JP2012033089A - Cross statistical system, utterance terminal, cross statistical method - Google Patents

Abstract

PROBLEM TO BE SOLVED: To realize cross statistics and mining, while concealing attribute information of users, by using the information owned by multiple administrators.SOLUTION: A cross statistical system comprises: space provision means having space provision recording section for providing extended real space; multiple concealment calculation means; and analysis means. The analysis means has an arithmetic acquisition section, a plaintext calculation section and a concealment calculation request section. The plaintext calculation section performs arithmetic, which are calculations performed without using encrypted attribute information (pieces). The concealment calculation request section requests all the concealment calculation means to perform arithmetic, which are concealment calculations performed by using encrypted attribute information (pieces). The analysis means analyses speakers recorded in the different space provision means as the same person if there is information to identify the speakers, which corresponds to each other. All the concealment calculation means performs concealment calculations in cooperation by using the encrypted attribute information (pieces).

Description

  The present invention relates to a cross-statistic system and a speech terminal considering privacy protection when there are a plurality of space providing means for providing an augmented reality space in which a geographical object and a speech object linked to the geographical object are arranged. , Related to crossing statistics.

  With the widespread use of GPS-equipped personal network terminals and various sensors, services using augmented reality space (AR space) such as “Sekai Camera (registered trademark)” (Non-Patent Document 1) have begun to spread. In addition, the information provider has started to provide information by extending the framework. The augmented reality space realization technology targeted by this patent is a geographical network such as the latitude / longitude (/ height) of the real world using a personal network terminal equipped with a position acquisition sensor such as GPS. This is a technique for registering utterances and information called an air tag or the like for a special place, and allowing the air tag at the current position to be viewed on a video from a camera.

  In “Sekai Camera (registered trademark)”, attribute information (profile information) of a speaker (user) can be displayed, but disclosure control and encryption are not performed. A function called Air Profile floats a profile tag around the overhead of the user who started Sekai Camera (registered trademark), but is not encrypted. A function group called Sekai Life can refer to a speaker's list of statements for a certain statement, but not a speaker's profile.

Next, the secret calculation will be described. There are various methods of secret calculation, and there are the following two typical ones.
1) A secret key is distributed and held by a plurality of secret computation entities, the data provider encrypts data using a public key corresponding to the secret key, and a plurality of secret computation entities use the encrypted data. A method to perform computations in cooperation with encryption (Non-Patent Documents 2 and 3)
2) A method in which two secret data pieces (encrypted attribute information pieces to be described later) are created, and the secret calculation subject holding each of them performs the calculation while keeping it secret. Although the method is covered, the explanation will be complicated, so the following description will focus on the method 2.

  FIG. 1 shows a configuration example of a system that realizes a conventional augmented reality space. FIG. 2 shows a processing flow of a system for realizing a conventional augmented reality space. The system that realizes the augmented reality space includes an augmented reality space providing server 800-n that provides the augmented reality space (where n = 1,..., N and N are positive integers) and an augmented reality space providing server 800-n. , A utterance terminal 900-m for registering a utterance object (where m = 1,..., M and M are positive integers) and a network 1000. The augmented reality space providing server 800-n includes a space providing communication unit 880-n and a space providing recording unit 890-n. The speech terminal 900-m includes a browsing unit 910-m, a speech unit 920-m, and a speech communication unit 980-m, and may also include a camera 930-m. The space providing recording unit 890-n records a geographical object and a remark object associated with the geographical object.

  First, the browsing process (S910) will be described. When the speaking terminal 900-m includes the camera 930-m, the speaking terminal 900-m acquires a real-time video (S901). Next, the browsing unit 910-m of the speech terminal 900-m acquires GPS information and the like (S902), and selects a geographic object that enters the augmented reality space display screen (S903). The browsing unit 910-m requests data such as a speech object associated with the geographical object selected by the augmented reality space providing server 800-n (S911). The space providing communication unit 880-n of the augmented reality space providing server 800-n extracts the requested data from the space providing recording unit 890-n and provides it to the speech terminal 900-m (S801). Based on the data obtained in step S801, the browsing unit 910-m creates and displays an augmented reality space screen by overlaying the speech object based on the data obtained in step S801 (S912). Some display information (for example, a speaker ID, etc.) for identifying a speaker together with the speaker object. The browsing process (S910) is configured from step S901 to step S912.

  Next, the registration process (S920) will be described. The speech terminal 900-m acquires a speech declaration during browsing the augmented reality space (S921). The utterance terminal 900-m specifies a geographical object to which the utterance object is linked (S922). The utterance unit 920-m acquires a utterance object (S923), transmits information on the geographical object and the utterance object to the augmented reality space providing server 800-n, and requests registration (S924). The augmented reality space providing server 800-n records a speech object in association with the geographical object of the space providing recording unit 890-n (S802). The registration process (S920) is configured from step S921 to step S802.

Tonchi (Tonchidot) Co., Ltd., “Sekai Camera Support Center” [searched on June 29, 2010], Internet <URL: http://support.sekaicamera.com/en>. B. Schoenmakers and P. Tuyls, “Efficient binary conversion for Paillier encrypted values, Advances in Cryptography”, EUROCRYPT 2006, Lecture Notes in Computer Science 4004, pp.522-537, Spring-Verlag, 2006. Ronald Cramer, Ivan Damgard, Jesper Buus Nielsen “Multiparty Computation from Threshold Homomorphic Encryption”, EUROCRYPT 2001, pp. 280-300, 2000.

