JP5244864B2 - Product identification apparatus, method and program - Google Patents

Description

  The present invention relates to a product identification device, a method, and a program.

  In recent years, various techniques have been proposed for grasping the details of a product in detail based on the appearance of the product or a code attached to the product.

  For example, it has been proposed to attach a two-dimensional code displaying URL information of an Internet site for confirming information related to a product to the outer surface of the product container (see, for example, Patent Document 1). According to this technique, the user can grasp the detailed information about the product by photographing the two-dimensional code attached to the product and accessing the URL obtained from the photographed two-dimensional code.

  Further, the processing server is connected to the terminal device via the network, and the product image data and the item data of the product information (product name, price, product details) are stored in advance and transmitted from the terminal device. In response to receiving the product image, product information acquisition means for acquiring product image data similar to the received product image and item data corresponding to the product image data, and the acquired data to the terminal device 2. Description of the Related Art A processing server including a transmitting unit that transmits data is known (for example, see Patent Document 2). According to this processing server, the user of the terminal can grasp the product name, the price, and the product details by photographing the product and transmitting an image of the photographed product to the processing server.

JP 2009-234656 A JP 2006-285654 A

However, in the invention described in Patent Document 1, when the merchandise provider intends to accurately convey detailed information to the user about the attributes of various merchandise displayed in the store, the merchandise provider individually assigns a two-dimensional code to all merchandise. It must be attached, and it alone imposes a great burden on the product provider. Also, in order to reduce this burden, based on Patent Document 2, a product image is captured instead of a two-dimensional code, and the product image is transmitted to the server. The server tries to convey correct and detailed information about the product to the user. In the invention described in Patent Document 2, a large amount of product image data must be stored in a database.
When a large amount of product image data is stored in a terminal device, that is, a mobile phone terminal or other portable terminal, the storage capacity inside the terminal device is limited, so a large amount of product image data is stored in the storage area inside the terminal device. There is a limit to saving. As a result, as shown in FIG. 1 and FIG. 2 of Patent Document 2, the user cannot acquire product information unless the terminal device is connected to the processing server. Furthermore, since the invention described in Patent Document 2 is based on image identification technology for product image data, it is problematic to uniquely identify the product from the captured image. This is because many very similar products are sold only from the appearance of the product. Further, each time new product image data is added to the database, the accuracy when the product information acquisition unit acquires item data corresponding to the product image data may be reduced.

  The present invention provides a product identification device, a method, and a program for uniquely identifying a product by photographing a product label attached to a product in a predetermined range and providing the product attribute to a user. Objective.

  The present invention provides the following solutions.

  (1) A product attribute database in which a predetermined number of product label image divided images and product attribute corresponding to the product label image are stored, and an identification request product label that requests identification of which product label Identification request product label image receiving means for receiving the image of the identification request, and an identification request product label image for dividing the image of the identification request product label received by the identification request product label image receiving means a predetermined number of times according to a predetermined division rule An identification request product label split image for selecting one identification request product label split image from a plurality of identification request product label split images divided by the identification request product label image split means according to a predetermined selection rule Selecting means; and the predetermined number of product label divided images stored in the product attribute database; The product label division image most similar to the identification request product label division image is extracted by comparing with the identification request product label division image selected by the different request product label division image selection means, and the extracted product label division A product identification device comprising: product identification means for extracting an attribute of the product stored in association with an image.

  According to the invention described in (1), in response to a request for identification of which product label is the product label, the identification request product label image receiving means receives the image of the identification request product label related to the request. Accept. Further, the identification request product label image dividing unit divides the image of the identification request product label received by the identification request product label image receiving unit a predetermined number of times according to a predetermined division rule stored in the product attribute database, The identification request product label divided image selection means selects one identification request product label divided image from among the plurality of identification request product label divided images divided by the identification request product label image division means according to a predetermined selection rule. The product identification unit compares the predetermined number of product label divided images stored in the product attribute database with the identification request product label divided image selected by the identification request product label divided image selection unit, and determines the identification request product label. The product label divided image most similar to the divided image is extracted, and the attribute of the product stored in association with the extracted product label divided image is extracted, so that the product with the identification request product label attached is a predetermined number of products. Which product is identified. Thereby, the product identification apparatus which identifies the said product and provides the product attribute to a user can be provided by photographing the product label attached to the product of the predetermined range.

