JP2009116820A - Image analysis apparatus, and information exchange system having this image analysis apparatus - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide an image analysis apparatus that analyzes photographic styles of image data registered by photographers, and an information exchange system with the image analysis apparatus that enables photographers having the same photographic style to exchange information. <P>SOLUTION: The image analysis apparatus 10 of the information exchange system 100 includes a storage part 12 for storing images subjected to analysis, and an analysis part 13 for categorizing each subject image stored in the storage part 12 and calculating a pertinent category that is a category indicating a photographic style of the subject images stored in the storage part from the categories of the subject images stored in the storage part. <P>COPYRIGHT: (C)2009,JPO&INPIT

Description

  The present invention relates to an image analysis apparatus and an information exchange system having the image analysis apparatus.

2. Description of the Related Art Conventionally, a system has been proposed in which images taken by a photographer for various subjects are transmitted to an image server via a network such as the Internet, and stored and disclosed. In addition, a system for evaluating whether or not these images are taken under optimum conditions (for example, aperture, shutter speed, ISO sensitivity, strobe light amount, camera shake, composition, etc.) has been proposed (for example, patents). Reference 1).
JP 2003-87442 A

  However, the conventional system only evaluates the quality of a plurality of image data photographed by the photographer, and there is no system that statistically analyzes the photographer's photographing tendency and shows it. This is the subject that the photographer prefers, for example, when a certain photographer has photographed many people, such as women, men, and children, or has photographed many group photos, landscapes, railways, etc. It shows the taste and field of expertise. Therefore, in the past, it was difficult for photographers to grasp their own shooting trends, and further information exchange and information collection with other photographers with common tastes could not be actively performed, and the photographer's willingness to photograph did not improve much There was a problem.

  The present invention has been made in view of such a problem, and an image analysis apparatus that analyzes a shooting tendency of image data registered by a photographer, and a photographer having the same shooting tendency having the image analysis apparatus It is an object to provide an information exchange system that enables information exchange.

  In order to solve the above problems, an image analysis apparatus according to the present invention determines a category to which the analysis target image belongs for each storage target image stored in the storage unit and a storage unit that stores the analysis target image. The analysis unit calculates an affiliation category that is a category representing the photographing tendency of the analysis target image stored in the storage unit from the categories to which the analysis target image stored in the storage unit belongs.

  In such an image analysis apparatus according to the present invention, the analysis unit has sample information corresponding to each of a plurality of categories, and uses pattern recognition or object recognition by a neural network from the analysis target image and sample information, It is preferable to be configured to determine a category to which each of the analysis target images belongs.

Further, in such an image analysis apparatus according to the present invention, for each analysis target image, the analysis unit calculates the weight W from the photographing date X, the current date Y, and the coefficients a and b of the analysis target image by the following formula W = A (Y-X) + b
Further, for each category, among the analysis target images stored in the storage unit, the degree of preference of the category is calculated as the sum of the weights W of the analysis target images belonging to the category, and the category belongs from the degree of preference. Preferably configured to determine a category.

  At this time, it is preferable that the category having the highest degree of preference among the categories to which the analysis target images stored in the storage unit belong is set as the belonging category. Alternatively, it is preferable that the categories to which the analysis target images stored in the storage unit belong are rearranged in descending order based on the preference level, and a predetermined number of categories from the category with the highest preference level are set as the belonging category.

  Further, the image analysis apparatus according to the present invention provides an analysis result presentation that presents the photographing tendency of the analysis target image stored in the storage unit by classifying the number of analysis target images stored in the storage unit for each category. It is preferable to have a part. Alternatively, the storage unit is configured to store the analysis target image together with the shooting date when the analysis target image is captured, and classifies the analysis target image stored in the storage unit for each shooting date, and for each category with respect to the classification It is preferable to have an analysis result presentation unit that presents the photographing tendency of the analysis target image stored in the storage unit by calculating the number of images in each category based on the time series.

  Furthermore, the image analysis apparatus according to the present invention is configured such that the storage unit stores the analysis target image together with the identification information of the registrant who registered the analysis target image, and the analysis unit selects the belonging category for each identification information. It is preferably configured to calculate.

