Feature vector generation and encryption

1. A method of generating an image vector, the method comprising:

extracting feature data from an image, the feature data including:

location data that indicates a plurality of locations within the image; and

a plurality of feature vectors that each correspond to one of the plurality of locations of the location data in which each feature vector represents a feature of an object depicted in the image at its corresponding location;

determining a feature-data plot of each of the feature vectors extracted from the image based on relative positions of the plurality of locations with respect to each other;

selecting a grid that overlays the feature-data plot of the feature vectors extracted from the image based on a particular probability of distribution of the plurality of feature vectors with respect to cells of the grid;

generating a first representative feature vector for a first cell of the grid based on a first feature vector set of one or more first feature vectors of the feature-data plot that are included in the first cell;

generating a second representative feature vector for a second cell of the grid based on a second feature vector set of one or more second feature vectors of the feature-data plot that are included in the second cell;

generating a single feature vector that represents the object in the image based on the first representative feature vector and the second representative feature vector; and

encrypting the single feature vector.

2. The method of claim 1 , wherein the generating the first representative feature vector comprises averaging the one or more first feature vectors of the first feature vector set.

3. The method of claim 2 , wherein the averaging the one or more first feature vectors of the first feature vector set comprises averaging the one or more first feature vectors of the first feature vector set based on a weighted average of the first feature vector set.

4. The method of claim 1 , wherein the selecting the grid comprises selecting the grid based on the first cell of the grid and the second cell of the grid having an even probability of including one or more feature vectors.

5. The method of claim 1 , wherein the generating the single feature vector comprises concatenating the first representative feature vector and the second representative feature vector.

6. The method of claim 5 , wherein the concatenating the first representative feature vector and the second representative feature vector comprises ordering the first representative feature vector and the second representative feature vector based on an ordering sequence of cells within the grid.

7. The method of claim 6 , wherein the ordering the first representative feature vector and the second representative feature vector based on the ordering sequence of cells within the grid comprises ordering the first representative feature vector and the second representative feature vector based on a particular permutation associated with a particular security level.

8. The method of claim 1 , wherein:

the image is a first image;

the object is a first object;

the encrypted single feature vector is an encrypted first single feature vector; and

the method further comprises:

comparing the encrypted first single feature vector to an encrypted second single feature vector, wherein the encrypted second single feature vector represents a second object of a second image; and

determining whether the first object and the second object are the same based on the comparing the encrypted first single feature vector to the encrypted second single feature vector.

9. A system of generating an image vector, the system comprising:

one or more non-transitory computer-readable storage media having instructions stored thereon; and

one or more processors communicatively coupled to the one or more computer-readable storage media and configured to cause the system to perform operations in response to executing the instructions stored on the one or more computer-readable storage media, the operations comprising:

extracting feature data from an image, the feature data including:

location data that indicates a plurality of locations within the image; and

a plurality of feature vectors that each correspond to one of the plurality of locations of the location data in which each feature vector represents a feature of an object depicted in the image at its corresponding location;

determining a feature-data plot of each of the feature vectors extracted from the image based on relative positions of the plurality of locations with respect to each other;

selecting a grid that overlays the feature-data plot of the feature vectors extracted from the image based on a particular probability of distribution of the plurality of feature vectors with respect to cells of the grid;

generating a first representative feature vector for a first cell of the grid based on a first feature vector set of one or more first feature vectors of the feature-data plot that are included in the first cell;

generating a second representative feature vector for a second cell of the grid based on a second feature vector set of one or more second feature vectors of the feature-data plot that are included in the second cell;

generating a single feature vector that represents the object in the image based on the first representative feature vector and the second representative feature vector; and

encrypting the single feature vector.

10. The system of claim 9 , wherein the first representative feature vector comprises a first averaged feature vector of the one or more first feature vectors of the first feature vector set.

11. The system of claim 10 , wherein the first averaged feature vector of the one or more first feature vectors of the first feature vector set is based on a weighted average of the one or more first feature vectors of the first feature vector set.

12. The system of claim 9 , wherein the grid is selected based on the first cell of the grid and the second cell of the grid having an even probability of including one or more feature vectors.

13. The system of claim 9 , wherein the single feature vector comprises a concatenated vector that includes the first representative feature vector and the second representative feature vector.

14. The system of claim 13 , wherein the concatenated vector that includes the first representative feature vector and the second representative feature vector is based on an ordering sequence of cells within the grid.

15. The system of claim 14 , wherein the ordering sequence of cells within the grid is based on a particular permutation associated with a particular security level.

16. The system of claim 9 , wherein:

the image is a first image;

the object is a first object;

the encrypted single feature vector is an encrypted first single feature vector; and

the operations further comprise:

comparing the encrypted first single feature vector to an encrypted second single feature vector, wherein the encrypted second single feature vector represents a second object of a second image; and

determining whether the first object and the second object are the same based on the comparing the encrypted first single feature vector to the encrypted second single feature vector.