  (2) Product information receiving means for receiving the predetermined number of product label images and product attributes corresponding to the product label images, and products for dividing each of the predetermined number of product label images according to the predetermined division rule For each of the label image dividing unit and the predetermined number of product label images, one product label divided image is selected from the plurality of divided images divided by the product label image dividing unit according to the predetermined selection rule. Product label divided image selecting means, product label divided image determining means for comparing all the predetermined number of product label divided images selected by the product label divided image selecting means and determining whether or not they can be distinguished from each other, and the product label If it is determined that the product can be identified by the divided image discriminating means, the product label image dividing means The product label division image selection means and the product label division image discrimination means repeatedly repeat the selection means, and the product label division image discrimination means determines that the product label division image discrimination means is not identifiable. The predetermined number of product label divided images finally determined to be identifiable by the image discriminating means and the product attributes corresponding to the product label divided images are stored in the product attribute database in association with each other, and the identification The product identification device according to (1), further comprising storage control means for storing the number of repetitions last determined to be possible in the predetermined storage area as the predetermined number of times.

  According to the invention described in (2), the product identification device repeats the division of a predetermined number of product label images as many times as necessary until the product label divided image determination unit determines that the product label is not identifiable. When it is determined that the product cannot be identified by the divided image discriminating unit, the predetermined number of product label segmented images that are discriminated last by the product label segmented image discriminating unit and the product corresponding to the product label segmented image The attributes are associated with each other and stored in the product attribute database, and the number of repetitions last determined to be identifiable is stored in a predetermined storage area as a predetermined number of times.

  As a result, the user can use the product attribute database with the minimum amount of data provided by the product identification device to display the product label image on the mobile terminal or the like even in an environment where the user cannot connect to the network. You can grasp the product attributes just by taking a picture. Furthermore, when a new product label image is added to the product attribute database, the predetermined division rule, the predetermined selection rule, and the number of repetitions are updated, and the minimum number of products that can always distinguish the predetermined number of product label images from each other. Since the label divided image is stored, the identification accuracy of the product label image can be maintained at a certain level even when a new product label image is added.

  (3) A method of extracting a product attribute by a product identification apparatus having a product attribute database in which a predetermined number of divided product label images and product attributes corresponding to the product label image are stored. An identification request product label image receiving step for receiving an image of an identification request product label requesting identification of whether it is a product label, and an image of the identification request product label received in the identification request product label image receiving step being divided into predetermined segments An identification request product label image dividing step for dividing a predetermined number of times according to a rule, and one identification request product label divided image from a plurality of identification request product label divided images divided in the identification request product label image division step Identification request product label divided image selection step to be selected according to the selection rule, and stored in the product attribute database The predetermined number of product label split images and the identification request product label split image selected in the identification request product label split image selection step are compared to extract the product label split image most similar to the identification request product label split image And a product identification step of extracting the attribute of the product stored in association with the extracted product label divided image.

  (4) Product information receiving step for receiving the predetermined number of product label images and product attributes corresponding to the product label images, and products for dividing each of the predetermined number of product label images according to the predetermined division rule For each of the label image dividing step and the predetermined number of product label images, a product for selecting one product label divided image from the plurality of divided images divided in the product label image dividing step according to the predetermined selection rule A label divided image selecting step, a product label divided image determining step for comparing all the predetermined number of product label divided images selected in the product label divided image selecting step and determining whether or not they can be distinguished from each other, and the product label divided image If it is determined that the product can be identified in the determination process, the product label image is divided by the product label image dividing step, the product label When it is determined that the selection by the split image selection step and the determination by the product label split image discrimination step are further repeated and the product label split image discrimination step cannot be identified, the product label split image discrimination step identifies The predetermined number of product label divided images that are finally determined to be possible and the attributes of the products corresponding to the product label divided images are stored in the product attribute database in association with each other, and finally, if the product can be identified The method according to (3), further comprising: a storage control step of storing the determined number of repetitions in the predetermined storage area as the predetermined number of times.

  (5) A program for causing a product identification apparatus having a product attribute database in which a predetermined number of divided product label images and product attributes corresponding to the product label image are stored to extract product attributes, An identification request product label image receiving step for receiving an image of an identification request product label requesting identification of which product label, and an image of the identification request product label received in the identification request product label image receiving step are predetermined. The identification request product label image division step for dividing the identification request product label image by a predetermined number of times according to the division rule, and one identification request product label division image from among the plurality of identification request product label division images divided in the identification request product label image division step An identification request product label divided image selection step for selecting the product according to a predetermined selection rule, and the product attribute database. The predetermined number of product label divided images stored in the image and the identification request product label divided image selected in the identification request product label divided image selection step are compared, and the product most similar to the identification request product label divided image A product identification step of extracting a label divided image and extracting an attribute of the product stored in association with the extracted product label divided image.

  According to the method described in (3), the same effect as in (1) can be expected by using this method. According to the method described in (4), the same effect as in (2) can be expected by using the method. Further, according to the program described in (5), the same effect as in (1) can be expected by applying for the production, use, transfer, etc., export, import or transfer of the program.

