Method and apparatus for video coding

1. A method of processing point cloud data at a decoder, comprising:

receiving three dimensional (3D) coordinates of a set of points of a point cloud including first points and a current point, each of the first points being associated with a reconstructed attribute value;

determining a plurality of neighboring points of the current point from the first points;

receiving an indicator and determining whether the indicator indicating that reconstruction of the current point in a prediction step in a lifting scheme is based on a weighted sum of all of the reconstructed attribute values of the plurality of neighboring points or based on a single one of the reconstructed attribute values of the plurality of neighboring points;

in response to the received indicator being determined as indicating that the reconstruction of the current point is based on the weighted sum of all of the reconstructed attribute values of the plurality of neighboring points, reconstructing the current point in the prediction step in the lifting scheme based on the weighted sum of all of the reconstructed attribute values of the plurality of neighboring points; and

in response to the received indicator being determined as indicating that the reconstruction of the current point is based on the single one of the reconstructed attribute values of the plurality of neighboring points, reconstructing the current point in the prediction step in the lifting scheme based on a reconstructed attribute value of a selected one of the plurality of neighboring points.

2. The method of claim 1 wherein

the plurality of neighboring points of the current point is a subset of the first points that are nearest to the current point in terms of geometric distances.

3. The method of claim 1 , wherein

the indicator is generated based on a determination that a measurement of variability of the reconstructed attribute values of the plurality of neighboring points is below a threshold.

4. The method of claim 3 , wherein

the variability of the reconstructed attribute values of the plurality of neighboring points is a maximum difference of the reconstructed attribute values among the plurality of neighboring points.

5. The method of claim 1 , further comprising:

when the indicator is determined as indicating that the reconstruction of the current point in the prediction step in the lifting scheme is based on the single one of the reconstructed attribute values of the plurality of neighboring points, receiving an index indicating the reconstructed attribute value of the selected one of the plurality of neighboring points, the index being determined based on analyzing the reconstructed attribute values of the plurality of neighboring points according to a rate-distortion decision function; and

determining the reconstructed attribute value indicated by the index to be an attribute prediction of the current point in the prediction step in the lifting scheme.

6. The method of claim 1 , wherein the weighted sum of all of the reconstructed attribute values of the neighboring points is calculated based on geometric weights of the reconstructed attribute values of the neighboring points.

7. The method of claim 3 wherein the determination that the measurement of variability of the reconstructed attribute values of the plurality of neighboring points is below the threshold indicates that a neighboring region of the current point is homogeneous.

8. A method of processing point cloud data at a decoder, comprising:

receiving three dimensional (3D) coordinates of a set of points of a point cloud including first points and a current point, each of the first points being associated with a reconstructed attribute value;

determining a plurality of neighboring points of the current point from the first points;

calculating variability of the reconstructed attribute values of the plurality of neighboring points;

in response to the calculated variability of the reconstructed attribute values of the plurality of neighboring points being below a threshold, determining an attribute value predictor for the current point used in a prediction step in a lifting scheme based on a weighted sum of all of the reconstructed attribute values of the plurality of neighboring points; and

in response to the calculated variability of the reconstructed attribute values of the plurality of neighboring points being greater than or equal to the threshold, determining the attribute value predictor for the current point used in the prediction step in the lifting scheme based on a single one of the reconstructed attribute values of the plurality of neighboring points that is a reconstructed attribute value of a selected one of the plurality of neighboring points, wherein

the prediction step is performed before an updating step in the lifting scheme.

9. The method of claim 8 , further comprising:

in response to the calculated variability of the reconstructed attribute values of the plurality of neighboring points being greater than or equal to the threshold, receiving an index indicating the reconstructed attribute value of the selected one of the plurality of neighboring points, the index being determined based on analyzing the reconstructed attribute values of the plurality of neighboring points according to a rate-distortion decision function; and

determining the reconstructed attribute value indicated by the index to be the attribute value predictor of the current point in the lifting scheme.

10. The method of claim 8 , wherein the variability of the reconstructed attribute values of the plurality of neighboring points is a maximum difference of the reconstructed attribute values among the plurality of neighboring points.

11. The method of claim 8 , wherein the weighted sum of all of the reconstructed attribute values of the neighboring points is calculated based on geometric weights of the reconstructed attribute values of the neighboring points.

