JP2005332404A - Content providing system - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a content in accordance with the main topic of an object person in real time without requiring huge hardware resources. <P>SOLUTION: The content providing system is provided with a means for recognizing a setting keyword by recognizing voice captured from a microphone, a means for recording a recognition time in association with the keyword or the like or for updating the recorded recognition time to the recognition time of the above recognition in response to the recognition of the keyword, a means for recognizing a prescribed key set wherein a elapsed time from the oldest recognition time to the latest recognition time in the recognition of a plurality of kinds of keywords configuring the key set is within a recognition setting time, a means for calling the content set in association with the recognized prescribed key set from a content DB, and a display type device which is installed near a table and reproduces and outputs the content called in substantially real time with the latest recognition time. <P>COPYRIGHT: (C)2006,JPO&NCIPI

Description

  The present invention provides a content providing system that captures voice from, for example, a conversation of a speaker or the like who is provided with content, specifies a topic based on a keyword included in the voice, and provides content such as an advertisement corresponding to the topic in real time About.

  Patent Document 1 discloses a publicly known document relating to a configuration in which a voice is taken from a dialogue of a recipient or the like who is provided with content, a topic is specified based on a keyword included in the voice, and content corresponding to the topic is provided in real time. is there. Patent Document 1 discloses information that recognizes a keyword by recognizing a dialogue such as a video conference, identifies a topic based on the keyword, determines content based on the identified topic, and outputs the content to a display or the like. A providing device is disclosed, and the information providing device identifies a topic corresponding to the keyword or acquires a distribution of application frequencies of a plurality of keywords appearing in the topic in accordance with recognition of a characteristic keyword, A set appearance frequency distribution with the most similar appearance frequency distribution is determined, and a topic corresponding to the determined set appearance frequency distribution is specified as a corresponding topic.

  Further, there is Patent Document 2 as another known document. In Patent Document 2, a keyword is recognized by voice recognition of a dialog of a videophone and recognized from advertisements distributed in advance and stored in a videophone terminal. A videophone terminal that calls the corresponding advertisement and outputs the advertisement on a display is disclosed.

  Further, there is Patent Document 3 as a publicly known document related to the above configuration. In Patent Document 3, a voice recognition sensor detects a voice of a surrounding person, analyzes the frequency band of the voice and a fault cloak, and analyzes the surrounding person. An advertisement display device is disclosed that displays the advertisement information according to the determination result of the sex and age group on the display.

JP-A-11-203295 JP 2002-271507 A JP 2002-244606 A

  However, as disclosed in Patent Documents 1 and 2, according to the recognition of a keyword, content corresponding to the keyword is provided, and content corresponding to a keyword that appears by chance or a keyword that appears in an incidental topic is provided. Therefore, it is difficult to specify a central topic of a speaker and provide content in accordance with the central topic and interest in real time. Also, as in Patent Document 1, the distribution of the application frequency of a plurality of keywords appearing in the topic is acquired, the set appearance frequency distribution with the most similar appearance frequency distribution is determined, and the determined set appearance frequency distribution is supported. In this configuration, the central topic of the conversation can be accurately identified, but a large amount of hardware resources are required to calculate the similarity of the appearance frequency distribution. Costs are increased, and the speed of provision of content and real-time performance are impaired due to accumulation of keyword appearance frequency data and similarity calculation, and there are many situations in which content is provided after a central topic or interest changes. . Such a defect becomes very remarkable as the number of keywords and topics to be set increases.

  The present invention is proposed in view of the above problems, and provides a content providing system capable of identifying a central topic of a speaker and quickly providing content along the central topic in real time. With the goal. Another object of the present invention is to provide a content providing system capable of realizing real-time content provision suitable for the central topic or interest of the recipient at low cost without requiring a large amount of hardware resources. There is.

  The content providing system of the present invention is configured by a computer or a network computer, recognizes a voice captured from a microphone, recognizes a keyword set from the recognized voice, and determines the keyword according to the keyword recognition. A key set is configured and means for recording the recognition corresponding time in correspondence with the keyword or in correspondence with the key corresponding to the keyword, and updating the recognition corresponding time recording according to the recognition of the corresponding keyword Means for recognizing a predetermined key set whose recognition elapsed time from the earliest recognition correspondence time to the last recognition correspondence time for a plurality of types of keywords or keys is within the recognition setting time, and correspondingly set to each key set The content database that stores the content A means for calling content set in correspondence with a predetermined key set, means for playing back the called content in substantially real time, and means for outputting the content to be played back, To do. The means for outputting the content can be a display, a speaker, or both, and the content such as an image and / or sound to be reproduced corresponding to the output means is set in the content database.

  Further, the content providing system of the present invention is constituted by a computer or a network computer, recognizes a voice taken in from a microphone, recognizes a keyword set from the recognized voice, and according to the keyword recognition, A key set that records a recognition time of a keyword corresponding to the keyword or a key corresponding to the keyword, and updates a record of the recognition time according to the recognition of the corresponding keyword; Means for recognizing a predetermined key set in which the recognition elapsed time from the earliest cognitive response time to the last cognitive response time for the plurality of types of keywords or keys constituting the key is within the cognitive setting time, and the corresponding setting From the content database storing stored content. Characterized in that it comprises and means for invoking the content with a predetermined key set is associated Configuration, and means for transmitting the call content cognitive corresponding time and near real time said last to the playback output device.

  For example, according to the keyword recognition, the content providing system records the keyword recognition corresponding time corresponding to the keyword or the key corresponding to the keyword, and recognizes the same keyword or the keyword corresponding to the same key. In response to the update, the recorded cognitive response time is updated, and for a key set composed of a plurality of types of keywords or keys, a predetermined key set that recognizes all of a plurality of types of keywords or keys within a set recognition time is recognized. Or by recognizing a predetermined key set in which the recognition elapsed time from the earliest recognition correspondence time to the last recognition correspondence time for all keywords or all keys is within the recognition setting time, according to the passage of time Track the topics of speakers that change in complex ways in parallel at any given time. It is possible to easily, flexibly and quickly identify a highly concentrated or central topic or interest that does not require complex processing, and provide content using multiple types of keywords or keys. The topic and interest can be specified with sufficient and sufficient accuracy. Further, for example, by calling a content corresponding to a key set for each topic for which a keyword or key corresponding to the topic is set, the content is reproduced and output in substantially real time, or transmitted to a reproduction output device. Promptly accept psychologically very acceptable content according to the topic and interest to the speaker, the subject listening to the speaker's dialogue, or the subject listening to the speaker's dialogue with the playback output device Can be provided in real time. In addition, it requires a lot of hardware resources because it recognizes the set keyword and records / updates the cognitive response time and recognizes the key set that recognized the keyword or key within the recognition set time. It can be realized at a low cost, and a content provision system can be installed ubiquitously. In addition, for example, when the number of topics or keywords set in the key set increases, the speed of content provision and real-time performance are not increased while there is no increase in the required hardware resources or cost. Can be secured. Furthermore, even when the number of key sets or keywords increases due to the recording and updating of the cognitive response time, complicated processing is required for the recognition elapsed time from the earliest cognitive response time to the last cognitive response time. For example, complicated processing such as time measurement according to keyword recognition for each of a large number of key sets corresponding to the recognized keyword is not necessary. Furthermore, since the content is provided based on the recognition of the set keyword, it is possible to provide the content moderately and prevent excessive information provision. Furthermore, it is possible for the speaker to consciously interact with a certain topic and to receive or provide the content corresponding to the topic. Sex and entertainment can be improved.

