Image processing and authentication of unclonable functions
A method of authenticating an item is disclosed which includes applying light to an item, wherein the item includes a random distribution of fluorescent particles disposed thereon, capturing one or more images from the item, generating a cryptographic pattern from the one or more captured images by a host processing system, communicating the cryptographic pattern to a remote processing system having a plurality of cryptographic keys in a database each uniquely associated with a corresponding item, comparing the cryptographic pattern with the plurality of cryptographic keys in the database, and communicating a positive evaluation for authentication to the host processing system if a match is found between one of the plurality of cryptographic keys and the communicated cryptographic pattern.
1 . A method of authenticating an item, comprising:
randomly distributing one or more types of edible fluorescent particles on to an area of an item (Area), thus generating a physical unclonable function on the item based on locations of the edible fluorescent particles and illumination wavelength;
applying light to the Area;
capturing one or more images from the Area;
binarizing each of the one or more images by associating a “1” or a “0” to each pixel in the one or more images whereby a “1” is assigned to a pixel representing a local highest fluorescent intensity within each of the edible fluorescent particles and a “0” is assigned to all other pixels, thereby determining a spatial location based on location of a pixel with the maximum intensity for each edible fluorescent particle;
generating a cryptographic pattern from the one or more captured images based on the determined spatial locations by a host processing system;
communicating the cryptographic pattern to a remote processing system having a plurality of cryptographic keys in a database each uniquely associated with a corresponding item;
comparing the cryptographic pattern with the plurality of cryptographic keys in the database; and
communicating a positive evaluation for authentication to the host processing system if a match is found between one of the plurality of cryptographic keys and the communicated cryptographic pattern, else communicating a negative evaluation,
wherein the one or more types of edible fluorescent particles are chosen from the group consisting of enhanced cyan fluorescent protein (eCFP), enhanced green fluorescent protein (eGFP), enhanced yellow fluorescent protein (eYFP), and mKate2 (far-red) fluorescent protein.
2 . The method of claim 1 , wherein the evaluation for authentication is based on a statistical match between one of the plurality of cryptographic keys and the cryptographic pattern.
3 . The method of claim 2 , wherein the statistical match is based on linear regression with a predetermined threshold for a P-value.
4 . The method of claim 1 , further comprising:
reducing noise in the one or more captured images;
detecting the one or more types of edible fluorescent particles amongst N×M pixels in the one or more captured images; and
compressing the binarized data stream in order to generate the cryptographic pattern.
5 . The method of claim 4 , wherein the compression of the binarized data is based on a von Neumann compression.
6 . The method of claim 4 , further comprising filtering the light source to provide selective wavelengths.
7 . The method of claim 6 , wherein the one or more types of edible fluorescent particles include a plurality of fluorescent compounds, each generating a different fluorescence in response to a selected light wavelength, whereby the cryptographic pattern is a linear combination of the compressed binarized data stream associated with sequentially generated fluorescence responses.
8 . A system of authenticating an item, comprising:
a remote processing system, configured to hold a plurality of cryptographic keys in a database each uniquely associated with a corresponding item;
a light source, configured to apply light to an area of an item (Area), wherein the Area includes a random distribution of one or more types of edible fluorescent particles disposed thereon, thus generating a physical unclonable function on the item based on locations of the edible fluorescent particles and illumination wavelength;
an image capture device having an N×M imaging sensor, configured to capture one or more images each having NM pixels from the Area;
a host processing system, configured to:
capture one or more images from the Area utilizing the image capture device;
binarize each of the one or more images by associating a “1” or a “0” to each pixel in the one or more images whereby a “1” is assigned to a pixel representing a highest fluorescent intensity within each of the edible fluorescent particles and a “0” is assigned to all other pixels, thereby determine a spatial location based on location of a pixel with the maximum intensity for each edible fluorescent particle;
generate a cryptographic pattern from the captured image; and
communicate the cryptographic pattern to the remote processing system,
wherein the remote processing system is configured to compare the cryptographic pattern with the plurality of cryptographic keys in the database, and communicate a positive evaluation for authentication to the host processing system if a match is found between one of the plurality of cryptographic keys and the received cryptographic pattern, else communicate a negative evaluation,
wherein the one or more types of edible fluorescent particles are chosen from the group consisting of enhanced cyan fluorescent protein (eCFP), enhanced green fluorescent protein (eGFP), enhanced yellow fluorescent protein (eYFP), and mKate2 (far-red) fluorescent protein.
