JP2012118623A - Advertisement distribution system, advertisement distribution method and advertisement distribution program - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide an advertisement distribution system capable of properly collecting and utilizing feedback from viewers belonging to a target class of an advertisement distributed by digital signage.SOLUTION: An advertisement distribution system comprises: customer information storage means for storing customer information; advertisement display means for displaying an advertisement; voice input means for inputting conversations of viewers viewing the advertisement; voice recognition means for performing voice recognition for the input conversations; collection means for collecting voice recognition results; and update means for updating the customer information stored in the customer information storage means based on the collected voice recognition results.

Description

  The present invention relates to an advertisement distribution system, an advertisement distribution method, and an advertisement distribution program.

  Advertisement distribution services using voice recognition technology are known. For example, in the following Patent Document 1 (Japanese Patent Laid-Open No. 2002-183579), in a voice operation type advertisement distribution method, a configuration in which character information included in displayed advertisement information is converted into voice information and output as voice, A configuration is described in which an advertisement is displayed in response to an instruction input by voice.

  Japanese Patent Laid-Open No. 2004-271615 discloses text data suitable for characteristics of a speech recognition target that analyzes language information of text data and based on the analyzed language information. An information processing apparatus that generates a statistical language model based on converted text data is disclosed.

JP 2002-183579 A JP 2004-271615 A

  By the way, in recent years, advertisement distribution services using digital signage (electronic signage) that displays video and information on a display in a timely manner using network technology have appeared. Such digital signage enables the target layer setting in consideration of the installation location in advance or the display content to be switched easily, thereby enabling advertisement distribution focused on the target layer. It is expected that it can increase the cost effectiveness of advertising.

  Here, in order to increase the cost-effectiveness of the advertisement, it is conceivable to collect and use feedback from the viewer for the distributed advertisement. However, there are many cases where viewers who actually watch the advertisement include viewers who are not the target audience. However, if such non-target audience responses are captured, There is a problem that accuracy is lowered.

  Therefore, a new mechanism that can more appropriately collect and use feedback from viewers belonging to the target layer of the distributed advertisement is desired. However, neither of the above-mentioned Patent Document 1 and Patent Document 2 considers such a problem in advertisement distribution using digital signage.

  Accordingly, an object of the present invention is to provide an advertisement distribution system that can more appropriately collect and use feedback from viewers belonging to the target layer of distributed advertisements in advertisement distribution using digital signage. To do.

  The advertisement distribution system of the present invention includes customer information storage means for storing customer information, advertisement display means for displaying advertisements, voice input means for inputting conversations of viewers who view the advertisements, and the input conversations. Voice recognition means for performing voice recognition, collection means for collecting the voice recognition results that have been voice-recognized, and updating customer information stored in the customer information storage means based on the collected voice recognition results Updating means.

  The present invention is an advertisement distribution system configured such that a digital signage device and an advertisement management device can communicate with each other, wherein the digital signage device displays an advertisement transmitted from the advertisement management device, and Voice input means for inputting a conversation of a viewer who views an advertisement, voice recognition means for performing voice recognition on the input conversation, and transmission for transmitting the voice recognition result recognized by the voice to the advertisement management apparatus Means for transmitting the advertisement to the digital signage device, customer information storage means for storing customer information, and receiving means for receiving the transmitted voice recognition result. And updating means for updating customer information stored in the customer information storage means based on the received voice recognition result.

  The present invention is an advertisement distribution method in an advertisement distribution system, comprising: an advertisement display step for displaying an advertisement; a voice input step for inputting a conversation of a viewer who views the advertisement; and a voice for the input conversation. A speech recognition step for performing recognition, a collection step for collecting the speech recognition result obtained by speech recognition, and an update step for updating customer information stored in a customer information database based on the collected speech recognition result. Prepare.

  The present invention relates to an advertisement distribution method in an advertisement distribution system configured such that a digital signage device and an advertisement management device can communicate with each other, wherein the advertisement management device transmits the advertisement to the digital signage device. In the digital signage device, an advertisement display step for displaying an advertisement transmitted from the advertisement management device, a voice input step for inputting a conversation of a viewer who views the advertisement, and a voice for the input conversation A speech recognition step for performing recognition, a transmitting step for transmitting the speech recognition result recognized by the speech to the advertisement management device, a receiving step for receiving the transmitted speech recognition result in the advertisement management device, and the reception Update the customer information stored in the customer information database based on the voice recognition result Includes an update step, the.

  The present invention is also realized as a program for causing a computer to execute each step described in the advertisement distribution method.

  According to the present invention, it is possible to provide an advertisement distribution system that can more appropriately collect and use feedback from viewers belonging to the target layer of the distributed advertisement in advertisement distribution using digital signage. Become.

