WO2017138088A1 - Color classification device, color classification method, program, and non-transient computer-readable information recording medium - Google Patents

Color classification device, color classification method, program, and non-transient computer-readable information recording medium Download PDF

Info

Publication number
WO2017138088A1
WO2017138088A1 PCT/JP2016/053783 JP2016053783W WO2017138088A1 WO 2017138088 A1 WO2017138088 A1 WO 2017138088A1 JP 2016053783 W JP2016053783 W JP 2016053783W WO 2017138088 A1 WO2017138088 A1 WO 2017138088A1
Authority
WO
WIPO (PCT)
Prior art keywords
color
region
product
area
unit
Prior art date
Application number
PCT/JP2016/053783
Other languages
French (fr)
Japanese (ja)
Inventor
淳 片川
Original Assignee
楽天株式会社
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 楽天株式会社 filed Critical 楽天株式会社
Priority to JP2016536793A priority Critical patent/JP6028130B1/en
Priority to PCT/JP2016/053783 priority patent/WO2017138088A1/en
Publication of WO2017138088A1 publication Critical patent/WO2017138088A1/en

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q30/00—Commerce
    • G06Q30/06—Buying, selling or leasing transactions

Definitions

  • the present invention relates to an amber color classification apparatus, a color classification method, a program, and a non-transitory computer-readable information recording medium.
  • sales sites for electronic commerce have become popular on the Internet.
  • a sales site switching mall site
  • stores virtual stores
  • a user searches for a target product by searching for a target product by narrowing down a category or by setting a product name, a model number, or the like as a search condition.
  • Patent Literature 1 discloses the prior art of a sales site (Internet shopping system) that searches for products by color beyond the types of sellers and products.
  • Patent Document 1 it is possible to select a color visually confirmed by the user from a predetermined number of color samples displayed on the color menu and instruct a search for a product. ing. In consideration of a simpler user interface, it may be possible to search for a product by designating an arbitrary color name from among a predetermined number of color names. In any case, one of the roughly classified representative colors (for example, “white”, “yellow”, “orange”, “pink”, “red”, “beige”,%) Designation (selection) is performed to search for products.
  • the roughly classified representative colors for example, “white”, “yellow”, “orange”, “pink”, “red”, “beige”, etc.
  • the product color corresponds to an intermediate color that can be said to be “red” or “orange”.
  • the color of a product is registered as one color (for example, “red”), even if the other color (for example, “orange”) is designated and searched, the product is included in the search results. There will be no. For this reason, simply, if it corresponds to an intermediate color, it may be included in the search result regardless of which color is specified.
  • the present invention has been made in view of the above circumstances, and provides a color classification device, a color classification method, a program, and a non-transitory computer-readable information recording medium that can appropriately classify the color of a product.
  • the purpose is to do.
  • a color classification device provides: A storage unit that stores region information about a plurality of regions that divide the color space and associates the range of the region with a representative color indicating the region; A classifying unit that classifies pixels for products included in the product image according to which of the plurality of regions the color of the pixels belongs; An acquisition unit that acquires the first region and the second region in descending order of the number of pixels classified into each of the plurality of regions; The first area and the second area are in contact with each other in the color space, and a ratio between the number of pixels classified into the first area and the number of pixels classified into the second area is a predetermined ratio.
  • a determination unit that determines whether or not the intermediate color condition is satisfied, A search unit capable of searching for the product in any of the first representative color associated with the first region and the second representative color associated with the second region, if the intermediate color condition is satisfied; A first response to the product of the searcher after the product is searched with the first representative color and a second response to the product of the searcher after the product is searched with the second representative color And a correction unit that corrects a range of the region in the region information so as to change a boundary between the first region and the second region, It is characterized by providing.
  • the correction unit corrects the range of the region in the region information so that the first region is narrowed and the second region is widened when the second reaction is greater in degree than the first reaction. , It is characterized by that.
  • the correction unit is configured to return the boundary between the first region and the second region to the initial boundary when the first reaction is larger than the second reaction. Modify the range, It is characterized by that.
  • a registration unit that registers both the first representative color and the second representative color in the product information of the product when the intermediate color condition is satisfied;
  • the search unit includes a first probability that the search is hit when the specified color specified during the search matches the first representative color in the product information, and the specified color is the second representative in the product information. Search for products according to the second probability of hitting the search when it matches the color, It is characterized by that.
  • the search unit calculates the first probability and the second probability based on the number of pixels classified into the first region and the number of pixels classified into the second region. It is characterized by that.
  • the color classification method is: A color classification method in a color classification apparatus having a storage unit that stores area information for associating a range of the area with a representative color indicating the area, which is area information about a plurality of areas that divide the color space, A classification step in which the color classification device classifies pixels of products included in a product image according to which of the plurality of regions a color of the pixels belongs; An acquisition step in which the color classification device acquires the first region and the second region in descending order of the number of pixels classified into each of the plurality of regions; In the color classification device, the first area and the second area are in contact with each other in the color space, and the number of pixels classified into the first area and the number of pixels classified into the second area A determination step of determining whether or not the intermediate color condition is satisfied, wherein the ratio exceeds a predetermined ratio; If the color classification device satisfies the intermediate color condition, the product is searched for in either the first representative color associated with the first region or the second representative color associated with
  • the color classification device includes a first response to the product of the searcher after the product is searched with the first representative color and a searcher after the product is searched with the second representative color.
  • the program according to the third aspect of the present invention is: A computer having a storage unit that stores area information about a plurality of areas that divide the color space and that associates the range of the area with a representative color indicating the area, A classifying unit that classifies pixels for products included in the product image according to which of the plurality of regions the color of the pixels belongs; An acquisition unit that acquires the first region and the second region in descending order of the number of pixels classified into each of the plurality of regions; The first area and the second area are in contact with each other in the color space, and a ratio between the number of pixels classified into the first area and the number of pixels classified into the second area is a predetermined ratio.
  • a determination unit that determines whether or not the intermediate color condition is satisfied, A search unit capable of searching for the product in any of the first representative color associated with the first region and the second representative color associated with the second region, if the intermediate color condition is satisfied; A first response to the product of the searcher after the product is searched with the first representative color and a second response to the product of the searcher after the product is searched with the second representative color And a correction unit that corrects a range of the region in the region information so as to change a boundary between the first region and the second region, It is made to function as.
  • the above program can be distributed and sold via a computer communication network independently of the computer on which the program is executed.
  • a computer-readable recording medium is provided.
  • a computer having a storage unit that stores area information about a plurality of areas that divide the color space and that associates the range of the area with a representative color indicating the area, A classifying unit that classifies pixels for products included in the product image according to which of the plurality of regions the color of the pixels belongs; An acquisition unit that acquires the first region and the second region in descending order of the number of pixels classified into each of the plurality of regions; The first area and the second area are in contact with each other in the color space, and a ratio between the number of pixels classified into the first area and the number of pixels classified into the second area is a predetermined ratio.
  • a determination unit that determines whether or not the intermediate color condition is satisfied, A search unit capable of searching for the product in any of the first representative color associated with the first region and the second representative color associated with the second region, if the intermediate color condition is satisfied; A first response to the product of the searcher after the product is searched with the first representative color and a second response to the product of the searcher after the product is searched with the second representative color And a correction unit that corrects a range of the region in the region information so as to change a boundary between the first region and the second region, A program characterized by functioning as a program is recorded.
  • the recording medium may be a non-transitory recording medium and can be distributed and sold independently of the computer.
  • the non-temporary recording medium refers to a tangible recording medium.
  • Non-temporary recording media are, for example, compact disks, flexible disks, hard disks, magneto-optical disks, digital video disks, magnetic tapes, semiconductor memories, and the like.
  • the transitory recording medium refers to the transmission medium (propagation signal) itself.
  • the temporary recording medium is, for example, an electric signal, an optical signal, an electromagnetic wave, or the like.
  • the temporary storage area is an area for temporarily storing data and programs, and is, for example, a volatile memory such as a RAM (Random Access Memory).
  • FIG. 6 is a schematic diagram for explaining a “white” region. It is a schematic diagram for demonstrating the area
  • Embodiments of the present invention will be described below.
  • a sales site shopping mall site
  • shops virtual stores
  • the present invention can be appropriately applied to a sales site by one shop (trader).
  • the following embodiment is for description and does not limit the scope of the present invention. Therefore, those skilled in the art can employ embodiments in which each or all of these elements are replaced with equivalent ones, and these embodiments are also included in the scope of the present invention.
  • a sales system 100 is configured by connecting a color classification device 200 and each user terminal 300 via the Internet 900 as shown in FIG. Although simplified in the figure, it is assumed that there are a large number of user terminals 300 according to the users to be used.
  • the color classification device 200 includes, for example, a sales server (server computer) or the like, and presents information related to products to the user terminal 300 and sells products desired by the user.
  • the color classification device 200 presents a product list page or the like to the accessing user terminal 300.
  • the product information about the product is provided with a color tag indicating the color of the product.
  • the color classification device 200 displays the color tag.
  • the product list page narrowed down is presented to the user terminal 300.
  • the user terminal 300 includes, for example, a personal computer, a smartphone, and the like, accesses the color classification device 200 via the Internet 900, acquires information about the product, accepts a user operation, and receives a user operation from the color classification device 200. Make a purchase.
  • the information processing apparatus 400 includes a CPU (Central Processing Unit) 401, a ROM (Read Only Memory) 402, a RAM (Random Access Memory) 403, a NIC (Network Interface Card) 404, an image A processing unit 405, an audio processing unit 406, a DVD-ROM (Digital Versatile Disc Disc ROM) drive 407, an interface 408, an external memory 409, a controller 410, a monitor 411, and a speaker 412 are provided.
  • a CPU Central Processing Unit
  • ROM Read Only Memory
  • RAM Random Access Memory
  • NIC Network Interface Card
  • the CPU 401 controls the overall operation of the information processing apparatus 400 and is connected to each component to exchange control signals and data.
  • the ROM 402 records an IPL (Initial Program Loader) that is executed immediately after the power is turned on, and when this is executed, a predetermined program is read into the RAM 403 and the CPU 401 starts executing the program.
  • the ROM 402 stores an operating system program and various data necessary for operation control of the information processing apparatus 400 as a whole.
  • the RAM 403 is for temporarily storing data and programs, and holds programs and data read from the DVD-ROM and other data necessary for communication.
  • the NIC 404 is used to connect the information processing apparatus 400 to a computer communication network such as the Internet, and conforms to the 10BASE-T / 100BASE-T standard used when configuring a LAN (Local Area Network).
  • Analog modem for connecting to the Internet using a telephone line, ISDN (Integrated Services Digital Network) modem, ADSL (Asymmetric Digital Subscriber Line) modem, cable modem for connecting to the Internet using a cable television line, etc. These are configured by an interface (not shown) that mediates between these and the CPU 401.
  • the image processing unit 405 processes the data read from the DVD-ROM or the like by a CPU 401 or an image arithmetic processor (not shown) provided in the image processing unit 405, and then processes this data in a frame memory provided in the image processing unit 405. (Not shown).
  • the image information recorded in the frame memory is converted into a video signal at a predetermined synchronization timing and output to the monitor 411. Thereby, various page displays are possible.
  • the audio processing unit 406 converts audio data read from a DVD-ROM or the like into an analog audio signal, and outputs the analog audio signal from the speaker 412 connected thereto. Further, under the control of the CPU 401, a sound to be generated during the progress of the processing performed by the information processing apparatus 400 is generated, and a sound corresponding to the sound is output from the speaker 412.
  • the DVD-ROM mounted on the DVD-ROM drive 407 stores, for example, a program for realizing the color classification device 200 according to the embodiment. Under the control of the CPU 401, the DVD-ROM drive 407 performs a reading process on the DVD-ROM loaded therein, reads necessary programs and data, and these are temporarily stored in the RAM 403 or the like.
  • External memory 409, controller 410, monitor 411, and speaker 412 are detachably connected to interface 408.
  • the external memory 409 stores data relating to the user's personal information in a rewritable manner.
  • the controller 410 accepts operation inputs performed at various settings of the information processing apparatus 400.
  • the user of the information processing apparatus 400 can record these data in the external memory 409 as appropriate by inputting instructions through the controller 410.
  • the monitor 411 presents the data output by the image processing unit 405 to the user of the information processing apparatus 400.
  • the speaker 412 presents the audio data output by the audio processing unit 406 to the user of the information processing apparatus 400.
  • the information processing apparatus 400 uses a large-capacity external storage device such as a hard disk so as to perform the same functions as the DVD-ROM mounted on the ROM 402, RAM 403, external memory 409, and DVD-ROM drive 407. You may comprise.
  • each program causing the information classification apparatus 200 to function as the color classification apparatus 200 according to the present embodiment is executed, and the color classification apparatus 200 according to the present embodiment. Is realized.
  • Each user terminal 300 is also realized in the information processing apparatus 400 in the same manner, but these configurations are omitted, and the most characteristic color classification apparatus 200 in the present embodiment will be described below.
  • FIG. 3 is a block diagram illustrating an example of a schematic configuration of the color classification device 200 according to the present embodiment.
  • the color classification device 200 includes a receiving unit 210, a storage unit 220, a control unit 230, and a presentation unit 240.
  • the accepting unit 210 accepts various information sent from each user terminal 300 via the Internet 900.
  • the accepting unit 210 accepts from the user terminal 300 an instruction to transition to a list page of products and various operation instructions in the list page (for example, a mouse click operation by the user).
  • the above-described NIC 404 or the like can function as such a reception unit 210.
  • the storage unit 220 stores area information about a plurality of areas that divide the color space, in addition to various information related to products and various information related to users. For example, as illustrated in FIG. 4A, the storage unit 220 associates, with respect to a plurality of regions (each region divided by a broken line) dividing the color space CS, the range of the region and a representative color indicating the region.
  • regions each region divided by a broken line
  • the representative color of each area are white, yellow, orange, pink, red, beige, silver, gold, gray, purple, brown, green, blue, and black.
  • the representative color is an example and can be changed as appropriate.
  • multi-colors may be included in the representative colors so as to be able to deal with striped patterns of different colors.
  • the range of R: 196 to 255, G: 0 to 50, and B: 0 to 50 in the RGB value is defined.
  • the RGB values are R: 255, G: 0, and B: 0.
  • the range of “red” is set to 196 in the region information. ⁇ R ⁇ 255, 0 ⁇ G ⁇ 50, and 0 ⁇ B ⁇ 50 are all satisfied.
  • the range of area information is defined for other representative colors. That is, a range having a certain width is also defined for the area information of other representative colors.
  • Such area information is used for classifying the color of the product from the image (image data) of the product, as will be described later. Many images are not monochromatic. Also, it may appear white due to reflection of illumination. These color differences are classified into the same color by giving a certain width in the region information. Further, as described later, the area information may define a range of intermediate colors.
  • the above-described RAM 403, external memory 409, and the like can function as such a storage unit 220.
  • control unit 230 controls the entire color classification apparatus 200.
  • the control unit 230 includes a classification unit 231, an acquisition unit 232, a determination unit 233, a search unit 234, and a correction unit 235, and supports purchase of products by the user.
  • the classification unit 231 classifies which region the pixel color of the product image belongs to.
  • product images are divided into grids (griding) to classify which region each cell belongs to.
  • grids grids
  • one dod (1 pixel) pixel without gridding is described.
  • the region to which each pixel belongs may be classified.
  • the classification unit 231 forms a grid on the product image PI including the product P (suitcase) as illustrated in FIG. 5 so that the number of cells is 100 ⁇ 100 (vertical ⁇ horizontal). After color abstraction is performed for each square, the area to which each square belongs is classified. Since such a product image PI usually includes a background, the classification unit 231 classifies which region each cell for the product image PI excluding the background belongs to.
  • a single color is used for the background.
  • a product image PI including a pop (POP) image G in addition to the product P, there is a product image PI including a pop (POP) image G, and a single color is often used for such a pop image G.
  • POP pop
  • the classification unit 231 determines that the adjacent squares have the same RGB value as a background or a pop image G, and excludes it from the classification target.
  • the classification unit 231 excludes the background using the feature that the product P does not hang over 25% or more around the product image PI (upper, lower, left, and right sides).
  • the product image PI illustrated in FIG. 7A shows a case where the lower part of the product P is hooked on one side of the product image PI.
  • the ratio based on the entire four sides is about 12%.
  • the product image PI illustrated in FIG. 7B a case where a part of the product P is hung on two sides of the product image PI is shown.
  • the lower part of the product P is about 30% on the lower side of the product image PI
  • the right part of the product P is about 10% on the right side of the product image PI. About 16%.
  • FIG. 7C and FIG. 7D are examples of the case where the product P takes 25% or more on the four sides of the product image PI, but it is extremely unnatural and is not considered to be actually used. In fact, as long as about 1,000 product images were randomly extracted and verified at the sales site managed by the applicant, they did not exist. Further, even when a plurality of products P are included in the image as in the product image PI illustrated in FIG. 7E, it is also conceivable that each product P takes 25% or more on four sides of the product image PI. Even so, if each product P of different colors is included in one product image PI, it is difficult to know which color product is which product P, so it is considered that it is not actually used. From these things, the classification
  • the color of the product P and the background color are usually characterized by a color balance having a relatively large contrast. Based on these characteristics, when a color with a width of 1 to 3 pixels is extracted along the periphery of the product image PI, a color that occupies 75% or more of the color can be estimated as a background color and exists in the product image PI. It can be determined that there is a high possibility that all the same-colored parts are the background. For example, in the case of the product image PI shown in FIG.
  • the classification unit 231 determines that all the same color portions as the color indicated by L1 occupying 88% of the entire four sides are the background, and excludes them from the classification target.
  • the product image PI may have a single color border around the product image PI. For this reason, even if the color is extracted along the periphery of the product image PI and occupies 75% or more, if the proportion of the same color in the product image PI is low, the case of bordering is also considered.
  • the background may be determined by another method.
  • the classification unit 231 classifies which region the pixel (mass) for the product image PI excluding the background belongs to. Specifically, when the product P shown in FIG. 5 is a red suitcase, the classification unit 231 classifies 2500 cells as “red” and 900 cells as “gray” as an example. Then, the 200 cells are classified as “pink”. Other squares are also classified but omitted. Further, when the product P shown in FIG. 5 described above is a suitcase having an intermediate color between red and orange with different colors, the classification unit 231 classifies 1400 cells as “red” as an example. Are classified as “orange” and 900 are classified as “gray”. Other squares are also classified but omitted.
  • the acquisition unit 232 acquires the first region and the second region in descending order of the number of pixels (mass) classified by the classification unit 231. For example, when the classification unit 231 classifies “red”: 2500, “gray”: 900, “pink”: 200,..., The acquisition unit 232 acquires “red” as the first region. Then, “gray” is acquired as the second region. When the classification unit 231 classifies “red”: 1400, “orange”: 1100, “gray”: 900,..., The acquisition unit 232 acquires “red” as the first region. Then, “orange” is acquired as the second area.
  • the determination unit 233 determines whether or not the first region and the second region acquired by the acquisition unit 232 satisfy the intermediate color condition.
  • the determination unit 233 includes a first area and a second area that are in contact with each other in the color space, and the number of pixels (mass) classified into the first area and the pixels (mass cells) classified into the second area. ) Is greater than a predetermined ratio (for example, 5: 3), it is determined that the intermediate color condition is satisfied. Specifically, when “red” is acquired as the first area and “gray” is acquired as the second area by the acquisition unit 232, “red” and “gray” are not in contact with each other in the color space. Therefore, the determination unit 233 determines that the intermediate color condition is not satisfied.
  • the determination unit 233 determines that the intermediate color condition is satisfied.
  • the intermediate color condition is an example, and may be determined based on other conditions. For example, as described later, a range for determining the intermediate color is further determined, and it is determined whether the intermediate color condition is satisfied based on the result of classification by the classification unit 231 according to the two types of ranges. May be.
  • the search unit 234 selects the first representative color associated with the first region and the second representative color associated with the second region. Search is possible in any case. For example, the search unit 234 searches for a first probability of hitting the search when the specified color specified during the search matches the first representative color, and when the specified color matches the second representative color. The product is searched according to the second probability to be hit.
  • the first probability is calculated by, for example, “f (number classified into the first region) / f (number classified into the first region + number classified into the second region)”.
  • the second probability is calculated by, for example, “f (the number classified into the second region) / f (the number classified into the first region + the number classified into the second region)”.
  • the search unit 234 makes the product hit based on the first probability when the specified color specified at the time of the search matches the first representative color, or When the designated color matches the second representative color, the product is hit based on the second probability. For this reason, when the designated color matches the first representative color, it is probabilistically advantageous and the product is easily searched. However, when the designated color matches the second representative color, the product is appropriately searched. Will be.
  • the search unit 234 searches for the product only when the first representative color associated with the first region is designated. To do.
  • the correcting unit 235 performs a first reaction on a product by the user (searcher) after the product is searched for with the first representative color, and a second response for the product by the user after the product is searched for with the second representative color. Based on the reaction, the range of the region in the region information is modified so that the boundary between the first region and the second region is changed. For example, when the product P shown in FIG. 5 is a suitcase having an intermediate color between red and orange, the user is instructed to search for a “red” suitcase, and the search unit 234 searches for this suitcase. After presenting it in the results (list of search results), the user is instructed to search for the suitcase of “orange” and the number of cases where this suitcase was selected by the user (instructed to transition to the product page). The search unit 234 changes the boundary between “red” and “orange” based on the number of the suitcases selected by the user after presenting the suitcases in the search results. The range of the area in the area information is corrected.
  • the correction unit 235 reduces the range of the region in the region information so that the “red” region RA is narrowed and the “orange” region OA is expanded accordingly, as shown in FIG. Correct it.
  • products in the middle color between red and orange will be more classified into “orange” than “red” in the future, and it will be easier to search when “orange” is designated as the designated color.
  • the correction unit 235 corrects the range of the region in the region information so that the boundary between the “red” region and the “orange” region is returned to the initial boundary. That is, as shown in FIG. 8, if the “red” area RA becomes narrower and the “orange” area OA expands by that amount, the areas are returned to the original. If the “red” area RA and the “orange” area OA remain the initial areas, the correction unit 235 does not perform any correction.
  • the control unit 230 having such a configuration appropriately performs a so-called A / B test on the intermediate color product, and changes the boundary between corresponding regions according to the result (user reaction). As a result, the intermediate color products are classified into the representative colors that are more responsive to nature, and are easily presented to the user.
  • the above-described CPU 401 and the like can function as the control unit 230 having such a configuration.
  • the presentation unit 240 presents various information to the user terminal 300 via the Internet 900.
  • the presentation unit 240 presents the search results (search result list) that the search unit 234 of the control unit 230 has searched for products to the user terminal 300. That is, when the user designates a designated color and instructs to search for a product, a search result list narrowed down by the designated color is presented to the user terminal 300.
  • FIG. 9 is a flowchart showing the flow of color classification processing executed by the color classification device 200. This color classification process is started, for example, when the user designates a designated color and instructs to search for a product.
  • the color classification device 200 classifies which region the color of the pixel (mass) of the product image belongs to (step S11). For example, the control unit 230 (classification unit 231) determines that the number of cells is 100 ⁇ 100 (vertical ⁇ horizontal) for the product image PI including the product P (suitcase) as shown in FIG. After the grid is formed and color abstraction is performed for each cell, the region to which each cell belongs is classified. As described above, since the product image PI includes a background, the classification unit 231 classifies which area each cell for the product image PI excluding the background belongs to.
  • the classification unit 231 classifies 2500 cells as “red” and 900 cells as “gray” as an example. Then, the 200 cells are classified as “pink”. Other squares are also classified but omitted. Further, when the product P shown in FIG. 5 described above is a suitcase having an intermediate color between red and orange with different colors, the classification unit 231 classifies 1400 cells as “red” as an example. Are classified as “orange” and 900 are classified as “gray”. Other squares are also classified but omitted.
  • the color classification device 200 acquires the first area and the second area in descending order of the number of classified pixels (mass) (step S12). Specifically, in step S11, when the classification is “red”: 2500, “gray”: 900, “pink”: 200,..., The control unit 230 (acquisition unit 232) “Red” is acquired as the area, and “gray” is acquired as the second area. In Step S11, when it is classified as “red”: 1400, “orange”: 1100, “gray”: 900,..., The acquisition unit 232 acquires “red” as the first area. Then, “orange” is acquired as the second area.
  • the color classification device 200 determines whether or not the first area and the second area satisfy the intermediate color condition (step S13).
  • the control unit 230 determines whether or not the first area and the second area satisfy the intermediate color condition (step S13).
  • the control unit 230 determines whether or not the first area and the second area satisfy the intermediate color condition (step S13).
  • the control unit 230 determines whether or not the first area and the second area satisfy the intermediate color condition (step S13).
  • the control unit 230 determines whether or not the first area and the second area satisfy the intermediate color condition.
  • the control unit 230 determines whether or not the intermediate color condition is not satisfied.
  • the determination unit 233 determines that the intermediate color condition is satisfied.
  • the color classification device 200 determines that the first region and the second region satisfy the intermediate color condition (step S13; Yes)
  • the first probability of hitting the search when the designated color matches the first representative color The product is searched according to the second probability of hitting the search when the designated color matches the second representative color (step S14). That is, the control unit 230 (search unit 234) causes the product to be hit based on the first probability when the designated color designated at the time of retrieval matches the first representative color, or the designated color is the second color. If it matches the representative color, the product is hit based on the second probability.
  • the color classification device 200 performs a search when the designated color matches the first representative color.
  • a search for hit is performed (step S15). That is, the control unit 230 (search unit 234) includes the product in the search result only when the first representative color associated with the first region is designated.
  • the color classification device 200 presents the search result to the user terminal 300 (step S16). That is, the presentation unit 240 presents the search results (search result list) searched in step S14 or step S15 described above to the user terminal 300.
  • the color classification device 200 collects user responses when the intermediate color condition is satisfied (step S17). That is, the control unit 230 collects user reactions when the search results searched in step S14 described above are presented to the user terminal 300. That is, the first response to the product by the user after the product is searched with the first representative color and the second response to the product after the product is searched with the second representative color are collected.
  • the product P shown in FIG. 5 is a suitcase of an intermediate color between red and orange
  • the user is instructed to search for a “red” suitcase, and the suitcase is searched in step S14.
  • the user is instructed to search for the number of suitcases selected by the user (first reaction) and the suitcase of “orange”.
  • step S14 the suitcases are searched. After the presentation is included in the result, the number of the suitcases selected by the user (second reaction) is totaled.
  • the color classification device 200 determines whether or not the second reaction has a greater degree of reaction than the first reaction (step S18). If the color classification device 200 determines that the second reaction has a greater degree of reaction than the first reaction (step S18; Yes), the region information is corrected so that the first region is narrowed and the second region is widened. (Step S19). For example, when the search result includes the second representative color “orange”, the suitcase is selected more by the user than when the first representative color “red” is included in the search result. In this case, as shown in FIG. 8 described above, the control unit 230 (correction unit 235) makes the “red” region RA narrow and the “orange” region OA widens accordingly. Correct the range of the area. As a result, products in the middle color between red and orange will be more classified into “orange” than “red” in the future, and it will be easier to search when “orange” is designated as the designated color.
  • the region information is corrected so that the boundary between the first region and the second region is restored.
  • Step S20 For example, when the first representative color “red” is included in the search results, more suitcases are selected from the user than when the second representative color “orange” is included in the search results.
  • the correction unit 235 corrects the range of the region in the region information so that the boundary between the “red” region and the “orange” region is returned to the initial boundary. That is, in step S19, if the “red” area RA is narrowed and the “orange” area OA is expanded by that amount, the areas are returned to the original. If the “red” area RA and the “orange” area OA remain the initial areas, the correction unit 235 does not perform any correction.
  • the color of the product is an intermediate color, it can be appropriately classified into any of the representative colors. Further, since the boundary between adjacent regions is changed according to the reaction of the user, the intermediate color can be classified into representative colors that are easily supported by many users. As a result, the product colors can be appropriately classified.
  • the control unit 230 further includes a registration unit, and the registration unit registers the first representative color and the second representative color in the product information of the product. You may do it.
  • the search unit 234 has a first probability that the search is hit when the specified color specified at the time of the search matches the first representative color in the product information, and the specified color is the second representative in the product information. The product is searched according to the second probability of hitting the search when it matches the color.
  • the classification unit 231 further classifies the product image pixel including the color of the pixel (mass) belonging to the intermediate force region. Then, the determination unit 233 determines whether or not the intermediate color condition is satisfied based on the result of the classification by the classification unit 231 according to the two types of ranges. Thereby, it can be determined in more detail whether the color of the product satisfies the intermediate color condition.
  • a color classification device As described above, according to the present invention, it is possible to provide a color classification device, a color classification method, a program, and a non-transitory computer-readable information recording medium that can appropriately classify the color of a product. Can do.

