Content selection based on image content

• US 9,305,025 B2
• Filed: 03/14/2013
• Issued: 04/05/2016
• Est. Priority Date: 12/14/2012
• Status: Active Grant

• Current Assignee: Google LLC (Alphabet Inc.)
• Original Assignee: Google Inc. (Alphabet Inc.)
• Inventors: Wang, Jingbin, Chen, Xiangrong, Rosenberg, Charles J.
• Primary Examiner(s): Hoang, Son T
• Application Number:
  US13/827,000
• Publication Number:
  US 20140172884A1
• Time in Patent Office: 1,118 Days
• Field of Search: None
• US Class Current: 1/1
• CPC Class Codes: G06F 16/24578 using ranking
• CPC Class Codes: G06F 16/5846 using extracted text
• CPC Class Codes: G06F 16/5866 using information manually generated, e.g. tags, keywords, comments, manually generated location and time information
• CPC Class Codes: G06F 16/9566 URL specific, e.g. using aliases, detecting broken or misspelled links

Abstract

Methods, systems, and apparatus, include computer programs encoded on a computer-readable storage medium, for determining keywords for an image that supports an overlay content item. A method includes identifying, using one or more processors, an image that is to support an overlay content item, the image being presented on a web site and including a portion that is designated as being enabled to receive and display the overlay content item; evaluating pixel data associated with the image including determining one or more labels that are associated with content included within the image; and determining one or more keywords for the image based at least in part on the one or more labels.

Claims

1. A computer-implemented method using one or more computing processors for content selection, the method comprising:

identifying a landing page associated with a creative, the creative being a content item that is to be displayed on a web site responsive to a received request for content, wherein the creative includes first associated keywords and wherein the landing page includes at least one image and other content;

determining a prominence level of the at least one image included in the landing page based at least in part on a relative size, location or both of the at least one image;

when the prominence level exceeds a predetermined level, evaluating pixel data associated with the at least one image including determining one or more labels that are associated with the content included within the at least one image;

determining one or more additional keywords for the creative based at least in part on the one or more labels;

replacing one or more of the first associated keywords with the one or more additional keywords producing an updated set of keywords; and

storing the landing page including the updated set of keywords to enable searching and serving of the creative responsive to subsequent received requests for the content.

2. The method of claim 1, further comprising ranking the first associated keywords and the one or more additional keywords and selecting one or more top-ranked keywords including at least one of the one or more additional keywords as selection criteria to be associated with the creative.

3. The method of claim 2, wherein ranking the first associated keywords and the one or more additional keywords includes:

providing as an input to an evaluation engine the first associated keywords, the one or more additional keywords, and the image;

receiving as an output of the evaluation engine a relevance score for each of the first associated keywords and for each of the or more additional keywords that indicates a relevance of a given keyword to the image; and

ranking the first associated keywords and the one or more additional keywords based on the relevance scores.

4. The method of claim 1, further comprising augmenting the first associated keywords including adding the one or more additional keywords to the first associated keywords and using the augmented keywords as selection criteria to be associated with the creative.

5. The method of claim 1, wherein the determining the prominence level for the image further includes determining a relative portion of the content that is embodied in the image.

6. A computer-implemented method using one or more computing processors for content selection, the method comprising:

identifying a creative associated with an online campaign, the creative being a content item that is to be displayed on a web site responsive to a received request for content, wherein the creative includes first associated keywords;

determining a prominence level of an image included in the creative based at least in part one a relative size, location or both of the image;

when the prominence level exceeds a predetermined level, providing as an input to an evaluation engine the first associated keywords and the image and receiving as an output a relevance score for each of the first associated keywords that indicates a relevance of a given keyword to the image;

identifying one or more additional different keywords based on pixel data associated with the image; and

ranking the one or more additional different keywords;

comparing rankings of the first associated keywords and the one or more additional different keywords;

suggesting keywords for use with the creative including at least one of the one or more additional different keywords that are more relevant to the image than at least one of the first associated keywords;

replacing one or more of the first associated keywords with the one or more additional keywords producing an updated set of keywords;

storing the creative including the updated set of keywords to enable searching; and

serving of the landing page responsive to subsequent received requests for the content.

7. The method of claim 6, wherein the suggesting keywords for use with the creative further includes:

evaluating the pixel data associated with the image including determining one or more labels that are associated with the content included within the image; and

determining the suggested keywords for the image based at least in part on the one or more labels.

8. The method of claim 6, wherein the determining the prominence level of the image further includes determining a relative portion of the content that is embodied in the image.

9. A computer-implemented method using one or more computing processors, the method comprising

identifying a content item that is to be displayed on or as a web site responsive to a received request for content wherein the content item includes at least two images;

determining a prominence level of each image included in the content item based at least in part on a relative size, location or both of each image;

determining a most prominent image included in the content item based on the determined prominence level for each image;

when a prominence level for the most prominent image exceeds a predetermined level, evaluating pixel data associated with the most prominent image including determining one or more labels that are associated with the content included within the most prominent image;

determining one or more keywords for the content item based at least in part on the one or more labels; and

storing the content item including the determined one or more keywords to enable searching and serving of the content responsive to subsequent received requests for the content.

10. The method of claim 9, wherein the content item is a creative.

11. The method of claim 9, wherein the content item is a landing page associated with a creative.

12. The method of claim 9, wherein the content item is an advertisement.

13. The method of claim 9, wherein the determining the prominence level of a given image further includes determining a relative portion of the content that is embodied in the given image.

14. The method of claim 9, further comprising:

evaluating one or more supplied keywords associated with the content item including;

providing as an input to an evaluation engine the one or more supplied keywords and the image;

receiving as an output of the evaluation engine a relevance score for each supplied keyword that indicates a relevance of a given supplied keyword to the image; and

accepting the one or more supplied keywords based on the relevance scores or alternatively suggesting using the determined keywords based on the one or more labels as being keywords for use as selection criteria to be associated with the content item.