  Furthermore, the content providing system of the present invention is characterized by comprising means for deleting the record of the cognitive response time when the cognitive set time has elapsed from the current time. By deleting the record of cognitive response time when the cognitive set time has elapsed from the current time, hardware resources can be used effectively, and each keyword or key belonging to the keyset is recognized or cognitively supported. According to the time recording, it is possible to automatically recognize the key set for calling the content without requiring any special post-processing.

  Furthermore, the content providing system of the present invention is characterized in that there are three or more types of keywords or a plurality of types of keys constituting the key set. Multiple types of different keywords or keys that make up the key set are appropriate as long as there are two or more types, such as three, four, five, six, etc. It is possible to identify and provide content suitable for the topic, which is preferable. The same number of keywords or keys that make up the key set may be set for each key set, but the balance between the identification of the target topic and the content of the content to be provided, and the appearance of keywords that are set corresponding to the key set A required number of keywords or keys may be set as appropriate for each key set according to the possibility of performing the operation.

  Furthermore, the content providing system of the present invention is characterized in that a means for recognizing a captured voice and recognizing a keyword set from the recognized voice recognizes only the set keyword. For example, by using word spotting speech recognition technology and recognizing only the set keywords, content can be provided in real time with a fast response speed, the system can be further reduced in cost, and predetermined keywords can be provided. Since it does not recognize anything other than, it is possible to protect the privacy of the speaker who captures the voice.

  Further, the content providing system according to the present invention includes means for determining whether the content is being reproduced or transmitted based on the recognition of the predetermined key set, and the recognition of the predetermined key set or the reproduction or transmission of the content. Elapsed time from the determination to the end of content playback or transmission end is within the set time. Based on the recognition of playback end or transmission end, the content set corresponding to the predetermined key set is played back or played back to the playback output device. And a means for not playing back the content set corresponding to the predetermined key set or transmitting to the playback output device when the end of playback within the set time or the recognition of the end of transmission cannot be obtained. It is characterized by that. For example, when playing back or transmitting content corresponding to the key set recognized later, if the content corresponding to the key set recognized earlier is played back or transmitted, When the reproduction or transmission of the content of the current key ends, the content of the subsequent key set is played or transmitted. When the content does not end within the set time, the content of the subsequent key set is not played or transmitted. , The topic or interest of the speaker is likely to have changed over time. The content is not provided after a predetermined period of time. Only content that is psychologically acceptable can be reliably provided.

  Further, the content providing system of the present invention includes a unit for determining whether the content is being reproduced or transmitted based on the recognition of the predetermined key set, and the predetermined key based on the determination of the content being reproduced or transmitted. A means for recording and holding the set recognition, and when there is a recognition of a key set that has already been recorded and held, a means for updating and recording and holding the predetermined key set, and a recognition of the end of playback or transmission of the content And a means for reproducing or transmitting the content set corresponding to the predetermined key set recorded and held to the reproduction output device. The key set recognized during content playback or transmission is sequentially updated and recorded, and the content corresponding to the last key set recognized closest to the current time is provided. Adapting to changes over time, it is possible to provide content that is psychologically acceptable to the target audience according to the topic and interest of the speaker at the time of providing the content or in the near future, and also recognizes the key set. It is possible to effectively use the hardware resources by updating and storing the information.

  Further, the content providing system of the present invention comprises a computer or a network computer, recognizes a voice captured from a microphone, recognizes a keyword set from the recognized voice, a plurality of types of keywords or the keywords, Means for recognizing a predetermined key set based on recognition of a keyword from a key set composed of corresponding keys; means for determining whether content is being reproduced or transmitted based on recognition of the predetermined key set; The predetermined key set based on the recognition of the end of reproduction or the end of transmission in which the elapsed time from the determination during the reproduction or transmission of the content to the recognition of the end of content reproduction or the end of transmission is within a set time Play content that is set to support When transmitting to the playback output device and when the end of playback within the set time or the recognition of the end of transmission cannot be obtained, the content set in correspondence with the predetermined key set is not played back or transmitted to the playback output device. Means.

  The above content provision system recognizes a key set based on keyword recognition by voice recognition from a key set composed of a plurality of types of keywords or keys, thereby paralleling topics of speakers that change in a complex manner over time. Can be tracked at any time, and can easily and flexibly and quickly identify highly concentrated or central topics and interests without requiring complex processing. You can identify interests. Further, for example, by calling a content corresponding to a key set for each topic for which a keyword or key corresponding to the topic is set, and reproducing / outputting the content or transmitting it to a reproduction output device, the central topic or interest of the speaker is obtained. Thus, content that is psychologically acceptable according to the user can be provided to a speaker, a subject who is listening to the conversation of the speaker, or a subject who is listening to the conversation of the speaker using the reproduction output device. Furthermore, since it is a simple configuration that recognizes a set keyword and a key set that recognizes a keyword or key to which it belongs, it can be realized at low cost without requiring a large amount of hardware resources, and ubiquitous content It is possible to install a provision system. For example, even when the number of topics or keywords set with a key set increases, there is no increase in required hardware resources, an increase in cost, or a very slight increase. it can. In addition, the content of the speaker's topic or interest is likely to have changed over time. After a predetermined time has elapsed, the content is not provided, and the subject is subject to the content and the latest speaker's topic and interest. It is possible to reliably provide only psychologically acceptable content. Furthermore, since the content is provided based on the recognition of the set keyword, it is possible to provide the content appropriately and prevent excessive information provision. Furthermore, it is possible for the speaker to consciously interact with a certain topic and to receive or provide the content corresponding to the topic. Sex and entertainment can be improved.

  Further, the content providing system of the present invention comprises a computer or a network computer, recognizes a voice captured from a microphone, recognizes a keyword set from the recognized voice, a plurality of types of keywords or the keywords, Means for recognizing a predetermined key set based on keyword recognition from a key set composed of corresponding keys, means for determining whether the content is being reproduced or transmitted based on recognition of the predetermined key set, Based on the determination during playback or transmission, the recognition of the predetermined key set is recorded and held, and if there is a recognition of the key set that has already been recorded and held, it is updated to the recognition of the predetermined key set and recorded. Based on the means to hold and the recognition of the end of playback or transmission of content. It can, characterized in that it comprises a means for transmitting a predetermined key set being the record holder to reproduce or playback output apparatus content that is compatible set.

