JP4313638B2 - Image management method, image management program, and image management system - Google Patents

Description

The present invention relates to an image management method for managing image data, relating to images management program and an image management system.

  As a method of searching for a desired image from an image database in which a plurality of images are accumulated, a method of searching for a similar image based on the feature amount of a designated search source image and the feature amount of each image in the image database is known. It has been. Various methods have been proposed as methods for extracting and using feature amounts used for such similar image search (see, for example, Patent Document 1 and Patent Document 2).

Furthermore, a method is known in which a search request is accepted via a network such as the Internet, a similar image search is performed, and a search result is provided (see, for example, Patent Document 3).
JP 2000-187731 A (page 4-10) JP 2000-285141 A (pages 10-14) JP 2002-245045 A (page 3-5)

  However, in the conventional search method described above, even images related to the same subject may not be extracted as similar images depending on the form and configuration of the subject in each image. For example, when a swan image is desired and a single swan image floating on the water is designated as a search source image, there is a high possibility that an image of a group of swan flying in the sky is not extracted as a similar image. As described above, the determination may be different between the case where the similarity is determined by focusing on the meaning of the subject and the case where the similarity is determined based only on the feature amount of the image.

  For this reason, when providing a similar image in consideration of the meaning of the subject, for example, it is conceivable to use text data. However, if text data is used, it is necessary to associate and register appropriate text data with an image, and the burden on the user for registration and association is large. In this case, management becomes complicated as the number of text data for explaining each image increases.

The present invention has been made to solve the above problems, and an object, an image management method that can provide a variety of images based on the similar image search, the images management program and an image management system It is to provide.

In order to solve the above-mentioned problems, the invention described in claim 1 includes an image data storage unit that stores image data, an association information storage unit that stores association information of similar image data, and a selection history storage unit. , using an image management system including a computer to manage these, an image management method for managing an image to be provided to the user, the computer, the first step of accepting designation of retrieval source image from the user , Using a second step of extracting image data of a similar image having a feature amount approximate to the feature amount of the search source image from the image data storage unit, and association information recorded in the association information storage unit, a third step of extracting image data of the related image related to the similar image from the image data storage means, the similar image and before said extracted The image data of the related image presented to the user, and a fourth step of storing the association information of the selected image by the user with the search source image on the association information storage unit, the image in the fourth stage The selection identification information specifying that the selection is made under the same condition by the same user is stored in the selection history storage means in association with the selected image, and the selection history storage means is used. Then, the gist of executing the step of deleting the association information about the image that has not been selected by the same user under the same condition within the predetermined period from the association information storage means .

  According to a second aspect of the present invention, in the image management method according to the first aspect, the first stage accepts a search source image from the user, and the image management method includes: And a step of storing the image data of the search source image in the image data storage means.

  The invention according to claim 3 is the image management method according to claim 1 or 2, wherein the feature amount includes at least a feature amount related to any of an overall color, a color distribution, a contour, and a pattern. To do.

  According to a fourth aspect of the present invention, in the image management method according to any one of the first to third aspects, the image management system includes a simultaneous extraction history storage unit, and the image management method is performed by the computer. When a plurality of similar images are extracted in the second step, the plurality of similar images are associated with each other and stored together with the extraction time information in the simultaneous extraction history storage unit, and the simultaneous extraction history storage unit is used. And a step of deleting from the association information storage means association information about images that were not simultaneously extracted within a predetermined period.

A fifth aspect of the present invention is the image management method according to any one of the first to fourth aspects, wherein the selection identification information is a user identifier and a selection time of a user who has selected the image. And
The invention of claim 6 includes an image data storage means for storing image data, and the association information storage means for storing the association information of the image data to be similar, and the selection history storage unit, and a computer to manage these An image management program for managing an image to be provided to a user using an image management system, wherein the computer has a first means for accepting designation of a search source image from the user, and features of the search source image A second means for extracting image data of a similar image having a feature quantity approximate to the quantity from the image data storage means, and association information recorded in the association information storage means, and third means and the image of the extracted the similar image and the associated image to extract image data from the image data storage means Presenting over data to the user, and the fourth means for storing the association information between the selected image by the user the search source image on the association information storage unit, an image is selected by said fourth means The selection identification information for specifying that the selection is made by the same user under the same condition is stored in the selection history storage unit in association with the selected image, and the selection history storage unit The gist of the present invention is to cause the association information about an image that has not been selected under the same condition by the same user within a period to function as a means for deleting from the association information storage means .

According to a seventh aspect of the present invention, in the image management program according to the sixth aspect , the first means functions as a means for receiving a search source image from the user, and the computer further includes the search source image. The gist of the present invention is to function as means for calculating the feature amount and means for storing the image data of the search source image in the image data storage means.

The invention according to claim 8 is the image management program according to claim 6 or 7 , wherein the feature amount includes at least a feature amount related to any of an overall color, a color distribution, a contour, and a pattern. To do.

According to a ninth aspect of the present invention, in the image management program according to any one of the sixth to eighth aspects, the image management system includes a simultaneous extraction history storage unit, and the simultaneous extraction history storage unit is used. The computer further includes means for associating the plurality of similar images with each other and storing them in the simultaneous extraction history storage means together with the extraction time information when a plurality of similar images are extracted by the second means. The gist of the present invention is to use the extraction history storage means to function as means for deleting association information about images that have not been simultaneously extracted within a predetermined period from the association information storage means.