9 . The system of claim 8 , wherein the evaluation for authentication is based on a statistical match between one of the plurality of cryptographic keys and the cryptographic pattern.
10 . The system of claim 8 , wherein the statistical match is based on linear regression with a predetermined threshold for a P-value.
11 . The system of claim 8 , wherein the host processing system is further configured to:
reduce noise in the one or more captured images;
detect the one or more types of edible fluorescent particles amongst the N×M pixels in the one or more captured images; and
compress the binarized data stream in order to generate the cryptographic pattern.
12 . The system of claim 11 , wherein the compression of the binarized data is based on a von Neumann compression.
13 . The system of claim 11 , wherein the light source is filtered to provide selective wavelengths.
14 . The system of claim 13 , wherein the one or more types of edible fluorescent particles include a plurality of fluorescent compounds, each generating a different fluorescence in response to a selected light wavelength, whereby the cryptographic pattern is a linear combination of the compressed binarized data stream associated with sequentially generated fluorescence responses.
15 . A method of authenticating an item, comprising:
applying light to an are of an item (Area), wherein the Area includes a distribution of colored particles disposed thereon, thus generating a physical unclonable function on the item based on locations of the colored particles and illumination wavelength;
capturing one or more RGB images from the Area by an image capturing device, each RGB image producing a channel data stream from one of Red, Green, and Blue channels of the device;
binarizing each of the one or more images by associating a “1” or a “0” to each pixel in the one or more images whereby a “1” is assigned to a pixel representing a local highest fluorescent intensity within each of the edible fluorescent particles and a “0” is assigned to all other pixels, thereby determining a spatial location based on location of a pixel with the maximum intensity for each edible fluorescent particle;
generating a cryptographic pattern from the one or more captured RGB images by a host processing system, based on:
receiving a hyperspectral dataset representing a priori hyperspectral data of items of a population of interest;
pairing the corresponding Red, Green, and Blue data streams with the hyperspectral dataset,
obtaining a transformation matrix adapted to convert an item-specific RGB image dataset into an item-specific hyperspectral dataset for the image capturing device,
generating an item-specific hyperspectral dataset using the transformation matrix,
determining the cryptographic pattern from the item-specific hyperspectral dataset;
communicating the cryptographic pattern to a remote processing system having a plurality of cryptographic keys in a database each uniquely associated with a corresponding item;
comparing the cryptographic pattern with the plurality of cryptographic keys in the database; and
communicating a positive evaluation for authentication to the host processing system if a match is found between one of the plurality of cryptographic keys and the communicated cryptographic pattern, else communicating a negative evaluation,
wherein the colored particles are chosen from the group consisting of enhanced cyan fluorescent protein (eCFP), enhanced green fluorescent protein (eGFP), enhanced yellow fluorescent protein (eYFP), and mKate2 (far-red) fluorescent protein.
16 . The method of claim 15 , wherein the evaluation for authentication is based on a statistical match between one of the plurality of cryptographic keys and the cryptographic pattern.
17 . The method of claim 16 , wherein the statistical match is based on linear regression with a predetermined threshold for a P-value.
18 . The method of claim 15 , further comprising:
reducing noise in the one or more captured RGB images;
detecting the colored particles amongst NM pixels in the one or more captured RGB images; and
compressing the binarized data stream in order to generate the cryptographic pattern.
19 . The method of claim 18 , wherein the compression of the binarized data is based on a von Neumann compression.
20 . The method of claim 15 , wherein the colored particles include a plurality of fluorescent compounds, each generating a different fluorescence in response to a selected light wavelength, whereby the cryptographic pattern is a linear combination of the compressed binarized data stream associated with sequentially generated fluorescence responses.