The above content provision system recognizes a key set based on keyword recognition by voice recognition from a key set composed of a plurality of types of keywords or keys, thereby paralleling topics of speakers that change in a complex manner over time. Can be tracked at any time to identify highly concentrated or central topics and interests easily, flexibly and quickly without complicated processing, and with the accuracy required to provide content, You can identify interests. Further, for example, by calling a content corresponding to a key set for each topic for which a keyword or key corresponding to the topic is set, and reproducing / outputting the content or transmitting it to a reproduction output device, the central topic or interest of the speaker is obtained. Thus, content that is psychologically acceptable according to the user can be provided to a speaker, a subject who is listening to the conversation of the speaker, or a subject who is listening to the conversation of the speaker using the reproduction output device. Furthermore, since it is a simple configuration that recognizes a set keyword and a key set that recognizes a keyword or key to which it belongs, it can be realized at low cost without requiring a large amount of hardware resources, and ubiquitous content It is possible to install a provision system. For example, even when the number of topics or keywords set with a key set increases, there is no increase in required hardware resources, an increase in cost, or a very slight increase. it can.
In addition, it adapts to changes in the speaker's topic and interest over time with a simple structure, and provides content that is psychologically acceptable to the target audience according to the topic and interest of the speaker at the time of providing the content or in the immediate vicinity. In addition, it is possible to effectively use hardware resources by updating the key set recognition and recording and holding it. Furthermore, since the content is provided based on the recognition of the set keyword, it is possible to provide the content moderately and prevent excessive information provision. Furthermore, it is possible for the speaker to consciously interact with a certain topic and to receive or provide the content corresponding to the topic. Sex and entertainment can be improved.

  When the content providing system of the present invention is configured by a network computer, any appropriate network capable of transmitting data, such as a LAN, the Internet, a communication network, and a broadcast network, can be used as the network. Alternatively, a dedicated or general purpose, internal or external network may be used. Furthermore, each constituent means of the present invention can be provided in an appropriate computer of a plurality of computers connected by a network, and the required constituent means and other constituent means are provided in a computer at another place such as a remote place. It is good also as a structure. For example, the content providing system is a personal computer, a video phone, a mobile phone, a mobile information terminal, a TV mobile phone, a TV receiver, a TV receiver, etc., with one or more display-type devices, speaker-type devices, and a configuration capable of real-time voice chat. And other terminals such as set-top boxes or game machines, or remote servers connected to various devices via a communication network, etc. Based on the call command transmitted to the server and based on the call command received by the server, the corresponding content is extracted from the content DB and transmitted to the terminal, or based on the audio data captured by the various terminals from the microphone or the captured audio Send recognized keyword data to server A configuration in which predetermined processing is executed based on audio data or keyword data received by the server, content corresponding to a predetermined key set is extracted from the content DB and transmitted to the terminal, or audio that the server captures from the microphone It is possible to adopt a configuration in which predetermined processing is executed on the basis of the content, content corresponding to the predetermined key set is extracted from the content DB, and transmitted to the terminal. In addition, you may provide each structural means of a content provision system integrally or in one place in the above various terminals or various apparatuses.

  In addition, each means such as a means for reproducing and outputting content can be provided at an appropriate place. For example, a device that captures sound from a microphone and reproduces and outputs an image and sound or an image or sound can be provided at a store such as a restaurant. , Taxis, buses, trains, airplanes, etc., department stores, shopping malls, entertainment facilities, etc., gathering places such as station gathering places, seats and passenger seats, near tables, on tables, near meeting places, etc. Installed in the vicinity of a space where the target person or customer stays for a certain amount of time and interacts with the other target person or the other party on the mobile phone, and recognizes the keyword from the voice of the target person or customer's conversation. A configuration that provides content corresponding to the topic of conversation in real time, a configuration in which the device is installed at home, or a transfer of the device. It can be configured to be portable, etc., providing advertisement information, guidance information, tourism information, entertainment programs, etc. content at an appropriate time, with respect to the entertainment and convenience of the target audience and customers, information content The degree of impression can be increased, and when it is installed in a store, vehicle, facility, etc., the rate of attracting customers can be improved. In particular, when a system is configured with a thin display type device or the like, and installed in public spaces such as stores, facilities, vehicles, etc. It is possible to increase the marketing effect by providing information by capturing as many times as possible.

  Further, the means for reproducing and outputting the microphone and the content is not limited to a configuration in which the display, the speaker, and the microphone are integrally provided, such as a display-type device or a mobile phone, and for example, a store A reproduction output means having a large screen display and a speaker provided at a predetermined distance from the seat and a microphone provided in the vicinity of the seat or in the vicinity of the table corresponding to each seat unit on which one or a plurality of persons can sit. Installed so that seat customers can view and process the sound captured by each microphone separately to execute key set recognition processing, and according to the recognition of the key set based on the sound captured by a predetermined microphone, the reproduction output means A system microphone that plays and outputs content It is good also as a structure which provides a one-to-one correspondence with the phone and the means for reproducing and outputting the content. For example, a microphone is provided in a seat unit that can seat two to four people such as an amusement facility, and the seat unit. Based on the sound taken from one microphone provided with a reproduction output means having a display and a speaker so that the target person can view each seat, and a reproduction output means corresponding to each seat is provided for each seat. It may be configured to provide one-to-multiple means for reproducing and outputting content from a single system, such as a configuration for reproducing and outputting the same content, and to reproduce and output content from a single system microphone. It is good also as a structure which provides the means to respond | correspond in multiple-to-multiple correspondence.

  In addition, when a speaker is provided as a means for outputting audio content, for example, a speaker that outputs at the ear of the subject sitting on a reclining chair and outputs at a volume that can be heard only by the subject, or an audio that exceeds the ear of the subject. With a directional speaker or the like that is transported by sound waves, it is possible to exclude in advance the output sound of a speaker placed at a short distance from the microphone from the sound captured by the microphone. It is preferable that recognition can be continuously performed. The microphone that captures the sound is appropriate as long as it can pick up the sound of the target person by excluding the sound other than the target person. For example, a directional microphone having directivity toward the mouth of the target person or a predetermined volume It is assumed that a microphone or the like that acquires only the above-described sound is acquired, and the sound of one or a plurality of target persons around is acquired. Further, for example, a structure for capturing and recognizing a target person's voice at a predetermined position using an existing means for specifying the position of a sound source, such as specifying the position of a sound source by performing frequency analysis on sounds acquired from a plurality of microphones. Etc.

  Also, display-type devices and speaker-type devices such as display-type devices and speaker-type devices placed in public spaces such as in the vicinity of seats in tables, near tables, on tables, near seats in facilities, near seats in public vehicles, and wall surfaces of meeting places Various devices provided with means for reproducing and outputting content provided in a space, etc., various devices provided with a microphone provided in a public space, etc., or various devices provided with means for reproducing and outputting a microphone and content provided in a public space etc. In addition, a sensor such as an infrared sensor that detects the presence of a target person at a predetermined target location is provided, and a predetermined unit such as a control unit of various apparatuses is configured as a control program in response to detection of the sensor for the presence of the target person at the predetermined location. Collaborate to output a predetermined control command, and based on the control command, content or settings to be transmitted Plays and outputs specified content that is remembered, or recognizes a keyword from the voice captured from the microphone based on the control command, calls the content based on the recognized keyword, and plays and outputs the content providing process. It is good also as a structure etc. which the predetermined part of a provision system performs.