A tenth aspect of the present invention is the image management program according to any one of the sixth to ninth aspects, wherein the selection identification information is a user identifier and a selection time of a user who has selected the image. And
The invention described in claim 11 includes image data storage means for storing image data, association information storage means for storing association information of similar image data, selection history storage means, and a computer for managing these. An image management system for managing images to be provided to a user, wherein the computer has first means for accepting designation of a search source image from the user, and a feature amount approximate to the feature amount of the search source image Using the second means for extracting the image data of the similar image from the image data storage means and the association information recorded in the association information storage means, the image data of the related image relating to the similar image is stored in the image data storage means. A third means for extracting the image and the image data of the extracted similar image and the related image are presented to the user; The fourth means for storing the association information between the image selected by the user and the search source image in the association information storage means, and when the image is selected by the fourth means, under the same conditions by the same user The selection history information is stored in the selection history storage unit in association with the selected image, and the selection history storage unit is used to select the same condition by the same user within a predetermined period. The gist is provided with means for deleting association information regarding an image that has not been recorded from the association information storage means.
The invention according to claim 12 is the image management system according to claim 11, wherein the selection identification information is a user identifier and a selection time of a user who has selected the image.

(Function)
According to the present invention , an image selected by the user from images searched based on a search source image can be associated with the search source image. When one of the associated images is extracted as a similar image, the related image associated with the image can be extracted. Therefore, the images can be associated based on the selection by the user, and the images can be extracted based on the association. For this reason, a similar image is searched using the feature amount, and an image that is similar to the similar image can be searched by the user. Therefore, various images can be provided based on a similar image search that searches for a similar image using a feature amount based on a search source image.

According to the present invention, it is possible to newly register a search source image and associate the search source image with an image selected by the user from images searched based on the search source image. Therefore, a search source image can be newly registered, and various images can be provided based on a similar image search related to a newly registered image.

According to the present invention , a similar image can be searched based on at least a feature amount related to any one of the overall color, color distribution, outline, and pattern.
According to the present invention , it is possible to unlink an image that has not been simultaneously extracted as a similar image based on a search source image within a predetermined period. Therefore, it is possible to unlink the images having low similarity to each other in relation to the search source image, and to suppress an unlimited increase in the association.

ADVANTAGE OF THE INVENTION According to this invention , the association about the image which was not selected on the same conditions by the same user within the predetermined period can be cancelled | released, and it can suppress that an association increases unlimitedly.

ADVANTAGE OF THE INVENTION According to this invention , the correlation of the image which was not selected within the same period by the same user within a predetermined period can be cancelled | released. Therefore, the association can be solved for the images that the same user did not show interest in the same period within the predetermined period, and the association can be prevented from increasing indefinitely.

  According to the present invention, various images can be provided based on a similar image search.

(First embodiment)
Hereinafter, a first embodiment of the present invention will be described with reference to FIGS. In the present embodiment, associating the image based on the selection by the user, if the one of the image is extracted as a similar image, the image managing method for also extracted the other image, as images management program and an image management system explain. In this embodiment, a similar image search is performed based on a search source image designated by the user, a search result is presented, and an image selected by the user from the presented images is associated with the search source image. When one of the associated images is extracted as a similar image, the other image is also extracted. In the present embodiment, a similar image refers to an image that is searched based on a search source image by a similar image search that uses feature amounts relating to the overall color, color distribution, contour, and pattern of the image. To do.

In the present embodiment, as shown in FIG. 1, the user uses the client terminal 30 to designate a search source image, makes a similar image search request, and selects a desired image from the search result images.
The client terminal 30 is a user computer terminal having a function of transmitting data via a network, a function of displaying received data, and the like. In addition to a CPU, RAM, and ROM (not shown), the client terminal 30 includes input means such as a keyboard and a mouse, output means such as a display, communication means such as a modem, and the like.

  The client terminal 30 is connected to the image search server 20 of the image management system via the Internet 40 as shown in FIG. In the present embodiment, the image search server 20 is a server that performs various data processing such as management of various data such as image data related to similar image search and provision of image data. The image search server 20 includes a management computer 21.

  The management computer 21 performs data transmission / reception with each client terminal 30, data management processing for managing various information, and the like. The management computer 21 includes a CPU, a RAM, a ROM, and the like (not shown), and performs processing described later (processing including a first stage, a second stage, a third stage, a fourth stage, and the like). By executing the image management program for that purpose, the management computer 21 functions as first means, second means, third means, fourth means, and the like.

  Further, the image search server 20 includes an image data storage unit 22 as image data storage means and a search data storage unit 23 as association data storage means. Image data is recorded in the image data storage unit 22. In the present embodiment, a plurality of image data are registered in advance by the administrator of the image search server 20. Further, when the search source image is received from the client terminal 30 and image search is performed and associated with the image selected by the user, the image data of the search source image is recorded.

  As shown in FIG. 2, search data 230 relating to information for similar image search is recorded in the search data storage unit 23 for each image data. The search data 230 is recorded when image data is registered in the image data storage unit 22. The search data 230 includes data relating to an image ID, overall color, color distribution, contour, pattern, storage destination, and related image ID for each image data.

