Organizing search results based upon clustered content

• US 10,691,755 B2
• Filed: 06/05/2017
• Issued: 06/23/2020
• Est. Priority Date: 06/11/2010
• Status: Active Grant

• Current Assignee: Microsoft Technology Licensing LLC (Microsoft Corporation)
• Original Assignee: Microsoft Technology Licensing LLC (Microsoft Corporation)
• Inventors: Wang, Xin-Jing, Zhang, Lei, Ma, Wei-Ying
• Primary Examiner(s): Wong, Leslie
• Application Number:
  US15/613,697
• Publication Number:
  US 20170270222A1
• Time in Patent Office: 1,114 Days
• CPC Class Codes: G06F 16/51 Indexing; Data structures therefor; Storage structures
• CPC Class Codes: G06F 16/583 using metadata automatically derived from the content
• CPC Class Codes: G06F 16/9038 Presentation of query results

Abstract

Many users make use of search engines to locate desired internet content by submitting search queries. For example, a user may search for photos, applications, websites, videos, documents, and/or information regarding people, places, and things. Unfortunately, search engines may provide a plethora of information that a user may be left to sift through to find relevant content. Accordingly, one or more systems and/or techniques for organizing search results are disclosed herein. In particular, user generated content, such as photos, may be retrieved based upon a search query. The user generated content may be grouped into clusters of user generated content having similar features. Search results of the search query may be obtained and organized based upon comparing the search results with the clusters. The organized search results and/or a table of content including the clusters may be presented to provide an enhanced user experience.

Claims

1. A method performed by a computing device, the method comprising:

performing an image querying process to retrieve a plurality of positively-rated images that are related to a particular query, the plurality of positively-rated images having associated positive characterizations provided by users;

performing image processing on the plurality of positively-rated images to extract visual features from the plurality of positively-rated images;

defining an image cluster using the extracted visual features, the image cluster comprising particular positively-rated images retrieved in the image querying process that are associated with a particular topic related to the particular query;

performing a search result querying process to retrieve a plurality of search result web pages, separate from the plurality of positively-rated images retrieved via the image querying process, that are responsive to the particular query;

using the image cluster as a comparison model to perform a ranking process on the plurality of search result web pages, wherein the ranking process involves ranking the search result web pages relative to one another based at least on similarity of the search result web pages to the particular positively-rated images of the image cluster retrieved in the image querying process; and

providing a response to the particular query based at least on the ranked search result web pages.

2. The method of claim 1, further comprising:

extracting textual features from the plurality of positively-rated images; and

using the textual features with the visual features to define the image cluster.

3. The method of claim 2, wherein the extracted visual features include a scale-invariant feature transform (“

SIFT”

) feature of the plurality of positively-rated images.

4. The method of claim 2, further comprising:

executing a multimodal Dirichlet Process Mixture Sets model upon both the visual features and the textual features, the multimodal Dirichlet Process Mixture Sets model outputting the image cluster.

5. The method of claim 1, wherein the plurality of positively-rated images are retrieved from an image repository of user-generated images.

6. The method of claim 5, wherein the search result web pages are retrieved from a general-purpose search engine.

7. A system comprising:

a processing unit; and

a volatile or non-volatile memory storing instructions that, when executed by the processing unit, cause the processing unit to;

perform an image query to retrieve a plurality of positively-rated images that are related to a particular query for use in a comparison model, the plurality of positively-rated images having associated positive characterizations provided by users;

process the plurality of positively-rated images to extract features;

define a first image cluster of the comparison model using the extracted features, the first image cluster comprising first positively-rated images retrieved via the image query that are associated with a first topic related to the particular query;

define a second image cluster of the comparison model using the extracted features, the second image cluster comprising second positively-rated images retrieved via the image query that are associated with a second topic related to the particular query;

perform a search query to retrieve separate image search results, other than the plurality of positively-rated images retrieved for use in the comparison model, that are responsive to the particular query;

use the first image cluster and the second image cluster as the comparison model to rank the separate image search results; and

output the ranked separate image search results in response to the particular query.

8. The system of claim 7, wherein the instructions, when executed by the processing unit, cause the processing unit to:

organize the separate image search results using the first image cluster to identify first image search results that are associated with the first topic and using the second image cluster to identify second image search results that are associated with the second topic.

9. The system of claim 8, wherein the instructions, when executed by the processing unit, cause the processing unit to:

based at least on first features of the first positively-rated images, determine a first descriptive name of the first image cluster;

based at least on second features of the second positively-rated images, determine a second descriptive name of the second image cluster; and

provide a user interface that presents the first descriptive name together with a particular positively-rated first image from the first image cluster and the second descriptive name together with a particular positively-rated second image from the second image cluster.

10. The system of claim 9, wherein the instructions, when executed by the processing unit, cause the processing unit to:

responsive to first user input selecting the first image cluster, present the first image search results and not the second image search results on the user interface; and

responsive to second user input selecting the second image cluster, present the second image search results and not the first image search results on the user interface.

11. The system of claim 7, wherein the instructions, when executed by the processing unit, cause the processing unit to:

select the plurality of positively-rated images for the comparison model based at least on positive user comments characterizing the plurality of positively-rated images.

12. The system of claim 7, wherein the instructions, when executed by the processing unit, cause the processing unit to:

select the plurality of positively-rated images for the comparison model based at least on frequency of access by users to the plurality of positively-rated images relative to frequency of access by the users to other images that are also related to the particular query.

13. The system of claim 7, wherein the instructions, when executed by the processing unit, cause the processing unit to:

select the plurality of positively-rated images for the comparison model based at least on user voting data for the plurality of positively-rated images relative to user voting data for other images that are also related to the particular query.

14. The system of claim 7, wherein the instructions, when executed by the processing unit, cause the processing unit to:

define the first image cluster based at least on first visual features of the first positively-rated images and define the second image cluster based at least on second visual features of the second positively-rated images.

15. The system of claim 7, wherein the instructions, when executed by the processing unit, cause the processing unit to:

define the first image cluster based at least on first text features of the first positively-rated images and define the second image cluster based at least on second text features of the positively-rated second images.

16. A volatile or non-volatile memory or storage device storing instructions which, when executed by a processing unit, cause the processing unit to perform acts comprising:

querying an image repository to retrieve a plurality of positively-rated images that are related to a particular query, the plurality of positively-rated images having associated positive characterizations provided by users;

performing image processing on the plurality of positively-rated images to extract features from the plurality of positively-rated images;

defining different image clusters using the extracted features, the different image clusters comprising respective positively-rated images retrieved from the image repository that are associated with respective topics related to the particular query;

querying a search engine separate from the image repository to retrieve a plurality of search result web pages that are responsive to the particular query;

using the different image clusters as a comparison model to rank the search result web pages retrieved from the search engine based at least on similarity to the respective positively-rated images of the different image clusters; and

outputting the ranked search result web pages in response to the particular query.

17. The volatile or non-volatile memory or storage device of claim 16, the acts further comprising:

presenting an interactive user interface having representations of the different image clusters; and

responsive to user selection of a selected image cluster from the interactive user interface, presenting selected search result webpages associated with a selected topic of the selected image cluster on the interactive user interface.

18. The volatile or non-volatile memory or storage device of claim 17, the acts further comprising:

presenting the respective topics on the interactive user interface together with the representations of the different image clusters and the selected search result webpages.

19. The volatile or non-volatile memory or storage device of claim 17, the acts further comprising:

as user input is received selecting different respective image clusters associated with different topics, presenting different sets of search result webpages associated with the different topics on the interactive user interface.