  According to the present invention, the storage capacity for specifying the product by providing the user with the product attribute by photographing the product label attached to the product in a predetermined range, and for specifying the product It is possible to provide a product identification device, method, and program that provide a product attribute database with a small number of items.

It is a figure showing a schematic structure for explaining merchandise item identification device 1 concerning an embodiment of the present invention. It is a figure which shows the example of the division | segmentation and selection of a product label image. It is a figure following FIG. It is a figure following FIG. It is a figure which shows the flowchart of the goods attribute DB creation task by the goods identification device. It is a figure which shows the flowchart of the goods identification task by the goods identification device. It is a figure which shows the example of the goods attribute database 31 produced using the goods identification apparatus. It is a figure which shows the example of the goods attribute database 31 which concerns on a modification.

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

[overall structure]
FIG. 1 is a diagram showing a schematic configuration for explaining a product identification device 1 according to the present embodiment.

  The product identification device 1 is a device for identifying which product among a predetermined number of products, and in that the product label found in the store displaying the product can be photographed on the spot, A mobile phone terminal with a camera or other mobile terminal with a photographing function is preferable. The commodity identification device 1 includes a control unit 10 and a storage unit 30. The control part 10 is a part which controls each function which concerns on the merchandise item identification apparatus 1, and as a main structure regarding a merchandise item identification function, the identification request | requirement merchandise label image reception part 11, the identification request | requirement merchandise label image division | segmentation part 12, An identification request product label divided image selection unit 13 and a product identification unit 14 are provided. On the other hand, the storage unit 30 is a part that stores various programs (not shown) and a database for functioning as the product identification device 1, and includes a product attribute database 31, for example.

  First, from the storage unit 30, as illustrated in FIG. 7 described later, the product attribute database 31 includes a predetermined number of divided product label images and product attributes corresponding to the product label images (for example, The URL of the Internet site for confirming detailed information on the product).

  Next, the control unit 10 will be described. The identification request product label image receiving unit 11 receives an image of an identification request product label that requests identification of which product among the predetermined number of product labels. The identification request product label image receiving unit 11 functions as an identification request product label image receiving unit. In order to accept this image, the user selects a product he / she is interested in from among the products displayed at the store, photographs the label of the selected product, and which product's label is the photographed label This is done by requesting identification. It is preferable to shoot the product from the front, but depending on the location of the product lined up in the store, the product cannot be taken from the front and the product must be taken from the side, diagonally or upside down. Therefore, the identification request product label image accepting unit 11 performs a certain correction on the image captured by the user, corrects the product label image from the front, and accepts it as an image of the identification request product label. It is preferable.

  The identification request product label image dividing unit 12 divides the image of the identification request product label received by the identification request product label image receiving unit 11 according to a predetermined division rule stored in the product attribute database 31. The “predetermined division rule” will be described in detail later with reference to FIGS. The identification request product label image dividing unit 12 functions as an identification request product label image dividing unit.

  The identification request product label divided image selection unit 13 stores one identification request product label divided image among the plurality of identification request product label divided images divided by the identification request product label image division unit 12 in the product attribute database 31. Select according to a predetermined selection rule. The “predetermined selection rule” will be described in detail later with reference to FIGS. The identification request product label divided image selection unit 13 functions as an identification request product label divided image selection unit.

  The product identification unit 14 first compares the predetermined number of product label divided images stored in the product attribute database 31 with the identification request product label divided image selected by the identification request product label divided image selection unit 13 to request identification. The product label divided image most similar to the product label divided image is extracted. Then, the product identification unit 14 extracts the product attribute stored in the product attribute database 31 in association with the extracted product label divided image, and which product the product to which the identification request product label is attached is. Identify The product identification unit 14 functions as a product identification unit.

  By the way, in addition to the configuration described above, the control unit 10 includes, as main components related to the product attribute DB creation function, a product information reception unit 21, a product label image division unit 22, a product label division image selection unit 23, a product label division. An image discriminating unit 24, a repeating unit 25, a storage control unit 26, and the like are provided. On the other hand, the storage unit 30 is a part that stores various programs (not shown) and a database for functioning as the product identification device 1 and includes a product attribute database 31 and the like.

  The product information receiving unit 21 receives a predetermined number of product label images and product attributes corresponding to the product label images. The product information receiving unit 21 functions as a product label image receiving unit. In the present embodiment, the product is “wine” and the product information receiving unit 21 receives a wine product label image. However, the product is not limited to “wine”, and a label is attached. As long as it is a product, it may be any product such as sake or shochu.