  Marketing analysis and research in augmented reality space, where services are beginning to spread, is considered to be very effective. However, the investigator's request to investigate what kind of tendency the person made a statement (matching specific conditions) at that location and the speaker's side not wanting to disclose their profile information to unspecified majority These requests cannot be realized at the same time.

  In addition, when managers of stores and the like install an augmented reality space providing server individually and transmit information on the premises, the number of managers who prompt user registration increases as user enclosure. is expected. That is, registration information such as attribute information is managed for each manager (for each augmented reality space providing server), and a list of registrant's messages (in the management area) is accumulated. Useful marketing data in a way that is desirable for each manager, if cross-counting statistics and mining can be performed by name identification while hiding the information managed by each manager (which may also include action history such as purchasing). / Study data is expected to be obtained, but that is not currently done. Further, the speaker does not necessarily register as a user for all managers using the same user identifier.

  The cross-statistic system of the present invention is for an speaker who satisfies a condition such as a statement in a specific geographical object in a situation where a plurality of managers respectively hold user attribute information in an augmented reality space. Name identification, and the attribute information of each user (or management information of each administrator) is concealed, and cross-sectional statistics and mining are realized using information held by multiple administrators. With the goal. Further, even when there is no user identification information, by using encrypted profile information, location information, content of statements, etc., the same user can be estimated, and cross-sectional statistics and mining can be realized.

  The crossing statistics system according to the present invention is characterized in that when there are a plurality of space providing means for providing an augmented reality space in which a geographical object and a speech object associated with the geographical object are arranged, the attribute of the speaker of the speech object Statistical processing is performed across the information managed by the plurality of space providing means while keeping the information secret. The transverse statistical system of the present invention has a space providing recording unit, and includes a plurality of space providing means for providing an augmented reality space, a secret calculation means, and an analysis means. There are a plurality of secret calculation means, and they perform the secret calculation in cooperation. In the case of the above-mentioned secure calculation method 1, the encryption attribute information has one secret calculation recording unit, and in the case of the secret calculation method 2, there are two secret calculation means, and each of the secret calculation recording units has the encryption attribute. It has a piece of information. Usually, one secret calculation means is arranged on the same hardware as the space providing means. One secret calculation means may be arranged in the speech terminal or the analysis terminal. These means may be connected via a network. The analysis unit is normally arranged in the analysis server, but may be arranged in the same hardware as the speech terminal, the analysis terminal, the space providing unit, or the secret calculation unit. The space providing recording unit of each space providing unit records at least the geographical object and the speech object managed by each space providing unit.

  The analysis means includes a calculation acquisition unit, a plaintext calculation unit, and a secret calculation request unit. The calculation acquisition unit receives a request for calculation using information recorded by the plurality of space providing means. The plaintext calculation unit performs calculation without using the encryption attribute information in the calculation. The secret calculation requesting unit requests the secret calculation means for the secret calculation using the encrypted attribute information among the calculations. The analysis means analyzes the speakers as the same person when the speakers are recorded in different space providing means and there is information that identifies the speakers and they match.

  The secret calculation means uses the encrypted attribute information (piece) to perform secret calculation in a coordinated manner without revealing the attribute information.

  In addition, a speech terminal that registers a speech object in the crossing statistical system of the present invention includes at least a browsing unit, a speech unit, an encryption unit, and an attribute management unit. The browsing unit selects a geographic object, acquires data including a speech object linked to the geographic object from a space providing unit that manages the geographic object, and acquires the selected geographic object and the speech object that has been acquired. Is displayed as an augmented reality space. The remark part specifies a geographical object, and transmits a remark object associated with the geographical object and the geographical object to a space providing unit that manages the geographical object. The encryption unit encrypts the attribute information of the speaker of the message object, and obtains encrypted attribute information or a plurality of pieces of encrypted attribute information pieces for recording the encrypted attribute information in a distributed manner. The attribute management unit manages attribute information and encrypted attribute information or a plurality of pieces of encrypted attribute information. Management of plaintext attribute information is not essential.

  The analysis means further includes speakers recorded in different space providing means, and 1) the attribute information of the speakers is equal to or more than a predetermined ratio. 2) (the position is within a predetermined range. A) The number of utterances for geographic objects of the same category is equal to or greater than the predetermined number for either space providing means. 3) The expression of the remarked object recorded in either space providing means is predetermined. If all or some of the criteria are similar, the speakers are analyzed as the same person. “The attribute information matches a predetermined ratio or more” may be, for example, that 90% or more of the information in the attribute information matches. “The position is within a predetermined range” may be determined in advance such that the distance is within the range of the same town name. “Category” is a classification of geographical objects such as restaurants and amusement parks. The “predetermined number of times” may be appropriately determined such as 10 times. “The expression is similar to a predetermined criterion or more” means that the wording is similar, and it may be a case where a difference in some feature amount obtained by analyzing the sentence is equal to or less than a threshold value.

