WO2015129331A1 - 商品検索装置、方法及びシステム - Google Patents

商品検索装置、方法及びシステム Download PDF

Info

Publication number
WO2015129331A1
WO2015129331A1 PCT/JP2015/051242 JP2015051242W WO2015129331A1 WO 2015129331 A1 WO2015129331 A1 WO 2015129331A1 JP 2015051242 W JP2015051242 W JP 2015051242W WO 2015129331 A1 WO2015129331 A1 WO 2015129331A1
Authority
WO
WIPO (PCT)
Prior art keywords
block
product
sensitivity
category
image
Prior art date
Legal status (The legal status 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 status listed.)
Ceased
Application number
PCT/JP2015/051242
Other languages
English (en)
French (fr)
Japanese (ja)
Inventor
野口　幸典
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Fujifilm Corp
Original Assignee
Fujifilm Corp
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 Fujifilm Corp filed Critical Fujifilm Corp
Priority to CN201580007680.2A priority Critical patent/CN105981012A/zh
Priority to EP15754715.9A priority patent/EP3113045A4/en
Publication of WO2015129331A1 publication Critical patent/WO2015129331A1/ja
Priority to US15/205,403 priority patent/US10229177B2/en
Anticipated expiration legal-status Critical
Ceased legal-status Critical Current

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
    • G06F16/20—Information retrieval; Database structures therefor; File system structures therefor of structured data, e.g. relational data
    • G06F16/25—Integrating or interfacing systems involving database management systems
    • G06F16/258—Data format conversion from or to a database
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
    • G06F16/20—Information retrieval; Database structures therefor; File system structures therefor of structured data, e.g. relational data
    • G06F16/28—Databases characterised by their database models, e.g. relational or object models
    • G06F16/284—Relational databases
    • G06F16/285—Clustering or classification
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
    • G06F16/50—Information retrieval; Database structures therefor; File system structures therefor of still image data
    • G06F16/51—Indexing; Data structures therefor; Storage structures
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
    • G06F16/50—Information retrieval; Database structures therefor; File system structures therefor of still image data
    • G06F16/58—Retrieval characterised by using metadata, e.g. metadata not derived from the content or metadata generated manually
    • G06F16/5866—Retrieval characterised by using metadata, e.g. metadata not derived from the content or metadata generated manually using information manually generated, e.g. tags, keywords, comments, manually generated location and time information
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
    • G06F16/90—Details of database functions independent of the retrieved data types
    • G06F16/95—Retrieval from the web
    • G06F16/953—Querying, e.g. by the use of web search engines
    • G06F16/9535—Search customisation based on user profiles and personalisation
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
    • G06F16/90—Details of database functions independent of the retrieved data types
    • G06F16/95—Retrieval from the web
    • G06F16/953—Querying, e.g. by the use of web search engines
    • G06F16/9538—Presentation of query results
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; 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
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; 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
    • G06Q30/0601—Electronic shopping [e-shopping]
    • G06Q30/0641—Electronic shopping [e-shopping] utilising user interfaces specially adapted for shopping
    • G06Q30/0643—Electronic shopping [e-shopping] utilising user interfaces specially adapted for shopping graphically representing goods, e.g. 3D product representation

