Tile Tracks for Geometry-Based Point Cloud Data

Abstract

This invention provides a method for providing efficient partial access to non-time-specified geometry-based point cloud compression (G-PCC) data transported in ISO-based media file format (ISOBMFF) files. [Solution] The method includes a decoding device receiving a time-specified metadata track that identifies point cloud tiles corresponding to one or more spatial regions in a point cloud scene, determining one or more point cloud tiles to be used to render an image, and retrieving one or more geometry tile tracks corresponding to the determined one or more point cloud tiles via a communication network. Each geometry tile track contains point cloud geometry data for its respective tile. The method also processes the retrieved geometry tile tracks.

Claims

1. It is a method, Receiving a time-specified metadata track that identifies point cloud tiles corresponding to one or more spatial regions within a point cloud scene, In the decoding device, determine one or more point cloud tiles to be used to render the image, Retrieving one or more geometry tile tracks from the communication network corresponding to one or more determined point cloud tiles, wherein each geometry tile track includes point cloud geometry data for each tile. A method comprising processing the extracted geometric tile track.

2. The method according to claim 1, further comprising identifying a set of tile tracks that carry information necessary to render a specific spatial region or tile within the point cloud scene.

3. The method according to claim 1 or 2, wherein each geometry tile track is linked to one or more attribute tile tracks.

4. The method according to any one of claims 1 to 3, wherein, when data files are transported using tile tracks, the tile base track includes tile inventory information within the tile base track sample, and the geometry tile track includes a sample group for signaling the tile inventory of tiles present within the geometry tile track sample.

5. The method according to any one of claims 1 to 4, wherein a G-PCC component tile track belonging to an alternative group is referenced by a G-PCC base track or the respective G-PCC geometry tile track.

6. Receiving a formatted container containing geometry-based point cloud data, which includes one or more point cloud tiles, Retrieving the time-specified metadata track from the formatted container, wherein the time-specified metadata track includes multiple tile identifiers, and each tile identifier corresponds to each of the one or more point cloud tiles. The selection of at least one selected tile from the one or more point cloud tiles, wherein the at least one selected tile corresponds to at least one tile identifier. Identifying at least one geometric tile track associated with the at least one tile identifier, Identifying a base track containing initialization data for the at least one selected tile using a first track reference type associated with the at least one geometry tile track, The method according to any one of claims 1 to 5, further comprising decoding the at least one selected tile into at least one decoded tile using the at least one geometry tile track and the initialization data.

7. Further includes identifying at least one attribute tile track associated with the at least one selected tile, The method according to any one of claims 1 to 6, wherein decoding the at least one selected tile includes using the at least one geometry tile track, the at least one attribute tile track, and the initialization data to obtain the at least one decoded tile.

8. The method according to any one of claims 1 to 7, wherein the decoding is performed without decoding all of the geometry-based point cloud data.

9. Identifying the viewport for the client, Identifying at least one 3D region associated with the aforementioned viewport, When the information for the at least one 3D region is dynamically changing, the information for the at least one 3D region present in the 3D spatial region information time-specified metadata track sample is identified, Based on the available 3D region information, identify which of the at least one 3D region is associated with the viewport, From the aforementioned 3D spatial region information time-specified metadata track sample, at least one tile associated with at least one target 3D region is identified, By using the information present within each tile track, it is possible to identify at least one tile track associated with the at least one tile associated with the at least one target 3D region, The method according to any one of claims 1 to 8, further comprising extracting the identified tile track from a G-PCC data file, decoding the identified tile track, and displaying the decoded tile track in the current viewport or based on the viewport.

10. The method according to any one of claims 1 to 9, wherein the time-specified metadata track sets a sample as either a synchronous sample or an asynchronous sample, the asynchronous sample in the time-specified metadata track carries only updated 3D spatial region information by referring to 3D spatial region information available in the nearest preceding synchronous sample, and the asynchronous sample in the time-specified metadata track signals only updated 3D spatial region information by referring to 3D spatial region information available in the nearest preceding synchronous sample, including updated dimensions or associated 3D tiles and any added or canceled 3D spatial regions.

11. The method according to any one of claims 1 to 10, wherein different encoded versions of a cloud tile are signaled using multiple tile-based tracks and have the same group identification.

12. The method according to any one of claims 1 to 11, wherein the frames of the point cloud data are distributed among a plurality of identified time layers, each frame is assigned to one of the plurality of identified time layers, a geometry tile track signals at least one time layer identifier of G-PCC samples present in the geometry tile track, and samples of G-PCC components in the geometry tile track are grouped based on the time level of each sample.

13. The method according to claim 12, wherein frames of individual time layers among the plurality of identified time layers are decoded and rendered without decoding and rendering any other time layers.

14. A non-temporary computer-readable storage medium storing instructions, wherein when the instructions are executed by a processor, the processor causes the processor to perform the method according to any one of claims 1 to 13.

15. It is a device, At least one processor, The device comprises at least one memory storing instructions, and when the instructions are executed by the at least one processor, Receive a time-specified metadata track that identifies point cloud tiles corresponding to one or more spatial regions within a point cloud scene. In the decoding device, one or more point cloud tiles are determined to be used to render the image. From the communication network, retrieve one or more geometry tile tracks, each containing point cloud geometry data for its respective tile, corresponding to the one or more point cloud tiles determined, and A device that can be operated to process the extracted geometric tile track.

Images