Multimodal authentication and liveliness detection
A robust and reliable multimodal authentication is provided by a multimodal authentication device. The multimodal authentication device utilizes an audio authentication, a video authentication, an audio liveliness authentication, and a video liveliness authentication to determine the authentication of a user. By including a liveliness component in the authentication determination reduces the risk of fraud by factoring in live movement and orientation into the authentication determines. For example, various image/location combinations are displayed to the user and the user is instructed to track and verbally identify the various images. In this way, the user is authenticated based not only on, for example, facial and audio recognition but also a liveliness associated with each.
1. A multimodal authentication device to provide a multimodal authentication, comprising:
a memory storing one or more computer-readable instructions; and
a processor configured to execute the one or more computer-readable instructions to perform one or more operations to:
receive, from a user, a request for access to any of one or more devices, one or more services, one or more resources associated with the multimodal authentication device, or any combination thereof;
display one or more images at one or more locations of a display device associated with the multimodal authentication device;
receive one or more user images of the user in response to the display of the one or more images, wherein the user image is received from an image capture device associated with the multimodal authentication device;
receive one or more user audio inputs in response to the display of the one or more images, wherein the user audio is received from an audio capture device;
determine a visual authentication based on the one or more user images;
determine a visual liveliness authentication based on a tracking of the one or more user images to the display of the one or more images;
determine an audio authentication based on the one or more user audio inputs;
determine an audio liveliness authentication based on a comparison of a first timestamp associated with one or more lip movements of the user from the one or more user images and a second timestamp associated with a start of a command from the one or more user audio inputs; and
provide a multimodal authentication based on the visual authentication, the visual liveliness authentication, the audio authentication, and the audio liveliness authentication so that the user has the requested access.
2. The multimodal authentication device of claim 1 , wherein the processor is further configured to execute the one or more computer-readable instructions to perform one or more further operations to:
generate a location random number series, wherein the one or more locations are based on the location random number series; and
generate an image random number series, wherein the one or more images are based on the image random number series.
3. The multimodal authentication device of claim 1 , wherein the processor is further configured to execute the one or more computer-readable instructions to perform one or more further operations to:
set a visual authentication result based on the visual authentication;
set a visual liveliness authentication result based on the visual liveliness authentication;
set an audio authentication result based on the audio authentication; and
set an audio liveliness authentication result based on the audio liveliness authentication.
4. The multimodal authentication device of claim 3 , wherein the processor is further configured to execute the one or more computer-readable instructions to perform one or more further operations to:
compare the visual authentication result to a visual authentication result threshold;
compare the visual liveliness authentication result to a visual liveliness authentication threshold;
compare the audio authentication result to an audio authentication result threshold; and
compare the audio liveliness authentication result to an audio liveliness authentication result threshold, wherein providing the multimodal authentication is further based on each of the comparisons.
5. The multimodal authentication device of claim 1 , wherein the determining the visual liveliness authentication comprises at least one of:
determining a face angle associated with the one or more user images; and
determining a gaze angle associated with the one or more user images.
6. The multimodal authentication device of claim 5 , wherein the determining the visual liveliness authentication further comprises at least one of:
determining that the face angle tracks the displaying of the one or more images at each of the one or more locations; and
determining that the gaze angle tracks the displaying of the one or more images at each of the one or more locations.
7. The multimodal authentication device of claim 1 , wherein the processor is further configured to execute the one or more computer-readable instructions to perform one or more further operations to:
provide one or more instructions at the display device;
receive a user input in response to the displaying the one or more instructions; and
wherein the displaying the one or more images is based on the user input.
8. A method for a multimodal authentication device to provide a multimodal authentication, the method comprising:
receiving, from a user, a request for access to any of one or more devices, one or more services, one or more resources associated with the multimodal authentication device, or any combination thereof;
displaying one or more images at one or more locations of a display device associated with the multimodal authentication device;
receiving one or more user images of the user in response to the display of the one or more images, wherein the user image is received from an image capture device associated with the multimodal authentication device;
receiving one or more user audio inputs in response to the display of the one or more images, wherein the user audio is received from an audio capture device;
determining a visual authentication based on the one or more user images;
determining a visual liveliness authentication based on a tracking of the one or more user images to the display of the one or more images;
determining an audio authentication based on the one or more user audio inputs;
determining an audio liveliness authentication based on a comparison of a first timestamp associated with one or more lip movements of the user from the one or more user images and a second timestamp associated with a start of a command from the one or more user audio inputs; and
providing a multimodal authentication based on the visual authentication, the visual liveliness authentication, the audio authentication, and the audio liveliness authentication so that the user has the requested access.
