modwtmra

Multiresolution analysis based on MODWT

Syntax

mra = modwtmra(w)

mra = modwtmra(w,wname)

mra = modwtmra(w,Lo,Hi)

mra = modwtmra(___,'reflection')

Description

mra = modwtmra(w) returns the multiresolution analysis (MRA) of the maximal overlap discrete wavelet transform (MODWT) matrix, w. The MODWT matrix, w, is the output of the modwt function. By default, modwtmra assumes that you obtained w using the 'sym4' wavelet with periodic boundary handling.

mra = modwtmra(w,wname) constructs the MRA using the wavelet corresponding to wname. The wname wavelet must be the same wavelet used to obtain the MODWT.

mra = modwtmra(w,Lo,Hi) constructs the MRA using the scaling filter Lo and wavelet filter Hi. The Lo and Hi filters must be the same filters used to obtain the MODWT.

mra = modwtmra(___,'reflection') uses the reflection boundary condition in the construction of the MRA using any of the arguments from previous syntaxes. If you specify 'reflection', modwtmra assumes that the column dimension of w is even and equals twice the length of the original signal.

You must enter the entire character vector 'reflection'. If you added a wavelet named 'reflection' using the wavelet manager, you must rename that wavelet prior to using this option. 'reflection' may be placed in any position in the input argument list after x. By default, modwtmra uses periodic extension at the boundary.

Examples

Perfect Reconstruction with the MODWTMRA

Obtain the MODWTMRA of a simple time-series signal and demonstrate perfect reconstruction.

Create a time-series signal

t = 1:10;
x = sin(2*pi*200*t);

Obtain the MODWT and the MODWTMRA and sum the MODWTMRA rows.

m = modwt(x);
mra = modwtmra(m);
xrec = sum(mra);

Use the maximum of the absolute values to show that the difference between the original signal and the reconstruction is extremely small. The largest absolute value is on the order of $$10^{-25}$$, which demonstrates perfect reconstruction.

max(abs(x-xrec))

ans = 5.5738e-25

MRA Using Non-Default Wavelet

Construct an MRA of an ECG signal down to level four using the db2 wavelet. The data are taken from Percival & Walden (2000), p.125 (data originally provided by William Constantine and Per Reinhall, University of Washington). The sampling frequency for the ECG signal is 180 hertz.

load wecg;
lev = 4;
wtecg = modwt(wecg,'db2',lev);
mra = modwtmra(wtecg,'db2');

Plot the ECG waveform and the MRA.

t = (0:numel(wecg)-1)/180;
subplot(6,1,1)
plot(t,wecg)
for kk = 2:lev+2
    subplot(6,1,kk)
    plot(t,mra(kk-1,:))
end
xlabel('Time (s)')
set(gcf,'Position',[0 0 500 700])

MRA Using the Default Wavelet

Construct a multiresolution analysis for the Southern Oscillation Index data. The sampling period is one day. Plot the level eight details corresponding to a scale of $$2^8$$ days. The details at this scale capture oscillations on a scale of approximately one year.

load soi
wtsoi = modwt(soi);
mrasoi = modwtmra(wtsoi);
plot(mrasoi(8,:))
title('Level 8 Details')

MRA Using Minimum Bandwidth Scaling and Wavelet Filters

Obtain the MRA for the Deutsch Mark - U.S. Dollar exchange rate data using the minimum bandwidth scaling and wavelet filters with four coefficients.

load DM_USD;
Lo = [0.4801755, 0.8372545, 0.2269312, -0.1301477];
Hi = qmf(Lo);
wdm = modwt(DM_USD,Lo,Hi);
mra = modwtmra(wdm,Lo,Hi);

MRA Using Fejér-Korovkin Filters

Load the ECG data.

load wecg

Obtain the MODWT of the signal using the filters associated with the 8-coefficient Fejér-Korovkin filters.

[~,~,Lo,Hi] = wfilters("fk8");
wtecg = modwt(wecg,Lo,Hi);

Obtain the MRA of the signal using the filters.

mra = modwtmra(wtecg,Lo,Hi);

Obtain a second MRA of the signal using the wavelet name. Confirm the multiresolution analyses are equal.

mra2 = modwtmra(wtecg,"fk8");
max(abs(mra(:)-mra2(:)))

ans = 0

MRA Using Reflection Boundary

Obtain the MRA for an ECG signal using 'reflection' boundary handling. The data are taken from Percival & Walden (2000), p.125 (data originally provided by William Constantine and Per Reinhall, University of Washington).

load wecg;
wtecg = modwt(wecg,'reflection');
mra = modwtmra(wtecg,'reflection');

Show that the number of columns in the MRA is equal to the number of elements in the original signal.

isequal(size(mra,2),numel(wecg))

ans = logical
   1

MRA of Multisignal

Load the 23 channel EEG data Espiga3 [3]. The channels are arranged column-wise. The data is sampled at 200 Hz.

load Espiga3

Obtain the MRA of the multisignal.

w = modwt(Espiga3);
mra = modwtmra(w);

Comparing MODWT and MODWTMRA

This example demonstrates the differences between the MODWT and MODWTMRA. The MODWT partitions a signal's energy across detail coefficients and scaling coefficients. The MODWTMRA projects a signal onto wavelet subspaces and a scaling subspace.