It is a block diagram which shows schematic structure of the advertisement delivery system which concerns on 1st Embodiment. It is a figure which shows an example of a function structure of an advertisement delivery system. It is a figure showing an example of a data structure of advertisement corresponding information DB. It is a figure showing an example of a data structure of advertisement information DB. It is a figure showing an example of a data structure of customer information DB. It is a figure showing an example of the flow of an advertisement delivery process. It is a figure showing an example of the selection process of a language model and an acoustic model. It is a figure which shows an example of a function structure of the advertisement delivery system which concerns on 2nd Embodiment.

  Hereinafter, embodiments of the present invention will be described in detail with reference to the drawings. In addition, the same code | symbol is attached | subjected to the same element and the overlapping description is abbreviate | omitted.

[First Embodiment]
With reference to FIG. 1, the schematic configuration of the advertisement distribution system in the first embodiment will be described. The advertisement distribution system 1 includes a digital signage device 11 (11-1, 11-2... 11-n), an advertisement data input device 41, an advertisement management device 51, and a voice recognition management device 61. ing. The digital signage device 11 is installed at a predetermined installation location such as a station or a shopping center, and has directional microphones 21 (21-1, 21-2,... 21-n). The advertisement management device 51 has an advertisement information database (advertisement information DB) 52 and a customer information database (customer information DB) 53, and the speech recognition management device 61 has a language model / acoustic model database (model DB) 62. ing. The digital signage device 11 is also called an advertisement display device having a voice recognition function.

  Here, an outline of the operation of the advertisement distribution system 1 will be described. First, the advertisement data input device 41 transmits the advertisement data to the advertisement management device 51 based on the advertiser's operation input. The advertisement data includes advertisement main body data and advertisement-related data related to the advertisement. The advertisement-related data includes content specifying data for specifying the content (type) of the advertisement and target layer data for specifying the target layer.

  The advertisement management apparatus 51 registers the received advertisement data in the advertisement information DB 52 and transmits advertisement-related data included in the advertisement data to the voice recognition management apparatus 61.

  The speech recognition management device 61 receives advertisement-related data and selects a language model / acoustic model (hereinafter referred to as “language / acoustic model”) corresponding thereto from the model DB. For example, if “for household appliances” is set for the advertisement content and “20s female / standard language” is set for the target audience, the language model suitable for “for household appliances” and “20th female / standard language” A suitable acoustic model is selected. The advertisement distribution system 1 executes the above process for each advertisement.

  The speech recognition management device 61 transmits the language / acoustic model selected for each advertisement to the corresponding digital signage device 11. Further, the advertisement management device 51 transmits the advertisement main body data to the corresponding digital signage device 11.

  The digital signage device 11 receives the advertisement main body data and displays the advertisement on the display based on the advertisement body data. The directional microphone 21 inputs a conversation (audio information) of a viewer who views the advertisement displayed on the display.

  The digital signage device 11 executes speech recognition processing on the speech of the viewer inputted by voice using the language / acoustic model corresponding to the displayed advertisement transmitted from the speech recognition management device 61, and the speech recognition result Is transmitted to the advertisement management apparatus 51. The advertisement management apparatus 51 updates the customer information DB 53 based on the received voice recognition result.

  As described above, according to the advertisement distribution system shown in FIG. 1, since the speech recognition process is executed using the language / acoustic model suitable for the advertisement content and the target layer, the conversation regarding the advertisement content by the target layer viewers. Are easily recognized (the recognition rate is high), while conversations of non-target audiences and conversations unrelated to the advertisement content are difficult to be recognized (the recognition rate is low). Therefore, by using only the speech recognition result whose recognition rate is equal to or higher than a predetermined threshold, it is possible to take out the reaction of the target audience from among the viewers who viewed the advertisement and appropriately reflect it in the customer information DB. As a result, the accuracy of advertisement feedback can be further improved.

  Next, the functional configuration of the advertisement distribution system 1 will be further described with reference to FIG. The advertisement distribution system 1 includes a digital signage device 100, a voice recognition management device 200, and an advertisement management device 300. The advertisement distribution system 1 is configured to be able to communicate with the advertisement provider terminal device 400 via a network.

  The digital signage apparatus 100 includes an advertisement display unit 101, a voice input unit 102, and a voice recognition unit 103. The digital signage device 100 includes a CPU that controls processing and operations of the digital signage device 100, a memory such as a ROM and a RAM, an external storage device that stores various information, an input / output device, a communication device, and a bus that connects these. For example, the CPU implements the various function realizing means by executing a predetermined program stored in a memory or the like.

  The advertisement display unit 101 has a function of displaying an advertisement based on advertisement data transmitted from the advertisement management apparatus 300, and corresponds to, for example, a display.

  The voice input unit 102 has a function of inputting a viewer's voice (utterance), and corresponds to, for example, a directional microphone.

