IMAGE IDENTIFIERS AND METHODS AND SYSTEMS OF PRESENTING IMAGE IDENTIFIERS

• US 20140093122A1
• Filed: 10/02/2013
• Published: 04/03/2014
• Est. Priority Date: 10/02/2012
• Status: Active Grant

• Current Assignee: Simetricka Systems Ltd.
• Original Assignee: Nitzan Rabinowitz
• Inventors: RABINOWITZ, Nitzan
• Granted Patent: US 9,349,152 B2
• Time in Patent Office: Days
• US Class Current: 382/100
• CPC Class Codes: G06F 16/5854 using shape and object relationship
• CPC Class Codes: G06F 18/23 Clustering techniques
• CPC Class Codes: G06T 1/0021 Image watermarking
• CPC Class Codes: G06V 10/457 by analysing connectivity, e.g. edge linking, connected component analysis or slices
• CPC Class Codes: G06V 10/464 using a plurality of salient features, e.g. bag-of-words [BoW] representations

Abstract

A method of creating an image identifier of a target image. The method comprises providing a target image, identifying a plurality of connected component clusters in the target image, each the connected component cluster comprises a plurality of neighboring pixels having at least one common property, calculating a center of gravity for each the connected component cluster, generating an image identifier of the target image by using a plurality of elements to map the connected component clusters in a space, each the element represents a respective the at least one common property, a respective the center of gravity, and a size which corresponds with a number of the plurality of neighboring pixels of the respective the connected component cluster, and outputting the image identifier.

Claims

1. A method of creating an image identifier of a target image, comprising:

providing a target image;

identifying a plurality of connected component clusters in said target image, each said connected component cluster comprises a plurality of neighboring pixels having at least one common property;

calculating a center of gravity for each said connected component cluster;

generating an image identifier of said target image by using a plurality of elements to map said connected component clusters in a space, each said element represents a respective said at least one common property, a respective said center of gravity, and a size which corresponds with a number of said plurality of neighboring pixels of said respective said connected component cluster; and

outputting said image identifier.

2. The method of claim 1, wherein said image identifier is a graphical representation of said target image wherein each said element is represented as an object located in relation to other objects according to the location of a respective said center of gravity, having a visual size which corresponds with said number of neighboring pixels and colored according to said at least one common property.

3. The method of claim 2, wherein each said object is a circular object centered around said respective center of gravity, having a radius which corresponds with said number of neighboring pixels, and colored according to said respective at least one common property.

4. The method of claim 1, said at least one common property is a color pixel value.

5. The method of claim 1, said at least one common property is a grayscale pixel value.

6. The method of claim 1, further comprising printing said image identifier and using said printed imager identifier as a fingerprint of said target image.

7. The method of claim 1, wherein said outputting comprises storing said image identifier as an entry in an index;

further comprising matching said image identifier with another image identifier received in a search query and retrieving said target image is match between said image identifier and said another image identifier.

8. The method of claim 1, wherein said generating comprises uniformly quantizing pixel values in said target image.

9. The method of claim 1, wherein said image identifier is a graphical representation of said target image wherein each said element is represented as a line object located to connect each said connected component cluster with a selected point in said target image, said line having a visual size which corresponds with said number of neighboring pixels and colored according to said at least one common property.

10. The method of claim 1, wherein said generating comprises multiplying values of each said element to generate a Goedel number and using said Goedel number as said image identifier.

11. The method of claim 1, further comparing adding said image identifier to at least one of metadata and a file name of a file containing said target image.

12. A method of creating an image identifier of a target image, comprising:

providing a target image;

identifying a plurality of connected component clusters in said target image, each said connected component cluster comprises a plurality of neighboring pixels having at least one common property;

calculating a center of gravity for each said connected component cluster;

identifying a plurality of connecting lines each connecting a pair of members of a cluster group of said connected component clusters;

identifying a set of one or more intersection points, each said intersection point is identified in a location wherein a line group of said plurality of connecting lines intersect;

selecting at least one high resolution point from said set;

generating an image identifier by mapping a plurality of distances between said at least one high resolution point and a plurality of centers of gravity of members of said cluster group; and

outputting said image identifier.

13. The method of claim 12, wherein each said distance is represented as a visual line spread between said at least one high resolution point and a respective said center of gravity, said visual line having a width corresponds with said number of neighboring pixels, and colored according to said respective at least one common property.

14. The method of claim 12, wherein said plurality of distances are represented as a plurality of visual lines perpendicular to a common axis, each said visual line has a length that corresponds with a respective said distance between said at least one high resolution point and a respective said center of gravity, said plurality of visual lines are ordered according to said respective number of neighboring pixels and colored according to said respective at least one common property.

15. The method of claim 12, further comparing printing said image identifier as a fingerprint of said target image.

16. A method of identifying a match between a target image and at least one of a plurality of images, comprising:

providing a dataset having a plurality of image identifiers each associated with one of a plurality of images;

receiving a query for matching between a target image and at least one of said plurality of images;

generating a target image identifier associated with said target image; and

identifying a match between said target image identifier and at least one of said plurality of image identifiers;

outputting at least one of said plurality of images according to said match;

wherein each said image identifier is generated by identifying in a respective said associated image a plurality of connected component clusters, each said connected component cluster comprises a plurality of neighboring pixels having at least one common property, each said image identifier having a plurality of elements, each said element maps for one of said respective plurality of connected component clusters, a center of gravity, a respective said at least one common property and a size which corresponds with a number of said plurality of neighboring pixels.

17. A system of creating an image identifier of a target image, comprising:

a processor;

an image acquiring module providing a target image;

a blob mapping module which identifies a plurality of connected component clusters in said target image, each said connected component cluster comprises a plurality of neighboring pixels having at least one common property, calculates, using said processor, a center of gravity for each said connected component cluster, and generates an image identifier of said target image by using a plurality of elements to map said connected component clusters in a space, each said element represents a respective said at least one common property, a respective said center of gravity, and a size which corresponds with a number of said plurality of neighboring pixels of said respective said connected component cluster.

18. The system of claim 17, wherein said target image is an image stored to be rendered when a webpage is loaded, said image acquiring module identifies said target image while crawling a plurality of websites.

19. A device of identifying an object, comprising:

a repository which includes a dataset having a plurality of image identifiers each associated with one of a plurality of images;

an image sensor which captures a target image;

a processor;

an image identifier module which generates, using said processor, a target image identifier associated with said target image and identifies a match between said target image identifier and at least one of said plurality of image identifiers;

a presentation unit generates an indication according to said match;

wherein each said image identifier includes a plurality of elements indicative of a plurality of connected component clusters, each said connected component cluster comprises a plurality of neighboring pixels having at least one common property, each said element maps for one of said respective plurality of connected component clusters, a center of gravity, a respective said at least one common property and a size which corresponds with a number of said plurality of neighboring pixels.

20. The device of claim 19, wherein said device is a handheld device having said image identifier module installed as an application.