Abstract

According to the present invention, a storage unit (220) stores region information about a plurality of divided regions of a color space (region information associating the area of each region with a representative color indicating the region). A classification unit (231) classifies the colors of the pixels of an image of a product (product color) so as to indicate to which region the color of each pixel belongs. An acquisition unit (232) ascertains a first region to which the largest number of classified pixels belong and a second region to which the second largest number of classified pixels belong. A determination unit (233) determines whether or not the ascertained first and second regions collectively satisfy neutral color conditions. If it is determined that the first and second regions collectively satisfy the neutral color conditions, a search unit (234) allows the product to be found through search based on either one of a first representative color associated with the first region and a second representative color associated with the second region. A modification unit (235) modifies the areas of regions indicated by the region information, so as to change the boundary between the first region and the second region, on the basis of a first response to the product, which was made after the product was found through search based on the first representative color, and on the basis of a second response to the product, which was made after the product was found through search based on the second representative color.

Description

色分類装置、色分類方法、プログラム、ならびに、非一時的なコンピュータ読取可能な情報記録媒体Color classification apparatus, color classification method, program, and non-transitory computer-readable information recording medium
　本発明は、 色分類装置、色分類方法、プログラム、ならびに、非一時的なコンピュータ読取可能な情報記録媒体に関する。 The present invention relates to an amber color classification apparatus, a color classification method, a program, and a non-transitory computer-readable information recording medium.
　近年、インターネット上において、電子商取引を行う販売サイトが人気を博している。例えば、多くの店舗（仮想店舗）が集まって構成される販売サイト（ショッピングモールサイト）では、全体で幅広いカテゴリ（ジャンル）の商品を取り扱っているだけでなく、店舗同士の価格競争も行われていることもあり、多くのユーザを獲得している。
　このような販売サイトにおいて、ユーザは、カテゴリを絞って目的の商品を探したり、商品名や型番等を検索条件に設定して検索することで目的の商品を探していた。 In recent years, sales sites for electronic commerce have become popular on the Internet. For example, in a sales site (shopping mall site) composed of many stores (virtual stores), not only handles products of a wide category (genre) as a whole, but also price competition between stores is held. Sometimes it has gained many users.
In such a sales site, a user searches for a target product by searching for a target product by narrowing down a category or by setting a product name, a model number, or the like as a search condition.
　ところで、販売サイトで扱われる商品の中には、種々のカラーバリエーションが用意された商品（色違い商品）も多い。例えば、鞄や衣料品などの商品では、商品ページ（商品の販売ページ）において、その商品にどのような色の種類があるかや、各色の商品在庫数などが表示されるようになっている。 By the way, among the products handled at the sales site, there are many products with different color variations (products with different colors). For example, in products such as bags and clothing, on the product page (product sales page), what kind of color the product has, the number of products in each color, etc. are displayed. .
　また、検索条件に色を指定して、商品を検索できる販売サイトも知られている。例えば、特許文献１には、販売者や商品の種別を超えて色によって商品を検索する販売サイト（インターネットショッピングシステム）の先行技術が開示されている。 Also, sales sites that can search for products by specifying colors as search conditions are also known. For example, Patent Literature 1 discloses the prior art of a sales site (Internet shopping system) that searches for products by color beyond the types of sellers and products.
特開２００２－９２０２０号公報Japanese Patent Laid-Open No. 2002-92020
　特許文献１に開示された先行技術では、色メニューに表示される所定数の色見本の中からユーザが目で見て確認した任意の色を選択して、商品の検索を指示できるようになっている。また、より簡易なユーザインタフェースを考慮すると、所定数の色の名称の中から任意の色の名称を指定して、商品の検索を行うことも考えられる。
　いずれにしても、検索時において、大まかに分類した代表色（例えば、「ホワイト」、「イエロー」、「オレンジ」、「ピンク」、「レッド」、「ベージュ」、・・・）の何れかを指定（選択）して、商品の検索を行うことになる。 In the prior art disclosed in Patent Document 1, it is possible to select a color visually confirmed by the user from a predetermined number of color samples displayed on the color menu and instruct a search for a product. ing. In consideration of a simpler user interface, it may be possible to search for a product by designating an arbitrary color name from among a predetermined number of color names.
In any case, one of the roughly classified representative colors (for example, “white”, “yellow”, “orange”, “pink”, “red”, “beige”,...) Designation (selection) is performed to search for products.
　それでも、実際の商品において、その商品の色が、代表色の何れに該当するのかが、明確に決められない場合も少なくない。例えば、商品の色が、「レッド」とも「オレンジ」とも言える中間色に該当する場合などである。この場合、商品の色が一方の色（例えば、「レッド」）に登録されると、他方の色（例えば、「オレンジ」）が指定されて検索されても、その商品が検索結果に含まれないことになる。
　そのため、単純に、中間色に該当する場合には、どちらの色が指定されても、検索結果に含めてしまうことも考えられる。 Nevertheless, in actual products, there are many cases in which it is not possible to clearly determine which color the product corresponds to. For example, the product color corresponds to an intermediate color that can be said to be “red” or “orange”. In this case, if the color of a product is registered as one color (for example, “red”), even if the other color (for example, “orange”) is designated and searched, the product is included in the search results. There will be no.
For this reason, simply, if it corresponds to an intermediate color, it may be included in the search result regardless of which color is specified.
　しかしながら、その場合、せっかく色を指定して商品を絞り込んでいるにも拘わらず、大量の商品が検索されてしまったり、ユーザが色の指定を変えて検索しても、同じ商品（中間色の商品）が再び検索されてしまうなどの不具合も起こり得る。
　このため、中間色等の商品であっても、代表色の何れかに適切に分類することのできる技術が求められていた。 However, in that case, even if the product is narrowed down by specifying the color, even if a large number of products are searched or the user changes the color specification and searches, the same product (intermediate color product) ) May be searched again.
For this reason, there has been a demand for a technique that can appropriately classify a product such as an intermediate color into one of representative colors.
　本発明は、上記実状に鑑みてなされたもので、商品の色を適切に分類することのできる色分類装置、色分類方法、プログラム、ならびに、非一時的なコンピュータ読取可能な情報記録媒体を提供することを目的とする。 The present invention has been made in view of the above circumstances, and provides a color classification device, a color classification method, a program, and a non-transitory computer-readable information recording medium that can appropriately classify the color of a product. The purpose is to do.
　本発明の第１の観点に係る色分類装置は、
　色空間を分割する複数の領域についての領域情報であり、当該領域の範囲と当該領域を示す代表色とを対応付ける領域情報を記憶する記憶部、
　商品画像に含まれる商品についての画素を、当該画素が有する色が前記複数の領域のいずれに属するかにより、分類する分類部、
　前記複数の領域のそれぞれに分類された画素の数が多い順に第１領域および第２領域を取得する取得部、
　前記第１領域と前記第２領域とが前記色空間において接しており、かつ、前記第１領域に分類された画素の数と前記第２領域に分類された画素の数との比率が所定比率を超える、中間色条件を満たすか否かを判別する判別部、
　前記中間色条件を満たせば、前記商品を、前記第１領域に対応付けられた第１代表色および前記第２領域に対応付けられた第２代表色のいずれにても検索可能とする検索部、
　前記第１代表色にて前記商品が検索された後の検索者の前記商品に対する第１反応と、前記第２代表色にて前記商品が検索された後の検索者の前記商品に対する第２反応と、に基づいて、前記第１領域と前記第２領域との境界を変化させるように、前記領域情報における前記領域の範囲を修正する修正部、
　を備えることを特徴とする。 A color classification device according to a first aspect of the present invention provides:
A storage unit that stores region information about a plurality of regions that divide the color space and associates the range of the region with a representative color indicating the region;
A classifying unit that classifies pixels for products included in the product image according to which of the plurality of regions the color of the pixels belongs;
An acquisition unit that acquires the first region and the second region in descending order of the number of pixels classified into each of the plurality of regions;
The first area and the second area are in contact with each other in the color space, and a ratio between the number of pixels classified into the first area and the number of pixels classified into the second area is a predetermined ratio. A determination unit that determines whether or not the intermediate color condition is satisfied,
A search unit capable of searching for the product in any of the first representative color associated with the first region and the second representative color associated with the second region, if the intermediate color condition is satisfied;
A first response to the product of the searcher after the product is searched with the first representative color and a second response to the product of the searcher after the product is searched with the second representative color And a correction unit that corrects a range of the region in the region information so as to change a boundary between the first region and the second region,
It is characterized by providing.
　また、上記観点に係る色分類装置において、
　前記修正部は、前記第２反応が前記第１反応よりも度合いが大きい場合に、前記第１領域が狭くなり、前記第２領域が広がるように、前記領域情報における前記領域の範囲を修正する、
　ことを特徴とする。 In the color classification apparatus according to the above aspect,
The correction unit corrects the range of the region in the region information so that the first region is narrowed and the second region is widened when the second reaction is greater in degree than the first reaction. ,
It is characterized by that.
　また、上記観点に係る色分類装置において、
　前記修正部は、前記第１反応が前記第２反応よりも度合いが大きい場合に、前記第１領域と前記第２領域との境界を初期の境界に戻すように、前記領域情報における前記領域の範囲を修正する、
　ことを特徴とする。 In the color classification apparatus according to the above aspect,
The correction unit is configured to return the boundary between the first region and the second region to the initial boundary when the first reaction is larger than the second reaction. Modify the range,
It is characterized by that.
　また、上記観点に係る色分類装置において、
　前記中間色条件を満たす場合に、前記第１代表色および前記第２代表色の両方を前記商品の商品情報に登録する登録部、を更に備え、
　前記検索部は、検索時に指定された指定色が、前記商品情報における前記第１代表色と一致する際に検索をヒットさせる第１確率と、当該指定色が、前記商品情報における前記第２代表色と一致する際に検索をヒットさせる第２確率とに応じて、商品の検索を行う、
　ことを特徴とする。 In the color classification apparatus according to the above aspect,
A registration unit that registers both the first representative color and the second representative color in the product information of the product when the intermediate color condition is satisfied;
The search unit includes a first probability that the search is hit when the specified color specified during the search matches the first representative color in the product information, and the specified color is the second representative in the product information. Search for products according to the second probability of hitting the search when it matches the color,
It is characterized by that.
　また、上記観点に係る色分類装置において、
　前記検索部は、前記第１領域に分類された画素の数と前記第２領域に分類された画素の数とに基づいて、前記第１確率および第２確率を算定する、
　ことを特徴とする。 In the color classification apparatus according to the above aspect,
The search unit calculates the first probability and the second probability based on the number of pixels classified into the first region and the number of pixels classified into the second region.
It is characterized by that.
　本発明の第２の観点に係る色分類方法は、
　色空間を分割する複数の領域についての領域情報であり、当該領域の範囲と当該領域を示す代表色とを対応付ける領域情報を記憶する記憶部を有する色分類装置における色分類方法であって、
　前記色分類装置が、商品画像に含まれる商品についての画素を、当該画素が有する色が前記複数の領域のいずれに属するかにより、分類する分類ステップ、
　前記色分類装置が、前記複数の領域のそれぞれに分類された画素の数が多い順に第１領域および第２領域を取得する取得ステップ、
　前記色分類装置が、前記第１領域と前記第２領域とが前記色空間において接しており、かつ、前記第１領域に分類された画素の数と前記第２領域に分類された画素の数との比率が所定比率を超える、中間色条件を満たすか否かを判別する判別ステップ、
　前記色分類装置が、前記中間色条件を満たせば、前記商品を、前記第１領域に対応付けられた第１代表色および前記第２領域に対応付けられた第２代表色のいずれにても検索可能とする検索ステップ、
　前記色分類装置が、前記第１代表色にて前記商品が検索された後の検索者の前記商品に対する第１反応と、前記第２代表色にて前記商品が検索された後の検索者の前記商品に対する第２反応と、に基づいて、前記第１領域と前記第２領域との境界を変化させるように、前記領域情報における前記領域の範囲を修正する修正ステップ、
　を備えることを特徴とする。 The color classification method according to the second aspect of the present invention is:
A color classification method in a color classification apparatus having a storage unit that stores area information for associating a range of the area with a representative color indicating the area, which is area information about a plurality of areas that divide the color space,
A classification step in which the color classification device classifies pixels of products included in a product image according to which of the plurality of regions a color of the pixels belongs;
An acquisition step in which the color classification device acquires the first region and the second region in descending order of the number of pixels classified into each of the plurality of regions;
In the color classification device, the first area and the second area are in contact with each other in the color space, and the number of pixels classified into the first area and the number of pixels classified into the second area A determination step of determining whether or not the intermediate color condition is satisfied, wherein the ratio exceeds a predetermined ratio;
If the color classification device satisfies the intermediate color condition, the product is searched for in either the first representative color associated with the first region or the second representative color associated with the second region. Search step to enable,
The color classification device includes a first response to the product of the searcher after the product is searched with the first representative color and a searcher after the product is searched with the second representative color. A correction step of correcting a range of the region in the region information so as to change a boundary between the first region and the second region based on a second response to the product;
It is characterized by providing.
　本発明の第３の観点に係るプログラムは、
　色空間を分割する複数の領域についての領域情報であり、当該領域の範囲と当該領域を示す代表色とを対応付ける領域情報を記憶する記憶部を有するコンピュータを、
　商品画像に含まれる商品についての画素を、当該画素が有する色が前記複数の領域のいずれに属するかにより、分類する分類部、
　前記複数の領域のそれぞれに分類された画素の数が多い順に第１領域および第２領域を取得する取得部、
　前記第１領域と前記第２領域とが前記色空間において接しており、かつ、前記第１領域に分類された画素の数と前記第２領域に分類された画素の数との比率が所定比率を超える、中間色条件を満たすか否かを判別する判別部、
　前記中間色条件を満たせば、前記商品を、前記第１領域に対応付けられた第１代表色および前記第２領域に対応付けられた第２代表色のいずれにても検索可能とする検索部、
　前記第１代表色にて前記商品が検索された後の検索者の前記商品に対する第１反応と、前記第２代表色にて前記商品が検索された後の検索者の前記商品に対する第２反応と、に基づいて、前記第１領域と前記第２領域との境界を変化させるように、前記領域情報における前記領域の範囲を修正する修正部、
　として機能させることを特徴とする。 The program according to the third aspect of the present invention is:
A computer having a storage unit that stores area information about a plurality of areas that divide the color space and that associates the range of the area with a representative color indicating the area,
