Method and system to determine the location and/or orientation of a head

1. A method for determining an absolute depth map to monitor the location and pose of a head ( 100 ) being imaged by a camera, the method comprising:

acquiring ( 20 ) an image from the camera ( 110 ) including a head with a facial region;

determining ( 23 ) a distance from the camera ( 110 ) to a base plane, the base plane being a plane located at an average distance from the camera ( 110 ) to the facial region, the average distance being based at least on one distance from the camera ( 110 ) to a facial feature of the facial region using a distance measuring sub-system ( 120 );

determining ( 24 ) a relative depth map of facial features within the facial region with respect to the base plane; and

adding ( 25 ) the distance from the camera to the base plane to the relative depth map to form an absolute depth map for the facial region.

2. The method of claim 1 , wherein the at least one distance from the camera ( 110 ) to a facial feature of the facial region comprises an array of distance measurements for a plurality of locations within the facial region, and

wherein combining the relative depth map with the at least one distance to form an absolute depth map comprises:

determining a head-pose from the array of distance measurements; and

refining the relative depth map using the determined head-pose.

3. The method of claim 1 , wherein the distance measuring sub-system ( 120 ) comprises a stereo camera comprising the camera ( 110 ) and an additional camera offset from the camera ( 110 ) and with a substantially overlapping field of view; and

wherein the distance to at least one feature imaged in the camera ( 110 ) and the additional camera is determined from a disparity in the at least one feature when imaged in the camera ( 110 ) relative to the additional camera.

4. The method of claim 1 , wherein the distance measuring sub-system ( 120 ) comprises a radar system and the at least one distance comprises a point cloud of distance measurements.

5. The method of claim 1 , wherein at least some of the determined facial information is communicated to the distance measuring sub-system ( 120 ), and the method comprises the distance measuring sub-system ( 120 ) adjusting an operational parameter according to the communicated information.

6. The method of claim 5 , wherein the distance measuring sub-system ( 120 ) comprises a laser transceiver and the method comprises using the communication facial information to configure the distance measuring sub-system to direct a laser source into a region of the face being tracked that does not comprise an eye.

7. The method of claim 1 , further comprising:

determining an inter-pupillary distance from said acquired image, and

confirming the at least one distance by comparing the determined inter-pupillary distance to expected inter-pupillary distance values for a head located at the at least one distance.

8. A system for determining an absolute depth map to monitor the location and pose of a head ( 100 ), the system comprising:

a camera ( 110 ) configured to acquire an image including a head with a facial region;

a distance measuring sub-system ( 120 ) for measuring the distance from a point on the head ( 100 ) to the camera ( 110 ); and

a computational module ( 6 ) configured to perform the steps of claim 1 .

9. The system of claim 8 , wherein the computational module ( 7 ) comprises:

a residual neural network ( 70 ) with an encoder-decoder structure with skip connections ( 7024 ) that is configured to produce an initial relative depth map ( 7025 )

in response to receiving an image ( 71 ) from the camera ( 110 ) and the at least one distance; and

a guided filter block ( 703 ) configured to refine the initial relative depth map ( 7025 ) with the at least one distance to form the absolute depth map ( 73 ).

10. The system of claim 9 , wherein the computational module ( 7 ) has been trained by using the distance measuring sub-system ( 120 ) to provide truth labels for depth measurements produced from analysis of the camera ( 110 ).

11. The system of claim 8 , wherein the computational module ( 6 ) is further configured to load a stored depth map to provide an initial estimate of the depth map.

12. The system of claim 11 , wherein the system is configured to gather and process facial images of a particular user to refine the stored depth map.

13. The system of claim 9 , wherein the computational module ( 7 ) comprises a facial expression classifier and the computational module is configured to adjust the initial relative depth map ( 7025 ) to account for the current facial expression.

14. A driver monitoring system for a vehicle comprising the system of claim 8 .