Method and apparatus for video encoding and decoding with bi-directional optical flow adapted to weighted prediction
Different implementations are described, particularly implementations for video encoding and decoding are presented. According to an implementation, in a method for encoding or decoding a part of an image, at least an equation of the bi-directional optical flow is modified for unequal weights in a weighted prediction of the part of the image. According to non-limiting examples, modifying the bi-directional optical flow equation comprises bit depth shifting the unequal weights used in the weighted prediction, quantifying the unequal weighs used in the weighted prediction or deriving modified weighs to apply to predictions of the part of the image from the unequal weights used in weighted prediction.
1. A method, comprising:
performing sample-wise motion refinement using bi-directional optical flow of a block-wise motion compensated prediction of at least a part of an image, the block-wise motion compensated prediction being a bi-prediction;
wherein the bi-directional optical flow comprises bit depth shifting unequal weights used in the bi-prediction of the at least one part of the image by a bit depth of the unequal weights and processing a sample-wise motion refinement correction on a bit depth increased by the bit depth of the unequal weights; and
encoding or decoding the at least a part of the image using the sample-wise motion refined prediction.
2. The method of claim 1 , wherein the bi-prediction using unequal weights is one of a generalized bi-prediction, a bi-prediction with coding units level weights, a bi-prediction with image level weights, or a bi-prediction with an image distance based weights.
3. A method, comprising:
performing sample-wise motion refinement using bi-directional optical flow of a block-wise motion compensated prediction of at least a part of an image, the block-wise motion compensated prediction being a bi-prediction;
wherein the bi-directional optical flow comprises quantifying unequal weights used in the bi-prediction of the at least one part of the image; and
encoding or decoding the at least a part of the image using the sample-wise motion refined prediction.
4. The method of claim 1 , wherein the bi-directional optical flow comprises deriving modified weights to apply to the block-wise motion compensated prediction of the part of the image from the unequal weights used in the bi-prediction.
5. An apparatus, comprising,
a memory; and
one or more processors configured to:
perform sample-wise motion refinement using bi-directional optical flow of a block-wise motion compensated prediction of at least a part of an image, the bi-directional optical flow comprising a bit depth shift of unequal weights used in a bi-prediction of the block-wise motion compensated prediction by a bit depth of the unequal weights and a sample-wise motion refinement correction processed on a bit depth increased by the bit depth of the unequal weights; and
encode or decode the at least a part of the image using the sample-wise motion refined prediction.
6. The apparatus of claim 5 , wherein the bi-prediction using unequal weights is one of a generalized bi-prediction, a bi-prediction with coding units level weights, a bi-prediction with image level weights, or a bi-prediction with an image distance based weights.
7. An apparatus, comprising:
a memory; and
one or more processors configured to:
perform sample-wise motion refinement using bi-directional optical flow of a block-wise motion compensated prediction of at least a part of an image, the block-wise motion compensated prediction being a bi-prediction;
wherein the bi-directional optical flow comprises a quantification of the unequal weights used in the bi-prediction of the at least one part of the image; and
encode or decode the at least a part of the image using the sample-wise motion refined prediction.
8. The apparatus of claim 5 , wherein the bi-directional optical flow comprises derivation of modified weights to apply to the block-wise motion compensated prediction of the part of the image from the unequal weights used in bi-prediction.
9. A non-transitory computer readable medium having stored thereon instructions that when executed by one or more processors perform the method of claim 1 .
10. A non-transitory computer readable medium containing data content generated according to the method of claim 3 for playback using a processor.
11. A non-transitory computer readable medium containing data content generated according to the method of claim 1 for playback using a processor.
12. The method of claim 1 , wherein the bi-directional optical flow comprises quantifying the unequal weights used in the bi-prediction.
13. The method of claim 3 , wherein the bi-prediction using unequal weights is one of a generalized bi-prediction, a bi-prediction with coding units level weights, a bi-prediction with image level weights, or a bi-prediction with an image distance based weights.
14. The method of claim 3 , wherein the bi-directional optical flow comprises bit depth shifting the unequal weights used in the bi-prediction.
15. The method of claim 3 , wherein the bi-directional optical flow comprises deriving modified weights to apply to the block-wise motion compensated prediction of the part of the image from the unequal weights used in the bi-prediction.
16. A non-transitory computer readable medium having stored thereon instructions that when executed by one or more processors perform the method of claim 3 .
17. The apparatus of claim 5 , wherein the bi-directional optical flow comprises a quantification of the unequal weights used in the bi-prediction.
18. The apparatus of claim 7 , wherein the bi-prediction using unequal weights is one of a generalized bi-prediction, a bi-prediction with coding units level weights, a bi-prediction with image level weights, or a bi-prediction with an image distance-based weights.
19. The apparatus of claim 7 , wherein the bi-directional optical flow comprises a bit depth shift of the unequal weights used in the bi-prediction.
20. The apparatus of claim 7 , wherein the bi-directional optical flow comprises derivation of modified weights to apply to the block-wise motion compensated prediction of the part of the image from the unequal weights used in bi-prediction.