Calibration of inertial measurement units in alignment with a skeleton model to control a computer system based on determination of orientation of an inertial measurement unit from an image of a portion of a user
A method to calibrate orientation measurements of an inertial measurement unit of a sensor device based on an image of a portion of a user to which the sensor device is attached. For example, the sensor device can be configured to be attached to the middle phalange of the index finger and configured with a touch pad. In response to the determination that the thumb of the user is placed on the touch pad of the sensor device, the camera of the system can capture the image showing that the hand of the user. A convolutional neural network is configured to determine, from the image, orientations of predefined features of the hand of the user. A further artificial neural network is configured to determine the orientation of the sensor device based on the orientations of the predefined features to calibrate the orientation measurements of the inertial measurement unit.
1. A method, comprising:
receiving an image showing a portion of the user, wherein a sensor device is attached on the portion of the user, the sensor device including an inertial measurement unit configured to measure an orientation;
determining orientations of predefined features of the portion of the user based on the image;
determining, using a first artificial neural network (ANN), the orientation of the sensor device based on the orientations of the predefined features; and
calibrating orientation measurements generated by the inertial measurement unit relative to a skeleton model of the user based on the orientation of the sensor device determined using the first artificial neural network.
2. The method of claim 1 , wherein the orientations of the predefined features are determined from the image using a second artificial neural network (ANN).
3. The method of claim 2 , wherein the second ANN is a convolutional neural network.
4. The method of claim 3 , wherein the portion of the user includes a hand of the user; and the predefined features include vectors aligned with bones in the hand of the user.
5. The method of claim 4 , further comprising:
capturing the image using a camera in a head mounted display, in response to a determination that a thumb on the hand is on a touch pad on the sensor device worn on a finger on the hand.
6. The method of claim 1 , wherein the orientations of the predefined features are calculated relative to a reference system of the skeleton model of the user.
7. A system, comprising:
one or more processors; and
a non-transitory computer-readable medium including one or more sequences of instructions that, when executed by the one or more processors, cause:
receiving an image showing a portion of the user, wherein a sensor device is attached on the portion of the user, the sensor device including an inertial measurement unit configured to measure an orientation;
determining orientations of predefined features of the portion of the user based on the image;
determining, using a first artificial neural network (ANN), the orientation of the sensor device based on the orientations of the predefined features; and
calibrating orientation measurements generated by the inertial measurement unit relative to a skeleton model of the user based on the orientation of the sensor device determined using the first artificial neural network.
8. The system of claim 7 , wherein the orientations of the set of features are determined from the image using a second artificial neural network (ANN).
9. The system of claim 8 , wherein the second ANN is a convolutional neural network.
10. The system of claim 9 , wherein the portion of the user includes a hand of the user; and the predefined features include vectors aligned with bones in the hand of the user.
11. The system of claim 10 , further comprising:
capturing the image using a camera in a head mounted display, in response to a determination that a thumb on the hand is on a touch pad on the sensor device worn on a finger on the hand.
12. The system of claim 7 , wherein the orientations of the set of features are calculated relative to a reference system of the skeleton model of the user.
13. A non-transitory computer storage medium storing instructions which, when executed by a computing device, cause the computing device to perform a method, the method comprising:
receiving an image showing a portion of the user, wherein a sensor device is attached on the portion of the user, the sensor device including an inertial measurement unit configured to measure an orientation;
determining orientations of predefined features of the portion of the user based on the image;
determining, using a first artificial neural network (ANN), the orientation of the sensor device based on the orientations of the predefined features; and
calibrating orientation measurements generated by the inertial measurement unit relative to a skeleton model of the user based on the orientation of the sensor device determined using the first artificial neural network.
14. The non-transitory computer storage medium of claim 13 , wherein the orientations of the set of features are determined from the image using a second artificial neural network (ANN).
15. The non-transitory computer storage medium of claim 14 , wherein the second ANN is a convolutional neural network.
16. The non-transitory computer storage medium of claim 15 , wherein the portion of the user includes a hand of the user; and the predefined features include vectors aligned with bones in the hand of the user.
17. The non-transitory computer storage medium of claim 16 , further comprising:
capturing the image using a camera in a head mounted display, in response to a determination that a thumb on the hand is on a touch pad on the sensor device worn on a finger on the hand.