JP2009116107A - Information processing device and method - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To present confirmation information for confirmation results in a manner easy for a user to understand while allowing recognition of an optional utterance by voice recognition using subword grammar. <P>SOLUTION: A voice recognition part 102 performs, to an input voice, voice recognition using keyword grammar based on a keyword for a state related to registration, and voice recognition using subword grammar based on a subword, and outputs recognition results of the recognition having a more satisfactory recognition score of the both. When the voice recognition part 102 outputs the recognition results of the voice recognition using the keyword grammar, a confirmation information generation part 106 generates confirmation information composed of a keyword string of the recognition results. At the same time, when the recognition part 102 outputs the recognition results of the voice recognition using the subword grammar, the generation part 106 generates confirmation information composed of attribute information associated with the state related to registration. <P>COPYRIGHT: (C)2009,JPO&INPIT

Description

  The present invention relates to speech recognition technology.

  Voice recognition technology is being put to practical use in devices such as car navigation systems. Voice recognition has the advantage that one item can be selected by voice from a plurality of selection items, and the number of buttons and operation steps of the apparatus can be reduced.

  As one of the voice recognition applications, it is possible to register the state and setting of another application by voice and call the registered state by voice. FIG. 15 is a diagram illustrating an example of an application that displays an electronic program guide on a television. In the figure, reference numeral 1501 denotes a program guide. 1502 is a program (broadcast frame) in the program guide, and 1503 is a focused program. In the program guide, programs are arranged in order of broadcast time for each channel. Then, each program can be focused and selected, or the focus can be moved to a nearby program. The user can register a state in which a certain program is focused in the electronic program guide using voice recognition. For example, when the user utters “Taro Yamada's King David”, the state is registered in association with the speech recognition result. Then, by calling the same utterance from another state, it is possible to transition to the registered state.

  In speech recognition, as a method of uttering and registering speech and calling the registered information by speech, there is a method of determining by proximity of speech by DP matching of speech feature values at the time of registration and at the time of calling. However, at present, speech recognition by HMM (Hidden Markov Model) has become mainstream because recognition accuracy, recognition is possible regardless of the speaker, and user registration is not essential.

  In voice recognition using the HMM, voice registration / calling is performed in the following procedure. At the time of registration, speech recognition using a word or subword (phoneme) that is a speech component to be registered as a grammar is executed, and the recognized word string or subword string is registered as a word in the registration grammar for registration. At the time of calling, voice recognition is executed using the registered grammar, and the state and setting corresponding to the word of the recognition result are called.

  At this time, there are two possible grammar types used during registration. One is a keyword grammar, which is a grammar based on words prepared in advance or extracted from state and setting information. The other is a subword grammar, for example, a grammar based on phonemes, which is the minimum unit of phonemes.

  FIG. 4 shows an example of the keyword grammar, and FIG. 5 shows an example of the subword grammar. The keyword grammar in FIG. 4 is a grammar generated by extracting a keyword from program information included in the program 1503 focused on the program table in FIG. FIG. 5 is a grammar based on a predetermined phoneme set representing all phonemes when speaking a language without depending on program information. FIG. 16 shows a speech recognition result obtained using these grammars. When the user utters “Taro Yamada's King David”, in the case of speech recognition using the keyword grammar, the recognition result is “Taro Yamada's King King”, which is a keyword string that combines the words of the keyword grammar. On the other hand, in the case of sub-word grammar, the recognition result is “y a m a d a t a r o o n o d a b i d e o o”, which is a sub-word string combining sub-words. In each case, such a recognition result is registered in the registration grammar as one word. When the same utterance is made at the time of calling, a registered word that matches a word (“Taro Yamada's King of David” or “y amad a t o r o o n o d a b d e o o”) included in the registration grammar can be obtained as a recognition result.

  Here, the recognition by the keyword grammar can register only a word string included in the keyword grammar. For example, if you say “King David of Yama-chan”, you can convert it to “y a m a ch a X n o d a b i d e o o” in the subword grammar. On the other hand, since there is no word corresponding to “Yama-chan” in the keyword grammar, an incorrect recognition result is registered or a recognition result cannot be obtained. That is, in order for a user to register with an arbitrary utterance, it is necessary to use a subword grammar.