A classifying unit that classifies pixels for products included in the product image according to which of the plurality of regions the color of the pixels belongs;
An acquisition unit that acquires the first region and the second region in descending order of the number of pixels classified into each of the plurality of regions;
The first area and the second area are in contact with each other in the color space, and a ratio between the number of pixels classified into the first area and the number of pixels classified into the second area is a predetermined ratio. A determination unit that determines whether or not the intermediate color condition is satisfied,
A search unit capable of searching for the product in any of the first representative color associated with the first region and the second representative color associated with the second region, if the intermediate color condition is satisfied;
A first response to the product of the searcher after the product is searched with the first representative color and a second response to the product of the searcher after the product is searched with the second representative color And a correction unit that corrects a range of the region in the region information so as to change a boundary between the first region and the second region,
It is made to function as.
　上記プログラムは、当該プログラムが実行されるコンピュータとは独立して、コンピュータ通信網を介して配布・販売することができる。 The above program can be distributed and sold via a computer communication network independently of the computer on which the program is executed.
　本発明の第４の観点に係るコンピュータ読取可能な記録媒体は、
　色空間を分割する複数の領域についての領域情報であり、当該領域の範囲と当該領域を示す代表色とを対応付ける領域情報を記憶する記憶部を有するコンピュータを、
　商品画像に含まれる商品についての画素を、当該画素が有する色が前記複数の領域のいずれに属するかにより、分類する分類部、
　前記複数の領域のそれぞれに分類された画素の数が多い順に第１領域および第２領域を取得する取得部、
　前記第１領域と前記第２領域とが前記色空間において接しており、かつ、前記第１領域に分類された画素の数と前記第２領域に分類された画素の数との比率が所定比率を超える、中間色条件を満たすか否かを判別する判別部、
　前記中間色条件を満たせば、前記商品を、前記第１領域に対応付けられた第１代表色および前記第２領域に対応付けられた第２代表色のいずれにても検索可能とする検索部、
　前記第１代表色にて前記商品が検索された後の検索者の前記商品に対する第１反応と、前記第２代表色にて前記商品が検索された後の検索者の前記商品に対する第２反応と、に基づいて、前記第１領域と前記第２領域との境界を変化させるように、前記領域情報における前記領域の範囲を修正する修正部、
　として機能させることを特徴とするプログラムを記録する。 A computer-readable recording medium according to a fourth aspect of the present invention is provided.
A computer having a storage unit that stores area information about a plurality of areas that divide the color space and that associates the range of the area with a representative color indicating the area,
A classifying unit that classifies pixels for products included in the product image according to which of the plurality of regions the color of the pixels belongs;
An acquisition unit that acquires the first region and the second region in descending order of the number of pixels classified into each of the plurality of regions;
The first area and the second area are in contact with each other in the color space, and a ratio between the number of pixels classified into the first area and the number of pixels classified into the second area is a predetermined ratio. A determination unit that determines whether or not the intermediate color condition is satisfied,
A search unit capable of searching for the product in any of the first representative color associated with the first region and the second representative color associated with the second region, if the intermediate color condition is satisfied;
A first response to the product of the searcher after the product is searched with the first representative color and a second response to the product of the searcher after the product is searched with the second representative color And a correction unit that corrects a range of the region in the region information so as to change a boundary between the first region and the second region,
A program characterized by functioning as a program is recorded.
　上記記録媒体は、非一時的な（non-transitory）記録媒体であってよく、コンピュータとは独立して配布・販売することができる。ここで、非一時的な記録媒体とは、有形な（tangible）記録媒体をいう。非一時的な記録媒体は、例えば、コンパクトディスク、フレキシブルディスク、ハードディスク、光磁気ディスク、ディジタルビデオディスク、磁気テープ、半導体メモリ等である。また、一時的な（transitory）記録媒体とは、伝送媒体（伝搬信号）それ自体を示す。一時的な記録媒体は、例えば、電気信号、光信号、電磁波等である。なお、一時的な（temporary）記憶領域とは、データやプログラムを一時的に記憶するための領域であり、例えば、ＲＡＭ（Random Access Memory）等の揮発性メモリである。 The recording medium may be a non-transitory recording medium and can be distributed and sold independently of the computer. Here, the non-temporary recording medium refers to a tangible recording medium. Non-temporary recording media are, for example, compact disks, flexible disks, hard disks, magneto-optical disks, digital video disks, magnetic tapes, semiconductor memories, and the like. The transitory recording medium refers to the transmission medium (propagation signal) itself. The temporary recording medium is, for example, an electric signal, an optical signal, an electromagnetic wave, or the like. The temporary storage area is an area for temporarily storing data and programs, and is, for example, a volatile memory such as a RAM (Random Access Memory).
　本発明によれば、商品の色を適切に分類することができる。 According to the present invention, it is possible to appropriately classify product colors.
本実施形態に係る販売システムの全体構成の一例を示すブロック図である。It is a block diagram which shows an example of the whole structure of the sales system which concerns on this embodiment. 色分類装置、及び、ユーザ端末が実現される典型的な情報処理装置の概要構成の一例を示すブロック図である。It is a block diagram which shows an example of schematic structure of the typical information processing apparatus with which a color classification device and a user terminal are implement | achieved. 本実施形態に係る色分類装置の概要構成の一例を示すブロック図である。It is a block diagram which shows an example of schematic structure of the color classification apparatus which concerns on this embodiment. 色空間が各領域に分けられている様子を説明するための模式図である。It is a schematic diagram for demonstrating a color space being divided into each area | region. 「ホワイト」の領域を説明するための模式図である。FIG. 6 is a schematic diagram for explaining a “white” region. 「レッド」の領域を説明するための模式図である。It is a schematic diagram for demonstrating the area | region of "red". 商品画像をグリッド化する様子を説明するための模式図である。It is a schematic diagram for demonstrating a mode that a merchandise image is made into a grid. ポップ画像が含まれている商品画像の一例を示す模式図である。It is a schematic diagram which shows an example of the product image containing the pop image. 背景を判別する手法を説明するための模式図である。It is a schematic diagram for demonstrating the method of discriminating a background. 背景を判別する手法を説明するための模式図である。It is a schematic diagram for demonstrating the method of discriminating a background. 背景を判別する手法を説明するための模式図である。It is a schematic diagram for demonstrating the method of discriminating a background. 背景を判別する手法を説明するための模式図である。It is a schematic diagram for demonstrating the method of discriminating a background. 背景を判別する手法を説明するための模式図である。It is a schematic diagram for demonstrating the method of discriminating a background. 背景を判別する手法を説明するための模式図である。It is a schematic diagram for demonstrating the method of discriminating a background. 領域情報における領域同士の境界を変動させる様子を説明するための模式図である。It is a schematic diagram for demonstrating a mode that the boundary of the area | regions in area | region information is fluctuate | varied. 本実施形態に係る色分類処理を説明するためのフローチャートである。It is a flowchart for demonstrating the color classification process which concerns on this embodiment.
　以下に本発明の実施形態を説明する。本発明の実施形態では、多くのショップ（仮想店舗）が集まって構成される販売サイト（ショッピングモールサイト）を一例として説明するが、１つのショップ（業者）による販売サイトにおいても適宜適用可能である。
　なお、以下の実施形態は説明のためのものであり、本願発明の範囲を制限するものではない。したがって、当業者であればこれらの各要素または全要素をこれと均等なものに置換した実施形態を採用することが可能であるが、これらの実施形態も本発明の範囲に含まれる。 Embodiments of the present invention will be described below. In the embodiment of the present invention, a sales site (shopping mall site) composed of many shops (virtual stores) will be described as an example. However, the present invention can be appropriately applied to a sales site by one shop (trader). .
In addition, the following embodiment is for description and does not limit the scope of the present invention. Therefore, those skilled in the art can employ embodiments in which each or all of these elements are replaced with equivalent ones, and these embodiments are also included in the scope of the present invention.
（全体構成）
　本発明の実施形態に係る販売システム１００は、図１に示すように、色分類装置２００と各ユーザ端末３００とがインターネット９００を介して接続されて構成される。なお、図中では簡略化しているが、ユーザ端末３００は、利用するユーザに応じて、多数存在しているものとする。 (overall structure)
A sales system 100 according to an embodiment of the present invention is configured by connecting a color classification device 200 and each user terminal 300 via the Internet 900 as shown in FIG. Although simplified in the figure, it is assumed that there are a large number of user terminals 300 according to the users to be used.
　色分類装置２００は、例えば、販売サーバ（サーバコンピュータ）等からなり、ユーザ端末３００に対して商品に関する情報を提示すると共に、ユーザが希望する商品を販売する。
　一例として、色分類装置２００は、アクセスしてきたユーザ端末３００に、商品の一覧ページ等を提示する。後述するように、商品についての商品情報には、商品の色を示すカラータグが付与されており、ユーザ（検索者）が任意のカラータグを選択すると、色分類装置２００は、そのカラータグにて絞り込んだ商品の一覧ページをユーザ端末３００に提示する。 The color classification device 200 includes, for example, a sales server (server computer) or the like, and presents information related to products to the user terminal 300 and sells products desired by the user.
As an example, the color classification device 200 presents a product list page or the like to the accessing user terminal 300. As will be described later, the product information about the product is provided with a color tag indicating the color of the product. When the user (searcher) selects an arbitrary color tag, the color classification device 200 displays the color tag. The product list page narrowed down is presented to the user terminal 300.
　ユーザ端末３００は、例えば、パソコンやスマートフォン等からなり、インターネット９００を介して色分類装置２００にアクセスし、商品に関する情報を取得すると共に、ユーザの操作を受け付け、色分類装置２００から所望の商品の購入を行う。 The user terminal 300 includes, for example, a personal computer, a smartphone, and the like, accesses the color classification device 200 via the Internet 900, acquires information about the product, accepts a user operation, and receives a user operation from the color classification device 200. Make a purchase.
（情報処理装置の概要構成）
　本発明の実施形態に係る色分類装置２００、及び、ユーザ端末３００が実現される典型的な情報処理装置４００について説明する。 (Outline configuration of information processing device)
A typical information processing apparatus 400 in which the color classification apparatus 200 and the user terminal 300 according to the embodiment of the present invention are realized will be described.
　情報処理装置４００は、図２に示すように、ＣＰＵ（Central Processing Unit）４０１と、ＲＯＭ（Read Only Memory）４０２と、ＲＡＭ（Random Access Memory）４０３と、ＮＩＣ（Network Interface Card）４０４と、画像処理部４０５と、音声処理部４０６と、ＤＶＤ－ＲＯＭ（Digital Versatile Disc ROM）ドライブ４０７と、インターフェース４０８と、外部メモリ４０９と、コントローラ４１０と、モニタ４１１と、スピーカ４１２と、を備える。 As shown in FIG. 2, the information processing apparatus 400 includes a CPU (Central Processing Unit) 401, a ROM (Read Only Memory) 402, a RAM (Random Access Memory) 403, a NIC (Network Interface Card) 404, an image A processing unit 405, an audio processing unit 406, a DVD-ROM (Digital Versatile Disc Disc ROM) drive 407, an interface 408, an external memory 409, a controller 410, a monitor 411, and a speaker 412 are provided.
　ＣＰＵ　４０１は、情報処理装置４００全体の動作を制御し、各構成要素と接続され制御信号やデータをやりとりする。 The CPU 401 controls the overall operation of the information processing apparatus 400 and is connected to each component to exchange control signals and data.
　ＲＯＭ　４０２には、電源投入直後に実行されるＩＰＬ（Initial Program Loader）が記録され、これが実行されることにより、所定のプログラムをＲＡＭ　４０３に読み出してＣＰＵ　４０１による当該プログラムの実行が開始される。また、ＲＯＭ　４０２には、情報処理装置４００全体の動作制御に必要なオペレーティングシステムのプログラムや各種のデータが記録される。 The ROM 402 records an IPL (Initial Program Loader) that is executed immediately after the power is turned on, and when this is executed, a predetermined program is read into the RAM 403 and the CPU 401 starts executing the program. The ROM 402 stores an operating system program and various data necessary for operation control of the information processing apparatus 400 as a whole.
　ＲＡＭ　４０３は、データやプログラムを一時的に記憶するためのもので、ＤＶＤ－ＲＯＭから読み出したプログラムやデータ、その他、通信に必要なデータ等が保持される。 The RAM 403 is for temporarily storing data and programs, and holds programs and data read from the DVD-ROM and other data necessary for communication.
　ＮＩＣ　４０４は、情報処理装置４００をインターネット等のコンピュータ通信網に接続するためのものであり、ＬＡＮ（Local Area Network）を構成する際に用いられる１０ＢＡＳＥ－Ｔ／１００ＢＡＳＥ－Ｔ規格にしたがうものや、電話回線を用いてインターネットに接続するためのアナログモデム、ＩＳＤＮ（Integrated Services Digital Network）モデム、ＡＤＳＬ（Asymmetric Digital Subscriber Line）モデム、ケーブルテレビジョン回線を用いてインターネットに接続するためのケーブルモデム等と、これらとＣＰＵ　４０１との仲立ちを行うインターフェース（図示せず）等により構成される。 The NIC 404 is used to connect the information processing apparatus 400 to a computer communication network such as the Internet, and conforms to the 10BASE-T / 100BASE-T standard used when configuring a LAN (Local Area Network). Analog modem for connecting to the Internet using a telephone line, ISDN (Integrated Services Digital Network) modem, ADSL (Asymmetric Digital Subscriber Line) modem, cable modem for connecting to the Internet using a cable television line, etc. These are configured by an interface (not shown) that mediates between these and the CPU 401.
　 画像処理部４０５は、ＤＶＤ－ＲＯＭ等から読み出されたデータをＣＰＵ　４０１や画像処理部４０５が備える画像演算プロセッサ（図示せず）によって加工処理した後、これを画像処理部４０５が備えるフレームメモリ（図示せず）に記録する。フレームメモリに記録された画像情報は、所定の同期タイミングでビデオ信号に変換され、モニタ４１１に出力される。これにより、各種のページ表示が可能となる。 The image processing unit 405 processes the data read from the DVD-ROM or the like by a CPU 401 or an image arithmetic processor (not shown) provided in the image processing unit 405, and then processes this data in a frame memory provided in the image processing unit 405. (Not shown). The image information recorded in the frame memory is converted into a video signal at a predetermined synchronization timing and output to the monitor 411. Thereby, various page displays are possible.
　音声処理部４０６は、ＤＶＤ－ＲＯＭ等から読み出した音声データをアナログ音声信号に変換し、これに接続されたスピーカ４１２から出力させる。また、ＣＰＵ　４０１の制御の下、情報処理装置４００が行う処理の進行の中で発生させるべき音を生成し、これに対応した音声をスピーカ４１２から出力させる。 The audio processing unit 406 converts audio data read from a DVD-ROM or the like into an analog audio signal, and outputs the analog audio signal from the speaker 412 connected thereto. Further, under the control of the CPU 401, a sound to be generated during the progress of the processing performed by the information processing apparatus 400 is generated, and a sound corresponding to the sound is output from the speaker 412.
