COGNITIVE SNAPSHOTS FOR VISUALLY-IMPAIRED USERS

摘要

Systems, methods, and computer-readable media are described for providing machine-learning assistance to a user who may be visually-impaired in the form of cognitive snapshots taken of an environment. The visually-impaired user can be guided by a user application executing on a user device to position the user device to capture an image of a desired portion of an environment. Object recognition processing is performed on the captured image to identify objects present in the environment. An audible listing of the identified objects can be presented to the user. A delta between objects identified across different snapshots can be determined and an indication thereof can be presented to the user. The user application can guide the user towards a particular desired object by outputting an audible or haptic signal that increases in intensity as the user scans the user device towards a location of the object in the environment.