JP-A-10-097284 Acoustical Society of Japan 2006 Spring Research Presentation 1-1-6

  However, the recognition result of the subword grammar is difficult to present to the user. The application preferably presents the recognition result to the user by display or voice for confirmation at the time of registration or calling. If the speech recognition is based on keyword grammar, the word string “Taro Yamada's King of David” can be obtained from the recognition result, so the user can present it by displaying it, or it can also present it by speech synthesis. . On the other hand, it is difficult for the user to understand even if the sub-word string “y m a dat a r o o n o d a b i d e o o” is displayed. In addition, even if speech synthesis is performed, there is no linguistic information, so synthesis cannot be performed with correct intonation.

  Therefore, if the recognition result of the keyword grammar is preferable while using the subword grammar to recognize any user's utterance, information that matches the user's utterance as much as possible is presented as natural as possible. It is desirable.

  As speech recognition using both subword grammar and keyword grammar, there are Patent Document 1 and Non-Patent Document 1. In Patent Document 1, the recognition result of the keyword grammar and the recognition result of the subword grammar are compared, and if the score of the keyword grammar is high, the recognition result is output, and if the score of the subword grammar is high, the recognition result is rejected. . Further, in Non-Patent Document 1, as a first and last name input interface, first name surname dictionary recognition and syllable string recognition are used together to output each recognition result, and the user selects from the recognition results.

  However, Patent Document 1 does not disclose using the recognition result of the subword grammar for presentation. Non-Patent Document 1 may be used as an input of syllables, but even if a syllable string is presented when predetermined information is presented to the user, it is difficult for the user to understand.

  An object of the present invention is to make it possible to present confirmation information about a recognition result in an easy-to-understand manner to a user while making it possible to recognize an arbitrary utterance by speech recognition using a subword grammar.

  According to one aspect of the present invention, information on an application selected by a user can be registered according to a user's registration operation, and the registered state can be called according to a user's call operation to transition to the state. A speech processing unit for inputting a user's speech, speech recognition using a keyword grammar based on a keyword related to the state, and a subword grammar based on a subword for speech input from the speech input unit A speech recognition unit that performs speech recognition using a word and outputs a recognition result of a person with a good recognition score; and when the speech recognition unit outputs a recognition result of speech recognition using the keyword grammar, the recognition While the confirmation information composed of the resulting keyword string is presented to the user, the voice recognition means When a recognition result of speech recognition using word grammar is output, a presentation means for presenting confirmation information made up of attribute information associated with the state to the user, and a recognition result output by the speech recognition means There is provided an information processing apparatus comprising registration means for associating the state with the state and registering them in a storage means.

  According to the present invention, it is possible to present confirmation information about a recognition result in an easy-to-understand manner to a user while making it possible to recognize an arbitrary utterance by speech recognition using subword grammar.

  DESCRIPTION OF EMBODIMENTS Hereinafter, preferred embodiments of the present invention will be described in detail with reference to the drawings. In addition, this invention is not limited to the following embodiment, It shows only the specific example advantageous for implementation of this invention. In addition, not all combinations of features described in the following embodiments are indispensable as means for solving the problems of the present invention.

(Embodiment 1)
Hereinafter, as an example, an information processing apparatus will be described in which the state of an application that handles an electronic program guide on a television is registered by voice and the registered state can be called by voice. In this information processing apparatus, for example, a state in which the user has selected a desired program frame from the displayed electronic program guide can be registered in accordance with the user's registration operation. After registration, the system can be quickly transitioned to that state according to the user's call operation. Voice recognition can be used for the registration operation and the call operation. Specifically, when registering, the user utters a word or phrase as a title for specifying the state. Then, voice recognition is performed on this utterance, and the recognition result and the state related to registration are associated and registered. In order to call the state registered in this way, the user may utter a word or phrase as the title of the state. Then, voice recognition is performed on the utterance, the state associated with the recognition result is read, and the system transitions to that state.

  However, it goes without saying that the present invention is not limited to applications that handle electronic program guides.

