JP4964574B2 - Information processing apparatus and method for registering speech reading vocabulary - Google Patents

Description

  The present invention relates to an information processing apparatus and a method for registering a speech-to-speech vocabulary. More specifically, the present invention relates to a text-to-speech (TTS (Text To) text information including kana-kanji characters such as e-mail (E-mail) texts and news texts. The present invention relates to a technique adapted to register a vocabulary (text) for reading aloud in an information processing apparatus equipped with a speech engine.

  2. Description of the Related Art Conventionally, in-vehicle devices such as an in-vehicle navigation device are equipped with a function that enables various voice guidance by voice output. For example, voice guidance such as “Returning a parrot” that utters the result of voice recognition, “Talkback” that informs by voice when a task is executed according to the operation result, for example, “Set the scale to 100 m” There is. Also, text information mixed with kana and kanji such as e-mail texts and news texts can be read aloud (voice guidance) by using a speech synthesis technology using a TTS engine. These voice guidance (“Parrot return”, “Talkback”, “Reading out e-mail texts and news texts”) can be made by methods such as “Recording voice”, “TTS engine”, “Recording voice and TTS engine”, respectively. It has been realized.

  The TTS engine is a system (conversion software) that reads out Japanese texts by voice when it is entered, and is widely used in mobile phone answering services. Recently, it is a portable service for individuals, and it is close to the user. Even in the absence of a computer, it is used in voice services that read out information such as emails, news, and market trends when you make a call. The TTS engine is for converting the input text into speech and reading it out, but it is practically impossible to read the text mixed with kana and kanji 100% accurately. In order to alleviate such inconvenience, the current TTS engine is provided with a user dictionary, and the user registers kanji “reading” in the user dictionary so that kanji that could not be read can be read correctly.

  FIG. 1 shows an example of the system (TTS engine). As shown in FIG. 1, a TTS reading unit 1 including a language analyzing unit 2 and a waveform generating unit 3 and a language analysis to be referred to when performing TTS processing. A dictionary 4, a user dictionary 5 and a speech synthesis dictionary 6 are provided. When using this TTS engine to read “text” mixed with kana and kanji, the language analysis unit 2 analyzes the text information based on the language analysis dictionary 4 and the user dictionary 5 and writes an intermediate language (with readings and accents). String). A kana-kanji mixed text that cannot be read correctly even if the language analysis dictionary 4 is used can be read out correctly by registering the text (vocabulary) and the “reading” in the user dictionary 5. The intermediate language generated through the language analysis unit 2 is input to the waveform generation unit 3 and converted into speech (waveform) data corresponding to the text information with reference to the speech synthesis dictionary 6. Since the user dictionary 5 can be customized unlike the language analysis dictionary 4, it is necessary in advance based on the judgment of the designer or developer who builds the application using the TTS engine, the request from the customer, etc. Vocabulary can be registered.

As a technique related to the above-described conventional technique, for example, as described in Patent Document 1, in an information processing apparatus having a function of reading out an arbitrary text and outputting a voice, the vocabulary of different vocabulary with the same notation is read. Some of the kana readings give priority to the reading method used by the user so that more accurate reading can be performed.
Japanese Patent Laid-Open No. 9-245023

  As described above, in the conventional technique (FIG. 1), text mixed with kana and kanji characters that cannot be read correctly using the language analysis dictionary 4 is registered in the user dictionary 5 together with the “reading” so that it can be read correctly. The state in which text (kanji) that cannot be read correctly using only the language analysis dictionary and its “reading” registered in the user dictionary are referred to as “default state” hereinafter. FIG. 2 shows an example thereof, and shows an example of user dictionary registration in a default state using a conventional TTS process. In the example shown in the figure, based on the result of checking whether it can be read correctly or not using only the language analysis dictionary, texts that could not be read correctly (“Mikamocho”, “Mitsumacho”, “Oita Expressway”) Along with the notation (kanji), “reading” (“Mikachou”, “Mimachou”, “Oitaji Dodo Shadow”) is registered in the user dictionary. As a result, as shown on the right side of the figure, finally, all the texts (text corresponding to readings such as city names and street names) can be correctly read using the language analysis dictionary and the user dictionary.