  The voice recognition unit 103 has a function of performing voice recognition processing on the viewer's conversation input from the voice input unit 102. Specifically, a dictionary (language model / acoustic model) corresponding to the advertisement transmitted from the speech recognition management apparatus 200 prior to the advertisement display is received and stored in a predetermined storage area. When a voice is input from the voice input unit 102 during advertisement display (which may include a predetermined time before and after display), voice recognition processing is executed on the voice using the above dictionary. Then, the voice recognition result is transmitted to the advertisement management apparatus 300.

  The speech recognition management apparatus 200 includes a model data input unit 201, a language model DB 202, an acoustic model DB 203, an advertisement correspondence information DB 204, a model selection unit 205, and a dictionary transmission unit 206. The voice recognition management device 200 includes a CPU that controls processing and operations of the voice recognition management device 200, a memory such as a ROM and a RAM, an external storage device that stores various types of information, an input / output device, a communication device, and a bus connecting them. A general-purpose server computer provided can be applied. For example, the CPU implements various functions by executing predetermined programs stored in a memory or the like.

  The model data input unit 201 inputs a language model and an acoustic model used for speech recognition processing. The model may be input directly by an operator via a predetermined input device, or may be input from another device via a communication line.

  The language model DB 202 is a database (storage means) that stores language models used for speech recognition processing. The language model is a model related to linguistic features, and represents, for example, restrictions on how phonemes are arranged, and can be expressed by an N-gram language model, a context free grammar, or the like. Here, one or more language models that can correspond to each advertisement content are stored.

  The acoustic model DB 203 is a database (storage means) that stores an acoustic model used for speech recognition processing. The acoustic model is a model related to acoustic features, and represents, for example, what frequency characteristics each phoneme to be recognized has. Here, one or more acoustic models that can correspond to target layers that can be recognized are stored.

  The advertisement correspondence information DB 204 is a database (storage means) that manages advertisement contents and language models, and target layers and acoustic models in association with each advertisement. FIG. 3 shows an example of the data configuration of the advertisement correspondence information DB 204. The language model correspondence table shown in FIG. 6A includes, as an example of a data structure, an advertisement content specifying information item (eg, “for home appliances”, “for beverages”, etc.) for specifying the content of the target advertisement, and a language model. And a stored location information item. In addition, the acoustic model correspondence table shown in FIG. 5B is a target layer specifying information item (for example, “20s female / standard language”, “20s female” that identifies the target layer of the target advertisement as an example of the data structure.・ Kansai dialect etc.) and location information items in which acoustic models are stored. The data structure is not limited to that shown in the figure, and arbitrary data items can be provided depending on the design.

  The model selection unit 205 refers to the advertisement correspondence information DB 204 to select a language model that matches the content of the target advertisement from the language model DB 202 and selects an acoustic model that matches the target layer from the acoustic model DB 203. The model selection unit 205 dynamically selects a language model / acoustic model corresponding to each advertisement, and if there is a change in the advertisement content or target layer, a new language model / acoustic model that matches the changed content is newly created. Select

  The dictionary transmission unit 206 transmits the language model DB 202 and the acoustic model 203 selected by the model selection unit 205 to the digital signage apparatus 100 as a dictionary used for speech recognition of the target advertisement. Specifically, the model data body or information (URL or the like) specifying the model data is transmitted.

  The advertisement management apparatus 300 includes an advertisement data registration unit 301, an advertisement information DB 302, an advertisement distribution unit 303, a customer information collection unit 304, and a customer information DB 305. The advertisement management apparatus 300 includes a CPU that controls processing and operations of the advertisement management apparatus 300, a memory such as a ROM and a RAM, an external storage device that stores various types of information, an input / output device, a communication device, and a bus that connects these. For example, the CPU implements various functions by executing a predetermined program stored in a memory or the like.

  The advertisement data registration unit 301 receives advertisement data provided from the advertisement provider and registers it in the advertisement information DB 302. The advertisement data includes, for example, information distributed to viewers as an advertisement (hereinafter referred to as “advertising body data”) and information related to the advertisement (hereinafter referred to as “advertisement related data”). It is. The advertisement main body data is, for example, moving image / still image data and video / audio data. The advertisement-related data includes information that identifies the advertisement content and information about the target of the advertisement. For example, an advertisement ID that uniquely identifies the advertisement, information about the target segment (target age, target gender), advertisement), target field , Keywords, language used, delivery region, delivery time zone, etc.

  The advertisement information DB 302 is a database (storage means) that stores advertisement data. The advertisement information DB 302 illustrated in FIG. 4 includes, as an example of a data structure, an advertisement ID that uniquely identifies an advertisement, information on a target layer (target age, target sex), an advertisement provider (advertiser), a target area of an advertisement, Each item includes a keyword, a language used, a delivery region, a delivery time zone, and location information in which advertisement main body data is stored.