Definitions

  • the present invention relates to a product search apparatus, method, and system that enable a consumer to smoothly make a decision to purchase a plurality of products.
  • Patent Document 1 a correlation between a sensory feature amount used to specify an impression felt by a designer from an arbitrary image and a physical feature amount extracted by calculation from an image processing result is obtained in advance by a statistical method.
  • image search at the time of design, a configuration is disclosed in which an image search is performed by obtaining a spatial distance (Euclidean distance) between the sensitivity feature amount of the input image and the sensitivity feature amount of the registered image.
  • the coordinate position of the specific similar designated image designated by the user from the example image menu and the specific The center point of the image search is obtained based on the line segment determined by the coordinate position of the dissimilar designated image, and n registered images having a spatial distance from the center point are searched from the database in order from the nearest registered image.
  • Patent Document 2 a plurality of registered images are stored in advance in a database together with physical feature amounts as a pattern sample, and in the image search at the time of design, the coordinate axes of physical features (for example, line thickness axis, left-right symmetry axis) , And a density axis of the space), when a sensitivity word (for example, “a little sharper”) that defines a shift rule for the reference coordinate position is input by the user, it corresponds to a physical feature amount of a specific image.
  • a shift corresponding to the input sensitivity word from the reference coordinate position (for example, a shift to a coordinate position having a smaller line thickness value) is performed, and an image with a short spatial distance is centered on the shifted reference coordinate position.
  • a configuration for searching from a database is disclosed. As a result, it is possible to accurately search and display an image that is almost the same as the human impression compared to the specific image specified by the user.
  • client device When displaying images of recommended products on a consumer's terminal device (hereinafter referred to as “client device”) via a network, images of multiple recommended products are randomly displayed without considering the consumer's impression of the product image.
  • client device When displayed, there are problems that the images are displayed randomly and the consumer cannot narrow down the product he / she wants to buy, and that the consumer gets tired before finding a favorite product. That is, there is a problem that it is difficult for consumers to make purchase decisions simply by displaying images of recommended products at random.
  • consumers are confused by the coordination, and the probability of not making a purchase decision for online shopping increases.
  • the consumer desires to purchase a plurality of different products in one category the consumer may be confused by the coordination.
  • Patent Documents 1 and 2 It is also conceivable to apply a known image search technique for designers as disclosed in Patent Documents 1 and 2 to product search for product recommendation for general consumers.
  • the conventional techniques described in Patent Documents 1 and 2 described above are image search techniques for designers, and may certainly be efficient for designers who want to create similar designs one after another from designs already created.
  • image search technology for product recommendation suitable for consumers who want to purchase a plurality of products at the same time is no image search technology for product recommendation suitable for consumers who want to purchase a plurality of products at the same time.
  • the present invention has been made in view of such circumstances, and an object of the present invention is to provide a product search apparatus, method, and system that enable a consumer to smoothly make a decision to purchase a plurality of products.
  • the present invention is a product search apparatus using a product database that stores a plurality of images respectively corresponding to a plurality of products, a physical quantity of product images, and a product category in association with each other.
  • the physical quantity acquisition unit for acquiring the physical quantity of the image of the specific product from the product database, and the physical quantity of the acquired image of the specific product for the plurality of blocks in the sensitivity space in which a plurality of sensitivity words representing human sensitivity are arranged
  • a first conversion unit that converts information indicating a specific product sensitivity block that is a block corresponding to an image of the specific product, and a sensitivity space based on the information indicating the specific product sensitivity block obtained by the first conversion unit.
  • a second conversion unit that converts information indicating a target block selected from the plurality of blocks into information indicating a range of a physical quantity of an image corresponding to the target block;
  • a category selection unit that selects a search target category from a plurality of categories stored in the product database based on a category of a specific product acquired from the database, a search target category selected by the category selection unit, and a second
  • a product search device including a search unit that searches an image corresponding to a search target category and a block of interest from a product database based on information indicating a physical quantity range obtained by the conversion unit.
  • a category search range is selected based on a category of a specific product, so that a consumer can purchase a plurality of different products in one category. Whether you wish or if you want to purchase multiple products across multiple categories, consumers can make product recommendations based on specific products and consumers can purchase multiple products. It becomes possible to make the decision smoothly.
  • the category selection unit selects the same category as the specific product as a search target category. Thereby, when the consumer desires to purchase a plurality of products in one category, it is possible to prompt the consumer to make a purchase decision.
  • the category selection unit selects a category of a product having the same body position to be worn as a search target category as compared with the specific product. Accordingly, when the consumer desires to purchase a plurality of products over a plurality of categories, it is possible to prompt the consumer to make a purchase decision.
  • the category selection unit selects a category of a product in which the body position to be worn is adjacent, a category of adjacent products, or a category of products partially overlapping as compared with a specific product. Select as. Accordingly, when the consumer desires to purchase a plurality of products over a plurality of categories having different wearing positions, it is possible to prompt the consumer to make a purchase decision.
  • a category-to-body position database that stores category-to-body position information indicating a correspondence relationship between each category stored in the product database and a person's body position.
  • a search target category is selected based on the body position information. Accordingly, when the consumer desires to purchase a plurality of products over a plurality of categories having different wearing positions, it is possible to prompt the consumer to make a purchase decision.
  • an attention block selection unit that selects at least one of a block adjacent to a specific product sensitivity block, a block adjacent to the specific product sensitivity block, and a partially overlapping block in the sensitivity space.
  • an attention block selection unit that selects an opposite word sensitivity block, which is a block including a sensitivity word having a meaning opposite to the sensitivity word corresponding to the image of the specific product, as the attention block is provided.
  • the block-of-interest selection unit also selects a block arranged between the specific product sensitivity block and the opposite word sensitivity block in the sensitivity space as the block of interest.
  • a user-specified information receiving unit that receives information indicating at least one of a user-specified sensitivity word and a user-specified block, and a block of a sensitivity space corresponding to the user-specified sensitivity word or
  • An attention block selection unit is provided that selects a user-specified sensitivity block, which is a block of the sensitivity space designated by the user, as the attention block. This makes it possible to recommend a product with an image that is not similar to the specific product in accordance with user designation.