  The number of product label images received by the product information receiving unit 21 can be arbitrarily determined, but is preferably as large as possible so that the product identification device 1 can identify many products. On the other hand, considering that the storage area of the product identification device 1 is limited, the number of product label images received by the product information receiving unit 21 is preferably as small as possible. If both are balanced, the product information reception unit 21 receives as many product label images as possible for one type of product, and the product identification device 1 creates a product attribute database 31 based on the product label image. Is preferably repeated for each product. That is, a product attribute database reflecting as many product label images as possible for the product “wine”, a product attribute database reflecting as many product label images as possible for the product “sake”, and as many product label images as possible for the product “shochu” It is preferable that the product identification device 1 creates a product attribute database that reflects as many product label images as possible for one product for each product.

  The merchandise label image may be received in any manner, for example, an image captured using a photographing apparatus may be captured, or an image scanned with a scanner may be captured.

  An attribute of a product refers to a characteristic or characteristic of the product and other features that express the feature of the product. For example, information on the URL of an Internet site that provides information about the product may be used.

  The product label image dividing unit 22 divides each of the predetermined number of product label images received by the product information receiving unit 21 according to a predetermined division rule. The product label divided image selection unit 23 then selects one product label divided image from among the plurality of divided images divided by the product label image division unit 22 for each of a predetermined number of product label images as a predetermined selection rule. Therefore, select. The product label image dividing unit 22 functions as a product label image dividing unit, and the product label divided image selecting unit 23 functions as a product label divided image selecting unit.

  Here, the “predetermined division rule” and the “predetermined selection rule” will be described. The division of the product label image and the selection of the product label divided image may be any as long as they follow a certain division rule and selection rule. For example, as illustrated in FIGS. 2 to 4, as a division rule, it is conceivable to divide a product label image into m rows and n columns in the horizontal direction and the vertical direction. Here, three rows and three columns are illustrated, but the present invention is not limited to three rows and three columns. As a selection rule, it is conceivable to select any combination that can be selected from 9 squares divided into 3 rows and 3 columns shown in FIG. In addition, this method is only an example, and is not limited to this method.

  As in this embodiment, the size of the divided image can be further reduced by examining the identifiability of any combination of a plurality of divided images, not limited to a single cell. The number of products that can be stored in the identification device 1 can be increased. And when using the divided image of several places, the size of a divided image can be made further smaller by using the divided image of a distant place, As a result, the goods which can be memorize | stored in the goods identification apparatus 1 You can increase the number.

  The product label divided image determination unit 24 compares all the predetermined number of product label divided images selected by the product label divided image selection unit 23 and determines whether or not they can be distinguished from each other. The product label divided image discrimination unit 24 functions as a product label divided image discrimination unit.

  The determination of whether or not identification is possible may be made by directly comparing the product label divided images, but the feature amount of the product label divided image (for example, area, second moment, absolute maximum length, pattern width, equivalent circle diameter, Number of holes, outer perimeter, total perimeter, maximum chord length (horizontal), maximum chord length (vertical), ferret length (horizontal), ferret length (vertical), projection length (horizontal), projection length (vertical), for the time being (Diameter (horizontal), equivalent diameter (vertical), total number of holes, Euler number, degree of roundness, circularity, acicularity or fractal dimension, or a combination thereof).

  2 to 4 will be described in more detail. First, as shown in FIG. 2, the product label image dividing unit 22 divides the image in m rows and n columns (here, 3 rows and 3 columns, but not limited to 3 rows and 3 columns) in the horizontal direction and the vertical direction. . Subsequently, the merchandise label divided image selection unit 23 performs 2. As shown in (1), for each of a predetermined number of product label images, a central cell is selected from among a plurality of divided images divided by the product label image dividing unit 22. Then, the product label divided image determination unit 24 compares all the predetermined number of product label divided images selected by the product label divided image selection unit 23 and determines whether or not they can be distinguished from each other. If it is identifiable, see 3 in FIG. If it is not identifiable, go to 2 in FIG. Move to (2).

  If the product cannot be identified by the central cell, the product label divided image selection unit 23 displays the item 2 in FIG. As shown in (2), the upper right cell is selected, and the product label divided image discriminating unit 24 compares the product label divided images corresponding to the division method, the number of divisions, and the divided image positions for all the predetermined number of product labels. It is then determined whether or not they can be distinguished from each other. If it is identifiable, see 3 in FIG. Move on. If the product label is not identifiable, the product label divided image selection unit 23 performs 2. As shown in (2), the right center cell is selected, and the product label divided image discriminating unit 24 selects the product label divided images corresponding to the division method, the number of divisions, and the divided image positions for all the predetermined number of product labels. A comparison is made to determine whether they can be distinguished from each other. The product label divided image selection unit 23 repeats the selection of one square clockwise until the product label divided image determination unit 24 determines that it can be identified.

