JP2009181468A - Image search log collection system, image search log collection method and program - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To collect information showing a potential need for a commodity of a user who is considering purchase of the commodity from a commodity sales site. <P>SOLUTION: The image search log collection system connected to a user terminal comprises: a storage means which stores images of commodities and commodity identification information of the commodities; a transmitting means which transmits a list of commodity images to the user terminal; a receiving means which receives from the user terminal a similar image search request containing a selected image selected from the list of commodity images and parameter information; a similar image search means which performs similar image search to the commodity images stored in the storage means, and transmits the search result to the user terminal; and a log collection means which stores the commodity identification information corresponding to the selected image contained in the similar image search request and the parameter information in a log storage means as a log. <P>COPYRIGHT: (C)2009,JPO&INPIT

Description

  The present invention relates to a technology that makes it possible to collect information about potential needs of a user's product using an image of the product provided by a product sales site on a communication network.

  Establishing a product sales site on a communication network such as the Internet and selling products to users is widely used as an existing technology.

  In such an existing technology, a user terminal connected to a communication network accesses a product sales site, a user performs a product search or the like via the user terminal, selects a desired product, and is determined at the product sales site. The product is purchased by performing the purchase operation. In the process related to such product purchase, the product sales site can leave a history (sales history, click history) of what products are purchased from the user terminal through what kind of page browsing.

  By using such history and user registration information, the merchant of the product purchases which product purchased what product (sales performance trend) and what page was viewed. (Product browsing lead), purchase product ranking by terminal and time zone, etc. can be grasped, and these can be utilized for the next product development.

Note that Patent Documents 1 and 2 are prior art documents related to the present application.
Japanese Patent Laid-Open No. 2005-133228 JP 2004-220267 A

  In general, users who access product sales sites have needs for product colors, shapes, etc., and search for products that meet those needs from the product group displayed on the product sales site and consider purchasing. To do.

  Some users who access the product sales site purchase products that meet their needs, while there are no products that meet their needs, but users who purchase such products because they need them. In addition, since there is no product that meets the needs, it is considered that there is a user who abandons the purchase of the product from the product sales site.

  If each of the above-mentioned users who access the product sales site can grasp what color and shape the product was searched for at the stage of considering the purchase, it will be very useful in the next product development etc. Helpful. In other words, it is thought that the next product development etc. can be performed more effectively by grasping the potential needs (referred to as potential needs) of the product color and shape shown before the user purchases the product. .

  However, in the above-described conventional technology, only the sales data and history data of the browsing conductor of the user who made the purchase can be collected, and information on the latent needs for the user's products cannot be collected. For this reason, in the prior art, it was not possible to develop the next product in consideration of the potential needs of users.

  The present invention has been made in view of the above points, and makes it possible to collect information representing potential needs of a user who is considering purchasing a product from a product sales site on a communication network. The purpose is to provide technology.

  The above problem is an image search log collection system that performs a similar image search based on a product image received from a user terminal connected via a communication network and collects information used in the similar image search as a log. Product information storage means for associating and storing at least an image of the product and product identification information of the product, and a list of product images acquired from the product information storage means to the user terminal via the communication network A product image transmitting means for transmitting, a similar image search request receiving means for receiving a similar image search request including a selected image selected from a list of product images and parameter information from the user terminal, and the similar Products stored in the product information storage means using the selected image and the parameter information included in the image search request Similar image search means for performing a similar image search on an image and transmitting a search result to the user terminal via the communication network, identification information of a product corresponding to the selected image included in at least the similar image search request, and the This can be solved by an image search log collection system comprising log collection means for storing parameter information as a log in the log storage means.

  In the image search log collection system, when the user terminal accesses the product sales site corresponding to the product selected from the search result transmitted to the user terminal by the similar image search unit, The log collection unit may store information indicating that the product sales site has been accessed in the log storage unit in association with the product identification information.

  Further, the parameter information is information indicating a color, and the similar image search means may search for an image of a product having at least a color similar to the color indicated by the parameter information, and similar as the parameter information. The information indicating that the shape is important in the image search may be used, and the similar image search means may search for an image of a product having a shape similar to the shape of the product indicated by the selected image.

  Further, the parameter information is information indicating a partial area selected by the user from the selected image, and the similar image search unit is an image of a product having a part similar to the part indicated by the image corresponding to the partial area. It is good also as searching.

  The parameter information includes information indicating a partial area selected by the user from the selected image, and a partial image combined with the partial area instead of the image of the partial area in the selected image. The image search means may search for an image similar to an image formed by combining the partial image with the partial area of the selected image.

  Further, when the partial area is used, the selected image is divided into a plurality of unit areas, and the frequency at which the unit area is selected as the partial area for each unit area is stored in the log data stored in the log storage unit. Further, an analysis result output unit that assigns a color corresponding to the frequency to each unit region and displays the color assigned to the unit region on the selected image may be further provided.