  According to the crossing statistical system of the present invention, it is possible to perform name identification for a speaker for a geographical object managed separately by a plurality of space providing means. Therefore, cross-sectional statistics and mining can be realized by using the information possessed by a plurality of managers while concealing the attribute information of each talker (or the management information of each manager). Further, when name identification is performed using encrypted attribute information, position information, content of a statement, etc., it is possible to estimate whether the person is the same even when there is no information specifying the speaker.

The figure which shows the structural example of the system which implement | achieves the conventional augmented reality space. The figure which shows the processing flow of the system which implement | achieves the conventional augmented reality space. The figure which shows the structural example of the crossing statistics system of this invention, and its periphery. The figure which showed the example of the specific arrangement | positioning of the space provision means 100-n, the secret calculation means 200-p, and the analysis means 300. FIG. The figure which shows the 1st processing flow of a browsing process and a registration process. The figure which shows the 2nd processing flow of a browsing process and a registration process. The figure which shows the 3rd processing flow of a browsing process and a registration process. The figure which shows the processing flow of an analysis process. The figure which shows the processing flow of a name collation function.

  Hereinafter, embodiments of the present invention will be described in detail. In addition, the same number is attached | subjected to the structure part which has the same function, and duplication description is abbreviate | omitted.

  The crossing statistics system according to the present invention is characterized in that when there are a plurality of space providing means for providing an augmented reality space in which a geographical object and a speech object associated with the geographical object are arranged, the attribute of the speaker of the speech object Statistical processing is performed across the information managed by the plurality of space providing means while keeping the information secret. FIG. 3 shows an example of the configuration of the crossing statistical system of the present invention and its periphery. The crossing statistical system of the present invention has a space providing recording unit 190, and N space providing means 100-n that provide an augmented reality space (where N is an integer of 1 or more and n is an integer of 1 to N) ), A secret calculation means 200-p (where P is an integer greater than or equal to 1 and p is an integer greater than or equal to 1 and less than or equal to P), and an analysis means 300. These means may be connected via the network 1000. Further, the network 1000 includes M speech terminals 500-m (where M is an integer greater than or equal to 1 and m is an integer greater than or equal to 1 and less than or equal to M) for registering speech objects in the space providing unit 100-n, statistics, An analysis terminal 400 that requests the analysis means 300 to perform operations for processing and mining is connected.

  The space providing unit 100-n includes a space providing communication unit 180-n and a space providing recording unit 190-n. The secret calculation unit 200-p includes a secret calculation unit 210-p, a secret calculation communication unit 280-p, and a secret calculation recording unit 290-p. The analysis unit 300 includes an operation acquisition unit 310, a plaintext calculation unit 340, a secret calculation request unit 350, and an analysis communication unit 380. The speech terminal 500-m includes a browsing unit 910-m, a speech unit 520-m, an attribute management unit 540-m, an encryption unit 550-m, a speech communication unit 580-m, and also includes a camera 930-m. Also good. The space providing communication unit 180-n, the secret calculation communication unit 280, the analysis communication unit 380, and the message communication unit 580-m exchange necessary data via the network 1000.

[Geographic object]
A geographical object refers to an object that can be specified geographically as follows. However, it is not necessarily limited to these, and includes geographically specified items as well.
(1) Point expressed by latitude and longitude (2) Range of constant distance from (1) (3) Point expressed by (1) + height (4) Range of constant distance from (3) 5) Static facilities such as buildings and open spaces (6) A part of (5) (ex. XX building 1st floor, XX department store menswear department, △△ amusement park ferris wheel, etc.)
(7) Municipally defined areas (eg, Japan, Kanagawa Prefecture, Yokohama City, Shinjuku-ku, Tokyo, Yokohama City ○○ Town, Yokohama City ○○ Town ×× Chome, Yokohama City ○○ Town × × Chome △△ address)
(8) Dynamic objects whose current position is known (humans, animals, vehicles, luggage, etc. with GPS and RFID tags)

[Remark Object]
The speech object refers to the following target, for example.
(1) Text (2) Image (3) Audio (4) Video (5) Program (For example, there is a voting program, and there are choices such as “I am glad I came here” or “I was not good” with respect to the geographical object. You can choose the option that reflects your opinion from the list and display the results of each person's vote.)
(6) Other electronic data (electronic file) (for example, PDF file)

[Distribution of space providing means, confidential calculation means, and analysis means]
FIG. 4 is a diagram illustrating a specific arrangement example of the space providing unit 100-n, the secret calculation unit 200-p, and the analysis unit 300. FIG. 4A shows the space providing means 100-1,..., N and the secret calculation means 200-1,..., N in the augmented reality space providing servers 101-1,. In this example, N + 1 is placed in the secure calculation server 201 and the analysis means 300 is placed in the analysis server 301. In this case, P = N + 1. In this arrangement, the image of FIG. 3 is directly associated with specific hardware. However, in the case of the secret calculation method 1, the secret calculation server 201 is not necessary. In the case of the secret calculation method 2, only one augmented reality space providing server has a secret calculation means. However, the space providing means 100-1,..., N, the secret calculation means 200-p, and the analysis means 300 are not necessarily arranged on separate hardware. For example, as shown in FIG. 4B, space providing means 100-1,..., N and secret calculation means 200-1,..., N and analyzing means 300 are arranged in one augmented reality space providing server 102. Alternatively, as shown in FIG. 4C, the secret calculation means 200-N + 1 and the analysis means 300 may be arranged in one secret calculation server 202. Such an arrangement may be appropriately designed based on resources such as a hardware CPU and memory, and a necessary calculation amount and storage capacity. Furthermore, as shown in FIG. 4D, any one of the secret calculation means 200-p can be arranged inside the speech terminal 500-m. If arranged in this way, statistical processing and mining cannot be performed without the consent of the speaker (when the speaking terminal 500-m does not cooperate).