  The advertisement distribution unit 303 distributes the advertisement main body data included in the advertisement information stored in the advertisement information DB 302 to the digital signage device 100. The advertisement main body data is distributed, for example, according to the distribution region and distribution time specified in the advertisement-related data.

  The customer information collection unit 304 collects the voice recognition result of the voice recognition unit 103 as customer information (viewer feedback information). The collected customer information is stored in the customer information DB 305.

  The customer information DB 305 is a database (storage unit) that stores customer information (customer feedback) collected by the customer information collection unit 304. The customer information DB 305 shown in FIG. 5 includes, as an example of the data structure, an advertisement ID, a predicted age of collected viewers, a predictive sex, a language used for conversation, a collection time, a conversation content (conversation 1, conversation 2,... Items such as conversation N) are included.

  The advertisement provider terminal device 400 includes advertisement data input / transmission means 401. The advertisement data input transmission unit 401 receives input of advertisement data from the advertisement provider, and transmits the input advertisement data to the advertisement management apparatus 300 for registering in the advertisement information DB 302. The advertisement provider terminal device 400 includes a CPU that controls processing and operations of the advertisement provider terminal device 400, a memory such as a ROM and a RAM, an external storage device that stores various information, an input / output device, a communication device, and the like. A general-purpose computer having a bus connecting the two can be applied. For example, the CPU implements the function realizing unit by executing a predetermined program stored in a memory or the like.

  Next, the flow of processing in the advertisement distribution system 1 will be described with reference to the flowcharts of FIGS. In addition, each process step shown in the flowchart to be described later can be executed in any order or in parallel within a range in which there is no contradiction in processing contents. Moreover, you may add another step between each process step. Further, a step described as one step for convenience can be executed by being divided into a plurality of steps, while a step described as being divided into a plurality of steps for convenience can be grasped as one step.

  The advertisement provider's terminal device 400 accepts input of advertisement data from the advertisement provider (step S100). Then, the input advertisement data is transmitted to the advertisement management apparatus 300 (step S101). The advertisement data includes advertisement main body data and advertisement-related data.

  The advertisement management apparatus 300 receives the advertisement data, and the advertisement data registration unit 301 registers it in the advertisement information DB 302. Hereinafter, as an example of advertisement distribution, a case where the first advertisement and the second advertisement are distributed to the digital signage apparatus 100 and customer information (feedback) is collected will be described.

  The advertisement management apparatus 300 transmits the advertisement-related data of the first advertisement to the voice recognition management apparatus 200 (Step S102).

  The model selection means 205 of the speech recognition management apparatus 200 selects a language / acoustic model suitable for the first advertisement by referring to the advertisement correspondence information DB 204 based on the received advertisement related data (S103). As shown in FIG. 3, for example, when the first advertisement is an advertisement ID “0001”, the target field is “home appliance”, the target layer is “20-35 years old, female (F), Japanese, Tokyo "Region" is set. Therefore, as shown in FIG. 4, a language model corresponding to “for home appliances” and an acoustic model corresponding to “20s female / standard language” are selected as the language model of the first advertisement.

  The dictionary transmission unit 206 reads the selected language model from the language model DB 202, reads the selected acoustic model from the acoustic model DB 203, and reads the read language model and acoustic model (hereinafter referred to as “dictionary”) of the first advertisement. In association with the advertisement ID, it is transmitted to the digital signage apparatus 100 (step S104).

  When the dictionary transmission is completed, the dictionary transmission unit 206 transmits information indicating that the dictionary transmission is completed to the advertisement management apparatus 300 (step S105). In addition, the transmission timing of a language model and an acoustic model should just be before the delivery time of a 1st advertisement. In addition, the speech recognition management apparatus 200 transmits position information of the language model and the acoustic model in advance, and the digital signage apparatus 100 recognizes the language model and the acoustic model (dictionary) before the first advertisement delivery time. You may acquire from the management apparatus 200. FIG.

  Here, the language model selection processing will be described in detail with reference to FIG.

  The advertisement management apparatus 300 reads the advertisement data of the target advertisement (first advertisement) from the advertisement information DB (Step S201), and based on the advertisement-related data, distributes the language model (Step S202) and distributes the acoustic model. An instruction (step S203) is transmitted to the speech recognition management apparatus 200, respectively.