If any of the squares cannot be identified, the product label divided image selection unit 23 selects 2. (3) A: As shown by the two combinations, two squares are selected, and the product label divided image determination unit 24 sets the product label divided images corresponding to the division method, the number of divisions, and the divided image positions to a predetermined number of products. All labels are compared to determine whether they can be distinguished from each other. The product label divided image selection unit 23 repeats the selection of the two cells for every case (here, ₉ C ₂ = 36) until the product label divided image determination unit 24 determines that the two can be identified.

If any of the squares cannot be identified, the product label divided image selection unit 23 selects 2. (3) B: As shown by the three combinations, three squares are selected, and the product label divided image discriminating unit 24 compares all of the selected predetermined number of product label divided images and discriminates whether or not they can be distinguished from each other. To do. The product label divided image selection unit 23 repeats the selection of the three cells in all cases (here, ₉ C ₃ = 84) until the product label divided image determination unit 24 determines that it can be identified. Subsequently, the product label divided image selection unit 23 selects all four cells, selects five cells, and so on, until every time the product label divided image determination unit 24 determines that it can be identified (here, ₉ C ₄ = 126, ₉ C ₅ = 126,...

  When it is determined that the product label divided image determination unit 24 can identify the repetition unit 25, the repetition by the product label image division unit 22, the selection by the product label divided image selection unit 23, and the product label divided image determination unit The determination by 24 is further repeated. The repeating unit 25 functions as a repeating unit.

  Here, in FIG. As illustrated in FIG. 6, the product label image dividing unit 22 further applies the horizontal direction and the divided image selected by the product label divided image selecting unit 23 when the product label divided image determining unit 24 determines that the product label divided image determining unit 24 is identifiable. The product label divided image selection unit 23 divides the image into three rows and three columns in the vertical direction by dividing the image into m rows and n columns (in this example, three rows and three columns are illustrated, but the present invention is not limited to three rows and three columns). All possible combinations can be selected as described above for the divided image of 9 squares divided into columns.

  When it is determined that the product label divided image discriminating unit 24 cannot identify the storage control unit 26, the storage control unit 26 has a predetermined number of product label divided images that are finally discriminated by the product label divided image discriminating unit 24. The product attribute database 31 stores the product attribute corresponding to the product label divided image in association with each other. In addition, the storage control unit 26 stores a predetermined division rule, a predetermined selection rule, and the number of repetitions determined last by the product label divided image determination unit 24 in a predetermined storage area of the storage unit 30. Let For example, until the second determination by the product label divided image determination unit 24, it was possible to identify with the central square, but in the third determination, if it is determined that it cannot be identified even in any case, The storage control unit 26 corresponds to the product label divided image and the product label divided image which are determined to be identifiable for the second time by the product label divided image determination unit 24, that is, the square of 9, that is, 81 divided. The product attributes are stored in the product attribute database 31. In addition, the storage control unit 26 stores in a predetermined storage area of the storage unit 30 the division rule of dividing into three in the horizontal direction and the vertical direction, the selection rule of selecting the central square, and the number of repetitions of repeating twice. Let

[Hardware and Software Configuration of this Embodiment]
The hardware and software configurations of this embodiment will be described. In addition to the CPU as a control device (control unit), the hardware of the commodity identification device 1 includes a storage device (storage unit), a communication device (communication unit), a display device (display unit), and an input device (input unit). included. Examples of the storage device (storage unit) include a memory (RAM, ROM, etc.), a hard disk drive (HDD), an optical disk (CD, DVD, etc.) drive, and the like. Examples of the communication device (communication unit) include various wired or wireless interface devices. Examples of the display device (display unit) include various displays such as a liquid crystal display or a plasma display. Examples of the input device (input unit) include a keyboard, a pointing device (such as a mouse), a remote control device, and the like.

  The software of the product identification device 1 includes a program and data for controlling the hardware. Programs and data are stored in a storage device (storage unit), and are appropriately executed and referenced by a control device (control unit). Further, the computer program and data can be distributed by an electromagnetic method, or can be recorded and distributed on a computer-readable medium such as a CD-ROM.

[flowchart]
FIG. 5 is a flowchart of a product attribute database creation task by the product identification device 1 according to the present embodiment, and FIG. 6 is a flowchart of the product identification task by the product identification device 1.

  First, a flowchart of a product attribute database creation task by the product identification device 1 will be described with reference to FIG. In step S1, the product information receiving unit 21 performs a product information receiving process. In this process, the product information receiving unit 21 performs a process of receiving a predetermined number of product label images and product attributes corresponding to the product label images.

  In step S2, the product label image dividing unit 22 performs a product label image dividing process. In this process, the product label image dividing unit 22 performs a process of dividing each of a predetermined number of product label images received by the product information receiving unit 21 according to a predetermined division rule.

  In step S <b> 3, the product label divided image selection unit 23 performs a product label divided image selection process. In this process, the product label divided image selection unit 23 selects a predetermined product label divided image from a plurality of divided images divided by the product label image division unit 22 for each of a predetermined number of product label images. Select according to the rules.

