SYSTEMS AND METHODS FOR ENCODING AND DECODING VIDEO WITH REGIONS OF GLOBAL MOTION
A decoder includes circuitry configured to receive a bitstream, extract a header, determine, using the header, a global motion model, and decode a current block of a current frame using the global motion model. Related apparatus, systems, techniques and articles are also described.
1 . A decoder comprising circuitry configured to:
receive a bitstream, the bitstream comprising a coded picture, the coded picture comprising a first region with global motion and a second region with local motion bordering the first region, the first region with global motion comprising a first plurality of coding units with most of the coding units of the picture, the second region with local motion comprising a second plurality of coding units of the picture, the first plurality of coding units all having the same affine motion motion model,
for each block in the first region,
determine, a motion model, the motion model being the same for all of the blocks in the first region and being one of 4-parameter affine motion or 6-parameter affine motion, the parameters of the motion model for each block in the first region being determined from at least one motion vector signaled in the bitstream; and
decode each block in the first region using the parameters of the motion model to reconstruct the global motion in the first region; and
for each block in the second region, decode each block using motion information determined individually for each block to reconstruct the local motion in the second region.
2 . The decoder of claim 1 , wherein the motion model in the 1 first region is 4-parameter affine motion.
3 . The decoder of claim 1 , wherein the motion model in the first region is 6-parameter affine motion.
4 . The decoder of claim 1 , wherein the at least one motion vector is a control point motion vector.
5 . The decoder of claim 1 , wherein the motion model in the first region is 4-parameter affine motion and the at least one motion vector includes two control point motion vectors.
6 . The decoder of claim 1 , wherein the motion model in the first region is 6-parameter affine motion and the at least one motion vector includes three control point motion vectors.
7 . A method of transmitting an encoded bitstream comprising:
receiving a video signal;
generating an encoded bitstream representing the video signal comprising a coded picture, the coded picture having a first region with global motion and a second region with local motion bordering the first region, the first region with global motion comprising a first plurality of coding units with most of the coding units of the picture, the second region with local motion comprising a second plurality of coding units of the picture, the first plurality of coding units all having the same affine motion motion model;
the bitstream being further configured to be decodable by a decoding method comprising:
for each block in the first region determine a motion model, the motion model being the same for all of the blocks in the first region and being one of 4-parameter affine motion or 6-parameter affine motion, the parameters of the motion model for each block in the first region being determined from at least one motion vector signaled in the bitstream; and
decode each block in the first region using the parameters of the motion model to reconstruct the global motion in the first region; and
for each block in the second region, decode each block using motion information determined individually for each block to reconstruct the local motion in the second region; and
transmit the encoded bitstream over a channel to a decoder.
8 . The method of claim 7 , wherein the motion model in the first region is 4-parameter affine motion.
9 . The method of claim 7 , wherein the motion model in the first region is 6-parameter affine motion.
10 . The decoder of claim 7 , wherein the at least one motion vector is a control point motion vector.
11 . The decoder of claim 7 , wherein the motion model in the first region is 4-parameter affine motion and the at least one motion vector includes two control point motion vectors.
12 . The decoder of claim 7 , wherein the motion model in the first region is 6-parameter affine motion and the at least one motion vector includes three control point motion vectors.
13 . An encoder comprising circuitry configured to:
receive a video signal;
generate an encoded bitstream representing the video signal, the encoded bitstream
including a coded picture, the coded picture having a first region with global motion and a second region with local motion bordering the first region, the first region with global motion comprising a first plurality of coding units with most of the coding units of the picture, the second region with local motion comprising a second plurality of coding units of the picture, the first plurality of coding units all having the same affine motion motion model;
the bitstream being further configured by the encoder to be decodable by a decoding method comprising:
for each block in the first region determine a motion model, the motion model being the same for all of the blocks in the first region and being one of 4-parameter affine motion or 6-parameter affine motion, the parameters of the motion model for each block in the first region being determined from at least one motion vector signaled in the bitstream; and
decode each block in the first region using the parameters of the motion model to reconstruct the global motion in the first region; and
for each block in the second region, decode each block using motion information determined individually for each block to reconstruct the local motion in the second region; and
transmit the encoded bitstream over a channel.
14 . The encoder of claim 13 , wherein the motion model in the first region is 4-parameter affine motion.
15 . The encoder of claim 13 , wherein the motion model in the first region is 6-parameter affine motion.
16 . The encoder of claim 13 , wherein the at least one motion vector is a control point motion vector.
17 . The encoder of claim 13 , wherein the motion model in the first region is 4-parameter affine motion and the at least one motion vector includes two control point motion vectors.
18 . The encoder of claim 13 , wherein the motion model in the first region is 6-parameter affine motion and the at least one motion vector includes three control point motion vectors.