12. The method of claim 1 , wherein

the indicator is determined based on a determination of whether a cost of the weighted sum of all of the reconstructed attribute values of the plurality of neighboring points is lower than a lowest cost among individual costs of the reconstructed attribute values of the plurality of neighboring points, and

the cost of the weighted average of the reconstructed attribute values of the plurality of neighboring points and the individual costs of the reconstructed attribute values of the plurality of neighboring points are determined based on a rate-distortion decision function.

13. The method of claim 12 , wherein

the indicator indicates a reconstructed attribute value having the lowest cost among the individual costs of the reconstructed attribute values of the plurality of neighboring points as the reconstructed attribute value of the selected one of the plurality of neighboring points when the cost of the weighted average of the reconstructed attribute values of the plurality of neighboring points is higher or equal to the lowest cost among the individual costs of the reconstructed attribute values of the plurality of neighboring points.

14. A method of processing point cloud data at a decoder, comprising:

receiving three dimensional (3D) coordinates of a set of points of a point cloud including first points and a current point, each of the first points being associated with a reconstructed attribute value;

determining a plurality of neighboring points of the current point from the first points;

receiving an indicator and determining whether the indicator indicating that reconstruction of the current point in a prediction step in a lifting scheme is based on a weighted sum of all of the reconstructed attribute values of the plurality of neighboring points or based on a single one of the reconstructed attribute values of the plurality of neighboring points;

in response to the received indicator being determined as indicating that the reconstruction of the current point is based on the weighted sum of all of the reconstructed attribute values of the plurality of neighboring points, reconstructing the current point in the prediction step in the lifting scheme based on the weighted sum of all of the reconstructed attribute values of the plurality of neighboring points; and

in response to the received indicator being determined as indicating that the reconstruction of the current point is based on the single one of the reconstructed attribute values of the plurality of neighboring points, reconstructing the current point in the prediction step in the lifting scheme based on a reconstructed attribute value having the lowest cost among the reconstructed attribute values of the plurality of neighboring points.

15. The method of claim 14 , wherein

the prediction step is performed before an updating step in the lifting scheme.

16. The method of claim 14 , wherein

the plurality of neighboring points of the current point is a subset of the first points that are nearest to the current point in terms of geometric distances.

17. The method of claim 14 , wherein the weighted sum of all of the reconstructed attribute values of the neighboring points is calculated based on geometric weights of the reconstructed attribute values of the neighboring points.

18. The method of claim 14 , wherein

the indicator is determined based on a determination of whether a cost of the weighted sum of all of the reconstructed attribute values of the plurality of neighboring points is lower than a lowest cost among individual costs of the reconstructed attribute values of the plurality of neighboring points, and

when the cost of the weighted sum of all of the reconstructed attribute values of the plurality of neighboring points is determined as lower than the lowest cost among the individual costs of the reconstructed attribute values of the plurality of neighboring points, the indicator is determined as indicating that the current point is reconstructed based on the weighted sum of all of the reconstructed attribute values of the plurality of neighboring points.

19. An apparatus for processing point cloud data, the apparatus comprising:

processing circuitry configured to

receive three dimensional (3D) coordinates of a set of points of a point cloud including first points and a current point, each of the first points being associated with a reconstructed attribute value;

determine a plurality of neighboring points of the current point from the first points;

receive an indicator and determine whether the indicator indicating that reconstruction of the current point in a prediction step in a lifting scheme is based on a weighted sum of all of the reconstructed attribute values of the plurality of neighboring points or based on a single one of the reconstructed attribute values of the plurality of neighboring points;

in response to the received indicator being determined as indicating that the reconstruction of the current point is based on the weighted sum of all of the reconstructed attribute values of the plurality of neighboring points, reconstruct the current point in the prediction step in the lifting scheme based on the weighted sum of all of the reconstructed attribute values of the plurality of neighboring points; and

in response to the received indicator being determined as indicating that the reconstruction of the current point is based on the single one of the reconstructed attribute values of the plurality of neighboring points, reconstruct the current point in the prediction step in the lifting scheme based on a reconstructed attribute value of a selected one of the plurality of neighboring points.

20. The apparatus of claim 19 , wherein

the indicator is generated based on a determination that a measurement of variability of the reconstructed attribute values of the plurality of neighboring points is below a threshold, and

the variability of the reconstructed attribute values of the plurality of neighboring points is a maximum difference of the reconstructed attribute values among the plurality of neighboring points.

21. The apparatus of claim 19 , wherein the weighted sum of all of the reconstructed attribute values of the neighboring points is calculated based on geometric weights of the reconstructed attribute values of the neighboring points.