  One of the above-described image analysis devices connected to a network, and an information exchange device connected to this network and having information exchanged by a user having identification information, the information exchange device, An information exchange system having a group information storage unit for grouping identification information having the same belonging category among belonging categories.

  Such an information exchange system preferably has a bulletin board manager for each group where users having identification information belonging to the group exchange messages.

  When the image analysis apparatus according to the present invention and the information exchange system including the image analysis apparatus are configured as described above, the photographer can analyze the image data based on the category to which the image data belongs, Thus, it is possible to know the photographing tendency of a plurality of image data that have been photographed, and to improve the willingness of the photographer to photograph.

  Hereinafter, preferred embodiments of the present invention will be described with reference to the drawings. First, the configuration of the image analysis apparatus 10 according to the present embodiment will be described with reference to FIGS. 1 to 6, but the present invention is not limited to this embodiment. As shown in FIG. 1, the image analysis device 10 is connected to a network 30 such as the Internet, and receives image data from various terminal devices 20 that are also connected to the network 30 to perform management and analysis. It is configured. Examples of the terminal device 20 include a PDA (Personal Data Assistant Terminal) 21, a personal computer 22, a mobile phone 23, a digital camera 24 having a network connection function, and the like.

  As shown in FIG. 2, the image analysis apparatus 10 is connected to the network 30 via the input / output port 11, and stores the image data (analysis target image) transmitted from the user (photographer) and the like. The control unit 13 that controls the image analysis apparatus 10, the analysis unit 14 that determines the category to which the analysis target image belongs, and statistically analyzes the analysis target image based on the determined category, and the analysis unit 14 And an analysis result presentation unit 15 for presenting the analysis result of For example, a general web browser is used for browsing image information and various displays by the analysis result presentation unit 15 in the terminal device 20. The storage unit 12 can also store the analysis result by the analysis unit 14. The input / output port 11, the storage unit 12, the control unit 13, the analysis unit 14, and the analysis result presentation unit 15 are connected via an internal bus 16.

  The storage unit 12 includes a user management information storage unit 60 for performing user authentication and management, an image information storage unit 70 for storing image data, and management information for image data stored in the image information storage unit 70. The image management information storage unit 73 stores the statistical information storage unit 80 that stores the analysis result of the analysis unit 14 as statistical information, and the group information storage unit 90 stores the group information classified by the analysis unit 14. ing.

  The user management information storage unit 60 has the data structure shown in FIG. 3 and includes a user ID area 61, a user password area 62, a category presence / absence information area 63, and a belonging category area 64. Here, the user ID area 61 is a part in which a user ID that is identification information for specifying a user is stored, and a user password corresponding to the user ID is stored in the user password area 62. These are used for user authentication. The category presence / absence information area 62 is an area for storing whether or not there is a main category to which the user belongs as a result of analyzing the image data registered by the user (for example, “1” is assumed to be a belonging category). This information is managed with “0” belonging to no category). When there is a belonging category, the name of the main category to which the user belongs (or identification information for identifying the category) is stored in the belonging category area 64. As shown in FIG. 3, the affiliation category area 64 can be configured to manage a plurality of categories.

  The image information storage unit 70 has a data structure shown in FIG. 4A, and includes an attribute information area 71 and an image data area 72. Here, the attribute information area 71 is a part in which attribute information of the image data such as tag information of the image data stored in the image data area 72 is stored, and the image information area 72 is the image data itself (for example, The digital image itself taken by the digital camera 24) is stored. The image information storage unit 70 is configured to manage, for example, image data including tag information as one file. In this case, the data structure including the tag information described above is standardized by JEIDA. Exif (Exchangeable image file format) is known.