　ＤＶＤ－ＲＯＭドライブ４０７に装着されるＤＶＤ－ＲＯＭには、例えば、実施形態に係る色分類装置２００等を実現するためのプログラムが記憶される。ＣＰＵ　４０１の制御によって、ＤＶＤ－ＲＯＭドライブ４０７は、これに装着されたＤＶＤ－ＲＯＭに対する読み出し処理を行って、必要なプログラムやデータを読み出し、これらはＲＡＭ　４０３等に一時的に記憶される。 The DVD-ROM mounted on the DVD-ROM drive 407 stores, for example, a program for realizing the color classification device 200 according to the embodiment. Under the control of the CPU 401, the DVD-ROM drive 407 performs a reading process on the DVD-ROM loaded therein, reads necessary programs and data, and these are temporarily stored in the RAM 403 or the like.
　インターフェース４０８には、外部メモリ４０９、コントローラ４１０、モニタ４１１、及びスピーカ４１２が、着脱可能に接続される。 External memory 409, controller 410, monitor 411, and speaker 412 are detachably connected to interface 408.
　外部メモリ４０９には、ユーザの個人情報に関するデータなどが書き換え可能に記憶される。 The external memory 409 stores data relating to the user's personal information in a rewritable manner.
　コントローラ４１０は、情報処理装置４００の各種の設定時などに行われる操作入力を受け付ける。情報処理装置４００のユーザは、コントローラ４１０を介して指示入力を行うことにより、これらのデータを適宜外部メモリ４０９に記録することができる。 The controller 410 accepts operation inputs performed at various settings of the information processing apparatus 400. The user of the information processing apparatus 400 can record these data in the external memory 409 as appropriate by inputting instructions through the controller 410.
　モニタ４１１は、画像処理部４０５により出力されたデータを情報処理装置４００のユーザに提示する。 The monitor 411 presents the data output by the image processing unit 405 to the user of the information processing apparatus 400.
　スピーカ４１２は、音声処理部４０６により出力された音声データを情報処理装置４００のユーザに提示する。 The speaker 412 presents the audio data output by the audio processing unit 406 to the user of the information processing apparatus 400.
　この他、情報処理装置４００は、ハードディスク等の大容量外部記憶装置を用いて、ＲＯＭ ４０２、ＲＡＭ ４０３、外部メモリ４０９、ＤＶＤ－ＲＯＭドライブ４０７に装着されるＤＶＤ－ＲＯＭ等と同じ機能を果たすように構成してもよい。 In addition, the information processing apparatus 400 uses a large-capacity external storage device such as a hard disk so as to perform the same functions as the DVD-ROM mounted on the ROM 402, RAM 403, external memory 409, and DVD-ROM drive 407. You may comprise.
　以下、上記情報処理装置４００において実現される色分類装置２００の構成等について、図３～図８を参照して説明する。情報処理装置４００に電源が投入され、例えば、ユーザ端末３００からのアクセスを受け付けると、本実施形態に係る色分類装置２００として機能させるそれぞれのプログラムが実行され、本実施形態に係る色分類装置２００が実現される。
　なお、各ユーザ端末３００も同様に情報処理装置４００において実現されるが、これらの構成については省略し、本実施形態において最も特徴的な色分類装置２００について、以下説明する。 Hereinafter, the configuration and the like of the color classification apparatus 200 realized in the information processing apparatus 400 will be described with reference to FIGS. When the information processing apparatus 400 is turned on, for example, when an access from the user terminal 300 is received, each program causing the information classification apparatus 200 to function as the color classification apparatus 200 according to the present embodiment is executed, and the color classification apparatus 200 according to the present embodiment. Is realized.
Each user terminal 300 is also realized in the information processing apparatus 400 in the same manner, but these configurations are omitted, and the most characteristic color classification apparatus 200 in the present embodiment will be described below.
（色分類装置の概要構成）
　図３は、本実施形態に係る色分類装置２００の概要構成の一例を示すブロック図である。図示するように、色分類装置２００は、受付部２１０と、記憶部２２０と、制御部２３０と、提示部２４０とを備える。 (Outline configuration of color classification device)
FIG. 3 is a block diagram illustrating an example of a schematic configuration of the color classification device 200 according to the present embodiment. As illustrated, the color classification device 200 includes a receiving unit 210, a storage unit 220, a control unit 230, and a presentation unit 240.
　受付部２１０は、インターネット９００を介して各ユーザ端末３００から送られる種々の情報を受け付ける。
　例えば、受付部２１０は、商品の一覧ページへの遷移指示や、一覧ページ中における種々の操作指示（一例として、ユーザによるマウスのクリック操作等）を、ユーザ端末３００から受け付ける。
　上述したＮＩＣ　４０４等が、このような受付部２１０として機能しうる。 The accepting unit 210 accepts various information sent from each user terminal 300 via the Internet 900.
For example, the accepting unit 210 accepts from the user terminal 300 an instruction to transition to a list page of products and various operation instructions in the list page (for example, a mouse click operation by the user).
The above-described NIC 404 or the like can function as such a reception unit 210.
　記憶部２２０は、商品に関する種々の情報や、ユーザに関する種々の情報に加え、色空間を分割する複数の領域についての領域情報を記憶する。
　例えば、記憶部２２０は、図４Ａに示すように、色空間ＣＳを分割する複数の領域（破線にて区分けされる各領域）について、その領域の範囲と領域を示す代表色とを対応付ける領域情報を記憶する。
　以下、一例として、各領域の代表色を、ホワイト、イエロー、オレンジ、ピンク、レッド、ベージュ、シルバー、ゴールド、グレー、パープル、ブラウン、グリーン、ブルー、および、ブラックとする場合について説明するが、これら代表色は一例であり、適宜変更可能である。また、異なる色の縞模様の商品等に対応できるように、代表色に、マルチカラーを含めてもよい。 The storage unit 220 stores area information about a plurality of areas that divide the color space, in addition to various information related to products and various information related to users.
For example, as illustrated in FIG. 4A, the storage unit 220 associates, with respect to a plurality of regions (each region divided by a broken line) dividing the color space CS, the range of the region and a representative color indicating the region. Remember.
Hereinafter, as an example, the case where the representative colors of each area are white, yellow, orange, pink, red, beige, silver, gold, gray, purple, brown, green, blue, and black will be described. The representative color is an example and can be changed as appropriate. In addition, multi-colors may be included in the representative colors so as to be able to deal with striped patterns of different colors.
　具体的に、領域情報において、代表色が「ホワイト」の場合、図４Ｂに示すように、ＲＧＢ値におけるＲ，Ｇ，Ｂ（各８ビット）の全ての値が２３０～２５５の範囲を規定している。なお、純粋な「ホワイト」であれば ＲＧＢ値は、Ｒ：２５５，Ｇ：２５５，Ｂ：２５５であるが、オフホワイト（Ｒ：２５５，Ｇ：２５５，Ｂ：２３０）や、ややグレーがかったホワイト（Ｒ：２３０，Ｇ：２３０，Ｂ：２３０）なども、広義的に「ホワイト」に含めることができる。そこで、ある程度の幅を持たせるため、領域情報において、「ホワイト」の範囲を、Ｒ，Ｇ，Ｂの全ての値が２３０以上と規定している。 Specifically, in the region information, when the representative color is “white”, as shown in FIG. 4B, all values of R, G, and B (8 bits each) in the RGB value define a range of 230 to 255. ing. In the case of pure “white”, the RGB value is R: 255, G: 255, B: 255, but off-white (R: 255, G: 255, B: 230) or slightly grayish. White (R: 230, G: 230, B: 230) and the like can be broadly included in “white”. Therefore, in order to provide a certain amount of width, the range of “white” is defined as all values of R, G, and B being 230 or more in the region information.
　また、代表色が「レッド」の場合、図４Ｃに示すように、ＲＧＢ値におけるＲ：１９６～２５５，Ｇ：０～５０，Ｂ：０～５０の範囲を規定している。なお、純粋な「レッド」であれば、ＲＧＢ値は、Ｒ：２５５，Ｇ：０，Ｂ：０であるが、ある程度の幅を持たせるため、領域情報において、「レッド」の範囲を、１９６≦Ｒ≦２５５，０≦Ｇ≦５０，０≦Ｂ≦５０の全てを満たすものと規定している。 Further, when the representative color is “red”, as shown in FIG. 4C, the range of R: 196 to 255, G: 0 to 50, and B: 0 to 50 in the RGB value is defined. In the case of pure “red”, the RGB values are R: 255, G: 0, and B: 0. However, in order to provide a certain width, the range of “red” is set to 196 in the region information. ≦ R ≦ 255, 0 ≦ G ≦ 50, and 0 ≦ B ≦ 50 are all satisfied.
　同様に、他の代表色にも、領域情報の範囲が規定されている。つまり、他の代表色の領域情報についても、ある程度の幅を持たせた範囲が規定されている。
　このような領域情報は、後述するように、商品の画像（画像データ）から、その商品の色を分類するために使用される。なお、多くの画像は、単色ではない。また、照明の反射などにより白く写ることもある。それらの色の違いを、領域情報において、ある程度の幅を持たせることで、同色に分類されるようにしている。
　また、後述するように、領域情報には、中間色の範囲が規定されていてもよい。
　上述したＲＡＭ　４０３や外部メモリ４０９等が、このような記憶部２２０として機能しうる。 Similarly, the range of area information is defined for other representative colors. That is, a range having a certain width is also defined for the area information of other representative colors.
Such area information is used for classifying the color of the product from the image (image data) of the product, as will be described later. Many images are not monochromatic. Also, it may appear white due to reflection of illumination. These color differences are classified into the same color by giving a certain width in the region information.
Further, as described later, the area information may define a range of intermediate colors.
The above-described RAM 403, external memory 409, and the like can function as such a storage unit 220.
　図３に戻って、制御部２３０は、色分類装置２００全体を制御する。この制御部２３０は、分類部２３１、取得部２３２、判別部２３３、検索部２３４、及び、修正部２３５を含んでおり、ユーザによる商品の購入を支援する。 3, the control unit 230 controls the entire color classification apparatus 200. The control unit 230 includes a classification unit 231, an acquisition unit 232, a determination unit 233, a search unit 234, and a correction unit 235, and supports purchase of products by the user.
　分類部２３１は、商品画像の画素の色がどの領域に属するか分類する。なお、以下では、商品画像を格子状に区分け（グリッド化）して、各マスがどの領域に属するかを分類する場合について説明するが、グリッド化することなく、１ドッド（１ピクセル）の画素について、各画素がどの領域に属するかを分類してもよい。 The classification unit 231 classifies which region the pixel color of the product image belongs to. In the following, a case will be described in which product images are divided into grids (griding) to classify which region each cell belongs to. However, one dod (1 pixel) pixel without gridding is described. The region to which each pixel belongs may be classified.
　分類部２３１は、図５に示すような商品Ｐ（スーツケース）を含んだ商品画像ＰＩに対して、一例として、マス目の数が１００×１００（縦×横）となるようにグリッド化し、１マス毎に色の抽象化を行った後に、各マスがどの領域に属するかを分類する。このような商品画像ＰＩには、通常、背景が含まれているため、分類部２３１は、背景を除いた商品画像ＰＩに対する各マスがどの領域に属するかを分類する。 As an example, the classification unit 231 forms a grid on the product image PI including the product P (suitcase) as illustrated in FIG. 5 so that the number of cells is 100 × 100 (vertical × horizontal). After color abstraction is performed for each square, the area to which each square belongs is classified. Since such a product image PI usually includes a background, the classification unit 231 classifies which region each cell for the product image PI excluding the background belongs to.
　背景には、多くの場合、単色が使用される。また、図６に示すように、商品Ｐの他に、ポップ（ＰＯＰ）画像Ｇを含んだ商品画像ＰＩもあるが、このようなポップ画像Ｇにも単色が使用される場合が多い。これら単色が使用される場合には、隣り合ったマス同士のＲＧＢ値が全く同じになるという特徴がある。そのため、分類部２３１は、隣り合ったマス同士のＲＧＢ値が同じ場合に、背景やポップ画像Ｇであると判別し、分類対象から除外する。 In many cases, a single color is used for the background. As shown in FIG. 6, in addition to the product P, there is a product image PI including a pop (POP) image G, and a single color is often used for such a pop image G. When these single colors are used, there is a feature that the RGB values of adjacent cells are exactly the same. For this reason, the classification unit 231 determines that the adjacent squares have the same RGB value as a background or a pop image G, and excludes it from the classification target.
　また、単色の背景の代わりに、利用シーンをイメージした背景が使用される場合もある。この場合、分類部２３１は、商品Ｐが商品画像ＰＩの周囲（上下左右の４辺）に２５％以上掛かることがないという特徴を利用して、背景を除外する。
　例えば、図７Ａに例示する商品画像ＰＩでは、商品Ｐの下部が、商品画像ＰＩの１辺に掛かっている場合を示す。この図７Ａでは、商品Ｐの下部が商品画像ＰＩの下辺に３０％ほど掛かっているものの、４辺全体を基準とした割合では、およそ１２％程度となる。
　また、図７Ｂに例示する商品画像ＰＩでは、商品Ｐの一部が、商品画像ＰＩの２辺に掛かっている場合を示す。この図７Ｂでは、商品Ｐの下部が商品画像ＰＩの下辺に３０％ほど掛かり、商品Ｐの右部が商品画像ＰＩの右辺に１０％ほど掛かっているものの、４辺全体を基準とした割合では、およそ１６％程度となる。 In some cases, a background reflecting the use scene is used instead of a single color background. In this case, the classification unit 231 excludes the background using the feature that the product P does not hang over 25% or more around the product image PI (upper, lower, left, and right sides).
For example, the product image PI illustrated in FIG. 7A shows a case where the lower part of the product P is hooked on one side of the product image PI. In FIG. 7A, although the lower part of the product P is hung on the lower side of the product image PI by about 30%, the ratio based on the entire four sides is about 12%.
Moreover, in the product image PI illustrated in FIG. 7B, a case where a part of the product P is hung on two sides of the product image PI is shown. In FIG. 7B, the lower part of the product P is about 30% on the lower side of the product image PI, and the right part of the product P is about 10% on the right side of the product image PI. About 16%.
　なお、商品Ｐが商品画像ＰＩの４辺に２５％以上掛かる場合として、図７Ｃや図７Ｄが挙げられるが、極めて不自然となり、現実には使用されないと考えられる。実際に、出願人が管理している販売サイトにおいて、約１，０００件の商品画像を無作為に抽出して検証した限りにおいて、存在しなかった。
　また、図７Ｅに例示する商品画像ＰＩのように、画像中に複数の商品Ｐが含まれる場合にも、各商品Ｐが商品画像ＰＩの４辺に２５％以上掛かることも考えられる。それでも、色違いの各商品Ｐを１つの商品画像ＰＩ中に含めてしまうと、どちらがどの色の商品なのかが分かり難くなるため、現実には使用されないと考えられる。
　これらのことから、分類部２３１は、商品画像ＰＩの４辺に２５％以上掛かる場合に、背景（利用シーンをイメージした背景）であると判別し、分類対象から除外する。 Note that FIG. 7C and FIG. 7D are examples of the case where the product P takes 25% or more on the four sides of the product image PI, but it is extremely unnatural and is not considered to be actually used. In fact, as long as about 1,000 product images were randomly extracted and verified at the sales site managed by the applicant, they did not exist.
Further, even when a plurality of products P are included in the image as in the product image PI illustrated in FIG. 7E, it is also conceivable that each product P takes 25% or more on four sides of the product image PI. Even so, if each product P of different colors is included in one product image PI, it is difficult to know which color product is which product P, so it is considered that it is not actually used.
From these things, the classification | category part 231 discriminate | determines that it is a background (background which imaged the utilization scene), and excludes it from a classification | category object, when it covers 25% or more on 4 sides of the product image PI.
　更に、商品画像ＰＩにおいて、商品を引き立たせる必要があるため、通常、商品Ｐの色と背景の色とは、比較的コントラストの大きいカラーバランスが採用されるという特徴もある。
　これらの特徴から、商品画像ＰＩの周囲に沿って、幅１～３ピクセル程度の色を抽出した場合、その中で、７５％以上を占める色が背景色と推定でき、商品画像ＰＩ内に存在する同色の部分が、全て背景である可能性が高いと判断できる。
　例えば、図７Ｆに示す商品画像ＰＩの場合に、商品画像ＰＩの周囲に沿って、幅１～３ピクセル程度の色を抽出すると、Ｌ１で示される色の部分が４辺全体の８８％となり、Ｌ２で示される色の部分が４辺全体の１２％となったとする。この場合、分類部２３１は、４辺全体の８８％を占めるＬ１で示される色と同色の部分を全て背景であると判別して、分類対象から除外する。 Furthermore, since it is necessary to enhance the product in the product image PI, the color of the product P and the background color are usually characterized by a color balance having a relatively large contrast.
Based on these characteristics, when a color with a width of 1 to 3 pixels is extracted along the periphery of the product image PI, a color that occupies 75% or more of the color can be estimated as a background color and exists in the product image PI. It can be determined that there is a high possibility that all the same-colored parts are the background.
For example, in the case of the product image PI shown in FIG. 7F, if a color having a width of about 1 to 3 pixels is extracted along the periphery of the product image PI, the color portion indicated by L1 becomes 88% of the entire four sides. Assume that the color portion indicated by L2 is 12% of the entire four sides. In this case, the classification unit 231 determines that all the same color portions as the color indicated by L1 occupying 88% of the entire four sides are the background, and excludes them from the classification target.
