Method for decoding immersive video and method for encoding immersive video
An image encoding method according to the present disclosure may include generating an atlas based on a plurality of viewpoint images; encoding the atlas; and encoding metadata for the atlas. In this case, the metadata may include first information showing whether a viewpoint image is a background image.
1 . An image encoding method, the method comprising:
generating atlases based on a plurality of viewpoint images;
encoding the atlases; and
encoding metadata for the plurality of viewpoint images and the atlases,
wherein the metadata comprises a first flag indicating whether a viewpoint image is a background image or not, and
wherein the first flag is encoded for each of the plurality of viewpoint images.
2 . The method of claim 1 , wherein the background image is generated by merging spatial background images extracted from images captured at different positions.
3 . The method of claim 1 , wherein the background image is generated by merging temporal background images extracted from images captured on different timepoints.
4 . The method of claim 3 , wherein one of the temporal background images is designated as a representative background image, and
wherein the background image is generated by updating an empty pixel in the representative background image based on a pixel value of a reference background image,
the empty pixel being included in a region occupied by a foreground object in an image from which the representative background image is extracted.
5 . The method of claim 4 , wherein:
a pixel adjacent to a region occupied by the foreground object in the representative background image is set as a dirty pixel, and
wherein the dirty pixel is updated by the reference background image.
6 . The method of claim 1 , wherein the atlases comprise a background atlas and a foreground atlas, and
wherein the background atlas comprises only viewpoint images that are background images.
7 . The method of claim 1 , wherein
the metadata further comprises a second flag indicating separate processing of a background and a foreground is enabled, and
wherein the first flag is present in the metadata only when the second flag is encoded to indicate that separate processing of the background and the foreground is enabled.
8 . An image decoding method, the method comprising:
decoding atlases;
decoding metadata for a plurality of viewpoint images and the atlases; and
synthesizing a target viewpoint image by using the decoded atlases and the decoded metadata,
wherein the metadata comprises a first flag indicating whether a viewpoint image is a background image or not, and
wherein the first flag is decoded for each of the plurality of viewpoint images.
9 . The method of claim 8 , wherein
the metadata further comprises a second flag indicating separate processing of a background and a foreground is enabled, and
wherein the first flag is present in the metadata only when the second flag indicates that separate processing of the background and the foreground is enabled.
10 . A non-transitory computer readable recording medium comprising instructions when executed cause a computer to carry out:
generating atlases based on a plurality of viewpoint images;
encoding the atlases; and
encoding metadata for the plurality of viewpoint images and the atlases,
wherein the metadata comprises a first flag indicating whether a viewpoint image is a background image or not, and
wherein the first flag is encoded for each of the plurality of viewpoint images.
11 . The method of claim 8 , wherein the atlases comprise a background atlas and a foreground atlas, and
wherein the background atlas comprises only viewpoint images that are background images.