  However, when a new “text mixed with kana / kanji” is added to the user dictionary 5 by the user, the following problems arise. The state where the user adds text to the user dictionary in the default state in this way is hereinafter referred to as “user customization state”. FIG. 3 shows an example of this, and shows an example of user dictionary registration in a user customization state using conventional TTS processing. In the example shown in the figure, the user registered the vocabulary “Kawauchi” together with the reading “Kawauchi” (the part enclosed by A and B in the figure), but the same vocabulary “Kawauchi” in the default text-to-speech correspondence "Satsuma Kawauchi City" is included, so the user-customized text that can be read aloud can be read correctly for the vocabulary "Kawauchi" added by the user (the part enclosed in C in the figure) ) For the text “Satsuma Kawauchi City” that was in the default state, “Satsuma Sendai”, which is the original reading, is mistakenly read as “Satsuma Kawauchi” (D in the figure) Enclosed part). That is, there may be a problem that the text corresponding to reading that is correctly read in the default state cannot be read correctly in the user customization state.

  Such inconvenience can be solved, for example, by registering all “reading” such as city names and road names in the user dictionary, but if all vocabularies are registered in the user dictionary, the size (memory capacity) of the user dictionary is reduced. There is a problem that the amount of resources (memory, CPU load, etc.) required for speech synthesis increases.

  Such a problem of the prior art is not unique to an in-vehicle device such as an in-vehicle navigation device, but as long as it is an information processing device having a function (TTS engine) for reading out text information mixed with kana and kanji as described above. For example, it can occur in an information processing apparatus such as a personal computer (PC) or a personal digital assistant such as a PDA.

  The present invention was created in view of the problems in the prior art, and even when a dictionary of a vocabulary is registered based on a user instruction, the reading of the vocabulary in a text to be read out can always be kept in a correct state. It is an object to provide an information processing apparatus and a method for registering a speech reading vocabulary.

In order to solve the above-described problems of the prior art, according to an aspect of the present invention, there is provided a speech reading unit having a function of converting text mixed with kana and kanji into speech data and outputting the speech data, and a storage unit, A language analysis dictionary to be referred to when analyzing the text to be read, a vocabulary based on default and user instructions for registering in advance text that cannot be read correctly even if the language analysis dictionary is used User dictionary divided for customization for registering , text-to-speech text including text-to-speech text created using the language analysis dictionary and the user dictionary, and reference when executing speech-to-speech and what is the speech synthesis dictionary stored, input means for inputting a user's instruction, the speech reading means, said storage means and Control means operatively connected to the input means, and when the control means obtains a vocabulary to be registered based on a user instruction, the control means refers to the text corresponding to reading a part of the vocabulary. the text is extracted if it contains text that contains as, after obtaining readings in the default state of the text the extracted, registered readings of the vocabulary and the vocabulary in the user dictionary for the customization and customize state, by comparing the readings in the default state of the reading and the text in the text including the vocabulary acquired by the customization state as a part thereof, both reading of the relevant text in the case of disagreement said default An information processing apparatus is provided in which readings in a state are registered in the default user dictionary.

According to the information processing apparatus according to this embodiment, when the user is included in the corresponding text-to-speech is text containing a vocabulary has been specified as a registration subject, after obtaining readings in the default state of the text, the vocabulary and The vocabulary reading is registered in the user dictionary for customization to be in a customized state, a reading of the text including the vocabulary in the customized state as a part thereof is acquired, and the acquired reading and the default state of the text are acquired. When the readings are compared and the readings do not match, the reading of the text in the default state is registered in the default user dictionary. In other words, even when a vocabulary based on a user instruction and its reading (for example, “Kawauchi”) are registered in a dictionary, reading of a text (for example, “Satsuma Kawauchi”) that includes the vocabulary as a part thereof is performed. Is inconsistent with the correct reading in the default state (Satsuma Sendai) (when the reading of the text changes), the correct reading in the default state is also registered in the dictionary.

