wxfrm.3

VC小波应用程序

.TH WXFRM 3wvlt "1 May 1995"
.SH NAME
wxfrm_[df]a[1n]d \- perform a wavelet transform
.SH SYNOPSIS
.ft B
.nf
#include <wvlt.h>
.sp .5
void wxfrm_da1d(a, nA, isFwd, wfltr, aXf);
\s-1double\s0 *a, *aXf;
\s-1int\s0 nA;
\s-1bool\s0 isFwd;
\s-1waveletfilter\s0 *wfltr;
.sp .5
void wxfrm_fa1d(a, nA, isFwd, wfltr, aXf);
\s-1float\s0 *a, *aXf;
\s-1int\s0 nA;
\s-1bool\s0 isFwd;
\s-1waveletfilter\s0 *wfltr;
.sp .5
void wxfrm_dand(a, nA, nD, isFwd, isStd, wfltr, aXf);
\s-1double\s0 *a, *aXf;
\s-1int\s0 nA[], nD;
\s-1bool\s0 isFwd, isStd;
\s-1waveletfilter\s0 *wfltr;
.sp .5
void wxfrm_fand(a, nA, nD, isFwd, isStd, wfltr, aXf);
\s-1float\s0 *a, *aXf;
\s-1int\s0 nA[], nD;
\s-1bool\s0 isFwd, isStd;
\s-1waveletfilter\s0 *wfltr;
.ft R
.fi
.SH DESCRIPTION
.LP
These functions, which are part of the UBC Imager Wavelet Library,
perform wavelet transforms on data.
Every function uses a particular wavelet filter (or basis), specified by the
.I wfltr
pointer.
The possible choices are described on the
.I wfltr(3wvlt)
man page.
.LP
.I nA
is the number of elements in
.IR a[] .
It must be a power of two.
.LP
A non-zero (i.e. TRUE)
.I isFwd
performs a forward transform and a zero
(i.e. FALSE)
.I isFwd
performs an inverse transform.
.LP
The individual functions are:
.TP 20
.B wxfrm_da1d()
performs a 1-dimensional transformation on the data in
.I a[]
and leaves the result in
.IR aXf[] .
.TP 20
.B wxfrm_fa1d()
is identical to
.IR wxfrm_da1d() ,
except that it acts on
.B float
data.
.TP 20
.B wxfrm_dand()
performs an n-dimensional wavelet transform on the data in
.I a[]
and leaves the result in
.IR aXf[] .
Note that even though the transform is n-dimensional,
.I a[]
itself is not.
Instead, the
.I nA[]
array is used to specify the multidimensionality.
The transform of an
.I m
row by
.I n
column array would be specified by setting
.I nD
to 2 and
.I nA
to { 
.IR m ,
.I n
}.
This is consistent with declaring a with
the C multidimensional array declaration
.I "double a[m][n]"
(with
.I n
and
.I m
constants)
and passing
.I &a[0][0]
in place of
.IR a[] .
.I isStd
refers to whether or not you wish to use the standard (Cartesian)
or non-standard (Beylkin-Coifman-Rokhlin) multidimensional wavelet bases.
.TP 20
.B wxfrm_fand()
is identical to
.IR wxfrm_dand() ,
except that it acts on
.B float
data.
.SH DIAGNOSTICS
If
.I nA
is less than 2, these functions return quietly.
.SH FILES
.TP 20
.B libwvlt.a