  The image management information storage unit 73 stores management information for managing the image data stored in the image information storage unit 70. The image management information storage unit 73 has a data structure shown in FIG. 4B, and includes a user ID area 74, an image ID area 75, an image path area 76, a shooting date / time area 77, a category area 78, and a weight area. 79, corresponding to the number of image data stored in the image data storage unit 70. Here, the user ID area 74 is a part in which a user ID for specifying a user who registered the image data stored in the image data storage unit 70 is stored. The image ID area 75 is a part in which an image ID, which is identification information for specifying image data, is stored. The image path area 76 is a location where the image data is stored in the image information storage unit 70. This is the part where the path shown is stored. The shooting date / time area 77 is a portion in which the date / time when the image data was shot or the date / time uploaded to the image information storage unit 70 of the storage unit 12 is stored. As the shooting date and time, for example, the shooting date and time of Exif information stored in the attribute information area 71 of the image information storage unit 70 can be used. The category area 78 is a portion in which the name of the category to which the image data determined by the analysis unit 14 belongs is stored. The weight area 79 is a portion in which the weight for the image data is stored. This weight is calculated by the analysis unit 14 and is used to calculate the user's preference degree as will be described later. .

  The statistical information storage unit 80 has the data structure shown in FIG. 5 and includes a user ID area 81 and a statistical information area 82. Here, the user ID area 81 is a part where a user ID is stored, and the statistical information area 82 stores a plurality of analysis results statistically analyzed by the analysis unit 14 for image data registered by each user. This is the part.

  The group information storage unit 90 has the data structure shown in FIG. 6 and includes a user ID area 91 and a group information area 92. Here, the user ID area 91 is a part where a user ID is stored. The group information area 92 further includes a group name area 93 and a registration date / time area 94. When there are a plurality of groups for one user ID, the group name area 93 and the registration date / time area 94 are displayed. Multiple pairs are stored. The group name area 93 of the group information area 92 is a part in which a group name in which the user can participate is stored, and is determined by the analysis unit 14. The registration date / time area 94 of the group information area 92 is a part in which the registration date / time for the group is stored. The user selects a group desired to participate from among the groups stored in the group name area 93, and the registration is performed. The registration date and time when this is performed is recorded.

  The image data registered in the image analysis apparatus 10 and the analysis result thereof are combined with the information exchange apparatus 40 as shown in FIG. 1, for example, to share information and exchange information between users having the same taste. It can be performed. Here, as shown in FIG. 2, the information exchange device 40 is connected to the network 30 via the input / output port 41 and performs grouping of users based on the analysis result obtained by the image analysis device 10. The bulletin board management unit 42 that enables the users belonging to the same group to exchange information with each other when the user accesses the bulletin board, and stores the messages by the user on the bulletin board. And the like, and a notification unit 44 that notifies a bulletin board in which the user can participate based on the group to which the user belongs, obtained by the analysis unit 13.

  Now, a description will be given of processing by the image analysis apparatus 10 and a system (hereinafter referred to as “information exchange system 100”) configured by the information exchange apparatus 40 that exchanges information using the analysis result of the image analysis apparatus 10. . In the configuration of the present embodiment illustrated in FIG. 1, the information exchange system 100 is mounted on two servers each including the image analysis apparatus 10 and the information exchange apparatus 40, but the information exchange system 100 is illustrated. May be mounted on one server.

(Image registration process)
First, the processing at the time of image registration by the information exchange system 100 will be described with reference to FIG. This image registration process is started, for example, when the user accesses the image analysis device 10 from the terminal device 20 and registers image data. First, the control unit 13 communicates with the terminal device 20, and this terminal device 20. User authentication is performed by inputting a user ID and a user password (step S101). This user authentication is performed by the control unit 13 collating the input user ID and user password with the user ID and user password stored in advance in the user management information storage unit 60 of the storage unit 12. When the user is authenticated in step S101, the control unit 13 stores all image data uploaded from the user via the terminal device 20 in the attribute information area 71 and the image data area 72 of the image information storage unit 70. At the same time, the management information of the image data is stored in the image management information storage unit 73 (step S102).

  Next, the analysis unit 14 reads out one of the image data stored in the image information storage unit 70 by the above processing as an analysis target image (step S103), determines a category of the read analysis target image, The information is stored in the category area 78 of the management information storage unit 73 (step S104). As a specific category determination method, pattern recognition or object recognition using a neural network is used. For example, sample information is prepared in advance for each category, a 256-dimensional Euclidean space distance between the sample information and the evaluation target image is calculated, and an optimum one is selected and determined based on the calculated value. (For example, a method of determining the category of the sample information with the closest distance as the category of the image to be evaluated, or a predetermined threshold is provided, and the category of the sample information at a distance within the threshold is selected. There is a method of determining the category of the image to be evaluated).

