Constraint for template matching in decoder side motion derivation and refinement

1. A method for video decoding in a decoder, comprising:

receiving a coded bitstream;

determining a template of a first block in a picture in the received coded bitstream in response to a determination that decoder-side motion vector derivation (DMVD) is applied to the first block for reconstruction;

determining a first region of the template that overlaps with a second block that is decoded prior to the first block;

determining a second region of the template by subtracting the first region of the template that overlaps with the second block from the template; and

deriving motion information of the first block using template matching based on reconstruction samples of the second region,

wherein the first region and the second region of the template are adjacent to a same side of the first block.

2. The method of claim 1 , wherein

the template matching does not use reconstruction samples of the first region to derive the motion information of the first block.

3. The method of claim 1 , further comprising:

reconstructing reconstruction samples of the first region of the template using padding or extrapolation.

4. The method of claim 1 , further comprising:

receiving an indicator in a sequence parameter set, a picture parameter set, or a slice header in the coded bitstream, wherein the indicator indicates whether a candidate motion vector generated using the template matching is constructed based on reconstruction samples of both of the first region of the template and the second region of the template.

5. The method of claim 4 , wherein the indicator indicates that the candidate motion vector generated using the template matching is constructed based on the reconstruction samples of both of the first region of the template and the second region of the template in response to a determination that the second block is intra-coded.

6. The method of claim 4 , wherein the indicator indicates that the candidate motion vector generated using the template matching is constructed based on the reconstruction samples of both of the first region of the template and the second region of the template in response to a determination that prediction information of the second block is used for deriving the motion information of the first block.

7. The method of claim 6 , wherein the second block is inter-coded and is not coded with a DMVD mode.

8. The method of claim 1 , wherein a candidate motion vector generated using the template matching is constructed based on reconstruction samples of both of the first region of the template and the second region of the template in response to a determination that constrained intra prediction is disabled.

9. The method of claim 1 , wherein the a candidate motion vector generated using the template matching is constructed based on reconstruction samples of both of the first region of the template and the second region of the template in response to a determination that (i) a left portion of the template is in a first coding tree unit, slice, or tile, (ii) an upper portion of the template is in a second coding tree unit, slice or tile, and (iii) the first block is in a third coding three unit, slice or tile.

10. The method of claim 1 , wherein a cost of performing template matching for the first block is calculated as:

Cost_TM= w _left*Cost_TM_left+ w _above*Cost_TM above,

wherein Cost_TM_left is a cost using only a left portion of the template in the template matching, Cost_TM_above is a cost using only an upper portion of the template in the template matching, w_left is a first weight associated with Cost_TM_left, and w_above is a second weight associated with Cost_TM_above.

11. An apparatus, comprising:

processing circuitry configured to

receive a coded bitstream,

determine a template of a first block in a picture in the received coded bitstream in response to a determination that decoder-side motion vector derivation (DMVD) is applied to the first block for reconstruction,

determine a first region of the template that overlaps with a second block that is decoded prior to the first block,

determine a second region of the template by subtracting the first region of the template that overlaps with the second block from the template, and

derive motion information of the first block using template matching based on reconstruction samples of the second region,

wherein the first region and the second region of the template are adjacent to a same side of the first block.

12. The apparatus of claim 11 , wherein

the template matching does not use reconstruction samples of the first region to derive the motion information of the first block.

13. The apparatus of claim 11 , wherein the processing circuitry is configured to

reconstruct reconstruction samples of the first region of the template using padding or extrapolation.

14. The apparatus of claim 11 , wherein the processing circuitry is configured to

receive an indicator in a sequence parameter set, a picture parameter set, or a slice header in the coded bitstream, and the indicator indicates whether a candidate motion vector generated using the template matching is constructed based on reconstruction samples of both of the first region of the template and the second region of the template.

15. The apparatus of claim 14 , wherein the indicator indicates that the candidate motion vector generated using the template matching is constructed based on the reconstruction samples of both of the first region of the template and the second region of the template in response to a determination that the second block is intra-coded.

16. The apparatus of claim 14 , wherein the indicator indicates that the candidate motion vector generated using the template matching is constructed based on the reconstruction samples of both of the first region of the template and the second region of the template in response to a determination that prediction information of the second block is used for deriving the motion information of the first block.

17. The apparatus of claim 16 , wherein the second block is inter-coded and is not coded with a DMVD mode.

18. The apparatus of claim 11 , wherein a candidate motion vector generated using the template matching is constructed based on reconstruction samples of both of the first region of the template and the second region of the template in response to a determination that constrained intra prediction is disabled.

19. The apparatus of claim 11 , wherein a candidate motion vector generated using the template matching is constructed based on reconstruction samples of both of the first region of the template and the second region of the template in response to a determination that (i) a left portion of the template is in a first coding tree unit, slice, or tile, (ii) an upper portion of the template is in a second coding tree unit, slice or tile, and (iii) the first block is in a third coding three unit, slice or tile.

20. A non-transitory computer-readable medium storing instructions which when executed by a computer for video decoding cause the computer to perform a method, the method comprising:

receiving a coded bitstream;

determining a template of a first block in a picture in the received coded bitstream in response to a determination that decoder-side motion vector derivation (DMVD) is applied to the first block for reconstruction;

determining a first region of the template that overlaps with a second block that is decoded prior to the first block;

determining a second region of the template by subtracting the first region of the template that overlaps with the second block from the template; and

deriving motion information of the first block using template matching based on reconstruction samples of the second region,

wherein the first region and the second region of the template are adjacent to a same side of the first block.