Avatar animation in virtual conferencing

1. A method comprising:

receiving, by a computing device:

a photo of a virtual conference participant; and

a depth map based on the photo;

generating, by the computing device, a plurality of synthesized images based on the photo and the depth map, the plurality of synthesized images having respective simulated gaze directions of the virtual conference participant;

receiving, by the computing device during a virtual conference, an indication of a current gaze direction of the virtual conference participant, the indication of the current gaze direction includes changes in the current gaze direction;

filtering the indication of the current gaze direction using a low-pass filter; and

animating, in a display of the virtual conference, an avatar corresponding with the virtual conference participant,

the avatar being based on the photo, and

the animating the avatar being based on:

the photo, the depth map and at least one synthesized image of the plurality of synthesized images, the at least one synthesized image corresponding with the current gaze direction; and

the filtered indication of the current gaze direction.

2. The method of claim 1 , further comprising defining a 3-dimensional (3D) mesh from the depth map, wherein animating the avatar includes applying at least a portion of the photo and at least a portion of the at least one synthesized image as a texture to the 3D mesh.

3. The method of claim 2 , wherein the 3D mesh includes a mesh of geometric shapes representing a 3D model of the photo.

4. The method of claim 2 , wherein animating the avatar further includes performing at least one geometric transformation on the 3D mesh based on the current gaze direction.

5. The method of claim 4 , wherein the at least one geometric transformation includes one or more of:

rotation of the 3D mesh;

scaling of the 3D mesh;

translation of the 3D mesh; or

skewing of the 3D mesh.

6. The method of claim 1 , wherein the depth map includes information indicating respective distances of surfaces of the photo from a fixed viewpoint.

7. The method of claim 1 , wherein the plurality of synthesized images is a plurality of images generated from the photo using a machine learning model.

8. The method of claim 1 , wherein animating the avatar includes blending a first synthesized image of the plurality of synthesized images with a second synthesized image of the plurality of synthesized images.

9. The method of claim 8 , wherein the current gaze direction is between a simulated gaze direction of the first synthesized image and a simulated gaze direction of the second synthesized image.

10. The method of claim 8 , wherein blending the first synthesized image of the plurality of synthesized images with the second synthesized image of the plurality of synthesized images includes alpha blending.

11. The method of claim 1 , wherein the respective simulated gaze directions are at respective equally spaced angles over 360 degrees.

12. The method of claim 1 , wherein animating the avatar includes animating a periocular region of the virtual conference participant based on a segmentation mask.

13. The method of claim 1 , wherein the plurality of synthesized images is a first plurality of synthesized images, the method further including:

generating, by the computing device, a second plurality of synthesized images based on the photo and the depth map, the second plurality of synthesized images having respective simulated mouth shapes of the virtual conference participant; and

receiving, by the computing device during the virtual conference, an audio signal indicating speech of the virtual conference participant,

wherein animating the avatar is further based on the audio signal and at least one synthesized image of the second plurality of synthesized images corresponding with the audio signal.

14. The method of claim 13 , further comprising:

selecting one or more synthesized images of the second plurality of synthesized images based on a pitch of the audio signal and an amplitude of the audio signal; and

animating, in the avatar, a mouth region of the virtual conference participant based on the selected one or more synthesized images of the second plurality of synthesized images.

15. The method of claim 13 , wherein animating the avatar includes blending a first synthesized image of the second plurality of synthesized images with a second synthesized image of the second plurality of synthesized images.

16. The method of claim 1 , wherein animating the avatar includes animating a mouth region of the virtual conference participant based on a segmentation mask.

17. A computing device, comprising:

at least one processor; and

a non-transitory computer-readable medium storing executable instructions that, when executed by the at least one processor, cause the computing device to:

receive:

a photo of a virtual conference participant; and

a depth map based on the photo;

generate a plurality of synthesized images based on the photo, the plurality of synthesized images having respective simulated gaze directions of the virtual conference participant;

receive, during a virtual conference, an indication of a current gaze direction of the virtual conference participant, the indication of the current gaze direction includes changes in the current gaze direction;

filtering the indication of the current gaze direction using a low-pass filter; and

animate, in a display of the virtual conference, an avatar corresponding with the virtual conference participant,

the avatar being based on the photo, and

the animating the avatar being based on:

the photo, the depth map and at least one synthesized image of the plurality of synthesized images, the at least one synthesized image corresponding with the current gaze direction; and

the filtered indication of the current gaze direction.

18. The computing device of claim 17 , wherein the executable instructions include instructions that, when executed by the at least one processor, cause the computing device to:

define a 3-dimensional (3D) mesh from the depth map, wherein animating the avatar includes applying at least a portion of the photo and at least a portion of the at least one synthesized image as a texture to the 3D mesh.

19. The computing device of claim 18 , wherein animating the avatar further includes performing at least one geometric transformation of the 3D mesh based on the current gaze direction.

20. The computing device of claim 17 , wherein the plurality of synthesized images is a first plurality of synthesized images, and the executable instructions include instructions that, when executed by the at least one processor, cause the computing device to:

generate a second plurality of synthesized images based on the photo, the second plurality of synthesized images having respective simulated mouth shapes of the virtual conference participant; and

receive, during the virtual conference, an audio signal indicating speech of the virtual conference participant,

wherein animating the avatar is further based on the audio signal and at least one synthesized image of the second plurality of synthesized images corresponding with the audio signal.

21. The computing device of claim 20 , wherein the executable instructions include instructions that, when executed by the at least one processor, cause the computing device to:

select one or more synthesized images of the second plurality of synthesized images based on a pitch of the audio signal and an amplitude of the audio signal; and

animate a mouth region of avatar of the virtual conference participant based on the selected one or more synthesized images of the second plurality of synthesized images.

22. A non-transitory computer-readable medium storing executable instructions that, when executed by at least one processor, cause a computing device to:

receive:

a photo of a virtual conference participant; and

a depth map based on the photo;

generate a plurality of synthesized images based on the photo, the plurality of synthesized images having respective simulated gaze directions of the virtual conference participant;

receive, during a virtual conference, an indication of a current gaze direction of the virtual conference participant, the indication of the current gaze direction includes changes in the current gaze direction;

filtering the indication of the current gaze direction using a low-pass filter; and

animate, in a display of the virtual conference, an avatar corresponding with the virtual conference participant,

the avatar being based on the photo, and

the animating the avatar being based on:

the photo, the depth map and at least one synthesized image of the plurality of synthesized images, the at least one synthesized image corresponding with the current gaze direction; and

the filtered indication of the current gaze direction.

23. The non-transitory computer-readable medium of claim 22 , wherein the plurality of synthesized images is a first plurality of synthesized images, and the executable instructions include instructions that, when executed by the at least one processor, cause the computing device to:

generating a second plurality of synthesized images based on the photo, the second plurality of synthesized images having respective simulated mouth shapes of the virtual conference participant; and

receive, during the virtual conference, an audio signal indicating speech of the virtual conference participant,

wherein animating the avatar is further based on the audio signal and at least one synthesized image of the second plurality of synthesized images corresponding with the audio signal.

24. The computing device of claim 17 , wherein the respective simulated gaze directions are at respective equally spaced angles over 360 degrees.

25. The non-transitory computer-readable medium of claim 22 , wherein the respective simulated gaze directions are at respective equally spaced angles over 360 degrees.