Verification based on input comprising captured images, captured audio and tracked eye movement

摘要

In one embodiment, a method comprises receiving a verification request, sending one or more patterns for display at one or more specified locations on a screen of a device, receiving verification input responsive to the one or more patterns, and granting or denying the verification request based at least in part on the verification input. The verification input comprises one or more captured images, captured audio and tracked eye movement. The verification request may be generated in conjunction with an attempted launch of a designated application on a first device, with user access to the designated application on the first device being controlled responsive to the granting or denying of the verification request.

主权项

1. A method comprising:
receiving a verification request; responsive to receiving the verification request, sending one or more randomly generated patterns for display at one or more randomly specified locations on a screen of a device; responsive to the sending, receiving verification input responsive to the displayed one or more randomly generated patterns, the verification input comprising one or more captured images, captured audio and tracked eye movement, wherein the one or more captured images, the captured audio and the tracked eye movement are concurrently collected from a user of the device; and granting or denying the verification request based at least in part on the verification input; wherein granting the verification request comprises verifying liveliness of the user by:
verifying, utilizing the captured audio, that a stored voice profile for the user matches user spoken phrases or digits corresponding to the one or more randomly generated patterns;verifying, utilizing the tracked eye movement, that one or more locations of the screen the user was viewing while speaking the phrases or digits corresponding to the one or more randomly generated patterns match the one or more randomly specified locations; andverifying, utilizing the one or more captured images, that a stored facial profile for the user matches a facial profile captured while the user was speaking the phrases or digits corresponding to the one or more randomly generated patterns.