Data relating to an identifier for specifying image data is recorded in the image ID data area. In the overall color data area, data relating to the feature value obtained by digitizing the overall color is recorded. In the color distribution data area, data relating to a feature value obtained by quantifying the color distribution is recorded. In the contour data area, data relating to a feature value obtained by digitizing the contour is recorded. In the pattern data area, data relating to the feature value obtained by digitizing the pattern is recorded. Data relating to the location of the corresponding image data is recorded in the storage destination data area. In the related image ID data area, data related to the image ID of the image data associated based on the user's selection is recorded. These feature amounts are calculated using a known similar image search technique.

  In the system configured as described above, a processing procedure for searching for a similar image based on the search source image and associating the image selected by the user among the images presented as the search result with the search source image is illustrated. This will be described with reference to FIGS. Here, a description will be given by dividing into an associating stage for associating an image selected by the user from the search results with the search source image and a searching stage for the associated image. In the present embodiment, an image owned by a user newly registered in the image data storage unit 22 or an image recorded in the image data storage unit 22 is used as the search source image.

<Association stage>
Here, a case where an image owned by the user is used as a search source image and the image is newly registered in the image data storage unit 22 will be described.

  A user who wishes to register an image in the image search server 20 first accesses the image search server 20. Here, a request for search request screen data is transmitted to the image search server 20 via the Internet 40 using the client terminal 30. The management computer 21 that has received this transmits the search request screen data to the client terminal 30 via the Internet 40.

  The display screen 50 shown in FIG. 3 is output to the display of the client terminal 30 that has received the search request screen data. The display screen 50 includes a search source image selection field 51, a reference button 52, and a send button 53. When the reference button 52 is clicked, a file selection screen is displayed. When a file (search source image) is selected on the file selection screen, the location of the selected file is displayed in the search source image selection field 51. When the transmission button 53 is clicked, the client terminal 30 transmits the image data of the selected search source image to the image search server 20 via the Internet 40.

  As shown in FIG. 7, when receiving the image data of the search source image from the client terminal 30 (step S1-1), the management computer 21 of the image search server 20 calculates the feature amount of the search source image (step S1- 2). Specifically, the management computer 21 calculates feature amounts for each of the overall color, the color distribution, the contour, and the pattern using a known similar image search technique based on the image data of the search source image.

  Then, the management computer 21 extracts a similar image using a known similar image search technique based on the feature amount of the search source image (step S1-3). Specifically, the management computer 21 first obtains the degree of similarity related to the feature amount related to each feature amount of the search source image and each searched image. Here, first, the distance between the image to be searched and the image to be searched that is plotted in the feature amount space based on each feature amount is obtained using a vector distance definition formula, and the distance is specified for those distances. Weighting according to importance. In the initial search, it is assumed that there is no bias in the importance of each feature amount.

Next, the management computer 21 calculates the total similarity by summing the distances calculated for each of the overall color, color distribution, contour, and pattern. In this way, the management computer 21 calculates the total similarity for all the searched images. Then, the management computer 21 extracts, as the search data 230 of similar images, the search data 230 whose total similarity is equal to or higher than a predetermined value in descending order of the total similarity. Then, the management computer 21 extracts image data from the image data storage unit 22 based on the data relating to the storage location included in the extracted search data 230.

  Next, the management computer 21 extracts related images (step S1-4). Here, the related image refers to an image related by being recorded in the search data 230 as a related image ID. Specifically, the management computer 21 extracts a related image ID included in the extracted search data 230. Then, the management computer 21 extracts the search data 230 from the search data storage unit 23 using the image ID extracted as the related image ID. Then, the management computer 21 extracts image data from the image data storage unit 22 based on the data relating to the storage location included in the search data 230.

  Then, the management computer 21 transmits search result list screen data displaying the search result of similar images to the client terminal 30 via the Internet 40 (step S1-5).

  The display screen 60 shown in FIG. 4 is output to the display of the client terminal 30 that has received the search result list screen data. The display screen 60 includes a search source image display field 61, an importance level designation field 63, a re-search request button 65, and a search result list 66. In the search source image display column 61, the image image 62 of the search source image is displayed based on the image data of the search source image received by the image search server 20 in step S1-1. The importance level designation column 63 includes an importance level designation bar 64 for the entire color, color distribution, outline and pattern. Each importance designation bar 64 is used when the relative importance is designated by sliding the designation portion. Then, weighting for each feature amount is performed based on the designated importance. Note that, in the first search, as described above, it is assumed that the importance level of each feature amount is not biased, and the designation portion is located at the center of each importance level designation bar 64. The re-search request button 65 is used when a re-search is requested based on the designation of the importance designation bar 64. The search result list 66 displays a list of thumbnail images of similar images and related images that have been searched. Here, similar images are arranged and displayed in descending order of the total similarity with the search source image. When there is a related image for a similar image, the related image is arranged next to the corresponding similar image.

  Here, when there is an image in which the user is interested in the search result list 66, the user clicks a thumbnail image in the search result list 66 for the image. Here, the search source image is an image of “a state where swans on the lake are spreading their wings”. On the other hand, the search result list 66 includes an image of “swan on the lake in a state where the wings are not spread” of the image ID “20015”. Here, it is assumed that the user desires an image with the image ID “20015”. In this case, the user clicks on the thumbnail image 67. Then, the client terminal 30 transmits display request data including the image ID “20015” corresponding to the clicked thumbnail image 67 to the image search server 20 via the Internet 40.

