NAG CL Interface
c09ecc (dim2_​multi_​fwd)

Settings help

CL Name Style:


1 Purpose

c09ecc computes the two-dimensional multi-level discrete wavelet transform (DWT). The initialization function c09abc must be called first to set up the DWT options.

2 Specification

#include <nag.h>
void  c09ecc (Integer m, Integer n, const double a[], Integer lda, Integer lenc, double c[], Integer nwl, Integer dwtlvm[], Integer dwtlvn[], Integer icomm[], NagError *fail)
The function may be called by the names: c09ecc, nag_wav_dim2_multi_fwd or nag_mldwt_2d.

3 Description

c09ecc computes the multi-level DWT of two-dimensional data. For a given wavelet and end extension method, c09ecc will compute a multi-level transform of a matrix A, using a specified number, nfwd, of levels. The number of levels specified, nfwd, must be no more than the value lmax returned in nwlmax by the initialization function c09abc for the given problem. The transform is returned as a set of coefficients for the different levels (packed into a single array) and a representation of the multi-level structure.
The notation used here assigns level 0 to the input matrix, A. Level 1 consists of the first set of coefficients computed: the vertical (v1), horizontal (h1) and diagonal (d1) coefficients are stored at this level while the approximation (a1) coefficients are used as the input to a repeat of the wavelet transform at the next level. This process is continued until, at level nfwd, all four types of coefficients are stored. The output array, C, stores these sets of coefficients in reverse order, starting with anfwd followed by vnfwd , hnfwd , dnfwd , v nfwd-1 , h nfwd-1 , d nfwd-1 , , v1 , h1 , d1 .

4 References

None.

5 Arguments

1: m Integer Input
On entry: number of rows, m, of data matrix A.
Constraint: this must be the same as the value m passed to the initialization function c09abc.
2: n Integer Input
On entry: number of columns, n, of data matrix A.
Constraint: this must be the same as the value n passed to the initialization function c09abc.
3: a[lda×n] const double Input
Note: the (i,j)th element of the matrix A is stored in a[(j-1)×lda+i-1].
On entry: the m×n data matrix A.
4: lda Integer Input
On entry: the stride separating matrix row elements in the array a.
Constraint: ldam.
5: lenc Integer Input
On entry: the dimension of the array c. c must be large enough to contain, nct, wavelet coefficients. The maximum value of nct is returned in nwct by the call to the initialization function c09abc and corresponds to the DWT being continued for the maximum number of levels possible for the given data set. When the number of levels, nfwd, is chosen to be less than the maximum, lmax, then nct is correspondingly smaller and lenc can be reduced by noting that the vertical, horizontal and diagonal coefficients are stored at every level and that in addition the approximation coefficients are stored for the final level only. The number of coefficients stored at each level is given by 3× m¯/2 × n¯/2 for mode=Nag_Periodic in c09abc and 3× (m¯+nf-1)/2 × (n¯+nf-1)/2 for mode=Nag_HalfPointSymmetric, Nag_WholePointSymmetric or Nag_ZeroPadded, where the input data is of dimension m¯×n¯ at that level and nf is the filter length nf provided by the call to c09abc. At the final level the storage is 4/3 times this value to contain the set of approximation coefficients.
Constraint: lencnct, where nct is the total number of coefficients that correspond to a transform with nwl levels.
6: c[lenc] double Output
On exit: the coefficients of the discrete wavelet transform. If you need to access or modify the approximation coefficients or any specific set of detail coefficients then the use of c09eyc or c09ezc is recommended. For completeness the following description provides details of precisely how the coefficient are stored in c but this information should only be required in rare cases.
Let q(i) denote the number of coefficients (of each type) at level i, for i=1,2,,nfwd, such that q(i) = dwtlvm[nfwd-i] × dwtlvn[nfwd-i] . Then, letting k1=q(nfwd) and kj+1=kj+q(nfwd-j/3+1), for j=1,2,,3nfwd, the coefficients are stored in c as follows:
c[i-1], for i=1,2,,k1
Contains the level nfwd approximation coefficients, anfwd.
c[i-1], for i=kj+1,,kj+1
Contains the level nfwd-j/3+1 vertical, horizontal and diagonal coefficients. These are:
  • vertical coefficients if j mod 3=1;
  • horizontal coefficients if j mod 3=2;
  • diagonal coefficients if j mod 3=0,
for j=1,,3nfwd.
7: nwl Integer Input
On entry: the number of levels, nfwd, in the multi-level resolution to be performed.
Constraint: 1nwllmax, where lmax is the value returned in nwlmax (the maximum number of levels) by the call to the initialization function c09abc.
8: dwtlvm[nwl] Integer Output
On exit: the number of coefficients in the first dimension for each coefficient type at each level. dwtlvm[i-1] contains the number of coefficients in the first dimension (for each coefficient type computed) at the (nfwd-i+1)th level of resolution, for i=1,2,,nfwd. Thus for the first nfwd-1 levels of resolution, dwtlvm[nfwd-i] is the size of the first dimension of the matrices of vertical, horizontal and diagonal coefficients computed at this level; for the final level of resolution, dwtlvm[0] is the size of the first dimension of the matrices of approximation, vertical, horizontal and diagonal coefficients computed.
9: dwtlvn[nwl] Integer Output
On exit: the number of coefficients in the second dimension for each coefficient type at each level. dwtlvn[i-1] contains the number of coefficients in the second dimension (for each coefficient type computed) at the (nfwd-i+1)th level of resolution, for i=1,2,,nfwd. Thus for the first nfwd-1 levels of resolution, dwtlvn[nfwd-i] is the size of the second dimension of the matrices of vertical, horizontal and diagonal coefficients computed at this level; for the final level of resolution, dwtlvn[0] is the size of the second dimension of the matrices of approximation, vertical, horizontal and diagonal coefficients computed.
10: icomm[180] Integer Communication Array
On entry: contains details of the discrete wavelet transform and the problem dimension as setup in the call to the initialization function c09abc.
On exit: contains additional information on the computed transform.
11: fail NagError * Input/Output
The NAG error argument (see Section 7 in the Introduction to the NAG Library CL Interface).