[Browsing process, registration process]
5 to 7 show the processing flow of the browsing process and the registration process. The browsing process (S910) is the same as the conventional one. However, the data related to the geographical object acquired in the browsing process (S910) is in a wider range than the range (minimum necessary data) necessary for the first display in preparation for the processing after the browsing process (S910). A geographic object may be selected.

Next, the registration process (S520) will be described. The speech terminal 500-m acquires a speech declaration while browsing the augmented reality space (S921). The utterance terminal 500-m specifies a geographical object to which the utterance object is linked (S922). For example, the geographical object may be specified as follows.
(1) The current position is determined from the GPS information (and the angle sensor) of the speaking terminal 500-m, and the geographical object is specified. In addition, when there are a plurality of candidates, a speaker list may be selected by presenting a candidate list such as a name representing a geographical object or a figure.
(2) A geographic object is specified by clicking an augmented reality screen in the speech terminal 500-m. In addition, when there are a plurality of candidates, a speaker list may be selected by presenting a candidate list such as a name representing a geographical object or a figure.

(3) A list of geographical objects on the screen is shown, and a geographical object is specified by selecting a speaker from the list. For example, a name representing a geographical object may be displayed on the list.
(4) A list of geographical objects that match the search conditions set by the speaker is shown, and the geographical object is specified by the speaker selecting from the list. For example, a name representing a geographical object may be displayed on the list.
The speech unit 520-m of the speech terminal 500-m acquires a speech object (S923).

  From the next step, the process differs depending on the method of confidential calculation. First, the case of the processing flow of the secret calculation method 2 shown in FIG. 5 will be described. The configuration includes a speech terminal 500-m, an augmented reality space providing server 101-n having a space providing unit 100-n and a secret calculation unit 200-n, and a secret calculation unit 200-N + 1 (same configuration as FIG. 4A). To do. In the case of this configuration (in the case where each augmented reality space providing server includes one secret calculation means and one common independent secret calculation means), P = N + 1. The number may be higher. For example, two or more common independent secret calculation means may be used. The encryption unit 550-m encrypts the attribute information of the speaker, obtains two encrypted attribute information pieces A and B for recording the distributed attribute information in a distributed manner, and the attribute management unit 540-m records the encrypted attribute information pieces A and B. (S551). Note that the attribute information to be encrypted may be changed for each geographical object or each speech object. For example, normal profile information can be used for serious utterances, and profile information specialized for the hobbies can be used for hobby utterances.

  The speech terminal 500-m is an augmented reality space providing server that manages the geographical object identified in step S922, the speech object acquired in step S923, and the encrypted attribute information piece A created in step S551. 101-n and request registration (S524). The space providing unit 100-n of the augmented reality space providing server 101-n records a speech object in association with a geographical object in the space providing recording unit 190-n, and performs the secret calculation of the augmented reality space providing server 101-n. The means 200-n records the encrypted attribute information piece A in the secret calculation recording unit 290-n in association with the message object (S102). The utterance terminal 500-m transmits the encrypted attribute information piece B created in step S551 to the secret calculation means 200-N + 1 together with information that can specify the utterance object (for example, the ID of the utterance object) (S525). The secret calculation means 200-N + 1 records the encrypted attribute information piece B in the secret calculation recording unit 290-N + 1 in association with the message object (S202). For example, it is recorded together with the ID of the statement object. The registration process (S520) is configured from step S921 to step S202.

  If there are a plurality of augmented reality space providing servers, a plurality of space providing means may exist. A plurality of secret calculation means may exist. When there are a plurality of means as described above, the space providing means and the secret calculation means may be properly used by performing some grouping. For example, if the theme park operator allocates various rides as geographical objects and operates the space providing means only in the theme park, the registration in the theme park is registered with the theme park operator's space providing means. Can be considered. And registration outside the theme park can be done for global space provision means. In this way, the rule of which space providing means to select for a geographical object may be designed as appropriate. In addition to the theme park example, a rule may be considered in which priority is negotiated and determined for each operator of the space providing means, and the server with the highest priority is selected as the registration destination server.

  Furthermore, the secret calculation means 200-p may link the encrypted attribute information (piece) recorded by the speaking terminal 500-m, instead of recording the encrypted attribute information (piece). If linked, if one speaker registers a plurality of speaking objects, the encrypted attribute information (piece) can be shared by the plurality of speaking objects. Therefore, it is convenient when it is necessary to update the encrypted attribute information (piece). To simplify the description, in this specification, when simply “recording” is described, in addition to recording the target data itself, it is linked so that the target data can be obtained. Shall also be included.

