signal processing - algorithm for antialiasing an image with sync filter -
i have been reading ways antialiasing , since not processed in real time antialising signal processing seems ideal especial against artifacts.
however have read not mention step turning a bitmap image signal , again,so i'm looking algorithm or code examples demonstrate that.
the usual way things handled apply filter independently in both x , y directions. way, overall filter g(x,y) = f(x) * f(y)
.
in kind of situation, g(x,y)
called separable filter, , advantage that, applying x- , y- filters separately, straightforward filter convolution takes o(x y f) time, x , y dimensions of image, , f support width of filter f()
. arbitrary nonseparable filter of same size (which have o(f^2) samples) requires o(x y f^2) time...
if want apply full sinc() (== sin(x)/x
) filter image, unlimited support of sinc() function make straightforward convolution slow. in case, faster 2d fft of image, filter in frequency domain, , transform back.
in practice, though, people use windowing or other modification, finite filter can practically applied in spatial domain.
Comments
Post a Comment