  When receiving the display request data, the management computer 21 of the image search server 20 transmits download request screen data to the client terminal 30 via the Internet 40.

The display screen 70 shown in FIG. 5 is output to the display of the client terminal 30 that has received the download request screen data. The display screen 70 includes an image image 71 of an image requested to be displayed and a download request button 72. The user confirms the image image 71 and examines whether or not the image data needs to be stored in the client terminal 30. When the user desires to save the image data, the user clicks the download request button 72. In response to this, a storage location selection screen is displayed on the display of the client terminal 30.

  When the storage location is designated, the client terminal 30 transmits download request data including the image ID to the image search server 20. When the management computer 21 that has received this downloads the designated image data to the client terminal 30, the client terminal 30 records the image data in the designated storage location.

  On the other hand, when receiving the download request data (YES in step S1-6), the management computer 21 stores the search source image in association with the image requested to be downloaded (step S1-7). Specifically, the management computer 21 stores the image data of the search source image in the image data storage unit 22, assigns an image ID, and stores the search data 230 for the search source image in the search data storage unit 23. Record. Here, for the overall color, color distribution, contour, and pattern, the respective feature values calculated in step S1-2 are recorded. As for the storage destination, data relating to the location of the search source image in the image data storage unit 22 is recorded. As for the related image ID, the image ID of the image requested to be downloaded is recorded. Here, “30001” is assigned as the image ID of the search source image.

  Further, the management computer 21 searches the search data storage unit 23 based on the image ID of the image requested to be downloaded. Then, the image ID of the search source image is recorded in the related image ID data area of the search data 230 of the image ID.

  On the other hand, when there is no image desired by the user in the search result list 66 on the display screen 60, it is possible to change the weighting of the feature amount and perform a search again. For example, the overall color and the color distribution need to be approximated to the search source image, but the outline and the pattern may not be very close to the search source image. In such a case, the designation part of the importance designation bar 64 for the entire color and the color distribution is slid to the right (in the direction of increasing importance), and the designation part of the importance designation bar 64 for the outline and the pattern is respectively changed. Slide left (to decrease importance). Then, when the re-search request button 65 is clicked, the client terminal 30 transmits re-search request data including designation of importance for each feature amount to the image search server 20.

  If download request data has not been received (NO in step S1-6) and re-search request data has been received (YES in step S1-8), management computer 21 is designated. The search process is performed again according to the importance. Specifically, the management computer 21 extracts similar images in step S1-3 based on the importance of each designated feature amount.

  Here, higher importance is specified for the overall color and the color distribution. On the other hand, a lower importance is specified for the outline and the pattern. Therefore, for the distance between the image to be searched and the image to be searched obtained based on each feature amount, the overall color and color distribution are weighted corresponding to higher importance, and the outline and pattern are more Perform weighting corresponding to low importance. It should be noted that the weighting corresponding to the direction of increasing the importance acts in the direction of shortening the distance between the images, and the weighting corresponding to the direction of decreasing the importance acts in the direction of increasing the distance between the images. Therefore, the weighting for the whole color and the color distribution works in a direction to shorten the distance between the search source image and the searched image, and the weighting for the outline and the pattern increases the distance between the search source image and the searched image. Work in the direction.

Then, the management computer 21 calculates the total similarity by summing the distances calculated for each of the overall color, color distribution, outline, and pattern. In this way, the management computer 21 calculates the total similarity for all the searched images. Then, the management computer 21 extracts, as the search data 230 of similar images, the search data 230 whose total similarity is equal to or higher than a predetermined value in descending order of the total similarity.

  Then, as in the case described above, the management computer 21 extracts the search data 230 whose total similarity is greater than or equal to a predetermined value in order from the highest total similarity as the search data 230 of similar images. Then, the management computer 21 extracts image data from the image data storage unit 22 based on the data relating to the storage location included in the extracted search data 230.

And the management computer 21 performs the process after step S1-4 similarly to said case.
<Related image search stage>
Next, the case where one of the images registered in association with each other is extracted as a similar image will be specifically described. Here, the image ID “30001” “swan on the lake spreading wings” associated with the image and the image ID “20015” “swan on the lake not spreading wings” are associated. ”Will be described.

  In this case, as described above, in the related image ID data area in the search data 230 related to the image of the “registered swan on the lake” with the newly registered image ID “30001”, the image ID “ 20015 "is recorded. On the other hand, the image ID “30001” is recorded in the related image ID data area in the search data 230 regarding the image of “swan on the lake in a state where the wings are not spread” of the image ID “20015”.

  Here, a case where a search is performed using an image of “swan on the lake” different from the above two images as a search source image will be described. Here, an image registered in the image search server 20 is used as a search source image. In the present embodiment, a thumbnail image list screen is provided to the user for the image data stored in the image data storage unit 22, and the user selects an image as a search source image from the thumbnail image list screen. When an image registered in the image search server 20 is used as a search source image, the image data of the image selected as the search source image and the search data 230 are used instead of the processing of step S1-1 and step S1-2 in FIG. And extract. And the data regarding the feature-value contained in the extracted search data 230 is used.

