word and the replacement word, and stores the replacement word list in the HDD 107. As a result, the replacement word list 31 is stored in the HDD 107.

  FIG. 4 is a flowchart showing an example of the flow of replacement processing executed by the speech conversion apparatus 102. Referring to FIG. 4, in the replacement process, it is determined whether or not a replacement level has been input (step S01). If a replacement level is input, the process proceeds to step S02, and if not, a standby state is entered. In step S02, the process waits until voice data is input (NO in step S02). When voice data is input, the process proceeds to step S03. That is, the replacement process is a process executed on condition that a replacement level and audio data are input. The input voice data is recognized as voice and converted to text data (step S03). Then, text data is acquired (step S04).

  In the next step S05, the text of the text data is sequentially scanned to detect a replacement target word. Specifically, a replacement target word is acquired from the replacement level input in step S01 and the replacement word table 35 read from the ROM 103, and it is determined whether the text is the replacement target word acquired. If the text is a replacement target word, the process proceeds to step S06 assuming that the replacement target word is detected, and the process proceeds to step S11 because it is not detected if it is not a replacement target word.

  In step S06, an alternative word corresponding to the replacement target word detected in step S05 is determined from the replacement level input in step S01 and the alternative word table read from the ROM 103. Then, it is determined whether or not the replacement level input in step S01 is “1” (step S07). If yes, the process proceeds to step S08, and if no, the process skips step S08 and proceeds to step S09. . In step S08, the combination of the replacement target word and the alternative word is added to the replacement word list 31 stored in the HDD 107, and the process proceeds to step S09.

  In step S09, the alternative word determined in step S06 is synthesized with speech to generate speech data (step S09). Then, the portion corresponding to the replacement target word detected in step S05 in the speech data input in step S02 is replaced with the substitute word speech data generated in step S09 (step S10).

  Next, it is determined whether or not the text to be scanned exists in the text data (step S11). If it exists, the process returns to step S05, and if not, the process proceeds to step S12. In step S12, it is determined whether or not the replacement level input in step S01 is “1”. If yes, the process proceeds to step S13, and if no, the process skips step S13 and proceeds to step S14. In step S13, the replacement word list stored in HDD 107 is output. The output may be from the display unit 113 or may be printed on a recording medium such as paper by a printer.

  In step S14, the voice data replaced with the alternative word voice data in step S10 is output. In the case of the audio processing circuit 116, the audio data is output from the speaker. In the case of the HDD 107, the audio data 33 is stored in the HDD 107.

  FIG. 5 is a diagram showing voice data input to the voice conversion device in text. In the figure, the left column shows the speakers of the conversation shown in the right column. Dr1 indicates the qualification “doctor” of the name “Tanaka”, Ns5 indicates the qualification “nurse” with the name “ya”, and Ns2 indicates the qualification “2 years of experience” with the name “Meito” PtD3 indicates the qualification “patient with nursing degree D3” of the name “Kuwahara”.

  FIG. 6 is a diagram showing, as text, the voice data after the voice data shown in FIG. 5 is voice-converted at the replacement level “1”. The left column in the figure is the same as that shown in FIG. FIG. 7 is a diagram showing a replacement word list generated corresponding to the voice data shown in FIG. FIG. 8 is a diagram showing text data of the voice data after the voice data shown in FIG. 5 is voice-converted at the replacement level “2”.

  Referring to FIG. 5, FIG. 6, and FIG. 8, the number of words replaced with alternative words is larger in the voice data shown in FIG. 8 than in the voice data shown in FIG. This is because the replacement word is made different in the replacement word table 35 according to the replacement level. Further, there are more types of alternative words in the voice data shown in FIG. 6 than alternative words in the voice data shown in FIG. This is because, in the alternative word table 35, the alternative words corresponding to the replacement target word are made different according to the replacement level. Furthermore, since the replacement word list shown in FIG. 7 is shown for the voice data of FIG. 6, medical personnel who listen to the voice data while looking at the replacement word list can reproduce the voice data before conversion.

<Modification>
In the speech conversion apparatus 100 described above, speech data in which a portion corresponding to a replacement target word in the input speech data is replaced with speech data of an alternative word is output. The speech conversion apparatus 100 according to the modification does not output the speech data replaced with the speech data of the alternative word, but outputs the text data in which the replacement target word is replaced with the alternative word in the text data. .

  FIG. 9 is a functional block diagram showing an overview of the functions of the CPU of the speech conversion apparatus according to the modification. Referring to FIG. 9, the difference from the functional block diagram shown in FIG. 2 is that a replacement unit 27 is included instead of speech synthesis unit 19 and synchronization unit 21. Other configurations are the same as those shown in FIG. 2, and therefore, description thereof will not be repeated here. Referring to FIG. 9, the replacement unit 27 receives the replacement word from the replacement word determination unit 17, receives the text data from the speech recognition unit 13, and includes the replacement target word detection unit 15 in the text data of the replacement target word. Position information is input. The replacement unit 27 replaces a portion corresponding to the replacement target word in the text data with an alternative word. Then, the replacement unit 27 stores the text data replaced with the alternative word in the HDD 107. As a result, the replaced text data 37 is stored in the HDD 107.

  FIG. 10 is a flowchart illustrating an example of the flow of replacement processing executed by the speech conversion apparatus according to the modification. Referring to FIG. 10, the difference from the replacement process shown in FIG. 4 is that step S21 is executed instead of steps S09 and S10, and step S22 is executed instead of step S14. The other processes are the same as those shown in FIG. 4, and therefore description thereof will not be repeated here. In step S21, the replacement target word detected in step S05 in the text data acquired in step S04 is replaced with the alternative word determined in step S06.