  Next, the process flow of the secret calculation method 1 shown in FIG. 6 will be described. Steps up to step S923 are the same as those in FIG. In the case of FIG. 6, the encryption unit obtains encrypted attribute information by encrypting the attribute information of the speaker of the message object (S551 '). Note that the attribute information to be encrypted may be changed for each geographical object or each speech object. The utterance terminal 500-m includes the augmented reality space providing server 101 that manages the geographic object identified in step S922, the utterance object acquired in step S923, and the encrypted attribute information created in step S551 ′. -N to request registration (S524 '). The space providing unit 100-n of the augmented reality space providing server 101-n records a speech object in association with a geographical object in the space providing recording unit 190-n, and performs the secret calculation of the augmented reality space providing server 101-n. The means 200-n records the encrypted attribute information in the secret calculation recording unit 290-n in association with the message object (S102 ′). The registration process (S520 ') is configured from step S921 to step S102'.

Next, the case of the processing flow shown in FIG. 7 will be described. The process up to step S551 ′ is the same as in FIG. In the case of FIG. 7, the utterance terminal 500-m transmits the geographic object identified in step S922 and the utterance object acquired in step S923 to the augmented reality space providing server 101-n that manages the identified geographic object. Registration is requested (S524 "). The space providing unit 100-n of the augmented reality space providing server 101-n records a speech object in association with a geographical object in the space providing recording unit 190-n (S102"). ). The utterance terminal 500-m transmits the encryption attribute information created in step S551 ′ to the secret calculation means 200-N + 1 together with information (for example, ID of the utterance object) that can specify the utterance object (S525 ″). The calculation means 200-N + 1 records the encrypted attribute information in the secret calculation recording unit 290-N + 1 in association with the message object (S202 ″). The registration process (S520 ″) is configured from Step S921 to Step S202 ″.
Further, as can be said with respect to both FIG. 6 and FIG. 7, the encryption information may be held at the speaking terminal and used at the time of the secret calculation.

[Analysis processing]
FIG. 8 shows a process flow of the analysis process (S300). An arithmetic expression for statistical processing and mining is input to the analysis terminal 400 (S410). If the input arithmetic expression can be executed with information of only one space providing unit 100-n, the subject of the present invention does not exist. Therefore, in the following description, it is assumed that the information separately recorded in the plurality of space providing units 100-1,... The analysis terminal 400 requests the analysis means 300 for calculation (S420). The calculation acquisition unit 310 of the analysis unit 300 acquires the requested calculation (requested calculation) (S311). In statistical processing, it is often necessary to distinguish between speech of the same person or speech of different persons. Therefore, the analysis unit 300 performs the name identification process S320 and then performs the analysis process S340.

  In the name identification process S320, the plaintext calculation unit 340 of the analysis unit 300 performs a calculation using unencrypted information such as the contents of the speech object and the geographical object (S341). For example, since the encryption attribute information is not required for the calculation for obtaining the similarity of the expression between the speech objects, the plaintext calculation unit 340 may perform the calculation. The secret calculation request unit 350 transmits the contents of the secret calculation to the secret calculation means 200-1,..., N + 1, and requests the secret calculation (S351). The secret calculation units 210-1,..., N + 1 of the respective secret calculation means 200-1,..., N + 1 cooperate to perform the requested secret calculation without revealing the attribute information, and send the result to the analysis means 300 ( S111). Note that the plaintext calculation in step S341 and the secret calculation in step S111 may be reversed in order as necessary, or may be performed in parallel, or divided as “plaintext calculation → secret calculation → plaintext calculation”. You may do it.

  Next, also in the analysis process S340, the plaintext calculation unit 340 of the analysis unit 300 performs a calculation using unencrypted information (S342). For example, since the encryption attribute information is not required in the calculation for obtaining the number of statement objects associated with a certain geographical object, the plaintext calculation unit 340 may perform the calculation. The secret calculation request unit 350 transmits the contents of the secret calculation to the secret calculation means 200-1,..., N + 1, and requests the secret calculation (S352). The secret calculation units 210-1,..., N + 1 of the respective secret calculation means 200-1,..., N + 1 cooperate to perform the requested secret calculation without revealing the attribute information, and send the result to the analysis means 300 ( S112). The analysis unit 300 obtains a result for the requested calculation from the result of step S342 and the result of step S112, and transmits the result to the analysis terminal 400 (S312). The analysis terminal 400 outputs the calculation result (S430). Note that the plaintext calculation in step S342 and the secret calculation in step S112 may be reversed in order as necessary, or may be performed in parallel, or divided as “plaintext calculation → secret calculation → plaintext calculation”. You may do it.