  In step S <b> 1-3, the management computer 21 searches for similar images based on the feature amounts of the search data 230 for the search source image and the feature amounts of the search data 230. Here, the extracted similar image includes an image of “swan on the lake in a state where the wings are not spread” of the image ID “20015”. Here, it is assumed that the extracted similar image does not include an image with the image ID “30001” “a swan on the lake spreading its wings”.

  In this case, in step S1-4, based on the search data 230 with the image ID “20015”, an image with the image ID “30001” “a swan on the lake spreading its wings” is extracted as a related image. The Specifically, the management computer 21 first extracts the image ID “30001” from the related image ID data area of the search data 230 of the image ID “20015”. Based on this, the management computer 21 extracts the search data 230 with the image ID “30001”. Then, the management computer 21 extracts the image data of the image “a state where the swans on the lake are spreading their wings” from the image data storage unit 22 based on the data relating to the storage location included in the search data 230.

In this case, the search result list screen data transmitted to the client terminal 30 in step S1-5 includes the image with the image ID “20015” and the image with the image ID “30001”.

  The display screen 60 shown in FIG. 6 is output to the display of the client terminal 30 that has received the search result list screen data. The display screen 60 includes a search source image display field 61, an importance level designation field 63, a re-search request button 65, and a search result list 66. In the search source image display field 61, an image image 81 of the search source image is displayed. The search result list 66 displays a list of thumbnail images of similar images and related images that have been searched. In the present embodiment, the image extracted as the related image of the similar image is displayed next to the related similar image in the search result list 66. Therefore, in this case, next to the thumbnail image 86 of the image ID “20015” “Swan on the lake without spreading wings”, the related image “Swan on the lake” of image ID “30001” spreads its wings. A thumbnail image 87 of “in a state of being” is arranged.

  If the related image extracted in step S1-4 includes the similar image extracted in step S1-3, only one image is displayed. Here, the related images are arranged in the search result list 66 so that the order of the positions arranged as the related images and the positions arranged as the similar images comes first.

  Specifically, when the total similarity of the related images is higher than the total similarity of the similar images associated with the related images, the related images are arranged in the search result list 66 according to the total similarity of the related images. . On the other hand, when the total similarity of the related images is lower than the total similarity of the similar images associated with the related images, the related images are arranged next to the similar images associated with the related images in the search result list 66. To do.

As described above, according to the present embodiment, the following effects can be obtained.
In the above embodiment, the management computer 21 accepts designation of the search source image, and extracts image data of a similar image having a feature amount approximate to the feature amount of the search source image from the image data storage unit 22. Then, an image (related image) associated with the similar image is specified based on the related image ID included in the search data 230, and the image data of the related image is extracted from the image data storage unit 22. Then, the extracted similar image and related image are presented to the user, a download request for an image selected by the user from among these images is received, and the image data is downloaded. For this reason, not only a similar image but the related image linked | related with the similar image can be provided to a user.

  Then, association information between the search source image and the image requested to be downloaded is stored in the search data storage unit 23. For this reason, the image selected by the user from the images searched based on the search source image can be associated with the search source image. When one of the associated images is extracted as a similar image, the related image associated with the image can be extracted. Therefore, the images can be associated based on the selection by the user, and the images can be extracted based on the association. For this reason, various images can be provided based on a similar image search in which a similar image is searched using feature amounts.

In the above embodiment, the management computer 21 receives the search source image from the client terminal 30 and stores the image data of the search source image in the image data storage unit. The feature amount of the search source image is calculated and recorded in the search data storage unit 23, and the image ID of the image requested to be downloaded is recorded in the search data storage unit 23 as the related image ID for the search source image. Therefore, it is possible to newly register a search source image and associate the search source image with an image selected by the user from images searched based on the search source image. Therefore, a search source image can be newly registered, and various images can be provided based on a similar image search related to a newly registered image.

  In the above-described embodiment, similar images are searched using feature amounts relating to the overall color, color distribution, outline, and pattern. Then, for each feature amount, each feature amount is weighted according to the importance set by the user, and a similar image is searched. For this reason, it is possible to search for an image similar to a case where the overall color, color distribution, contour, and pattern are comprehensively determined. In addition, by weighting each feature amount, it is possible to increase the accuracy of similar image search performed based on the search source image.

  In the above embodiment, when a download request for downloading and saving an image to the client terminal 30 is received, image association is performed. For this reason, only the image required by the user can be associated with the search source image.

(Second Embodiment)
Hereinafter, a second embodiment of the present invention will be described with reference to FIG. In this embodiment, the association is deleted in order to prevent the association from increasing indefinitely.

In the second embodiment, the image search server 20 of the first embodiment is further provided with a simultaneous extraction history data storage unit 91 as a simultaneous extraction history storage unit.
In the simultaneous extraction history data storage unit 91, as shown in FIG. 8, simultaneous extraction history data 910 relating to the extraction history of images simultaneously extracted as similar images is recorded. The simultaneous extraction history data 910 is recorded when an image is associated. In step S1-3, the associated image is updated when it is simultaneously extracted as a similar image based on the search source image. The simultaneous extraction history data 910 is configured to include data relating to the first image ID, the second image ID, and the use date for the associated images.

