Systems and methods for providing image-based security measures
AssigneeFacebook, Inc.
InventorsNikhil Johri, Trevor M. Pottinger, Balamanohar Paluri
Life of this patent
Estimating the term…
Abstract
Systems, methods, and non-transitory computer-readable media can detect an operation that causes a challenge response process to be initiated. An image category associated with a recognized category label can be identified. At least one image associated with the image category can be displayed during the challenge response process. The operation can be executed when the challenge response process, based on the at least one image, is successfully completed.
Claims20
A computer-implemented method comprising:
- determining, by a computing system, at least one image that contains an image object associated with an image category based at least in part on a trained image classifier;
- providing, by the computing system, the at least one image during a challenge response process, wherein the challenge response process includes a user input to indicate whether the at least one image is associated with a recognized category label and where in the at least one image the image object associated with the image category is located;
- receiving, by the computing system, the user input to indicate whether the at least one image is associated with the recognized category label and where in the at least one image the image object associated with the image category is located;
- and
- further training, by the computing system, the trained image classifier based at least in part on the user input.
A system comprising:
- at least one processor;
- and
- a memory storing instructions that, when executed by the at least one processor, cause the system to perform:
- determining at least one image that contains an image object associated with an image category based at least in part on a trained image classifier;
- providing the at least one image during a challenge response process, wherein the challenge response process includes a user input to indicate whether the at least one image is associated with a recognized category label and where in the at least one image the image object associated with the image category is located;
- receiving the user input to indicate whether the at least one image is associated with the recognized category label and where in the at least one image the image object associated with the image category is located;
- and
- further training the trained image classifier based at least in part on the user input.
A non-transitory computer-readable storage medium including instructions that, when executed by at least one processor of a computing system, cause the computing system to perform:
- determining at least one image that contains an image object associated with an image category based at least in part on a trained image classifier;
- providing the at least one image during a challenge response process, wherein the challenge response process includes a user input to indicate whether the at least one image is associated with a recognized category label and where in the at least one image the image object associated with the image category is located;
- receiving the user input to indicate whether the at least one image is associated with the recognized category label and where in the at least one image the image object associated with the image category is located;
- and
- further training the trained image classifier based at least in part on the user input.
Dependencies are inferred from the claim text.