Sample color selection for online retail items

• US 10,672,049 B1
• Filed: 09/23/2014
• Issued: 06/02/2020
• Est. Priority Date: 09/23/2014
• Status: Active Grant

• Current Assignee: Amazon Technologies, Inc. (Amazon.com, Inc.)
• Original Assignee: Amazon Technologies, Inc. (Amazon.com, Inc.)
• Inventors: Biggs, Jody David, Ravikumar, Srijith, Smith, Brent Russell
• Primary Examiner(s): Miller, Brittney N
• Assistant Examiner(s): Mitros, Anna Mae
• Application Number:
  US14/494,576
• Time in Patent Office: 2,079 Days
• Field of Search: None
• CPC Class Codes: G06Q 30/0621 Item configuration or customization

Abstract

Systems, methods, and computer-readable media are disclosed for identifying a set of images of child variants of a parent item, identifying a set of colors in the set of images, categorizing the set of colors into color subgroups, identifying a set of representative color(s) for the color subgroups, generating a color value distribution representation for an image of a child variant that indicates a respective number of pixels in the child variant image corresponding to each of one or more representative colors, identifying a color cluster in the color value distribution representation, scoring the color cluster, and selecting a particular color from the color cluster for inclusion in a sample image of the child variant if the color cluster score meets or exceeds a threshold value.

Claims

1. A method, comprising:

identifying a set of child variant images, each child variant image comprising an image of a respective child variant of a parent retail item;

determining a set of colors in the set of child variant images;

assigning, by a computer processor, each color in the set of colors to a respective color subgroup;

reducing, by the computer processor, the set of colors present in the set of child variant images to a smaller set of representative colors, wherein each respective color subgroup corresponds to a respective representative color in the set of representative colors;

generating, by the computer processor, a first color histogram for a particular child variant image indicative of one or more representative colors present in the particular child variant image;

generating, by the computer processor, an average color histogram for the set of child variant images;

modifying, by the computer processor, the first color histogram using the average color histogram to generate a second color histogram for the particular child variant image;

identifying, by the computer processor, a color cluster present in the second color histogram;

determining, by the computer processor, a score for the color cluster indicative of a number of pixels in the particular child variant image having color values corresponding to the color cluster;

determining, by the computer processor, that the score meets or exceeds a threshold value;

selecting, by the computer processor, a particular color from the color cluster as a dominant color for inclusion in a sample image corresponding to the particular child variant;

storing, by the computer processor, a color value corresponding to the particular color;

modifying webpage source content to cause the sample image to be generated using the color value; and

generating the sample image responsive to a request for webpage content, wherein the sample image represents the dominant color of the particular child variant of the parent retail item.

2. The method of claim 1, wherein assigning each color in the set of colors to a respective color subgroup comprises:

determining that a color distance between a first color value associated with a first color in a color space and a second color value associated with a second color in the color space is within a threshold color distance; and

assigning the first color and the second color to a same color subgroup.

3. The method of claim 2, wherein the color distance is a first color distance and the threshold color distance is a first threshold color distance, and wherein identifying a color cluster present in the second color histogram comprises:

determining that a second color distance between a third color value associated with a first representative color in the color space and a fourth color value associated with a second representative color in the color space is within a second threshold color distance; and

determining that the first representative color and the second representative color are included in the color cluster,

wherein the first threshold color distance is less than the second threshold color distance.

4. The method of claim 1, wherein determining a score for the color cluster comprises:

generating the score by dividing the number of pixels by a color distance between a color value associated with the color cluster and a color value of a color name associated with the particular child variant.

5. The method of claim 1, wherein modifying the first color histogram using the average color histogram comprises subtracting the average color histogram from the first color histogram to generate the second color histogram or subtracting the first color histogram from the average color histogram to generate the second color histogram.

6. A method, comprising:

identifying a set of images comprising a particular image of a particular child variant of a parent item;

determining a set of colors in the set of images;

assigning, by a computer processor, a first color and a second color in the set of colors to a color subgroup;

determining, by the computer processor, a representative color for the color subgroup;

generating, by the computer processor, a first color value distribution representation for the particular image, wherein the first color value distribution representation associates each pixel in the particular image having a color value corresponding to the first color or the second color with a color value of the representative color;

generating, by the computer processor, an aggregate color value distribution representation for the set of images;

modifying, by the computer processor, the first color value distribution representation based at least in part on the aggregate color value distribution representation to generate a second color value distribution representation for the particular image;

selecting, by the computer processor and based at least in part on the second color value distribution representation, the representative color as a dominant color for inclusion in a sample image corresponding to the particular child variant;

modifying webpage source content to cause the sample image to be generated using the dominant color; and

generating, responsive to a request for webpage content, the sample image comprising the dominant color, wherein the sample image represents the dominant color of the particular child variant of the parent item.