  First, the distribution of the acoustic model will be described. In step S204, the model selection unit 205 of the speech recognition management apparatus 200 makes an inquiry to the acoustic model DB 203 based on the target layer information, and searches for the corresponding acoustic model (step S205). If the corresponding acoustic model is in the acoustic model DB 203 (step S206: YES), the dictionary transmission unit 206 reads the position information of the acoustic model (step S207) and prepares for distribution of the acoustic model (step S208). The acoustic model is distributed by transmitting a message including the position information (URI) of the acoustic model (step S209). If the corresponding acoustic model is not in the acoustic model DB 203 (step S206: NO), it is determined that it is difficult to perform speech recognition of the target target layer (step 210), and the target advertisement is subject to speech recognition. Set as outside (step S211).

  Next, language model distribution will be described. In step S212, the model selection unit 205 of the speech recognition management apparatus 200 makes an inquiry to the language model DB 202 based on the advertisement content (keywords and the like) and searches for the corresponding language model (step S213). If the corresponding language model is in the language model DB 202 (step S214: YES), the dictionary transmission unit 206 reads the position information of the language model (step S215), and prepares to distribute the language model (step S216). The language model is distributed by transmitting a message including the location information (URI) of the language model (step S217). If the corresponding language model is not in the language model DB 202 (step S214: NO), it is determined that it is difficult to perform speech recognition of the target target layer (step 218), and the target advertisement is subject to speech recognition. Set as outside (Sstep 219).

  The speech recognition management apparatus 200 transmits the completion of selection and distribution of the dictionary (model) to the advertisement management apparatus 300 (step 218), thereby completing the preparation for advertisement distribution (step S219).

  Returning to FIG. 6, the description will be continued. When receiving the dictionary transmission completion notification, the advertisement management apparatus 300 notifies the digital signage apparatus 100 of the start of advertisement distribution (step S106).

  The voice recognition means 103 of the digital signage apparatus 100 starts voice recognition from this (step S107).

  The advertisement distribution unit 303 of the advertisement management apparatus 300 transmits the advertisement main body data of the first advertisement to the digital signage apparatus 100 in accordance with the distribution time (step S108).

  The advertisement display means 101 of the digital signage apparatus 100 displays the first advertisement on the display based on the received advertisement main body data (step S109).

  For example, when a woman A and a woman B who have passed around the digital signage device 100 have a conversation while viewing the first advertisement, the conversation is input by the voice input unit 102 (step S110).

  The voice recognition unit 103 executes a voice recognition process for the input conversation using the dictionary corresponding to the first advertisement (step S111). Note that speech recognition processing using a dictionary (language model / acoustic model) can be executed using a known technique.

  The voice recognition means 103 ends the voice recognition when the display of the first advertisement is finished (step S112), and transmits the voice recognition result to the advertisement management apparatus 300 (step S113). Note that only speech recognition results having a recognition rate equal to or higher than a predetermined threshold value can be adopted and transmitted.

  Upon receiving the voice recognition result for the first advertisement, the customer information collection unit 304 of the advertisement management apparatus 300 updates the customer information DB 305 based on this (step S115). Also, the voice recognition result is transmitted to the advertisement provider terminal device 400 (step S114). For example, for the first advertisement (advertisement ID = 0001), “home appliance” is set as the target field, and “20-35 years old, female (F), Japanese, Tokyo region” is set as the target layer (FIG. However, it can be seen that the customer information DB 305 shown in FIG. 5 contains the contents of conversations by a plurality of viewers estimated to be age “20-30” or “30-40”.

  Processing is performed for the next second advertisement. Specifically, the advertisement management apparatus 300 transmits the advertisement-related data of the second advertisement to the voice recognition management apparatus 200 (step S116).

  The model selection means 205 of the speech recognition management device 200 selects a language / acoustic model suitable for the second advertisement by referring to the advertisement correspondence information DB 204 based on the received advertisement-related data of the second advertisement ( S117). As shown in FIG. 3, for example, when the second advertisement has an advertisement ID “0002”, the target area of the advertisement is “beverage”, the target group is “10-25 years old, male (M), Japanese” , Kyoto region "is set. Therefore, as shown in FIG. 4, a language model corresponding to “for beverage” and an acoustic model corresponding to “male in the 20s / Kansai dialect” are selected.

  The dictionary transmission unit 206 reads the selected language model from the language model DB 202, reads the selected acoustic model from the acoustic model DB 203, and associates the read language model and acoustic model (dictionary) with the advertisement ID of the second advertisement. To the digital signage apparatus 100 (step S118).

  When the dictionary transmission is completed, the dictionary transmission unit 206 transmits information indicating that the dictionary transmission is completed to the advertisement management apparatus 300 (step S119). The selection and transmission processing of the language model and acoustic model of the second advertisement need only be before the advertisement distribution time, and for example, the first advertisement distribution processing may be executed in parallel.

  The advertisement distribution means 303 of the advertisement management apparatus 300 transmits the advertisement main body data to the digital signage apparatus 100 when the second advertisement distribution time comes (step S120).