  • the block-of-interest selection unit also selects a block arranged between the specific product sensitivity block and the user-specified sensitivity block in the sensitivity space as the block of interest.
  • the first sensitivity word block, the second sensitivity word block, and the specific product sensitivity block which are blocks corresponding to the first sensitivity word and the second sensitivity word having opposite meanings, respectively.
  • a block of interest selection unit that selects a block arranged on a curve connecting the three blocks as a block of interest.
  • a conversion database that stores conversion data that indicates correspondence between information indicating each block of the sensitivity space and a range of physical quantities of images of a plurality of products, and the second conversion unit includes the conversion data Conversion based on
  • the physical quantity is at least one of a color feature quantity, a shape feature quantity, a pattern feature quantity, and a texture feature quantity.
  • an output unit that outputs an image searched by the search unit is provided.
  • this invention is a goods search system provided with the server apparatus which comprises said goods search apparatus, and the client apparatus connected to a server apparatus via a network
  • a client apparatus shows a display part and specific goods.
  • a client input unit that receives input of information or information indicating an image of a specific product, a terminal transmission unit that transmits information received by the client input unit to the server device, and a search that is transmitted from the server device via the network
  • a terminal reception unit that receives the received image, and a control unit that causes the display unit to display the image received by the terminal reception unit.
  • the physical quantity acquisition unit of the server device includes information transmitted from the client device, Based on the information stored in the database, the physical quantity of the image of the specific product is acquired, and the server device clicks the searched image. To send to Ianto apparatus, to provide a product search system.
  • the present invention also relates to a product search method using a product database that stores a plurality of images corresponding to a plurality of products, physical quantities of product images, and product categories in association with each other.
  • the physical quantity of the image of the specific product and the physical quantity of the acquired image of the specific product correspond to the image of the specific product among the plurality of blocks in the sensitivity space in which a plurality of sensitivity words representing human sensitivity are arranged.
  • a first conversion step for converting into information indicating a specific product sensitivity block, which is a block, and a plurality of blocks in the sensitivity space are selected based on the information indicating the specific product sensitivity block obtained in the first conversion step.
  • a second conversion step for converting the information indicating the target block into information indicating the range of the physical quantity of the image corresponding to the target block.
  • a category selection step of selecting a search target category from a plurality of categories stored in the product database based on a category of the specific product, a search target category selected in the category selection step, and a second conversion step There is provided a product search method including a step of searching an image corresponding to a search target category and a block of interest from a product database based on information indicating a range of physical quantities obtained.
  • the consumer decides to purchase the plurality of products. It is possible to provide a product search apparatus, method, and system that enable smooth execution.
  • System configuration diagram of product search system The block diagram which shows the principal part structural example of a server apparatus (product search apparatus) The figure which shows the structural example of a goods database The figure which shows the structural example of a category versus body position database Conceptual diagram of Kansei space (Sensitive words grouped into blocks) Conceptual diagram of Kansei space (Aspect partitioned into Kansei words) A diagram showing an example of Kansei space Figure showing another example of Kansei space. The figure which shows the structural example of Kansei space database Diagram showing an example of the configuration of the conversion database Figure showing another example of the configuration of the conversion database 1st explanatory drawing used for description of the flow of a principal part process 2nd explanatory drawing used for description of the flow of a principal part process Explanatory drawing used to describe body position Block diagram showing a configuration example of a client device Explanatory drawing used for description of an example of attention block selection in the first embodiment Explanatory drawing used for explanation of another example of attention block selection in the first embodiment The figure which shows
  • Explanatory drawing used for explaining an example of attention block selection in the second embodiment The figure which shows the flow of a process of the whole goods search system in 3rd Embodiment.
  • Explanatory drawing used for description of an example of attention block selection in the third embodiment Explanatory drawing used for explaining an example of attention block selection in the fourth embodiment
  • First system configuration diagram used for explaining variations of the product search system Second system configuration diagram used for explaining variations of the product search system
  • FIG. 1 is a system configuration diagram of a product search system.
  • a product search system 1 is configured such that a server device 10 corresponding to an embodiment of a product search device of the present invention and a client device 11 that is a user terminal are connected via a network 12 such as the Internet. It is.
  • FIG. 2 is a block diagram illustrating a configuration example of a main part of the server device 10.
  • the server device 10 of this example is a server device that constitutes a product search device 100 that searches for product images from the product database 102, and includes a plurality of product images corresponding to a plurality of products, physical quantities of product images, and product categories.
  • a product database 102 that stores information in association with each other
  • a sensitivity space database 104 that stores various information in a sensitivity space in which a plurality of sensitivity words representing human sensitivity (hereinafter referred to as “sensitivity space information”) are stored in association with each other
  • a conversion database 106 that stores data for converting physical quantities of product images stored in the product database 102 and Kansei space information stored in the Kansei space database 104, and each category stored in the product database 102
  • Category-to-body position information indicating the correspondence between the body position
  • the memorized category versus body position database 108, the external input / output unit 112 having an input unit 112a for inputting information from the client device 11 and an output unit 112b for outputting information to
  • the first conversion unit 118 that converts the block corresponding to the image of the specific product (hereinafter referred to as “specific product sensitivity block”) and the sensitivity space database 104 into the information indicating the specific product sensitivity block. Based on the specific product sensitivity block among the multiple blocks of the sensitivity space based on The target block selecting unit 120 that selects the same or different block as the target block and the conversion database 106 are used to convert the information indicating the target block in the Kansei space into information indicating the range of the physical quantity of the product image corresponding to the target block.
  • the category selection unit 123 selects a search target category from a plurality of categories stored in the product database 102 based on the category of the specific product using the second conversion unit 122 that performs the above and the category versus body position database 108. And a product image corresponding to the search target category and the target block from the product database 102 based on the search target category selected by the category selection unit 123 and the information indicating the physical quantity range obtained by the second conversion unit 122. And a search unit 124. The product image searched by the search unit 124 is output to the client device 11 by the output unit 112b.
  • FIG. 3 is a diagram illustrating a configuration example of the product database 102.