  In step S4, the product label divided image determination unit 24 performs a product label divided image determination process. In this process, the product label divided image determination unit 24 compares all the predetermined number of product label divided images selected by the product label divided image selection unit 23, and determines whether or not these predetermined number of product label divided images can be distinguished from each other. The process which discriminates is performed.

  In step S5, the repetition unit 25 performs a repetition process. In this process, when it is determined that the product label divided image determination unit 24 can identify the repetition unit 25, the repetition by the product label image division unit 22, the selection by the product label divided image selection unit 23, and the product label are performed. Processing to further repeat the discrimination by the divided image discrimination unit 24 is performed.

  In step S6, the storage control unit 26 performs a product identification information storage process. In this process, the storage control unit 26, when it is determined that the product label divided image discriminating unit 24 cannot identify, the predetermined number of products that are finally discriminated by the product label divided image discriminating unit 24. The product attribute database 31 stores the label division image and the product attribute corresponding to the product label division image. In addition, the storage control unit 26 stores a predetermined division rule, a predetermined selection rule, and the number of repetitions determined last by the product label divided image determination unit 24 in a predetermined storage area of the storage unit 30. . When this process is finished, the control unit 10 finishes a series of processes related to the product attribute database creation task.

  Next, with reference to FIG. 6, a flowchart of a product identification task by the product identification device 1 will be described. In step S11, the identification request product label image reception unit 11 performs an identification request product label image reception process. In this process, the identification request product label image receiving unit 11 performs a process of receiving an image of the identification request product label related to the request when requesting identification of which product is the label of the product.

  In step S12, the identification request product label image division unit 12 performs an identification request product label image division process. In this process, the identification request product label image dividing unit 12 stores the image of the identification request product label received by the identification request product label image receiving unit 11 in a predetermined division rule stored in a predetermined storage area of the storage unit 30. Accordingly, a process of dividing the product label divided image discriminating unit 24 by the same number of times as the number of repetitions discriminated last is performed.

  In step S13, the identification request product label divided image selection unit 13 performs an identification request product label divided image selection process. In this process, the identification request product label split image selection unit 13 splits one or more identification request product label split images from among the plurality of identification request product label split images divided by the identification request product label image split unit 12. A process of selecting an image according to a predetermined selection rule stored in a predetermined storage area of the storage unit 30 is performed.

  In step S14, the product identification unit 14 performs a product identification process. In this process, the product identification unit 14 firstly includes a predetermined number of product label division images stored in the product attribute database 31 and one or more identification requests selected by the identification request product label division image selection unit 13. The product label divided image is compared, and the product label divided image most similar to the identification request product label divided image is extracted. And the goods identification part 14 identifies which goods the goods which affixed the identification request | requirement goods label are by extracting the attribute of the goods memorize | stored linked | related with the extracted goods label division | segmentation image. When this process ends, the control unit 10 ends a series of processes related to the product identification task.

[Example]
According to the product identification device 1 according to the present embodiment, first, the product information receiving unit 21 uses a predetermined number of product label images (for example, images of a predetermined number of labels photographed using a photographing device) and the product label images. Corresponding product attributes (for example, URL information of an Internet site that provides information on the product) are received. Then, the product label image dividing unit 22 divides each of the predetermined number of product label images received by the product information receiving unit 21 according to a predetermined division rule. The product label divided image selection unit 23 then selects one product label divided image from among the plurality of divided images divided by the product label image division unit 22 for each of a predetermined number of product label images as a predetermined selection rule. Therefore, select. Then, the product label divided image discriminating unit 24 compares all the predetermined number of product label divided images selected by the product label divided image selecting unit 23, and determines whether or not these predetermined number of product label divided images can be distinguished from each other. Determine. Then, when it is determined that the product label divided image determining unit 24 can identify the repeating unit 25, the repetition unit 25 performs the division by the product label image dividing unit 22, the selection by the product label divided image selecting unit 23, and the product label divided image. The determination by the determination unit 24 is further repeated. Then, when it is determined that the product label divided image discriminating unit 24 cannot identify the storage control unit 26, the storage control unit 26 finally determines a predetermined number of product label divisions that are discriminated by the product label divided image discriminating unit 24. The image and the product attribute corresponding to the product label divided image are associated with each other and stored in the product attribute database 31, and at the end, the image can be identified by the predetermined division rule, the predetermined selection rule, and the product label divided image determination unit 24. The number of repetitions determined in the above is stored in a predetermined storage area of the storage unit 30.