  In addition, various devices including means for reproducing and outputting content such as a display-type device and / or a speaker-type device do not provide content other than when providing content based on keyword recognition, or as appropriate. It is possible to adopt a configuration for providing the content. For example, when the playback processing unit of various devices plays back normal content that is stored or transmitted, stops playback of the content, or if playback of the content ends within a predetermined time, or the playback of the content ends Later, the content based on keyword recognition is played back. The normal content is, for example, a program such as a television or radio, advertisement information, guidance information, and the like, and also includes a menu screen in a display type device. Alternatively, the content based on the keyword recognition may be provided in parallel with the normal content. For example, the content based on the keyword recognition is displayed on a part of the display, or based on the normal content and the keyword recognition. A configuration that displays content in a split screen, a configuration that displays the content image based on keyword recognition while the output audio is normal content audio, and a normal content that displays output audio as content audio based on keyword recognition It is possible to adopt a configuration in which an image of the content is displayed on a display.

  Further, the content providing process based on the recognition of the keyword from the voice of the content providing system is executed based on, for example, the execution request of the content providing process based on the recognition of the keyword or the ON of the execution switch, If the execution switch is not turned on, the content providing process is not executed, so that it is possible to protect the privacy of the subject's conversation and the like. For example, a content providing system is configured by connecting a server and a display-type device via a network, etc., and a menu screen that is stored or held by the display-type device is displayed on a touch panel display and displayed on the menu screen. In response to the designated input of the execution request button of the content providing process, a predetermined unit of the system such as the control unit of the display type device or the control unit of the server cooperates with the control program, and outputs a predetermined execution control command, Based on the execution control command, a keyword is recognized from a voice captured from a microphone installed in or near a display type device, the content is called based on the recognized keyword, and the called content is displayed on the display type. Content provision processing to be played and output on the device is predetermined There may be configured such that execution.

  Further, the content providing system may be a game system or the like that provides highly entertaining content free of charge, such as providing content in which a specific character appears based on recognition of a combination of specific keywords. For example, when a favorite keyword of a specific character is recognized within a set time, content in which the specific character appears is provided, and when a keyword different from the favorite keyword is recognized within a set time, The content is provided such that another character set to correspond to a different keyword appears. Further, for example, the predetermined unit of the content providing system reproduces and displays a quiz-type guidance screen stored in the predetermined storage unit on the display of the display-type device, recognizes the keyword from the voice of the subject's answer to the quiz, and the keyword It is good also as a structure which provides content, such as a specific character appearing based on recognition. Further, the processing of billing information in the case of providing for a fee is configured, for example, by configuring the content providing system by connecting a server and a display type device via a network, and in accordance with the designation input of the execution request button for the content providing process. The billing processing unit of the device or server records the start of execution of the content providing process based on the recognition of the keyword as time data such as a real time clock, and a part of the screen or the menu at the time of providing the content of the touch panel display In response to the designated input of the execution end button of the content provision process displayed on the screen or the like, the billing processing unit records the execution end time of the content provision process and obtains the elapsed usage time from the execution start time to the execution end time Content is provided by multiplying the elapsed time by a predetermined unit price per unit time that is set and stored It may be configured to calculate the compensation for the storage and store the compensation, or to transmit the compensation to a charging system that stores the compensation in a predetermined storage unit, etc. An appropriate configuration can be adopted.

  In addition, the content providing system recognizes a keyword having been recognized last with a predetermined key set recognized by the predetermined unit or having a last recognized setting time, and stores the recognized keyword in a predetermined storage unit. The image data and / or audio data stored in the setting is extracted, and immediately before the reproduction and output of the content corresponding to the predetermined key set, the reproduction processing unit or the transmission processing unit extracts the image of the keyword. The data and / or audio data may be reproduced and transmitted or transmitted to attract the attention of the subject to whom the content is provided. In addition, in the present invention, specific items of other inventions are added to each invention, or specific items of each invention are changed to specific items of other inventions, or partial effects of the present invention are exhibited. In terms of limits, those obtained by deleting specific matters of each invention and making it a superordinate concept are also included in the present invention. Besides the system category, the inventions that define similar inventions as methods and programs are also included in the present invention. It is included, and the specific matters of the invention in each category can be appropriately specified as specific matters of the invention in other categories. Furthermore, the predetermined means and the predetermined unit in the present invention are appropriately realized by a CPU that cooperates with a set program, a memory that stores programs and data, and the like.

  The content providing system of the present invention tracks the topic of a speaker that changes in a complex manner as time passes, and at the same time, the topic of high concentration or the central topic or interest that is rising in a predetermined time is complicated. Can be identified easily and quickly without the need for special processing, and topics and interests can be identified with sufficient accuracy necessary and sufficient for content provision. Provide psychologically very acceptable content along the way to speakers, subjects who are listening to the speaker's dialogue, or subjects who are listening to the speaker's dialogue on the playback output device in real time. be able to. Furthermore, it can be realized at a low cost without requiring a large amount of hardware resources with a simple configuration, and a predetermined part of the content providing system such as a content providing system or a display-type device or a microphone can be installed ubiquitously. It becomes possible. In addition, for example, when the number of topics or keywords set in the key set increases, the speed of content provision and real-time performance are not increased while there is no increase in the required hardware resources or cost. Can be secured. Furthermore, even if the number of key sets, keywords, etc. increases due to the recording and updating of the cognitive response time, complicated processing of the elapsed time from the earliest cognitive response time to the last cognitive response time is required. Without being able to get very easily. Furthermore, since the content is provided based on the recognition of the set keyword, it is possible to provide the content moderately and prevent excessive information provision. Furthermore, it is possible for the speaker to consciously interact with a certain topic and to receive or provide the content corresponding to the topic. Sex and entertainment can be improved.

  A first embodiment of the content providing system of the present invention will be described. As shown in FIG. 1, the content providing system 100 according to the first embodiment includes a microphone 101, a feature extraction unit 102, a keyword recognition unit 103, a standard feature storage unit 104, a recognition management unit 105, and a content call unit. 106, a content database (content DB) 107, a playback processing unit 108, and a display 109. The feature extraction unit 102, keyword recognition unit 103, standard feature storage unit 104, recognition management unit 105, content call The predetermined units such as the unit 106, the content DB 107, and the reproduction processing unit 108 are realized by a CPU that cooperates with a set program, a memory that stores the program and data, and the like. The physical arrangement configuration of each of the units 101 to 109 of the content providing system 100 is appropriate. For example, a display type device in which 101 to 109 are integrally provided, or a display type device in which items other than the content DB 107 are provided integrally. And a server provided with the content DB 107, the display type device transmits / receives data to / from the server, or the display type device provided with the microphone 101, the reproduction processing unit 108, and the display 109 integrally, And a configuration in which the display-type device transmits audio to the server having the feature extraction unit 102, the server executes predetermined processing, and the reproduction processing unit 108 receives content data from the server. Can also be used, and appropriate predetermined parts or predetermined It can be constructed by connecting a network device having a.

