CODING AND DECODING METHODS FOR A PICTURE SIGNAL USING OVERLAPPING BLOCKS HAVING VARIABLE BLOCK SIZES AND APPARATUSES THEREFOR

摘要

A method of coding and decoding a bidimensional signal, comprising the steps of: at a coding side, (a) dividing a bidimensional input signal into rectangular blocks of a plurality of sizes; (b) providing window functions hAM(m) and hBN(n) on a direction and block size basis; (c) calculating, assuming that a block of interest has a M X N size, transformed signals Xp(k,i) from input signals xp(m, n) (m = 0, 1, ....., 2M-1; n = 0, 1, ....., 2N-1) by a transform equation: 2M-1 2N-1 Xp(k, i) = .SIGMA. .SIGMA. xp(m,n) hAM(m) hBN(n) m=0 n=0 x cos{(2m+M+1) (2k+1) .pi./(4M x cos{(2n+N+1) (2i+1) .pi./(4N (k = 0, 1, ..., M-1; i = 0, 1, ..., N-1) (d) transmitting block division information and said transformed signals to a decoding side; at said decoding side, (e) providing window functions fAM(m) and fBN(n) corresponding to said window functions hAM(m) and hBN(n), respectively; (f) calculating, when a block size of said transformed signals sent from said coding side is M X N, inverse-transformed signals yp(m,n) of 2M X 2N points from said transformed signals by using a transform equation: M-1 N-1 Yp(m, n) = .SIGMA. .SIGMA. Xp(k,i) fAM(m) fBN(n) / (MN) k=0 i=0 x cos{(2m+M+1) (2k+1) .pi./(4M x cos{(2n+N+1) (2i+1) .pi./(4N (m = 0, 1, ..., 2M-1; n = 0, 1, ..., 2N-1) (g) arranging said inverse-transformed signals according to said block division information sent from said coding side; and, (h) adding, in a portion where a plurality of blocks overlap, said inverse-transformed signals to produce a decoded signal.