  FIG. 7 shows an example of the product attribute database 31 at this time. In the “product label divided image” field, one or more images obtained by dividing the label of the product to be identified as small as possible within the identifiable range. Are stored for a predetermined number of products, and the URL of the Internet site for confirming information related to the products is stored in the “URL” field. Although not shown, the predetermined storage area of the storage unit 30 stores, as an image division condition (predetermined division rule), that the image is divided into three in both the horizontal direction and the vertical direction. As a condition (predetermined selection rule), it is stored that an image corresponding to the central cell is selected, and it is stored that the number of division repetitions is two.

  As described with reference to FIG. 1, the product identification device 1 includes any one of the product attribute database 31 created by the product attribute database creation task of the product identification device 1 and a predetermined number of product labels. An identification request product label image receiving unit 11 that receives an image of an identification request product label that requests identification of whether it is a product, and an image of the identification request product label received by the identification request product label image receiving unit 11 in a predetermined division The identification request product label image segmentation unit 12 that divides the product label segmented image discriminating unit 24 by the same number as the number of repetitions determined last, and the plurality of identification request product label segmented images that are divided Identification request product label split image selection for selecting one identification request product label split image from among the above-mentioned predetermined selection rules 13 and a predetermined number of product label divided images stored in the product attribute database 31 are compared with the identification request product label divided image selected by the identification request product label divided image selection unit 13 to obtain an identification request product label divided image. By extracting the most similar product label divided image and extracting the attribute of the product stored in association with the extracted product label divided image from the product attribute database 31, which product is attached with the identification request product label And a product identification unit 14 for identifying whether or not.

  According to the invention described in the present embodiment, the user configures the product attribute database 31 with the minimum amount of data provided by the product identification device 1 so that the product can be obtained simply by taking an image of the product label with a mobile terminal or the like. Can understand the attributes of Furthermore, when a new product label image is added to the product attribute database 31, the predetermined division rule, the predetermined selection rule, and the number of repetitions are updated, and the product attribute database 31 always identifies a predetermined number of product label images from each other. Since the minimum possible product label division image is stored, even if a new product label image is added, the identification accuracy of the product label image can be kept at a certain level.

  The effects described in the present embodiment only list the most preferable effects resulting from the present invention, and the effects according to the present invention are not limited to those described in the present embodiment.

  In this embodiment, the URL of the Internet site for confirming information related to the product is given as an example of the “product attribute”. However, if the URL includes the URL, a unique property of the product, You may include attributes and other features that show the characteristics of the product. For example, in addition to URL, product brand, type, place of origin, vintage year, price, taste information such as content, spiciness / sweetness, producer, winery, grape variety, cork type, purchaser's A comment etc. are mentioned.

  For example, FIG. 8 shows an example of the product attribute database 31 at this time, and the “product label divided image” field stores a predetermined number of images obtained by dividing the label of the product to be identified as small as possible within an identifiable range. In the “URL” field, the URL of the Internet site for confirming information related to the product is stored. The “brand” field stores the brand name of the product corresponding to the label, the “type” field stores the type of the product corresponding to the label, and the “production area” field stores the brand name of the product. The origin of the corresponding product is stored, the “vintage year” field stores the vintage year of the product corresponding to the label, and the “price” field stores the price of the product corresponding to the label. By storing these pieces of information in, for example, a portable terminal, it is possible to provide the user with simple information regarding the product without accessing the URL displayed on the portable terminal.

  In this embodiment, the product label image dividing unit 22 divides the image according to the “predetermined division rule”. The “predetermined division rule” is any rule as long as it is determined in advance. Alternatively, it may be a four-divided rectangle, or may be a three-divided or two-divided rectangle. Further, the shape of the divided image does not have to be a rectangle, and may be a circle or another shape.

  In the present embodiment, the product label divided image selection unit 23 selects an image at the center position of the image before the division. However, the present invention is not limited to this. Any rule may be used as long as it is a “selection rule”.

DESCRIPTION OF SYMBOLS 1 Product identification apparatus 10 Control part 11 Identification request | requirement product label image reception part 12 Identification request | requirement product label image division | segmentation part 13 Identification request | requirement product label division | segmentation image selection part 14 Product identification part 30 Memory | storage part 31 Product attribute database

Claims (3)