[Name identification function]
FIG. 9 is a diagram illustrating a processing flow of the name identification function. The analysis unit 300 confirms whether or not there is information specifying the speaker for the speaker recorded in the different space providing unit (S321). If it exists (if S321 is Yes), if the information specifying the speaker matches, the speaker is determined to be the same person (S322). In other words, the analysis unit 300 performs statistical processing with the speakers as the same person if the speakers are recorded in different space providing units and there is information that identifies the speakers and they match. . Further, when there is no information for identifying a speaker (when S321 is No), the analysis unit 300 narrows down the speakers to only those whose attribute information between the speakers matches a predetermined ratio or more ( S323). The analysis unit 300 narrows the speakers to only those whose number of utterances with respect to a geographical object of the same category (within a position within a predetermined range) is greater than or equal to a predetermined number of times for both space providing units ( S324). The analysis unit 300 narrows down the speakers to the speech objects recorded in either of the space providing units only to those whose expressions are similar to a predetermined standard or more (S325). Then, the narrowed speakers may be analyzed as the same person. The execution order of S323 to S325 may be arbitrary and may be skipped. When the information necessary for name identification is plain text, the above-described processing may be performed in the plain text calculation processing (step S341). If the information necessary for name identification is encrypted, the above-described processing may be performed in the confidential calculation processing (step S351 and step S111).

  “The attribute information matches a predetermined ratio or more” may be, for example, that 90% or more of the information in the attribute information matches. Alternatively, attributes having the same value may be a certain number (constant ratio) or more, or individual attributes having a certain number (constant ratio) or more may be approximate. “Individual attributes are approximate” means that in the case of an attribute having a number as a value, the difference between the attribute values may be a certain value or less. In the case of an attribute having a character string as a value, the number of matching character strings may be equal to or greater than a certain value (and the number of non-matching character strings is equal to or less than a certain value). “The position is within a predetermined range” may be determined in advance such that the distance is within the range of the same town name. “Category” is a classification of geographical objects such as restaurants and amusement parks. The “predetermined number of times” may be appropriately determined such as 10 times. "The expression is similar to a predetermined criterion or more" means that the wording or moustache is similar, and if the difference in some feature amount obtained by analyzing the sentence is below the threshold value Good. Attributes to be compared may be limited to only static attributes. Further, the location information and action history of the speaker may be narrowed down. “Position information” is information related to positions such as the current position, movement history, routine route, activity range, etc., and if they are similar, it can be the same person or different person with higher accuracy if it is added to the conditions to treat as the same person. Can be estimated. A “routine route” is a frequently used route such as a commuting route or a commuting route, and is expressed in the form of an ordered list of geographical objects indicating a departure point, a relay point, and a destination point. There may be a loop. As an example of the detection method, a routine route is created from a frequently used route using train getting on / off information. Voluntary offers are acceptable. The “activity range” is a geographical area composed of polygons that are normally active, and is expressed in the form of an ordered list of each geographical object that is a vertex constituting the outer periphery. As an example of the detection method, in order to exclude places that have just been visited for sightseeing, a polygon formed by connecting the utterance positions that have been uttered twice or more is set as the activity range. Voluntary offers are acceptable.

  Specific examples of refinement include: current location is within a certain distance, searcher's current location and candidate's current location are in the same category, searcher's current location is within the candidate's routine route or range of activity, The movement / behavior history cut out during the period of the searcher and the candidate overlaps more than a certain amount, and the routine route or activity range of the searcher and the candidate overlaps more than a certain amount.

[Confidential calculation]
As a specific example of the confidential calculation in step S110, for example, there is a method of combining Non-Patent Document 2 and Non-Patent Document 3. Non-Patent Document 2 is a method for generating a ciphertext in which a numerical value encrypted by Paillier encryption is encrypted for each digit in which the numerical value is expressed in binary notation without revealing the numerical value. Non-Patent Document 3 is a method that allows an arbitrary logical operation of a discussion numerical value to be executed without revealing the numerical value by encrypting a numerical value expressed in binary notation for each digit. If the method disclosed in Non-Patent Document 2 is used, it is possible to execute a secret calculation using a numerical ciphertext as an input. As a result, the device (the speaking terminal 500-m) that provides the input of the confidential calculation only needs to calculate and store the ciphertext of the numerical value, and calculates and stores the ciphertext for each digit in which the numerical value is expressed in binary. There is no need. Therefore, the method disclosed in Non-Patent Document 2 is a particularly effective method when the speech terminal 500-m is scarce in computing resources and memory resources.

  The process of step S110 depends on how the encrypted data is recorded during the registration process (S520, S520 ′, S520 ″), but using the encrypted attribute information (piece). All the secret calculation units perform the calculation in cooperation with each other.

[Other variations]
The data to be encrypted may be encrypted not only in attribute information but also in the contents of geographic objects and speech objects. In this case, an encrypted data storage unit 191 (not shown) for storing is provided, and the analysis unit 300 performs a confidential circuit calculation on the data, thereby improving the confidentiality of the space providing recording unit 190. Also good. As a method for specifying the position, not only the position specification by GPS but also the position specification by a sensor or a tag can be considered. Of course, there may be a cross-sectional statistical process that does not require name identification.

[Application example of the present invention]
The crossing statistical system of the present invention can be used in the following applications.
(1) The managers of sightseeing spots A to E grasp the attribute information such as the age, sex, and hobbies of members (in the case of hobbies, description formulation (= absorption of notation shaking, listing, etc.)) Under the circumstances, 1 case / 2 cases / 3 cases / 5 cases are visited and the tendency of the person who made some remarks is investigated. (And take measures to improve the ability to attract customers.)
(2) In the situation where the managers of shopping malls A to F grasp the same attribute information as in the first embodiment, 1 case / 2 cases / 3 cases / 5 cases / 6 cases visited and complaints Explore the tendency of those who have made the remarks (determined by whether or not the remarks contained in the pre-listed complaints are included). (And take measures to avoid complaining.)