  • the product database 102 stores a plurality of product images 132 corresponding to a plurality of products, a physical quantity 134 of the product image 132, and a product category 136 in association with each other.
  • the physical quantity 134 also referred to as “physical feature quantity”
  • the color feature quantity, shape feature quantity, pattern feature quantity, and texture feature quantity of the product image 132 are included.
  • the product database 102 may store various attribute information about the product such as the price of the product in association with the product image 132 in addition to the physical quantity 134 of the product image 132 and the product category 136.
  • the color feature amount is a feature amount related to the color of the product image. For example, the representative color or color scheme (color combination) of the product image is shown. In this example, the color feature amount measured by image analysis of the product image is stored in the product database 102.
  • the shape feature value is a feature value related to the shape of all or part of the product shown in the product image. For example, if the product is clothing, if the product is thin / thick, the length of the sleeve, the shape, size and / or angle of the collar, the size of the vacant area at the neck, the angle of the V-neck
  • the feature amount indicating the curvature of the U-neck and the like can be given as an example.
  • the shape feature amount includes a feature amount indicating the shape, size, and the like of a decorative article (for example, a ribbon).
  • the shape feature amount measured by image analysis of the product image is stored in the product database 102.
  • the pattern feature amount indicates the type (for example, floral pattern) and size of the pattern in the product image.
  • the pattern feature amount measured by image analysis of the product image is stored in the product database 102.
  • the texture feature amount indicates the degree of texture such as gloss of the product image.
  • the pattern feature amount measured by image analysis of the product image is stored in the product database 102.
  • the product category 136 is, in other words, the product type.
  • a product worn on a human body for example, a T-shirt, a sweater, a coat, jeans, a skirt, and the like.
  • FIG. 4 is a diagram illustrating a configuration example of the category versus body position database 108.
  • the category-to-body position database 108 in FIG. 4 is a database related to products worn on the human body, and includes category-to-body position information indicating the correspondence between each category stored in the product database 102 and the person's body position.
  • Body position is information indicating a position on a person's body, and indicates a position where the product can be worn. For example, the torso, legs, and head.
  • the body position may be defined according to the shape of the product. For example, if there are only four categories of T-shirts, sweaters, jeans, and skirts, only two body positions of the upper body and the lower body may be defined. If you add a hat to the category, you can add a “body position” such as a head, a hand if you add gloves to the category, or a foot if you add shoes to the category.
  • 5 and 6 are conceptual diagrams of the sensitivity space.
  • the sensitivity space in FIG. 5 and the sensitivity space in FIG. 6 have the same coordinate axes (first sensitivity feature amount axis and second sensitivity feature amount axis) and sensitivity words (WORD1 to WORD28).
  • the sensitivity space in FIG. 5 is a mode in which a plurality of sensitivity words are grouped to form blocks (B1 to B13)
  • the sensitivity space in FIG. 6 is a mode in which blocks are configured for each sensitivity word. It is different. It does not limit which aspect a block is made.
  • the number of coordinate axes (the number of dimensions of the sensitivity space), the types of sensitivity features that constitute the coordinate axes, the number and types of sensitivity words.
  • the number, shape and size of the blocks are not particularly limited.
  • the Kansei space in this specification has the following (Feature 1) to (Feature 3).
  • the Kansei space is a multi-dimensional coordinate space with a plurality of Kansei features as axes.
  • the sensitivity space illustrated in FIG. 5 and FIG. 6 is a two-dimensional space composed of the first sensitivity feature amount axis (X axis) and the second sensitivity feature amount axis (Y axis). Good.
  • the sensory feature amount indicates the degree of impression of a person who observes the image.
  • Feature 2 A plurality of sensitivity words are arranged in the sensitivity space in association with the sensitivity feature quantity constituting the axis of the sensitivity space.
  • sensitivity words WORD1 to WORD28 are arranged, but the number of sensitivity words is not particularly limited.
  • the sensitivity word is a word representing the impression of the person who observed the image.
  • the sensitivity space is divided into a plurality of blocks (hereinafter also referred to as “sensitivity blocks”), and one or more sensitivity words belong to each block.
  • the number and shape of the sensitivity blocks are not particularly limited.
  • Figures 7 to 8 show the image scales disclosed by Nippon Color Design Laboratory Co., Ltd. ("Color System” written by Shigejun Kobayashi, Kodansha), and http://www.ncd-ri.co.jp/ see about / image_system.html).
  • the “image scale” corresponds to a form of “Kansei space”.
  • the image scale shown in FIG. 7 is a horizontal axis indicating the degree of WARM (warm) / COOL (cold) corresponding to the axis of the first sensitivity feature amount, and a HARD corresponding to the axis of the second sensitivity feature amount. It consists of a vertical axis indicating the degree of (hard) / SOFT (soft). In addition, 66 sensitivity words are arranged and divided into 16 blocks.
  • the image scale shown in FIG. 8 is composed of two axes (WARM / COOL axis, HARD / SOFT axis) similar to those in FIG. 7, and a single color is associated as a color feature amount.
  • a single color is represented in grayscale, but in actuality it is represented in color.
  • WARM left side in the figure
  • COOL right side in the figure
  • SOFT light tone colors
  • HARD lower end side in the figure
  • colors that are close to each other are close to each other, and colors that are far from each other are close to each other.
  • FIG. 8 illustrates an emotional space in which monochromatic color feature amounts are associated
  • the emotional space used in the present invention is not particularly limited to such a case.
  • focusing on multiple color schemes as color feature quantities in the physical measurement space represented by physical quantities, and mapping multiple color schemes to the sensitivity space May be.
  • the color feature amount by using a plurality of color schemes as the color feature amount, it is possible to associate the color feature amount with a portion where it is impossible or inappropriate to arrange a single color in the sensitivity space.
  • the shape feature amount, the pattern feature amount, and the texture feature amount may be associated with the sensitivity space.
  • the emotional space database 104 stores emotional space information (for example, emotional feature quantities, emotional words, blocks, etc. constituting the axis) associated with each other and stored in the emotional space exemplified in FIGS.
  • the Kansei space database 104 illustrated in FIG. 9 stores the following information as Kansei space information.
  • ⁇ Block identification information of each block ⁇ Range information in the sensitivity space of each block (for example, coordinates of each vertex of the block or outline information) -Perspective information in the sensitivity space between blocks (for example, proximity / non-proximity between blocks, adjacent / non-adjacent, presence / absence of overlap) ⁇ Semantic correspondence information between Kansei words and Kansei words belonging to each block (for example, combinations of opposite words, combinations of similar words, etc.)
  • a block in which a plurality of sensitivity words are grouped hereinafter referred to as “large block” as shown in FIG.
  • small block a block for each sensitivity word (hereinafter referred to as “small block”) as shown in FIG.
  • large block Kansei space information block identification information, range information, distance information, Kansei word
  • small block Kansei space information block identification information, range information, distance information, Kansei
  • the conversion database 106 in FIG. 10 includes a conversion data table T1 indicating the correspondence between the physical quantity range and information indicating blocks.
  • information indicating the block for example, block identification information stored in the sensitivity space database 104 and range information in the sensitivity space of each block can be used.