  The feature extraction unit 102 takes in the voice input from the microphone 101, performs analog / digital conversion, and extracts a feature quantity for each unit time such as a cepstrum by acoustic analysis. In addition, the standard feature storage unit 104 stores standard feature amounts set corresponding to each registered keyword, and the keyword recognition unit 103 extracts the feature extraction unit 102 by, for example, continuous DP matching. A similarity distance is calculated by comparing the feature quantity with the standard feature quantity of each keyword stored in the standard feature storage unit 104, and it is determined whether the calculated similarity distance is equal to or less than a predetermined threshold value that is set and stored. When the similarity distance is equal to or smaller than a predetermined threshold, a predetermined keyword set corresponding to the standard feature amount for which the similarity distance is calculated is recognized. In the configuration of recognizing the keyword set from the speech recognized by recognizing the speech in the present invention, for example, when using the word spotting speech recognition technology as in the above configuration, in real time with a fast response speed. It is possible to realize a system that can provide content at a low cost, and since only keywords that are determined in advance are extracted without using a special configuration, it is possible to protect the privacy of people who capture audio. Although it is preferable, it is possible to adopt a configuration for recognizing a keyword set using an appropriate voice recognition technique.

  The recognition management unit 105 obtains the recognition time of the predetermined keyword based on the measurement time of the real time clock in accordance with the recognition of the predetermined keyword in the keyword recognition unit 103, and stores and holds the recognition time of the predetermined keyword in FIG. From the recognition table or the like that stores one or a plurality of recognition time recording areas corresponding to the predetermined keyword, and the recognition time of the predetermined keyword is recorded in the recognized recognition time recording area. When the recognition time is already recorded in the time recording area based on the recognition of the predetermined keyword, the new recognition time is updated and recorded based on the new recognition of the same predetermined keyword. The recognition management table of FIG. 2 has a plurality of key sets identified by identification IDs, and each key set is set with a plurality of different keywords, and a recognition time recording area for each keyword is provided. . In the example of FIG. 2, three or four keywords are set in each key set of the recognition management table. If the number of a plurality of different keywords set for each key set is three or more, for example, it is possible to provide content more suitable for the central topic of the voice input speech. Furthermore, the recognition management unit 105 sequentially recognizes and deletes the recognized recognition times of the recorded keywords that have passed the recognized setting time that has been set and stored from the current time acquired by a real time clock or the like. Furthermore, the recognition management unit 105 recognizes the identification ID of a predetermined key set in which the recognition time is recorded for all the set keywords, and outputs the recognized identification ID to the content extraction unit 106.

  The content calling unit 106 determines whether the content is being reproduced by the reproduction processing unit 108 in response to the input of the identification ID from the recognition management unit 105. If the content calling unit 106 determines that the content is not being reproduced, the content calling unit 106 is stored in the content DB 107. The content that is set to correspond to the identification ID is called out from the content that is being played, the called content is output to the playback processing unit 108, and if it is determined that the content is being played back, The measurement of the call setting time stored in the setting is started, and when the reproduction of the content is recognized within the call setting time, the content corresponding to the identification ID is called and output to the reproduction processing unit 108, If the end of content playback is not recognized within the set time, the content corresponding to the identification ID is called up. It is made as to no. The reproduction processing unit 108 decodes and reproduces the content extracted by the content calling unit 106, for example, reproduces the content of a moving image or a still image, and the display 109 displays the image reproduced by the reproduction processing unit 108. indicate. The content of the content to be reproduced may be, for example, advertisement information, but may be information on facilities, etc., explanations of matters highly relevant to keywords belonging to the key set, programs with high relevance, quizzes, etc. In addition, it is possible to have appropriate contents having relevance to the keyword.

  The flow of processing by the content providing system 100 of the first embodiment is shown in FIG. For example, the microphone 101 of the content providing system 100 configured as a display-type device or the like takes in the voice of a person located in the vicinity of the surroundings and continuously inputs voice (S1). The input speech voice is captured by the feature extraction unit 102 and A / D converted at any time, and the feature extraction unit 102 extracts the feature amount from the speech data at every unit time (S2), and the extracted feature amount is obtained. Output to the keyword recognition unit 103. The keyword recognizing unit 103 compares the feature quantity of the input speech with the standard feature quantity stored in the standard feature storage unit 104 as needed to calculate a similarity distance (S3), and the calculated similarity distance is a predetermined threshold value. It is determined at any time whether or not (S4), and when the similar distance is equal to or smaller than a predetermined threshold, a predetermined keyword set corresponding to the standard feature amount for which the similar distance is calculated is recognized (S5). .

  The recognition management unit 105 acquires a recognition time for the recognized predetermined keyword according to the recognition of the predetermined keyword in the keyword recognition unit 103 (S6), and recognizes the recognition management table corresponding to the predetermined keyword. Recognizing the time recording area, recording the recognition time of the predetermined keyword in the recognition time storage area (S7), and if the recognition time has already been recorded in the previous recognition time recording area, The new recognition time is updated and recorded (S7). In this case, for example, a plurality of the same keywords are set in different key sets, such as the keyword A of the key set with the identification ID 0001 in the recognition management table of FIG. 2 and the keyword A of the key set with the identification ID 0003. If it is, the recognized time is recorded in the recognized time recording area for the keyword for all the key sets to which the keyword belongs, as at 12: 30: 5 in FIG. Then, the recognition management unit 105 determines at any time whether the elapsed time from the current time to the recognized time exceeds the recognized set time such as 10 minutes with respect to the recorded recognized time of each keyword (S8). If it is determined that the elapsed time until the recognition time exceeds the recognition setting time, the determined recognition time record is deleted according to the determination (S9). The recognition setting time is, for example, a predetermined time of 3 to 30 minutes, preferably a predetermined time of 5 to 15 minutes, such as the number of keywords in the key set and the predicted appearance of multiple keywords in a natural conversation. It is possible to set appropriately according to time or the like. With such a recognition setting time, a predetermined process is performed by capturing voice from a natural conversation between surrounding subjects or a conversation of a surrounding subject using a mobile phone.

  Furthermore, the recognition management unit 105 recognizes the record of the recognition time for all keywords of a predetermined key set whose elapsed time from the current time to the earliest recognition time is within the recognition setting time (S10). Accordingly, the identification ID of the predetermined key set is recognized and output to the content calling unit 106 (S11), and the record of the recognition time of all keywords in the predetermined key set is deleted and initialized. In the example of the recognition management table in FIG. 2, for example, the recognition time for the keywords A, B, and C is recorded within the recognition setting time of 10 minutes, so that the identification ID: 0001 is output to the content calling unit 106. In addition, when the identification manager 105 outputs the identification ID of the predetermined key set, the recognition management unit 105 maintains the record of the recognition time in the other key sets as it is, while maintaining the recognition time of the predetermined key set. The recording process is started, and the recording of the recognition time based on the recognition of the keyword is continuously performed. It should be noted that the number of keywords set corresponding to each key set in the recognition management table may be the same for all keyword sets. For example, the identification IDs are identified as three IDs A, B, and C in FIG. The number of keywords set for each key set, such as D, E, F, and G of ID: 0002, may be different. In this case as well, all the keys set for each key set are all set. The identification ID is output based on the recording of the recognition time within the recognition setting time for the keyword. In addition, the cognitive management unit 105 stores the cognitive setting time that is the same time uniformly for each key set, and stores the cognitive setting time that is set separately for each key set, in addition to executing the processing. And it is good also as a structure which performs the said process.

