JP2017220807A - Voice data collection system - Google Patents

Abstract

PROBLEM TO BE SOLVED: To collect voice data naturally, efficiently and securely.SOLUTION: A voice data collection system 1, which collects voice data of the utterance of a user 2, includes: a voice data collection environment, including a PBX 10, a CTI server 20 and an IVR server 30 to acquire, as voice data, the utterance of the user 2 in speech communication on the telephone received from the user 2; and a voice analysis server 40 which is connected to each IVR server 30 through a network to process the voice data acquired from each IVR server 30 to enable an information processing terminal 53 of the user 2 to read the processing result. In the speech communication on the telephone received from the user 2, the voice data collection environment issues a query to the user 2 to ask for an answer, to record the utterance of the answer from the user 2 to obtain voice data, so as to repeat the issue of queries and the acquisition of answer voice data from the user 2 until the voice data reaches the predetermined number.SELECTED DRAWING: Figure 1

Description

  The present invention relates to a technique for collecting data from a large number of specific and unspecified users, and more particularly to a technique that is effective when applied to a voice data collection system that collects voice data of users.

  There is a speech pathological analysis technique that grasps a constantly changing mental state that is difficult for the person to understand based on speech that a person utters daily. In this technology, the involuntary component included in the voice of the person (the real intention) that the person naturally put out and the voluntary component that the person tried to convey to the other party (the erection) Quantify and visualize emotional states based on ingredients.

  Using this technology, MIMSYS (Mind Monitoring Systems, Non-Patent Document 1, Registered Trademark (hereinafter referred to as the “registered trademark”)) can collect and analyze voices uttered daily by a person, monitor the state of the mind, and visualize the analysis results. The same system is also developed (see, for example, JP-A-2015-128579 (Patent Document 1) for details). According to this, it is possible to measure and express the user's stress and the state of the heart as a numerical value, such as a state in which the state of the heart is normal, upward, and active, or a state in which the state is low and sometimes needs rest. And by widely utilizing this technology, it is possible to detect illnesses such as “depression” at an early stage before the person becomes aware of it, and prevent diseases by taking appropriate treatment and countermeasures. Expected to be.

Japanese Patent Laying-Open No. 2015-128579

"MIMOSYS | PST Corporation for Speech Pathology Analysis Technology", [online], PST Corporation, [March 18, 2016 search], Internet <URL: http://medical-pst.com/products-2/ mimosys>

  In the above-described conventional technology, for example, it is possible to perform from voice collection / analysis to visualization by the function of an application program installed in a mobile terminal such as a smartphone.

  On the other hand, when using an application installed in a portable terminal, the user must speak the voice actively after installing and starting the application. In this regard, for example, it is considered effective to acquire the contents of a call when the user makes a general call with a mobile phone, but it is uncertain when the call by the telephone itself is performed. Therefore, these methods alone are highly uncertain and efficient in collecting user voice data. In order to realize the purpose of preventing illness as a whole, it is necessary to have a voice data collection mechanism that can be widely used by various users in various situations and situations.

  Accordingly, an object of the present invention is to provide an audio data collection system that can collect audio data more naturally, efficiently and reliably.

  The above and other objects and novel features of the present invention will be apparent from the description of this specification and the accompanying drawings.

  Of the inventions disclosed in this application, the outline of typical ones will be briefly described as follows.

  An audio data collection system according to an exemplary embodiment of the present invention is an audio data collection system that collects audio data related to a user's utterance, and is configured for each of one or more sites. , A CTI server, and an IVR server, and a voice data collection environment for acquiring the user's utterance in a telephone call received from the user as the voice data, and each IVR in each voice data collection environment A voice analysis server that is connected to a server via a network and that processes the voice data acquired from each of the IVR servers and allows the processing result to be viewed on the information processing terminal of the user.

  The voice data collection environment issues a question asking the user for an answer in a telephone call received from the user, records an utterance related to the answer from the user to the question, and records the voice As the data, until the voice data reaches a predetermined number, the issuing of the question and the acquisition of the voice data related to the answer from the user are repeated.

  Among the inventions disclosed in the present application, effects obtained by typical ones will be briefly described as follows.

  That is, according to a typical embodiment of the present invention, it is an object to provide an audio data collection system capable of collecting audio data more naturally, efficiently and reliably.