6 Error Indicators and Warnings

NE_ALLOC_FAIL
Dynamic memory allocation failed.
See Section 3.1.2 in the Introduction to the NAG Library CL Interface for further information.
NE_BAD_PARAM
On entry, argument value had an illegal value.
NE_INITIALIZATION
Either the initialization function has not been called first or icomm has been corrupted.
Either the initialization function was called with wtrans=Nag_SingleLevel or icomm has been corrupted.
NE_INT
On entry, m=value.
Constraint: m=value, the value of m on initialization (see c09abc).
On entry, n=value.
Constraint: n=value, the value of n on initialization (see c09abc).
On entry, nwl=value.
Constraint: nwl1.
NE_INT_2
On entry, lda=value and m=value.
Constraint: ldam.
On entry, lenc=value.
Constraint: lencvalue, the total number of coefficents to be generated.
On entry, nwl=value and nwlmax=value in c09abc.
Constraint: nwlnwlmax in c09abc.
NE_INTERNAL_ERROR
An internal error has occurred in this function. Check the function call and any array sizes. If the call is correct then please contact NAG for assistance.
See Section 7.5 in the Introduction to the NAG Library CL Interface for further information.
NE_NO_LICENCE
Your licence key may have expired or may not have been installed correctly.
See Section 8 in the Introduction to the NAG Library CL Interface for further information.

7 Accuracy

The accuracy of the wavelet transform depends only on the floating-point operations used in the convolution and downsampling and should thus be close to machine precision.

8 Parallelism and Performance

Background information to multithreading can be found in the Multithreading documentation.
c09ecc is not threaded in any implementation.

9 Further Comments

The wavelet coefficients at each level can be extracted from the output array c using the information contained in dwtlvm and dwtlvn on exit (see the descriptions of c, dwtlvm and dwtlvn in Section 5). For example, given an input data set, A, denoising can be carried out by applying a thresholding operation to the detail (vertical, horizontal and diagonal) coefficients at every level. The elements c[k1] to c[knfwd+1-1], as described in Section 5, contain the detail coefficients, c^ij, for i=nfwd,nfwd-1,,1 and j=1,2,,3q(i), where q(i) is the number of each type of coefficient at level i and c^ij=cij+σεij and σεij is the transformed noise term. If some threshold parameter α is chosen, a simple hard thresholding rule can be applied as
c¯ ij = { 0, if ​ |c^ij| α c^ij , if ​ |c^ij| > α,  
taking c¯ij to be an approximation to the required detail coefficient without noise, cij. The resulting coefficients can then be used as input to c09edc in order to reconstruct the denoised signal.
See the references given in the introduction to this chapter for a more complete account of wavelet denoising and other applications.

10 Example

This example performs a multi-level resolution transform of a dataset using the Daubechies wavelet (see wavnam=Nag_Daubechies2 in c09abc) using half-point symmetric end extensions, the maximum possible number of levels of resolution, where the number of coefficients in each level and the coefficients themselves are not changed. The original dataset is then reconstructed using c09edc.

10.1 Program Text

Program Text (c09ecce.c)

10.2 Program Data

Program Data (c09ecce.d)

10.3 Program Results

Program Results (c09ecce.r)