  Thereby, the reading of the vocabulary (text) in the reading-ready text can always be maintained in a correct state. In addition, since it is not necessary to register all vocabularies in the dictionary as in the prior art, the size of the user dictionary can be kept small, and resources required for speech synthesis can be reduced.

Further, according to another aspect of the present invention, a language analysis dictionary to be referred to when analyzing a text to be read aloud, a text that cannot be read correctly even if the language analysis dictionary is used, and a correct reading thereof are registered in advance. Text mixed with kana and kanji using a user dictionary that is divided into a default for storing and a custom dictionary for registering vocabulary based on user instructions, and a speech synthesis dictionary that is referred to when performing speech reading the information processing apparatus having a text-to-speech function and outputs the converted into voice data, stored in the storage means corresponding text-to-speech including speech reading the text that is created by using the language analysis dictionary and the user dictionary When the vocabulary to be registered is acquired based on the user instruction, the word is referred to by referring to the reading-ready text. The search text that contains as a part, after the text is extracted if it contains text containing the vocabulary as a part, were obtained readings in the default state of the text the extracted, the and customize state reading vocabulary and the vocabulary registered in the user dictionary for the customization, the readings and in the default state of the reading and the text in the text including the vocabulary acquired by the customization state as part In comparison, there is provided a method for registering a speech-to-speech vocabulary characterized in that when the readings of the two do not match, the default reading of the text is registered in the default user dictionary.

  Other structural features and the detailed processing contents of the information processing apparatus and the like according to the present invention will be described with reference to embodiments of the invention described later.

  DESCRIPTION OF EXEMPLARY EMBODIMENTS Hereinafter, preferred embodiments of the invention will be described with reference to the accompanying drawings.

  FIG. 4 is a block diagram showing the configuration of an in-vehicle navigation apparatus incorporating an information processing apparatus according to an embodiment of the present invention.

  The in-vehicle navigation device 20 according to the present embodiment includes a hard disk drive (HDD) 10 as a storage medium, an operation unit 11 as a user interface, a GPS receiver 12, a self-contained navigation sensor 13, and a communication as illustrated. 14, display device 15, speaker 16, control unit 21, guidance route storage unit 22, image composition unit 23 that supplies image data to display device 15, TTS (Text To Speech) reading unit 24, And an audio output unit 25 for supplying audio data to the speaker 16. The control unit 21 is built in the navigation device body together with the guidance route storage unit 22, the image synthesis unit 23, the TTS reading unit 24, and the voice output unit 25. The information processing apparatus related to the present invention includes three types of dictionaries and reading-ready text in the HDD 10 to be described later, a TTS reading unit 24, and a control unit 21 (part of its functions) for controlling input / output of these. It is comprised by. The TTS reading unit 24 has the same configuration (language analysis unit 2 and waveform generation unit 3) and functions as the TTS reading unit 1 shown in FIG.

  A disk (not shown) driven by the HDD 10 includes map data (map database 10a) used when the navigation function is executed, and three types of dictionaries (when the speech synthesis (TTS) function is executed). A language analysis dictionary 10b, a user dictionary 10c, a speech synthesis dictionary 10d), and a text-to-speech text 10e prepared in advance are stored in assigned storage areas. The map database 10a stores map data divided into longitude and latitude widths of appropriate sizes according to each scale level (1/12500, 1/25000, 1/50000, etc.). Includes road data for display and road data for route search, as well as data related to various facilities (convenience store, gas station, entertainment facility, etc.) (name of facility, reading, telephone number, address, etc.) Each data is represented by a coordinate set of points (nodes) expressed by longitude and latitude.