It is the figure which showed the outline | summary about the structural example of the audio | voice data collection system which is one embodiment of this invention. It is the figure which showed the outline | summary about the example of the flow of a process concerning collection and analysis of the audio | voice data in one embodiment of this invention. (A), (b) is the figure which showed the outline | summary about the example of the screen displayed on an information processing terminal as an analysis result in the audio | voice analysis server in one embodiment of this invention. It is the figure which showed the outline | summary about the other example of the flow of the process which concerns on collection and analysis of the audio | voice data in one embodiment of this invention. It is the figure which showed the outline | summary about the example of a screen of the application which collects audio | voice data with the smart phone in one embodiment of this invention. It is the figure which showed the outline | summary about the example of a data structure of authentication master DB of the IVR server in one embodiment of this invention. It is the figure which showed the outline | summary about the example of the data structure of audio | voice data DB of the IVR server in one embodiment of this invention. It is the figure which showed the outline | summary about the example of a data structure of audio | voice data DB of the audio | voice analysis server in one embodiment of this invention. It is the figure which showed the outline | summary about the example of a data structure of user master DB of the audio | voice analysis server in one embodiment of this invention. It is the figure which showed the outline | summary about the example of the data structure of the analysis result DB of the audio | voice analysis server in one embodiment of this invention.

  Hereinafter, embodiments of the present invention will be described in detail with reference to the drawings. Note that components having the same function are denoted by the same reference symbols throughout the drawings for describing the embodiment, and the repetitive description thereof will be omitted. On the other hand, parts described with reference numerals in some drawings may be referred to with the same reference numerals although not illustrated again in the description of other drawings.

<System configuration>
FIG. 1 is a diagram showing an outline of a configuration example of an audio data collection system according to an embodiment of the present invention. The voice data collection system 1 according to the present embodiment acquires and collects voices uttered by the user 2 via the telephone 51 and the smartphone 52, and analyzes the voices using a voice pathological analysis engine such as the above-mentioned MIMSSYS, and the result. Is a system for visualizing the user 2 so that the user 2 can view it.

  The voice data collection system 1 includes a voice data collection environment including, for example, a PBX (Private Branch eXchange) 10, a CTI (Computer Telephony Integration) server 20, an IVR (Interactive Voice Response) server 30, and This is an information processing system including each subsystem of the voice analysis server 40. For example, a plurality of voice data collection environments can be constructed independently for each base such as an office building.

  The PBX 10 is an exchange that accepts calls by the telephones 51 from a plurality of users 2 corresponding to a predetermined telephone number. The telephone number may be, for example, a free dial (registered trademark) or an extension number in an office building or the like. A commonly used PBX device can be used as the PBX 10 as appropriate. Note that the telephone 51 is not particularly limited, such as a fixed phone, a mobile phone, and a smartphone. It may be an extension telephone.

  The CTI server 20 is a server device having a function of linking access by a call from the telephone 51 received by the PBX 10 to an information processing system such as the IVR server 30. Also for the CTI server 20, commonly used CTI equipment can be used as appropriate. In the present embodiment, the CTI server 20 includes, for example, a collection control unit 21 implemented as a software program or script that runs on middleware such as an OS (Operating System), a DBMS (DataBase Management System), or a CTI software (not shown). Have The collection control unit 21 executes and controls a series of processing flows for acquiring and collecting voice data associated with the utterance from the user 2 in cooperation with the IVR server 30 described later.

  The IVR server 30 is a server device that cooperates with the CTI server 20 and has a function of outputting a predetermined voice guidance based on contents set in advance with respect to a call made by the telephone 2 of the user 2. In the present embodiment, unlike in a general CTI / IVR voice guidance system or call center system, the user 2's voice data is acquired from the call contents and recorded in the voice data DB 34. It provides voice guidance that encourages utterances.

  The IVR server 30 includes various units such as an authentication unit 31 and a voice collection unit 32 that are implemented as software programs or scripts that run on middleware such as an OS, DBMS, or IVR software (not shown). Each data store includes an authentication master database (DB) 33 and a voice data DB 34 implemented as a database.

  The authentication unit 31 has a function of authenticating the user 2 who is accessing by telephone from the telephone 51. In the present embodiment, as will be described later, the user 2 inputs the ID code by operating the telephone 51 and pushing a series of numbers, and whether or not the input ID code is registered in the authentication master DB 33. Is verified by the authentication unit 31. The authentication method is not limited to this. For example, information other than the ID code may be input and collated. Further, based on information of the voice of the user 2 acquired by the voice collecting unit 32 described later, authentication may be performed using a known voiceprint authentication technique or the like.

  The voice collection unit 32 has a function of automatically performing voice guidance for a call from the user 2 by the telephone 51 based on a voice guidance pattern registered in the voice guidance information 35 in advance. Then, the content of the voice uttered by the user 2 as a response to the guidance is recorded as digital data and recorded in the voice data DB 34.

