Computer-implemented method, a computer program product and a computer system for image processing

Patent metadata

Field	Value
Publication number	US-9349077-B2
Application number	US-201113807629-A
Country	US
Kind code	B2
Filing date	Jun 30, 2011
Priority date	Jul 2, 2010
Publication date	May 24, 2016
Grant date	May 24, 2016

Abstract

The present description refers in particular to a computer-implemented method, a computer program product and a computer system for image processing, the method comprising: receiving at least one user image; identifying a plurality of image classification elements of the user image by: assigning an initial classification to the user image, wherein the initial classification is based on temporal data associated with the user image; determining at least one image label that globally describes content of the user image; calculating a label correctness value for each image label; recognizing at least one image component of the user image; calculating a component correctness value for each image component; correlating the image label and the image component using the label correctness value and the component correctness value, whereby a correlated image label and a correlated image component are identified; applying a rule to determine a category of the user image, wherein the rule is based on at least one of the following: the temporal data, the correlated image label and the correlated image component; and producing a final classification of the user image including the following image classification elements: the initial classification, the correlated image label, the correlated image component, and the category.

First claim

The invention claimed is:

1. A method comprising:
receiving, by one or more processors, a user image;
identifying, by the one or more processors, a plurality of image classification elements of the user image by:
assigning an initial classification to the user image,
the initial classification being based on temporal data associated with the user image;

determining at least one image label that globally describes content of the user image;
calculating a label correctness value for each image label of the at least one image label;
segmenting the user image to identify at least one image component of the user image;
calculating a component correctness value for each image component of the at least one image component;
correlating the at least one image label and the at least one image component using the label correctness value for each image label of the at least one image label and the component correctness value for each image component of the at least one image component,
correlating the at least one image label and the at least one image component including selecting an image label, of the at least one image label, as a correlated image label based on the at least one image component and selecting an image component, of the at least one image component, as a correlated image component,
another image label, of the at least one image label, being excluded based on the at least one image component;

applying a rule to determine a category of the user image,
the rule being based on at least one of the temporal data, the correlated image label, or the correlated image component; and

producing, by the one or more processors, a final classification of the user image,
the final classification including the initial classification, the correlated image label, the correlated image component, and the category.

2. The method of claim 1 , wherein identifying the plurality of image classification elements comprises:
receiving a geographic location associated with the image; and
determining a place name associated with the geographic location,
wherein the final classification further includes the place name.

3. The method of claim 2 , wherein identifying the plurality of image classification elements comprises:
determining an event based on the temporal data and the geographic location,
wherein the final classification further includes the event.

4. The method of claim 1 , wherein identifying the plurality of image classification elements comprises:
deriving a weather indicator from the temporal data,
wherein the final classification further includes the weather indicator.

5. The method of claim 1 , wherein segmenting the user image to identify the at least one image component of the user image comprises recognizing a plurality of image components, and
wherein identifying the plurality of image classification elements comprises:
associating an image component classification with a first image component of the plurality of image components,
wherein the final classification further includes the image component classification.

6. The method of claim 5 , wherein the first image component is recognized as a face,
the method further comprising:
associating a name with the face; and
determining a mood based on an expression of the face,
wherein the final classification further includes the name and the mood.

7. The method of claim 1 , further comprising:
verifying the initial classification;
verifying the final classification of the user image,
wherein training information is received for producing a subsequent final classification of a subsequent image.

8. The method of claim 1 , further comprising at least one of:
associating the user image with a stored image based on the initial classification; or
associating the user image with a stored image based on the final classification.

9. The method of claim 1 , wherein receiving the user image further comprises:
receiving a plurality of user images; and
retrieving at least one of the plurality of user images from an image sharing network.

10. The method of claim 9 , further comprising:
displaying a plurality of image classification elements,
wherein each image classification element is displayed based on a quantity of the plurality of user images associated with the image classification element;

receiving user input selecting an image classification element from the plurality of image classification elements; and
displaying a preview of a selected image from the plurality of user images,
wherein the selected image classification element is included in the final classification of the selected image.

11. The method of claim 9 , further comprising:
receiving a query including at least one query term;
matching the query term to a matching classification element; and
retrieving a matching image from the plurality of user images based on matching the query term to the matching classification element,
wherein the matching classification element is included in a final classification of the matching image.

12. The method of claim 9 , wherein the plurality of user images comprises a query image and a response image,
the method further comprising:
receiving a query comprising the query image;
matching a classification element of the query image with a classification element of the response image; and
retrieving the response image based on the query.

13. A non-transitory computer-readable medium storing instructions, the instructions comprising:
one or more instructions which, when executed by one or more processors, cause the one or more processors to:
receive a user image from a client,
the user image being associated with an initial classification that is based on temporal data associated with the user image;

identify a plurality of image labels that describes the user image;
calculate a label correctness value for each image label of the plurality of image labels;
segment the user image to identify an image component of the user image;
calculate a component correctness value for the image component;
correlate the plurality of image labels and the image component, using the label correctness value for each image label of the plurality of image labels and the component correctness value, to identify a correlated image label of the plurality of image labels,
another image label, of the plurality of image labels, being excluded;

apply a rule to determine a category of the user image,
the rule being applied to at least one of the temporal data, the correlated image label, or the image component; and

produce a final classification based on the initial classification, the correlated image label, the image component, and the category,
the final classification being transmitted to the client.

14. The non-transitory computer-readable medium of claim 13 , where the one or more instructions to produce the final classification include one or more instructions to produce the final classification further based on a list of image components depicted in the user image and names of people depicted in the user image.

15. A system comprising:
a server to:
segment the user image to identify a user image from a client,
the user image being associated with an initial classification that is based on temporal data associated with the user image;

identify a plurality of image labels that globally describes the user image;
calculate a label correctness value for each image label of the plurality of image labels;

recognize an image component of the user image;
calculate a component correc

Assignees

• Capuozzo Giuseppe
• Borghini Marco
• Mammoliti Fabio
• Accenture Global Services Ltd

Inventors

• Capuozzo Giuseppe
• Borghini Marco
• Mammoliti Fabio

Classifications

• G06V20/70 Primary

  Labelling scene content, e.g. deriving syntactic or semantic representations · CPC title

• G06F18/241

  relating to the classification model, e.g. parametric or non-parametric approaches · CPC title

• G06F16/587

  using geographical or spatial information, e.g. location · CPC title

• G06K9/00288

  Physics · mapped topic

• G06K9/00677

  Physics · mapped topic