  The present invention is also an image search log collection system connected to a user terminal and a product sales site providing system via a communication network, receiving an image of the product from the product sales site providing system, and an image of the product Product information registration means for storing the product identification information of the product in the product information storage means, a selection image that is an image of the product selected by the user from a list of product images, and parameter information. A similar image that is received from the image, performs a similar image search on the product image stored in the product information storage means using the selected image and parameter information, and transmits the search result to the user terminal via the communication network. Search means, at least product identification information corresponding to the selected image received from the user terminal, and the parameter. It is also possible to configure the information as an image search log acquisition system characterized by comprising a logging means for storing in the log storage unit as a log.

  Furthermore, the present invention can be configured as an image search log collection method in each of the above image search log collection systems, and can also be configured as a program for causing a computer to realize each means in the above image search log collection system.

  ADVANTAGE OF THE INVENTION According to this invention, it becomes possible to collect the information showing the latent needs with respect to the goods of the user who is considering purchasing goods from the goods sales site on a communication network.

  Next, embodiments of the present invention will be described with reference to the drawings.

(System configuration)
FIG. 1 shows the overall configuration of the marketing system in the present embodiment. As shown in FIG. 1, the marketing system includes a user terminal 1, an image search log collection system 3, and a merchandise sales site providing system 5, each of which is connected to a communication network 7 such as the Internet. In FIG. 1, one user terminal 1 and one product sales site providing system 5 are shown, but a plurality of user terminals 1 and one product sales site providing system 5 may exist. Not too long. The outline of the processing in this marketing system is as follows.

  In this marketing system, as pre-processing, a product image is first registered in the image search log collection system 3 from the product sales site providing system 5 side, and the image search log collection system 3 is used for similar image search. The quantity is extracted for each registered product image, and the extracted feature quantity is stored in the database in association with the product. In this specification, “image” is used to mean “image data”.

  Then, the user who is considering purchasing the product accesses the image search log collection system 3 from the user terminal 1 and selects a product of interest from the product images provided from the image search log collection system 3. A search for a product image similar to a product image is requested to the image search log collection system 3 by specifying specific parameter information, and a search result is obtained. Then, if there is a desired product in the search result, the user accesses the product sales site providing system 5 that sells the product from the user terminal 1 and purchases the product. The image search log collection system 3 collects information about product images and parameter information used in the above-described similar image search as a log, and stores it in a database. This log is data that is utilized as an indication of the potential needs of the shape and color of the user's product.

  Next, each apparatus which comprises the marketing system shown in FIG. 1 is demonstrated in detail.

  FIG. 2 is a functional configuration diagram of the image search log collection system 3. As shown in FIG. 2, the image search log collection system 3 includes a user connection interface unit 31, a similar image search unit 32, an image feature amount extraction unit 33, a product information registration unit 34, a log collection analysis unit 35, and a log information provision unit. 36, a category DB 37 (DB is an abbreviation of a database), a site DB 38, a feature DB 39, a product DB 40, a log DB 41, and a user DB 42.

  The user connection interface unit 31 is a functional unit that provides various screens to the user terminal 1 and performs data transmission / reception with the user terminal 1. The user connection interface unit 31 also has a function of searching various DBs and acquiring data corresponding to the request. The user connection interface unit 31 has a Web server function, for example, and performs data communication based on http with the user terminal 1 in which a Web browser is installed. In addition, the user connection interface unit 31 may include an e-mail transmission / reception function. In this case, the user connection interface unit 31 transmits / receives data to / from the user terminal 1 by e-mail.

  The similar image search unit 32 is a functional unit that searches the feature amount DB 39 based on a similar image search request received from the user terminal 1 via the user connection interface unit 31 and acquires a search result. The image feature amount extraction unit 33 is a functional unit that extracts a feature amount used for similar image search from the product image received from the product sales site providing system 5 and stores it in the feature amount DB 39.

  The similar image retrieval technique itself is an existing technique. As the feature quantity, for example, a feature quantity related to color, a feature quantity related to shape (contour), and a feature quantity related to pattern are extracted and stored in the feature quantity DB 39 together with the product ID. In the similar image search, feature amounts similar to those described above are extracted from the image that is the basis of the search (referred to as a key image), and a product ID corresponding to a feature amount close to the feature amount of the key image is extracted. Will be. In addition, when comparing feature quantities, it is possible to specify in advance which feature quantity is important. For example, in the case of emphasizing color, a key image and a specific color are given. In particular, the feature amount of the color is close to the feature amount of the specific color, and other feature amounts are close to some extent (predetermined extent). Things are extracted as search results. Further, if importance is attached to the shape, a product having a feature amount that is particularly close to the feature amount of the key image shape is selected. There are various similar image search techniques, but the type of the similar image search technique is not particularly limited in the present invention, and the search can be performed with emphasis on parameters such as color and shape as described above. Anything is acceptable.