  In order to accurately analyze the voice pathological condition by the voice analysis server 40 described later based on the voice information, a plurality of voice data (for example, about 7 or 8) in units that can be divided as utterances are acquired. Is desirable. In this embodiment, as will be described later, a plurality of questions are issued to the user 2 as voice guidance so that a plurality of voice data can be acquired, and the answers are recorded as voice data.

  Questions with a predetermined pattern may be used each time, or a required number of questions may be selected from among candidate questions prepared in advance at random or based on a predetermined standard. The user attribute information may be acquired from the authentication master DB 33 or the like based on the ID code of the user 2 and the question may be switched accordingly. Also, the answer contents may be immediately recognized and analyzed, and the question may be switched according to the answer contents.

  The voice data collection environment including the PBX 10, the CTI server 20, and the IVR server 30 provides a service for collecting voice data when the user 2 makes a call using the telephone 51. In the present embodiment, the voice data collection means is not limited to this. For example, voice data recorded on the smartphone 52 by the user 2 using the smartphone 52 can be used. In this case, it is possible to acquire voice data regardless of the call quality of the carrier of the telephone service. The voice data collected by these means is commonly used for analysis by the voice analysis server 40 to be described later without being distinguished by the collection means. Thus, by providing a plurality of voice data collection means, it is possible to collect voice data with simple and appropriate means according to the situation of the user 2.

  The voice analysis server 40 acquires voice data from the IVR server 30 of each voice data collection environment, the smartphone 52 used by the user 2, and the like and records it as the voice data DB 45, and the contents thereof are voice pathological analysis such as the above-mentioned MIMSSYS. It is a server device that analyzes the engine and visualizes the result so that the user 2 can view it. For example, it can be configured by a virtual server constructed in a cloud computing environment, and can provide a voice pathological analysis function as a cloud service for the voice data collection environment at each site.

  The voice analysis server 40 according to the present embodiment includes, for example, a voice data acquisition unit 41, a voice analysis unit 42, and an analysis result processing unit implemented as a software program that runs on middleware such as an OS, a DBMS, and a Web server program (not shown). 43 and a user interface (IF) unit 44. Each data store includes a voice data DB 45, a user master DB 46, and an analysis result DB 47 implemented as a database.

  The voice data acquisition unit 41 has a function of acquiring voice data from the IVR server 30 in the voice data collection environment, the smartphone 52 used by the user 2, and the like and recording it as the voice data DB 45. For example, the unprocessed audio data DB 34 acquired by the IVR server 30 at each site is transmitted to the audio analysis server 40 by batch processing at regular intervals (for example, every hour). When transmission in real time or near this is possible due to the size of audio data, network bandwidth, etc., the audio data acquired by the IVR server 30 is transmitted to the audio analysis server 40 in real time each time. A configuration is also possible.

  Further, audio data transmitted in real time from an application (not shown) on the smartphone 52 may be received and recorded in the audio data DB 45. The voice data acquisition unit 41 performs processing and processing such as format conversion on voice data transmitted from the IVR server 30 in the voice data collection environment or the smartphone 52 of the user 2 as necessary, and records the data in the voice data DB 45. .

  The voice analysis unit 42 has a function of performing voice pathological analysis on the voice data for each user 2 recorded in the voice data DB 45 to grasp the state of mind (current state, time-series change state). In the present embodiment, the speech pathological analysis engine such as the above-described MIMMOS is included. Triggered by newly recorded audio data in the audio data DB 45, the unprocessed audio data may be analyzed in real time each time, or at regular intervals (for example, every hour). The unprocessed audio data stored in the database may be analyzed collectively.

  The analysis result processing unit 43 has a function of acquiring the result of the voice pathological analysis performed by the voice analysis unit 42, visualizing the result for each user 2, and recording the result in the analysis result DB 47. Visualization refers to, for example, conversion of output data of analysis results into character information, numerical information, evaluation information, etc. that can be presented to the user 2, and screen data or image data that can be displayed on the information processing terminal 53. Including the process of generating At this time, for example, the user master DB 46 that holds attribute information for each user 2, unique setting information, and the like may be referred to, and visualization according to the user 2 may be performed.

  The user IF unit 44 has a function of accepting a request from the user 2 via the information processing terminal 53 and displaying the visualized data recorded in the analysis result DB 47 on the screen of the information processing terminal 53. It may have a function of newly registering account information and issuing an ID code for an unspecified user 2 who has not yet registered as a user and does not have account information such as an ID code.

  As the information processing terminal 53, for example, a general-purpose information processing terminal such as a PC (Personal Computer), a smartphone, or a tablet terminal can be used as appropriate. When the smartphone 52 for recording voice data is provided, it may be used as the information processing terminal 53 as it is. One user 2 may use different types of information processing terminals 53 depending on the situation. For example, the user 2 accesses the voice analysis server 40 on the cloud computing environment using a web browser (not shown) on the information processing terminal 53.

