Methods and apparatus for orientation keypoints for complete 3D human pose computerized estimation

1. An apparatus, comprising:

a processor; and

a memory storing instruction which, when executed by the processor, causes the processor to:

receive an image depicting at least one subject;

predict at least one orientation keypoint associated with a section of the body part of the at least one subject via a neural network detector; and

compute a three-axis joint rotation associated with the section of the body part of the at least one subject based on at least one orientation keypoint associated with the body part of the at least one subject and at least one joint keypoint associated with the body part of the at least one subject.

2. The apparatus of claim 1 , wherein the memory storing instructions which, when executed by the processor, further causes the processor to:

train a neural network based on an orientation key point from the at least one orientation keypoint.

3. The apparatus of claim 1 , wherein the memory storing instructions which, when executed by the processor, further causes the processor to:

produce training data for a neural network to estimate a subject pose.

4. The apparatus of claim 1 , wherein the memory storing instructions which, when executed by the processor, further causes the processor to:

predict at least one subject pose in a three-dimensional space base on an estimated two-dimensional position associated with the at least one orientation keypoint.

5. The apparatus of claim 1 , wherein the memory storing instructions which, when executed by the processor, further causes the processor to:

compute a change in position of the at least one orientation keypoint over time; and

calculate a rotational velocity or acceleration associated with the at least one orientation keypoint based on the change in position of the at least one orientation keypoint.

6. The apparatus of claim 1 , wherein the at least one orientation keypoint is predicted in at least one of a two-dimensional space and a three-dimensional space.

7. The apparatus of claim 1 , wherein the image is produced by at least one of a camera and a video camera.

8. The apparatus of claim 1 , wherein the image comprises depth information.

9. The apparatus of claim 1 , wherein the at least one orientation keypoint is located outside the section of the body part of the at least one subject and is indicative of a kinematic rotation.

10. The apparatus of claim 9 , where in the kinematic rotation comprises at least one of a forward direction from the center of the section of the body part, an outward direction from the center of the section of the body part, an inward direction from the center of the section of the body part, a backward direction from the center of the section of the body part, and a lower direction from the center of the section of the body part.

11. The apparatus of claim 1 , wherein the three-axis joint rotation associated with the section of the body part of the at least one subject is determined via a at least one of a perspective-n-point computational technique, a Procrustes computational technique, a Kabsch computational technique, and a regression technique.

12. An apparatus, comprising:

a processor; and

a memory storing instruction which, when executed by the processor, causes the processor to:

receive an image depicting at least one subject;

predict at least one orientation keypoint associated with a section of a body part of the at least one subject via the neural network detector; and

predict an aspect of a pose associated with the at least one subject based on the at least one orientation keypoint, the aspect of the pose associated with the at least one subject comprises at least one of a position associated with the at least one subject, size associated with at least one subject, and a movement associated with the at least one subject.

13. The apparatus of claim 12 , wherein the pose of the at least one subject comprises at least one joint position and at least one joint angle associated with the section of the body part of the at least one subject.

14. The apparatus of claim 12 , wherein the at least one orientation keypoint is predicted in at least one of a two-dimensional space and a three-dimensional space.

15. The apparatus of claim 12 , wherein at least one orientation keypoint is used for training or supervising a network or model.

16. The apparatus of claim 12 , wherein the memory storing instructions which, when executed by the processor, further causes the processor to:

train a neural network based on an orientation key point from the at least one orientation keypoint.

17. The apparatus of claim 12 , wherein the memory storing instructions which, when executed by the processor, further causes the processor to:

predict the at least one subject pose in a three-dimensional space base on an estimated two-dimensional position associated with the at least one orientation keypoint.

18. The apparatus of claim 12 , wherein the memory storing instructions which, when executed by the processor, further causes the processor to:

compute a change in position of the at least one orientation keypoint over time; and

compute at least one of a rotational velocity associated with the at least one orientation keypoint and acceleration associated with the at least one orientation keypoint based on the change in position of the at least one orientation keypoint.

19. A method, comprising:

receiving an image depicting at least one subject;

predicting at least one orientation keypoint associated with a section of the body part of the at least one subject via a neural network detector; and

computing a three-axis joint rotation associated with the section of the body part of the at least one subject based on at least one orientation keypoint associated with the body part of the at least one subject and at least one joint keypoint associated with the body part of the at least one subject.

20. The method of claim 19 , wherein the method further comprises:

rendering a graphical representation on a display comprising a plurality of marks indicative of the three-axis joint rotation associated with the section of the body part of the at least one subject.