Digital fingerprinting object authentication and anti-counterfeiting system

1. A computer implemented method comprising:

acquiring digital image data representing an image of at least a portion of a physical object that belongs to a class of objects;

accessing a predetermined object feature template associated with the class of objects, the feature template identifying a plurality of individual authentication regions;

processing at least portions of the digital image data corresponding to the identified authentication regions so as to form a digital fingerprint of the object, wherein the processing comprises the following steps—

extracting at least one object feature to characterize the image data in each of the authentication regions;

for each of the extracted object features, forming at least one fingerprint feature vector that describes the object feature extracted from the image data; and

storing the fingerprint feature vectors of the object in an object database as part of a digital fingerprint of the object.

2. The method of claim 1 wherein, for each extracted object feature, the corresponding feature vector identifies a corresponding location and a shape of the extracted object feature.

3. The method of claim 1 wherein the object feature template comprises a stored data structure that specifies a class of objects, at least one authentication region definition for the class of objects, and at least one feature location in the authentication region for digital fingerprinting the object.

4. The method of claim 1 wherein the object feature template defines a method of extracting features from a region of interest to form corresponding feature vectors, and identifies methods for comparing the corresponding feature vectors to other feature vectors for authenticating the object.

5. The method of claim 1 including storing the digital fingerprint without storing the acquired digital image data in the database.

6. The method of claim 1 wherein extracting an object feature utilizes an algorithm among a class of algorithms characterized by one or more of edge detection, corner detection, blob detection, wavelet features; Gabor, gradient and steerable output filter histograms, scale-invariant feature transformation, active contours, shape contexts and parameterized shapes.

7. The method of claim 1 and further comprising:

accessing a digital fingerprint of a target object;

comparing the digital fingerprint of the target object to the stored digital fingerprint; and

determining an identity or authenticity of the target object based on the comparison; wherein comparing the digital fingerprint of the target object to the stored digital fingerprint comprises comparing fingerprint feature vectors of the digital fingerprint of the target object to the stored fingerprint feature vectors of the stored digital fingerprint.

8. The method of claim 7 and wherein the comparing fingerprint feature vectors includes utilizing an algorithm of a set of algorithms characterized by nearest neighbor, hashing, indexing feature sets with visual vocabularies, support vector machines, multilayer preceptor and random forests and ferns.

9. The method of claim 1 , wherein at least one of the extracted object features is subject to a predetermined manufacturing specification, and further comprising comparing the extracted object features to the predetermined manufacturing specification to determine whether the physical object complies with the predetermined manufacturing specification.

10. The method of claim 1 wherein the object comprises any one of government-issued documents, legal and financial documents, mail pieces, parcels, art, photographs, coins, currency, precious metals, gems, jewelry, apparel, mechanical parts, consumer goods, electronics, apparel, toys, integrated circuits, weapons, pharmaceuticals, drugs, chemicals, alcohol, tobacco and food and beverages.

11. An apparatus comprising:

a scanner arranged to acquire digital image data representing an image of at least a portion of a physical object;

a digital processor coupled to the scanner to receive the digital image data;

a memory accessible to the computer processor and storing an object feature template comprising data that specifies a class of objects, at least one authentication region for the class of objects, and at least one feature location in the authentication region for digital fingerprinting a object of the class of objects;

the computer processor configured to execute instructions to—

locate an authentication region of the scanned object based on accessing the stored object feature template;

process the digital image data to select only image data corresponding to the located authentication region;

in the selected image data, identify the object feature location based on the object feature template;

extract an object feature from the image data at the identified object feature location; and

store the extracted object feature as part of a digital fingerprint to identify the object without storing the acquired digital image data.

12. The apparatus of claim 11 wherein

the digital processor is further configured to store the extracted object feature as a feature vector.

13. The method of claim 1 wherein the object feature template is provided for a first class of objects and the template defines substantially all of the image as an authentication region.

14. The method of claim 13 wherein the first class of objects comprises a photograph.

15. The method of claim 13 wherein the object feature template is provided for a second class of objects and the template defines several regions selected where significant variations take place among different objects in the same class.