Identification and/or verification by a consensus network using sparse parametric representations of biometric images
Image data is run through a neural network, and the neural network produces a vector representation of the image data. Random sparse sampling masks are created. The vector representation of the image data is masked with each of the random sparse sampling masks, the masking generating corresponding sparsely sampled vectors. The sparsely sampled vectors are transmitted to nodes of a consensus network, wherein a sparsely sampled vector of the sparsely sampled vectors is transmitted to a node of the consensus network. Votes from the nodes of the consensus network are received. Whether a consensus is achieved in the votes is determined. Responsive to determining that the consensus is achieved, at least one of identification and verification of the image data may be provided.
1. A computer-implemented method, comprising:
receiving image data;
running the image data through a neural network, the neural network producing a vector representation of the image data;
creating random sparse sampling masks;
masking the vector representation of the image data with each of the random sparse sampling masks, the masking generating corresponding sparsely sampled vectors;
transmitting the sparsely sampled vectors to nodes of a consensus network, wherein a sparsely sampled vector of the sparsely sampled vectors is transmitted to a node of the consensus network;
receiving votes from the nodes of the consensus network;
determining whether a consensus is achieved in the votes; and
responsive to determining that the consensus is achieved, providing at least one of identification and verification of the image data.
2. The method of claim 1 , wherein the neural network comprises a convolutional neural network (CNN), and the vector representation is obtained from a final layer of the CNN.
3. The method of claim 1 , wherein the neural network comprises a deep neural network, the vector representation obtained from a final layer of the deep neural network.
4. The method of claim 1 , wherein each of the random sparse sampling masks is unique.
5. The method of claim 1 , wherein each of the sparsely sampled vectors is unique.
6. The method of claim 1 , wherein the image data is of a biometric feature.
7. The method of claim 1 , wherein the image data is of an object feature.
8. A computer readable storage medium storing a program of instructions executable by a machine to perform a method comprising:
receiving image data;
running the image data through a neural network, the neural network producing a vector representation of the image data;
creating random sparse sampling masks;
masking the vector representation of the image data with each of the random sparse sampling masks, the masking generating corresponding sparsely sampled vectors;
transmitting the sparsely sampled vectors to nodes of a consensus network, wherein a sparsely sampled vector of the sparsely sampled vectors is transmitted to a node of the consensus network;
receiving votes from the nodes of the consensus network;
determining whether a consensus is achieved in the votes; and
responsive to determining that the consensus is achieved, providing at least one of identification and verification of the image data.
9. The computer readable storage medium of claim 8 , wherein the neural network comprises a convolutional neural network (CNN), and the vector representation is obtained from a final layer of the CNN.
10. The computer readable storage medium of claim 8 , wherein the neural network comprises a deep neural network, the vector representation obtained from a final layer of the deep neural network.
11. The computer readable storage medium of claim 8 , wherein each of the random sparse sampling masks is unique.
12. The computer readable storage medium of claim 8 , wherein each of the sparsely sampled vectors is unique.
13. The computer readable storage medium of claim 8 , wherein the image data is of a biometric feature.
14. The computer readable storage medium of claim 8 , wherein the image data is of an object feature.
15. A system comprising:
at least one hardware processor coupled to a computer network;
a memory device coupled to the at least one hardware processor; and
a trained neural network stored on the memory device;
the at least one hardware processor operable to perform at least:
receive image data;
run the image data through the neural network, the neural network producing a vector representation of the image data;
create random sparse sampling masks;
mask the vector representation of the image data with each of the random sparse sampling masks, the masking generating corresponding sparsely sampled vectors;
transmit the sparsely sampled vectors to nodes of a consensus network, wherein a sparsely sampled vector of the sparsely sampled vectors is transmitted to a node of the consensus network;
receive votes from the nodes of the consensus network;
determine whether a consensus is achieved in the votes; and
responsive to determining that the consensus is achieved, provide at least one of identification and verification of the image data.
16. The system of claim 15 , wherein the neural network comprises a convolutional neural network (CNN), and the vector representation is obtained from a final layer of the CNN.
17. The system of claim 15 , wherein the neural network comprises a deep neural network, the vector representation obtained from a final layer of the deep neural network.
18. The system of claim 15 , wherein each of the random sparse sampling masks is unique.
19. The system of claim 15 , wherein each of the sparsely sampled vectors is unique.
20. The system of claim 15 , wherein the image data is of a biometric feature.