9. The method of claim 8 , further comprising:
generating a location random number series, wherein the one or more locations are based on the location random number series; and
generating an image random number series, wherein the one or more images are based on the image random number series.
10. The method of claim 8 , further comprising:
setting a visual authentication result based on the visual authentication;
setting a visual liveliness authentication result based on the visual liveliness authentication;
setting an audio authentication result based on the audio authentication; and
setting an audio liveliness authentication result based on the audio liveliness authentication.
11. The method of claim 10 , wherein providing the multimodal authentication comprises:
comparing the visual authentication result to a visual authentication result threshold;
comparing the visual liveliness authentication result to a visual liveliness authentication threshold;
comparing the audio authentication result to an audio authentication result threshold; and
comparing the audio liveliness authentication result to an audio liveliness authentication result threshold.
12. The method of claim 8 , wherein determining the visual liveliness authentication comprises at least one of:
determining a face angle associated with the one or more user images; and
determining a gaze angle associated with the one or more user images.
13. The method of claim 12 , wherein determining the visual liveliness authentication further comprises at least one of:
determining that the face angle tracks the displaying of the one or more images at each of the one or more locations; and
determining that the gaze angle tracks the displaying of the one or more images at each of the one or more locations.
14. The method of claim 8 , further comprising:
providing one or more instructions at the display device;
receiving a user input in response to the displaying the one or more instructions; and
wherein the displaying the one or more images is based on the user input.
15. A non-transitory computer-readable medium of a multimodal authentication device storing one or more instructions for providing a multimodal authentication, which when executed by a processor of the multimodal authentication device, cause the multimodal authentication device to perform one or more operations comprising:
receive, from a user, a request for access to any of one or more devices, one or more services, one or more resources associated with the multimodal authentication device, or any combination thereof;
displaying one or more images at one or more locations of a display device associated with the multimodal authentication device;
receiving one or more user images of the user in response to the display of the one or more images, wherein the user image is received from an image capture device associated with the multimodal authentication device;
receiving one or more user audio inputs in response to the display of the one or more images, wherein the user audio is received from an audio capture device;
determining a visual authentication based on the one or more user images;
determining a visual liveliness authentication based on a tracking of the one or more user images to the display of the one or more images;
determining an audio authentication based on the one or more user audio inputs;
determining an audio liveliness authentication based on a comparison of a first timestamp associated with one or more lip movements of the user from the one or more user images and a second timestamp associated with a start of a command from the one or more user audio inputs; and
providing a multimodal authentication based on the visual authentication, the visual liveliness authentication, the audio authentication, and the audio liveliness authentication so that the user has the requested access.
16. The non-transitory computer-readable medium of claim 15 , the one or more instructions when executed by the processor further cause the multimodal authentication device to perform one or more further operations comprising:
generating a location random number series, wherein the one or more locations are based on the location random number series; and
generating an image random number series, wherein the one or more images are based on the image random number series.
17. The non-transitory computer-readable medium of claim 15 , the one or more instructions when executed by the processor further cause the multimodal authentication device to perform one or more further operations comprising:
setting a visual authentication result based on the visual authentication;
setting a visual liveliness authentication result based on the visual liveliness authentication;
setting an audio authentication result based on the audio authentication; and
setting an audio liveliness authentication result based on the audio liveliness authentication.
18. The non-transitory computer-readable medium of claim 17 , the one or more instructions when executed by the processor further cause the multimodal authentication device to perform one or more further operations comprising:
comparing the visual authentication result to a visual authentication result threshold;
comparing the visual liveliness authentication result to a visual liveliness authentication threshold;
comparing the audio authentication result to an audio authentication result threshold; and
comparing the audio liveliness authentication result to an audio liveliness authentication result threshold, wherein providing the multimodal authentication is further based on each of the comparisons.
19. The non-transitory computer-readable medium of claim 17 , the one or more instructions when executed by the processor further cause the multimodal authentication device to perform one or more further operations comprising:
providing one or more instructions at the display device;
receiving a user input in response to the displaying the one or more instructions; and
wherein the displaying the one or more images is based on the user input.
20. The non-transitory computer-readable medium of claim 15 , wherein determining the visual liveliness authentication comprises at least one of:
determining a face angle associated with the one or more user images;
determining a gaze angle associated with the one or more user images; and
determining that the face angle and the gaze angle tracks the displaying of the one or more images at each of the one or more locations.