  The language analysis dictionary 10b is a dictionary that is referred to when performing syntactic analysis of text information to be subjected to speech synthesis processing. Details such as notation, kana reading, accent information, and part-of-speech information in advance for each vocabulary. Registered in association with each other. The user dictionary 10c is a dictionary (dictionary that can be customized by the user) for registering vocabulary (text information) that cannot be read correctly even if the language analysis dictionary 10b is used. The vocabulary in which the “reading” changes due to the vocabulary registered by the user (or due to the update of the map data) or text information including the vocabulary is registered. The speech synthesis dictionary 10d is a dictionary that is referred to when text information analyzed through the language analysis unit in the TTS reading unit 24 is converted into corresponding speech (waveform) data. In the text to be read-read 10e, texts (city names, road names, etc. as illustrated in FIG. 2) that can be read correctly in a default state are stored in a list format in advance. Also stored is a text that can correctly read a vocabulary (text) in which “reading” changes.

  The operation unit 11 is for operating the navigation apparatus body, and has, for example, a remote controller (remote control) transmitter. The remote control transmitter has various operations for displaying various menus on the screen of the display device 15 to be described later, selecting various menus and various items on the display screen, and executing the selected menus. Keys, buttons, joysticks, etc. are provided. Furthermore, the operation unit 11 is provided with a “TTS reading” key used as a trigger when the name of the road on which the vehicle is currently traveling is output as an operation key related to the present invention, as will be described later. . In addition to the remote control transmitter, the operation unit 11 may be a touch panel provided as appropriate on the screen of the display device 15.

  The GPS receiver 12 receives radio waves (GPS signals) transmitted from GPS satellites and detects the longitude and latitude of the current position of the vehicle. The self-contained navigation sensor 13 is for detecting the direction and traveling speed of the own vehicle, and is composed of an angle sensor such as a 3D gyro, a distance sensor that generates a pulse for every certain traveling distance based on the number of revolutions, and the like. Has been. The communicator 14 includes a mobile phone (attached with a dedicated modem), an in-vehicle phone, and the like for communicating with an information center outside the vehicle, a mobile phone held by a friend, and the like.

  The display device 15 is composed of an LCD monitor or the like, and is installed at a substantially middle position of the center console so that at least the display screen can be viewed from the driver's seat. On the screen of the display device 15, guidance information related to navigation (a map around the vehicle position, the vehicle position mark, the vehicle position mark, the To the navigation route). Although only one speaker 16 is shown, in practice, a predetermined number of speakers 16 in a predetermined position in the passenger compartment, for example, if the rear seats are in a single row, at least the vicinity of the left and right of the rear seat and the left and right of the front seat. In total, four speakers are installed in the vicinity of each. The speaker 16 basically outputs the guidance information related to the navigation through the voice output unit 25 based on the control from the control unit 21. Further, as the guidance information related to the present invention, The voice data (text information such as e-mail text) converted through the TTS reading unit 24 is output as voice.

  The control unit 21 is constituted by a microcomputer or the like, and various control unit programs that define various processes (drawing process of map and vehicle position mark, route search process, map matching and vehicle position correction process based on this, A control program for registering a vocabulary (text) for TTS reading that characterizes the invention is stored in a built-in memory (not shown). The control unit 21 performs processing in accordance with these control programs, and basically detects the current position of the own vehicle (own vehicle position) based on the signal output from the GPS receiver 12 or the self-contained navigation sensor 13. Detecting the direction and traveling speed of the vehicle based on the signal output from the vehicle, or guiding from the departure point (own vehicle position) to the destination according to the search conditions set using the map data read from the map database 10a Various processes related to navigation, such as searching for a route, are performed. Further, as a process related to the present invention, the control unit 21 controls speech synthesis processing based on language analysis and waveform generation in cooperation with the TTS reading unit 24 as described later, and also uses a vocabulary (text for TTS reading). ) Is controlled.

