Point cloud occupancy map compression

A method for compressing a point cloud, the method comprising:

• determining a plurality of patches, each corresponding to a respective portion of the point cloud;
• generating, for each patch, one or more patch images;
• packing the generated patch images into an image frame;
• generating an occupancy map for the image frame, wherein the occupancy map indicates pixels of the image frame that are occupied by one of the patch images and pixels of the image frames that are un-occupied;
• and
• encoding a compressed version of the occupancy map, wherein said encoding comprises:
• indicating empty blocks of the occupancy map that correspond to respective portions of the image frame that include only un-occupied pixels;
• assigning binary values to sub-blocks of each non-empty block of the occupancy map, wherein occupied sub-blocks and non-occupied sub-blocks are assigned different binary values;
• selecting a traversal order for encoding the sub-blocks of the occupancy map such that sub-blocks with similar binary values are traversed sequentially;
• and
• encoding the traversal order and sequential binary values for the sub-blocks of the occupancy map traversed according to the traversal order.

A method for decompressing a compressed version of a point cloud, the method comprising:

• receiving an image frame comprising:
• occupied portions occupied with patch images;
• and
• un-occupied portions,
• wherein the patch images comprise attribute information or geometry information for respective patches of the point cloud;
• receiving a compressed version of an occupancy map for the patch image;
• decompressing the compressed version of the occupancy map, wherein said decompressing comprises:
• determining non-empty blocks of the occupancy map;
• and
• for each non-empty block:
• determining a traversal order for traversing sub-blocks of the non-empty block;
• and
• assigning sequential binary values from a set of encoded binary values to the sub-blocks according to the traversal order to indicate whether the sub-blocks are occupied or non-occupied sub-blocks;
• and
• identifying patch images included in the image frame, based at least in part on the decompressed occupancy map.