  The product information registration unit 34 has a function of transmitting / receiving data to / from the product sales site providing system 5, a function of performing authentication processing for the product sales site providing system 5, and a category name and a product name from the product sales site providing system 5. It is a functional unit having a function of receiving an image and a URL indicating a Web page corresponding to the product and storing them in the product DB 40.

  The log collection / analysis unit 35 performs a similar image search after the user terminal 1 accesses the image search log collection system 3, selects a product, and uses the images and images used until the product sales site providing system 5 is accessed. This is a functional unit having a function of collecting various information such as parameter information as a log and storing them in the log DB 41 and a function of performing a data analysis process using the log stored in the log DB 41.

  The log information providing unit 36 has a function of transmitting log information stored in the log DB 41 to the external device based on a request received from the external device (for example, the merchandise sales site providing system 5) via the communication network 7, an external device The log collection / analysis unit 35 is instructed to perform log analysis processing based on the request received from the log, and the log analysis processing result executed by the log collection / analysis unit is received and provided to an external device.

  FIG. 3 shows an example of data stored in the category DB 37. As shown in FIG. 3, the category DB 37 stores a category ID (eg, “0001”) and a category name (eg, “shoes”) in association with each other. FIG. 4 shows an example of data stored in the site DB 38. As shown in FIG. 4, the site DB 38 stores the merchandise sales site ID, password, and site name. These pieces of information are stored in advance when registering the merchandise sales site. Each DB in the image search log collection system 3 is realized by data storage means (memory, hard disk, etc.) in the computer.

  FIG. 5 shows an example of data stored in the product DB 40. As shown in FIG. 5, the product DB 40 stores a product ID, a category ID, an image, and a site URL in association with each other. The site URL is a URL indicating a page on which an image of the corresponding product ID is posted. Further, as shown in FIG. 6, the feature amount DB 39 stores the product ID and the feature amount in association with each other.

  FIG. 7 shows basic information recorded in the log DB 41. As shown in FIG. 7, the log DB 41 is connected to a user ID, a category ID, a product ID, information about parameters used in similar image search, information about images used in similar image search, access time, and a product sales site ( Hereinafter, information indicating whether or not a link has been made is recorded. The information stored in the log DB 41 varies depending on the type of similar image search, and details thereof will be described in a specific example of log collection analysis processing described later. Although not shown, the user DB 42 stores data including a user ID and a password.

  The image search log collection system 3 can be realized by using a single computer (so-called server) having a storage device such as a CPU, a memory, and a hard disk, and each functional unit or a group of a plurality of functional units. Can be realized by using different computers, and each computer can be connected by a communication network.

  When the image search log collection system is realized by one computer or by a plurality of computers, the function of each computer is realized by causing the computer to execute a program for realizing the function. It is what is done. The program can be installed in the computer from a recording medium such as a portable memory, or can be downloaded from an external device via a communication network and installed.

  The user terminal 1 is a general PC including a web browser and a mailer. A mobile terminal such as a mobile phone can also be used as the user terminal 1. The product sales site providing system 5 is a system for providing a general product sales site including, for example, a Web server, a database server, and a management terminal for providing the product sales site. Each of the Web server, database server, management terminal, and the like is connected to a communication network.

(System operation)
Next, the operation of the image search log collection system 3 will be described with reference to sequence charts shown in FIGS.

  First, an operation when a product image is registered from the product sales site providing system 5 to the image search log collecting system 3 will be described with reference to FIG.

  For example, a log-in request including an ID and a password is transmitted from the management terminal in the product sales site providing system 5 to the product information registration unit 34 (step 1).

  The product information registration unit 34 that has received the login request searches the site DB 38 using the ID and password, and if it detects that an ID and password that match the ID and password are registered in the site DB 38, the login is OK. It is transmitted to the product sales site providing system 5 (steps 2 and 3). If the login is NG, a message refusing the subsequent access is transmitted to the product sales site providing system 5.

  Next, the product information registration unit 34 acquires a list of category names from the category DB 37 (step 4), and transmits it to the product sales site providing system 5 (step 5). Subsequently, an image of the product registered in the image search log collection system 3, its category name, and a URL corresponding to the image are transmitted from the product sales site providing system 5, and the product information registration unit 34 receives the data. (Step 6). The merchandise information registration unit 34 searches the category DB 37 using the received category name, thereby acquiring a category ID corresponding to the category name (step 7). The product information registration unit 34 determines a product ID corresponding to the product image received from the product sales site providing system 5 (step 8). The product ID can be determined using, for example, a serial number.

  Subsequently, the product information registration unit 34 transmits an image feature amount extraction request including the image received from the product sales site providing system 5 and the product ID to the image feature amount extraction unit 33 (step 9). The image feature amount extraction unit 33 that has received the image feature amount extraction request extracts the feature amount used for the similar image search process from the image, and stores the feature amount in the feature amount DB 39 in association with the product ID (step 10). . Further, the product information registration unit 34 stores the product ID, category ID, image, site URL, and the like in the product DB 40 in association with each other (step 11).