  The content calling unit 106 determines whether the content corresponding to the key set already recognized by the reproduction processing unit 108 is being reproduced based on the input of the identification ID from the recognition management unit 105 (S12). If it is determined that the content is not being played back by acquiring data that has not been played back, the content set corresponding to the identification ID is called from the content DB 107 in response to the determination (S15), and the playback processing unit 108 Output to. Further, when the content calling unit 106 obtains data with reproduction from the reproduction processing unit 108 and determines that the content is being reproduced, the content calling unit 106 is set and stored by a real time clock or the like according to the determination. The measurement of the call set time is started (S13), it is determined whether the reproduction of the content is completed during the measurement of the call set time (S14), and the content is not reproduced or reproduced from the reproduction processing unit 108 during the measurement of the call set time. When the end data is acquired to determine or recognize the end of playback of the content, the content set corresponding to the identification ID is called from the content DB 107 according to the determination or the like (S15), and the playback processing unit 108 is called. Outputs and initializes the measurement time for the call setup time. On the other hand, if the content calling unit 106 does not determine or recognize the end of the content reproduction without obtaining the content reproduction end or the reproduction end data from the reproduction processing unit 108 during the call setting time measurement, In response to the completion of the measurement of the set time, the process is terminated without executing the call and reproduction of the content set corresponding to the identification ID, and the measurement time is initialized.

  Then, the reproduction processing unit 108 decodes and reproduces the content such as the advertisement video input from the content calling unit 106, and displays the video to be reproduced on the display 109 (S16). The reproduction and display are performed within the real time or almost the call setting time with the recognition of the last keyword of the recognized predetermined key set, and the video of the content to be reproduced and output is a display type device constituting the content providing system 100 The voice input person or the like that is located in the vicinity of the voice and the like and has the voice of the spoken voice taken in is provided. The content providing system 100 recognizes, for example, “overseas travel”, “summer”, and “bali” as keywords within a set time, and provides advertisement content for a Bali trip in summer corresponding to the keywords based on this. Is possible.

  Note that, for example, in response to the input of the identification ID from the recognition management unit 105, the content calling unit 106 calls the content corresponding to the identification ID and outputs the content to the reproduction processing unit 108, and the reproduction processing unit 108 is input. If the playback processing unit 108 or the predetermined unit determines whether the content is being played back and determines that the content is not being played back, the playback processing unit 108 Accordingly, the content that is output and stored is decrypted and played back, and when the playback processing unit 108 or the predetermined unit determines that the content is being played back, playback is performed according to the determination. Start measuring the set time, determine whether the content playback has ended while measuring the playback set time, and recognize the end of content playback while measuring the playback set time. In this case, the reproduction processing unit 108 decrypts and reproduces the content that is output and stored in accordance with the recognition, and the reproduction processing unit 108 or the predetermined unit initializes the measurement time for the reproduction set time. On the other hand, when the playback processing unit 108 or the predetermined unit does not recognize the end of the playback of the content during the measurement of the playback set time, the playback processing unit 108 outputs the output in response to the end of the playback set time measurement. The stored content may be erased without being reproduced, and the reproduction processing unit 108 or the predetermined unit may be configured to initialize the measurement time. In addition, the content calling unit 106 obtains data during playback or playback completion from the playback processing unit 108 that is installed separately by network connection, and calls the content or starts measuring the set time accordingly. The reproduction processing unit 108 may reproduce the content received from the content calling unit 106.

  Next, a second embodiment of the content providing system will be described. As shown in FIG. 4, the content providing system 100 according to the second embodiment includes a microphone 101, a feature extraction unit 102, a keyword recognition unit 103, a standard feature storage unit 104, and recognition management, as in the first embodiment. Unit 105, content calling unit 106, and content database (content DB) 107, but unlike the first embodiment, the content called by the content calling unit 106 is reproduced and output outside the content providing system 100. The transmission processing unit 110 that transmits to the apparatus 200 is provided, and the predetermined unit is realized by a CPU that cooperates with a set program, a memory that stores the program and data, and the like. The functions and operations of the microphone 101, the feature extraction unit 102, the keyword recognition unit 103, the standard feature storage unit 104, the recognition management unit 105, and the content DB 107 are the same as those in the first embodiment.

  In response to the input of the identification ID from the recognition management unit 105, the content calling unit 106 determines whether the content is being transmitted by the transmission processing unit 110, and determines that the content is not being transmitted by the transmission processing unit 110. The content that is set in correspondence with the identification ID is called from the content stored in the content DB 107, the content that is called is output to the transmission processing unit 110, and the content is being transmitted by the transmission processing unit 110. When the determination is made, the measurement of the call setting time set and stored by the real time clock or the like is started, and when the content transmission end is recognized within the call setting time, the content corresponding to the identification ID is called. Output to the transmission processing unit 110, and on the other hand, if the end of content transmission is not recognized within the call setup time , So as not to call the contents corresponding to the identification ID.

  The transmission processing unit 110 is configured to store and hold the content called and output by the content calling unit 106 and sequentially transmit the content data to the reproduction output device 200 in time series. The content data is, for example, the reproduction output device. It is assumed that the image data is partially displayed on the 200 display. The reproduction output device 200 decodes and sequentially outputs content data to be transmitted, and outputs, for example, a digital television receiver that receives, reproduces, and outputs the transmitted image data and audio data through a display and a speaker. For example, a moving image or still image of the content is arranged on a part of the display screen using a picture-in-picture or the like, and the content image is displayed as a sub-image together with the main image. The content may be, for example, an explanation of matters highly relevant to the keywords belonging to the key set, advertisement information, quizzes, guidance information, etc., but other appropriate content that is relevant to the keywords. It is possible.

  The flow of processing by the content providing system 100 of the second embodiment is shown in FIG. When the content providing system 100 of the second embodiment transmits a live broadcast program to the reproduction output device 200 such as a television receiver, for example, the microphone 101 captures the voice of the performer's voice from the program being broadcast. The keyword is recognized from the voice, and based on the recognized keyword, it is used to transmit the explanation of the matters related to the keyword and the advertisement information as the content. As shown in FIG. Until the identification ID of a predetermined key set is output, the same processing as in the first embodiment is executed (S1 to S11).

  Then, in response to the input of the identification ID of the predetermined key set from the recognition management unit 105, the content calling unit 106 determines whether content corresponding to the key set already recognized by the transmission processing unit 110 is being transmitted. (S17) When it is determined that the data not transmitted is acquired from the transmission processing unit 110 and the content is not being transmitted, the content corresponding to the identification ID is called from the content DB 107 according to the determination. (S15) Output to the transmission processing unit 110, the transmission processing unit 110 stores and holds the input content in a predetermined storage area, and sequentially stores the content data stored and held to the reproduction output device 200 in time series Transmit (S19). Further, when the content calling unit 106 acquires data with transmission from the transmission processing unit 110 and determines that the content is being transmitted, the content calling unit 106 is set and stored by a real time clock or the like according to the determination. The measurement of the call setup time is started (S13), it is determined whether the content transmission is completed during the measurement of the call setup time (S18), and the content is not transmitted or transmitted from the transmission processing unit 110 during the measurement of the call setup time. When the end data is acquired and the transmission end of the content is determined or recognized, the content set corresponding to the identification ID is called from the content DB 107 according to the determination or the like (S15), and sent to the transmission processing unit 110. At the same time, the transmission processing unit 110 initializes the measurement time for the call setup time, and stores the input content in a predetermined storage area. Held sequentially transmits to the playback output apparatus 200 the data of the content to be stored and held in time series (S19). On the other hand, if the content calling unit 106 does not determine or recognize the content transmission end without acquiring the content transmission transmission data or the transmission end data from the transmission processing unit 110 during the call setting time measurement, In response to the completion of the measurement of the set time, the process is terminated without executing the call or transmission of the content set corresponding to the identification ID, and the measurement time is initialized.