　なお、商品画像ＰＩの中には、商品画像ＰＩの周囲を単色で縁取りする場合もある。このため、商品画像ＰＩの周囲に沿って抽出して、７５％以上を占める色であっても、商品画像ＰＩに占める同色の割合が低いなどの場合には、縁取りである場合も考慮して、別の手法によって背景を判別してもよい。 In some cases, the product image PI may have a single color border around the product image PI. For this reason, even if the color is extracted along the periphery of the product image PI and occupies 75% or more, if the proportion of the same color in the product image PI is low, the case of bordering is also considered. The background may be determined by another method.
　このようにして、分類部２３１は、背景を除いた商品画像ＰＩに対する画素（マス）がどの領域に属するかを分類する。
　具体的に、上述した図５に示す商品Ｐが真っ赤なスーツケースである場合に、分類部２３１は、一例として、２５００のマスを「レッド」に分類し、９００のマスを「グレー」に分類し、２００のマスを「ピンク」に分類する。なお、他のマスも分類されているが省略する。
　また、色違い商品で、上述した図５に示す商品Ｐが赤と橙との中間色のスーツケースである場合に、分類部２３１は、一例として、１４００のマスを「レッド」に分類し、１１００のマスを「オレンジ」に分類し、９００のマスを「グレー」に分類する。なお、他のマスも分類されているが省略する。 In this way, the classification unit 231 classifies which region the pixel (mass) for the product image PI excluding the background belongs to.
Specifically, when the product P shown in FIG. 5 is a red suitcase, the classification unit 231 classifies 2500 cells as “red” and 900 cells as “gray” as an example. Then, the 200 cells are classified as “pink”. Other squares are also classified but omitted.
Further, when the product P shown in FIG. 5 described above is a suitcase having an intermediate color between red and orange with different colors, the classification unit 231 classifies 1400 cells as “red” as an example. Are classified as “orange” and 900 are classified as “gray”. Other squares are also classified but omitted.
　図３に戻って、取得部２３２は、分類部２３１によって分類された画素（マス）の数が多い順に第１領域および第２領域を取得する。
　例えば、分類部２３１によって、「レッド」：２５００，「グレー」：９００，「ピンク」：２００，・・・と分類されている場合に、取得部２３２は、第１領域として「レッド」を取得し、第２領域として「グレー」を取得する。
　また、分類部２３１によって、「レッド」：１４００，「オレンジ」：１１００，「グレー」：９００，・・・と分類されている場合に、取得部２３２は、第１領域として「レッド」を取得し、第２領域として「オレンジ」を取得する。 Returning to FIG. 3, the acquisition unit 232 acquires the first region and the second region in descending order of the number of pixels (mass) classified by the classification unit 231.
For example, when the classification unit 231 classifies “red”: 2500, “gray”: 900, “pink”: 200,..., The acquisition unit 232 acquires “red” as the first region. Then, “gray” is acquired as the second region.
When the classification unit 231 classifies “red”: 1400, “orange”: 1100, “gray”: 900,..., The acquisition unit 232 acquires “red” as the first region. Then, “orange” is acquired as the second area.
　判別部２３３は、取得部２３２によって取得された第１領域および第２領域が、中間色条件を満たしているか否かを判別する。
　例えば、判別部２３３は、第１領域と第２領域とが色空間において接しており、かつ、第１領域に分類された画素（マス）の数と前記第２領域に分類された画素（マス）の数との比率が所定比率（一例として、５：３）を超える場合に、中間色条件を満たしていると判別する。
　具体的に、取得部２３２によって、第１領域として「レッド」が取得され、第２領域として「グレー」が取得されている場合において、「レッド」と「グレー」とが色空間において接していないことから、判別部２３３は、中間色条件を満たしていないと判別する。
　一方、取得部２３２によって、第１領域として「レッド」（マスの数１４００）が取得され、第２領域として「オレンジ」（マスの数１１００）が取得されている場合において、「レッド」と「オレンジ」とが色空間において接しており、かつ、所定比率を超えていることから、判別部２３３は、中間色条件を満たしていると判別する。
　なお、中間色条件は、一例であり、他の条件で判別してもよい。例えば、後述するように、中間色を判定するための範囲を更に定めておき、分類部２３１にて２種類の範囲に従って分類した結果に基づいて、中間色条件を満たしているか否かを判別するようにしてもよい。 The determination unit 233 determines whether or not the first region and the second region acquired by the acquisition unit 232 satisfy the intermediate color condition.
For example, the determination unit 233 includes a first area and a second area that are in contact with each other in the color space, and the number of pixels (mass) classified into the first area and the pixels (mass cells) classified into the second area. ) Is greater than a predetermined ratio (for example, 5: 3), it is determined that the intermediate color condition is satisfied.
Specifically, when “red” is acquired as the first area and “gray” is acquired as the second area by the acquisition unit 232, “red” and “gray” are not in contact with each other in the color space. Therefore, the determination unit 233 determines that the intermediate color condition is not satisfied.
On the other hand, when the acquisition unit 232 acquires “red” (the number of squares 1400) as the first region and “orange” (the number of squares 1100) as the second region, “red” and “ Since “orange” is in contact with the color space and exceeds the predetermined ratio, the determination unit 233 determines that the intermediate color condition is satisfied.
The intermediate color condition is an example, and may be determined based on other conditions. For example, as described later, a range for determining the intermediate color is further determined, and it is determined whether the intermediate color condition is satisfied based on the result of classification by the classification unit 231 according to the two types of ranges. May be.
　検索部２３４は、判別部２３３によって中間色条件を満たすと判別された場合に、当該商品を、第１領域に対応付けられた第１代表色および第２領域に対応付けられた第２代表色のいずれにても検索可能とする。
　例えば、検索部２３４は、検索時に指定された指定色が、第１代表色と一致する際に検索をヒットさせる第１確率と、同指定色が、第２代表色と一致する際に検索をヒットさせる第２確率とに応じて、商品の検索を行う。
　ここで、第１確率は、例えば、「ｆ（第１領域に分類された数）／ｆ（第１領域に分類された数＋第２領域に分類された数）」により算定される。一方、第２確率は、例えば、「ｆ（第２領域に分類された数）／ｆ（第１領域に分類された数＋第２領域に分類された数）」により算定される。
　なお、関数ｆ（）は、例えば、ｆ（ｘ）＝ｘ，ｆ（ｘ）＝　ｘ2，ｆ（ｘ）＝ｌｏｇ2（ｘ），ｆ（ｘ）＝２xなど、１以上の引数ｘに対して単調増加し、非負値を返す関数であれば、適宜採用できる。
　また、第１確率＋第２確率＝１となる場合に限られず、第１確率＋第２確率＞１となる場合や、極端な例では、第１確率＝第２確率＝１となる場合も、適宜採用することができる。 When the determining unit 233 determines that the intermediate color condition is satisfied, the search unit 234 selects the first representative color associated with the first region and the second representative color associated with the second region. Search is possible in any case.
For example, the search unit 234 searches for a first probability of hitting the search when the specified color specified during the search matches the first representative color, and when the specified color matches the second representative color. The product is searched according to the second probability to be hit.
Here, the first probability is calculated by, for example, “f (number classified into the first region) / f (number classified into the first region + number classified into the second region)”. On the other hand, the second probability is calculated by, for example, “f (the number classified into the second region) / f (the number classified into the first region + the number classified into the second region)”.
Note that the function f () has one or more arguments x such as f (x) = x, f (x) = x 2 , f (x) = log 2 (x), f (x) = 2 x, etc. Any function that monotonically increases and returns a non-negative value can be used as appropriate.
In addition, the present invention is not limited to the case where the first probability + the second probability = 1, and the case where the first probability + the second probability> 1 or in the extreme example, the case where the first probability = the second probability = 1. Can be employed as appropriate.
　このように、中間色条件を満たすと判別されると、検索部２３４は、検索時に指定された指定色が第１代表色と一致する場合に、第１確率に基づいて当該商品をヒットさせ、若しくは、同指定色が第２代表色と一致する場合に、第２確率に基づいて当該商品をヒットさせる。そのため、指定色が第１代表色と一致する際に、確率的に有利となり、当該商品が検索され易くなるものの、同指定色が第２代表色と一致する際にも、適宜当該商品が検索されることになる。 As described above, when it is determined that the intermediate color condition is satisfied, the search unit 234 makes the product hit based on the first probability when the specified color specified at the time of the search matches the first representative color, or When the designated color matches the second representative color, the product is hit based on the second probability. For this reason, when the designated color matches the first representative color, it is probabilistically advantageous and the product is easily searched. However, when the designated color matches the second representative color, the product is appropriately searched. Will be.
　なお、判別部２３３によって中間色条件を満たさないと判別された場合に、検索部２３４は、第１領域に対応付けられた第１代表色が指定された際に限り、当該商品を検索するものとする。 When the determination unit 233 determines that the intermediate color condition is not satisfied, the search unit 234 searches for the product only when the first representative color associated with the first region is designated. To do.
　修正部２３５は、第１代表色にて商品が検索された後のユーザ（検索者）による商品に対する第１反応と、第２代表色にて商品が検索された後のユーザによる商品に対する第２反応と、に基づいて、第１領域と第２領域との境界を変化させるように、領域情報における領域の範囲を修正する。
　例えば、上述した図５に示す商品Ｐが赤と橙との中間色のスーツケースである場合において、「レッド」のスーツケースについての検索がユーザに指示され、検索部２３４が、このスーツケースを検索結果（検索結果一覧）に含めて提示した後に、ユーザからこのスーツケースが選択された（商品ページへの遷移が指示された）数と、「オレンジ」のスーツケースについての検索がユーザに指示され、検索部２３４が、このスーツケースを検索結果に含めて提示した後に、ユーザからこのスーツケースが選択された数と、に基づいて、「レッド」と「オレンジ」との境界を変化させるように、領域情報における領域の範囲を修正する。 The correcting unit 235 performs a first reaction on a product by the user (searcher) after the product is searched for with the first representative color, and a second response for the product by the user after the product is searched for with the second representative color. Based on the reaction, the range of the region in the region information is modified so that the boundary between the first region and the second region is changed.
For example, when the product P shown in FIG. 5 is a suitcase having an intermediate color between red and orange, the user is instructed to search for a “red” suitcase, and the search unit 234 searches for this suitcase. After presenting it in the results (list of search results), the user is instructed to search for the suitcase of “orange” and the number of cases where this suitcase was selected by the user (instructed to transition to the product page). The search unit 234 changes the boundary between “red” and “orange” based on the number of the suitcases selected by the user after presenting the suitcases in the search results. The range of the area in the area information is corrected.
　具体的には、第２代表色である「オレンジ」を検索結果に含めたときの方が、第１代表色である「レッド」を検索結果に含めたときよりも、このスーツケースがユーザからより多く選択された場合に、修正部２３５は、図８に示すように、「レッド」の領域ＲＡが狭くなり、その分「オレンジ」の領域ＯＡが広がるように、領域情報における領域の範囲を修正する。これにより、今後、赤と橙との中間色の商品は、「レッド」よりも「オレンジ」により多く分類され、指定色に「オレンジ」が指定された際により検索され易くなる。
　一方、第１代表色である「レッド」を検索結果に含めたときの方が、第２代表色である「オレンジ」を検索結果に含めたときよりも、このスーツケースがユーザからより多く選択された場合に、修正部２３５は、「レッド」の領域と「オレンジ」の領域との境界を初期の境界に戻すように、領域情報における領域の範囲を修正する。つまり、図８に示すように、「レッド」の領域ＲＡが狭くなり、その分「オレンジ」の領域ＯＡが広がっていれば、互いの領域を元に戻す。なお、「レッド」の領域ＲＡと「オレンジ」の領域ＯＡとが初期の領域のままであれば、修正部２３５は、何ら修正を行わない。 Specifically, when the second representative color “orange” is included in the search results, the suitcase is received from the user rather than when the first representative color “red” is included in the search results. When more selections are made, the correction unit 235 reduces the range of the region in the region information so that the “red” region RA is narrowed and the “orange” region OA is expanded accordingly, as shown in FIG. Correct it. As a result, products in the middle color between red and orange will be more classified into “orange” than “red” in the future, and it will be easier to search when “orange” is designated as the designated color.
On the other hand, when the first representative color “red” is included in the search results, this suitcase is selected more by the user than when the second representative color “orange” is included in the search results. In this case, the correction unit 235 corrects the range of the region in the region information so that the boundary between the “red” region and the “orange” region is returned to the initial boundary. That is, as shown in FIG. 8, if the “red” area RA becomes narrower and the “orange” area OA expands by that amount, the areas are returned to the original. If the “red” area RA and the “orange” area OA remain the initial areas, the correction unit 235 does not perform any correction.
　このような構成の制御部２３０は、中間色の商品について、いわゆるＡ／Ｂテストを適宜行うことになり、その結果（ユーザの反応）に応じて、対応する領域同士の境界を変動させる。これによって、中間色の商品は、自然と反応のよい方の代表色に分類され、ユーザに提示され易くなる。
　上述したＣＰＵ　４０１等が、このような構成からなる制御部２３０として機能しうる。 The control unit 230 having such a configuration appropriately performs a so-called A / B test on the intermediate color product, and changes the boundary between corresponding regions according to the result (user reaction). As a result, the intermediate color products are classified into the representative colors that are more responsive to nature, and are easily presented to the user.
The above-described CPU 401 and the like can function as the control unit 230 having such a configuration.
　提示部２４０は、インターネット９００を介してユーザ端末３００に種々の情報を提示する。
　例えば、提示部２４０は、制御部２３０の検索部２３４が商品を検索した検索結果（検索結果一覧）をユーザ端末３００に提示する。すなわち、ユーザが指定色を指定して商品の検索を指示した際に、指定色で絞り込んだ検索結果一覧を、ユーザ端末３００に提示する。 The presentation unit 240 presents various information to the user terminal 300 via the Internet 900.
For example, the presentation unit 240 presents the search results (search result list) that the search unit 234 of the control unit 230 has searched for products to the user terminal 300. That is, when the user designates a designated color and instructs to search for a product, a search result list narrowed down by the designated color is presented to the user terminal 300.
　なお、上述したＣＰＵ　４０１及びＮＩＣ　４０４等が、このような提示部２４０として機能しうる。 Note that the above-described CPU 401, NIC 404, and the like can function as such a presentation unit 240.
（色分類装置の動作）
　以下、このような構成の色分類装置２００の動作について図９を参照して説明する。図９は、色分類装置２００が実行する色分類処理の流れを示すフローチャートである。この色分類処理は、例えば、ユーザが指定色を指定して商品の検索を指示した際に開始される。 (Operation of color classification device)
Hereinafter, the operation of the color classification apparatus 200 having such a configuration will be described with reference to FIG. FIG. 9 is a flowchart showing the flow of color classification processing executed by the color classification device 200. This color classification process is started, for example, when the user designates a designated color and instructs to search for a product.
　まず、色分類装置２００は、商品画像の画素（マス）の色がどの領域に属するか分類する（ステップＳ１１）。
　例えば、制御部２３０（分類部２３１）は、上述した図５に示すような商品Ｐ（スーツケース）を含んだ商品画像ＰＩに対して、マス目の数が１００×１００（縦×横）となるようにグリッド化し、１マス毎に色の抽象化を行った後に、各マスがどの領域に属するかを分類する。
　なお、上述したように、商品画像ＰＩには、背景が含まれているため、分類部２３１は、背景を除いた商品画像ＰＩに対する各マスがどの領域に属するかを分類する。 First, the color classification device 200 classifies which region the color of the pixel (mass) of the product image belongs to (step S11).