  In the example of FIG. 1, the PBX 10, the CTI server 20, the IVR server 30, and the voice analysis server 40 are configured as individual devices and server systems, respectively, but the configuration is not limited thereto. Furthermore, the functions may be configured to be distributed among a plurality of server systems, or conversely, the functions of the plurality of server systems may be integrated into one server system.

<Processing flow (voice data collection via telephone call)>
FIG. 2 is a diagram showing an outline of an example of the flow of processing related to the collection and analysis of audio data in the present embodiment. Here, a case where the voice data collection environment acquires and collects voice data from a call made by the user 2 using the telephone 51 will be described. First, the user 2 makes a call to a predetermined telephone number by the telephone 51 (S01). When this call is received by the corresponding base, that is, the PBX 10 in the voice data collection environment, first, the collection control unit 21 of the corresponding CTI server 20 determines whether it is within the service time (S02). When it is outside the service time, although not shown, the voice guidance to that effect is responded to the telephone 51 of the user 2 in cooperation with the voice collecting unit 32 of the corresponding IVR server 30, that is, the entire call is terminated. The process ends.

  If it is within the service time, an authentication guidance consisting of voice prompting the input of identification information such as an ID code is performed on the telephone 51 of the user 2 in cooperation with the voice collecting unit 32 of the IVR server 30. A response is made (S03). Here, for example, the user 2 is instructed to push and input a series of numbers forming an ID code by using a push button or a numeric keypad of the telephone 51.

  When the user 2 inputs an ID code or the like according to the authentication guidance (S04), the collection control unit 21 of the CTI server 20 registers the input ID code or the like in the authentication master DB 33 in cooperation with the authentication unit 31 of the IVR server 30. The user is authenticated depending on whether or not it is done (S05). As described above, the authentication method is not limited to this, and instead of this, or in addition to this, for example, the user 2 is urged to speak, and the voice print or the like of the acquired voice data is analyzed and authenticated. You may use the method of.

  If the authentication fails, that is, if the input ID code or the like is not registered in the authentication master DB 33, a predetermined number of times (for example, three times) is allowed, although not shown. If the predetermined number of retries is exceeded, voice guidance indicating that the authentication could not be performed is responded to the telephone 51 of the user 2 and the call is terminated.

  When the authentication is successful, in cooperation with the voice collecting unit 32 of the IVR server 30, voice guidance for making a predetermined question to the user 2 is returned to the telephone 51 of the user 2 (S06). When the user 2 utters an answer to the question using the telephone 51 (S07), the collection control unit 21 of the CTI server 20 records the utterance content as voice data in cooperation with the voice collection unit 32 of the IVR server 30 (S08). . The obtained voice data is recorded in the voice data DB 34. If the audio data cannot be properly acquired or recorded, retrying a predetermined number of times (for example, 3 times) is allowed, although not shown. When the number of retries is exceeded, voice guidance indicating that voice data could not be acquired is responded to the telephone 51 of the user 2 and the call is terminated.

  If the voice data can be recorded, it is determined whether or not the voice data has been recorded a predetermined number of times during the call, that is, whether or not a predetermined number or more of voice data has been acquired (S09). If it is less than the predetermined number, the process returns to step S06 to ask the next (nth) question. The predetermined number for acquiring the voice data is appropriately set according to the specifications of the voice pathological analysis engine used in the voice analysis unit 42 of the voice analysis server 40, the required accuracy, and the like. For example, in the case of MIMOSYS, it is sufficient that there are currently seven or more audio data.

  When a predetermined number of voice data can be acquired, in cooperation with the voice collection unit 32 of the IVR server 30, a termination guidance for terminating the question and terminating the call is returned to the telephone 51 of the user 2 (S10). ). Then, the user 2 ends the telephone call using the telephone 51 (S11). The collection control unit 21 of the CTI server 20 may lead and end the call. If the call with the user 2 is disconnected before the predetermined number of audio data can be acquired, only the audio data acquired up to that time may be used.

  As described above, in the present embodiment, a predetermined number of questions are repeated to obtain answer voice data. As described above, the question may be used in a predetermined pattern every time, or a necessary number of questions may be selected from random candidate questions or selected based on a predetermined criterion. Good. The user attribute information may be acquired based on the ID code of the user 2 and the question may be switched according to the user's attribute information, or the answer content may be immediately recognized and analyzed, and the question may be switched according to the answer content. Good. By giving such a change, it is possible to prevent the user 2 from getting bored as much as possible.

