GROUPING IMAGES BY LOCATION

• US 20090161962A1
• Filed: 12/20/2007
• Published: 06/25/2009
• Est. Priority Date: 12/20/2007
• Status: Active Grant

• Current Assignee: Monument Peak Ventures, LLC (Dominion Harbor Enterprises, LLC)
• Original Assignee: Eastman Kodak Company
• Inventors: Loui, Alexander, Farmer, Jens C., Gallagher, Andrew C., Das, Madirakshi
• Granted Patent: US 8,150,098 B2
• Time in Patent Office: Days
• US Class Current: 382/203
• CPC Class Codes: G06F 16/58 Retrieval characterised by using metadata, e.g. metadata not derived from the content or metadata generated manually
• CPC Class Codes: G06F 16/583 using metadata automatically derived from the content
• CPC Class Codes: G06V 10/757 Matching configurations of points or features

Abstract

A method of grouping images captured in a common location, including receiving a collection of images; classifying the images into a set of events, where each image in the collection belongs to no more than one event; analyzing background region(s) of images from each event to determine one or more features that represent the event; and comparing features from at least two events to determine which events occurred in a common location.

Claims

1. A method of grouping images captured in a common location, comprising:

(a) receiving a collection of images;

(b) classifying the images into a set of events, where each image in the collection belongs to no more than one event;

(c) analyzing background region(s) of images from each event to determine one or more features that represent the event; and

(d) comparing features from at least two events to determine which events occurred in a common location.

2. The method of claim 1, further including:

(e) including a common location description to all events which have a common location.

3. The method of claim 2, further including:

(f) labeling one or more events with a descriptive label.

4. The method of claim 3, further including:

(g) determining if the descriptive label is a location descriptor, and then including that descriptive label for any events that occurred in a common location.

5. The method of claim 2, using a common location description to search the image collection or other image collections.

6. The method of claim 4, using a label with a common location description to search the image collection or other image collections.

7. The method of claim 1, further including augmenting the image collection with images obtained from an external source.

8. The method of claim 7, wherein the external source includes the Internet, private networks, or social networks.

9. The method of claim 3, wherein the descriptive label includes a GPS coordinate.

10. The method of claim 4, wherein a user initiates a query of the image collection, and the descriptor is used to find images that satisfy the query.

11. A method of matching background regions between a reference image and a target image, comprising

(a) determining the background regions for the reference image and the target image;

(b) locating feature points in the background of the reference and target images;

(c) matching feature points in the reference and target images;

(d) clustering matched points into spatially compact regions;

(e) filtering matched points that do not correspond to regions on the same objects; and

(f) computing a match score between the reference and target images.

12. The method of claim 11, wherein step (e) includes:

(i) removing clusters of points lacking strong correlation with matching cluster in target image;

(ii) removing points that are not coherent with the global trajectory of points; and

(iii) removing points that violate the compactness of clusters.

13. The method of claim 11, wherein the match score in step (f) can incorporate:

(i) number of matched points remaining after step 11(e); and

(ii) geographic distance between reference and target images, if available.