Selective image backup using trained image classifier
View Patent ↗Image backup using a trained image classifier is disclosed. In various embodiments, an image classifier is applied to a plurality of images to identify one or more images to be included in a save set of images. The save set of images are selectively stored to a second location according to one or more policies.
1. A system, comprising:
a processor configured to:
train an image classifier to identify an image in a search space, wherein the search space comprises a plurality of storage locations across a plurality of devices associated with a user, wherein the plurality of storage locations store multiple instances of the image;
apply the trained image classifier to a plurality of images stored across the plurality of devices associated with the user to identify one or more images to be included in a save set of images; and
selectively store the save set of images to a second location according to one or more policies, wherein the save set of images is deduplicated at the second location that is different than the plurality of storage locations, wherein the one or more policies indicate at least one of a frequency that the save set of images is selectively stored to the second location, a number of remote storage locations to store the save set of images, a type of storage that is to be used to store the save set of images, and/or a length of time in which the save set of images are to be stored at the second location; and
a memory coupled to the processor and configured to provide the processor with instructions.
2. The system of claim 1 , wherein the processor is further configured to receive an image selection criteria comprising an object or pattern within the image.
3. The system of claim 1 , wherein the processor is further configured to receive an image selection criteria comprising an object or pattern within the image, wherein to train the image classifier further includes to receive a set of positive images that correspond to the image selection criteria and a set of negative images that do not correspond to the image selection criteria.
4. The system of claim 1 , wherein the processor is further configured to receive an image selection criteria comprising an object or pattern within the image, wherein to train the image classifier further includes to receive a set of positive images that correspond to the image selection criteria and a set of negative images that do not correspond to the image selection criteria, wherein the set of positive images are extracted from frames of one or more videos containing the object or pattern.
5. The system of claim 1 , wherein the processor is further configured to:
provide a user interface to allow a user to review the save set of images; and
receive an indication from the user that one or more images of the save set of images is incorrectly identified as satisfying an image selection criteria.
6. The system of claim 1 , wherein the processor is further configured to:
provide a user interface to allow a user to review the save set of images;
receive an indication from the user that one or more images of the save set of images is incorrectly identified as satisfying an image selection criteria; and
retrain the image classifier based at least in part on the image selection criteria and the indication.
7. The system of claim 1 , wherein the one or more images of the save set of images are identified with a tag or other metadata associated with an image selection criteria.
8. The system of claim 1 , wherein the stored save set of images are searchable using at least one of an index, a tag, or other metadata.
9. The system of claim 1 , wherein the processor is further configured to:
receive an indication to apply the image classifier to the plurality of storage locations to return a second save set of images; and
store the second save set of images to the second location according to the one or more policies, wherein the second save set of images does not include images from the save set of images.
10. The system of claim 1 , wherein a set of positive images and a set of negative images used to train the image classifier are selected by a user.
11. A method comprising:
training an image classifier to identify an image in a search space, wherein the search space comprises a plurality of storage locations across a plurality of devices associated with a user, wherein the plurality of storage locations store multiple instances of the image;
applying the trained image classifier to a plurality of images stored across the plurality of devices associated with the user to identify one or more images to be included in a save set of images; and
selectively storing the save set of images to a second location according to one or more policies, wherein the save set of images is deduplicated at the second location that is different than the plurality of storage locations, wherein the one or more policies indicate at least one of a frequency that the save set of images is selectively stored to the second location, a number of remote storage locations to store the save set of images, a type of storage that is to be used to store the save set of images, and/or a length of time in which the save set of images are to be stored at the second location.
12. The method of claim 11 , further comprising receiving an image selection criteria comprising an object or pattern within the image.
13. The method of claim 11 , further comprising receiving an image selection criteria comprising an object or pattern within the image, wherein training the image classifier further includes receiving a set of positive images that correspond to the image selection criteria and a set of negative images that do not correspond to the image selection criteria.
14. The method of claim 11 , further comprising receiving an image selection criteria comprising an object or pattern within the image, wherein training the image classifier further includes receiving a set of positive images that correspond to the image selection criteria and a set of negative images that do not correspond to the image selection criteria, wherein the set of positive images are extracted from frames of one or more videos containing the object or pattern.
15. The method of claim 11 , further comprising:
providing a user interface that allows a user to review the save set of images; and
receiving an indication from the user that one or more images of the save set of images are incorrectly identified as satisfying an image selection criteria.
16. The method of claim 11 , further comprising:
providing a user interface that allows a user to review the save set of images;
receiving an indication from the user that one or more images of the save set of images are incorrectly identified as satisfying an image selection criteria; and
retraining the image classifier based at least in part on the image selection criteria and the indication.
17. A computer program product, the computer program product being embodied in a tangible non-transitory computer readable storage medium and comprising computer instructions for:
training an image classifier to identify an image in a search space, wherein the search space comprises a plurality of storage locations across a plurality of devices associated with a user, wherein the plurality of storage locations store multiple instances of the image;
applying the trained image classifier to a plurality of images stored across the plurality of devices associated with the user to identify one or more images to be included in a save set of images; and
selectively storing the save set of images to a second location according to one or more policies, wherein the save set of images is deduplicated at the second location that is different than the plurality of storage locations, wherein the one or more policies indicate at least one of a frequency that the save set of images is selectively stored to the second location, a number of remote storage locations to store the save set of images, a type of storage that is to be used to store the save set of images, and/or a length of time in which the save set of images are to be stored at the second location.