Method and apparatus for encoding and decoding three-dimensional scenes in and from a data stream

1. A method of decoding a three-dimensional scene from a data stream, the method comprising:

obtaining from said data stream:

color pictures packed in a color image, wherein the color pictures are parts of at least two views acquired from at least two points of view;

depth pictures packed in a depth image; and

metadata comprising, for a color picture of said color image:

de-projection data;

a color data comprising a description of a location of a color picture in the color image; and

a geometry data of a first geometry type comprising a description of a location of a depth picture in the depth image, or a geometry data of a second geometry type comprising an identifier of a three-dimensional object represented as a three-dimensional mesh; and

for a color picture of said color image:

on a condition that the geometry data is of the first geometry type, using de-projection data for de-projecting color of pixels of the color picture described by the color data at a depth value determined according to pixels of the depth picture described by the geometry data; and

on a condition that the geometry data is of the second geometry type, using de-projection data for de-projecting color of pixels of the color picture described by the color data at a depth value determined according to vertices and faces of the mesh.

2. The method of claim 1 , wherein vertices and faces of meshes identified in geometry data of the second type of geometry data are obtained from the stream.

3. The method of claim 1 , wherein a color data further comprises a description of a size and/or of a shape of the color picture in the color image.

4. The method of claim 1 , wherein the stream comprises a sequence of color images, depth images and associated metadata, a depth image and related metadata being associated with a color image; the color images being structured by temporal information.

5. A method for encoding a three-dimensional scene in a data stream, the method comprising:

obtaining a set of patches, a patch comprising:

de-projection data;

a color picture; and

geometry data of a first geometry type comprising a depth picture, or geometry data of a second geometry type comprising an identifier of a three-dimensional object represented as a three-dimensional mesh;

wherein the color pictures of said patches are parts of at least two views acquired from at least two points of view;

generating a color image by packing color pictures of said patches and a color data comprising a description of a location of the color picture as packed in the color image;

generating a depth image by packing depth pictures of patches having a geometry data of the first geometry type and replacing the depth picture in geometry data of corresponding patch, by a description of a location of the depth picture as packed in the depth image; and

encoding in the stream:

the generated color image;

the generated depth image; and

metadata comprising the de-projection data, the color data and the geometry data of every patch.

6. The method of claim 5 , wherein vertices and faces of meshes identified in geometry data of the second type of geometry data are further encoded in a separate element of syntax of the stream.

7. The method of claim 5 , wherein a color data further comprises a description of a size and/or of a shape of the color picture in the color image.

8. The method of claim 5 , wherein the stream comprises a sequence of color images, depth images and associated metadata, a depth image and related metadata being associated with a color image; the color images being structured by temporal information.

9. A device for decoding a three-dimensional scene from a data stream, the device comprising a memory associated with a processor configured to:

obtain from said data stream:

color pictures packed in a color image, wherein the color pictures are parts of at least two views acquired from at least two points of view;

depth pictures packed in a depth image; and

metadata comprising, for a color picture of said color image:

de-projection data;

a color data comprising a description of a location of a color picture in the color image; and

a geometry data of a first geometry type comprising a description of a location of a depth picture in the depth image, or a geometry data of a second geometry type comprising an identifier of a three-dimensional object represented as a three-dimensional mesh; and

for a color picture of said color image:

on a condition that the geometry data is of the first geometry type, use de-projection data for de-projecting color of pixels of the color picture described by the color data at a depth value determined according to pixels of the depth picture described by the geometry data; and

on a condition that the geometry data is of the second geometry type, use de-projection data for de-projecting color of pixels of the color picture described by the color data at a depth value determined according to vertices and faces of the mesh.

10. The device of claim 9 , wherein vertices and faces of meshes identified in geometry data of the second type of geometry data are obtained from the stream.

11. The device of claim 9 , wherein a color data further comprises a description of a size and/or of a shape of the color picture in the color image.

12. The device of claim 9 , wherein the stream comprises a sequence of color images, depth images and associated metadata, a depth image and related metadata being associated with a color image; the color images being structured by temporal information.

13. A device for encoding a three-dimensional scene in a data stream, the device comprising a memory associated with a processor configured to:

obtain a set of patches, a patch comprising:

de-projection data;

a color picture; and

geometry data of a first geometry type comprising a depth picture, or geometry data of a second geometry type comprising an identifier of a three-dimensional object represented as a three-dimensional mesh;

wherein the color pictures of said patches are parts of at least two views acquired from at least two points of view;

generate a color image by packing color pictures of said patches and a color data, the color data comprising a description of a location of the color picture as packed in the color image;

generate a depth image by packing depth pictures of patches having a geometry data of the first geometry type and replacing the depth picture in geometry data of corresponding patch, by a description of a location of the depth picture of said patch as packed in the depth image; and

encode in the stream:

the generated color image;

the generated depth image; and

metadata comprising the de-projection data, the color data and the geometry data of every patch.

14. The device of claim 13 , wherein vertices and faces of meshes identified in geometry data of the second type of geometry data are further encoded in a separate element of syntax of the stream.

15. The device of claim 13 , wherein a color data further comprises a description of a size and/or of a shape of the color picture in the color image.

16. The device of claim 13 , wherein the stream comprises a sequence of color images, depth images and associated metadata, a depth image and related metadata being associated with a color image; the color images being structured by temporal information.

17. A non-transitory computer readable medium comprising computer-executable instructions to enable a processor to perform the method of claim 1 .

18. A non-transitory computer readable medium comprising computer-executable instructions to enable a processor to perform the method of claim 5 .

19. A non-transitory computer readable medium storing data content generated according to the method of claim 5 .

20. A non-transitory computer readable medium storing data content generated by the device of claim 13 .