(3) In the situation where the manager of the amusement parks A to G grasps the routine route or the moving route in addition to the same attribute information as in the first embodiment, what kind of attribute is the person who Investigate the trend of going around amusement facilities in order.
(4) In the situation where the manager of the restaurants A to H knows which genre menu is eaten in addition to the same attribute information as in the first embodiment, what genre menu is and how Explore trends in how people are chosen.
(5) In the situation where the manager of the bookstores A to I grasps the purchase history of each person in addition to the attribute information similar to that in the first embodiment, on the whole, what kind of person purchases on average Explore trends in what you are doing.

[Program, recording medium]
The various processes described above are not only executed in time series according to the description, but may also be executed in parallel or individually as required by the processing capability of the apparatus that executes the processes. Needless to say, other modifications are possible without departing from the spirit of the present invention.

  Further, when the above-described configuration is realized by a computer, processing contents of functions that each device should have are described by a program. The processing functions are realized on the computer by executing the program on the computer.

  The program describing the processing contents can be recorded on a computer-readable recording medium. As the computer-readable recording medium, for example, any recording medium such as a magnetic recording device, an optical disk, a magneto-optical recording medium, and a semiconductor memory may be used.

  The program is distributed by selling, transferring, or lending a portable recording medium such as a DVD or CD-ROM in which the program is recorded. Furthermore, the program may be distributed by storing the program in a storage device of the server computer and transferring the program from the server computer to another computer via a network.

  A computer that executes such a program first stores, for example, a program recorded on a portable recording medium or a program transferred from a server computer in its own storage device. When executing the process, the computer reads a program stored in its own recording medium and executes a process according to the read program. As another execution form of the program, the computer may directly read the program from a portable recording medium and execute processing according to the program, and the program is transferred from the server computer to the computer. Each time, the processing according to the received program may be executed sequentially. Also, the program is not transferred from the server computer to the computer, and the above-described processing is executed by a so-called ASP (Application Service Provider) type service that realizes the processing function only by the execution instruction and result acquisition. It is good. Note that the program in this embodiment includes information that is used for processing by an electronic computer and that conforms to the program (data that is not a direct command to the computer but has a property that defines the processing of the computer).

  In this embodiment, the present apparatus is configured by executing a predetermined program on a computer. However, at least a part of these processing contents may be realized by hardware.

100 Space providing means 101, 102, 800 Augmented reality space providing server 180, 880 Space providing communication unit 190, 890 Space providing recording unit 191 Encrypted data storage unit 200 Secret calculation means 201, 202 Secret calculation server 210 Secret calculation unit 280 Secret Calculation communication unit 290 Secret calculation recording unit 300 Analysis means 301 Analysis server 310 Operation acquisition unit 340 Plain text calculation unit 350 Secret calculation request unit 380 Analysis communication unit 400 Analysis terminal 500, 900 Statement terminal 520, 920 Statement unit 540 Attribute management unit 550 Encryption Part 580, 980 message communication part 910 browsing part 930 camera 1000 network

Claims (9)