  Next, processing related to a similar image search based on a request from the user terminal 1 will be described with reference to FIG. As a premise of the processing shown in FIG. 9, user authentication based on the user ID and password sent from the user terminal 1 to the image search log collection system 3 is performed using registration data stored in the user DB 42, and the user authentication is successful. Suppose you are. In addition, the user connection interface unit 31 stores an identifier (address or the like) of the user terminal 1 that has been successfully authenticated in the user DB 42, and an identifier that is attached to data that is subsequently transmitted from the user terminal 1 in the session. Thus, it is assumed that the user ID related to the subsequent communication with the user terminal 1 can be acquired from the user DB 42 as necessary. The data transmitted from the user terminal 1 may be given a user ID, thereby acquiring the user ID.

  In FIG. 9, first, a category name desired by the user is input to the user terminal 1, and is transmitted from the user terminal 1 to the user connection interface unit 31 (step 21). Receiving the category name, the user connection interface unit 31 acquires a category ID corresponding to the category name from the category DB 37 (step 22). The user connection interface unit 31 transmits the current time, user ID, and category ID to the log collection analysis unit 35, and the log collection analysis unit 35 holds the received current time, user ID, and category ID (step). 23).

  Subsequently, the user connection interface unit 31 obtains a list of product images (a collection of images obtained by reducing product images, including product IDs corresponding to the images) from the product DB 40 (step 24). It transmits to the user terminal 1 (step 25). Further, screen information including information requesting parameter input to the user terminal 1 is sent from the user connection interface unit 31 (step 26), and the user terminal 1 is provided with a list of product images and information requesting parameter input. A screen is displayed.

  Note that the parameters in the present embodiment are information indicating the values (color, shape, region, etc.) that are important in similar image search, and will be described in detail later.

  A user who browses the screen displayed on the user terminal 1 selects a product image to be used for a similar image search from the product image list and sets search parameters. Then, the user terminal 1 transmits a similar image search request including the selected image (selected image), a product ID corresponding to the selected image, and parameter information to the user connection interface unit 31 (step 27).

  In step 25, after the product image list is displayed on the user terminal 1 and a certain image is selected by the user, the image search log collection system 3 that has received the selected image information is the product image selected by the user and reduced. The product image that has not been sent and the parameter input request information may be transmitted to the user terminal 1, and the product image that has not been reduced in the user terminal 1 and the parameter input request information may be displayed. Thereafter, after the parameters are set in the user terminal 1, a similar image search request including the same information as the information transmitted to the user connection interface unit 31 in step 27 of FIG. 9 is sent to the user connection interface unit 31. Of course, it is also possible to perform processing in a procedure other than the procedure shown here.

  Upon receiving the similar image search request, the user connection interface unit 31 transmits the similar image search request to the similar image search unit (step 28), and logs the product ID, parameter information, and selected image included in the similar image search request. The information is transmitted to the collection analysis unit 35 (step 29), and the log collection analysis unit 35 holds the information together with the information received in step 23.

  The similar image search unit 32 that has received the similar image search request is similar to the image feature amount extraction processing for an image that becomes a search key (a composite image that will be described later such as a selected image may be a search key). Image feature amount extraction processing is performed to extract feature amounts, feature amounts close to the feature amounts are searched from the feature amount DB 39, and product IDs corresponding to the searched feature amounts are acquired (steps 30 and 31). Note that parameter information is taken into account when searching the feature amount DB 39. As described above, for example, if the parameter information is information indicating a specific color, the product ID corresponding to the feature amount that is particularly close to the feature amount of the color is extracted.

  The similar image search unit 32 acquires a product ID corresponding to one or a plurality of images as a search result from the feature amount DB 39 (step 31), and acquires a product image, URL, and the like corresponding to those product IDs from the product DB 40. (Step 32), information including the product ID, product image, and URL is passed to the user connection interface unit 31 as a similar image search result (Step 33). The user connection interface unit 31 transmits the similar image search result received from the similar image search unit 32 to the user terminal 1 (step 34).

  When a user who browses a list of product images as a similar image search result on the user terminal 1 selects a specific product image from the list, information indicating the fact and the URL are the user connection interface unit 31. (Step 35). The user connection interface unit 31 that has received the information transmits information indicating that the product sales site 5 is connected to the log collection analysis unit 35 (step 36). The log collection / analysis unit 35 stores the information received in Step 23, Step 29, and Step 36 in the log DB 41 (Step 37).

  Further, the user connection interface unit 31 transmits an instruction for connecting to the URL to the user terminal 1 (step 38), and the user terminal 1 that has received the instruction provides the merchandise sales site indicated by the URL. The system 5 is accessed (step 39).

  After step 34, when a specific product is not selected from the list of product images that are similar image search results and logout or the like is performed, the user connection interface unit 31 detects that logout when it detects the logout. The log collection / analysis unit 35 notifies the log collection / analysis unit 35, and the log collection / analysis unit 35 that has received the notification stores the information held in step 23 and step 29 and the information indicating that the product sales site 5 is not connected in the log DB 41.

