Reference processing using advanced motion models for video coding

1. A method for processing a reference picture, comprising:

i) performing motion estimation on an original picture to be encoded, based on a reconstructed reference picture, and a processed reference picture estimate to provide motion model parameters;

ii) further processing the motion estimated picture to obtain a further processed reference picture estimate; and

iterating the motion estimation and the further processing to provide a processed reference picture,

wherein:

the reconstructed reference picture is divided into a plurality of regions,

the performing of the motion estimation provides motion model parameters for the plurality of regions, and

for at least one region of the reconstructed reference picture in the plurality of regions, the performing of the motion estimation comprises applying a plurality of motion models to the at least one region of the reconstructed reference picture to provide motion model parameters associated with the at least one region of the reconstructed reference picture, wherein each motion model among the plurality of motion models is applied to the at least one region of the reconstructed reference picture separately, the motion model parameters associated with the at least one region being an average or weighted average of motion model parameters provided by applying each of the plurality of motion models to the at least one region of the reconstructed reference picture separately, wherein the applying a plurality of motion models to the at least one region of the reconstructed reference picture comprises:

applying a first motion model on the at least one region to generate a first set of motion model parameters; and

applying a second motion model on the at least one region to generate a second set of motion model parameters,

wherein the motion model parameters associated with the at least one region is obtained by an average or a weighted average of the first set of motion model parameters with the second set of motion model parameters.

2. The method as recited in claim 1 , wherein the further processing comprises further motion estimation and estimating at least one of interpolation filter parameters, intensity compensation parameters, and denoising filter parameters, and applying the motion model parameters and the at least one of interpolation filter parameters, intensity compensation parameters, and denoising filter parameters to the reconstructed reference picture.

3. The method as recited in claim 1 , wherein the reconstructed reference picture comprises an original reference picture retrieved from the reference picture buffer, devoid of any processing subsequent to being retrieved from the reference picture buffer and prior to the performing of motion estimation.

4. The method as recited in claim 1 , wherein the reconstructed reference picture comprises a previously processed reference picture, the previously processed reference picture being a reference picture that has been processed subsequent to being retrieved from the reference picture buffer and prior to the performing of motion estimation.

5. The method as recited in claim 4 , wherein the previously processed reference picture has been processed by applying at least one of alternative motion model parameters, alternative interpolation filter parameters, alternative intensity compensation parameters, and alternative denoising filter parameters.

6. The method as recited in claim 5 , wherein the at least one of alternative motion model parameters, alternative interpolation filter parameters, alternative intensity compensation parameters, and alternative denoising filter parameters is estimated from a previous original picture to be encoded.

7. The method as recited in claim 5 , wherein the at least one of alternative motion model parameters, alternative interpolation filter parameters, alternative intensity compensation parameters, and alternative denoising filter parameters is estimated from a previous original picture to be encoded and the previously processed reference picture.

8. The method as recited in claim 1 , wherein the processed reference picture is used as a reference for subsequent motion estimation in the iteration.

9. The method as recited in claim 1 , wherein the motion estimation comprises at least one of:

a block-based motion estimation;

a frequency correlation motion estimation;

a phase correlation motion estimation;

a pixel-level motion estimation; or

a gradient descent estimation.

10. The method as recited in claim 1 , wherein the at least one region of the reconstructed reference picture is formed by an overlap of at least a first region and a second region among the plurality of regions of the reconstructed reference picture, and wherein the applying a plurality of motion models to the at least one region of the reconstructed reference picture comprises:

applying a first motion model on the first region to generate a first set of motion model parameters; and

applying a second motion model on the second region to generate a second set of motion model parameters, thus applying the first and second motion models to the at least one region of the reconstructed reference picture,

wherein the motion model parameters associated with the at least one region is obtained by an average or a weighted average of the first set of motion model parameters with the second set of motion model parameters.

11. The method as recited in claim 2 , wherein at least one of the intensity compensation parameters is derived based only on luma components.

12. The method as recited in claim 2 , wherein a first intensity compensation parameter is derived based only on a luma component and a second intensity compensation parameter is derived based on the first intensity compensation parameter.

13. A non-transitory computer readable storage medium that tangibly stores encoded instructions, which when executing over at least one processor, causes, programs, controls, or configures the processor to perform, the method as recited in claim 1 .

14. A video encoder, comprising:

at least one processor; and

a non-transitory computer readable storage medium comprising instructions that are stored therewith, which when executing over the at least one processor, causes, programs, controls, or configures the processor to:

perform the method as recited in claim 1 .