  If not all questions are required to be different, one or more of each question may be the same content as the others. Also, instead of letting the user 2 speak only for the purpose of collecting voice data, for example, it is combined with other services and services by telephone such as telephone reservation, and the contents spoken at that time are acquired and diverted. Also good.

  The voice data recorded by the above series of processes is transferred from the IVR server 30 to the voice analysis server 40 at regular time intervals such as every hour (S21). In the voice analysis server 40, the voice data acquisition unit 41 records the acquired voice data in the voice data DB 45 (S22). Then, for example, the speech analysis unit 42 performs speech pathological analysis processing on unprocessed speech data at regular time intervals (S23). It is desirable to delete the processed audio data in consideration of privacy and the like. Thereafter, the content of the analysis result is visualized by the analysis result processing unit 43 and recorded in the analysis result DB 47 (S24). With this series of processes, the user 2 is ready to display and browse the analysis results. You may make it notify to the target user 2 by e-mail, a push notification, etc. that the analysis was completed.

  The user 2 accesses the user IF unit 44 of the voice analysis server 40 by using a Web browser or the like on the information processing terminal 53 and inputs account information at an arbitrary timing after the analysis by the voice analysis server 40 is completed. And login (S31). As the account information, for example, the ID code and password of the user 2 can be used, but other account information may be used, and authentication information for other methods such as biometric authentication may be included. Good.

  The voice analysis server 40 performs user authentication based on the registered contents of the user master DB 46 (S32). If the authentication fails, that is, if the input ID code, password, etc. are not registered in the user master DB 46 or do not match the registered contents, a screen indicating that the authentication could not be performed is not shown. The information is displayed on the processing terminal 53 and the processing is terminated.

  When the authentication is successful, the analysis result data relating to the target user 2 is acquired from the analysis result DB 47, and a display screen (for example, HTML (HyperText Markup Language) data) is generated and output (S33). The information processing terminal 53 displays the output analysis result screen on a web browser (not shown) (S34). Thereby, the user 2 can grasp the state of the mind as a result of performing the speech pathological analysis on the speech data related to his / her speech.

<Screen example>
FIG. 3 is a diagram showing an outline of an example of a screen displayed on the information processing terminal 53 as an analysis result in the voice analysis server 40. After the user 2 logs in by entering account information such as an ID code and a password on a login screen (not shown), for example, an analysis result screen as shown in FIG. 3A is displayed. This screen shows an evaluation based on the state of change in time series (for example, the latest two weeks) as the “heart activity value”. Here, for example, the range of the measured value is divided into several stages, and the content of the category corresponding to the measured value (in the example shown in the figure, “Glitter ★”) makes it easy for the user 2 to understand intuitively. I have to.

  Further, when the user 2 switches the screen, an analysis result screen as shown in FIG. 3B is displayed. On this screen, as the “original pressure”, the evaluation of the spirit at the time when the target call is made, that is, when the voice data is acquired, is displayed on the meter image so that the user 2 can easily grasp intuitively. Show. In any of these screens, it is possible to browse the history of past analysis results. Note that the screen shown in FIG. 3 is merely an example, and it goes without saying that other display formats may be used.

<Process flow (voice data collection by smartphone)>
FIG. 4 is a diagram showing an outline of another example of the flow of processing relating to the collection and analysis of audio data in the present embodiment. Here, a case where the user 2 acquires and collects voice data using the voice recording function of the smartphone 52 will be described. First, the user 2 starts a dedicated application on the smartphone 52 (S41, S42). When the application on the smartphone 52 is activated, first, an authentication screen is displayed (S43).

  When the user 2 inputs account information such as a user ID and a password according to the instruction on the authentication screen (S44), the application of the smartphone 52 makes an authentication request to the voice analysis server 40 (S45). The voice analysis server 40 performs user authentication based on the registered contents of the user master DB 46 (S46). If the authentication fails, that is, if the input ID code, password, etc. are not registered in the user master DB 46 or do not match the registered contents, a screen indicating that the authentication could not be performed is displayed on the smartphone. It displays on the application of 52, and complete | finishes a process.

  When the authentication is successful, the application of the smartphone 52 displays a screen for asking a predetermined question to the user 2 (S47). The contents of the question can be appropriately changed by updating the application or downloading from a server such as the voice analysis server 40. When the user 2 utters an answer to the question to the microphone of the smartphone 52 (S48), the smartphone 52 records the utterance content as voice data by the voice recording function that the user 52 has (S49). The obtained audio data is recorded in a memory or the like. If the audio data cannot be properly acquired or recorded, retrying a predetermined number of times (for example, 3 times) is allowed, although not shown. If the number of retries is exceeded, a screen indicating that voice data could not be acquired is displayed on the application of the smartphone 52, and the process ends.

