A method and apparatus for encoding/decoding the geometry of a point cloud representing a 3d object

A method comprising:

• determining an octree-based structure representing points of a point cloud included in a leaf cube associated with a leaf node;
• obtaining a predictor of said determined octree-based structure from an already coded octree-based structure;
• encoding a first information representative of said determined octree-based structure by using a first context dependent entropy encoding;
• and
• encoding a second information indicating if a leaf cube associated with a leaf node of said determined octree-based structure includes a point representative of a part of the point cloud by using a second context dependent entropy encoding,
• wherein the context of at least one of said first and second context dependent entropy encodings is depending on said predictor.

A method comprising:

• decoding a first information of a second type indicating that an octree-based structure has been determined for approximating points of a point cloud included in a leaf cube associated with a leaf node;
• decoding an information representative of a predictor of said octree-based structure;
• obtaining an octree-based structure for said leaf cube by context adaptive entropy decoding a second information of the second type using a first context;
• determining if a leaf cube of said octree-based structure includes a point of the point cloud to be decoded by context adaptive entropy decoding an information of a first type using a second context, and in that case, adding a point to the leaf cube of said octree-based structure, wherein at least one of the first and second context is depending on said predictor.

A signal carrying:

• an information representative of a predictor of an octree-based structure representing points of a point cloud included in a leaf cube associated with a leaf node;
• an information of a second type encoded in the signal using a first context dependent entropy encoding and indicating that an octree-based structure has been determined for said leaf cube;
• an information of a first type encoded in the signal using a second context dependent entropy encoding and representative of said determined octree-based structure;
• and
• at least one second information of the second type indicating if a leaf cube of said octree-based structure includes a point representative of a part of the point cloud,
• wherein the context of at least one of said first and second context dependent entropy encodings is depending on said predictor.

A device comprising electronic circuitry adapted for determining an octree-based structure representing points of a point cloud included in a leaf cube associated with a leaf node;

• obtaining a predictor of said determined Octree-based structure from an already coded Octree-based structure;
• encoding a first information of a first type representative of said determined octree-based structure by using a first context dependent entropy encoding;
• and
• encoding a first information of a second type indicating if a leaf cube associated with a leaf node of said determined octree-based structure includes a point representative of a part of the point cloud by using a second context dependent entropy encoding,
• wherein the context of at least one of said first and second context dependent entropy encodings is depending on said predictor.

A device configured for

• decoding an information of a second type indicating that an octree-based structure has been determined for approximating points of a point cloud included in a leaf cube associated with a leaf node;
• decoding, an information representative of a predictor of said octree-based structure;
• obtaining an octree-based structure for said leaf cube by context adaptive entropy decoding an information of a second type using a first context;
• determining if a leaf cube of said local octree-based structure includes a point of the point cloud to be decoded by context adaptive entropy decoding an information of a second type using a second context, and in that case, adding a point to the leaf cube of said octree-based structure, wherein the at least one of the first and the second context is depending on said predictor.