  The reproduction output device 200 that receives the content data receives, reproduces, and outputs, for example, a program transmitted by a normal broadcast, and the reproduction processing unit decodes and reproduces the content data in response to reception. For example, a content image is displayed on a part of the display. The transmission and display are performed in real time or almost within the call setting time with the recognition of the last keyword of the recognized predetermined key set, and the content image to be reproduced and output is, for example, viewing the program on the reproduction output device 200. Provided to viewers. With this content providing system 100, for example, it is possible to recognize a keyword from the dialogue of performers of a TV program and provide content such as information and explanation related to the keyword.

  For example, in response to the input of the identification ID from the recognition management unit 105, the content calling unit 106 calls the content set corresponding to the identification ID and outputs the content to the transmission processing unit 110, and the transmission processing unit 110 is input. The transmission processing unit 110 or the predetermined unit determines whether the content is being transmitted and determines that the content is not being transmitted. In response to the determination, the content that is output and stored is transmitted to the reproduction output device 200, and when the transmission processing unit 110 or the predetermined unit determines that the content is being transmitted, the determination is made. Start the transmission set time measurement, determine whether the content transmission has been completed during the transmission set time measurement, and When recognizing the end of transmission, the transmission processing unit 110 transmits the content that has been output and stored in response to the recognition, and the transmission processing unit 110 or the predetermined unit determines the measurement time for the transmission set time. On the other hand, when the transmission processing unit 110 or the predetermined unit does not recognize the end of content transmission during the measurement of the transmission set time, the transmission processing unit 110 responds to the end of the measurement of the transmission set time according to the end of the measurement of the transmission set time. The content that is output and stored may be erased without being played back, and the transmission processing unit 110 or the predetermined unit may be configured to initialize the measurement time.

  Next, modified examples of the content providing system 100 according to the first and second embodiments will be described. In the content providing system 100 according to the first and second embodiments, the content to be provided is an image. However, the content to be provided is not limited to this, for example, only a moving image without sound or an image of a still image. In addition, it is possible to use a moving image or still image image and sound, only sound without an image, or the like. When providing audio content or content having audio, for example, the reproduction processing unit 108 decodes and reproduces the audio content, and outputs the content from a speaker installed in the content providing system 100 or the reproduction processing unit of the reproduction output device 200 The audio of the content is provided to the display 109, the viewer of the reproduction output device 200, and the like by decoding and reproducing at 108 and outputting from a speaker installed in the reproduction output device 200.

  In the first and second embodiments, the key set is composed of a plurality of types of keywords, the recognition time for each keyword is recorded, and the predetermined key is set based on the recognition of all the keywords of the predetermined key set within the recognition setting time. Although it is configured to recognize a set, it is appropriate as long as it is a configuration that recognizes a key set corresponding to a plurality of types of keywords based on recognition of a plurality of types of keywords within a recognition setting time, and is set to correspond to a keyword. A key set is composed of a plurality of types of keys, and the recognition time of the keyword is recorded as the recognition time of the key corresponding to the keyword. Based on the recognition of all keys of the predetermined key set within the recognition setting time or the recording of the recognition time A configuration that recognizes a predetermined key set may be used. For example, as shown in FIG. 6, a single keyword or a plurality of keywords (for example, A1, A2, A3) having similar meanings and contents are set corresponding to a key representing a similar set unit or a representative keyword (for example, A), A key set specified by an identification ID is configured with keys (for example, A, B, C), and the recognition management unit 105 corresponds to the recognition of the keyword (for example, A1), and the key corresponding to the keyword's recognition time (For example, A) is recorded as the recognition time, and the recognition time of the key (for example, A) is updated and recorded according to the recognition of the keyword (for example, A1, A2, or A3) corresponding to the same key (for example, A). , The time from the current time to the recognition time is erased the recognition time for the recorded key that has passed the recognition setting time, and moreover, a plurality of types constituting the key set For all the keys (for example, A, B, C), when the recognition time is recorded within the recognition setting time, for example, a configuration that recognizes a predetermined key set such as a key set specified by the identification ID: 0001, etc. .

  Also, when recognizing a predetermined key set, all the keywords or all keys corresponding to the predetermined key set are recognized, and the elapsed time from the earliest recognition time for the keyword or key in the recognition to the last recognition time The configuration for recognizing that the time is within the recognition setting time is not limited to the configuration for deleting the recognition time when the elapsed time from the current time to the recognition time has passed the recognition setting time. The recognition time for the keyword or key is recorded and updated in the same manner as above without deleting the recognition time when the recognition setting time has passed, and the recognition time of all keywords or all keys corresponding to the predetermined key set is recorded. In accordance with the record, the earliest recognition time and the elapsed time until the last recognition time among the recognition times of all the keywords or all the keys are acquired, The elapsed time in comparison with cognitive set time, the elapsed time may be such configured that recognize the predetermined set of keys when it is within cognitive set time.

  In the first and second embodiments, the time recorded corresponding to the keyword recognition is the recognition time acquired by a real-time clock or the like according to the keyword recognition. It is appropriate as long as the time has a correspondence relationship. For example, it may be a time when a keyword recognition fact or a key recognition fact is recorded.

  Further, for example, the content calling unit 106 is reproducing or transmitting content corresponding to a key set that has already been recognized by the reproduction processing unit 108 or the transmission processing unit 110 based on the input of the identification ID from the recognition management unit 105. If it is determined that the content is not being reproduced or transmitted, the content DB 107 calls the content corresponding to the identification ID and outputs the content to the reproduction processing unit 108 or the transmission processing unit 110. When it is determined that the content is being reproduced or transmitted, the identification ID input from the recognition management unit 105 is recorded and held in a predetermined storage area, or is already recorded and held in a predetermined storage area. When there is an identification ID, the identification ID newly recorded from the recognition manager 105 is recorded and held. D is rewritten, the identification ID of the predetermined storage area is updated and recorded and held, and further, the content reproduction end data is obtained from the reproduction processing unit 108 or the content transmission end data is obtained from the transmission processing unit 110. When the playback end or transmission end is determined or recognized, the content set corresponding to the identification ID recorded and held in the predetermined storage area at the time of the determination or the like is called and played back according to the determination or the like A configuration for outputting to the processing unit 108 or the transmission processing unit 110 may be adopted.

  In addition, for example, the same content may be set corresponding to a plurality of different key sets or identification IDs thereof, and the content calling unit 106 stores and holds based on the input of any identification ID corresponding to the content. It is good also as a structure which calls the said content from existing content DB107.

