Three-dimensional data encoding method and device, and three-dimensional data decoding method and device
Provided is a method of processing and transmitting three-dimensional data represented as a point cloud. The method of encoding three-dimensional (3D) data includes: generating a geometry image indicating position information of points, a texture image indicating color information of the points, and an occupancy map indicating occupancy information of the position information of the points in the geometry image, by projecting the points included in the 3D data onto a two-dimensional (2D) plane; generating a filtered occupancy map by performing filtering on the occupancy map; performing image padding on the geometry image and the texture image, based on the filtered occupancy map; and generating and outputting a bitstream including a padded geometry image, a padded texture image, a downsampled occupancy map, and information about the filtering performed on the occupancy map.
1. A method of encoding three-dimensional (3D) data, the method comprising:
generating a geometry image indicating position information of points included in the 3D data, a texture image indicating color information of the points, and an occupancy map indicating occupancy information of the position information of the points in the geometry image, by projecting the points included in the 3D data onto a two-dimensional (2D) plane;
generating a filtered occupancy map by performing filtering on the occupancy map;
performing image padding on the geometry image and the texture image, based on the filtered occupancy map; and
generating and outputting a bitstream including a padded geometry image, a padded texture image, a downsampled occupancy map, and information about the filtering performed on the occupancy map.
2. The method of claim 1 , wherein the generating of the occupancy map comprises downsampling an occupancy map based on an occupancy precision value N.
3. The method of claim 2 , wherein, in the downsampling, when there is a projected point in a predetermined N*N pixel block, an entire area of the predetermined N*N pixel block of the occupancy map is filled with 1.
4. The method of claim 1 , wherein the generating of the filtered occupancy map comprises upsampling the occupancy map based on an occupancy precision value N.
5. The method of claim 4 , wherein the generating of the filtered occupancy map further comprises performing 2D filtering on the upsampled occupancy map.
6. The method of claim 5 , wherein, in the generating of the filtered occupancy map, whether to perform the 2D filtering is determined based on at least one of a size of an occupancy map patch or a number of 1's included in the occupancy map patch.
7. The method of claim 1 , wherein the filtered occupancy map is generated in units of patches.
8. A computer program product comprising one or more non-transitory computer readable recording media having stored therein a program that, when executed by a processor, performs the method of claim 1 .
9. An apparatus for encoding three-dimensional (3D) data, the apparatus comprising:
a memory;
a communicator; and
at least one processor,
wherein the at least one processor is configured to:
generate a geometry image indicating position information of points included in the 3D data, a texture image indicating color information of the points, and an occupancy map indicating occupancy information of the position information of the points in the geometry image, by projecting the points included in the 3D data onto a two-dimensional (2D) plane;
generate a filtered occupancy map by performing filtering on the occupancy map;
perform image padding on the geometry image and the texture image, based on the filtered occupancy map; and
generate and output a bitstream including a padded geometry image, a padded texture image, a downsampled occupancy map, and information about the filtering performed on the occupancy map.
10. A method of decoding three-dimensional (3D) data, the method comprising:
receiving a bitstream;
obtaining, from the bitstream, a texture image, a geometry image, an occupancy map, and filtering information;
generating a filtered occupancy map by performing filtering on the occupancy map, based on the filtering information, wherein the filtering comprises at least one of upsampling or downsampling;
reconstructing position information of points based on the geometry image and the filtered occupancy map; and
reconstructing 3D data by reconstructing color information of points with reconstructed position information, based on the texture image.
11. An apparatus for decoding three-dimensional (3D) data, the apparatus comprising:
a memory;
a communicator; and
at least one processor,
wherein the at least one processor is configured to:
receive a bitstream;
obtain, from the bitstream, a texture image, a geometry image, an occupancy map, and filtering information,
generate a filtered occupancy map by performing filtering on the occupancy map based on the filtering information, wherein the filtering comprises at least one of upsampling or downsampling;
reconstruct position information of points based on the geometry image and the filtered occupancy map, and
reconstruct 3D data by reconstructing color information of points with reconstructed position information, based on the texture image.