  Next, specific examples of log collection processing and analysis processing using logs will be described.

(Log collection analysis processing example 1)
This example is an example in which similar image search is performed with emphasis on color. That is, the color selected by the user is parameter information.

  In the present example, following step 25 shown in FIG. 9, after one image is selected from the product image list on the user terminal 1, for example, the screen shown in FIG. 10 is displayed on the user terminal 1. Here, “color” is selected as the search method. When a color is selected, for example, a screen including a color palette shown in FIG. 11 is displayed. The user selects a desired color from the color palette as the color of the product displayed on the left side.

  Thereafter, a color-oriented similar image search request including information on the selected image and the color selected by the user is sent to the image search log collection system 3. The image search log collection system 3 that has received the color-oriented similar image search request searches the feature amount DB 39 for a product that is similar to the shape of the selected product and that has a color similar to the selected color. To extract. In this search, the search is performed with a color similarity higher than the shape similarity.

  The result of similar image search is transmitted to the user terminal 1, and if there is an image of a favorite product from among them, the user connects to the product sales site and proceeds to purchase the product.

  FIG. 12 shows an example of log data recorded in the log DB 41 when this process is performed by a plurality of products and a plurality of users. The log data shown in FIG. 12 relates to a specific category. Similarly, log data shown in the following examples relates to a specific category.

  The log data shown in FIG. 12 includes a product ID, color name, color code, time, user ID, presence / absence of sales site link, and product image. Such log data represents the latent needs indicated in the process of selecting a product by the user. For example, the analysis result shown in FIG. 13 can be obtained by tabulating the log data in this example using products as keys.

  In the example of the analysis result shown in FIG. 13, for a product with a product ID of 001, green is selected in 70% of the number of similar image searches that emphasize color, blue is selected in 20%, and 10%. It is shown that white is selected. In addition, for the product with the product ID 001, 10% of the number of times the similar image search with emphasis on color is made is shown in the morning, 30% is in the afternoon, and 60% is in the night. ing. Furthermore, it is shown that 70% of the number of times the similar image search with emphasis on color is performed for the product with the product ID 001 is linked to the sales site. From such an analysis result, for example, for a product with a product ID of 001, it can be determined that there is a high potential need for a green product with a product ID of 001.

  Further, by collecting the log data in this example using the color as a key, for example, the analysis result shown in FIG. 14 can be obtained. In the example shown in FIG. 14, it is shown that the search is performed by selecting green in 70% of the number of times that the similar image search with emphasis on color is performed. Further, it is indicated that the product ID of the target product for which the search is performed with green selected is 001, 003, 005, and 80% of the number of times the search is performed with green selected is in the morning 0% was done in the afternoon and 20% was done in the evening. Further, it is shown that the link to the merchandise sales site is made in 10% of the number of times that the search is performed with green selected.

  From the analysis results shown in FIG. 14, for example, regarding the products in this category, there is a high need for green, but it can be seen that the rate of links to product sales sites is high in blue.

  The log data analysis as described above may be performed by the log collection / analysis unit 35 that has received an analysis request from an external device (product sales site providing system 5 or the like), and the result may be sent to the external device. Then, the image search log collection system 3 may transmit only log data to the external device, and the external device may perform analysis processing. The same applies to the examples described below. In addition to the above-described example, the analysis process can be performed from various viewpoints.

(Log collection analysis processing example 2)
This example is an example in which similar image search is performed with emphasis on shape. That is, the information indicating that the shape is important is the parameter information. In this example, the screen shown in FIG. 10 is displayed on the user terminal 1, and “form” is searched as a search method, whereby the screen shown in FIG. 15 is displayed, for example. Here, by shifting the button shown in FIG. 15 in the direction of high similarity, an image having a high degree of similarity between the selected product and the shape is searched. If the button is shifted in the direction of low similarity, the shape is similar. Those with a low degree will also be searched. In this example, the color does not need to be similar to the selected product image, but a color similar to a certain degree can be selected.

  In this example, a numerical value indicating the degree of similarity corresponding to the position of the button is sent from the user terminal 1 to the image search log collection system 3 together with information that importance is placed on the shape. Accordingly, similar image search processing is performed.

  In this example, log data shown in FIG. 16 is recorded in the log DB 41, for example. This log data includes a product ID, information indicating importance of shape including a degree of similarity (for example, indicated by a numerical value of 1 to 10), a user ID, information on presence / absence of a link to a product sales site, and a product image. Including.

  For example, the analysis result shown in FIG. 17 can be obtained by tabulating the log data in this example using products as keys.

  In the example shown in FIG. 17, 30% of the number of times that the similar image search with emphasis on the shape is performed is performed on the product with the product ID 001. According to the analysis result shown in FIG. 17, for example, since a similar image search that emphasizes more shapes has been performed on the product with the product ID 002, the product ID is 002 rather than the shape of the product with the product ID 001. It can be predicted that the shape of the product is the preferred shape.