  Data relating to the image ID of the associated image is recorded in the first image ID data area and the second image ID data area. In the present embodiment, a lower-numbered image ID is a first image ID, and a higher-numbered image ID is a second image ID. In the usage date data area, data relating to the date of association when the association is performed is recorded. In addition, when similar images are extracted, the data is updated with data relating to the date of extraction of similar images.

  Hereinafter, a processing procedure in the case of deleting the association of images that were not simultaneously extracted as similar images within a predetermined period will be described. Here, the description will be divided into a similar image extraction history recording stage and an association deletion stage.

<Similar image extraction history recording stage>
In the present embodiment, in addition to the processing of the first embodiment, simultaneous extraction history data 910 is recorded when associating a search source image with an image requested to be downloaded. Specifically, the management computer 21 first records the simultaneous extraction history data 910 for the search source image and the image requested to be downloaded in step S1-7 in FIG. Here, since the image requested to be downloaded has a smaller image ID number, the image ID of the image requested to be downloaded is set as the first image ID, and the image ID of the search source image is set as the second image ID. Record. The date of the day is recorded as the usage date.

And about the image extracted simultaneously as a similar image, the management computer 21 updates the utilization date of simultaneous extraction history data 910 with a simultaneous extraction date. That is, when a plurality of similar images are extracted in step S1-3 of FIG. 7, the management computer 21 updates the use date of the simultaneous extraction history data 910 for the similar images extracted at the same time. Specifically, the management computer 21 sets the image ID as the first image ID in order from the smallest image ID number of the extracted similar images, and sets the image ID with the larger number as the second image. The simultaneous extraction history data 910 is searched as an ID. Then, the management computer 21 updates the use date of the retrieved simultaneous extraction history data 910 with the date of the day (simultaneous extraction date).

<Association deletion stage>
Next, a processing procedure when deleting the association will be described. In the present embodiment, this processing is performed at a predetermined timing (for example, daily).

  In this process, the management computer 21 first extracts simultaneous extraction history data 910 whose use date is a predetermined number of days or more ago from the simultaneous extraction history data storage unit 91. Then, the management computer 21 first searches the search data storage unit 23 using the first image ID of the extracted simultaneous extraction history data 910 as the image ID. Then, the management computer 21 deletes the related image ID corresponding to the second image ID of the simultaneous extraction history data 910 from the related image ID data area of the searched search data 230. Next, the management computer 21 searches the search data storage unit 23 using the second image ID of the extracted simultaneous extraction history data 910 as an image ID. Then, the management computer 21 deletes the related image ID corresponding to the first image ID of the simultaneous extraction history data 910 from the related image ID data area of the searched search data 230.

As mentioned above, according to 2nd Embodiment, in addition to the effect which 1st Embodiment show | plays, the effect shown below can be acquired.
In the above embodiment, the management computer 21 associates the extracted similar images with each other and records them in the simultaneous extraction history data storage unit 91 together with the simultaneous extraction date. Then, using the simultaneous extraction history data storage unit 91, the related image ID is deleted from the related image ID data area of each search data 230 for images that have not been extracted simultaneously within a predetermined period. For this reason, it is possible to unlink images that have not been simultaneously extracted as similar images based on the search source image for a predetermined period. Therefore, it is possible to suppress the association from increasing indefinitely.

(Third embodiment)
A third embodiment embodying the present invention will be described below with reference to FIGS. In this embodiment, the association is deleted in order to prevent the association from increasing indefinitely. In the present embodiment, when an associated image is not downloaded by the same user on the same day within a predetermined period, the association is deleted.

In the third embodiment, the image search server 20 of the first embodiment is further provided with a download history data storage unit 92 as selection history storage means.
As shown in FIG. 9, the download history data storage unit 92 stores download history data 920 for an image for which a download request has been made. This download history data 920 is recorded when a download request for image data is received. The download history data 920 includes data related to an image ID, a download date, and a user ID for each image download request.

  In the image ID data area, data relating to the image ID of the image used as the search source image or the image ID of the image having received the download request is recorded. In the download date data area, data related to the date of reception of the download request is recorded.

Hereinafter, a processing procedure when deleting the association of images that have not been downloaded by the same user on the same day for a predetermined period will be described. Here, the description will be divided into a download history recording stage and an association deletion stage.

<Download history recording stage>
In the present embodiment, in addition to the processing of the first embodiment, when a download request is made, download history data 920 is recorded for each of the search source image and the requested image. In the present embodiment, a user given a user ID accesses the image search server 20 from the client terminal 30 using his / her user ID.

  Specifically, when a download request is received in step S1-6 in FIG. 7, the management computer 21 records download history data 920 for the image requested for download and the search source image, respectively. Here, for the image requested to be downloaded, the image ID requested to be downloaded is recorded as the image ID of the download history data 920. On the other hand, for the search source image, the image ID given in step S1-7 is recorded as the image ID of the download history data 920 as the image ID of the search source image.

When a download request is made for a plurality of images based on the same search source image, one download history data 920 is recorded for the search source image.
<Association deletion stage>
Next, a processing procedure when deleting the association will be described. In the present embodiment, this processing is performed at a predetermined timing (for example, daily). This association deletion process will be described with reference to FIG.