  FIG. 1A is a block diagram showing a configuration of a voice recognition function in the information processing apparatus according to the first embodiment of the present invention. In the figure, reference numeral 101 denotes a voice input unit for inputting voice. Reference numeral 102 denotes a voice recognition unit that recognizes voice. Reference numeral 103 denotes a keyword grammar generation unit that generates keyword grammar from predetermined information in a registered state. Reference numeral 104 denotes a keyword grammar which is a grammar for speech recognition generated by the keyword grammar generation unit 103. Reference numeral 105 denotes a subword grammar that is a grammar for speech recognition based on a subword such as a phoneme. A confirmation information generation unit 106 generates confirmation information about the recognition result of the voice recognition. 107 is a presentation unit that displays the confirmation information created by the confirmation information generation unit 106 to the user by display or voice. Reference numeral 108 denotes a registration unit that registers a speech recognition result and a state in association with each other. Reference numeral 109 denotes a registration grammar in which the registration unit 108 registers the recognition result of the speech recognition unit 102 as a recognition grammar word. Reference numeral 110 denotes registration state information in which words to be registered in the registration grammar 109 are associated with states.

  FIG. 1B is a block diagram illustrating a hardware configuration of the information processing apparatus according to the present embodiment. This information processing apparatus can be realized by a computer apparatus. Specifically, as shown in the figure, a CPU 1 that controls the entire information processing apparatus, a RAM 2 that functions as a main storage device and provides a work area to the CPU 1, a ROM 3 that stores a boot program and the like, and the following configuration Is provided.

  The HDD 4 is a hard disk device, in which a voice recognition program 42 is installed in addition to the OS 41. The voice recognition program 42 includes a voice recognition module 43, a keyword grammar generation module 44, a confirmation information generation module 45, and a registration module 46. Each of these modules includes program codes for realizing the functions of the speech recognition unit 102, the keyword grammar generation unit 103, the confirmation information generation unit 106, and the registration unit 108. The above-described keyword grammar 104, subword grammar 105, registration grammar 109, and registration status information 110 are also stored in the HDD 4. In addition, a TV application 47 is installed in the HDD 4 for receiving and displaying television programs, providing an electronic program guide, and the like.

  Reference numeral 5 denotes a microphone, which constitutes the voice input unit 101. Reference numeral 6 denotes a speaker that outputs sound, 7 denotes a display that displays characters and images, and these constitute a presentation unit 107. Reference numeral 8 denotes a communication interface (I / F) for receiving television program content data, electronic program guide data, and the like. An operation unit 9 is realized by an operation button, a keyboard, a mouse, and the like. Alternatively, the operation unit 9 may be realized by a remote controller. The user can perform an operation for starting / ending an application, an operation on the electronic program guide displayed on the display 7, and the like via the operation unit 9.

  The configuration of the information processing apparatus in the present embodiment is generally as described above.

  The keyword grammar generation unit 103 generates a keyword grammar from attribute information associated with a state related to registration. FIG. 2 is a diagram illustrating how keyword grammar is generated from a focused state in an electronic program guide on a television. In the figure, 201 is an electronic program guide for television. A program 202 is a focused program in the electronic program guide 201. Reference numeral 203 denotes program information as attribute information associated with the program 202. Reference numeral 204 denotes a keyword acquired from the program information 203. The keyword grammar generation unit 103 extracts the keyword 204 from the program information 203 of the program 202. Then, language analysis is performed on the extracted keyword 204 to obtain reading information, and a keyword grammar 104 is generated.

  FIG. 3 is a flowchart showing processing when registering a focused state in the program 202 in the present embodiment.

  First, in step S <b> 301, a user's utterance is input via the voice input unit 101. Next, in step S302, the speech recognition unit 102 performs speech recognition using the keyword grammar 104 and speech recognition using the subword grammar 105 on the input speech, and the recognition result of obtaining a good recognition score. Is output. If the recognition result of the speech recognition unit 102 is a keyword string obtained from the keyword grammar 104 in step S303, the confirmation information generation unit 106 generates confirmation information based on the keyword obtained as the recognition result in step S304. On the other hand, when the recognition result is a subword string obtained from the subword grammar 105, in step S305, the confirmation information generation unit 106 obtains confirmation information based on the attribute information (program information 203 in FIG. 2) associated with the state to be registered. Generate. In step S306, the presentation unit 107 presents confirmation information. In step S <b> 307, the registration unit 108 associates the recognition result with the state and registers them in the registration grammar 109 and the registration state information 110.

