Signaling camera parameters in point cloud coding

1. A method implemented by a decoder, the method comprising:

receiving a bitstream comprising a plurality of patches in an atlas frame and a camera offset for a camera;

decoding the patches in the atlas frame;

converting the patches to a three dimensional (3D) patch coordinate system to obtain point cloud content;

determining an offset matrix based on the camera offset, wherein the camera offset indicates a first offset corresponding to an X axis, a second offset corresponding to a Y axis, and a third offset corresponding to a Z axis, wherein the offset matrix comprises four rows and four columns, wherein the first offset is positioned at row one column four, wherein the second offset is positioned at row two column four, and wherein the third offset is positioned at row three column four; and

applying the offset matrix to the point cloud content.

2. The method of claim 1 , wherein the bitstream further comprises a camera rotation for the camera, and wherein the method further comprises:

determining a rotation matrix based on the camera rotation; and

applying the rotation matrix to the point cloud content.

3. The method of claim 2 , wherein the camera rotation specifies an X component for a rotation of the camera using quaternion representation, a Y component for the rotation of the camera using quaternion representation, and a Z component for the rotation of the camera using quaternion representation.

4. The method of claim 3 , wherein the rotation matrix includes the X component, the Y component, the Z component, and a W component for the rotation of the camera using quaternion representation.

5. The method of claim 4 , wherein the W component is calculated according to: qW=Sqrt(1−qX 2 +qY 2 +qZ 2 )), where qW, qX, qY, and qZ represents the W component, the X component, the Y component, and the Z component, respectively, and where Sqrt is a square root function.

6. The method of claim 1 , wherein the patches in the atlas frame are two dimensional (2D) patches.

7. A method implemented by an encoder, the method comprising:

encoding a plurality of two dimensional (2D) patches and a geometry describing a point cloud frame into a bitstream;

determining a camera offset describing a position of a camera relative to the point cloud frame to support generation of an offset matrix for rendering a camera view of the point cloud frame;

encoding the camera offset into the bitstream, wherein the camera offset indicates a first offset corresponding to an X axis, a second offset corresponding to a Y axis, and a third offset corresponding to a Z axis, wherein the offset matrix comprises four rows and four columns, wherein the first offset is positioned at row one column four, wherein the second offset is positioned at row two column four, and wherein the third offset is positioned at row three column four; and

storing the bitstream for communication toward a decoder.

8. The method of claim 7 , further comprising:

determining a camera rotation describing a viewing vector for the camera relative to the point cloud frame to support generation of a rotation matrix for rendering the camera view of the point cloud frame; and

encoding the camera rotation into the bitstream.

9. The method of claim 8 , wherein the camera rotation specifies an X component for a rotation of the camera using quaternion representation, a Y component for the rotation of the camera using quaternion representation, and a Z component for the rotation of the camera using quaternion representation.

10. The method of claim 9 , wherein the rotation matrix includes the X component, the Y component, the Z component, and a W component for the rotation of the camera using quaternion representation.

11. The method of claim 10 , wherein the W component is calculated according to: qW=Sqrt(1−qX 2 +qY 2 +qZ 2 )), where qW, qX, qY, and qZ represents the W component, the X component, the Y component, and the Z component, respectively, and where Sqrt is a square root function.

12. A video coding device comprising:

a receiver configured to obtain a bitstream comprising a plurality of patches in an atlas frame and a camera offset for a camera; and

a processor configured to:

decode the patches in the atlas frame;

convert the patches to a three dimensional (3D) patch coordinate system to obtain point cloud content;

determine an offset matrix based on the camera offset, wherein the camera offset indicates a first offset corresponding to an X axis, a second offset corresponding to a Y axis, and a third offset corresponding to a Z axis, wherein the offset matrix comprises four rows and four columns, wherein the first offset is positioned at row one column four, wherein the second offset is positioned at row two column four, and wherein the third offset is positioned at row three column four; and

apply the offset matrix to the point cloud content.

13. The video coding device of claim 12 , wherein the bitstream further comprises a camera rotation for the camera, and wherein the processor is further configured to:

determine a rotation matrix based on the camera rotation; and

apply the rotation matrix to the point cloud content.

14. The video coding device of claim 13 , wherein the camera rotation specifies an X component for a rotation of the camera using quaternion representation, a Y component for the rotation of the camera using quaternion representation, and a Z component for the rotation of the camera using quaternion representation.