  The advertisement display means 101 of the digital signage apparatus 100 displays the advertisement on the display based on the received advertisement main body data (step S121).

  The voice input unit 102 starts voice input. For example, when the male A and the male B who have passed around the digital signage apparatus 100 have a conversation while viewing an advertisement, the voice input unit 102 inputs the conversation (step S123).

  Then, the voice recognition unit 103 performs a voice recognition process on the input conversation using the dictionary corresponding to the second advertisement (step S124). Speech recognition processing using a dictionary (language model / acoustic model) can be executed using a known technique.

  For example, when the display of the second advertisement ends, the voice recognition unit 103 ends the voice recognition (step S125), and transmits the voice recognition result to the advertisement management apparatus 300 (step S126). Note that only speech recognition results having a recognition rate equal to or higher than a predetermined threshold value can be adopted and transmitted.

  When the customer information collection unit 304 of the advertisement management apparatus 300 receives the voice recognition result of the second advertisement, the customer information collection unit 304 updates the customer information DB 305 based on the result (step S127). Further, the voice recognition result of the second advertisement is transmitted to the advertisement provider terminal device 400 (step S128).

  At a predetermined timing, the advertisement management apparatus 300 ends the advertisement distribution (step S129).

  As described above, according to the first embodiment, a language model / acoustic model suitable for an advertisement content and a target layer is selected in advance, and voice recognition is executed for each advertisement using the selected language model / acoustic model. Therefore, it is possible to appropriately automatically collect responses (feedback) of viewers belonging to the target audience among viewers who have viewed the advertisement.

  Further, in the first embodiment, since the collection result is selected according to the recognition rate, the reaction by viewers in the non-target layer that tends to decrease the recognition rate is easily discarded. As a result, more accurate advertising feedback can be provided to advertisers in real time in a configuration in which an advertisement with a predetermined target layer that is segmented according to the installation location is distributed to digital signage. .

  In the first embodiment, the impression of the viewer who actually viewed the advertisement is collected in real time and automatically updated to the customer information DB as the viewer feedback. Therefore, for example, information via a web page or an email is used. As compared with the case of collecting the feedback information, the collection work and the update work of the feedback information can be simplified. In addition, by using viewer feedback as a conversation between viewers, it becomes possible to obtain an undecided opinion on advertisements and products.

  In addition, according to the first embodiment, the digital signage apparatus 100 may perform speech recognition using a language model / acoustic model corresponding to each advertisement, and needs to refer to a large number of language models / acoustic models. Therefore, for example, even when the vocabulary increases as the target layer is refined, speech recognition can be performed without reducing the recognition speed and the recognition rate.

[Second Embodiment]
The second embodiment is characterized in that a predetermined event process is executed when a voice recognition result by the voice recognition means matches a predetermined event condition set in advance. FIG. 8 is a diagram illustrating a schematic configuration of the advertisement distribution system 1 according to the second embodiment. The advertisement distribution system 1 according to the second embodiment is different from the advertisement distribution system 1 according to the first embodiment in that it includes a recognition event determination unit 306. Therefore, only a difference from the advertisement distribution system 1 according to the first embodiment will be described here.

  The recognition event determination unit 306 executes a predetermined event process when the voice recognition result collected by the customer information collection unit 304 matches a predetermined event condition set in advance. The predetermined event process can be set as appropriate according to the specification or design. For example, storing the voice recognition result in the customer information DB 305 is applicable. That is, the recognition event determination unit 306 determines whether or not the voice recognition result matches a predetermined event condition, and stores the voice recognition result in the customer information DB 305 when the determination result is correct. On the other hand, if the determination result is NO, the voice recognition result is not stored in the customer information DB 305.

  The content of the predetermined event condition can be set as appropriate according to the specification or design. For example, the condition that the viewer speaks a keyword related to the advertising product can be set. In the advertisement information DB 302 of FIG. 3, as an example, “personal computer, Internet...” Is set as a keyword for an advertisement whose target field is “home appliance”.

  When the speech recognition result collected by the customer information collection unit 304 includes an utterance “personal computer” or “Internet”, the recognition event determination unit 306 sends a conversation including the utterance to the customer information DB 305. Store. On the other hand, if the speech “personal computer” or “Internet” is not included, the voice recognition result is not stored in the customer information DB 305.

  The predetermined event conditions are not limited to those described above. For example, a plurality of keywords are spoken, a conversation time is a predetermined time or more, and a plurality of viewers are having a conversation. Something may be set.

  Further, the predetermined event processing is not limited to the above, and for example, the content of the advertisement being distributed may be changed, or a predetermined message may be superimposed on the advertisement being distributed.

  As described above, according to the second embodiment, when the voice recognition result matches a predetermined event condition, the customer information is updated based on the voice recognition result. Combined, more appropriate viewer feedback can be obtained.