  In addition, after presenting confirmation information in step S306, the user may confirm whether or not the state may be registered. If the user indicates that registration may be performed by this confirmation, registration is performed in step S307. On the other hand, if the user instructs not to register, the process ends without registering, or a re-voice is requested and the process returns to step S301.

  FIG. 4 shows an example of the keyword grammar 104. In the figure, 401 is a word ID assigned to each word, 402 is a word notation, and 403 is a word reading. Reference numeral 404 denotes a grammar start node, and reference numeral 405 denotes a grammar end node. The grammar is composed of a word and a network, and accepts a word string included in the path from the start node 404 to the end node 405 as a recognition sentence.

  Here, the keyword grammar includes a keyword 204 and a predetermined general word (a word to be used regardless of the state, such as a particle “NO”) prepared in advance, and a grammar in which these word lists can be looped. According to the grammar of FIG. 4, it is possible to accept sentences including keywords such as “Friday movie”, “Taro David Yamada” and “Hanako Sato”.

  FIG. 5 shows an example of the subword grammar 105. In the figure, phonemes are used as subwords, and the notation and the reading are the same, so the notation is omitted. The subword grammar is a recognition grammar using a subword such as a phoneme or triphone as a recognition vocabulary. An arbitrary utterance can be converted into a subword string by using a grammar of a subword of a unit constituting a language like a phoneme. For example, when “Yama-chan” is uttered, a subword string “y a m a ch a X” is obtained. The subword grammar can be used independently of the keywords that appear in each state.

  FIG. 6 and FIG. 7 show the behavior of steps S301 to S306.

  In FIG. 6, the user utters “Taro Yamada's King David” on the voice input unit 101. The speech recognition unit 102 outputs the recognition result and score of speech recognition using the keyword grammar 104 and the recognition result and score of speech recognition using the subword grammar 105, respectively. The recognition using the subword grammar 105 has a higher degree of freedom than the keyword and the score tends to be superior, so it is desirable to add an insertion score for each subword. In this way, when the utterance includes all the words in the keyword grammar, the score of the recognition result using the keyword grammar increases. Thus, in the example shown in the figure, the keyword string “Taro Yamada's King David” is obtained as a recognition result. Then, the confirmation information generation unit 106 uses this keyword string to indicate “Taro Yamada's King of David”? The confirmation unit 107 generates confirmation information, and presents the confirmation information to the user by means of display or speech synthesis.

  On the other hand, as shown in FIG. 7, it is assumed that the user utters “King Yamada” to the voice input unit 101. The keyword grammar 104 does not include the word “Yama-chan”. Therefore, in the recognition result of the keyword grammar, an erroneous recognition result is output with a low score or no recognition result is output as shown in FIG. On the other hand, in the recognition by the subword grammar, any utterance can be applied to the subword string. For this reason, the recognition result based on the subword grammar has a higher score, and the subword string “y a mach a Xnodabideoo” is obtained as the recognition result. In this case, the confirmation information generation unit 106 generates confirmation information based on information on the state to be registered, not from the subword string of the recognition result. In FIG. 7, based on the predetermined rule of combining the character strings described in the broadcast station, title, and subtitle items from the program information 203, ““ Maruko TV Friday Wide Theater David King ”? However, the present invention is not limited to this. For example, other items of the program information 203 may be used, or items to be used may be changed according to the amount of information. Alternatively, the program information 203 or the keyword 204 extracted from the program information 203 may be used as a search key to obtain information from the database as confirmation information.

  8 and 9 show the state registration process in step S307. The registration unit 108 registers two, a registration grammar 109 that represents a registered word by voice and registration state information 110 that represents a registration state. The registered grammar 109 includes, for each registered word, three items: (1) a word ID for identifying the registered word, (2) notation of the registered word, and (3) reading of the registered word. The registration status information 110 includes, for each status related to registration, (1) a registration ID for identifying the status related to registration, (2) a status attribute representing the content of the status, and (3) a registration grammar 109 corresponding to the status. 3 of the word ID representing the registered word.