  The guidance route storage unit 22 stores data relating to all nodes (coordinates of points expressed in longitude and latitude) from the starting point (for example, the vehicle position) of the guidance route searched by the control unit 21 to the destination. It is for keeping. The image compositing unit 23 performs map image drawing processing using the map data read from the map database 10a by the control unit 21, and various menu screens (operation screens), own vehicle position marks, Various kinds of marks such as a cursor are generated, or data on the guidance route is read from the guidance route storage unit 22 based on the control from the control unit 21, and the guidance route is displayed with a relatively conspicuous display mode (for example, a color change, a line (Such as making the width thicker)). That is, the image composition unit 23 has a function of superimposing a guidance route, an operation screen, various marks, and the like on the map image and displaying them on the screen of the display device 15. The voice output unit 25 converts a digital voice signal (guidance information related to navigation, text voice data converted through the TTS reading unit 24) supplied via the control unit 21 into an analog waveform signal, and appropriately performs power amplification. Output to the speaker 16.

  In the in-vehicle navigation device 20 of the present embodiment configured as described above, the HDD 10 is a “storage unit”, the operation unit 11 is an “input unit”, and the GPS receiver 12 is a “own vehicle position detection unit”. The communication device 14 corresponds to “communication means”, the control section 21 corresponds to “control means”, and the TTS reading section 24 corresponds to “voice reading means”.

  Hereinafter, TTS performed in the in-vehicle navigation apparatus 20 incorporating the information processing apparatus (language analysis dictionary 10b, user dictionary 10c, speech synthesis dictionary 10d, reading-ready text 10e, TTS reading unit 24, control unit 21) according to the present embodiment. Processing related to registration of a vocabulary (text) for reading will be described with reference to FIG. In addition, a supplementary explanation will be given with reference to FIG. The vocabulary (text) according to the present embodiment is registered when the vehicle is stopped in consideration of safety.

  First, as a precondition, the reading-ready text 10e in the HDD 10 stores in advance a list of texts (city names, road names, etc. as illustrated in FIG. 2) that can be correctly read in the default state. It shall be.

  In the first step S1 in this state, the control unit 21 acquires a vocabulary that the user wants to register. That is, when the user designates a desired vocabulary as a speech reading target by a remote control operation from the operation unit 11 (or a touch operation on the screen of the display device 15), the control unit 21 selects the designated vocabulary and the vocabulary Get "reading" information. In the example shown in the process P1 of FIG. 6, a reading pseudonym “Kawauchi” is designated for the vocabulary “Kawauchi” to be registered. Information acquired by the control unit 21 is temporarily stored in a built-in memory (not shown).

  In the next step S2, the control unit 21 searches the text including the vocabulary (in this case, “Kawauchi”) by referring to the language analysis dictionary 10b, the user dictionary 10c, and the reading-ready text 10e in the HDD 10 (FIG. Process P2 of 6).

  In the next step S3, the control unit 21 determines whether there is a text including the vocabulary (YES) or not (NO) based on the search result. If the determination result is yes, the process proceeds to step S5, and if the determination result is no, the process proceeds to step S4.

  In step S4 (when the text including the vocabulary is not included in the reading-ready text 10e), the vocabulary “Kawauchi” is read together with the reading “kawauchi” in the user dictionary 10c in the HDD 10 based on the control from the control unit 21. Register (user customization status). The user dictionary registration performed here is the same as in the conventional case illustrated in FIG. Then, this processing flow is “end”.

  On the other hand, in step S5 (when the text including the vocabulary is included in the reading-ready text 10e), the control unit 21 extracts the text including the vocabulary from the reading-ready text 10e, and the default state of the text Get "reading" on In the example shown in the process P3 of FIG. 6, the text “Satsuma Kawauchi City” including the vocabulary “Kawauchi” to be registered is extracted, and “reading” “Satsuma Sendai” in the default state is acquired. The “reading” obtained here is an accurate “reading”. The acquired information is temporarily stored in a memory (not shown) in the control unit 21.