  • the conversion database 106 in FIG. 11 includes a conversion data table T2 indicating the correspondence between the physical quantity range and the sensitivity word. That is, this is a data table using a sensitivity word as “information indicating a block” in the conversion data table T1 of FIG. For example, in the case of a sensitivity word (representative sensitivity word) representing each block in the aspect of the large block shown in FIG. 5, or in the case of the aspect of the small block shown in FIG. Can be used.
  • the main part process shown in FIG. 12 is a case where a product image having almost the same impression received by a person is searched.
  • the product search device 100 of the present invention performs the following steps S1 to S6.
  • step S1 the physical quantity acquisition unit 116 acquires the physical quantity of the image of the specific product from the product database 102.
  • the image of the specific product is a product image serving as a reference for product image search, and includes a mode in which an instruction is input by the user on the client device 11 and a mode in which the server device 10 determines. Either mode may be switched by the user.
  • the color feature amount a expressed in three dimensions of R, G, and B
  • the pattern feature amount b expressed in two dimensions of size and density
  • the two dimensions of transparency and gloss may be used as the physical feature value.
  • step S2 the first conversion unit 118 uses the conversion database 106 to specify the physical quantity of the image of the specific product acquired in step S1 among a plurality of blocks in the sensitivity space.
  • the information is converted into information indicating the specific product sensitivity block corresponding to the product image.
  • the information indicating the specific product sensitivity block is converted into block identification information of the specific product sensitivity block or a sensitivity word representing the specific product sensitivity block.
  • step S3 the attention block selection unit 120 selects a specific product sensitivity block as an attention block to be noticed for search.
  • step S4 the second conversion unit 122 uses the conversion database 106 to convert information indicating the target block of the sensitivity space into information indicating the range of the physical quantity of the product image corresponding to the target block. Convert. For example, the block identification information (or sensitivity word) of the block of interest is converted into physical quantity range information (for example, information indicating the upper limit and lower limit of the physical quantity).
  • the sensitivity word of the block of interest is converted into physical quantity range information.
  • step S5 the category selection unit 123 selects a search target category from a plurality of categories stored in the product database 102 based on the category of the specific product.
  • step S6 search step
  • the search unit 124 searches the product database 102 for the search target category and the target block based on the search target category selected in step S5 and the information indicating the physical quantity range obtained in step S4. A product image corresponding to is retrieved.
  • the main part processing shown in FIG. 13 is a case of searching for product images having different impressions received by people.
  • Steps S11 and S12 are the same as steps S1 and S2 in FIG.
  • step S13 the attention block selection unit 120 uses the sensitivity space database 104 and, based on the information indicating the specific product sensitivity block obtained in step S2, among a plurality of blocks in the sensitivity space. A block different from the specific product sensitivity block is selected. That is, in the sensitivity space, it is predicted that the specific product sensitivity block, which is a block corresponding to the specific product image, is a block corresponding to a product image that is not similar to the specific product image and is likely to be noticed by the user. It moves to the attention block which is the block to be done.
  • Steps S14 to S16 are the same as steps S4 to S6 in FIG.
  • the category selection unit 123 selects the same category as the specific product as a search target category. For example, when the category of the specific product is “sweater”, “sweater” is selected as the search target category.
  • the search target category can be selected without referring to the category-to-body position database 108.
  • the category selection unit 123 selects, as a search target category, a category of a product that has the same body position to be worn as compared with a specific product. For example, in the human body 70 of FIG. 14, if the body position of the category of the specific product is the body part 71, the category to be worn on the body part 71 is also selected as the search target category. Similarly, if the body position of the category of the specific product is the leg 72, the head 73, the hand 74, and the foot 75, the leg 72, the head 73, the hand 74, and the foot 75 are worn as the search target categories, respectively. Select the category that can be worn. In the second mode, the category selection unit 123 uses the category-to-body position database 108 of FIG. 4 to select a search target category based on the body position of the category of the specific product.
  • the category selection unit 123 compares the category of the product with which the body position to be worn is adjacent, the category of the adjacent product, or the category of the product that partially overlaps with each other as compared with the specific product. Select as.
  • the term “adjacent” means adjoining between products when the products are worn.
  • Proximity means proximity between products when the product is worn.
  • Proximity includes cases other than “adjacent” where there is no contact between products.
  • Partially overlapping means that there is an overlapping portion between the products when the products are worn.
  • the category selection unit 123 uses the category-to-body position database 108 of FIG. 4 to select a search target category based on the body position of the category of the specific product.
  • the category selection using the category-to-body position database 108 is not particularly limited to the above second and third modes. Correspondence information between a combination of a plurality of categories and a combination of a plurality of body positions is stored in the category-to-body position database in advance, and is referred to by the category selection unit 123, thereby combining various categories and body positions. Yes.
  • FIG. 15 is a block diagram illustrating a configuration example of the client device 11.
  • the client device 11 of this example includes a terminal reception unit 52a that receives and inputs information transmitted from the output unit 112b of the server device 10 and a terminal transmission unit 52b that transmits and outputs information received by the input unit 112a of the server device 10.
  • the client input unit 56 receives input of information indicating a specific product or information indicating an image of the specific product. Information received by the client input unit 56 is transmitted to the server device 10 by the terminal transmission unit 52b. The terminal reception unit 52a receives a product image as a search result transmitted from the server device 10 via the network 12.
  • the client device 11 is not particularly limited to the configuration example shown in FIG.
  • the client device 11 only needs to have a communication function for connecting to the network 12 for communication, a display function for displaying images, and a user input function for receiving user input.
  • various user terminals such as a mobile terminal, a personal computer, and a tablet terminal can be used as the client device 11.
  • the block-of-interest selecting unit 120 includes, in the sensitivity space, an adjacent sensitivity block that is a block adjacent to the specific product sensitivity block, a proximity sensitivity block that is a block adjacent to the specific product sensitivity block, and a specific product sensitivity block. At least one type of sensitivity block is selected as a target block of interest for product image search from among the overlapping sensitivity blocks that are partially overlapping blocks.
  • the adjacent emotion block is a sensitivity block in which the boundary is in contact with the specific product sensitivity block among the plurality of sensitivity blocks in the sensitivity space.
  • the proximity sensitivity block is a sensitivity block that is close to the specific product sensitivity block among the plurality of sensitivity blocks in the sensitivity space, and other than the adjacent sensitivity block, the sensitivity block whose boundary is not in contact with the specific product sensitivity block.
  • This example is a case where a large block in which sensitivity words are grouped and a small block for each sensitivity word are used in combination.