  FIG. 8 shows a registration process when a keyword string is obtained as a recognition result. Since the recognition results of the keyword grammar are all composed of keywords, the recognition results are used as they are for the registered words. Readings are taken from the keyword grammar and combined. In the figure, the keyword string is registered as it is as a registered word notation for the keyword string “Taro Yamada of King David” as the recognition result. Also, from the keyword grammar 104, the readings of “Taro Yamada”, “No”, and “David King” are acquired as subword strings, and these are combined to register the reading of the registered word “y a m a d o t o o n o d a b i d e o o”.

  FIG. 9 shows registration processing when a subword string is obtained as a recognition result. In the recognition result of the subword grammar, the subword string of the recognition result is used as it is for reading the registered word. As for the notation of the registered word, information generated by the confirmation information generation unit 106 based on the registration state is registered as shown in FIG.

  As described above, when the registered state is called up by voice recognition, the same confirmation information as that at the time of registration can be presented. FIG. 10 shows a processing flow at the time of calling in the present embodiment. When voice is input in step S1001 at the time of calling, the voice recognition unit 102 performs voice recognition in step S1002. At the time of calling, a registration grammar 109 different from that at the time of registration is used. In step S1003, the confirmation information generation unit 106 acquires confirmation information from the recognition result. Since the confirmation information is registered in the notation of the registration grammar 109 shown in FIGS. 8 and 9, this is acquired. At this time, the confirmation information generation unit 106 may further add confirmation information according to the situation at the time of calling in addition to the acquired confirmation information. The presenting unit 107 presents the confirmation information thus obtained in step S1004. Through the above steps, the CPU 1 as the control means reads the state attribute from the registered tie information including the word ID corresponding to the recognition result in step S1002, and changes the system state to the state relating to the call desired by the user. Transition.

  In the present embodiment, the confirmation information generated at the time of registration is described as being registered in the notation of the registration grammar 109. However, the present invention is not limited to this. For example, in the registration status information 110, for each status, Confirmation information may be registered by adding an item of confirmation information.

  According to the first embodiment described above, the method for generating confirmation information is changed between the case where the recognition result based on the keyword string is selected and the case where the recognition result based on the subword string is selected. As a result, it is possible to present information relating to the state registered to the user in a natural manner while allowing arbitrary words to be registered.

(Embodiment 2)
In the first embodiment, as shown in the processing flow of FIG. 3, when the state is registered, the confirmation information generation method is switched according to the recognition result of the keyword string and the subword string. However, the present invention is not limited to this, and the confirmation information generation method may be switched and presented when the state is called.

  In this case, a registration grammar flag indicating whether a word to be registered in the registration grammar 109 or the registration status information 110 is a word obtained from either the keyword grammar or the subword grammar is provided (for example, the keyword grammar is 0, the subword grammar). Is represented by 1 and a flag). At the time of registration, a flag is set according to the recognition result.

  A processing flow at the time of calling is shown in FIG. In the figure, first in step S1101, the voice input unit 101 inputs voice. Next, in step S1102, the speech recognition unit 102 recognizes speech using the registered grammar 109. In step S1103, a registration grammar flag is acquired. If the registration grammar flag represents keyword grammar in step S1104, the confirmation information generation unit 106 generates confirmation information based on the recognition result notation in step S1105. On the other hand, when the registration grammar flag represents the subword grammar, the confirmation information generation unit 106 generates confirmation information based on information obtained from the state corresponding to the recognition result in step S1106.

(Embodiment 3)
In the above embodiment, confirmation information is generated based on the program information 203 when a subword string is obtained as a recognition result. On the other hand, in this embodiment, when a subword string is obtained, the voice uttered by the user is reproduced as confirmation information. In the present embodiment, the presentation method of the presentation unit 107 is presentation by audio output via the speaker 6.

