Filter selection in search environments

• US 9,477,748 B2
• Filed: 12/20/2013
• Issued: 10/25/2016
• Est. Priority Date: 12/20/2013
• Status: Active Grant

• Current Assignee: Adobe Inc.
• Original Assignee: Adobe Systems Incorporated (Adobe Inc.)
• Inventors: Becherer, Nico
• Primary Examiner(s): Coby, Frantz
• Application Number:
  US14/136,920
• Publication Number:
  US 20150178381A1
• Time in Patent Office: 1,040 Days
• Field of Search: 707/711, 707/722, 707/732, 707/718, 707/748, 707/754
• US Class Current: 1/1
• CPC Class Codes: G06F 16/2246 Trees, e.g. B+trees
• CPC Class Codes: G06F 16/335 Filtering based on additional data, e.g. user or group profiles filtering in web context G06F16/9535, G06F16/9536
• CPC Class Codes: G06F 16/951 Indexing; Web crawling techniques
• CPC Class Codes: G06F 16/9532 Query formulation
• CPC Class Codes: G06F 16/9535 Search customisation based on user profiles and personalisation
• CPC Class Codes: G06F 16/9538 Presentation of query results

Abstract

Techniques for utilizing sets of filters to reduce a large number of searchable assets to a meaningful or reduced number of searchable assets. Feature information may be extracted from a particular asset of a set of training assets to create an artificial term. A reduction ratio may then be calculated by utilizing the created artificial term to filter the set of training assets. The reduction ratio may represent the ratio of training assets that contain a particular artificial term. A plurality of filters and their associated reduction ratios may be created this way by utilizing the set of training assets. This process can also involve receiving a requested reduction ratio associated with a set of searchable assets. A combination of filters may then be selected which, when applied to the set of searchable assets, results in a reduced number of searchable assets according to the requested reduction ratio.

Claims

1. A computer-implemented method, comprising:

identifying filter reduction ratios for a plurality of filters, each respective filter reduction ratio corresponding to a respective reduction of training assets based on use of a respective filter, wherein identifying a filter reduction ratio for a filter of the plurality of filters comprises;

creating an artificial term for the filter based on color information;

filtering the training assets using the created artificial term; and

calculating the filter reduction ratio for the filter based at least in part on the filtered training assets and the training assets;

receiving a requested reduction ratio; and

selecting, by a computing system, a subset of filters from the plurality of filters based at least in part on the requested reduction ratio and the filter reduction ratios, the subset of filters selected such that applying the subset of filters to a set of a number of media assets reduces the number of media assets in the set according to the requested reduction ratio.

2. The computer-implemented method of claim 1, wherein identifying a filter reduction ratio for a filter of the plurality of filters comprises:

creating an artificial term by identifying a weighted palette of colors in a color histogram based on color information of an image, the artificial term representing a dominant color of the weighted palette of colors;

filtering the training assets using the created artificial term; and

calculating the filter reduction ratio for the filter based at least in part on the filtered training assets and the training assets.

3. The computer-implemented method of claim 1, wherein selecting the subset of filters further comprises:

evaluating multiple combinations of the plurality of filters to identify a respective combined reduction ratio for each of the multiple combinations;

assigning a score to each of the multiple combinations, each respective score representing a correlation between the respective combined reduction ratio of each combination and the requested reduction ratio; and

selecting a combination of the multiple combinations based at least in part on the assigned score.

4. The computer-implemented method of claim 1, wherein selecting the subset of filters further comprises evaluating multiple combinations of the plurality of filters utilizing one or more Boolean operations to identify a respective combined reduction ratio for each of the multiple combinations.

5. The computer-implemented method of claim 1, wherein selecting the subset of filters further comprises evaluating multiple combinations of the plurality of filters utilizing a hierarchical filter combination algorithm to identify a respective combined reduction ratio for each of the multiple combinations.

6. The computer-implemented method of claim 1, wherein selecting the subset of filters further comprises:

evaluating multiple combinations of the plurality of filters; and

selecting a combination of the multiple combinations based at least in part on an overlap threshold representing filter similarity.

7. The computer-implemented method of claim 1, wherein the media assets include at least one of an image file, an audio file, a text file, or a movie file.

8. The computer-implemented method of claim 1 further comprising receiving search query criteria, wherein the subset of filters is selected from the plurality of filters based at least in part on the search query criteria.

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

receiving search query criteria comprising an image, wherein the subset of filters is selected from the plurality of filters based at least in part on the image.

10. The computer-implemented method of claim 1 further comprising providing the set of the number of media assets for display via a graphical user interface, wherein the graphical user interface is further configured to receive and communicate user preferences used to determine the requested reduction ratio.

11. A system, comprising:

one or more processors; and

memory including instructions executable by the one or more processors that, when executed by the one or more processors, cause the system to at least;

identify filter reduction ratios for a plurality of filters, each respective filter reduction ratio corresponding to a respective reduction of training assets based on use of a respective filter, by executing further instructions that cause the system to;

create an artificial term for the filter based on color information;

filter the training assets using the created artificial term; and

calculate the filter reduction ratio for the filter based at least in part on the filtered training assets and the training assets; and

in response to receiving a requested reduction ratio;

select a subset of filters from the plurality of filters based at least in part on the requested reduction ratio and the filter reduction ratios, the subset of filters selected such that applying the subset of filters to a set of a number of media assets reduces the number of media assets in the set according to the requested reduction ratio.

12. The system of claim 11, wherein the instructions cause the system to identify a reduction ratio for a filter by:

creating a hash term based on extracted feature information, wherein the hash term corresponds to the filter and is comprised of data of a fixed length;

filtering the training assets using the hash term; and

calculating the reduction ratio for the filter based at least in part on the filtered training assets and the training assets.

13. The system of claim 11, wherein the instructions cause the system to identify a reduction ratio for a filter by:

creating a histogram based on extracted feature information;

identifying at least one hash term in the histogram;

filtering the training assets using the at least one hash term; and

calculating the reduction ratio for the filter based at least in part on the filtered training assets and the training assets.

14. The system of claim 11, wherein the instructions cause the system to identify a reduction ratio for a filter by:

extracting frequency information from each training asset of the training assets;

creating a frequency histogram based on the extracted frequency information;

creating hash terms based on the frequency histogram;

filtering the training assets by the created hash terms; and

calculating the reduction ratio based at least in part on the filtered training assets and the training assets, wherein the filtered training assets are comprised of a subset of the training assets that include a similar proportion of hash terms as the created hash terms.

15. The system of claim 11, further comprising:

a user-interface configured to receive instructions that identify a reduction ratio.

16. A non-transitory computer-readable storage medium having stored thereon computer-executable instructions that, when executed by one or more processors of a computer system, cause the computer system to perform operations comprising:

identifying reduction ratios for each of a plurality of filters, wherein identifying the reduction ratios for each of the plurality of filters comprises;

creating an artificial term for the filter based on color information;

filtering training assets using the created artificial term; and

calculating the reduction ratio for a particular filter of the plurality of filters based at least in part on the filtered training assets and the training assets; and

selecting a subset of filters from the plurality of filters based at least in part on a particular reduction ratio such that applying the subset of filters to a set of a number of media assets reduces the number of media assets in the set according to the particular reduction ratio.

17. The non-transitory computer readable storage medium of claim 16 wherein creating the artificial term includes analyzing an image for the color information.