  In step S22, the text data replaced with the alternative word in step S21 (replaced text data) is stored in the HDD 107.

  As described above, the speech conversion apparatus 100 according to the present embodiment replaces the word portion including the replacement target word in the speech data with the speech data obtained by speech synthesis of the alternative word defined in the alternative word table 35. For this reason, for example, if the replacement target word is a word that can specify an individual such as a name or address, the portion that specifies the individual from the voice data is replaced with the voice data of the alternative word. Can be protected. Further, since the replacement target word is replaced with an alternative word without being deleted at all, the meaning of the conversation can be understood.

  In addition, since the replacement target word is made different according to the replacement level, it is possible to cope with the range of information that can be notified depending on the listener. For example, medical personnel can protect patient privacy by substituting replacement words with alternative words, but with as few replacement words as possible and listening to the doctor-nurse conversation as closely as possible. Can do. On the other hand, the patient's privacy can be protected as much as possible by letting a person who is not related to medical care hear the voice in which as many replacement words as possible are replaced with alternative words.

  Further, since the alternative word table defines alternative words for each replacement level corresponding to one replacement target word, it is possible to cope with the range of information that may be notified by the listener. For example, by replacing the replacement target word with a plurality of alternative words, the medical staff can protect the privacy of the patient and can hear the conversation between the doctor and the nurse as detailed as possible. More specifically, the difference in drug names can be understood only from conversation. Conversely, for a person who is not related to medical care, the patient's privacy can be protected as much as possible by listening to the voice in which one replacement word is replaced with a single replacement word. According to the alternative word, since the difference between the replacement target words can be distinguished, it is possible to distinguish whether it is a person name or a medicine name by listening to the conversation.

  In the present embodiment, the voice conversion device 100 has been described. However, the invention can be understood as a voice conversion program or a voice conversion method for causing a computer to execute the replacement process shown in FIG. 4 or FIG. Needless to say.

  The embodiment disclosed this time should be considered as illustrative in all points and not restrictive. The scope of the present invention is defined by the terms of the claims, rather than the description above, and is intended to include any modifications within the scope and meaning equivalent to the terms of the claims.

It is a block diagram which shows the hardware constitutions of the speech converter in one of the embodiments of this invention. It is a functional block diagram which shows the outline | summary of the function of CPU of a speech converter. It is a figure which shows an example of an alternative word table. It is a flowchart which shows an example of the flow of the replacement process performed with a speech converter. It is a figure which shows the audio | voice data input into an audio | voice conversion apparatus with a text. It is a figure which shows the audio | voice data after carrying out the audio | voice conversion of the audio | voice data shown in FIG. 5 by substitution level "1" with a text. It is a figure which shows the replacement word list | wrist produced | generated corresponding to the audio | voice data shown in FIG. It is a figure which shows the audio | voice data after carrying out the audio | voice conversion of the audio | voice data shown in FIG. 5 by substitution level "2" with a text. It is a functional block diagram which shows the outline | summary of the function of CPU of the audio | voice conversion apparatus in a modification. It is a flowchart which shows an example of the flow of the replacement process performed with the audio | voice conversion apparatus in a modification.

Explanation of symbols

  DESCRIPTION OF SYMBOLS 11 Speech input part, 13 Speech recognition part, 15 Replacement object word detection part, 17 Alternative word determination part, 19 Speech synthesis part, 21 Synchronization part, 23 Replacement level input part, 25 Replacement word list generation part, 27 Replacement part, 31 Replacement word list, 33 Voice data, 35 Substitution word table, 37 Text data after replacement, 100 Voice conversion device, 101 CPU, 103 ROM, 105 RAM, 107 HDD, 108 IC card, 109 card I / F, 111 Operation unit , 113 display unit, 115 communication I / F, 116 sound processing circuit, 117 microphone, 119 speaker, 120 bus.

Claims (7)

Voice data input means for inputting voice data;
A speech conversion apparatus comprising: replacement means for replacing a word portion including a predetermined replacement target word in the input speech data with alternative speech data corresponding to the word portion. It further comprises a replacement level input means for inputting a replacement level,
The speech conversion apparatus according to claim 1, wherein the replacement unit includes a replacement target word determination unit that determines the replacement target word according to the input replacement level.
Make different An alternative word table storage unit that stores an alternative word table that defines an alternative word for each replacement level corresponding to one word of the replacement target word;
The speech conversion apparatus according to claim 2, wherein the replacement means further includes alternative speech data generation means for generating alternative speech data according to the input replacement level using the alternative word table. Level input means for inputting a replacement level;
An alternative word table storage unit that stores an alternative word table that defines alternative words for each of the replacement levels corresponding to one word of the replacement target word;
The speech conversion apparatus according to claim 1, wherein the replacement unit includes a substitute speech data generation unit that generates substitute speech data according to the input replacement level using the substitute word table. The replacement means includes voice recognition means for voice recognition of the input voice data and outputting text data;
Replacement target word detecting means for extracting the replacement target word from the text data;
An alternative word table storage means for storing an alternative word table defining alternative words corresponding to the replacement target words;
Alternative word determining means for determining an alternative word corresponding to the extracted word;
Speech synthesis means for synthesizing the determined alternative word to generate alternative speech data;
The speech conversion apparatus according to claim 1, further comprising: speech data replacement means for replacing the word portion of the input speech data with the alternative speech data. Inputting audio data;
A speech conversion program for causing a computer to execute a step of replacing a word portion including a predetermined replacement target word in the input speech data with alternative speech data corresponding to the word portion. Inputting audio data;
Replacing a word part including a predetermined replacement target word in the input voice data with alternative voice data corresponding to the word part.

Images