  As shown in FIG. 10, the management computer 21 first extracts the image ID and related image ID of each search data 230 from the search data storage unit 23 (step S2-1). When a plurality of related image IDs are recorded in the related image ID data area of the search data 230, the following processing is performed for each related image ID. Further, the management computer 21 performs this process for all the search data 230 in the search data storage unit 23.

  Then, the management computer 21 searches the download history data 920 that has the same user ID and the same download date within the predetermined period for the extracted image ID and related image ID (step S2-2). Specifically, first, the management computer 21 searches the download history data 920 within the predetermined period from the current date using the extracted image ID. Further, the management computer 21 searches the download history data 920 within the predetermined period from the current date using the extracted related image ID as the image ID. Then, the management computer 21 checks whether or not the download history data 920 corresponding to the image ID and the related image ID of the search data 230 has the same user ID and the same download date.

  When the download history data 920 having the same user ID and the download date being the same date within the predetermined period is not extracted (NO in step S2-3), the management computer 21 deletes the association (step S2-4). Specifically, the management computer 21 deletes the processing target related image ID from the related image ID data area of the processing target search data 230. Further, the search data 230 is searched using the related image ID as an image ID, and the processing target image ID is deleted from the related image ID data area of the searched search data 230.

On the other hand, when the download history data 920 having the same user ID and the same download date within the predetermined period is extracted (YES in step S2-3), the management computer 21 ends this process.

As mentioned above, according to 3rd Embodiment, in addition to the effect which 1st Embodiment show | plays, the effect shown below can be acquired.
In the above-described embodiment, the management computer 21 records the download ID data storage unit 92 in association with the user ID of the user who requested the download of the image data and the download date. Then, using the download history data storage unit 92, the related image ID is deleted from the related image ID data area of each search data 230 for image data that has not been downloaded to the same user on the same day within a predetermined period. . For this reason, it is possible to unlink images that have not been requested to be downloaded by the same user on the same day for a predetermined period. Therefore, it is possible to unlink the images that the same user did not show interest in on the same day within a predetermined period, and it is possible to suppress the association from increasing without limit.

In addition, you may change the said embodiment into the following aspects.
In the first to third embodiments, when there is a download request for storing image data in the client terminal 30, the search source image and the image for which the download request has been made are associated. Alternatively, when there is another action for selecting an image, the search source image may be associated with the image selected by the action. For example, a list of thumbnail images may be displayed as a search result list, and when a thumbnail image is clicked, the image image may be displayed, and the search source image may be associated with the clicked image. By doing in this way, even when it is not necessary to store image data in the client terminal 30, the image in which the user showed interest can be associated with the search source image.

  In the first to third embodiments, when the search source image is registered, the image is associated when there is a download request for storing the searched image data in the client terminal 30. The method for associating images is not limited to this. For example, when registering a search source image, a means for a user to make an image association request may be provided, and the image association may be performed when the association request is made. For example, instead of the download request button 72, an association request button may be provided on the display screen 70 of FIG. 5, and when the association request button is clicked, only the association may be performed without downloading. By doing in this way, even when it is not necessary to download and save the image to be associated with the newly registered search source image to the client terminal 30, it is possible to associate the search source image with the selected image. Therefore, for example, when a user who registers an image is different from a user who uses the image, the user who registers the image can associate the images without downloading and storing the associated image.

  In the third embodiment, the user ID registered in advance is used. However, it is only necessary to confirm that the download request is made by the same user, and the user identifier is not limited to the user ID registered in advance. For example, when the client terminal 30 accesses the image search server 20, identification information is given and processing is performed using the same identification information until the session is disconnected, and this is used instead of the user ID in the download history data 920. Identification information may be recorded. In the association deletion stage, it may be determined whether or not a download request has been made by the same user using this identification information. In this way, the service according to the present invention can be provided to the user without giving a user ID in advance. Further, in this case, even if the download date is not recorded, the processing can be performed assuming that the interest is shown on the same day.

Further, when a new search source image is input from the client terminal 30 and a similar image search request is made, identification information may be newly added. In this way, when the same user performs a search by inputting a plurality of search source images on the same day, the download history data 920 can be recorded with different identification information for each search source image. For this reason, images that are interested based on the same conditions are recorded with the same identification information, and if they are based on different conditions, different identifications can be made even if a download request is made by the same user. Information can be added and recorded. Therefore, it is possible to delete the association for images that have not been selected by the same user under the same conditions, and the accuracy when deleting the association can be further increased.

  In the first to third embodiments, the search source image is input from the client terminal 30 or the search source image is selected, and the image search server 20 performs a similar image search based on the input. Alternatively, a computer terminal such as a personal computer used by the user may include each unit of the present invention, and manage image data owned by the user using the personal computer of the user. In this way, when managing image data owned by a user on a personal computer or the like, various images can be obtained based on a similar image search using a feature amount based on a search source image to search for a similar image. Can be provided.

The system schematic of one Embodiment of this invention. Explanatory drawing of the data memorize | stored in the search data memory | storage part. Explanatory drawing of the display screen displayed on the client terminal. Explanatory drawing of the display screen displayed on the client terminal. Explanatory drawing of the display screen displayed on the client terminal. Explanatory drawing of the display screen displayed on the client terminal. Explanatory drawing of the process sequence of one Embodiment of this invention. Explanatory drawing of the data memorize | stored in the simultaneous extraction log | history data storage part. Explanatory drawing of the data memorize | stored in the download history data storage part. Explanatory drawing of the process sequence of 3rd Embodiment.