[Other Embodiments]
In addition, this invention is not limited to the said embodiment, In the range which does not deviate from the summary of this invention, it can implement in other various forms. For this reason, the said embodiment is only a mere illustration in all points, and is not interpreted limitedly.

  The program of the present invention can be installed or loaded on a computer through various recording media such as an optical disk such as a CD-ROM, a magnetic disk, and a semiconductor memory, or via a communication network. .

  Further, in this specification and the like, the term “means” does not simply mean a physical means, but includes a case where the functions of the means are realized by software. Further, the function of one means may be realized by two or more physical means, or the functions of two or more means may be realized by one physical means.

  (1) In the above embodiment, the case where the digital signage apparatus 100 includes the voice recognition unit 103 has been described. However, the advertisement management apparatus 300 may include the voice recognition unit 103. In this case, the voice recognition management apparatus 200 transmits the selected dictionary to the advertisement management apparatus 300, and the digital signage apparatus 100 transmits the conversation (voice information) input from the voice input unit 102 to the advertisement management apparatus 300. The advertisement management apparatus 300 performs voice recognition on the received conversation using a dictionary.

  (2) The digital signage apparatus 100 may include a photographing unit that photographs the viewer, recognizes the captured face of the viewer, and feeds back the image recognition result as customer information. The speech recognition management apparatus 200 stores various models for image recognition in addition to the speech recognition model, and selects the image recognition model corresponding to the advertisement content and the target layer. The digital signage apparatus 100 recognizes the viewer's face and facial expression from the captured image using the selected image recognition model, and transmits the image recognition result to the advertisement management apparatus 300. This makes it possible to obtain viewer feedback on advertisements and products from a different perspective than conversation.

  (3) In the above embodiment, the digital signage device, the advertisement management device, and the voice recognition management device are described as separate devices, but two or more of the devices can be implemented by one device. In this case, one device includes the corresponding functions among the functions of the digital signage device, the advertisement management device, and the voice recognition management device in the above-described embodiment, and the device is described in the above-described embodiment. The digital signage device, the advertisement management device, and the voice recognition management device may be made to function in the same manner.

  (4) A part or all of the above embodiment can be described as in the following supplementary notes, but is not limited thereto.

(Appendix 1)
Customer information storage means for storing customer information, advertisement display means for displaying advertisements, voice input means for inputting a conversation of a viewer who views the advertisement, and voice for performing voice recognition on the inputted conversation Recognizing means, collecting means for collecting the voice recognition results recognized by the voice, and updating means for updating the customer information stored in the customer information storage means based on the collected voice recognition results. An advertisement distribution system characterized by

(Appendix 2)
A model storage means for storing a plurality of acoustic models and a plurality of language models, respectively, and a selection means for selecting an acoustic model that matches the target layer of the advertisement and a language model that matches the advertisement content of the target advertisement, The advertisement distribution system according to appendix 1, wherein the voice recognition means performs voice recognition on the input conversation using the selected acoustic model and language model.

(Appendix 3)
The updating unit determines whether or not the collected voice recognition result matches a predetermined event condition. When the determination result is positive, the updating unit stores the collected voice recognition result in the customer information storage unit based on the collected voice recognition result. The advertisement distribution system according to appendix 1 or 2, wherein the stored customer information is updated.

(Appendix 4)
The update unit determines that the event condition is satisfied when the conversation content included in the collected speech recognition result includes a preset keyword, and stores the conversation content including the keyword in the customer information storage unit. The advertisement distribution system according to attachment 3, wherein the advertisement distribution system is stored.

(Appendix 5)
An advertisement distribution system configured to allow communication between a digital signage device and an advertisement management device, wherein the digital signage device displays an advertisement transmitted from the advertisement management device, and views the advertisement. Voice input means for inputting a viewer's conversation; voice recognition means for performing voice recognition on the input conversation; and transmission means for transmitting the voice recognition result of the voice recognition to the advertisement management device. The advertisement management device includes an advertisement transmission unit that transmits the advertisement to the digital signage device, a customer information storage unit that stores customer information, a reception unit that receives the transmitted voice recognition result, and the reception And an updating means for updating the customer information stored in the customer information storage means based on the voice recognition result. Stem.

(Appendix 6)
An advertisement distribution method in an advertisement distribution system, comprising: an advertisement display step for displaying an advertisement; a voice input step for inputting a conversation of a viewer who views the advertisement; and a voice for performing voice recognition on the input conversation A recognition step, a collection step of collecting the voice recognition result recognized by the voice recognition, and an update step of updating customer information stored in a customer information database based on the collected voice recognition result. Ad delivery method.

