Tree-based deep entropy model for point cloud compression

A device comprising a memory associated with a processor configured for:

• obtaining a node of a tree-structured point cloud from a bitstream;
• for each node of the of tree-structured point cloud;
• initializing context information for the node;
• predicting an occupancy symbol distribution for the node by using a learning-based entropy model utilizing the context information of the node and feature information from neighboring nodes and a parent node previously obtained from the bitstream;
• decoding an occupancy symbol for the node, using an adaptive entropy decoder based on the occupancy symbol distribution;
• and
• outputting an expanded tree based on the occupancy symbol.

A method comprising:

• obtaining a point cloud structured as a tree;
• for each node of the tree;
• initializing context information;
• predicting an occupancy symbol distribution by using a learning-based entropy model utilizing the context information of the node and feature information from neighboring nodes and a parent node of the node;
• encoding, into a bitstream, an occupancy symbol by using an adaptive entropy encoder based on the occupancy symbol distribution;
• and
• generating a combined bitstream for the tree.

A device comprising a memory associated with a processor configured for:

• obtaining a point cloud structured as a tree;
• for each node of the tree;
• initializing context information;
• predicting an occupancy symbol distribution by using a learning-based entropy model utilizing the context information of the node and feature information from neighboring nodes and a parent node of the node;
• encoding, into a bitstream, an occupancy symbol by using an adaptive entropy encoder based on the occupancy symbol distribution;
• and
• generating a combined bitstream for the tree.