Motion vector prediction method based on affine motion model and device

A video data coding device comprising:

• a non-transitory memory storage, configured to store video data in a form of a bitstream;
• and
• a video coder coupled to the memory storage, and configured for:
• obtaining one spatial reference block of a to-be-processed picture block in the video data;
• determining a plurality of preset subblock locations of the spatial reference block;
• obtaining motion vectors corresponding to preset pixel locations of the to-be-processed picture block that are extrapolated from motion vectors corresponding to the preset subblock locations of the spatial reference block;
• obtaining motion vectors corresponding to a plurality of subblock locations of the to-be-processed picture block that are interpolated from the motion vectors corresponding to the preset pixel locations of the to-be-processed picture block;
• performing motion compensation for a plurality of subblocks of the to-be-processed picture block based on the motion vectors corresponding to the plurality of subblock locations of the to-be-processed picture block, to obtain predicted pixel values of the plurality of subblocks of the to-be-processed picture block;
• and
• wherein the affine motion model is one of the following (a), (b) or (c):
• (a) a 4-parameter affine motion model, and the plurality of preset subblock locations comprise a first preset location (x4+M/2, Y4+N/2) and a second preset location (x4+M/2+P, Y4+N/2), wherein x4 is a horizontal coordinate of a top-left pixel of the spatial reference block, y4 is a vertical coordinate of the top-left pixel of the spatial reference block, M is a subblock width, N is a subblock height, P is 2 raised to the power of K, K is a nonnegative integer, K is less than U, and U is a width of the spatial reference block;
• (b) a 4-parameter affine motion model, and the plurality of preset subblock locations comprise a first preset location (x4+M/2, y4+N/2) and a third preset location (x4+M/2, Y4+N/2+Q), wherein x4 is a horizontal coordinate of a top-left pixel of the spatial reference block, y4 is a vertical coordinate of the top-left pixel of the spatial reference block, M is a subblock width, N is a subblock height, Q is 2 raised to the power of R, R is a nonnegative integer, Q is less than V, and V is a height of the spatial reference block:
• or
• (c) a 6-parameter affine motion model, and the plurality of preset subblock locations comprise a first preset location (x4+M/2, y4+N/2), a second preset location (x4+M/2+P, Y4+N/2), and a third preset location (x4+M/2, Y4+N/2+Q), wherein x4 is a horizontal coordinate of a top-left pixel of the spatial reference block, y4 is a vertical coordinate of the top-left pixel of the spatial reference block, M is a subblock width, N is a subblock height, P is 2 raised to the power of K, Q is 2 raised to the power of R, K and R are nonnegative integers, P is less than U, Q is less than V, U is a width of the spatial reference block, and V is a height of the spatial reference block.

A non-transitory computer-readable medium having processor-executable instructions stored thereon which, when executed by a processor of a computer device, causes the computer device to perform operations including:

• obtaining one spatial reference block of a to-be-processed picture block in the video data;
• determining a plurality of preset subblock locations of the spatial reference block;
• obtaining motion vectors corresponding to preset pixel locations of the to-be-processed picture block that are extrapolated from motion vectors corresponding to the preset subblock locations of the spatial reference block;
• obtaining motion vectors corresponding to a plurality of subblock locations of the to-be-processed picture block that are interpolated from the motion vectors corresponding to the preset pixel locations of the to-be-processed picture block;
• performing motion compensation for a plurality of subblocks of the to-be-processed picture block based on the motion vectors corresponding to the plurality of subblock locations of the to-be-processed picture block, to obtain predicted pixel values of the plurality of subblocks of the to-be-processed picture block;
• and
• wherein the affine motion model is one of the following (a), (b) or (c):
• (d) 4-parameter affine motion model, and the plurality of preset subblock locations comprise a first preset location (x4+M/2, y4+N/2) and a second preset location (x4+M/2+P, y4+N/2), wherein x4 is a horizontal coordinate of a top-left pixel of the spatial reference block, y4 is a vertical coordinate of the top-left pixel of the spatial reference block, M is a subblock width, N is a subblock height, P is 2 raised to the power of K, K is a nonnegative integer, K is less than U, and U is a width of the spatial reference block;
• (e) a 4-parameter affine motion model, and the plurality of preset subblock locations comprise a first preset location (x4+M/2, Y4+N/2) and a third preset location (x4+M/2, Y4+N/2+Q), wherein x4 is a horizontal coordinate of a top-left pixel of the spatial reference block, y4 is a vertical coordinate of the top-left pixel of the spatial reference block, M is a subblock width, N is a subblock height, Q is 2 raised to the power of R, R is a nonnegative integer, Q is less than V, and V is a height of the spatial reference block:
• or
• a 6-parameter affine motion model, and the plurality of preset subblock locations comprise a first preset location (x4+M/2, Y4+N/2), a second preset location (x4+M/2+P, y4+N/2), and a third preset location (x4+M/2, y4+N/2+Q), wherein x4 is a horizontal coordinate of a top-left pixel of the spatial reference block, y4 is a vertical coordinate of the top-left pixel of the spatial reference block, M is a subblock width, N is a subblock height, P is 2 raised to the power of K, Q is 2 raised to the power of R, K and R are nonnegative integers, P is less than U, Q is less than V, U is a width of the spatial reference block, and V is a height of the spatial reference block.