  In addition, when the content is a moving image or sound, for example, the reproduction processing unit 108 or the transmission processing unit 110 reproduces and outputs or transmits the content until the completion of the cycle. A configuration may be adopted in which the reproduction or transmission of the content being transmitted is interrupted. For example, the content calling unit 106 calls the content set corresponding to the identification ID from the content DB 107 based on the input of the identification ID based on the recognition of the key set, and outputs the content to the reproduction processing unit 108 or the transmission processing unit 110 for reproduction processing. The unit 108 or the transmission processing unit 110 starts reproduction or transmission of the input content when there is no content being reproduced or transmitted based on the input of the content or the like, and on the other hand, during reproduction or transmission If there is such content, the reproduction or transmission is stopped and terminated, and the reproduction or transmission of the input content is started. Further, when the content is a still image, for example, the content is reproduced or transmitted over a uniform set time set and stored in the reproduction processing unit 108 or the transmission processing unit 110, or the content corresponds to the content. A configuration in which the playback processing unit 108 or the transmission processing unit 110 recognizes a set time that is set and stored in the DB 107 or the like and plays or transmits content over the set time may be employed.

  Also, a plurality of contents are set corresponding to the identification ID of one key set or one key set, and in addition to the key set or the identification ID of the key set, each content is associated with another condition item or designated ID. Is stored in the content DB 107, and the content calling unit 106 determines whether the predetermined key set or the predetermined key set is based on the recognition or input of the predetermined key set or the identification ID of the predetermined key set and other condition items. It is good also as a structure which calls the content set corresponding to identification ID and another condition item. For example, each content is set and stored in the content DB 107 corresponding to the identification ID of the key set and the time zone classification, and the content calling unit 106 uses a real time clock or the like according to the input of the identification ID of the predetermined key set. The current time at which the identification ID of the predetermined key set is input from the measurement time is acquired, and the predetermined time zone in which the current time is included in comparison with the time zone section stored and held in the predetermined storage area Recognizing the category, the content corresponding to the recognized predetermined time zone classification including the identification ID of the input key set and the current time is called from the content DB 107, etc. Different content may be provided according to the time at which is recognized.

  Further, each content is set and stored in the content DB 107 in correspondence with the key set identification ID and the environmental classification such as the location classification or the weather classification. For example, the content calling unit 106 inputs the identification ID of a predetermined key set. In response, a predetermined location section or a predetermined weather section stored and held in a predetermined storage area is recognized, and an identification ID of the input predetermined key set and content corresponding to the recognized predetermined location section or the predetermined weather section are It may be configured to call from the content DB 107, and the content corresponding to the place and the weather may be provided. The location classification can be appropriately set according to, for example, the type of store where the content reproduction output means is installed, or by region, and the weather classification is appropriately set according to temperature, humidity, weather, or the like. For example, a predetermined part of the content providing system cooperates with the control program, recognizes weather data such as temperature, humidity, weather or a combination thereof input on the day, and sets and stores the weather category and weather. It is possible to adopt a configuration in which a predetermined weather classification including the weather data is recognized by comparing the data, and the weather classification is stored and held in a predetermined storage area.

  Further, for example, each content is set and stored in the content DB 107 in correspondence with the identification ID and designated ID of the key set, and the content calling unit 106 outputs or reproduces a plurality of contents corresponding to the identification ID of the predetermined key set. The specified ID of the content set in the next order of the acquired content is acquired by adding a predetermined number such as 1 to the specified number that is the specified ID of the output or reproduced content, and the specified ID is obtained from the predetermined key set. Recognizing a specified ID stored and held with respect to the identification ID of the predetermined key set based on the input of the identification ID of the predetermined key set based on the input of the identification ID of the predetermined key set; It is configured such that the content corresponding to the identification ID of the predetermined key set and the recognized designated ID is called, and the same key set It may be provided a different content to the second or the like to.

  Further, for example, a plurality of contents that are continuously reproduced or continuously transmitted are associated with one key set or an identification ID of one key set as necessary, and set and stored in the contents DB 107, and the contents calling unit 106 stores a predetermined key set. Based on the input or the input of an identification ID of a predetermined key set, a plurality of contents set corresponding to the predetermined key set or the identification ID of the predetermined key set are called, and the reproduction processing unit 108 or the transmission processing unit 110 A configuration in which a plurality of contents are continuously reproduced or continuously transmitted may be provided, and one or a plurality of contents may be provided as necessary.

  The present invention, for example, captures voice from a conversation of a speaker or the like in a store, specifies a topic based on a keyword included in the voice, and provides content such as an advertisement corresponding to the topic to the speaker or the like in real time. Can do.

The block diagram which shows the whole structure of the content provision system of 1st Embodiment. The figure which shows the data structure of a recognition management table. The flowchart which shows the flow of the content provision process by the content provision system of 1st Embodiment. The block diagram which shows the whole structure of the content provision system of 2nd Embodiment. The flowchart which shows the flow of the content provision process by the content provision system of 2nd Embodiment. The figure which shows the modification of the data structure of a recognition management table.

Explanation of symbols

DESCRIPTION OF SYMBOLS 100 Content provision system 101 Microphone 102 Feature extraction part 103 Keyword recognition part 104 Standard feature memory | storage part 105 Recognition management part 106 Content calling part 107 Content DB
108 reproduction processing unit 109 display 110 transmission processing unit 200 reproduction output device

Claims (3)

A computer or network computer, which recognizes the voice taken from the microphone, recognizes the keyword set from the recognized voice, and corresponds the recognition corresponding time of the keyword to the keyword according to the keyword recognition Or a means for updating the recording of the recognition corresponding time according to the recognition of the corresponding keyword, and a plurality of types of keywords or keys constituting the key set. From means for recognizing a predetermined key set whose recognition elapsed time from the earliest recognition correspondence time to the last recognition correspondence time is within the recognition setting time, and a content database storing contents set corresponding to each key set Corresponding setting with the recognized predetermined key set Dividing the content based on the recognition of the normal content and the keyword, the means for calling the content, the normal content such as the transmitted program and the like, and the means for playing the called content in near real time and the real time A content providing system comprising means for displaying on a screen. It is composed of a computer or a network computer, and recognizes a voice taken from a microphone, recognizes a keyword set from the recognized voice, and a key set including a plurality of types of keywords or keys corresponding to the keywords. Means for recognizing a predetermined key set based on keyword recognition, means for determining whether content is being reproduced or transmitted based on recognition of the predetermined key set, recognition of the predetermined key set, or reproduction of the content Alternatively, based on the recognition of the end of playback or the end of transmission that is within the set time from the determination during transmission to the end of content playback or transmission end, playback or playback of the content set corresponding to the predetermined key set When sending to the output device and at the time of the setting And a means for not reproducing or transmitting the content set corresponding to the predetermined key set when the end of reproduction or transmission end is not recognized. system. It is composed of a computer or a network computer, and recognizes a voice taken from a microphone, recognizes a keyword set from the recognized voice, and a key set including a plurality of types of keywords or keys corresponding to the keywords. Based on recognition of a keyword, a means for recognizing a predetermined key set, a means for determining whether the content is being reproduced or transmitted based on the recognition of the predetermined key set, and based on a determination of whether the content is being reproduced or transmitted, Means for recording and holding the recognition of the predetermined key set and, if there is recognition of the key set already recorded and held, renewing and holding the recognition of the predetermined key set; Based on the recognition of the end of transmission, the predetermined key set that is stored Contents providing system characterized in that it comprises means for transmitting to the corresponding have been set reproduction or playback output apparatus content.

Images