Robust and efficient image identification

1. A method of analyzing a digital image to determine an unordered local descriptor describing the digital image, the method comprising:

selecting a subset of feature points within multiple local extrema,

wherein each feature point has an associated blur level and scale octave;

creating a vector for each of the feature points, each feature point vector based on the normalized intensity values of a patch around the feature point,

wherein the patch size is determined by the associated octave;

dividing the patch into multiple overlapping patches,

wherein each overlapping patch is represented by a vector;

concatenating each overlapping patch vector by projecting the multiple overlapping patch vectors onto an eigen space; and

creating an unordered local descriptor for the digital image,

wherein the unordered local descriptor is a compact representation of feature points defining the digital image, and

wherein the compact representation is a multi-vector descriptor formed from the concatenated patch vectors.

2. The method of claim 1 , further comprising:

generating a scale-space pyramid configured from a series of blur levels and scale octaves of the digital image; and

localizing multiple local extrema in the scale-space pyramid, wherein the multiple local extrema are greater than a predetermined intensity, and

wherein dividing the patch into multiple overlapping patches includes dividing the patch into five or more overlapping patches.

3. The method of claim 1 , wherein each descriptor uses a scale space determined by the associated octave.

4. The method of claim 3 , wherein each descriptor is invariant to scale.

5. The method of claim 1 , wherein the patch is re-sized to a constant size.

6. The method of claim 1 , further comprising:

constructing a hyperplane to provide a line of demarcation between a first region and second region,

wherein the demarcation line is determined by edge distributions of the first region and the second region, each of the first and second region differing in intensity; and

defining the digital image to be analyzed based on the demarcation line.

7. The method claim 6 , wherein the first region is a continuous tone region and the second region is a graphics region.

8. The method of claim 6 , wherein the hyperplane includes a set of hyperplanes.

9. The method of claim 6 , defining a functional margin measured by a distance between the constructed hyperplane and a nearest feature point defined in the digital image.

10. The method of claim 1 , wherein the descriptor is resilient to distortions in a digital image plane.

11. The method of claim 1 , further comprising determining an orthogonal coordinate system based on the largest variance between each feature point vector.

12. The method of claim 1 , further comprising: generating a catalog of images, each image having an associated unordered local descriptor.

13. A non-transitory computer readable medium including instructions that when performed by a processor perform a method comprising:

selecting a subset of feature points within multiple local extrema,

wherein each feature point has an associated blur level and scale octave;

creating a vector for each of the feature points,

wherein the vector is based on the normalized intensity values of a patch around the feature point, wherein the patch size is determined by the associated octave;

dividing the patch into multiple overlapping patches,

wherein each overlapping patch is represented by a vector;

concatenating each overlapping patch vector by projecting the overlapping patches onto an eigen space; and

creating a compact representation of feature points defining the digital image,

wherein the compact representation is a multi-vector descriptor formed from the concatenated patch vectors.

14. The non-transitory computer readable medium of claim 13 , the method further comprising:

generating a scale-space pyramid, the scale-spaced pyramid configured from a series of blur levels and scale octaves of the digital image; and

localizing multiple local extrema in the scale-space pyramid,

wherein the multiple local extrema are greater than a predetermined intensity, and

wherein dividing the patch into multiple overlapping patches includes dividing the patch into five or more overlapping patches.

15. The method of claim 13 , wherein each descriptor is invariant to scale.

16. The method of claim 13 , wherein the patch is re-sized to a constant size.

17. The method of claim 13 , wherein the descriptor is resilient to distortions in a digital image plane.

18. The method of claim 13 , further comprising: determining an orthogonal coordinate system based on the largest variance between each feature point vector.

19. The method of claim 13 , further comprising: generating a catalog of images, each image having an associated unordered local descriptor.

20. A system of analyzing a digital image to determine an unordered local descriptor describing the digital image, the system comprising:

means for selecting a subset of feature points within multiple local extrema,

wherein each feature point has an associated blur level and scale octave;

means for creating a vector for each of the feature points, the vector based on the normalized intensity values of a patch around the feature point,

wherein the patch size is determined by the associated octave;

means for dividing the patch into multiple overlapping patches,

wherein each overlapping patch is represented by a vector;

means for concatenating each overlapping patch vector by projecting the overlapping patch vectors onto an eigen space; and

means for creating an unordered local descriptor for the digital image,

wherein the descriptor is a compact representation of feature points defining the digital image, and

wherein the compact representation is a multi-vector descriptor formed from the concatenated patch vectors.