  If the audio data can be recorded, it is determined whether or not the audio data has been recorded a predetermined number of times during execution of the application, that is, whether or not the audio data has been acquired a predetermined number or more (S50). If it is less than the predetermined number, the process returns to step S47 to ask the next (nth) question. As in the case of the telephone 51 in FIG. 2 described above, the predetermined number for acquiring the voice data is appropriately determined according to the specifications of the voice pathological analysis engine used in the voice analysis unit 42 of the voice analysis server 40, required accuracy, and the like. Set. Basically, there is no difference based on the difference between the call quality at the telephone 51 and the recording quality at the smartphone 52. For example, in the case of MIMSSYS, it is sufficient that there are currently seven or more voice data.

  If a predetermined number of audio data can be acquired, a screen to end the question is displayed (S51). Thereafter, the voice data recorded by the series of processes described above is transferred to the voice analysis server 40 (S52). In the voice analysis server 40, the voice data acquisition unit 41 records the acquired voice data in the voice data DB 45 (S53). The speech analysis unit 42 performs speech pathological analysis processing on the unprocessed speech data (S54), and the analysis result contents are visualized by the analysis result processing unit 43 and recorded in the analysis result DB 47. (S55). With this series of processes, the user 2 is ready to display and browse the analysis results. The user 2 can display and refer to the analysis result as shown in FIG. 3 using the application of the smartphone 52 immediately or within a short time after recording the voice data (S56, S57). ).

  In the example of FIG. 4, the voice data recorded on the smartphone 52 in step S49 is transferred to the voice analysis server 40 at the time when all the questions are completed (step S52). The transfer method is not limited to this. For example, the voice data recorded in step S49 is transmitted to the voice analysis server 40 each time, thereby reducing the processing burden after the question is finished and reducing the processing time until the analysis result is output. Also good.

<Screen example>
FIG. 5 is a diagram showing an overview of an example screen of an application that collects audio data by the smartphone 52. The upper left screen is an authentication screen displayed at the time of activation, and the user 2 performs user authentication by inputting account information such as a user ID and a password via this screen. Note that the account information once input may be stored using Cookie or the like, and authentication may be automatically performed after the next time. When the user authentication is successful, a start and guidance screen as shown in the upper center is displayed. When the “Start” button is pressed here, a question screen as shown in the upper right is displayed, and the voice data acquisition process starts.

  The user 2 taps the “microphone” button on the screen, and utters an answer to the question item displayed on the screen. At this time, the screen is a recording screen as shown in the lower left. The application uses the voice recording function of the smartphone 52 to record the speech of the user 2 input from the microphone as voice data. Recording is ended when the user 2 taps a “stop” button on the lower left screen.

  When the recording is finished, the process for the next question is repeated. The number of questions that have been answered at this time and the remaining questions are indicated by the numbers of ● (answered) and ○ (unanswered) marks at the top of the screen. When answers to predetermined number of questions (recording of voice data) are completed, an end screen as shown in the lower center is displayed. At this time, the recorded voice data is automatically transmitted from the smartphone 52 to the voice analysis server 40, and the voice pathological analysis process is performed in real time by the voice analysis server 40.

  Thereafter, for example, the user 2 selects an item “go to analysis result site” from the menu screen as shown on the lower right side, so that the analysis result screen as shown in FIG. It can be displayed on the browser and the analysis result can be confirmed immediately.

<Data structure>
FIG. 6 is a diagram showing an overview of an example of the data configuration of the authentication master DB 33 of the IVR server 30. The authentication master DB 33 is a master table that holds authentication data for performing user authentication when a call is made by the telephone 51 for acquiring voice data for each user 2, and includes, for example, an ID code and a user name. Each item.

  The ID code item is a code value composed of several digits uniquely assigned to the target user 2. Since it is input by pushing a button on the telephone 51, it is assumed to be composed of numbers. However, if input is possible, it may include characters and symbols. The user name item holds user name and name information as attribute information of the target user 2. For example, when the user 2 is authenticated, the user 2 may confirm that the user 2 has been correctly authenticated by responding with the user name as a voice by a known text-to-speech function.

  FIG. 7 is a diagram showing an outline of an example of the data configuration of the audio data DB 34 of the IVR server 30. The voice data DB 34 is a table that holds recorded voice data for each utterance in the call of the user 2, and includes items such as an ID code, a base code, a call number, a data number, voice data, and a recording date and time. Have.

