One can obtain either exact realizations or useful approximations of l
inear systems or matrix-vector products that arise in many different a
pplications by implementing them in the form of multistage or multicha
nnel fractional Fourier-domain filters, resulting in space-bandwidth-e
fficient systems with acceptable decreases in accuracy. Varying the nu
mber and the configuration of filters enables one to trade off between
accuracy and efficiency in a flexible manner. The proposed scheme con
stitutes a systematic way of exploiting the regularity or structure of
a given linear system or matrix, even when that structure is not read
ily apparent. (C) 1998 Optical Society of America.