  FIG. 12 shows a processing flow for registering a state by voice recognition in this embodiment. Steps S1201 to S1203 are the same as steps S301 to S303 in FIG. In step S1203, if the recognition result is a keyword string, in step S1204, a voice of confirmation information is generated by voice synthesis based on the keyword string of the recognition result. On the other hand, if the recognition result is a subword string, in step S1205, the input voice input by the voice input unit 101 and used by the voice recognition unit 102 is used as confirmation information. For example, when the user utters “Yama-chan King David”, the input voice “Yama-chan King David” is used as confirmation information. Furthermore, synthesized speech may be connected before and after that. In step S <b> 1207, the registration unit 108 associates the recognition result with the state and registers them in the registration grammar 109 and the registration state information 110.

  FIG. 13 shows how confirmation information is generated based on input voice. In the figure, “Yama-chan's David King” is an arbitrary utterance of the user, so the input voice is used as it is. Since the preceding and following “programs” and “Are you sure?” Are standard sentences, they are recorded voices that the device has in advance or synthesized voices synthesized from texts. In step S1206, the presentation unit 107 outputs the confirmation information sound, using the sound formed by connecting this to the input sound as confirmation information.

(Embodiment 4)
In the embodiment described above, speech recognition based on the keyword grammar and subword grammar is executed, and the generation method of the confirmation information is switched depending on which grammar yields the recognition result with the highest score.

  However, since the subword grammar can recognize any utterance, even if the utterance is applicable to the keyword grammar, the subword string may obtain a higher score. For example, for the utterance “Taro Yamada of King David” in FIG. 6, the subword string may obtain a higher score.

  Therefore, in the present embodiment, weights are assigned to the recognition result score based on the keyword grammar and the recognition result score based on the subword grammar based on the past history. For example, in the past recognition results, when the number of times that the speech recognition unit 102 has output a keyword string exceeds the number of times that the speech recognition unit 102 has output a subword string, a predetermined constant weight is added to the recognition result. However, the past history is not limited to the ratio of the number of times, and may be frequency information, for example. Further, the method of applying the weight is not limited to a constant, and any method may be used.

  By applying weights in this way, it is possible to select from the past history whether the user can easily register a keyword string or a subword string.

(Embodiment 5)
When the user inputs a voice, the keyword registered in the keyword grammar may be displayed to the user. By displaying the keyword, it is possible to prompt the user to speak with the keyword as much as possible while accepting any utterance with the subword.

(Embodiment 6)
In the above-described embodiment, the keyword grammar is generated by acquiring the keyword 204 using all of the program information 203 obtained from the program 202 selected on the electronic program guide 201. Instead, it is also possible to generate a keyword grammar using only the keywords displayed to the user even in the same state, that is, to change the keywords to be acquired according to the display to the user.

  FIG. 14 shows how the keyword to be acquired is changed according to the display to the user. In the figure, reference numeral 1401 denotes a state in which a program on the program guide screen is selected. 1402 displays detailed information of the program on the screen while the same program is selected. 1403 is a keyword acquired in 1401. 1404 is a keyword acquired in 1402. As shown in the figure, in 1401, since the program information obtained from the display by the user is a channel number, a broadcasting station name, a date and time, and a program name, only the program information is acquired as a keyword 1403. On the other hand, since a genre and performers are also displayed as program information in 1402, these are also added and acquired as a keyword 1404.

  Thus, by acquiring a keyword according to information displayed to the user and registering it in the keyword grammar, information that can be uttered by the user as a keyword can be narrowed down and registered with high accuracy.

(Embodiment 7)
In the above-described embodiment, the keyword grammar 104 is generated by acquiring the keyword from the state registered by the keyword grammar generation unit 103. However, the keyword grammar generation unit 103 is not essential, and the keyword grammar 104 may be generated in advance by another device for each state. When the recognition result of the speech recognition unit 102 is a keyword string, confirmation information is generated based on a keyword. On the other hand, when the recognition result is a subword string, confirmation information is generated based on state information. It is an important feature of the invention.

(Other embodiments)
As mentioned above, although embodiment of this invention was explained in full detail, this invention may be applied to the system comprised from several apparatuses, and may be applied to the apparatus which consists of one apparatus.

  In the present invention, a program for realizing each function of the above-described embodiments is supplied directly or remotely to a system or apparatus, and a computer included in the system or apparatus reads and executes the supplied program code. Can also be achieved.