  The item of ID code holds a code value that identifies the user 2 who has made a speech related to the target voice data. This item is the same as the ID code item in the authentication master DB 33 of FIG. The base code item holds information such as an ID and a code for specifying a base where the target voice data is collected, that is, a voice data collection environment to which the PBX 10 or the like that has received a call related to the target voice data belongs. The telephone number corresponding to PBX10 may be sufficient. The call number item holds information such as a sequence number for uniquely identifying each call in each user 2. Further, the data number item holds information such as a sequence number for uniquely identifying each utterance in each call. That is, the target speech (voice data) can be uniquely specified in the entire voice data collection system 1 by the combination of the items of the ID code, the call number, and the data number.

  The audio data item holds information of audio data recorded in the target speech. The audio data itself may be held directly, or the audio data may be held separately as a file, and the file name and path information may be held. The recording date / time item holds time stamp information when the voice data related to the target speech is recorded.

  FIG. 8 is a diagram showing an outline of an example of the data configuration of the voice data DB 45 of the voice analysis server 40. The voice data DB 45 is a table for recording information of voice data sent from the IVR server 30 at each site or the smartphone 52 of each user 2, and basically the same data as the voice data DB 34 in FIG. It has a configuration. Since voice data from the smartphone 52 is also included, in place of the call number specifying the call on the telephone 51, the call number is replaced with the call number item so that information on the utterance number specifying each utterance on the smartphone 52 can be held.・ Has an utterance number field. Further, in addition to the data structure of the voice data DB 34, for example, items such as analysis status are included. The analysis status item holds information such as a code value and a flag for identifying whether the analysis of the target audio data has been performed or not.

  FIG. 9 is a diagram showing an outline of an example of the data configuration of the user master DB 46 of the voice analysis server 40. The user master DB 46 is a master table that holds the account information of the user 2 for browsing the result of voice pathological analysis based on the voice data. For example, the ID code, password, user name, user attribute information, and user setting information Each item.

  The ID code item holds a code value that uniquely identifies the target user 2. This item may be the same as the item of the ID code in the authentication master DB 33 of FIG. The password item holds password information corresponding to the ID code of the target user 2. In addition to or instead of the password, authentication information related to biometric authentication or other authentication methods may be held. Further, it may have an item for holding information such as password registration date and time and update date and time.

  The item of the user name holds information on the display name of the target user. The item of user attribute information holds various attribute information of the target user 2. For example, sex and age can be included. These attribute information can be used as reference information when performing processing such as voice pathological analysis and visualization of analysis results, for example. The user setting information item holds the contents of the unique setting information set by the target user 2. For example, each user 2 can customize the contents such as customization of conditions when performing processing such as voice pathological analysis and visualization of analysis results, and customization of display items and display methods on the screen displaying the analysis results as shown in FIG. Can be set individually. The data format of user attribute information and user setting information is not particularly limited, and can be arbitrary.

  FIG. 10 is a diagram showing an outline of an example of the data configuration of the analysis result DB 47 of the voice analysis server 40. The analysis result DB 47 is a table that holds data visualizing the result of the voice pathological analysis performed on the voice data for each user 2, and includes, for example, an ID code, a base code, a call / utterance number, analysis result data, and Each item includes analysis date and time.

  The ID code item holds a code value that uniquely identifies the target user 2. This item is the same as the ID code item in the user master DB 46 of FIG. 9 described above. The item of the base code holds information such as a sequence number and a code for identifying the base from which the voice data related to the target analysis result is collected. This item is the same as the item of the base code in the voice data DB 34 of FIG. 7 and the voice data DB 45 of FIG.

  The item of the call / speech number is a sequence number or the like for identifying a call or utterance that collects voice data related to the target analysis result. This item is the same as the call / speech number item in the voice data DB 34 of FIG. 7 and the voice data DB 45 of FIG. The item of the analysis result data holds information relating to the contents of the target analysis result visualized. The data itself may be held, or the path of the data location may be held. The data format and the like are not particularly limited and can be arbitrary. The item of analysis date and time holds time stamp information when the analysis processing and the analysis result visualization processing by the voice analysis unit 42 and the analysis result processing unit 43 are performed.

  Note that the data configuration (items) of each table shown in FIGS. 6 to 10 is merely an example, and other table configurations and data configurations may be used as long as similar data can be held and managed. It may be.

  As described above, according to the voice data collection system 1 which is an embodiment of the present invention, a necessary number of voice data is obtained from a large number of unspecified / specific users 2 in a more natural, efficient and reliable manner. And can be used as an input to a process that requires voice data such as voice pathological analysis.

  As mentioned above, the invention made by the present inventor has been specifically described based on the embodiments. However, the present invention is not limited to the above-described embodiments, and various modifications can be made without departing from the scope of the invention. Needless to say. For example, the above-described embodiment has been described in detail for easy understanding of the present invention, and is not necessarily limited to the one having all the configurations described. In addition, it is possible to add, delete, and replace other configurations for a part of the configuration of each of the above embodiments.

