Encoding of patch temporal alignment for mesh compression

1. A method for mesh encoding, comprising:

receiving at least a first mesh frame associated with a first time instance and a second mesh frame associated with a second time instance, the first mesh frame and the second mesh frame representing a surface of an object with polygons, the first mesh frame comprising a first two dimension (2D) map with first patches having three-dimension (3D) information mapped to 2D, the second mesh frame comprising a second 2D map with second patches having 3D information mapped to 2D;

identifying that a first patch in the first patches is a reference matching patch for a second patch in the second patches;

determining first re-map transform parameters for a temporal alignment of the second patch to the first patch in the first 2D map; and

generating a new second 2D map corresponding to the second 2D map, the new second 2D map comprising a transformed second patch that is transformed from the second patch according to the first re-map transform parameters.

2. The method of claim 1 , wherein the identifying that the first patch is the reference matching patch for the second patch further comprises:

determining first matching metrics of the second patch respectively to the first patches; and

selecting the first patch as a best matching patch from the first patches according to the first matching metrics.

3. The method of claim 2 , further comprising:

determining second matching metrics of the first patch respectively to the second patches; and

determining that the first patch is the reference matching patch for the second patch in response to the second patch being a best matching patch according to the second matching metrics.

4. The method of claim 2 , wherein the determining the first matching metrics of the second patch respectively to the first patches further comprises at least one of:

determining a matching metric of the second patch to a specific patch in the first patches according to matched feature points of the second patch and the specific patch;

determining a matching metric of the second patch to a specific patch in the first patches according to a difference between a center of 3D coordinates of the second patch and a center of 3D coordinates of the specific patch; and

determining a matching metric of the second patch to a specific patch in the first patches according to a Hausdorff distance of 3D coordinates between the second patch and the specific patch.

5. The method of claim 1 , wherein the determining the first re-map transform parameters for the temporal alignment of the second patch to the first patch further comprises:

determining at least one of a rotation parameter, a reflection parameter and a spatial translation parameter.

6. The method of claim 1 , further comprising:

encoding the first re-map transform parameters for the second patch into a bitstream that carries the at least the first mesh frame and the second mesh frame.

7. The method of claim 1 , further comprising:

updating a geometry 2D map of the second mesh frame according to the first re-map transform parameters.

8. The method of claim 1 , further comprising:

classifying the second patches into a first class and a second class, the first class comprising a first subset of the second patches with corresponding reference matching patches identified in the first patches, the second class comprising a second subset of the second patches with no reference matching patch identified in the first patches.

9. The method of claim 8 , further comprising:

ordering the first subset of the second patches by size into a first patch sequence in a size order; and

determining, respective re-map transform parameters for the first subset of the second patches according to the first patch sequence.

10. The method of claim 9 , wherein the determining the respective re-map transform parameters for patches in the first patch sequence further comprises:

selecting the second patch from the first patch sequence according to the size order, the first patch sequence including at least a third patch larger than the second patch by size, second re-amp transform parameters for the third patch being determined, transformed third patch according to the second re-amp transform parameters being added into a packed patch list;

determining a first rotation angle for generating a rotated second patch with a highest similarity to the first patch; and

determining a first spatial translation value for moving the rotated second patch to avoid overlapping with transformed patches in the packed patch list.

11. The method of claim 10 , further comprising:

determining a transformed second patch of the second patch according to the first rotation angle and the first spatial translation value; and

adding the transformed second patch in the packed patch list.

12. The method of claim 10 , further comprising:

determining whether to reflect the second patch based a similarity comparison of the highest similarity with a second highest similarity of a reflected second patch and the first patch.

13. The method of claim 10 , further comprising:

determining to recategorize the second patch into the second class in response to no spatial translation value in a predefined window allowing overlapping free.

14. The method of claim 9 , further comprising:

ordering the second subset of the second patches by size into a second patch sequence in the size order; and

determining, respective re-map transform parameters for the second subset of the second patches according to the second patch sequence.

15. The method of claim 9 , further comprising determining a size of a patch according to at least one of:

a number of pixels in the patch including a boundary of the patch;

a number of pixels in the patch excluding a boundary of the patch; and

an area of a smallest rectangular bounding box that comprises the patch.

16. An apparatus for mesh processing, comprising processing circuitry configured to:

receive at least a first mesh frame associated with a first time instance and a second mesh frame associated with a second time instance, the first mesh frame and the second mesh frame representing a surface of an object with polygons, the first mesh frame comprising a first two dimension (2D) map with first patches having three dimensional (3D) information mapped to 2D, the second mesh frame comprising a second 2D map with second patches having 3D information mapped to 2D;

identify that a first patch in the first patches is a reference matching patch for a second patch in the second patches;

determine first re-map transform parameters that align the second patch to the first patch in the first 2D map; and

generate a new second 2D map with a transformed second patch that is transformed from the second patch according to the first re-map transform parameters.

17. The apparatus of claim 16 , wherein the processing circuitry is configured to:

determine first matching metrics of the second patch respectively to the first patches;

select the first patch as a best matching patch from the first patches according to the first matching metrics;

determine second matching metrics of the first patch respectively to the second patches; and

determine that the first patch is the reference matching patch for the second patch in response to the second patch being a best matching patch according to the second matching metrics.

18. The apparatus of claim 16 , wherein the processing circuitry is configured to:

classify the second patches into a first class and a second class, the first class comprising a first subset of the second patches with corresponding reference matching patches identified in the first patches, the second class comprising a second subset of the second patches with no reference matching patch identified in the first patches.

19. The apparatus of claim 18 , wherein the processing circuitry is configured to:

order the first subset of the second patches by size into a first patch sequence in a size order; and

determine, respective re-map transform parameters for the first subset of the second patches according to the first patch sequence.

20. The apparatus of claim 19 , wherein the processing circuitry is configured to:

select the second patch from the first patch sequence according to the size order, the first patch sequence including at least a third patch larger than the second patch by size, second re-amp transform parameters for the third patch being determined, transformed third patch according to the second re-amp transform parameters being added into a packed patch list;

determine a first rotation angle for generating a rotated second patch with a highest similarity to the first patch; and

determine a first spatial translation value for moving the rotated second patch to avoid overlapping with transformed patches in the packed patch list.