For example, the control unit 230 (classification unit 231) determines that the number of cells is 100 × 100 (vertical × horizontal) for the product image PI including the product P (suitcase) as shown in FIG. After the grid is formed and color abstraction is performed for each cell, the region to which each cell belongs is classified.
As described above, since the product image PI includes a background, the classification unit 231 classifies which area each cell for the product image PI excluding the background belongs to.
　具体的に、上述した図５に示す商品Ｐが真っ赤なスーツケースである場合に、分類部２３１は、一例として、２５００のマスを「レッド」に分類し、９００のマスを「グレー」に分類し、２００のマスを「ピンク」に分類する。なお、他のマスも分類されているが省略する。
　また、色違い商品で、上述した図５に示す商品Ｐが赤と橙との中間色のスーツケースである場合に、分類部２３１は、一例として、１４００のマスを「レッド」に分類し、１１００のマスを「オレンジ」に分類し、９００のマスを「グレー」に分類する。なお、他のマスも分類されているが省略する。 Specifically, when the product P shown in FIG. 5 is a red suitcase, the classification unit 231 classifies 2500 cells as “red” and 900 cells as “gray” as an example. Then, the 200 cells are classified as “pink”. Other squares are also classified but omitted.
Further, when the product P shown in FIG. 5 described above is a suitcase having an intermediate color between red and orange with different colors, the classification unit 231 classifies 1400 cells as “red” as an example. Are classified as “orange” and 900 are classified as “gray”. Other squares are also classified but omitted.
　色分類装置２００は、分類された画素（マス）の数が多い順に第１領域および第２領域を取得する（ステップＳ１２）。
　具体的に、ステップＳ１１にて、「レッド」：２５００，「グレー」：９００，「ピンク」：２００，・・・と分類されている場合に、制御部２３０（取得部２３２）は、第１領域として「レッド」を取得し、第２領域として「グレー」を取得する。
　また、ステップＳ１１にて、「レッド」：１４００，「オレンジ」：１１００，「グレー」：９００，・・・と分類されている場合に、取得部２３２は、第１領域として「レッド」を取得し、第２領域として「オレンジ」を取得する。 The color classification device 200 acquires the first area and the second area in descending order of the number of classified pixels (mass) (step S12).
Specifically, in step S11, when the classification is “red”: 2500, “gray”: 900, “pink”: 200,..., The control unit 230 (acquisition unit 232) “Red” is acquired as the area, and “gray” is acquired as the second area.
In Step S11, when it is classified as “red”: 1400, “orange”: 1100, “gray”: 900,..., The acquisition unit 232 acquires “red” as the first area. Then, “orange” is acquired as the second area.
　色分類装置２００は、第１領域および第２領域が、中間色条件を満たしているか否かを判別する（ステップＳ１３）。
　例えば、制御部２３０（判別部２３３）は、第１領域と第２領域とが色空間において接しており、かつ、第１領域に分類された画素（マス）の数と前記第２領域に分類された画素（マス）の数との比率が所定比率（一例として、５：３）を超える場合に、中間色条件を満たしていると判別する。
　具体的に、ステップＳ１２にて、第１領域として「レッド」が取得され、第２領域として「グレー」が取得されている場合において、「レッド」と「グレー」とが色空間において接していないことから、判別部２３３は、中間色条件を満たしていないと判別する。
　一方、ステップＳ１２にて、第１領域として「レッド」（マスの数１４００）が取得され、第２領域として「オレンジ」（マスの数１１００）が取得されている場合において、「レッド」と「オレンジ」とが色空間において接しており、かつ、所定比率を超えていることから、判別部２３３は、中間色条件を満たしていると判別する。 The color classification device 200 determines whether or not the first area and the second area satisfy the intermediate color condition (step S13).
For example, the control unit 230 (determination unit 233) has the first area and the second area in contact with each other in the color space, and the number of pixels (mass) classified in the first area and the second area are classified. When the ratio with the number of pixels (mass) that has been exceeded exceeds a predetermined ratio (for example, 5: 3), it is determined that the intermediate color condition is satisfied.
Specifically, when “red” is acquired as the first area and “gray” is acquired as the second area in step S12, “red” and “gray” are not in contact with each other in the color space. Therefore, the determination unit 233 determines that the intermediate color condition is not satisfied.
On the other hand, when “red” (number of squares 1400) is acquired as the first area and “orange” (number of squares 1100) is acquired as the second area in step S12, “red” and “ Since “orange” is in contact with the color space and exceeds the predetermined ratio, the determination unit 233 determines that the intermediate color condition is satisfied.
　色分類装置２００は、第１領域および第２領域が、中間色条件を満たしていると判別すると（ステップＳ１３；Ｙｅｓ）、指定色が第１代表色と一致する際に検索をヒットさせる第１確率と、同指定色が第２代表色と一致する際に検索をヒットさせる第２確率とに応じて、商品の検索を行う（ステップＳ１４）。
　すなわち、制御部２３０（検索部２３４）は、検索時に指定された指定色が第１代表色と一致する場合に、第１確率に基づいて当該商品をヒットさせ、若しくは、同指定色が第２代表色と一致する場合に、第２確率に基づいて当該商品をヒットさせる。 If the color classification device 200 determines that the first region and the second region satisfy the intermediate color condition (step S13; Yes), the first probability of hitting the search when the designated color matches the first representative color The product is searched according to the second probability of hitting the search when the designated color matches the second representative color (step S14).
That is, the control unit 230 (search unit 234) causes the product to be hit based on the first probability when the designated color designated at the time of retrieval matches the first representative color, or the designated color is the second color. If it matches the representative color, the product is hit based on the second probability.
　一方、第１領域および第２領域が、中間色条件を満たしていないと判別した場合（ステップＳ１３；Ｎｏ）に、色分類装置２００は、指定色が、第１代表色と一致する際に検索をヒットさせる検索を行う（ステップＳ１５）。
　すなわち、制御部２３０（検索部２３４）は、第１領域に対応付けられた第１代表色が指定された際に限り、当該商品を検索結果に含める。 On the other hand, when it is determined that the first area and the second area do not satisfy the intermediate color condition (step S13; No), the color classification device 200 performs a search when the designated color matches the first representative color. A search for hit is performed (step S15).
That is, the control unit 230 (search unit 234) includes the product in the search result only when the first representative color associated with the first region is designated.
　色分類装置２００は、検索結果を、ユーザ端末３００に提示する（ステップＳ１６）。
　すなわち、提示部２４０は、上述したステップＳ１４又は、ステップＳ１５にて検索された検索結果（検索結果一覧）をユーザ端末３００に提示する。 The color classification device 200 presents the search result to the user terminal 300 (step S16).
That is, the presentation unit 240 presents the search results (search result list) searched in step S14 or step S15 described above to the user terminal 300.
　色分類装置２００は、中間色条件を満たす場合におけるユーザの反応を収集する（ステップＳ１７）。
　すなわち、制御部２３０は、上述したステップＳ１４にて検索された検索結果をユーザ端末３００に提示した場合におけるユーザの反応を収集する。つまり、第１代表色にて商品が検索された後のユーザによる商品に対する第１反応と、第２代表色にて商品が検索された後のユーザによる商品に対する第２反応とを収集する。
　例えば、上述した図５に示す商品Ｐが赤と橙との中間色のスーツケースである場合において、「レッド」のスーツケースについての検索がユーザに指示され、ステップＳ１４にて、このスーツケースを検索結果に含めて提示した後に、ユーザからこのスーツケースが選択された数（第１反応）と、「オレンジ」のスーツケースについての検索がユーザに指示され、ステップＳ１４にて、このスーツケースを検索結果に含めて提示した後に、ユーザからこのスーツケースが選択された数（第２反応）と、を集計する。 The color classification device 200 collects user responses when the intermediate color condition is satisfied (step S17).
That is, the control unit 230 collects user reactions when the search results searched in step S14 described above are presented to the user terminal 300. That is, the first response to the product by the user after the product is searched with the first representative color and the second response to the product after the product is searched with the second representative color are collected.
For example, when the product P shown in FIG. 5 is a suitcase of an intermediate color between red and orange, the user is instructed to search for a “red” suitcase, and the suitcase is searched in step S14. After the presentation is included in the results, the user is instructed to search for the number of suitcases selected by the user (first reaction) and the suitcase of “orange”. In step S14, the suitcases are searched. After the presentation is included in the result, the number of the suitcases selected by the user (second reaction) is totaled.
　色分類装置２００は、第２反応の方が第１反応よりも反応度合いが大きいか否かを判別する（ステップＳ１８）。
　色分類装置２００は、第２反応の方が第１反応よりも反応度合いが大きいと判別すると（ステップＳ１８；Ｙｅｓ）、第１領域を狭くし、第２領域を広げるように、領域情報を修正する（ステップＳ１９）。
　例えば、第２代表色である「オレンジ」を検索結果に含めたときの方が、第１代表色である「レッド」を検索結果に含めたときよりも、このスーツケースがユーザからより多く選択された場合に、制御部２３０（修正部２３５）は、上述した図８に示すように、「レッド」の領域ＲＡが狭くなり、その分「オレンジ」の領域ＯＡが広がるように、領域情報における領域の範囲を修正する。これにより、今後、赤と橙との中間色の商品は、「レッド」よりも「オレンジ」により多く分類され、指定色に「オレンジ」が指定された際により検索され易くなる。 The color classification device 200 determines whether or not the second reaction has a greater degree of reaction than the first reaction (step S18).
If the color classification device 200 determines that the second reaction has a greater degree of reaction than the first reaction (step S18; Yes), the region information is corrected so that the first region is narrowed and the second region is widened. (Step S19).
For example, when the search result includes the second representative color “orange”, the suitcase is selected more by the user than when the first representative color “red” is included in the search result. In this case, as shown in FIG. 8 described above, the control unit 230 (correction unit 235) makes the “red” region RA narrow and the “orange” region OA widens accordingly. Correct the range of the area. As a result, products in the middle color between red and orange will be more classified into “orange” than “red” in the future, and it will be easier to search when “orange” is designated as the designated color.
　一方、第２反応の方が第１反応よりも反応度合いが大きくないと判別した場合（ステップＳ１８；Ｎｏ）、第１領域と第２領域との境界を元に戻すように、領域情報を修正する（ステップＳ２０）。
　例えば、第１代表色である「レッド」を検索結果に含めたときの方が、第２代表色である「オレンジ」を検索結果に含めたときよりも、このスーツケースがユーザからより多く選択された場合に、修正部２３５は、「レッド」の領域と「オレンジ」の領域との境界を初期の境界に戻すように、領域情報における領域の範囲を修正する。つまり、ステップＳ１９にて、「レッド」の領域ＲＡが狭くなり、その分「オレンジ」の領域ＯＡが広がっていれば、互いの領域を元に戻す。なお、「レッド」の領域ＲＡと「オレンジ」の領域ＯＡとが初期の領域のままであれば、修正部２３５は、何ら修正を行わない。 On the other hand, when it is determined that the degree of reaction is not greater in the second reaction than in the first reaction (step S18; No), the region information is corrected so that the boundary between the first region and the second region is restored. (Step S20).
For example, when the first representative color “red” is included in the search results, more suitcases are selected from the user than when the second representative color “orange” is included in the search results. In this case, the correction unit 235 corrects the range of the region in the region information so that the boundary between the “red” region and the “orange” region is returned to the initial boundary. That is, in step S19, if the “red” area RA is narrowed and the “orange” area OA is expanded by that amount, the areas are returned to the original. If the “red” area RA and the “orange” area OA remain the initial areas, the correction unit 235 does not perform any correction.
　このような色分類処理によって、商品の色が中間色であっても、代表色の何れかに適切に分類することができる。また、ユーザの反応に応じて、隣り合う領域同士の境界を変動させるため、中間色が多くのユーザから支持を受け易い代表色に分類することができる。
　この結果、商品の色を適切に分類することができる。 By such color classification processing, even if the color of the product is an intermediate color, it can be appropriately classified into any of the representative colors. Further, since the boundary between adjacent regions is changed according to the reaction of the user, the intermediate color can be classified into representative colors that are easily supported by many users.
As a result, the product colors can be appropriately classified.
（他の実施形態）
　上記の実施形態では、検索の度に、商品の色を分類する場合について説明したが、一旦分類した結果を商品情報に登録するようにし、所定の見直しタイミングで、再度分類するようにしてもよい。
　例えば、上述した図３に示す色分類装置２００において、制御部２３０に登録部を更に備えるようにし、この登録部が、第１代表色および第２代表色を、当該商品の商品情報に登録するようにしてもよい。
　この場合、検索部２３４は、検索時に指定された指定色が、商品情報における第１代表色と一致する際に検索をヒットさせる第１確率と、同指定色が、同商品情報における第２代表色と一致する際に検索をヒットさせる第２確率とに応じて、商品の検索を行う。 (Other embodiments)
In the above embodiment, the case has been described in which the color of the product is classified for each search. However, the classification result may be registered in the product information, and may be classified again at a predetermined review timing. .
For example, in the color classification apparatus 200 shown in FIG. 3 described above, the control unit 230 further includes a registration unit, and the registration unit registers the first representative color and the second representative color in the product information of the product. You may do it.
In this case, the search unit 234 has a first probability that the search is hit when the specified color specified at the time of the search matches the first representative color in the product information, and the specified color is the second representative in the product information. The product is searched according to the second probability of hitting the search when it matches the color.
　上記の実施形態では、領域情報に、代表色の領域の範囲だけが規定されている場合について説明したが、更に、中間色の領域の範囲が領域情報に規定されていてもよい。
　この場合、分類部２３１は、更に、商品画像の画素（マス）の色が中間力の領域に属するかも含めて分類する。そして、判別部２３３は、分類部２３１にて２種類の範囲に従って分類した結果に基づいて、中間色条件を満たしているか否かを判別する。
　これにより、商品の色が、中間色条件を満たしているかどうかを、より詳細に判別することができる。 In the above-described embodiment, the case where only the range of the representative color region is defined in the region information has been described, but the range of the intermediate color region may be defined in the region information.
In this case, the classification unit 231 further classifies the product image pixel including the color of the pixel (mass) belonging to the intermediate force region. Then, the determination unit 233 determines whether or not the intermediate color condition is satisfied based on the result of the classification by the classification unit 231 according to the two types of ranges.
Thereby, it can be determined in more detail whether the color of the product satisfies the intermediate color condition.
　以上説明したように、本発明によれば、商品の色を適切に分類することのできる色分類装置、色分類方法、プログラム、ならびに、非一時的なコンピュータ読取可能な情報記録媒体を提供することができる。 As described above, according to the present invention, it is possible to provide a color classification device, a color classification method, a program, and a non-transitory computer-readable information recording medium that can appropriately classify the color of a product. Can do.
　１００　販売システム
　２００　色分類装置
　２１０　受付部
　２２０　記憶部
　２３０　制御部
　２３１　分類部
　２３２　取得部
　２３３　判別部
　２３４　検索部
　２３５　修正部
　２４０　提示部
　３００　ユーザ端末
　４００　情報処理装置
　４０１　ＣＰＵ
　４０２　ＲＯＭ
　４０３　ＲＡＭ
　４０４　ＮＩＣ
　４０５　画像処理部
　４０６　音声処理部
　４０７　ＤＶＤ－ＲＯＭドライブ
　４０８　インターフェース
　４０９　外部メモリ
　４１０　コントローラ
　４１１　モニタ
　４１２　スピーカ
　９００　インターネット DESCRIPTION OF SYMBOLS 100 Sales system 200 Color classification apparatus 210 Reception part 220 Storage part 230 Control part 231 Classification part 232 Acquisition part 233 Determination part 234 Search part 235 Correction part 240 Presentation part 300 User terminal 400 Information processing apparatus 401 CPU
402 ROM
403 RAM
404 NIC
405 Image processing unit 406 Audio processing unit 407 DVD-ROM drive 408 Interface 409 External memory 410 Controller 411 Monitor 412 Speaker 900 Internet