  • a sensitivity block (for example, a sensitivity block with reference numerals 202 to 206) including a coordinate position where the difference in sensitivity feature amount is equal to or smaller than a threshold value is selected as a target block.
  • the coordinate position of the sensitivity word corresponding to the physical quantity of the specific product image is used as the specific product sensitivity space position.
  • each sensitivity block is a proximity sensitivity block. For example, with respect to a circle 221 (where a circle includes an ellipse) centered on a specific product sensitivity space position 220 in the sensitivity space, the sensitivity block has a portion overlapping the circle 221 and the inner side of the circle 221. Sensitivity blocks 202 to 206 other than the specific product sensitivity block 201 are determined as proximity sensitivity blocks.
  • the range of the proximity sensitivity block in the sensitivity space may be changed by switching the diameter of the circle 221 by the block-of-interest selection unit 120 according to the number of product images.
  • the sensitivity block given the reference B7 is a specific product sensitivity block
  • the reference product B3, the reference B6, the reference B8, and the reference B11 whose boundaries are adjacent to the specific product sensitivity block B7 are attached.
  • the sensitivity block is the adjacent sensitivity block.
  • the specific product sensitivity space position is included.
  • the sensitivity block with reference numeral 201 is a specific product sensitivity block, and the sensitivity blocks with reference numerals 202 to 204 that are in contact with the specific product sensitivity block 201 are adjacent sensitivity blocks. That is, in FIG. 17, the sensitivity block included in the area within the dotted line denoted by reference numeral 230 is selected as the target block.
  • sensitivity word “feminine” for example, in the specific product sensitivity block 201, for example, the sensitivity word “homely” in the adjacent sensitivity block 202, and “ Sensitivity words such as “sophisticated”, “smart”, and “modern” are identified as sensitivity words corresponding to the product image to be searched.
  • the number of product images in the image search result is inappropriately reduced.
  • the number of product images as an image search result tends to be small. That is, when a product image as a search result is displayed on the client device 11, the number of products recommended to the user is inappropriately reduced. Therefore, depending on the number of product images, whether the proximity sensitivity block is the attention block as described with reference to FIG. 16 or only the adjacent sensitivity block is the attention block as described with reference to FIG. Switching may be performed by the block selection unit 120.
  • the emotional spaces illustrated in FIGS. 5, 16, and 17 are divided in such a manner that the emotional blocks do not overlap with each other.
  • a plurality of sensitivity blocks are partitioned in the space.
  • the target block selection unit 120 adds a sensitivity block (“overlap sensitivity” that partially overlaps the designated product sensitivity block in addition to the adjacent sensitivity block and / or the proximity sensitivity block. Block)) may be selected as the target block.
  • FIG. 18 is a flowchart showing the flow of product search processing in the product search system 1 shown in FIG. In this example, the search is performed using the product image specified by the user on the client device 11 as the specific product image.
  • the representative product list is generated as screen information by the screen information generation unit 114 of the server device 10, and the representative product list is transmitted and output to the client device 11 by the output unit 112b of the server device 10 (step S102).
  • the representative product list includes reduced images of a plurality of representative products.
  • product image identification information for each representative product is added to the representative product list.
  • the representative product list is received and input by the terminal receiving unit 52a of the client device 11, the representative product list is displayed and output on the display unit 54 of the client device 11 under the control of the control unit 58 of the client device 11 (step S104). .
  • the client input unit 56 of the client device 11 accepts specification of a product from the user, and the terminal transmission unit 52b of the client device 11 receives product image identification information (or) corresponding to the product specified by the user to the server device 10. Are transmitted and output (step S106).
  • the physical quantity acquisition unit 116 of the server device 10 acquires the image of the specific product from the product database 102 based on the product image identification information. A physical quantity is acquired (step S110).
  • the first conversion unit 118 of the server device 10 converts the acquired physical quantity of the image of the specific product into information indicating a specific product sensitivity block that is a block corresponding to the image of the specific product (step S112). .
  • the first conversion unit 118 of the present example acquires the sensitivity space information (for example, block identification information and / or sensitivity word) corresponding to the specific product sensitivity block from the conversion database 106 based on the physical quantity of the image.
  • the block of interest selection unit 120 of the server device 10 selects the block of interest using the emotional space database 104 (step S122).
  • the target block selection unit 120 can also include the specific product sensitivity block denoted by reference number 201 in the target block.
  • the second conversion unit 122 of the server device 10 converts the information indicating the block of interest into information indicating the range of the physical quantity of the product image corresponding to the block of interest (hereinafter, “physical quantity range information”) (step) S124).
  • the information indicating the target block for example, block identification information of the target block is used.
  • a sensitivity word corresponding to the block of interest is used.
  • the physical quantity range information includes color feature amount ranges (upper limit value and lower limit value), shape feature amount ranges (upper limit value and lower limit value), and pattern features in the search target product image.
  • One of a quantity range (upper limit and lower limit) and a texture feature quantity range (upper limit and lower limit) are included.
  • a combination of two or more types of feature amounts among color feature amount, shape feature amount, pattern feature amount, and texture feature amount may be used.
  • the category selection unit 123 of the server device 10 selects a search target category from among a plurality of categories registered in the product database 102 based on the category of the specific product (step S125).
  • the search unit 124 of the server device 10 selects the target block from the product database 102 based on the search target category selected by the category selection unit 123 and the physical quantity range information obtained by the second conversion unit 122. A corresponding product image is searched (step S126).
  • the output unit 112b of the server device 10 transmits and outputs the search result of the search unit 124 to the client device 11 (step S128).
  • the search result is displayed and output on the display unit 54 of the client device 11 under the control of the control unit 58 of the client device 11 (step S130).
  • the search result includes a product image corresponding to the block of interest.
  • the search result is not limited when all the product images are displayed at once on the client device 11.
  • a plurality of product images are searched, first, a plurality of reduced images of each of the plurality of product images are transmitted and output to the client device 11 to be displayed and output on the client device 11.
  • the selection input is performed, the entire product image corresponding to the reduced image selected and input by the user may be transmitted to the client device 11 and displayed.
  • the attention block selection unit 120 selects an antonym sensitivity block that is a block including an anti-sensitivity word that is a sensitivity word having a meaning opposite to that of the specific product sensitivity word corresponding to the image of the specific product. , Select as the attention block.
  • the specific product sensitivity word is a word indicating an impression received from an image by a person who observed the image of the specific product.
  • the sensitivity block with the symbol B ⁇ b> 9 is the specific product sensitivity block and WORD 19 is the specific product sensitivity word.
  • the opposite sensitivity words having the opposite meaning to the specific product sensitivity word WORD19 are two words WORD11 and WORD25, two blocks of the sensitivity block of code B6 to which WORD11 belongs and the sensitivity block of code 12 to which WORD25 belongs are opposite. It is a word sensitivity block.