(Appendix 7)
An advertisement distribution method in an advertisement distribution system configured such that a digital signage device and an advertisement management device can communicate with each other, wherein the advertisement management device transmits the advertisement to the digital signage device; and the digital signage In the apparatus, an advertisement display step for displaying an advertisement transmitted from the advertisement management device, a voice input step for inputting a conversation of a viewer who views the advertisement, and a voice for performing voice recognition on the input conversation A recognition step, a transmission step of transmitting the voice recognition result recognized by the voice to the advertisement management device, a reception step of receiving the voice recognition result transmitted in the advertisement management device, and the received voice recognition result Update step to update customer information stored in the customer information database based on Ad delivery method which is characterized in that it comprises, and.

(Appendix 8)
A program for causing a computer to execute each step described in the advertisement distribution method according to attachment 6 or 7.

DESCRIPTION OF SYMBOLS 1 ... Advertisement delivery system, 11 ... Digital signage apparatus, 21 ... Directional microphone, 41 ... Advertisement data input device, 51 ... Advertisement management apparatus, 52 ... Advertisement information DB, 53 ... Customer information DB, 61 ... Voice recognition management device, 62 ... Model DB, 100 ... Digital signage device, 101 ... Advertisement display means, 102 ... Voice input means, 103 ... Voice recognition means, 200 ... voice recognition management apparatus, 201 ... model data input means, 203 ... acoustic model, 205 ... model selection means, 206 ... dictionary transmission means, 300 ... advertisement Management device 301... Advertisement data registration means 303... Advertisement distribution means 304... Customer information collection means 306... Recognition event determination means 400. 401 ... advertising data input transmission means, 202 ... language model DB, 203 ... acoustic model DB, 204 ... advertisement correspondence information DB, 302 ... advertisement information DB, 305 ... customer information DB

Claims (8)

Customer information storage means for storing customer information;
An advertisement display means for displaying an advertisement;
Voice input means for inputting a conversation of a viewer who views the advertisement;
Speech recognition means for performing speech recognition on the input conversation;
A collecting means for collecting the voice-recognized speech recognition result;
Updating means for updating customer information stored in the customer information storage means based on the collected voice recognition results;
An advertisement distribution system comprising: Model storage means for storing a plurality of acoustic models and a plurality of language models respectively;
A selection means for selecting an acoustic model that matches the target layer of the advertisement and a language model that matches the advertisement content of the target advertisement,
The voice recognition means
The advertisement distribution system according to claim 1, wherein speech recognition is performed on the input conversation using the selected acoustic model and language model. The updating means includes
It is determined whether or not the collected voice recognition result matches a predetermined event condition, and when the determination result is positive, the customer stored in the customer information storage unit based on the collected voice recognition result Information is updated, The advertisement delivery system of Claim 1 or 2 characterized by the above-mentioned. The updating means includes
When the conversation content included in the collected voice recognition result includes a preset keyword, it is determined that the event condition is satisfied, and the conversation content including the keyword is stored in the customer information storage unit. The advertisement distribution system according to claim 3. An advertisement distribution system configured to allow communication between a digital signage device and an advertisement management device,
The digital signage device is:
Advertisement display means for displaying an advertisement transmitted from the advertisement management device;
Voice input means for inputting a conversation of a viewer who views the advertisement;
Speech recognition means for performing speech recognition on the input conversation;
Transmitting means for transmitting the voice recognition result recognized by the voice to the advertisement management device,
The advertisement management device includes:
Advertisement transmitting means for transmitting the advertisement to the digital signage device;
Customer information storage means for storing customer information;
Receiving means for receiving the transmitted speech recognition result;
An advertisement distribution system comprising: update means for updating customer information stored in the customer information storage means based on the received voice recognition result. An advertisement distribution method in an advertisement distribution system,
An ad display step for displaying an ad;
A voice input step of inputting a conversation of a viewer who views the advertisement;
A voice recognition step for performing voice recognition on the input conversation;
A collecting step of collecting the speech recognition results recognized by the speech;
An update step for updating customer information stored in a customer information database based on the collected voice recognition results;
An advertisement delivery method comprising: An advertisement distribution method in an advertisement distribution system configured to allow communication between a digital signage device and an advertisement management device,
In the advertisement management device,
An advertisement transmission step of transmitting the advertisement to the digital signage device;
In the digital signage device,
An advertisement display step of displaying an advertisement transmitted from the advertisement management device;
A voice input step of inputting a conversation of a viewer who views the advertisement;
A voice recognition step for performing voice recognition on the input conversation;
A transmission step of transmitting the voice recognition result of the voice recognition to the advertisement management device;
In the advertisement management device,
A receiving step of receiving the transmitted speech recognition result;
An update step of updating customer information stored in a customer information database based on the received voice recognition result;
An advertisement delivery method comprising:   The program for making a computer perform each step as described in the advertisement delivery method of Claim 6 or 7.