  Here, specific examples of the category include a person, a landscape, a night view, a vehicle, and the like. These categories are determined in advance and set in the image analysis apparatus 10, and the evaluation target image is selected from these categories. The category to which the belongs is determined. Further, the person may be further subdivided into categories such as women, men, and children, and the vehicle may be further subdivided into categories such as SL, Shinkansen, and trains, and the optimum category may be selected from these. .

  Subsequently, the analysis unit 14 weights the analysis target image (step S105). The weight W calculated in this weighting process is calculated using the following equation (1) by taking the shooting date / time X of the analysis target image from the shooting date / time region 77 of the image management information storage unit 73. In Equation (1), a and b are predetermined coefficients. For example, a is a negative value and b is 1. Y indicates the current date.

W = a (Y−X) + b (1)

  The weight W obtained by this equation (1) compares the date of photographing of the analysis target image with the current date, and the closer to the current date, the closer the category of the image data is to the photographer's preference (current photographer). Represents the tendency of the subject that the user prefers).

  As described above, when the processing of the selected one piece of image data (analysis target image) is completed, the analysis unit 14 has analyzed all the image data of the user stored in the image information storage unit 70. It is determined whether or not (step S106), and the above-described steps S103 to S105 are repeated until all image data are analyzed.

  When the analysis of all the image data is completed, the sum S of the weights W is calculated for each category using the following equation (2) based on the categories and weights of all the image data obtained in steps S103 to S105. Ask. In this embodiment, the total sum S is defined as the degree of preference of the category (step S107).

S = Σ (W) (2)

  And based on this preference degree S, a user's photography tendency is analyzed. For example, the category with the highest preference degree S or the categories are sorted in descending order of the preference degree S, and several categories in order from the highest preference level are determined as the main categories (category categories) indicating the user's shooting tendency. Then, the information is stored in the affiliation category area 64 of the user management information storage unit 60, and the fact that there is an affiliation category is stored in the category presence / absence information area 62 (step S108). Further, the image data stored in the image information storage unit 70 is analyzed by a predetermined method, and the analysis result is stored in the statistical information storage unit 19 (step S109). As an analysis method, the image data stored in the image information storage unit 70 and the image management information storage unit 73 is extracted with a user ID, and the user is assigned to each category of the image data (for example, female, male, etc.). A method for obtaining the cumulative total (number of scenes) of image data photographed in the past and image data extracted by user ID are classified by photographing date (for example, classified by month or year), and for each classified unit. There is a method of accumulating for each category (for example, female, male, railway) and obtaining the change of the user's photographing tendency as time series information.

  In the information exchange system 100, when the image registration process is configured as described above, every time a user registers image data in the image analysis apparatus 10, the preference degree S indicating the shooting tendency of the user is recalculated. The latest image data can be reflected in the preference degree S. Therefore, it is possible to accurately grasp the photographing tendency of the photographer (the tendency of the subject that is most preferred at present, which is set as the above-mentioned affiliation category). In this embodiment, the user's shooting tendency is analyzed based on the weight W and the preference degree S. However, as another embodiment, the shooting tendency is determined by the number of images to be analyzed for each category. The analysis may be simpler.

(Group information exchange process)
Next, the group information exchange process performed by the information exchange device 40 will be described with reference to FIG. This group information exchange process is started, for example, when the user accesses the information exchange device 40 from the terminal device 20. At this time, the bulletin board management unit 42 performs user authentication based on the user ID and the user password input from the terminal device 20 (step S111). Note that the information exchange apparatus 40 can access the storage unit 12 of the image analysis apparatus 10 via the network 30.

  Next, the notification unit 44 performs grouping of users. First, at the first access, the user grouping is performed based on the affiliation category stored in the affiliation category area 64 of the user management information storage unit 60 in step S108 described above. And the obtained group is stored in the group name area 93 of the group information storage unit 90 (step S112). In the second and subsequent accesses, when a new group different from the previous time is added due to the addition of image data or the like, this additional group is stored in the vacant group information area 93. When the storage area reaches the maximum number and there is no area that can be stored, it is stored in a temporary storage area.