(Log collection analysis processing example 3)
In this example, the user selects a part of the selected image and searches for an image of a product having a part similar to that part. In this example, the area information of the selected part is parameter information.

  In this example, when “part” is selected on the screen shown in FIG. 10, for example, as shown in FIG. 18, a screen on which a desired part in the selected image can be selected is displayed on the user terminal 1. A selection of parts is made.

  After the portion is designated, a similar image search request including the selected image and region information of the selected portion (coordinate information indicating the region of the portion) is sent to the image search log collection system 3. The image search log collection system 3 that has received the similar image search request searches for a product having a portion similar to the selected portion in the received selected image, and returns the search result to the user terminal 1.

  FIG. 19 shows an example of log data stored in the log DB 41 when this process is performed by a plurality of products and a plurality of users. As shown in FIG. 19, in this log data, a product ID, partial area information, time, user ID, presence / absence of a link to a product sales site, and a product image are recorded.

  The concept of an example of log data analysis method for a certain product (product ID 001) in this example will be described with reference to FIG.

  The product image is divided into a predetermined number of divisions (for example, 25 divisions, 36 divisions, etc., hereinafter referred to as 36 divisions as an example). Then, for each divided area (referred to as a unit area), the number of times selected as a part is totaled, a color is assigned to the unit area according to the ratio of the selected number of times to the total number of times, and Display with color. For example, if the number of selected times is 70% or more of the total number of times (number of records corresponding to the product), red is assigned to the unit area, pink is assigned if the unit area is 50 to 70%, and 20% If it is below, assign gray and display. Other parts are not colored. In FIG. 20, diagonal lines correspond to red, and shading corresponds to gray. In the analysis result display of FIG. 20, colors can be displayed in addition to the hatched and shaded portions, but they are omitted in FIG. By viewing such a display screen, the merchandise sales site provider can grasp at a glance which part has been selected or not selected, and the part of the product that is of interest to the user (latent Needs) at a glance. Thereby, for example, it is possible to predict that the part of the product corresponding to the region that is frequently selected as a part is a part having a shape and color preferable for the user.

  FIG. 21 shows an example of a flowchart of processing when the log collection / analysis unit 35 performs this analysis processing in response to a request from the product sales site providing system 5. The process shown in FIG. 21 shows a process in the case where a product to be analyzed is designated and analysis is performed on the specific product.

  The log collection analysis unit 35 reads partial area information of the first record corresponding to the product from the log data stored in the log DB 41 (step 51). Then, it is determined which of the 36 sub-regions the partial region covers (each unit region is identified by a number from 1 to 36) (step 52). Here, for example, if the selected partial area covers a predetermined ratio (for example, 80%) or more of the unit areas, the partial area covers the unit area.

  Then, a variable M (i) (i = 1 to 36) having an initial value of 0 is prepared, and 1 is added to M (j) for each unit area number j covered by the partial area to obtain a new M (j (Step 53). That is, for example, if the partial area in the first record covers the unit areas 2, 3, and 4, M (2) = 1, M (3) = 1, and M (4) = 1.

  Subsequently, it is determined whether or not the current record is the last record for the analysis target product (step 54). If it is not the last record, the process proceeds to the next record (step 55), and the processes of steps 52 and 53 are performed.

If it is determined in step 54 that the record is the last record, the color to be displayed in each unit area is determined according to the size of M (i) (i = 1 to 36) for each unit area (step 56). For example, if the number of records related to the product is 100 and M (10) is 80, the unit area 10 is selected in 80% of the number of partial selections made. Here, for example, if the selected number is 70% or more, red is assigned to the unit area, pink is assigned if it is 50 to 70%, and gray is assigned if it is 20% or less. In this case, red is assigned to the unit area 10. In this way, the colors are determined for all the unit areas 1 to 36, and the result as shown in FIG. 20 is output to the merchandise sales site providing system 5 via the communication network (step 57).
(Log collection analysis processing example 4)
In this example, the image portion of the product prepared by the user is combined with the image portion of the selected product, and a similar image search is performed using the image obtained by the combination.

  In this example, when “combination” is selected on the screen shown in FIG. 10, for example, as shown in FIG. 22, a product image and an image held by the user are displayed, and a part of the image held by the user is displayed. A screen is displayed that allows the user to cut out and combine it with a portion of the product image.

  In the user terminal 1, after the combination image is created, the selection image, the additional image, the composite image, and the similarity including the region information indicating the portion where the additional image is combined in the selection image (the coordinate information indicating the region of the portion) An image search request is sent to the image search log collection system. In this case, the area information indicating the portion where the additional image is combined with the additional image, the composite image, and the selected image is the parameter information. Further, as the parameter information, the additional image and area information indicating a portion where the additional image is combined in the selected image may be transmitted. This is because if there is an additional image and the region information, a composite image can be created on the image search log collection system 3 side and a similar image search using the composite image can be performed.

