PRODUCT CATEGORY OPTIMIZATION FOR IMAGE SIMILARITY SEARCHING OF IMAGE-BASED LISTINGS IN A NETWORK-BASED PUBLICATION SYSTEM

• US 20110314031A1
• Filed: 03/28/2011
• Published: 12/22/2011
• Est. Priority Date: 03/29/2010
• Status: Active Grant

• Current Assignee: eBay Inc.
• Original Assignee: eBay Inc.
• Inventors: Chittar, Naren, Ghatare, Sanjay Pundlkrao, McDonald, Ryan, Roper, John, Gaikwad, Sandip
• Granted Patent: US 8,949,252 B2
• Time in Patent Office: Days
• US Class Current: 707/749
• CPC Class Codes: G06F 16/5838 using colour
• CPC Class Codes: G06F 16/5854 using shape and object relationship
• CPC Class Codes: G06Q 30/00 Commerce

Abstract

A weighted combination of attributes including but not limited to color, shape, pattern, brand, style, size, and material may be used to rank items when searching for similar items of a given category on an ecommerce system website. The weights on each attribute may be optimized by using feedback from users, the feedback indicating the importance of each attribute to a purchaser when searching for similar items of various categories. Alternatively, click-through patterns from the ecommerce system website can be mined and used to optimize the importance of each attribute to users by item category. The optimum attribute weights for that items category are used to rank the results in a search for similar items in that category.

Claims

1. A computer implemented method comprising:

using a computer processor, detecting a search query, the search query comprising an image of an item with specified attributes, the item being of a given category;

searching a database for a set of images of items of the given category, among a collection of images of items, that meet the specified attributes of the search query;

based on the searching, determining the set of images of items of the given category that meet the specified attributes; and

ranking at least some of the set of images of items of the given category that meet the specified attributes by weighting the at least some of the respective specified attributes with weights determined by the computer processor correlating feedback signals indicating relative importance of the specified attributes to prospective purchasers of the item of the given category.

2. The method of claim 1, wherein the feedback signals represent scores provided by a plurality of users.

3. The method of claim 1, further comprising determining the feedback signals by using training sets, the training sets for indicating importance of each attribute of the item of a given category to a prospective purchaser of the item.

4. The method of claim 3, further comprising providing the training sets to a plurality of users for evaluation, the training sets including target images, at least some of the target images having associated therewith a plurality of candidate images, wherein the feedback signals represent a score based on a similarity between each individual candidate image and the target image from the plurality of users.

5. The method of claim 4, wherein the individual target images correspond to stored categories in the database.

6. The method of claim 3, further comprising using a crowd sourcing algorithm to manage distribution of the training sets and collection of the feedback signals.

7. The method of claim 1, wherein the feedback signals comprise click-through patterns obtained by data mining click-through patterns from a website.

8. The method of claim 7, wherein the click-through patterns are received from client machines associated with purchasers of the item of the given category.

9. The method of claim 1, searching a database being based on a similarity search.

10. A non-transitory computer-readable storage medium having embedded therein a set of instructions which, when executed by one or more processors of a computer, causes the computer to execute operations comprising:

detecting a search query, the search query comprising an image of an item with specified attributes, the item being of a given category;

searching a database for a set of images of items of the given category, among a collection of images of items, that meet the specified attributes of the search query;

based on the searching, determining the set of images of items of the given category that meet the specified attributes; and

ranking at least some of the set of images of items of the given category that meet the specified attributes by weighting the at least some of the respective specified attributes with weights determined by the computer correlating feedback signals indicating relative importance of the specified attributes to prospective purchasers of the item of the given category.

11. The non-transitory computer-readable storage medium of claim 10, wherein the feedback signals represent scores provided by a plurality of users.

12. The non-transitory computer-readable storage medium of claim 10, wherein the operations further comprise determining the feedback signals by using training sets, the training sets for indicating importance of each attribute of the item of a given category to a prospective purchaser of the item.

13. The non-transitory computer-readable storage medium of claim 12, wherein the operations further comprise providing the training sets to a user for evaluation, the training sets including target images, at least some of the target images having associated therewith a plurality of candidate images, wherein the feedback signals represent a score based on a similarity between each individual candidate image and the target image.

14. The non-transitory computer-readable storage medium of claim 13, wherein the individual target images correspond to stored categories in the database.

15. The non-transitory computer-readable storage medium of claim 12, further comprising using a crowd sourcing algorithm to manage distribution of the training sets and collection of the feedback signals.

16. The non-transitory computer-readable storage medium of claim 10, wherein the feedback signals comprise click-through patterns obtained by data mining click-through patterns from a website.

17. The non-transitory computer-readable storage medium of claim 16, wherein the click-through patterns received from client machines are associated with purchasers of the item of the given category.

18. The non-transitory computer-readable storage medium of claim 10, wherein searching a database is based on a similarity search.

19. A system comprising:

a computer processor and storage configured to execute;

a detector module to detect a search query, the search query comprising an image of an item with specified attributes, the item bring of a given category;

a search module to search a database for a set of images of items of the given category, among a collection of images of items, that meet the specified attributes of the search query and, based on the searching, determining the set of images of items of the given category that meet the specified attributes; and

a ranking module to rank at least some of the set of images of items of the given category that meet the specified attributes by weighting the at least some of the respective specified attributes with weights determined by the computer processor correlating feedback signals indicating relative importance of the specified attributes to prospective purchasers of the item of the given category.

20. The system of claim 19, wherein the feedback signals represent scores provided by a plurality of users.