  The bulletin board management unit 42 of the information exchange apparatus 40 is provided with a bulletin board that can participate in each group to which the information exchange device 40 belongs, and the notification unit 44 includes an additional group temporarily stored in the user management information storage unit 60. It is determined whether or not there is a non-participating group (non-participating group) among bulletin boards (that is, bulletin boards in which this user can participate) corresponding to the belonging group set in the belonging category area 64 (step S113). . This determination is made by referring to the registration date / time area 94 of the group information storage unit 90. If the user has already participated, the registration date / time area 94 stores the registration date / time. If the user has not participated, the registration date / time area 94 remains the initial value (for example, blank or NUL).

  When the notification unit 44 determines that there is a non-participating group, the notification unit 44 notifies the terminal device 20 that there is a new group that can participate (step S114). Then, it is determined whether or not the user wishes to participate (step S115). If the user wishes to participate, the participation group is registered (step S116). This registration is performed by storing the current date and time as the registration date and time in the registration date and time area 94 corresponding to the group. When the number of groups reaches the maximum number that can be added, the group with the oldest registration date and time is deleted, and a new group name and registration date and time are registered. On the other hand, if the user does not wish to participate, step S116 is not performed.

  Next, the bulletin board management unit 42 of the information exchange device 40 displays a selection initial screen for the user to select a desired process on the terminal device 20 (step S117). On this initial selection screen, the user can select “display statistical information” or “transmit / receive information”. When the user selects “display statistical information” (step S118), the user is requested to instruct what statistical information is to be displayed next (step S119). Statistical information corresponding to the instruction content from the user is extracted from the statistical information storage unit 80 of the storage unit 12 by the analysis result presentation unit 15 and generated as displayable information, which is transmitted to the terminal device 20. The statistical information is displayed on the display unit (step S120).

  Display examples of this statistical information are shown in FIGS. In FIG. 9, as described above, the image data stored in the image information storage unit 70 and the image management information storage unit 73 are extracted by the user ID, and each category of the image data (for example, female, male, etc.) is extracted. The radar graph represents the result of obtaining the cumulative total (number of scenes) of image data taken by the user in the past. On the other hand, FIG. 10 classifies the image data extracted by the user ID by the shooting date (for example, by month or year), and for each category (for example, female, male, railway). 3 is a graph display of the analysis result in which the transition of the shooting tendency of the user is displayed as time series information.

  On the other hand, when the user selects “transmission / reception of information” (step S121), the bulletin board management unit 42 selects which of the plurality of groups to which the user belongs to transmit / receive information. An instruction to perform is requested (step S122). When a group is selected from the user, the bulletin board management unit 42 of the information exchange device 40 controls the user to exchange information with other users via the bulletin board in the selected group. (S123). As a result, the user can send and receive various information such as image data, analysis results, and messages to other users in the selected group via the bulletin board. Note that transmission / reception of image data between users is performed by the control unit 13 of the image analysis apparatus 10 via the information exchange apparatus 40. In addition, the message control unit 43 stores and manages messages input to the bulletin board by the user during information exchange.

  As described above, when grouped users (photographers) with the same shooting preferences are grouped from the analysis results statistically analyzed using group information exchange processing, and information exchange is possible between users belonging to the same group An example of this is shown in FIG. As shown in FIG. 11, for example, in the case of the “railway” group, the photographer A transmits a request to “Tell me about train shooting spots” to the bulletin board. Then, in response to the request from the photographer A, the photographer B transmits information indicating that “a new train is running at the station XX”. In response to the request from the photographer A, the photographer D transmits information indicating that “it is raining next week”. As described above, the photographer A can easily obtain information for photographing by exchanging information with a user having the same photographing preference, and the motivation for photographing is improved, and the action for photographing is efficiently performed. Can be done.

