Methods for improving the diversity of image search results

• US 8,171,043 B2
• Filed: 10/24/2008
• Issued: 05/01/2012
• Est. Priority Date: 10/24/2008
• Status: Active Grant

• Current Assignee: R2 Solutions LLC (Acacia Research Corporation)
• Original Assignee: Yahoo! Inc. (Apollo Global Management, Inc.)
• Inventors: Murdock, Vanessa, Van Zwol, Roelof, Pueyo, Lluis Garcia, Ramirez Camps, Georgina
• Primary Examiner(s): PHAM, KHANH B
• Application Number:
  US12/257,991
• Publication Number:
  US 20100114933A1
• Time in Patent Office: 1,285 Days
• Field of Search: 707/748, 707/750, 707/759, 707/765, 707/767
• US Class Current: 707/767
• CPC Class Codes: G06F 16/58 Retrieval characterised by using metadata, e.g. metadata not derived from the content or metadata generated manually

Abstract

Techniques are described to increase the diversity or focus of image search results. A user submits an original query to search for images. A server generates a first results set by executing the original query using metadata associated with each image. The server selects, from the first results set, a specified number of results ranked highest and generates a list of terms from the metadata of each of the results selected. The terms may be only the tags of the results. The server generates an updated query using terms in the list that may be weighted based on the frequency of the term in the list or include only a specified number of the highest occurring terms in the list. The server generates a second results set by executing the updated query using metadata associated with each image. The second results set is then stored and displayed to the user.

Claims

1. A computer-implemented method to generate a results set of images, comprising:

receiving an original query from a user to search a plurality of images, each of which is associated with metadata;

generating a first results set based on executing the original query relative to tags from the metadata associated with each image in said plurality of images;

selecting, from the first results set, a specified number of results ranked highest in the first results set;

generating a list comprising terms from tags from the metadata associated with each result of the specified number of results;

for each term in the list, calculating a weight for that term based on a frequency of occurrence of that term in the list;

generating a single updated query comprising terms that are in the list and that are weighted based on said weights;

generating a second results set by executing only the updated query relative to metadata associated with each image in said plurality of images;

storing the second results set on a non-volatile or volatile computer readable storage medium; and

displaying, as a result of the original query, only images that are associated with metadata containing terms in the second results set, without executing any query other than the updated query to generate the second results set.

2. The method of claim 1, further comprising:

generating a second updated query comprising terms that are in the list and that are weighted based on said weights;

assigning, to each term in the second updated query that is also a term of the original query, a specified weight that is greater than weights of all other non-original query terms in the second updated query;

generating a third results set based on executing the second updated query relative to all metadata associated with each image in said plurality of images; and

storing the third results set on a non-volatile or volatile computer readable storage medium.

3. The method of claim 1, further comprising:

generating a second updated query comprising terms that are in the list and that are weighted based on said weights;

assigning, to each term in the second updated query that is also a term of the original query, a specified weight that is less than weights of all other non-original query terms in the second updated query;

generating a third results set based on executing the second updated query relative to all metadata associated with each image in said plurality of images; and

storing the third results set on a non-volatile or volatile computer readable storage medium.

4. A computer-implemented method to generate a results set of images, comprising:

receiving an original query from a user to search a plurality of images, each of which is associated with metadata;

generating a first results set based on executing the original query relative to tags from the metadata associated with each image in said plurality of images;

selecting, from the first results set, a specified number of results ranked highest in the first results set;

generating a list comprising terms from tags from the metadata associated with each result of the specified number of results;

for each term in the list, calculating a weight for that term based on a quantity of separate users that tagged a same image with that term;

generating an updated query comprising terms that are in the list and that are weighted based on said weights;

excluding all terms of the original query from the updated query;

after said excluding, generating a second results set based on executing the updated query relative to tags from the metadata associated with each image in said plurality of images; and

storing the second results set on a non-volatile or volatile computer readable storage medium.

5. A computer-implemented method to generate a results set of images, comprising:

receiving an original query from a user to search a plurality of images, each of which is associated with metadata;

generating a first results set based on executing the original query relative to metadata associated with each image in said plurality of images;

selecting, from the first results set, a specified number of results ranked highest in the first results set;

generating a list comprising terms from metadata associated with each result of the specified number of results;

for each term in the list, calculating a weight for that term based on ratings that users, excluding users that tagged an image with that term, gave to that term as an assessment of that term'"'"'s appropriateness to an image;

generating an updated query comprising terms that are in the list and that are weighted based on said weights;

excluding all terms of the original query from the updated query;

after said excluding, generating a second results set based on executing the updated query relative to metadata associated with each image in said plurality of images; and

storing the second results set on a non-volatile or volatile computer readable storage medium.

6. The method of claim 5, wherein the list comprises terms from only tags, including tags that were associated with particular images by users other than users that uploaded the particular images.

7. The method of claim 5, wherein the specified number of results is based upon the number of relevant results generated by the original query.

8. The method of claim 5, wherein the specified number of results is based upon a percentage of a quantity of relevant results generated by the original query.

9. The method of claim 5, wherein metadata comprises title, description, and tags.

10. The method of claim 5, wherein the specified number of results is less than all of the results in the first results set.

11. A computer-readable storage medium carrying one or more sequences of instructions which, when executed by one or more processors, causes the one or more processors to:

receive an original query from a user to search a plurality of images, each of which is associated with metadata;

generate a first results set based on executing the original query relative to metadata associated with each image in said plurality of images;

select, from the first results set, a specified number of results ranked highest in the first results set;

generate a list comprising terms from metadata associated with each result of the specified number of results;

for each term in the list, calculate a weight for that term based on a frequency of occurrence of that term in the list;

generate a single updated query comprising terms that are in the list and that are weighted based on said weights;

generate a second results set by executing only the updated query relative to metadata associated with each image in said plurality of images;

store the second results set on a non-volatile or volatile computer readable storage medium; and

display, as a result of the original query, only images that are associated with metadata containing terms in the second results set, without executing any query other than the updated query to generate the second results set.

12. The computer-readable storage medium of claim 11, wherein the list comprises terms from only tags, including tags that were associated with particular images by users other than users than uploaded the particular images.

13. The computer-readable storage medium of claim 11, wherein the specified number of results is based upon the number of relevant results generated by the original query.

14. The computer-readable storage medium of claim 11, wherein the specified number of results is based upon a percentage of a quantity of relevant results generated by the original query.

15. The computer-readable storage medium of claim 11, wherein metadata comprises title, description, and tags.

16. The computer-readable storage medium of claim 11, wherein the specified number of results is less than all of the results in the first results set.