  Accordingly, since the functions and processes of the present invention are implemented by a computer, the program code itself installed in the computer also implements the present invention. That is, the computer program itself for realizing the functions and processes is also one aspect of the present invention.

  In this case, the program may be in any form as long as it has a program function, such as an object code, a program executed by an interpreter, or script data supplied to the OS.

  Examples of the computer-readable recording medium for supplying the program include a flexible disk, a hard disk, an optical disk, a magneto-optical disk, an MO, a CD-ROM, a CD-R, and a CD-RW. Examples of the recording medium include a magnetic tape, a non-volatile memory card, a ROM, a DVD (DVD-ROM, DVD-R), and the like.

  The program may be downloaded from a homepage on the Internet using a browser on a client computer. That is, the computer program itself of the present invention or a compressed file including an automatic installation function may be downloaded from a home page to a recording medium such as a hard disk. Further, it is also possible to divide the program code constituting the program of the present invention into a plurality of files and download each file from a different home page. That is, a WWW server that allows a plurality of users to download a program file for realizing the functions and processing of the present invention on a computer may be a constituent requirement of the present invention.

  The program of the present invention may be encrypted and stored in a computer-readable storage medium such as a computer-readable CD-ROM and distributed to users. In this case, only the user who cleared the predetermined condition is allowed to download the key information to be decrypted from the homepage via the Internet, decrypt the program encrypted with the key information, execute it, and install the program on the computer May be.

  Further, the functions of the above-described embodiments may be realized by the computer executing the read program. Note that an OS or the like running on the computer may perform part or all of the actual processing based on the instructions of the program. Of course, also in this case, the functions of the above-described embodiments can be realized.

  Furthermore, the program read from the recording medium may be written in a memory provided in a function expansion board inserted into the computer or a function expansion unit connected to the computer. Based on the instructions of the program, a CPU or the like provided in the function expansion board or function expansion unit may perform part or all of the actual processing. In this way, the functions of the above-described embodiments may be realized.

It is a block diagram which shows the structure of the speech recognition function in the information processing apparatus which concerns on Embodiment 1 of this invention. It is a block diagram which shows the hardware constitutions of the information processing apparatus which concerns on Embodiment 1 of this invention. It is a figure showing the process which produces | generates the keyword grammar in Embodiment 1 of this invention. It is a flowchart showing the flow of the state registration process in Embodiment 1 of this invention. It is a figure showing keyword grammar. It is a figure showing subword grammar. It is a figure explaining the process which produces | generates and shows confirmation information based on the recognition result of keyword grammar in Embodiment 1 of this invention. It is a figure explaining the process which produces | generates and shows confirmation information based on program information, when the recognition result turns into a subword sequence in Embodiment 1 of this invention. It is a figure explaining the process which registers the recognition result and state of keyword grammar in registration grammar and registration status information in Embodiment 1 of this invention. It is a figure explaining the process which registers the recognition result and state of subword grammar in registration grammar and registration status information in Embodiment 1 of this invention. It is a flowchart which shows the flow of the call process of a state in Embodiment 1 of this invention. It is a flowchart which shows the flow of the calling process of the state in Embodiment 2 of this invention. It is a flowchart which shows the flow of the registration process of the state in Embodiment 3 of this invention. It is a figure explaining the process of the confirmation information generation based on the input sound in Embodiment 3 of this invention. It is a figure explaining the process which changes the keyword acquired according to the display to a user in Embodiment 6 of this invention. It is a figure showing the application which registers the program on the electronic program guide of a television with an audio | voice. It is a figure showing the speech recognition result using each of keyword grammar and subword grammar.