  In addition, the photographer E transmits a request to the effect that “I want an SL photo” to the bulletin board of the group. Then, in response to the request from the photographer E, the photographer C transmits information indicating that the photograph of the SL is “opened by the photographer C itself”. Thereby, the photographer E can quickly obtain necessary information by, for example, browsing the photograph of the photographer C published on the website or downloading the image data registered in the image analyzer 10. Obtain high satisfaction. Such image data transmission / reception between photographers is managed by the image analysis device 10 via the information exchange device 40.

It is explanatory drawing which shows the structure of the information exchange system which concerns on this invention. It is a block diagram which shows the structure of the said information exchange system. It is explanatory drawing for demonstrating the data structure of a user management information storage part. It is explanatory drawing for demonstrating the data structure for managing image data, Comprising: (a) shows the data structure of an image information storage part, (b) shows the data structure of an image management information storage part. It is explanatory drawing for demonstrating the data structure of a statistics information storage part. It is explanatory drawing for demonstrating the data structure of a group information storage part. It is a flowchart which shows an image registration process. It is a flowchart which shows a group information exchange process. It is explanatory drawing for demonstrating an example of the statistical information displayed on a display part, and is the case where the ratio of a category is shown with the radar chart. It is explanatory drawing for demonstrating the other example of the statistical information displayed on a display part, and is a case where the change by the time series of a category is shown with the line graph. It is explanatory drawing for demonstrating the information exchange using an information exchange system.

Explanation of symbols

DESCRIPTION OF SYMBOLS 10 Image analyzer 12 Memory | storage part 14 Analysis part 15 Analysis result presentation part 20 Terminal device 40 Information exchange apparatus 41 Bulletin board management part 100 Information exchange system

Claims (10)

A storage unit for storing the analysis target image;
For each analysis target image stored in the storage unit, a category to which the analysis target image belongs is determined, and the category to which the analysis target image stored in the storage unit belongs is stored in the storage unit. And an analysis unit that calculates an affiliation category that is a category that represents a photographing tendency of the analysis target image. The analysis unit has sample information corresponding to each of the plurality of categories;
The image analysis apparatus according to claim 1, wherein a category to which each analysis target image belongs is determined from the analysis target image and the sample information using pattern recognition or object recognition using a neural network. The analysis unit
For each of the analysis target images, the weight W is expressed by the following formula W = a (Y−X) + b from the photographing date X, the current date Y, and the coefficients a and b of the analysis target image.
Calculated by
Further, for each category, among the analysis target images stored in the storage unit, the degree of preference of the category is calculated as a sum of the weights W of the analysis target images belonging to the category, and from the degree of preference The image analysis apparatus according to claim 1, wherein the image analysis apparatus is configured to determine the affiliation category.   The image analysis apparatus according to claim 3, wherein among the categories to which the analysis target image stored in the storage unit belongs, a category having the highest degree of preference is set as the belonging category.   4. The category to which the analysis target image stored in the storage unit belongs is rearranged in descending order based on the preference level, and a predetermined number of categories from the category with the highest preference level are set as the belonging category. The image analysis apparatus described.   The analysis result presenting unit that presents the photographing tendency of the analysis target image stored in the storage unit by classifying the number of the analysis target images stored in the storage unit for each category. The image analysis device according to any one of 5. The storage unit is configured to store the analysis target image together with a shooting date when the analysis target image is shot;
By classifying the analysis target images stored in the storage unit for each shooting date, and calculating the number of images for each category with respect to the classification, the number of images for each category based on the time series is stored in the storage unit. The image analysis apparatus according to claim 1, further comprising an analysis result presentation unit that presents a photographing tendency of the analysis target image. The storage unit is configured to store the analysis target image together with identification information of a registrant who has registered the analysis target image.
The image analysis apparatus according to claim 1, wherein the analysis unit is configured to calculate the affiliation category for each identification information. The image analysis apparatus according to claim 8 connected to a network;
A user having the identification information connected to the network and configured to exchange information using the identification information;
The information exchange system, wherein the information exchange device includes a group information storage unit that groups the identification information having the same affiliation category among the affiliation categories.   The information exchange system according to claim 9, further comprising: a bulletin board management unit that exchanges messages by users having the identification information belonging to the group for each group.