Claims (8)

  1. 　色空間を分割する複数の領域についての領域情報であり、当該領域の範囲と当該領域を示す代表色とを対応付ける領域情報を記憶する記憶部、
    　商品画像に含まれる商品についての画素を、当該画素が有する色が前記複数の領域のいずれに属するかにより、分類する分類部、
    　前記複数の領域のそれぞれに分類された画素の数が多い順に第１領域および第２領域を取得する取得部、
    　前記第１領域と前記第２領域とが前記色空間において接しており、かつ、前記第１領域に分類された画素の数と前記第２領域に分類された画素の数との比率が所定比率を超える、中間色条件を満たすか否かを判別する判別部、
    　前記中間色条件を満たせば、前記商品を、前記第１領域に対応付けられた第１代表色および前記第２領域に対応付けられた第２代表色のいずれにても検索可能とする検索部、
    　前記第１代表色にて前記商品が検索された後の検索者の前記商品に対する第１反応と、前記第２代表色にて前記商品が検索された後の検索者の前記商品に対する第２反応と、に基づいて、前記第１領域と前記第２領域との境界を変化させるように、前記領域情報における前記領域の範囲を修正する修正部、
    　を備えることを特徴とする色分類装置。     A storage unit that stores region information about a plurality of regions that divide the color space and associates the range of the region with a representative color indicating the region;
    A classifying unit that classifies pixels for products included in the product image according to which of the plurality of regions the color of the pixels belongs;
    An acquisition unit that acquires the first region and the second region in descending order of the number of pixels classified into each of the plurality of regions;
    The first area and the second area are in contact with each other in the color space, and a ratio between the number of pixels classified into the first area and the number of pixels classified into the second area is a predetermined ratio. A determination unit that determines whether or not the intermediate color condition is satisfied,
    A search unit capable of searching for the product in any of the first representative color associated with the first region and the second representative color associated with the second region, if the intermediate color condition is satisfied;
    A first response to the product of the searcher after the product is searched with the first representative color and a second response to the product of the searcher after the product is searched with the second representative color And a correction unit that corrects a range of the region in the region information so as to change a boundary between the first region and the second region,
    A color classification apparatus comprising:
  2. 　請求項１に記載の色分類装置であって、
    　前記修正部は、前記第２反応が前記第１反応よりも度合いが大きい場合に、前記第１領域が狭くなり、前記第２領域が広がるように、前記領域情報における前記領域の範囲を修正する、
    　ことを特徴とする色分類装置。     The color classification device according to claim 1,
    The correction unit corrects the range of the region in the region information so that the first region is narrowed and the second region is widened when the second reaction is greater in degree than the first reaction. ,
    A color classification device characterized by that.
  3. 　請求項１に記載の色分類装置であって、
    　前記修正部は、前記第１反応が前記第２反応よりも度合いが大きい場合に、前記第１領域と前記第２領域との境界を初期の境界に戻すように、前記領域情報における前記領域の範囲を修正する、
    　ことを特徴とする色分類装置。     The color classification device according to claim 1,
    The correction unit is configured to return the boundary between the first region and the second region to the initial boundary when the first reaction is larger than the second reaction. Modify the range,
    A color classification device characterized by that.
  4. 　請求項１に記載の色分類装置であって、
    　前記中間色条件を満たす場合に、前記第１代表色および前記第２代表色の両方を前記商品の商品情報に登録する登録部、を更に備え、
    　前記検索部は、検索時に指定された指定色が、前記商品情報における前記第１代表色と一致する際に検索をヒットさせる第１確率と、当該指定色が、前記商品情報における前記第２代表色と一致する際に検索をヒットさせる第２確率とに応じて、商品の検索を行う、
    　ことを特徴とする色分類装置。     The color classification device according to claim 1,
    A registration unit that registers both the first representative color and the second representative color in the product information of the product when the intermediate color condition is satisfied;
    The search unit includes a first probability that the search is hit when the specified color specified during the search matches the first representative color in the product information, and the specified color is the second representative in the product information. Search for products according to the second probability of hitting the search when it matches the color,
    A color classification device characterized by that.
  5. 　請求項４に記載の色分類装置であって、
    　前記検索部は、前記第１領域に分類された画素の数と前記第２領域に分類された画素の数とに基づいて、前記第１確率および第２確率を算定する、
    　ことを特徴とする色分類装置。     The color classification device according to claim 4,
    The search unit calculates the first probability and the second probability based on the number of pixels classified into the first region and the number of pixels classified into the second region.
    A color classification device characterized by that.
  6. 　色空間を分割する複数の領域についての領域情報であり、当該領域の範囲と当該領域を示す代表色とを対応付ける領域情報を記憶する記憶部を有する色分類装置における色分類方法であって、
    　前記色分類装置が、商品画像に含まれる商品についての画素を、当該画素が有する色が前記複数の領域のいずれに属するかにより、分類する分類ステップ、
    　前記色分類装置が、前記複数の領域のそれぞれに分類された画素の数が多い順に第１領域および第２領域を取得する取得ステップ、
    　前記色分類装置が、前記第１領域と前記第２領域とが前記色空間において接しており、かつ、前記第１領域に分類された画素の数と前記第２領域に分類された画素の数との比率が所定比率を超える、中間色条件を満たすか否かを判別する判別ステップ、
    　前記色分類装置が、前記中間色条件を満たせば、前記商品を、前記第１領域に対応付けられた第１代表色および前記第２領域に対応付けられた第２代表色のいずれにても検索可能とする検索ステップ、
    　前記色分類装置が、前記第１代表色にて前記商品が検索された後の検索者の前記商品に対する第１反応と、前記第２代表色にて前記商品が検索された後の検索者の前記商品に対する第２反応と、に基づいて、前記第１領域と前記第２領域との境界を変化させるように、前記領域情報における前記領域の範囲を修正する修正ステップ、
    　を備えることを特徴とする色分類方法。     A color classification method in a color classification apparatus having a storage unit that stores area information for associating a range of the area with a representative color indicating the area, which is area information about a plurality of areas that divide the color space,
    A classification step in which the color classification device classifies pixels of products included in a product image according to which of the plurality of regions a color of the pixels belongs;
    An acquisition step in which the color classification device acquires the first region and the second region in descending order of the number of pixels classified into each of the plurality of regions;
    In the color classification device, the first area and the second area are in contact with each other in the color space, and the number of pixels classified into the first area and the number of pixels classified into the second area A determination step of determining whether or not the intermediate color condition is satisfied, wherein the ratio exceeds a predetermined ratio;
    If the color classification device satisfies the intermediate color condition, the product is searched for in either the first representative color associated with the first region or the second representative color associated with the second region. Search step to enable,
    The color classification device includes a first response to the product of the searcher after the product is searched with the first representative color and a searcher after the product is searched with the second representative color. A correction step of correcting a range of the region in the region information so as to change a boundary between the first region and the second region based on a second response to the product;
    A color classification method comprising:
  7. 　色空間を分割する複数の領域についての領域情報であり、当該領域の範囲と当該領域を示す代表色とを対応付ける領域情報を記憶する記憶部を有するコンピュータを、
    　商品画像に含まれる商品についての画素を、当該画素が有する色が前記複数の領域のいずれに属するかにより、分類する分類部、
    　前記複数の領域のそれぞれに分類された画素の数が多い順に第１領域および第２領域を取得する取得部、
    　前記第１領域と前記第２領域とが前記色空間において接しており、かつ、前記第１領域に分類された画素の数と前記第２領域に分類された画素の数との比率が所定比率を超える、中間色条件を満たすか否かを判別する判別部、
    　前記中間色条件を満たせば、前記商品を、前記第１領域に対応付けられた第１代表色および前記第２領域に対応付けられた第２代表色のいずれにても検索可能とする検索部、
    　前記第１代表色にて前記商品が検索された後の検索者の前記商品に対する第１反応と、前記第２代表色にて前記商品が検索された後の検索者の前記商品に対する第２反応と、に基づいて、前記第１領域と前記第２領域との境界を変化させるように、前記領域情報における前記領域の範囲を修正する修正部、
    　として機能させることを特徴とするプログラム。     A computer having a storage unit that stores area information about a plurality of areas that divide the color space and that associates the range of the area with a representative color indicating the area,
    A classifying unit that classifies pixels for products included in the product image according to which of the plurality of regions the color of the pixels belongs;
    An acquisition unit that acquires the first region and the second region in descending order of the number of pixels classified into each of the plurality of regions;
    The first area and the second area are in contact with each other in the color space, and a ratio between the number of pixels classified into the first area and the number of pixels classified into the second area is a predetermined ratio. A determination unit that determines whether or not the intermediate color condition is satisfied,
    A search unit capable of searching for the product in any of the first representative color associated with the first region and the second representative color associated with the second region, if the intermediate color condition is satisfied;
    A first response to the product of the searcher after the product is searched with the first representative color and a second response to the product of the searcher after the product is searched with the second representative color And a correction unit that corrects a range of the region in the region information so as to change a boundary between the first region and the second region,
    A program characterized by functioning as
  8. 　色空間を分割する複数の領域についての領域情報であり、当該領域の範囲と当該領域を示す代表色とを対応付ける領域情報を記憶する記憶部を有するコンピュータを、
    　商品画像に含まれる商品についての画素を、当該画素が有する色が前記複数の領域のいずれに属するかにより、分類する分類部、
    　前記複数の領域のそれぞれに分類された画素の数が多い順に第１領域および第２領域を取得する取得部、
    　前記第１領域と前記第２領域とが前記色空間において接しており、かつ、前記第１領域に分類された画素の数と前記第２領域に分類された画素の数との比率が所定比率を超える、中間色条件を満たすか否かを判別する判別部、
    　前記中間色条件を満たせば、前記商品を、前記第１領域に対応付けられた第１代表色および前記第２領域に対応付けられた第２代表色のいずれにても検索可能とする検索部、
    　前記第１代表色にて前記商品が検索された後の検索者の前記商品に対する第１反応と、前記第２代表色にて前記商品が検索された後の検索者の前記商品に対する第２反応と、に基づいて、前記第１領域と前記第２領域との境界を変化させるように、前記領域情報における前記領域の範囲を修正する修正部、
    　として機能させることを特徴とするプログラムを記録したコンピュータ読取可能な記録媒体。     A computer having a storage unit that stores area information about a plurality of areas that divide the color space and that associates the range of the area with a representative color indicating the area,
    A classifying unit that classifies pixels for products included in the product image according to which of the plurality of regions the color of the pixels belongs;
    An acquisition unit that acquires the first region and the second region in descending order of the number of pixels classified into each of the plurality of regions;
    The first area and the second area are in contact with each other in the color space, and a ratio between the number of pixels classified into the first area and the number of pixels classified into the second area is a predetermined ratio. A determination unit that determines whether or not the intermediate color condition is satisfied,
    A search unit capable of searching for the product in any of the first representative color associated with the first region and the second representative color associated with the second region, if the intermediate color condition is satisfied;
    A first response to the product of the searcher after the product is searched with the first representative color and a second response to the product of the searcher after the product is searched with the second representative color And a correction unit that corrects a range of the region in the region information so as to change a boundary between the first region and the second region,
    A computer-readable recording medium having recorded thereon a program that is caused to function as:
PCT/JP2016/053783 2016-02-09 2016-02-09 Color classification device, color classification method, program, and non-transient computer-readable information recording medium WO2017138088A1 (en)

