Point cloud geometry padding

Patent metadata

Field	Value
Publication number	US-11334969-B2
Application number	US-201916655685-A
Country	US
Kind code	B2
Filing date	Oct 17, 2019
Priority date	Mar 19, 2019
Publication date	May 17, 2022
Grant date	May 17, 2022

Abstract

A method of point cloud geometry padding is described herein. The method searches for a reconstruction point in a compressed occupancy map to perform the geometry padding instead of using an uncompressed occupancy map.

First claim

What is claimed is:

1. A method programmed in a non-transitory memory of a device comprising:
generating geometry images from a point cloud;
placing the geometry images on a 2D canvas; and
filling empty spaces on the 2D canvas with a padding method which considers a 3D position of newly added points due to lossy compression of an occupancy map, wherein the padding method comprises searching in a 3D space for a value that, when reconstructed, minimizes a distance to the point cloud, wherein searching in the 3D space includes searching a limited range of values centered around a representative value of local points in the 2D canvas, wherein searching in the 3D space comprises starting with an average of three neighboring points and then sequentially adding or subtracting values starting with 1 to the average to determine results and repetitively comparing the results with an original point cloud value to determine a reconstructed point, wherein the padding method comprises adding the reconstructed point in an empty space of the empty spaces.

2. The method of claim 1 further comprising adding new values to a reconstructed point cloud due to the lossy compression of the occupancy map.

3. The method of claim 1 wherein searching in the 3D space comprises only searching for points within a specified range.

4. The method of claim 1 further comprising generating a compressed geometry image.

5. The method of claim 1 further comprising generating a compressed bitstream for the point cloud.

6. An apparatus comprising:
a non-transitory memory for storing an application, the application for:
generating geometry images from a point cloud;
placing the geometry images on a 2D canvas; and
filling empty spaces on the 2D canvas with a padding method which considers a 3D position of newly added points due to lossy compression of an occupancy map, wherein the padding method comprises searching in a 3D space for a value that, when reconstructed, minimizes a distance to the point cloud, wherein searching in the 3D space includes searching a limited range of values centered around a representative value of local points in the 2D canvas, wherein searching in the 3D space comprises starting with an average of three neighboring points and then sequentially adding or subtracting values starting with 1 to the average to determine results and repetitively comparing the results with an original point cloud value to determine a reconstructed point, wherein the padding method comprises adding the reconstructed point in an empty space of the empty spaces; and

a processor coupled to the memory, the processor configured for processing the application.

7. The apparatus of claim 6 wherein the application is further configured for adding new values to a reconstructed point cloud due to the lossy compression of the occupancy map.

8. The apparatus of claim 6 wherein searching in the 3D space comprises only searching for points within a specified range.

9. The apparatus of claim 6 wherein the application is further configured for generating a compressed geometry image.

10. The apparatus of claim 6 wherein the application is further configured for generating a compressed bitstream for the point cloud.

11. A system comprising:
one or more cameras for acquiring three dimensional content; and
an encoder for encoding the three dimensional content by:
generating geometry images from the three dimensional content;
placing the geometry images on a 2D canvas; and
filling empty spaces on the 2D canvas with a padding method which considers a 3D position of newly added points due to lossy compression of an occupancy map, wherein the padding method comprises searching in a 3D space for a value that, when reconstructed, minimizes a distance to the point cloud, wherein searching in the 3D space includes searching a limited range of values centered around a representative value of local points in the 2D canvas, wherein searching in the 3D space comprises starting with an average of three neighboring points and then sequentially adding or subtracting values starting with 1 to the average to determine results and repetitively comparing the results with an original point cloud value to determine a reconstructed point, wherein the padding method comprises adding the reconstructed point in an empty space of the empty spaces.

12. The system of claim 11 wherein the encoder is further configured for adding new values to a reconstructed point cloud due to the lossy compression of the occupancy map.

13. The system of claim 11 wherein searching in the 3D space comprises only searching for points within a specified range.

14. The system of claim 11 wherein the encoder is further configured for generating a compressed geometry image.

15. The system of claim 11 wherein the encoder is further configured for generating a compressed bitstream for the point cloud.

16. A method programmed in a non-transitory memory of a device comprising:
generating geometry images from a point cloud;
placing the geometry images on a 2D canvas;
filling empty spaces on the 2D canvas with a padding method which considers a 3D position of newly added points due to lossy compression of an occupancy map, wherein the padding method includes selecting a value for positions that generates a reconstructed point that minimizes a distance to the point cloud by starting with an average of three neighboring points and then sequentially adding or subtracting values starting with 1 to the average to determine results and repetitively comparing the results with an original point cloud value to determine the reconstructed point, wherein the empty spaces are each filled with reconstructed points, wherein the padding method comprises adding the reconstructed point in an empty space of the empty spaces.

Assignees

• Sony Corp
• Sony Group Corp

Inventors

• Graziosi Danillo

Classifications

• H04N19/597 Primary

  specially adapted for multi-view video sequence encoding · CPC title

• H04N19/96 Primary

  Tree coding, e.g. quad-tree coding · CPC title

• H04N19/56

  Motion estimation with initialisation of the vector search, e.g. estimating a good candidate to initiate a search · CPC title

• G06T9/00

  Image coding (bandwidth or redundancy reduction for static pictures H04N1/41; coding or decoding of static colour picture signals H04N1/64; methods or arrangements for coding, decoding, compressing or decompressing digital video signals H04N19/00) · CPC title

• G06T2207/10028

  Range image; Depth image; 3D point clouds · CPC title