  In the next step S6, the vocabulary “Kawauchi” is registered in the user dictionary 10c together with the reading “kawauchi” in the same manner as the processing performed in step S4 (processing P4 in FIG. 6).

  In the next step S7, the control unit 21 acquires “reading” in the user customization state of the text including the vocabulary extracted from the reading-ready text 10e. In the example shown in the process P5 of FIG. 6, “reading” “Satsuma Kawauchi” in the user customization state of the text “Satsuma Kawauchi City” including the extracted vocabulary “Kawauchi” is acquired. The “reading” obtained here is inaccurate “reading” because it is influenced by the vocabulary “kawauchi” registered by the user. The acquired information is temporarily stored in a memory (not shown) in the control unit 21.

  In the next step S8, the control unit 21 compares each acquired “reading”, that is, “reading” in the default state of the text with “reading” in the user customization state, and the two match (YES). ) Or not (NO). If the determination result is YES, the process proceeds to step S9, and if the determination result is NO, the process proceeds to step S10. In the example shown in the process P6 of FIG. 6, the “reading” “Satsuma Sendai” obtained in the default state is compared with the “reading” “Satsuma Kawauchi” obtained in the user customization state. Therefore, in this case, the determination result is NO, and the process proceeds to step S10.

  In step S9, similarly to the process performed in step S3, the control unit 21 determines whether there is another text (YES) or not (NO) including the vocabulary (vocabulary that the user wants to register). If the determination result is YES, the process returns to step S5 and the above processing is repeated. If the determination result is NO, the process flow is “end”.

  On the other hand, in step S10 (when the “reading” in the acquired default state and the “reading” in the user customization state do not match), the data is extracted in the user dictionary 10c in the HDD 10 based on the control from the control unit 21. Register the correct “reading” of the text. In the example shown in process P7 of FIG. 6, the correct “reading” of “Satsuma Sendai” in the default state of the text “Satsuma Kawauchi City” including the extracted vocabulary “Kawauchi” is registered in the user dictionary 10c. . The user dictionary registration performed here is processing that characterizes the present invention. Then, the process proceeds to step S9.

  As described above, according to the registration method (FIGS. 5 and 6) of the TTS reading vocabulary performed in the in-vehicle navigation device 20 (FIG. 4) according to the present embodiment, the vocabulary “Kawauchi” designated as the registration target by the user. When the text “Satsuma Kawauchi-shi” including “(kawauchi)” is included in the reading-ready text 10e, the reading “Satsuma Sendai” in the default state of the text is acquired and the vocabulary “Kawauchi” is Registered in the user dictionary 10c together with the reading “kawauchi” (user customization state), obtains the reading “Satsumakawauchi” of the text in this user customization state, and compares it with the reading “Satsuma Sendai” in the default state, If the readings do not match (“NO” in step S8), the text is correct in the default state So that to automatically register themselves "Satsuma Sendai death" to the user dictionary 10c. In other words, even when the user dictionary is registered for the vocabulary based on the user instruction, if the reading of the text containing the vocabulary does not match the correct reading in the default state, the correct reading in the default state is also added to the user. The dictionary is registered.

  Thereby, the reading of the vocabulary (text) in the reading-ready text 10e can be always kept in a correct state. In addition, since it is not necessary to register all vocabularies in the dictionary as in the past, the size of the user dictionary can be kept small, and resources (memory, CPU load, etc.) required for speech synthesis (TTS processing) can be reduced. Can do.