  • the block-of-interest selecting unit 120 is based on the sensitivity space database 104.
  • the opposite sensitivity word having the opposite meaning to “casual” as the specific product sensitivity word is selected from the sensitivity words arranged in the sensitivity space.
  • the three words “classy” belonging to the sensibility block denoted by reference numeral 302, “sophisticated” belonging to the sensibility block denoted by reference numeral 303, and “classy” belonging to the sensitivity block denoted by reference numeral 304 are “classy”.
  • the sensitivity blocks of reference numerals 302, 303, and 304 are selected as the target block.
  • the block-of-interest selection unit 120 also selects blocks (for example, reference numerals 305 to 307) arranged between the specific product sensitivity block 301 and the opposite word sensitivity blocks 302 to 304 in the sensitivity space as the target blocks. May be.
  • the case where the sensitivity space shown in FIG. 7 is used has been described as an example, but the present invention is not particularly limited to such a case. You may use the sensitivity space shown in FIG. 7
  • the external input / output unit 112 (corresponding to one form of the user designation information receiving unit) of the third embodiment is a user that indicates at least one of the sensitivity word specified by the user and the sensitivity block specified by the user on the client device 11.
  • the designation information is received and input from the client device 11 via the network 12.
  • the attention block selection unit 120 selects a user-specified sensitivity block corresponding to a user-specified sensitivity word or a user-specified sensitivity block as a user-specified sensitivity block as an attention block to be noted for image search. To do.
  • FIG. 20 is a flowchart showing a flow of an example of product search processing in the third embodiment. 20
  • the server apparatus 10 has the configuration shown in FIG. 2
  • the client apparatus 11 has the configuration shown in FIG.
  • steps S102 to S112 shown in FIG. 18 are the same as steps S102 to S112 shown in FIG. 18 and have already been described in the first embodiment, and thus description thereof is omitted in this embodiment.
  • the output unit 112b of the server device 10 transmits and outputs the sensory space selection screen information to the client device 11 (step S314).
  • Sensitivity space selection screen information is screen information for allowing the client device 11 to select one or more blocks from among a plurality of blocks in the sensitivity space.
  • Step S316 When the sensitivity space selection screen information is received and input by the external input / output unit 52 of the client device 11, the sensitivity space selection screen is displayed and output on the display unit 54 of the client device 11 under the control of the control unit 58 of the client device 11. (Step S316).
  • the client input unit 56 of the client device 11 accepts the selection of the sensitivity word or block from the user, and the terminal transmission unit 52b of the client device 11 sends the user designation information indicating the sensitivity word or block selected by the user to the server device 10. On the other hand, a transmission output is performed (step S318).
  • a block is specified by allowing a user to specify a sensitivity word. For example, a list of sensitivity words is displayed on the display unit 54 of the client device 11, and it is detected from the client input unit 56 which sensitivity word in the sensitivity word list is selected and input by the user.
  • a block is designated by allowing the user to designate a block. For example, the sensitivity space shown in FIG. 6 is displayed on the display unit 54 of the client device 11, and it is detected from the client input unit 56 which position in the sensitivity space the user inputs.
  • the block corresponding to the sensitivity word designated by the user based on the user designation information by the target block selection unit 120 of the server device 10
  • the user-designated sensitivity block which is a block designated by the user, is selected as the block of interest (step S322).
  • the attention block selection unit 120 in this example also selects a block arranged between the specific product sensitivity block and the user-specified sensitivity block as the attention block in the sensitivity space.
  • the sensitivity block indicated by reference numeral 401 is a specific product sensitivity block and the sensitivity block indicated by reference numeral 402 is a user-specified sensitivity block
  • the specific product sensitivity block 401 and the user-specified sensitivity For example, the blocks respectively indicated by reference numerals 403 and 404 arranged between the blocks 402 are also selected as the target block.
  • the axis width (W in FIG. 21) connecting the specific product sensitivity block 401 and the user-specified sensitivity block 402 may be changed according to the number of product images (or the number of products).
  • the width W of this axis indicates the width of the width for selecting the block between the specific product sensitivity block 401 and the user-specified sensitivity block 402 as the target block in the sensitivity space.
  • the width W is narrowed as the number of product images increases, and the width W is increased as the number of product images decreases.
  • the “number of product images” may be the number of product images in the entire sensitivity space, or may be the number of product images corresponding to a block selected when the width W is a specified value.
  • Steps S324 to S330 in FIG. 20 are the same as steps S124 to S130 shown in FIG. 18 and have already been described in the first embodiment, so description thereof will be omitted in this embodiment.
  • the present invention is not particularly limited, and the sensitivity space shown in FIG. 8 may be used.
  • the attention block selection unit 120 of the fourth embodiment includes a first sensitivity word block and a second sensitivity word block that are blocks corresponding to the first sensitivity word and the second sensitivity word, respectively, having opposite meanings. And the block arranged on the curve connecting the three blocks of the specific product sensitivity block are selected as the target block.
  • FIG. 22 is an explanatory diagram of this embodiment using the sensitivity space of FIG.
  • the block denoted by reference numeral 501 and the block denoted by reference numeral 502 are groups having sensitivity words having opposite meanings
  • the block denoted by reference numeral 503 is a specific product sensitivity block.
  • the block-of-interest selection unit 120 of the present embodiment connects a block 501 (corresponding to the first sensitivity word block), a specific product sensitivity block 503, and a block 502 (corresponding to the second sensitivity word block). Blocks 501 to 513 arranged on the curve are selected as the target block.
  • the server device 10 includes an image analysis server 14, a database server 15, a mail server 16, and a WEB server 17.
  • the image analysis server 14 measures the physical quantity of each product image by analyzing the product image obtained by imaging the appearance of the product by image processing.
  • the physical quantity of the product image measured by the image analysis server 14 is transmitted to and stored in the database server 15 corresponding to one form of the product database 102.
  • Examples of the physical measurement amount of the product image measured by the image analysis server 14 in this example include a color feature amount, a shape feature amount, a pattern feature amount, and a texture feature amount.
  • the mail server 16 communicates by e-mail with the client device 11 via the network 12 to accept and confirm a product order.
  • the WEB server 17 performs interactive communication with the client device 11 via the network 12, receives and inputs a specific product designation from the client device 11, and transmits and outputs a search result to the client device 11.
  • each unit (reference numerals 104 to 124) other than the product database 102 can be mainly configured by the WEB server 17.
  • the servers 14 to 17 are allocated.
  • the servers 14 to 17 may be arranged in countries A to D, respectively.
  • the client device 11 may be connected to the network 12 in the country E where none of the servers 14 to 17 is arranged.
  • 10 Server device, 11: Client device, 12: Network, 100: Product search device, 102: Product database, 104: Kansei space database, 106: Conversion database, 108: Category vs body position database, 112: Outside of server device Input / output unit, 114: screen information generation unit, 116: physical quantity acquisition unit, 118: first conversion unit, 120: attention block selection unit, 122: second conversion unit, 123: category selection unit, 124: search unit

