System and method for selective use of examples
Systems and methods for selective use of examples are provided. For example, information associated with external devices may be obtained, weights may be assigned to examples using the obtained information, and inference models may be generated using the assignment of weights. In some examples, update information associated with the external devices may be obtained, the assignment of weights may be updated using the update information, and updated inference models may be generated using the updated assignment of weights.
1. A system for image processing, the system comprising:
at least one communication device;
at least one storage device configured to store a plurality of examples; and
at least one processor configured to:
receive from an external device, using the at least one communication device, scene information including ambient noise information, the scene information is a result of analyzing one or more images captured using at least one image sensor included in the external device from an environment, wherein the ambient noise information includes a distribution of noise levels;
use the scene information to assign weights to the plurality of examples based on ambient noise corresponding to the examples; use a statistical distance between the distribution of the noise levels and a distribution corresponding to an example of the plurality of examples to assign weight to the example; and generate
an inference model based on the plurality of examples and the weights.
2. The system of claim 1 , wherein the at least one processor is further configured to:
receive from an external device, using the at least one communication device, an update to the scene information to obtain updated scene information;
use the updated scene information to update the assignment of weights to the plurality of examples;
generate an updated inference model based on the plurality of examples and the updated assignment of weights;
compare the assignment of weights with the updated assignment of weights;
based on a result of the comparison, determine that a difference between the assignment of weights and the updated assignment of weights is below a selected threshold; and
based on said determination, withhold generating the updated inference model.
3. The system of claim 1 , wherein the scene information comprises at least one of information associated with at least one object depicted in the one or more images, information associated with at least one person depicted in the one or more images, information associated with at least one position visible in the one or more images, and information differentiating between indoor scenes and outdoor scenes.
4. The system of claim 1 , wherein the at least one processor is further configured to:
generate one or more synthetic examples based on the scene information; and
further base the inference model on the one or more synthetic examples.
5. The system of claim 1 , the at least one processor is further configured to:
obtain settings information associated with the at least one image sensor; and
base the assignment of weights on the settings information.
6. The system of claim 1 , the at least one processor is further configured to:
obtain information related to geographical location associated with the at least one image sensor; and
base the assignment of weights on the information related to the geographical location.
7. A method for image processing, the method comprising:
receiving scene information including ambient noise information, the scene information is a result of analyzing one or more images captured using at least one image sensor from an environment, wherein the ambient noise information includes a distribution of noise levels;
accessing a plurality of examples;
using the scene information to assign weights to the plurality of examples based on ambient noise corresponding the examples; using a statistical distance between the distribution of the noise levels and a distribution corresponding to an example of the plurality of examples to assign weight to the example; and generating
an inference model based on the plurality of examples and the weights.
8. The method of claim 7 , wherein assigning the weights to the plurality of examples comprises analyzing the plurality of examples to determine the assignment of the weights.
9. The method of claim 7 , wherein the inference model is configured to be applied to one or more images captured using the at least one image sensor from the environment.
10. The method of claim 7 , further comprising:
receiving an update to the scene information to obtain updated scene information;
using the updated scene information to update the assignment of weights to the plurality of examples; and
generating an updated inference model based on the plurality of examples and the updated assignment of weights.
11. The method of claim 10 , further comprising:
comparing the assignment of weights with the updated assignment of weights;
based on a result of the comparison, determining that a difference between the assignment of weights and the updated assignment of weights is below a selected threshold; and
based on said determination, withholding generating the updated inference model.
12. The method of claim 7 , wherein the scene information comprises at least one of information associated with at least one object depicted in the one or more images and information associated with at least one person depicted in the one or more images.
13. The method of claim 7 , wherein the scene information comprises at least one of information associated with at least one position visible in the one or more images, and information differentiating between indoor scenes and outdoor scenes.
14. The method of claim 7 , wherein at least one of the weights is zero.
15. The method of claim 7 , wherein the scene information comprises a first distribution, at least one example of the plurality of examples comprises a second distribution, and the weight assigned to the at least one example is a function of a statistical distance between the first distribution and the second distribution.
16. The method of claim 7 , further comprising:
generating one or more synthetic examples based on the scene information; and
further basing the inference model on the one or more synthetic examples.
17. The method of claim 7 , further comprising:
obtaining settings information associated with the at least one image sensor; and
basing the assignment of weights on the settings information.
18. The method of claim 7 , further comprising:
obtaining information related to geographical location associated with the at least one image sensor; and
basing the assignment of weights on the information related to the geographical location.
19. A non-transitory computer readable medium storing data and computer implementable instructions for carrying out a method for image processing, the method comprising:
obtaining scene information including ambient noise information, the scene information is a result of analyzing one or more images captured using at least one image sensor from an environment, wherein the ambient noise information includes a distribution of noise levels;
accessing a plurality of examples;
using the scene information to assign weights to the plurality of examples based on ambient noise corresponding to the examples; using a statistical distance between the distribution of the noise levels and a distribution corresponding to an example of the plurality of examples to assign weight to the example; and generating
an inference model based on the plurality of examples and the weights.
20. The non-transitory computer readable medium of claim 19 , wherein the statistical distance is a Bhattacharyya distance.