7. The method of claim 6, further comprising:

identifying, by the computer processor, a color cluster present in the second color value distribution representation;

determining, by the computer processor, a score for the color cluster;

determining, by the computer processor, whether the score meets or exceeds a threshold value; and

selecting, by the computer processor, the representative color as the dominant color based at least in part on whether the score meets or exceeds the threshold value.

8. The method of claim 6, wherein assigning the first color and the second color to the color subgroup comprises determining that a color distance between a first color value of the first color and a second color value of the second color is within a threshold color distance.

9. The method of claim 6, wherein assigning the first color to the color subgroup comprises:

determining that a distance between a color value of the first color and a respective color value corresponding to each of one or more other colors in the set of colors that have been assigned to an existing color subgroup exceeds a threshold color distance; and

assigning the first color to a new color subgroup, wherein the new color subgroup is the color subgroup.

10. The method of claim 6, wherein determining a representative color for the color subgroup comprises reducing the set of colors to a smaller set of representative colors by performing bit-depth compression of the set of colors.

11. The method of claim 6, wherein a color value corresponding to the representative color is a statistical quantity generated from a respective color value corresponding to each color in the color subgroup.

12. The method of claim 6, wherein the first color value distribution representation comprises a first color histogram, the second color value distribution representation comprises a second color histogram, and the aggregate color value distribution representation comprises an average color histogram for the set of images, and wherein modifying the first color histogram comprises one of:

i) subtracting, for each representative color, a respective number of pixels for the representative color in the first color histogram from a respective number of pixels for the representative color in the average color histogram or ii) subtracting, for each representative color, a respective number of pixels for the representative color in the average color histogram from a respective number of pixels for the representative color in the first color histogram.

13. The method of claim 12, wherein the representative color is a first representative color, wherein a second representative color corresponds to a static element in the set of images, and wherein a number of pixels corresponding to the second representative color in the second color histogram is less than a number of pixels corresponding to the second representative color in the first color histogram.

14. The method of claim 6, wherein the representative color is a first representative color and the color value of the first representative color is a first color value, and wherein identifying the color cluster comprises:

determining that the first color value is within a threshold color distance from a second color value of a second representative color in the second color value distribution representation; and

determining that the color cluster comprises the first representative color and the second representative color.

15. The method of claim 14, wherein determining the score for the color cluster comprises:

determining a strength metric for the color cluster based at least in part on a total number of pixels having color values that correspond to one or more colors in the color cluster; and

dividing the strength metric by a color distance between the color cluster and a color value of a color name associated with the particular child variant.

16. The method of claim 15, wherein the color distance between the color cluster and the color value of the color name comprises one of:

i) a color distance between the color value of the color name and one of the first color value, the second color value, or a third color value of a third representative color in the color cluster or ii) a color distance between an average color value of the color cluster and the color value of the color name.

17. The method of claim 6, further comprising:

storing, by the computer processor, a color value of the representative color; and

directing, by the computer processor, transmission of the color value of the representative color for generating a sample image of the particular child variant.

18. A system, comprising:

at least one memory storing computer-executable instructions; and

at least one processor configured to access the at least one memory and execute the computer-executable instructions to;

identify a set of images, wherein each image corresponds to a respective child variant of a parent item;

generate a first color value distribution representation corresponding to a particular image in the set of images, wherein the particular image corresponds to a particular child variant;

generate an aggregate color value distribution representation for the set of images;

generate a second color value distribution representation by modifying the first color value distribution representation based at least in part on the aggregate color value distribution representation;

identify a color cluster present in the second color value distribution representation;

determine a score for the color cluster;

determine whether the score meets or exceeds a threshold value;

select a particular color from the color cluster for inclusion in a sample image corresponding to the particular child variant based at least in part on whether the score meets or exceeds the threshold value;

modify webpage source content to cause the sample image to be generated using the particular color; and

generate, responsive to a request for webpage content, the sample image comprising the particular color, wherein the sample image represents the particular color of the particular child variant of the parent item.

19. The system of claim 18, wherein the at least one processor is further configured to execute the computer-executable instructions to:

determine a set of colors in the set of images;

assign a first subset of the set of colors to a first color subgroup and a second subset of the set of colors to a second color subgroup; and

determine a first representative color for the first color subgroup and a second representative color for the second color subgroup, wherein each color in the first subset is represented by a first color value corresponding to the first representative color and each color in the second subset is represented by a second color value corresponding to the second representative color.

20. The system of claim 18, wherein the color cluster is a first color cluster and the score is a first score, wherein it is determined that the first score meets or exceeds the threshold value, wherein the particular color selected from the first color cluster is a first color, and wherein the at least one processor is further configured to execute the computer-executable instructions to:

identify a second color cluster in the second color value distribution representation;

determine a second score for the second color cluster;

determine that the second score meets or exceeds the threshold value; and

select a second color from the second color cluster for inclusion in the sample image.