Landscapes

  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • Databases & Information Systems (AREA)
  • Business, Economics & Management (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Finance (AREA)
  • Accounting & Taxation (AREA)
  • Data Mining & Analysis (AREA)
  • General Engineering & Computer Science (AREA)
  • Strategic Management (AREA)
  • Development Economics (AREA)
  • Economics (AREA)
  • Marketing (AREA)
  • General Business, Economics & Management (AREA)
  • Library & Information Science (AREA)
  • Software Systems (AREA)
  • Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
  • Management, Administration, Business Operations System, And Electronic Commerce (AREA)
PCT/JP2015/051242 2014-02-28 2015-01-19 商品検索装置、方法及びシステム Ceased WO2015129331A1 (ja)

Priority Applications (3)

Application Number Priority Date Filing Date Title
CN201580007680.2A CN105981012A (zh) 2014-02-28 2015-01-19 商品搜索装置、方法及系统
EP15754715.9A EP3113045A4 (en) 2014-02-28 2015-01-19 Product retrieval device, method, and system
US15/205,403 US10229177B2 (en) 2014-02-28 2016-07-08 Product search apparatus, method, and system

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
JP2014-038522 2014-02-28
JP2014038522A JP6163440B2 (ja) 2014-02-28 2014-02-28 商品検索装置及び方法、商品検索システム

Related Child Applications (1)

Application Number Title Priority Date Filing Date
US15/205,403 Continuation US10229177B2 (en) 2014-02-28 2016-07-08 Product search apparatus, method, and system

Publications (1)

Publication Number Publication Date
WO2015129331A1 true WO2015129331A1 (ja) 2015-09-03

Family

ID=54008660

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/JP2015/051242 Ceased WO2015129331A1 (ja) 2014-02-28 2015-01-19 商品検索装置、方法及びシステム

Country Status (5)

Country Link
US (1) US10229177B2 (enExample)
EP (1) EP3113045A4 (enExample)
JP (1) JP6163440B2 (enExample)
CN (1) CN105981012A (enExample)
WO (1) WO2015129331A1 (enExample)

Families Citing this family (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP2020135298A (ja) * 2019-02-18 2020-08-31 Ｉｎｓｉｇｈｔ Ｌａｂ株式会社 情報処理装置
CN112085553A (zh) * 2019-06-12 2020-12-15 阿里巴巴集团控股有限公司 一种特定商品检测方法及装置

Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JPH11232288A (ja) * 1998-02-13 1999-08-27 Fuji Xerox Co Ltd 検索装置及び文書画像登録装置
JP2011070412A (ja) * 2009-09-25 2011-04-07 Seiko Epson Corp 画像検索装置および画像検索方法

Family Cites Families (16)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JPH08249353A (ja) 1995-03-15 1996-09-27 Omron Corp 画像検索方法及び装置
JPH09114853A (ja) 1995-10-20 1997-05-02 Omron Corp 画像検索方法および画像検索装置
US20120290601A1 (en) * 2007-11-15 2012-11-15 Master Wave International Co., Ltd. Image-based Data Management Method and System
WO2010016281A1 (ja) * 2008-08-08 2010-02-11 株式会社ニコン 検索支援システム、検索支援方法及び検索支援プログラム
US8229912B2 (en) * 2009-05-06 2012-07-24 Mentis Technology, Llc Enhanced search engine
US8370337B2 (en) * 2010-04-19 2013-02-05 Microsoft Corporation Ranking search results using click-based data
CN102375856B (zh) * 2010-08-23 2016-08-31 腾讯科技（深圳）有限公司 一种商品搜索方法和装置
US20120158686A1 (en) * 2010-12-17 2012-06-21 Microsoft Corporation Image Tag Refinement
US8560517B2 (en) * 2011-07-05 2013-10-15 Microsoft Corporation Object retrieval using visual query context
CN102591972B (zh) * 2011-12-31 2017-09-29 北京百度网讯科技有限公司 提供商品搜索结果的方法及设备
US8676814B2 (en) * 2012-02-16 2014-03-18 Yahoo! Inc. Automatic face annotation of images contained in media content
US20140019431A1 (en) * 2012-07-13 2014-01-16 Deepmind Technologies Limited Method and Apparatus for Conducting a Search
JP5998807B2 (ja) * 2012-09-28 2016-09-28 株式会社メルコホールディングス 情報処理システム、情報処理装置、情報処理方法及び情報処理プログラム
US9460518B2 (en) * 2013-04-17 2016-10-04 Yahoo! Inc. Visual clothing retrieval
US9712627B2 (en) * 2013-07-31 2017-07-18 Ebay Inc. User interface provisioning system
US20150142566A1 (en) * 2013-11-19 2015-05-21 Microsoft Corporation Surfacing Product Listing Ads for Fresh Items

Patent Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JPH11232288A (ja) * 1998-02-13 1999-08-27 Fuji Xerox Co Ltd 検索装置及び文書画像登録装置
JP2011070412A (ja) * 2009-09-25 2011-04-07 Seiko Epson Corp 画像検索装置および画像検索方法

Non-Patent Citations (3)

* Cited by examiner, † Cited by third party
Title
See also references of EP3113045A4 *
TAKASHI INABA: "Shikisai Shitsukan Design ni yoru Kansei Hyogen no Toraekata", IMAGE LAB, vol. 21, no. 1, 10 January 2010 (2010-01-10), pages 61 - 71, XP008182031 *
TERUHISA HOCHIN: "Image Madia and Kansei Retrieval", IEICE TECHNICAL REPORT, vol. 112, no. 234, 5 October 2012 (2012-10-05), pages 1 - 6, XP008181832 *

Also Published As

Publication number Publication date
EP3113045A1 (en) 2017-01-04
JP2015162193A (ja) 2015-09-07
EP3113045A4 (en) 2017-03-22
US20160321334A1 (en) 2016-11-03
US10229177B2 (en) 2019-03-12
CN105981012A (zh) 2016-09-28
JP6163440B2 (ja) 2017-07-12

Similar Documents

Publication Publication Date Title
JP7196885B2 (ja) 検索システム、検索方法、及びプログラム
KR101955649B1 (ko) 이미지 특징 데이터 추출 및 사용
JP6065078B2 (ja) 検索支援システム、検索支援方法及び検索支援プログラム
US10216818B2 (en) Product search apparatus, method, and system
JP6212013B2 (ja) 商品推奨装置及び商品推奨方法
US20160321732A1 (en) Product search apparatus, product search system, server system, and product search method
WO2016152277A1 (ja) リコメンド装置、リコメンドシステム、リコメンド方法、及びプログラム
JP6300677B2 (ja) コーディネート提案装置及び方法
JP6163440B2 (ja) 商品検索装置及び方法、商品検索システム
JP6195550B2 (ja) コーディネート提案装置及び方法
JP6182125B2 (ja) コーディネート提案装置及び方法

Legal Events

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

Ref document number: 15754715

Country of ref document: EP

Kind code of ref document: A1

REEP Request for entry into the european phase

Ref document number: 2015754715

Country of ref document: EP

WWE Wipo information: entry into national phase

Ref document number: 2015754715

Country of ref document: EP

NENP Non-entry into the national phase

Ref country code: DE