  For example, in the above-described embodiment, a predetermined number of questions are asked in one call or utterance, and a predetermined number of voice data are all acquired as an answer to the question. Absent. If a predetermined number of voice data is obtained in units of one day, it is not necessary to acquire all the number of voice data in one call or utterance, and a predetermined number in total of multiple calls or utterances. The voice data may be acquired. Further, the audio data in this case is not limited to the data obtained by the series of processes shown in FIGS. For example, regardless of this, the contents of telephone calls made in daily life or business are automatically recorded, and the existing voice data obtained here is also diverted to a predetermined number of voices as a whole. It is also possible to secure data.

  Further, in the above embodiment, the user 2 needs to actively make a call with the telephone 51 or start an application with the smartphone 52, but these actions are performed more reliably. Such a mechanism may be provided separately. For example, the user 2 may be automatically called by the function of the CTI server 20 or the IVR server 30 from the voice data collection environment side at a predetermined time every day. If the telephone is not answered, a telephone call may be made after a predetermined time. Similarly, a dedicated application on the smartphone 52 may make a notification prompting execution at a predetermined time.

  Further, at a predetermined time every day, the analysis result by the voice analysis server 40 may be transmitted as a report to the user 2 by e-mail or the like for reminding. Alternatively, notification may be made by e-mail or the like when there is no use for a predetermined period from the most recent use date and time (that is, the date and time when voice data was recorded and analyzed).

  The present invention is applicable to a voice data collection system that collects user voice data.

1 ... voice data collection system, 2 ... user,
10 ... PBX,
20 ... CTI server, 21 ... collection control unit,
30 ... IVR server, 31 ... Authentication unit, 32 ... Voice collection unit, 33 ... Authentication master DB, 34 ... Voice data DB, 35 ... Voice guidance information,
DESCRIPTION OF SYMBOLS 40 ... Voice analysis server, 41 ... Voice data acquisition part, 42 ... Voice analysis part, 43 ... Analysis result process part, 44 ... User IF part, 45 ... Voice data DB, 46 ... User master DB, 47 ... Analysis result DB,
51 ... Telephone, 52 ... Smartphone, 53 ... Information processing terminal

Claims (8)

An audio data collection system for collecting audio data related to a user's utterance,
A PBX, a CTI server, and an IVR server configured for each of one or more locations, and acquiring the user's utterance in a telephone call received from the user as the voice data Voice data collection environment,
Voice connected to each IVR server in each voice data collection environment via a network, processing the voice data acquired from each IVR server, and enabling processing results to be viewed on the information processing terminal of the user An analysis server,
The voice data collection environment issues a question asking the user for an answer in a telephone call received from the user, and records the utterance related to the answer from the user to the question as the voice data. The voice data collection system that repeats issuing the question and obtaining the voice data related to the answer from the user until the voice data reaches a predetermined number. The voice data collection system according to claim 1,
The voice data collection system in which one or more of the questions issued in a telephone call received from the user by the voice data collection environment have the same contents as others. The voice data collection system according to claim 1,
When the voice data collection environment receives a telephone call from the user, one or more numbers input via the telephone by the user match one of one or more code values registered in advance. A voice data collection system that issues the question. The voice data collection system according to claim 1,
The voice analysis server performs voice pathological analysis on the voice data acquired from the IVR server, generates screen information for display based on the analysis result, and requests the user from the information processing terminal. And a voice data collection system for causing the information processing terminal to display contents based on the screen information. The voice data collection system according to claim 1,
The voice data collection system, wherein the voice data collection environment transmits the acquired voice data to the voice analysis server at predetermined time intervals. An audio data collection system for collecting audio data related to a user's utterance,
An information processing terminal having a voice recording function;
A voice analysis server that processes the voice data recording the user's utterance acquired from the information processing terminal, and enables a processing result to be viewed on the information processing terminal;
The information processing terminal displays a question asking the user for an answer, records an utterance related to the answer from the user to the question as the voice data, and until the voice data reaches a predetermined number, An audio data collection system that repeats display of the question and acquisition of the audio data related to an answer from the user. The voice data collection system according to claim 6,
Each of the questions displayed by the information processing terminal is an audio data collection system in which one or more of them have the same content as the others. The voice data collection system according to claim 6,
The speech analysis server performs speech pathological analysis on the speech data acquired from the information processing terminal, generates screen information for display based on the analysis result, and requests from the information processing terminal of the user And receiving the content based on the screen information on the information processing terminal.

Images