Explanation of symbols

101 voice input unit 102 voice recognition unit 103 keyword grammar generation unit 104 keyword grammar 105 subword grammar 106 confirmation information generation unit 107 presentation unit 108 registration unit 109 registration grammar 110 registration status information

Claims (13)

An information processing apparatus capable of registering a state of an application selected by a user according to a user's registration operation, calling the registered state according to a user's call operation, and transitioning to the state.
Voice input means for inputting user's voice;
Recognition of a person who obtains a good recognition score by performing speech recognition using a keyword grammar based on a keyword related to the state and speech recognition using a subword grammar based on a subword for the speech input from the speech input means Voice recognition means for outputting the results;
When the speech recognition means outputs a recognition result of speech recognition using the keyword grammar, confirmation information composed of a keyword string of the recognition result is presented to the user, while the speech recognition means When outputting a recognition result of speech recognition using grammar, a presentation means for presenting confirmation information made up of attribute information associated with the state to the user;
Registration means for registering the recognition result output by the voice recognition means and the state in association with each other;
An information processing apparatus comprising:   The information processing apparatus according to claim 1, further comprising a keyword grammar generation unit configured to generate the keyword grammar based on a keyword extracted from the attribute information.   When the speech recognition means outputs a recognition result of speech recognition using the keyword grammar, the presenting means outputs a synthesized speech of the keyword sequence of the recognition result, while the speech recognition means uses the subword grammar. The information processing apparatus according to claim 1 or 2, wherein when the recognition result of the used voice recognition is output, the voice input to the voice input unit is output.   The speech recognition means weights the recognition score based on a frequency at which a recognition result of speech recognition using the keyword grammar is output and a frequency at which a recognition result of speech recognition using the subword grammar is output. The information processing apparatus according to any one of claims 1 to 3, wherein the information processing apparatus is characterized in that:   5. The display device according to claim 1, further comprising a display unit configured to display a keyword registered in the keyword grammar to a user when a voice is input by the voice input unit. The information processing apparatus described.   The information processing apparatus according to claim 2, wherein the keyword grammar generation unit extracts a keyword corresponding to attribute information displayed by the application in the state. The registration means includes
A registered grammar including a word ID for identifying a registered word, a notation, and a reading for a registered word corresponding to the recognition result output by the voice recognition unit;
Registration state information including a registration ID for identifying the state related to registration, a state attribute representing the content of the state, and a word ID of a registered word in the registration grammar corresponding to the state;
The information processing apparatus according to any one of claims 1 to 6, characterized in that The registration means includes
When the speech recognition means outputs a recognition result of speech recognition using the keyword grammar, the keyword string that is the recognition result is represented as a registered word in the registered grammar,
When the speech recognition means outputs a recognition result of speech recognition using the subword grammar, the subword string that is the recognition result is read as a registered word in the registered grammar. The information processing apparatus described. When calling up the registered state,
The voice recognition means performs voice recognition using the registered grammar for the voice input by the voice input means,
The presenting means presents confirmation information at the time of calling based on the recognition result output by the voice recognition means,
The information processing apparatus transitions to the state based on the state attribute read from the registered state information including a word ID corresponding to a recognition result output by the voice recognition unit. Information processing device.   At the time of calling up the state, the presenting means confirms when the voice recognition means outputs a recognition result of voice recognition using the keyword grammar at the time of registration of the state, and is constituted by a keyword string of the recognition result When the speech recognition means outputs the recognition result of speech recognition using the subword grammar at the time of registration of the state while presenting information to the user, the confirmation made up of attribute information associated with the state The information processing apparatus according to claim 9, wherein information is presented to a user. An information processing method in an information processing apparatus capable of registering a state of an application selected by a user according to a user's registration operation and calling the registered state according to a user's calling operation and transitioning to the state. ,
A voice input step for inputting the user's voice;
Recognition of a person who obtains a good recognition score by performing speech recognition using a keyword grammar based on a keyword related to the state and speech recognition using a subword grammar based on a subword for the speech input in the speech input step A speech recognition step for outputting the results;
When a recognition result of speech recognition using the keyword grammar is output in the speech recognition step, confirmation information including a keyword string of the recognition result is presented to the user, while the speech recognition step When a recognition result of speech recognition using subword grammar is output, a presentation step of presenting confirmation information made up of attribute information associated with the state to the user;
A registration step in which the recognition result output in the voice recognition step and the state are associated with each other and registered in the storage unit;
An information processing method characterized by comprising:   A program for causing a computer to execute the information processing method according to claim 11.   A computer-readable storage medium storing the program according to claim 12.

Images