Explanation of symbols

  21 ... Management computer as computer, 22 ... Image data storage unit as image data storage unit, 23 ... Search data storage unit as association information storage unit, 91 ... Simultaneous extraction history data storage unit as simultaneous extraction history storage unit, 92: Download history data storage unit as selection history storage means.

Claims (12)

By using the image data storage means for storing image data, and the association information storage means for storing the association information of the image data to be similar, and the selection history storage means, the image management system comprising a computer for managing them, the user An image management method for managing images to be provided,
The computer is
A first step of accepting designation of a search source image from the user;
A second step of extracting from the image data storage means image data of a similar image having a feature amount approximate to the feature amount of the search source image;
A third step of extracting, from the image data storage unit, image data of the related image related to the similar image using the association information recorded in the association information storage unit;
A fourth step of presenting the extracted image data of the similar image and the related image to the user, and storing association information between the image selected by the user and the search source image in the association information storage means. When,
A step of storing, in the selection history storage unit, selection identification information for specifying that the selection is performed under the same condition by the same user in association with the selected image when an image is selected in the fourth step;
Performing , using the selection history storage unit, deleting association information about an image that has not been selected under the same condition by the same user within a predetermined period from the association information storage unit. An image management method characterized by the above. The first stage accepts a search source image from the user,
The image management method includes:
The computer is
Calculating a feature amount of the search source image;
The image management method according to claim 1, further comprising: storing image data of the search source image in the image data storage unit. The image management method according to claim 1, wherein the feature amount includes at least a feature amount related to any one of an overall color, a color distribution, a contour, and a pattern. The image management system includes simultaneous extraction history storage means,
The image management method includes:
The computer is
When a plurality of similar images are extracted in the second stage, the plurality of similar images are associated with each other and stored together with the extraction time information in the simultaneous extraction history storage unit;
2. The method according to claim 1, further comprising the step of deleting, from the association information storage unit, association information about images that have not been extracted simultaneously within a predetermined period using the simultaneous extraction history storage unit. The image management method according to any one of? The image management method according to claim 1, wherein the selection identification information is a user identifier and a selection time of a user who has selected the image. By using an image management system including an image data storage unit that stores image data, an association information storage unit that stores association information of similar image data, a selection history storage unit, and a computer that manages these, An image management program for managing images to be provided,
  The computer,
  First means for accepting designation of a search source image from the user;
  Second means for extracting, from the image data storage means, image data of a similar image having a feature amount approximate to the feature amount of the search source image;
  Third means for extracting the image data of the related image related to the similar image from the image data storage means using the association information recorded in the association information storage means;
  Fourth means for presenting the extracted image data of the similar image and the related image to the user, and storing association information between the image selected by the user and the search source image in the association information storage means. When,
  Means for storing, in the selection history storage means, selection identification information for specifying that the selection is made by the same user under the same condition when an image is selected by the fourth means, in association with the selected image;
  Means for deleting, from the association information storage means, association information for images that have not been selected by the same user under the same conditions within a predetermined period using the selection history storage means
Image management program to function as Causing the first means to function as a means for receiving a search source image from the user;
  Said computer further
  Means for calculating a feature amount of the search source image;
  Means for storing image data of the search source image in the image data storage means
The image management program according to claim 6, wherein the image management program is made to function as: The image management program according to claim 6 or 7 , wherein the feature amount includes at least a feature amount related to any of an overall color, a color distribution, a contour, and a pattern . The image management system includes simultaneous extraction history storage means, and uses the simultaneous extraction history storage means.
And
Said computer further
Means that when a plurality of similar images are extracted by the second means, the plurality of similar images are associated with each other and stored together with the extraction time information in the simultaneous extraction history storage unit;
Means for deleting from the association information storage means association information about images that have not been extracted simultaneously within a predetermined period using the simultaneous extraction history storage means;
9. The image management program according to claim 6, wherein the image management program is made to function as: The image management program according to any one of claims 6 to 9, wherein the selection identification information includes a user identifier and a selection time of a user who has selected the image. An image data storage unit that stores image data, an association information storage unit that stores association information of similar image data, a selection history storage unit, and a computer that manages them, and manages images to be provided to the user An image management system,
  The computer is
  First means for accepting designation of a search source image from the user;
  Second means for extracting, from the image data storage means, image data of a similar image having a feature amount approximate to the feature amount of the search source image;
  Third means for extracting the image data of the related image related to the similar image from the image data storage means using the association information recorded in the association information storage means;
  Fourth means for presenting the extracted image data of the similar image and the related image to the user, and storing association information between the image selected by the user and the search source image in the association information storage means. When,
  Means for storing, in the selection history storage means, selection identification information for specifying that the selection is made by the same user under the same condition when an image is selected by the fourth means, in association with the selected image;
  Means for deleting, from the association information storage means, association information about images that have not been selected by the same user under the same condition within a predetermined period using the selection history storage means;
An image management system comprising: The image management system according to claim 11, wherein the selection identification information is a user identifier and a selection time of a user who has selected the image.

Images