  The image search log collection system 3 that has received the similar image search request stores the received information in the log DB 41 as a log, searches for a product similar to the received composite image, and returns the search result to the user terminal 1.

  When this process is performed by a plurality of products and a plurality of users, for example, log data shown in FIG. 23 is recorded in the log DB 41.

As shown in FIG. 23, this log data includes the product ID, time, partial area information, product image, additional image, composite image, and presence / absence of a link to the product sales site.
The concept of an example of a log data analysis method for a certain product (product ID 001) in this example is the same as that shown in FIG. As in the case of the log collection analysis process 3, the product image is divided into a predetermined number of divisions (for example, 25 divisions, 36 divisions, etc., hereinafter referred to as 36 divisions as an example). Then, for each divided unit area, the number of times selected as an additional portion in the selected image is totaled, and the color of the area is assigned according to the ratio of the selected number to the total number of times, and the color is assigned to the area. Attached and displayed.

  Accordingly, it is possible to grasp at a glance a region where the number of times of selection is high, and for example, it is possible to predict that the shape of the part of the product corresponding to the portion where the number of times of selection is high is an undesirable shape for the user.

  The contents of this process are the same as the contents of the log collection analysis process 3 shown in the flowchart of FIG. 21, and the displayed results are also the same. However, in this example, the region indicated by red (a part that is selected a lot) in the displayed result is a part that has been replaced with another part, so that it can be predicted to be an unfavorable part for the user. .

(Effect of this embodiment)
As described above, in the image search log collection system 3 according to the present embodiment, the product search is performed using the similar image search technology, so that the user can search various image lists without considering the search word or the like. A product search can be performed using an image selected from the list.

  In addition, by using the image search log collection system 3 according to the present embodiment, it is possible to collect latent needs indicating what color and shape of the product the user likes. In addition, it is possible to collect not only the potential needs of the user who has reached the desired product and purchased the product, but also the potential needs of the user who is willing to purchase, but has given up without reaching the target product. In addition, the user can collect information on which part of the product (the color, shape, pattern, etc.) of the product is interested in the product and where the user feels dissatisfied with the product.

  The present invention is not limited to the above-described embodiment, and various modifications and applications can be made within the scope of the claims.

1 is an overall configuration diagram of a marketing system in an embodiment of the present invention. 3 is a functional configuration diagram of an image search log collection system 3. FIG. It is a figure which shows the example of the data stored in the category DB37. It is a figure which shows the example of the data stored in site DB38. It is a figure which shows the example of the data stored in goods DB40. It is a figure which shows the example of the data stored in feature-value DB39. It is a figure which shows the basic information stored in log DB41. 6 is a sequence chart for explaining the operation of the image search log collection system 3. 6 is a sequence chart for explaining the operation of the image search log collection system 3. 4 is a diagram illustrating an example of a screen displayed on the user terminal 1. FIG. It is a figure which shows the example of the screen containing a color palette. It is a figure which shows the example of the log data in the log collection analysis process example 1. FIG. It is a figure which shows the example of the analysis result in the log collection analysis process example 1. FIG. It is a figure which shows the example of the analysis result in the log collection analysis process example 1. FIG. It is a figure which shows the example of the screen displayed in the log collection analysis process example 2. FIG. It is a figure which shows the example of the log data in the log collection analysis processing example 2. It is a figure which shows the example of the analysis result in the log collection analysis process example 2. FIG. It is a figure which shows the example of the screen displayed in the log collection analysis process example 3. FIG. It is a figure which shows the example of the log data in the log collection analysis process example 3. FIG. It is a figure for demonstrating the concept of the analysis method in the log collection analysis process example 3. FIG. 14 is a flowchart showing an analysis process in a log collection analysis process example 3; It is a figure which shows the example of the screen displayed in the log collection analysis process example 4. FIG. It is a figure which shows the example of the log data in the log collection analysis processing example 4.

Explanation of symbols

DESCRIPTION OF SYMBOLS 1 User terminal 3 Image search log collection system 5 Product sales site provision system 7 Communication network 31 User connection interface part 32 Similar image search part 33 Image feature-value extraction part 34 Product information registration part 35 Log collection analysis part 36 Log information provision part 37 Category DB
38 Site DB
39 Feature DB
40 Product DB
41 Log DB
42 User DB

Claims (12)

