Color estimation device, color estimation method, and color estimation program

• US 10,692,133 B2
• Filed: 03/28/2014
• Issued: 06/23/2020
• Est. Priority Date: 03/28/2014
• Status: Active Grant

• Current Assignee: Rakuten Group, Inc.
• Original Assignee: Rakuten, Inc. (Rakuten Group, Inc.)
• Inventors: Masuko, Soh, Yoshimoto, Yoichi
• Primary Examiner(s): Thein, Marissa
• Assistant Examiner(s): Cai, Lance Y
• Application Number:
  US15/022,243
• Publication Number:
  US 20160232601A1
• Time in Patent Office: 2,279 Days
• CPC Class Codes: G06F 16/5838 using colour
• CPC Class Codes: G06F 18/21 Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
• CPC Class Codes: G06Q 30/0643 Graphical representation of items or shoppers
• CPC Class Codes: G06V 10/56 relating to colour

Abstract

A color estimation device includes an extraction means to extract, from product information associated with a product image showing a product and containing text information about the product to be presented to a user, a plurality of color text information indicating color names as candidate color names for the product, an estimation means to estimate a color of a product shown in the product image based on a pixel value set on the basis of each of the candidate color names and a pixel value of a pixel included in the product image, and an output means to output the color estimated by the estimation means.

Claims

1. A color estimation device comprising:

at least one memory configured to store computer program code;

at least one processor configured to read said computer program code and operate as instructed by said computer program code, said computer program code including;

extraction code configured to cause at least one of said at least one processor to extract, from text information contained in product information associated with a product image showing a product, a plurality of color text information indicating color names as candidate color names for the product;

estimation code configured to cause at least one of said at least one processor to estimate a color of the product shown in the product image among the candidate color names extracted based on a pixel value set to each of the candidate color names and a pixel value of a pixel included in the product image; and

output code configured to cause at least one of said at least one processor to output the color estimated, and

wherein the estimation code causes at least one of said at least one processor to calculate, for each of the candidate color names, the number of pixels of the product image included in a color range, the color range being a range of the pixel value set to each of the candidate color names, and estimate a candidate color name with the largest number of pixels as the color of the product shown in the product image, and

wherein the estimation code causes at least one of said at least one processor to set color ranges of the candidate color names not to overlap with each other based on the pixel value set to each of the candidate color names.

2. The color estimation device according to claim 1, wherein the estimation code causes at least one of said at least one processor to assign higher weights to a second pixel located at a position closer to a center of the product image than a first pixel included in the product image, and calculate the number of pixels for each of the candidate color names.

3. The color estimation device according to claim 1, wherein the estimation code causes at least one of said at least one processor to cluster pixels of the product image based on pixel values, and when a median value of a group is included in the color range of the candidate color name, calculate the number of pixels for each of the candidate color names, assuming the number of pixels included in the group as the number of pixels included in the color range of the candidate color name.

4. The color estimation device according to claim 1, wherein the estimation code causes at least one of said at least one processor to estimate the candidate color name where the calculated number of pixels is the largest and the candidate color name where a difference in the calculated number of pixels from that candidate color name is equal to or less than a specified number respectively as colors of a plurality of products shown in the product image.

5. The color estimation device according to claim 1, wherein, when a plurality of product images are associated with one product, the estimation code causes at least one of said at least one processor to estimate a color for each of the plurality of product images.

6. The color estimation device according to claim 1, wherein, when a plurality of product images are associated with one product, the estimation code causes at least one of said at least one processor to calculate, for each of the plurality of product images, the number of pixels of each of the plurality of product images included in the color range, the color range being the range of the pixel value set to one candidate color name among the plurality of candidate color names, and estimate a color of the product image with the largest number of pixels to be the one candidate color name.

7. The color estimation device according to claim 1, wherein

when a plurality of product images are associated with one product, the estimation code causes at least one of said at least one processor to calculate, for each of the candidate color names, the number of pixels of one product image included in the color range, the color range being the range of the pixel value set to the candidate color name, and estimate the candidate color name with the largest number of pixels as the color of the product shown in the one product image, and

when there is a candidate color name not estimated as the color of the product in any product image among the plurality of candidate color names, the estimation code causes at least one of said at least one processor to calculate, for each of the plurality of product images, the number of pixels of the product image included in the range of the pixel value set to one candidate color name among the plurality of candidate color names, and estimate the color of the product image with the largest number of pixels to be the one candidate color name.

8. The color estimation device according to claim 1, wherein, when the product information contains fields for receiving designation of the color of the product from the user, the extraction code causes at least one of said at least one processor to extract the color text information from the fields as the candidate color names.

9. The color estimation device according to claim 1, wherein the output code causes at least one of said at least one processor to store the estimated color in association with the product image into a product information storage that stores the product information and the product image in association with each other.

10. The color estimation device according to claim 9, further comprising:

search code configured to cause at least one of said at least one processor to refer to the product information storage in response to a search request containing designation of the color of the product from the user, and send a search result containing the product image with which the designated color is associated, back to the user.

11. The color estimation device according to claim 10, wherein, in a case where a plurality of product images are associated with one product, and one product image among the plurality of product images is set as a representative image for showing the product in a product page to present the product in the product information storage, when the search request containing designation of a color different from a color associated with the representative image is received from the user, the search code causes at least one of said at least one processor to send the search result containing the product image in which the designated color is associated with the representative image, back to the user.

12. A color estimation method performed by a computer, comprising:

extracting, from text information contained in product information associated with a product image showing a product, a plurality of color text information indicating color names as candidate color names for the product;

estimating a color of the product shown in the product image among the candidate color names extracted based on a pixel value set to each of the candidate color names and a pixel value of a pixel included in the product image; and

outputting the color estimated, and

wherein the estimating further comprises;

calculating, for each of the candidate color names, the number of pixels of the product image included in a color range, the color range being a range of the pixel value set to each of the candidate color names, and estimating a candidate color name with the largest number of pixels as the color of the product shown in the product image, and

setting color ranges of the candidate color names not to overlap with each other based on the pixel value set to each of the candidate color names.

13. A non-transitory computer-readable recording medium storing a color estimation program causing a computer to:

extract, from text information contained in product information associated with a product image showing a product, a plurality of color text information indicating color names as candidate color names for the product;

estimate a color of the product shown in the product image among the candidate color names extracted based on a pixel value set to each of the candidate color names and a pixel value of a pixel included in the product image; and

output the color estimated, and

wherein the estimating further comprises;

calculate, for each of the candidate color names, the number of pixels of the product image included in a color range, the color range being a range of the pixel value set to each of the candidate color names, and estimate a candidate color name with the largest number of pixels as the color of the product shown in the product image, and

set color ranges of the candidate color names not to overlap with each other based on the pixel value set to each of the candidate color names.