A product attribute database in which a predetermined number of divided images of product label images and product attributes corresponding to the product label images are stored;
An identification request product label image receiving means for receiving an image of an identification request product label requesting identification of which product label,
An identification request product label image dividing unit that divides an image of the identification request product label received by the identification request product label image receiving unit a predetermined number of times according to a predetermined division rule;
An identification request product label split image selection unit that selects one identification request product label split image from a plurality of identification request product label split images divided by the identification request product label image split unit according to a predetermined selection rule;
By comparing the predetermined number of product label division images stored in the product attribute database with the identification request product label division image selected by the identification request product label division image selection means, the identification request product label division image is the most. Product identification means for extracting similar product label divided images and extracting attributes of the products stored in association with the extracted product label divided images;
Product information accepting means for accepting the predetermined number of product label images and product attributes corresponding to the product label images;
Product label image dividing means for dividing each of the predetermined number of product label images according to the predetermined division rule;
Product label divided image selection means for selecting one product label divided image from a plurality of divided images divided by the product label image dividing means for each of the predetermined number of product label images according to the predetermined selection rule When,
Product label divided image discriminating means for comparing all the predetermined number of product label divided images selected by the product label divided image selecting means and discriminating whether or not they can be distinguished from each other;
If it is determined that the product label divided image discriminating means can identify, further division by the product label image dividing unit, selection by the product label divided image selecting unit, and discrimination by the product label divided image discriminating unit are further performed. Repeating means of repetition;
When it is determined that the product label divided image discriminating unit cannot identify the product label divided image discriminating unit, the predetermined number of product label divided images finally discriminated by the product label divided image discriminating unit and the product label divided unit Storage control means for associating and storing the product attribute corresponding to the image in the product attribute database, and further storing the number of repetitions last determined as identifiable in the predetermined storage area as the predetermined number of times When,
A product identification device comprising: A product identification apparatus having a product attribute database in which a predetermined number of divided product label images and product attributes corresponding to the product label images are stored is a method for extracting product attributes,
An identification request product label image receiving step for receiving an image of an identification request product label requesting identification of which product label,
An identification request product label image dividing step of dividing the image of the identification request product label received in the identification request product label image receiving step by a predetermined number of times according to a predetermined division rule;
An identification request product label split image selection step of selecting one identification request product label split image from a plurality of identification request product label split images divided in the identification request product label image split step according to a predetermined selection rule;
The predetermined number of product label division images stored in the product attribute database and the identification request product label division image selected in the identification request product label division image selection step are compared to most similar to the identification request product label division image. Extracting the product label divided image, and extracting the product attribute stored in association with the extracted product label divided image; and
A product information receiving step for receiving the predetermined number of product label images and product attributes corresponding to the product label images;
A product label image dividing step of dividing each of the predetermined number of product label images according to the predetermined division rule;
For each of the predetermined number of product label images, a product label divided image selection step for selecting one product label divided image from a plurality of divided images divided in the product label image division step according to the predetermined selection rule; ,
A product label split image discrimination step for comparing all the predetermined number of product label split images selected in the product label split image selection step and discriminating whether or not they can be distinguished from each other;
If it is determined that the product label division image determination step can be identified, the division by the product label image division step, the selection by the product label division image selection step, and the determination by the product label division image determination step are further repeated. An iterative process;
When it is determined that the product label divided image determining step is not identifiable, the predetermined number of product label divided images and the product label divided image that are finally determined to be identifiable by the product label divided image determining step A storage control step of associating and storing the corresponding product attribute in the product attribute database, and further storing the number of repetitions last determined as identifiable in the predetermined storage area as the predetermined number of times;
Including methods. A program for causing a product identification device having a product attribute database in which a predetermined number of divided product label images and product attributes corresponding to the product label image are stored to extract product attributes,
An identification request product label image receiving step for receiving an image of an identification request product label requesting identification of which product label,
An identification request product label image dividing step of dividing the image of the identification request product label received in the identification request product label image receiving step by a predetermined number of times according to a predetermined division rule;
An identification request product label split image selection step of selecting one identification request product label split image from a plurality of identification request product label split images divided in the identification request product label image split step according to a predetermined selection rule;
The predetermined number of product label division images stored in the product attribute database and the identification request product label division image selected in the identification request product label division image selection step are compared to most similar to the identification request product label division image. Extracting the product label divided image, and extracting the product attribute stored in association with the extracted product label divided image; and
A product information receiving step for receiving the predetermined number of product label images and product attributes corresponding to the product label images;
A product label image dividing step of dividing each of the predetermined number of product label images according to the predetermined division rule;
For each of the predetermined number of product label images, a product label divided image selection step for selecting one product label divided image from a plurality of divided images divided in the product label image division step according to the predetermined selection rule; ,
A product label split image discrimination step for comparing all the predetermined number of product label split images selected in the product label split image selection step and discriminating whether or not they can be distinguished from each other;
If it is determined that the product label division image determination step can be identified, the division by the product label image division step, the selection by the product label division image selection step, and the determination by the product label division image determination step are further repeated. An iterative process;
When it is determined that the product label divided image determining step is not identifiable, the predetermined number of product label divided images and the product label divided image that are finally determined to be identifiable by the product label divided image determining step A storage control step of associating and storing the corresponding product attribute in the product attribute database, and further storing the number of repetitions last determined as identifiable in the predetermined storage area as the predetermined number of times;
A program for running

Images