An image search log collection system that performs a similar image search based on an image of a product received from a user terminal connected via a communication network and collects information used in the similar image search as a log,
Product information storage means for storing at least an image of the product and product identification information of the product in association with each other;
Product image transmission means for transmitting a list of product images acquired from the product information storage means to the user terminal via the communication network;
Similar image search request receiving means for receiving a similar image search request including a selected image and parameter information selected from a list of images of the product from the user terminal;
Using the selected image and the parameter information included in the similar image search request, a similar image search is performed on the product image stored in the product information storage unit, and the search result is transmitted to the user terminal via the communication network. Similar image search means to send to
An image search log collection system comprising: log collection means for storing, in a log storage means, product identification information and parameter information corresponding to the selected image included in at least the similar image search request.   When the user terminal accesses the product sales site corresponding to the product selected from the search result transmitted to the user terminal by the similar image search unit, the log collecting unit 2. The image search log collection system according to claim 1, wherein information indicating that a site has been accessed is stored in the log storage unit in association with identification information of the product.   The parameter information is information indicating a color, and the similar image search unit searches for an image of a product having at least a color similar to the color indicated by the parameter information. Image search log collection system.   The parameter information is information indicating that importance is attached to a shape in a similar image search, and the similar image search means searches for an image of a product having a shape similar to the shape of the product indicated by the selected image. The image search log collection system according to claim 1 or 2.   The parameter information is information indicating a partial region selected by the user from the selected image, and the similar image search unit searches for an image of a product having a portion similar to the portion indicated by the image corresponding to the partial region. The image search log collection system according to claim 1 or 2, wherein   The parameter information includes information indicating a partial region selected by the user from the selected image, and a partial image combined with the partial region in place of the image of the partial region in the selected image. The image search log collection system according to claim 1 or 2, wherein the means searches for an image similar to an image formed by combining the partial image with the partial area of the selected image.   The selected image is divided into a plurality of unit areas, and the frequency at which the unit area is selected as the partial area for each unit area is calculated from the log data stored in the log storage unit, and the color corresponding to the frequency The image search log according to claim 5 or 6, further comprising: an analysis result output means for assigning to each unit area and displaying the color assigned to each unit area superimposed on the selected image. Collection system. An image search log collection system connected to a user terminal and a product sales site providing system via a communication network,
Product information registration means for receiving an image of the product from the product sales site providing system and storing the product image together with product identification information of the product in the product information storage means;
A selection image, which is an image of a product selected by the user from a list of product images, and parameter information are received from the user terminal and stored in the product information storage unit using the selected image and parameter information. A similar image search means for performing a similar image search on an image of a product that has been performed, and transmitting a search result to the user terminal via the communication network;
An image search log collection system comprising: log collection means for storing, in a log storage means, at least product identification information corresponding to the selected image received from the user terminal and the parameter information. An image search log collection system that performs a similar image search based on an image of a product received from a user terminal connected via a communication network and collects information used in the similar image search as a log. An image search log collection method executed by an image search log collection system comprising product information storage means for storing an image and product identification information of the product in association with each other,
A product image transmission step of transmitting a list of product images acquired from the product information storage means to the user terminal via the communication network;
A similar image search request receiving step for receiving a similar image search request including a selected image selected from a list of images of the product and parameter information from the user terminal;
Using the selected image and the parameter information included in the similar image search request, a similar image search is performed on the product image stored in the product information storage unit, and the search result is transmitted to the user terminal via the communication network. A similar image search step to send to
An image search log collection method comprising: a log collection step of storing, in a log storage unit, at least the product identification information and the parameter information corresponding to the selected image included in the similar image search request. An image search log collection method executed by an image search log collection system connected to a user terminal and a product sales site providing system via a communication network,
A product information registration step of receiving an image of the product from the product sales site providing system and storing the product image together with product identification information of the product in a product information storage unit;
A selection image, which is an image of a product selected by the user from a list of product images, and parameter information are received from the user terminal and stored in the product information storage unit using the selected image and parameter information. A similar image search step of performing a similar image search on the image of the product and transmitting a search result to the user terminal via the communication network;
An image search log collection method comprising: a log collection step of storing, in a log storage unit, at least product identification information corresponding to the selected image received from the user terminal and the parameter information as a log. A similar image search is performed based on an image of a product received from a user terminal connected via a communication network with a computer having a product information storage unit that stores at least an image of the product and product identification information of the product in association with each other. And a program that functions as an image search log collection system that collects information used in similar image search as a log, the computer comprising:
Commodity image transmission means for transmitting a list of commodity images acquired from the commodity information storage means to the user terminal via the communication network;
Similar image search request receiving means for receiving, from the user terminal, a similar image search request including a selection image selected from a list of images of the product and parameter information;
Using the selected image and the parameter information included in the similar image search request, a similar image search is performed on the product image stored in the product information storage unit, and the search result is transmitted to the user terminal via the communication network. Similar image search means to send to
Log collection means for storing, in a log storage means, product identification information and parameter information corresponding to at least the selected image included in the similar image search request;
Program to function as. A computer connected to a user terminal and a product sales site providing system via a communication network;
Product information registration means for receiving an image of the product from the product sales site providing system and storing the product image in the product information storage means of the computer together with product identification information of the product;
A selection image, which is an image of a product selected by the user from a list of product images, and parameter information are received from the user terminal and stored in the product information storage unit using the selected image and parameter information. A similar image search means for performing a similar image search on the image of the product that has been performed and transmitting a search result to the user terminal via the communication network;
Log collecting means for storing in the log storage means of the computer at least the identification information of the product corresponding to the selected image received from the user terminal and the parameter information;
Program to function as.

Images