Priority Applications (2)

Application Number Priority Date Filing Date Title
JP2016536793A JP6028130B1 (en) 2016-02-09 2016-02-09 Color classification apparatus, color classification method, program, and non-transitory computer-readable information recording medium
PCT/JP2016/053783 WO2017138088A1 (en) 2016-02-09 2016-02-09 Color classification device, color classification method, program, and non-transient computer-readable information recording medium

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
PCT/JP2016/053783 WO2017138088A1 (en) 2016-02-09 2016-02-09 Color classification device, color classification method, program, and non-transient computer-readable information recording medium

Publications (1)

Publication Number Publication Date
WO2017138088A1 true WO2017138088A1 (en) 2017-08-17

Family

ID=57326663

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/JP2016/053783 WO2017138088A1 (en) 2016-02-09 2016-02-09 Color classification device, color classification method, program, and non-transient computer-readable information recording medium

Country Status (2)

Country Link
JP (1) JP6028130B1 (en)
WO (1) WO2017138088A1 (en)

Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JPH11288465A (en) * 1998-02-06 1999-10-19 Fujitsu Ltd Color image processor and pattern extracting device
WO2011045920A1 (en) * 2009-10-16 2011-04-21 日本電気株式会社 Color analysis device, color analysis method, and color analysis program
JP2011520203A (en) * 2008-05-09 2011-07-14 エルティーユー テクノロジーズ エスエーエス Color matching tool box
JP2014160396A (en) * 2013-02-20 2014-09-04 Dainippon Printing Co Ltd Commodity recommendation device, commodity recommendation method, program, and commodity recommendation system
WO2015145766A1 (en) * 2014-03-28 2015-10-01 楽天株式会社 Color estimation device, color estimation method, and color estimation program

Patent Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JPH11288465A (en) * 1998-02-06 1999-10-19 Fujitsu Ltd Color image processor and pattern extracting device
JP2011520203A (en) * 2008-05-09 2011-07-14 エルティーユー テクノロジーズ エスエーエス Color matching tool box
WO2011045920A1 (en) * 2009-10-16 2011-04-21 日本電気株式会社 Color analysis device, color analysis method, and color analysis program
JP2014160396A (en) * 2013-02-20 2014-09-04 Dainippon Printing Co Ltd Commodity recommendation device, commodity recommendation method, program, and commodity recommendation system
WO2015145766A1 (en) * 2014-03-28 2015-10-01 楽天株式会社 Color estimation device, color estimation method, and color estimation program

Also Published As

Publication number Publication date
JP6028130B1 (en) 2016-11-16
JPWO2017138088A1 (en) 2018-02-15

Similar Documents

Publication Publication Date Title
CN107958030B (en) Video cover recommendation model optimization method and device
US9846809B2 (en) Information processing apparatus, method, and storage medium storing program
JP5919450B1 (en) SEARCH DEVICE, SEARCH METHOD, RECORDING MEDIUM, AND PROGRAM
JP2019148897A (en) Behavior pattern search system and behavior pattern search method
US9721291B1 (en) Identifying effective images associated with items
JP6922400B2 (en) Fashion analysis program, fashion analyzer and fashion analysis method
JP2004192156A (en) Image retrieval program, storage medium storing the program, image retrieval system and image retrieval method
JP6028130B1 (en) Color classification apparatus, color classification method, program, and non-transitory computer-readable information recording medium
CN112116102A (en) Method and system for expanding domain adaptive training set
JP2019050576A5 (en)
JPWO2013161086A1 (en) Image processing apparatus, image processing apparatus control method, program, and information storage medium
US8744179B2 (en) Image processing apparatus, image processing method, and non-transitory computer readable medium that utilizes representative colors
US10740815B2 (en) Searching device, searching method, recording medium, and program
TWI553571B (en) Sales support server, catalog page delivery method, recording media
KR101660984B1 (en) System and method for providing information according to image data
JP6750094B2 (en) Image evaluation apparatus, image evaluation method, and image evaluation program
JP5969158B1 (en) Server apparatus, control method, program, and recording medium
JP5778881B1 (en) Selection device, item selection method, recording medium, and program
JP3916802B2 (en) Video search method, video search device, and storage medium
WO2019229909A1 (en) Enthusiasm degree display system, stadium enthusiasm degree display method, program, and sensor terminal
KR101714932B1 (en) Apparatus and method for extracting items using a bloom filter
JP4457724B2 (en) Composite image generation method, composite image generation apparatus, and composite image generation program
JP6591000B1 (en) Image processing apparatus, image processing method, and image processing program
JP7129383B2 (en) Image processing device, image processing method, image processing program, and recording medium storing the program
JP6246987B2 (en) Product search device, search method, recording medium, and program

Legal Events

Date Code Title Description
ENP Entry into the national phase

Ref document number: 2016536793

Country of ref document: JP

Kind code of ref document: A

121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 16889790

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 16889790

Country of ref document: EP

Kind code of ref document: A1