Method and device for rendering point cloud-based data

An apparatus for rendering point cloud-based volumetric data, the apparatus comprising:

• a memory storing instructions;
• at least one processor configured to execute the instructions to:
• receive a plurality of two-dimensional (2D) image frames corresponding to a three-dimensional (3D) point cloud, and projection metadata comprising data of projecting the 3D point cloud into a 2D image plane,
• reconstruct a plurality of 3D points corresponding to the 3D point cloud, based on the plurality of 2D image frames and the projection metadata, and
• generate texture images based on the plurality of 3D points;
• and
• a display configured to render at least one quad image based on the texture images,
• wherein each of the plurality of 2D image frames comprises a plurality of pixels, and
• wherein the at least one processor is further configured to execute the instructions to:
• identify redundant pixels from each of the plurality of 2D image frames based on occupancy metadata corresponding to each of the plurality of 2D image frames, wherein the occupancy metadata comprises information of points located on a line lying in a direction of the projecting the 3D point cloud into the 2D image plane, information of the 3D points among the plurality of 3D points concentrated on a small area of which length is predetermined, information of a number of the points located on the line lying in the direction of the projecting the 3D point cloud into the 2D image plane, binary information about occupant pixels indicating pixels being valid occupied pixels, and binary information about the redundant pixels indicating pixels being redundant pixels;
• and
• remove the redundant pixels from the plurality of 2D image frames.

A non-transitory computer readable medium storing computer readable instructions that are executable by a processor to perform a method of rendering point cloud-based volumetric data, the method comprising:

• receiving, at a computing device, a plurality of two-dimensional (2D) image frames corresponding to a three-dimensional (3D) point cloud, and projection metadata comprising data of projecting the 3D point cloud into a 2D image plane;
• reconstructing a plurality of 3D points corresponding to the 3D point cloud, based on the plurality of 2D image frames and the projection metadata;
• generating texture images based on the plurality of 3D points;
• and
• rendering at least one quad image based on the texture images,
• wherein each of the plurality of 2D image frames comprises a plurality of pixels, and
• wherein the reconstructing the plurality of 3D points comprises:
• identifying redundant pixels from each of the plurality of 2D image frames based on occupancy metadata corresponding to each of the plurality of 2D image frames, wherein the occupancy metadata comprises information of points located on a line lying in a direction of the projecting the 3D point cloud into the 2D image plane, information of the 3D points among the plurality of 3D points concentrated on a small area of which length is predetermined, information of a number of the points located on the line lying in the direction of the projecting the 3D point cloud into the 2D image plane, binary information about occupant pixels indicating pixels being valid occupied pixels, and binary information about the redundant pixels indicating pixels being redundant pixels;
• and
• removing the redundant pixels from the plurality of 2D image frames.