Signaling of global motion relative to available reference frames
A decoder includes circuitry configured to receive a bitstream, extract a header including a list of reference frames available for global motion compensation, determine, using the header, a global motion model for a current block, the global motion relative to a reference frame contained in the list of reference frames, and decode the current block using the global motion model. Related apparatus, systems, techniques and articles are also described.
1. A decoder, the decoder comprising circuitry configured to:
receive a bitstream including a coded current picture, the coded current picture including multiple coding tree units, each coding tree unit including one or more coding units;
construct, using the bitstream, a list of stored previously decoded pictures for use as reference pictures for global motion compensation prediction for the coded current picture, wherein none of the stored reference pictures is obtained by pixel interpolation of a decoded picture using mapping parameters;
for a current coding unit in the current picture, detect from header information in the bitstream for the current coding unit, a flag indicating global motion compensation prediction is enabled for the current coding unit, thereby indicating that each motion vector of the current coding unit is determined relative to a reference picture in the list based on a motion vector of an adjacent coding unit;
if global motion compensation prediction is enabled for the current coding unit, determine based on the header of the current coding unit if a motion model for the current coding unit is translational motion, 4-parameter affine motion or 6-parameter affine motion;
determine at least a motion vector based on an adjacent coding unit, wherein:
if the motion model of the current coding unit is translational motion, determining the at least a motion vector further comprises determining a translational motion vector from the motion vector of an adjacent coding unit, and decode the current coding unit using the translational motion vector;
if the motion model of the current coding unit is 4-parameter affine motion, determining the at least a motion vector further comprises determining two control point motion vectors from the motion vectors of adjacent coding units, and decode the current coding unit using the two control point motion vectors; and
if the motion model of the current coding unit is 6-parameter affine motion, determining the at least a motion vector further comprises determining three control point motion vectors from the motion vectors of adjacent coding units, and decode the coding unit using the three control point motion vectors.
2. The decoder of claim 1 , where the reference frames are indexed in the list by an order relative to a current frame.
3. The decoder of claim 1 , wherein the header includes a picture parameter set or a sequence parameter set.
4. The decoder of claim 1 , wherein the header includes a flag characterizing whether global motion is present for the current block.
5. The decoder of claim 1 , wherein the global motion model includes a translational motion model.
6. The decoder of claim 1 , wherein the global motion model includes a four parameter affine motion model.
7. The decoder of claim 1 , wherein the global motion model includes a six parameter affine motion model.
8. The decoder of claim 1 , further comprising:
an entropy decoder processor configured to receive the bit stream and decode the bitstream into quantized coefficients;
an inverse quantization and inverse transformation processor configured to process the quantized coefficients including performing an inverse discrete cosine;
a deblocking filter;
a frame buffer; and
an intra prediction processor.
9. The decoder of claim 1 wherein the current picture contains a plurality of 128×128 coding tree units.
10. The decoder of claim 9 wherein one of the coding tree units has a coding unit with a geometric partition.
11. A method, the method comprising:
receiving, by a decoder, a bitstream including a coded current picture, the coded current picture including multiple coding tree units, each coding tree unit including one or more coding units;
constructing, using the bitstream, a list of stored previously decoded pictures for use as reference pictures for global motion compensation prediction for the coded current picture;
for a current coding unit in the current picture, detecting from header information in the bitstream for the current coding unit, a flag indicating global motion compensation prediction is enabled for the current coding unit, thereby indicating that each motion vector of the current coding unit is determined relative to a reference picture in the list based on a motion vector of an adjacent coding unit;
if global motion compensation prediction is enabled for the current coding unit, determining based on the header of the current coding unit if a motion model for the current coding unit is translational motion, 4-parameter affine motion or 6-parameter affine motion;
determining at least a motion vector based on an adjacent coding unit, wherein:
if the motion model of the current coding unit is translational motion, determining the at least a motion vector further comprises determining a translational motion vector from the motion vector of an adjacent coding unit, and decode the current coding unit using the translational motion vector;
if the motion model of the current coding unit is 4-parameter affine motion, determining the at least a motion vector further comprises determining two control point motion vectors from the motion vectors of adjacent coding units, and decode the current coding unit using the two control point motion vectors; and
if the motion model of the current coding unit is 6-parameter affine motion, determining the at least a motion vector further comprises determining three control point motion vectors from the motion vectors of adjacent coding units, and decode the coding unit using the three control point motion vectors.
12. The method of claim 11 , where the reference frames are indexed in the list by an order relative to a current frame.
13. The method of claim 11 , wherein the header includes a picture parameter set (PPS) or a sequence parameter set (SPS).
14. The method of claim 11 , wherein the header includes a flag characterizing whether global motion is present for the current block.
15. The method of claim 11 , wherein the global motion model includes a translational motion model.
16. The method of claim 11 , wherein the global motion model includes a four parameter affine motion model.
17. The method of claim 11 , wherein the global motion model includes a six parameter affine motion model.
18. The method of claim 11 , the decoder further comprising:
an entropy decoder processor configured to receive the bit stream and decode the bitstream into quantized coefficients;
an inverse quantization and inverse transformation processor configured to process the quantized coefficients including performing an inverse discrete cosine;
a deblocking filter;
a frame buffer; and
an intra prediction processor.
19. The method of claim 11 wherein the current picture contains a plurality of 128×128 coding tree units.
20. The method of claim 19 wherein one of the coding tree units has a coding unit with a geometric partition.