  FIG. 7 shows an example of user dictionary registration in the user customization state based on the processing (FIG. 5) according to the present embodiment. In the illustrated example, the user registers the vocabulary “Kawauchi” along with its reading “Kawauchi” (the part enclosed by A and B in the figure), and the same vocabulary “Kawauchi” as the vocabulary in the text corresponding to reading in the default state. Although the text “Satsuma Kawauchi City” is included, the correct “reading” (Satsuma Sendai) in the default state is also automatically registered (the part surrounded by E in the figure), so the user customization state In the reading-ready text, both the vocabulary “Kawauchi” added by the user and the text “Satsuma Kawauchi City” held in the default state can be read correctly (parts surrounded by C and D in the figure).

  Note that, as a method of using the vocabulary (text) registered in the reading-ready text 10e based on the processing (FIG. 5) according to the present embodiment, the following modes can be considered.

  One of them is to read out the name of the road on which the vehicle is currently traveling (voice guidance). Specifically, when there is a user instruction (operation of “TTS reading” key) through the operation unit 11, the control unit 21 refers to the data of the vehicle position detected by the GPS receiver 12. Then, the name (data) of the road on which the vehicle is currently traveling is read from the map database 10a, the read road name (text) is input to the TTS reading unit 24, and voice based on required language analysis and waveform generation is input. The sound converted by the synthesis is output as audio from the speaker 16 via the audio output unit 25.

  As another usage mode, it is conceivable to read out an e-mail text, a news text, or the like received from the outside by voice. For example, when an e-mail is received from the outside via a mobile phone (communication device 14), the received e-mail text (text) is converted into voice data through the TTS reading unit 24 based on control from the control unit 21. The sound is output from the speaker 16 via the sound output unit 25. Such voice guidance is useful in terms of safety because it can be automatically performed without requiring a special operation by the user even when the vehicle is running.

  In the above-described embodiment, the case has been described in which text information including the vocabulary whose “reading” changes due to the vocabulary registered by the user (FIGS. 5 and 6) is taken as an example. Factors that change "reading" exist not only in registration based on user instructions, but also in map data for navigation itself. That is, the map data for navigation is regularly updated (for example, every year for genuine products). During the period until this data update is performed, municipal mergers, etc. The name of the city may have been changed. Even in such a case, by registering the vocabulary according to the above-described processes of FIGS. 5 and 6, the updated data (vocabulary whose “reading” changes) can be read out correctly.

  In the above-described embodiment, the case where the information processing apparatus according to the present invention is incorporated as a part of the in-vehicle navigation apparatus 20 has been described as an example. However, as apparent from the gist of the present invention, the information processing apparatus is limited to in-vehicle use. Of course not. In short, as described above, a function for reading out text information mixed with kana and kanji by voice (TTS reading section 24) and a dictionary (language analysis dictionary 10b, user dictionary 10c, voice synthesis) referred to when the voice reading function is executed. An information processing device having a dictionary 10d) and a text-to-speech-reading text 10e set in advance so that it can be read correctly in a default state is sufficient. The present invention can be applied in the same way when reading aloud.

  In the embodiment described above, the HDD 10 is used as a storage medium for storing the map database 10a for navigation, the dictionaries 10b, 10c, 10d for reading TTS, and the text 10e for reading, but instead, Other storage media such as a DVD drive (DVD-ROM) and a CD drive (CD-ROM) may be used. Further, the map database 10a, the dictionaries 10b to 10d, and the reading-ready text 10e are not necessarily stored in the same storage medium, and may be stored separately in separate storage media as appropriate.

It is explanatory drawing of a typical TTS process. It is a figure which shows an example of the user dictionary registration in the default state using the conventional TTS process. It is a figure which shows an example of the user dictionary registration in the user customization state using the conventional TTS process. It is a block diagram which shows the structure of the vehicle-mounted navigation apparatus incorporating the information processing apparatus which concerns on one Embodiment of this invention. FIG. 5 is a flowchart showing an example of processing related to registration of a vocabulary (text) for TTS reading performed in the in-vehicle navigation device of FIG. 4. FIG. 6 is a diagram for supplementarily explaining the processing flow of FIG. 5. It is a figure which shows an example of the user dictionary registration based on the process of FIG.

