Systems and methods for large scale face identification and verification

1. A method of large-scale face representation comprising:

receiving, with an electronic processor, at least one image of a subject of interest;

applying, with the electronic processor, at least one subspace model as a splitting binary decision function on the at least one image of the subject of interest;

generating, with the electronic processor, at least one binary code from the at least one splitting binary decision function;

applying, with the electronic processor, a code aggregation model to combine the at least one binary codes generated by the at least one subspace model;

generating, with the electronic processor, an aggregated binary code from the code aggregation model;

using the aggregated binary code to provide a hashing scheme; and

executing the hashing scheme and performing a face recognition.

2. The method of claim 1 , further comprising executing the hashing scheme and performing a face verification.

3. The method of claim 1 , wherein applying, with the electronic image processor, at least one subspace model includes applying at least one subspace model using dictionary learning.

4. The method of claim 1 , further comprising applying the code aggregation model based on preserving mutual information.

5. The method of claim 1 , further comprising using the at least one subspace model as splitting nodes in a one or more tree-based decision system.

6. The method of claim 5 , wherein each tree in the one or more tree-based decision system includes one or more hierarchically connected split nodes and one or more leaf nodes.

7. The method of claim 6 , further comprising associating each split node with a weak learner.

8. The method of claim 6 , further comprising storing data associated with data points of one or more leaf nodes received during a training phase.

9. The method of 6 , further comprising evaluating each arriving data point and sending the data point to a left child or a right child based on the weak learner.

10. An image processing system comprising:

an electronic processor configured to

receive at least one image of a subject of interest;

apply at least one subspace model as a splitting binary decision function on the at least one image of the subject of interest;

generate at least one binary code from the at least one splitting binary decision function;

apply a code aggregation model to combine the at least one binary codes generated by the at least one subspace model;

generate an aggregated binary code from the code aggregation model;

use the aggregated binary code to provide a hashing scheme; and

execute the hashing scheme and perform a face recognition.

11. The image processing system of claim 10 , wherein the electronic processor is further configured to

apply at least one subspace model using dictionary learning.

12. The image processing system of claim 10 , wherein the electronic processor is further configured to

apply the code aggregation model based on preserving mutual information.

13. The image processing system of claim 10 , wherein the electronic processor is further configured to

use the at least one subspace model as splitting nodes in a one or more tree-based decision system.

14. The image processing system of claim 13 , wherein each tree in the one or more tree-based decision system includes one or more hierarchically connected split nodes and one or more leaf nodes.

15. The image processing system of claim 14 , wherein the electronic processor is further configured to

associate each split node with a weak learner.

16. The image processing system of claim 14 , wherein the electronic processor is further configured to

store data associated with data points of one or more leaf nodes received during a training phase.

17. The image processing system of claim 14 , wherein the electronic processor is further configured to

evaluate each arriving data point and send the data point to a left child or a right child based on the weak learner.