Method of coding and decoding image
View Patent ↗A method of simplifying the arithmetic operation in a global motion compensation process approximates the motion vector field of the whole image without using many parameters. Motion vectors in the global motion compensation are found by the interpolation and/or extrapolation of the motion vectors of a plurality of representative points 602, 603 and 604 having particular features in the spatial distance thereof. Since the shift operation can be substituted for the division for synthesizing a predicted image of global motion compensation, the processing using a computer or a dedicated hardware is simplified.
1. A video decoding method comprising:
storing a decoded image of a reference frame; and
synthesizing a predicted image of a present frame by using said decoded image and the information related to said present frame,
wherein said predicted image synthesizing comprises:
calculating motion vectors of 3 representative points having coordinates (i,j), (i+p,j), and (i,j+p) using motion vectors of corner points of said predicted image having coordinates (0,0), (r,0), and (0,s), where r, s, i, and j are integers, p and q are integer powers of 2, p is greater than or equal to r, p/2 is less than r, q is greater than or equal to 5, q/2 is less than s, and the sampling intervals of pixels are 1 in both horizontal and vertical directions;
calculating the motion vector of each pixel in said predicted image from said motion vectors of said representative points; and
synthesizing said predicted image from said motion vector of each pixel and said decoded image.
2. A video decoding method according to claim 1 ,
wherein said calculating motion vectors of 3 representative points and said calculating the motion vector of each pixel performs liner interpolation or extrapolation.
3. A video decoding method comprising:
storing a decoded image of a reference frame; and
synthesizing a predicted image of a present frame by using said decoded image and the information related to said present frame,
wherein said predicted image synthesizing comprises:
calculating motion vectors of 3 representative points having coordinates (i,j), (i+p,j), and (i,j+p) using motion vectors of corner points of said predicted image having coordinates (0,0), (r,0), and (0,s), where r, a, i, and j are integers, p and q are integer powers of 2, p is greater than or equal to r, p/2 is less than r, q is greater than or equal to s, q/2 is less than s, and the sampling intervals of pixels are 1 in both horizontal and vertical directions;
calculating the motion vector of each pixel in said predicted image from said motion vectors of said representative points; and
synthesizing said predicted image from said motion vector of each pixel and said decoded image, wherein:
the horizontal and vertical components of said motion vectors of said corner points are integer multiples of 1/n;
the horizontal and vertical components of said motion vector of each pixel in said predicted image are integer multiples of 1/m;
the horizontal and vertical components of said motion vectors of said representative points are integer multiples of 1/k;
in said calculating motion vectors of 3 representative points, said motion vectors of 3 representative points are calculated using equations:
u ′( x,y )=(( u 00 rs +( u 01 − u 00 )) xs +( u 02 − u 00 ) yr ) k )//( rsn ),
v ′( x,y )=(( v 00 rs +( v 01 − v 00 )) xs +( v 02 − v 00 ) yr ) k )//( rsn ),
u 0 = u ′( i,j ), v 0 + v ′( i,j ), u 1 = u ′( i+p,j ), v 1 = v ′( i+p,j ),
u 2 = u ′( i,j+q ), and v 2 = v ′( i,j+q ); and
in said calculating the motion vector of each pixel in said predicted image, a motion vector of a pixel in said predicted image is calculated using equations:
u ( x, y )=(( u 0 pq +( u 1 − u 0 )) xq +( u 2 − u 0 ) yp ) m )//( pqk ), and
v ( x, y )−(( v 0 pq +( v 1 − v 0 )) xq +( v 2 − v 0 ) yp ) m )//( pqk ),
where (u 00 ,v 00 ), (u 01 ,v 01 ), and (u 02 ,v 02 ) are n times said motion vectors of said corner points having coordinates ( 0 , 0 ), (r, 0 ), and ( 0 ,s), (u(x,y), v(x,y)) is m times the horizontal and vertical components of the motion vector of a pixel having coordinates (x,y) in said predicted image, (u 0 ,v 0 ), (u 1 ,v 1 ), and (u 2 ,v 2 ) are k times said motion vectors of said representative points having coordinates (i,j), (i+p,j), and (i,j+p), u 00 , v 00 , u 01 , v 01 , u 02 , v 02 , u(x,y), v(x,y), u 0 , v 0 , u 1 , v 1 , u 2 , and v 2 are integers, k, m, and n are integer powers of 2, and “///” and “//” represent integer divisions that round the guotient of ordinary division into an adjacent integer when said quotient of said ordinary division is not an integer, and their priority as an operator is the same as that of ordinary multiplication and division.
4. A video decoding method according to claim 3 ,
wherein said “///” and “//” round the quotient of ordinary division, when the result of said quotient of ordinary division is the sum of ½ and an integer, either into the nearest integer:
(1) away from 0; (2) toward 0; (3) away from 0 when the dividend is negative and toward 0 when the dividend is positive; or (4) away from 0 when the dividend is positive and toward 0 when the dividend is negative.