Visual search utilizing color descriptors

• US 10,346,684 B2
• Filed: 06/09/2017
• Issued: 07/09/2019
• Est. Priority Date: 03/15/2013
• Status: Active Grant

• Current Assignee: A9.com Incorporated (Amazon.com, Inc.)
• Original Assignee: A9.com Incorporated (Amazon.com, Inc.)
• Inventors: Dhua, Arnab Sanat Kumar, Arora, Himanshu, Ramesh, Sunil
• Primary Examiner(s): Youssef, Menatoallah
• Application Number:
  US15/618,946
• Publication Number:
  US 20170277948A1
• Time in Patent Office: 760 Days
• CPC Class Codes: G06F 2218/12 Classification; Matching
• CPC Class Codes: G06V 10/56 relating to colour

Abstract

Various embodiments provide a method for computing color descriptors of product images. For example, a number of fine color representatives can be determined to describe color variation in an image as a histogram by assigning a saturation value and a brightness value to a plurality of color hues. For each pixel of the image, the closest color among a defined fine color representative set is computed. In this example, each of the pixels is assigned a color ID corresponding to their closest matching fine color representative and at least one family color ID corresponding one or more pure color families. In this example, a histogram of the color representatives and a histogram for the color families are computed. A single color vector descriptor for the image is then determined by combining the family histogram with the color representative histogram.

Claims

1. A computer-implemented method, comprising:

defining, in a color space, individual representative colors within threshold distances from a color represented in a pixel;

determining color families for the individual representative colors;

determining a count for individual color families that is associated with the individual representative colors;

determining a first histogram for the count for respective color families for the individual representative colors and a second histogram for the threshold distances of the individual representative colors;

combining the first histogram and the second histogram to determine a descriptor; and

associating the descriptor to an image.

2. The computer-implemented method of claim 1, further comprising:

displaying the color on the pixel as part of a plurality of pixels on a screen of a computing device.

3. The computer-implemented method of claim 1, further comprising:

defining the color families as including a plurality of color values, the individual representative colors include a plurality of representative color values.

4. The computer-implemented method of claim 3, wherein determining the count for the individual color families further comprises:

comparing a color value of a color family of the individual color families with a representative color value for the individual representative colors; and

determining that the color value is a threshold distance from the representative color value; and

associating a representative color of the representative color value with the color family.

5. The computer-implemented method of claim 1, wherein the individual color families include a red, green, and blue (RGB) color values in an RGB color space.

6. The computer-implemented method of claim 1, wherein at least one representative color in the representative colors is assigned to two color families.

7. A computing device, comprising:

a processor; and

memory including instructions that, when executed by the processor, cause the computing device to;

define, in a color space, individual representative colors within threshold distances from a color represented in a pixel;

determine color families for the individual representative colors;

determine a count for individual color families that is associated with the individual representative colors;

provide a first histogram for the count and a second histogram for the threshold distances;

combine the first and the second histograms to provide a descriptor; and

attaching the descriptor to an image in an image database.

8. The computing device of claim 7, wherein each of the color families includes a plurality of red, green, blue (RGB) color values, each color value is an RGB color value, and the color space is an RGB color space.

9. The computing device of claim 7, wherein at least one representative color in the representative colors is assigned to two color families.

10. The computing device of claim 7, wherein the instructions, when executed by the processor, further cause the computing device to:

define the color families, each color family including a plurality of color values, and wherein each of representative colors include different representative color values.

11. The computing device of claim 10, wherein the instructions, when executed by the processor, in part to determine the count for each of the color families, further cause the computing device to:

compare at least a color value of each of the color families with a representative color value for each of the individual representative colors;

determine that the color value is a threshold distance from the representative color value; and

associate a representative color of the representative color value with the color family.

12. The computing device of claim 7, further comprising:

display the color on the pixel as part of a plurality of pixels on a screen of a the computing device.

13. The computing device of claim 7, wherein the instructions, when executed by the processor, further cause the computing device to:

provide at least an image in response to a requested color received at the computing device, the image comprising color similarity determined by a similarity measure between a first vector of the descriptor associated with the image and a second vector of a color descriptor of the requested color.

14. The computing device of claim 13, the similarity measure is determined by computing a dot product of the first vector and the second vector.

15. The computing device of claim 13, wherein each of the first vector and the second vector is computed based on a global histogram of the individual representative colors and a family histogram of the individual color families.

16. A non-transitory computer-readable storage medium storing instructions that, when executed by at least one processor, cause a computing device to:

define, in a color space, individual representative colors within threshold distances from a color represented in a pixel;

determine color families for the individual representative colors;

determine a count for individual color families that is associated with the individual representative colors;

determine a first histogram for the count for respective color families of the individual representative colors and a second histogram for the threshold distances of the individual representative colors;

combine the first histogram and the second histogram to determine a descriptor; and

associate the descriptor to an image.

17. The non-transitory computer-readable storage medium of claim 16, wherein the individual color families include red, green, and blue (RGB) color values in an RGB color space.

18. The non-transitory computer-readable storage medium of claim 16, wherein at least one representative color in the representative colors is assigned to two color families.

19. The non-transitory computer-readable storage medium of claim 16, wherein the instructions that, when executed by the at least one processor, further cause the computing device to:

determine, based at least in part on the count for the individual color families, color similarity between the descriptor of the image and a requested color; and

display the image in response to a request for the requested color.

20. The non-transitory computer-readable storage medium of claim 19, wherein the color similarity is determined by determining a similarity measure between a first vector of the descriptor and a second vector of a color descriptor of the requested color.