When there are a plurality of space providing means for providing an augmented reality space in which a geographical object and a speech object associated with the geographical object are arranged, a plurality of the attribute information of the speaker of the speech object is kept secret. A cross-sectional statistical system that performs cross-sectional statistical processing on information managed by each of the space providing means,
A plurality of space providing means for providing the augmented reality space; a plurality of secret calculation means; and an analysis means.
The space providing recording unit of each space providing unit records at least the geographical object and the speech object managed by each space providing unit,
The plurality of secret calculation means records encrypted attribute information pieces obtained by concealing and dividing the attribute information of the speaker of the speech object managed by the space providing means, and the analysis means includes:
A calculation acquisition unit that receives a request for calculation using information recorded by the plurality of space providing means;
A plaintext calculation unit for performing calculation without using the encrypted attribute information piece of the operation;
A secret calculation requesting unit that requests a secret calculation using the encrypted attribute information piece of the calculation to all the secret calculation means;
The analysis means includes
If the speakers are recorded in different space providing means and there is information that identifies the speakers, and they match, analyze the speakers as the same person,
All the secret calculation means
Using the encrypted attribute information piece, the secret calculation is performed in cooperation without revealing the contents of the attribute information. When there are a plurality of space providing means for providing an augmented reality space in which a geographical object and a speech object associated with the geographical object are arranged, a plurality of the attribute information of the speaker of the speech object is kept secret. A cross-sectional statistical system that performs cross-sectional statistical processing on information managed by each of the space providing means,
A plurality of space providing means for providing the augmented reality space; a plurality of secret calculation means; and an analysis means.
The space providing recording unit of each space providing unit records at least the geographical object and the speech object managed by each space providing unit,
One for each of the corresponding space providing means among the secret calculation means records the encrypted attribute information obtained by encrypting the attribute information of the speaker of the speech object managed by the space providing means,
The analysis means includes
A calculation acquisition unit that receives a request for calculation using information recorded by the plurality of space providing means;
A plaintext calculation unit that performs calculation without using the encryption attribute information among the operations;
A secret calculation requesting unit that requests a secret calculation using the encrypted attribute information of all the calculations to the secret calculation means;
The analysis means includes
If the speakers are recorded in different space providing means and there is information that identifies the speakers, and they match, analyze the speakers as the same person,
All the secret calculation means are
Using the encrypted attribute information, the secret calculation is performed in cooperation without revealing the content of the attribute information. The crossing statistical system according to claim 1 or 2,
The analysis means includes
The speakers recorded in different space providing means,
If the attribute information between the speakers matches a predetermined ratio or more,
If the number of utterances for geographic objects of the same category is greater than or equal to the predetermined number of times for either space provider,
In the case where the speech object recorded in either space providing means is similar in expression to a predetermined criterion or more,
A cross-statistical system characterized by analyzing speakers who satisfy any one, two or all as the same person. A crossing statistics system according to claim 3,
The cross-sectional statistics system, wherein the analysis means determines whether or not the speakers are the same person using position information and / or action history. A speech terminal for registering the speech object in the crossing statistical system according to claim 1,
A geographical object is selected, data including a speech object associated with the geographical object is acquired from the space providing means for managing the geographical object, and the selected geographical object and the acquired speech object are augmented. A viewing section to display as a space;
A remark part for identifying a geographical object and transmitting a remark object associated with the geographical object to the space providing means for managing the geographical object;
An encryption unit for obtaining a plurality of encrypted attribute information pieces for concealing and dividing the attribute information of a speaker of the speech object;
A speech terminal comprising the attribute information and an attribute management unit that manages the plurality of pieces of encrypted attribute information. A speech terminal for registering the speech object in the statistics / mining system according to claim 2,
A geographical object is selected, data including a speech object associated with the geographical object is acquired from the space providing means for managing the geographical object, and the selected geographical object and the acquired speech object are augmented. A viewing section to display as a space;
A remark part for identifying a geographical object and transmitting a remark object associated with the geographical object to the space providing means for managing the geographical object;
An encryption unit that encrypts attribute information of a speaker of the speech object to obtain encrypted attribute information; an attribute management unit that manages the attribute information and the encrypted attribute information;
A speech terminal comprising: When there are a plurality of space providing means for providing an augmented reality space in which a geographical object and a speech object associated with the geographical object are arranged, a plurality of the attribute information of the speaker of the speech object is kept secret. Statistical processing across the information respectively managed by the space providing means,
A cross-statistic method that has a space providing recording unit and uses a plurality of space providing means for providing the augmented reality space, a plurality of secret calculation means, and an analysis means,
Augmented reality recording step in which the space providing recording unit of each space providing unit records at least the geographical object and the speech object managed by each space providing unit;
A plurality of the secret calculation means, a secret information recording step for recording an encrypted attribute information piece obtained by concealing and dividing the attribute information of the speaker of the speech object managed by the space providing means;
A calculation acquisition step in which the analysis means receives a request for calculation using information recorded by the plurality of space providing means;
A plaintext calculation step in which the analysis means performs a calculation without using the encrypted attribute information piece in the calculation,
The analysis means requests a secret calculation using the encrypted attribute information piece of the calculation to a secret calculation request step for requesting all the space providing means and the secret calculation means,
The space providing means and the secret calculation means have a secret calculation step of performing the secret calculation in cooperation without revealing the attribute information using the encrypted attribute information,
The analysis means includes
A cross-statistical method characterized by analyzing speakers as the same person when there is information that identifies the speakers among the speakers recorded in different space providing means. When there are a plurality of space providing means for providing an augmented reality space in which a geographical object and a speech object associated with the geographical object are arranged, a plurality of the attribute information of the speaker of the speech object is kept secret. Statistical processing across the information respectively managed by the space providing means,
A cross-statistic method that has a space providing recording unit and uses a plurality of space providing means for providing the augmented reality space, a plurality of secret calculation means, and an analysis means,
Augmented reality recording step in which the space providing recording unit of each space providing unit records at least the geographical object and the speech object managed by each space providing unit;
A secret information recording step for recording encrypted attribute information obtained by encrypting the attribute information of a speaker of a speech object managed by the space providing unit, for each corresponding space providing unit among the secret calculation units,
A calculation acquisition step in which the analysis means receives a request for calculation using information recorded by the plurality of space providing means;
A plaintext calculation step in which the analysis means performs a calculation without using the encryption attribute information in the calculation;
The analysis means requests a secret calculation using the encrypted attribute information in the calculation to a secret calculation request step for requesting all the secret calculation means;
All the secret calculation means have a secret calculation step for performing the secret calculation in cooperation without revealing the contents of the attribute information using the encrypted attribute information,
The analysis means includes
A cross-statistical method characterized by analyzing speakers as the same person when there is information that identifies the speakers among the speakers recorded in different space providing means. The crossing statistics method according to claim 7 or 8,
The analysis means includes
The speakers recorded in different space providing means,
If the attribute information between the speakers matches a predetermined ratio or more,
If the number of utterances for geographic objects of the same category is greater than or equal to the predetermined number of times for either space provider,
In the case where the speech object recorded in either space providing means is similar in expression to a predetermined criterion or more,
A cross-statistical method characterized by analyzing speakers who satisfy any one, two, or all as the same person.

Images