Explanation of symbols

10: HDD (storage means),
10a ... Map database,
10b ... language analysis dictionary,
10c ... user dictionary,
10d: Speech synthesis dictionary,
10e ... Text to be read aloud,
11 ... operation part (input means),
12 ... GPS receiver (vehicle position detection means),
14 ... Communicator (communication means),
15 ... display device,
16 ... Speaker,
20 ... In-vehicle navigation device,
21 ... control unit (control means),
24 ... TTS reading section (speech reading section).

Claims (5)

A speech-to-speech means having a function of converting text mixed with kana and kanji into speech data and outputting;
A language analysis dictionary that is a storage means that is referred to when analyzing a text to be read aloud, a text that cannot be read correctly even if the language analysis dictionary is used, and a default for registering the correct reading in advance . And a user dictionary divided for customization for registering vocabulary based on user instructions, text-to-speech that includes the language analysis dictionary and text to be read-out created using the user dictionary, and voice-to-speech Storing a speech synthesis dictionary to be referred to when executing
Input means for inputting user instructions;
Control means operatively connected to the voice reading means, the storage means and the input means,
When the control means obtains a vocabulary to be registered based on a user instruction, if the text includes the vocabulary as a part thereof with reference to the reading-ready text, the control means extracts the text,
After obtaining the reading of the default state of extract out was the text, and customize state to register the reading of the vocabulary and the vocabulary to the user dictionary for the customized,
The reading of the text that includes the vocabulary acquired in the customization state as a part thereof is compared with the reading of the text in the default state, and if the readings of the text do not match, the reading of the text in the default state is compared. Is registered in the default user dictionary. The control means registers the vocabulary and the reading of the vocabulary in the user dictionary for customization when the text corresponding to the vocabulary is not included with reference to the reading-ready text. The information processing apparatus according to claim 1. Furthermore, a communication means and a speaker for outputting the voice data converted through the voice reading means as voice,
When the text information is received via the communication means, the control means converts the text information into voice data through the voice reading means, and outputs the voice data from the speaker. Item 4. The information processing apparatus according to Item 1. Furthermore, it is equipped with a storage means that is mounted on a vehicle and stores map data, and a vehicle position detection means that detects a current position of the vehicle,
The control means acquires the name of the road on which the host vehicle is currently traveling with reference to the detected current position of the host vehicle and the map data when the user gives an instruction via the input unit; 4. The information processing apparatus according to claim 3, wherein the name information is converted into voice data through the voice reading unit, and the voice data is output as voice from the speaker. Based on the language analysis dictionary to be referred to when analyzing the text to be read aloud, the text that cannot be read correctly even if the language analysis dictionary is used, and the default instructions for registering the correct reading in advance and user instructions A text-to-speech function that converts a kana-kanji mixed text into speech data and outputs it using a user dictionary that is divided into customizations for registering vocabularies and a speech synthesis dictionary that is referenced when speech-to-speech is executed. In an information processing apparatus comprising:
May be stored in the storage means corresponding text-to-speech including speech reading the text that is created by using the language analysis dictionary and the user dictionary,
When the vocabulary to be registered based on the user instruction is acquired, the text that includes the vocabulary as a part thereof is searched with reference to the reading-ready text,
When the text including the vocabulary as a part thereof is included, the text is extracted, and after reading the extracted text in a default state, the vocabulary and the reading of the vocabulary are used for the customization. Register in the user dictionary and customize it ,
Comparing the reading of the text that includes the vocabulary acquired in the customization state as a part thereof with the reading of the text in the default state;
A method for registering a speech-to-speech vocabulary, wherein when the readings of the two do not match, the reading of the text in the default state is registered in the default user dictionary.

Images