Choose the sym6 wavelet. Load and plot an electrocardiogram (ECG) signal. The sampling frequency for the ECG signal is 180 hertz. The data are taken from Percival and Walden (2000), p.125 (data originally provided by William Constantine and Per Reinhall, University of Washington).

load wecg
t = (0:numel(wecg)-1)/180;
wv = 'sym6';
plot(t,wecg)
grid on
title(['Signal Length = ',num2str(numel(wecg))])
xlabel('Time (s)')
ylabel('Amplitude')

Take the MODWT of the signal.

wtecg = modwt(wecg,wv);

The input data are samples of a function $$f(x)$$ evaluated at $$N$$ time points. The function can be expressed as a linear combination of the scaling function $$\varphi (x)$$ and wavelet $$\psi (x)$$ at varying scales and translations: $$f(x)=\sum _{k=0}^{N-1}{c}_k{2}^{-J_0/2}\varphi (2^{-J_0}x-k)+\sum _{j=1}^{J_0}{f}_j(x)$$, where $$f_j(x)=\sum _{k=0}^{N-1}{d}_{j,k}{2}^{-j/2}\psi (2^{-j}x-k)$$ and $$J_0$$ is the number of levels of wavelet decomposition. The first sum is the coarse scale approximation of the signal, and the $$f_j(x)$$ are the details at successive scales. MODWT returns the $$N$$ coefficients $$\{c_k\}$$ and the $$(J_0\times N)$$ detail coefficients $$\{d_{j,k}\}$$ of the expansion. Each row in wtecg contains the coefficients at a different scale.

When taking the MODWT of a signal of length $$N$$, there are $$\text{floor}({\log }_2(N))$$ levels of decomposition by default. Detail coefficients are produced at each level. Scaling coefficients are returned only for the final level. In this example, $$N=2048$$, $$J_0=\text{floor}(\log 2(2048))=11$$, and the number of rows in wtecg is $$J_0+1=11+1=12$$.

The MODWT partitions the energy across the various scales and scaling coefficients: $${||X||}^2=\sum _{j=1}^{J_0}{||W_j||}^2+{||V_{J_0}||}^2$$, where $$X$$ is the input data, $$W_j$$ are the detail coefficients at scale $$j$$, and $$V_{J_0}$$ are the final-level scaling coefficients.

Compute the energy at each scale, and evaluate their sum.

energy_by_scales = sum(wtecg.^2,2);
Levels = {'D1';'D2';'D3';'D4';'D5';'D6';...
    'D7';'D8';'D9';'D10';'D11';'A11'};
energy_table = table(Levels,energy_by_scales);
disp(energy_table)

    Levels     energy_by_scales
    _______    ________________

    {'D1' }         14.063     
    {'D2' }         20.612     
    {'D3' }         37.716     
    {'D4' }         25.123     
    {'D5' }         17.437     
    {'D6' }         8.9852     
    {'D7' }         1.2906     
    {'D8' }         4.7278     
    {'D9' }         12.205     
    {'D10'}         76.428     
    {'D11'}         76.268     
    {'A11'}         3.4192     

energy_total = varfun(@sum,energy_table(:,2))

energy_total=table
    sum_energy_by_scales
    ____________________

           298.28       

Confirm the MODWT is energy-preserving by computing the energy of the signal and comparing it with the sum of the energies over all scales.

energy_ecg = sum(wecg.^2);
max(abs(energy_total.sum_energy_by_scales-energy_ecg))

ans = 7.4402e-10

Take the MODWTMRA of the signal.

mraecg = modwtmra(wtecg,wv);

MODWTMRA returns the projections of the function $$f(x)$$ onto the various wavelet subspaces and final scaling space. That is, MODWTMRA returns $$\sum _{k=0}^{N-1}{c}_k{2}^{-J_0/2}\varphi (2^{-J_0}x-k)$$ and the $$J_0$$-many $$\{f_j(x)\}$$ evaluated at $$N$$ time points. Each row in mraecg is a projection of $$f(x)$$ onto a different subspace. This means the original signal can be recovered by adding all the projections. This is not true in the case of the MODWT. Adding the coefficients in wtecg will not recover the original signal.

Choose a time point, add the projections of $$f(x)$$ evaluated at that time point, and compare with the original signal.

time_point = 1000;
abs(sum(mraecg(:,time_point))-wecg(time_point))

ans = 3.0846e-13

Confirm that, unlike MODWT, MODWTMRA is not an energy-preserving transform.

energy_ecg = sum(wecg.^2);
energy_mra_scales = sum(mraecg.^2,2);
energy_mra = sum(energy_mra_scales);
max(abs(energy_mra-energy_ecg))

ans = 115.7053

The MODWTMRA is a zero-phase filtering of the signal. Features will be time-aligned. Show this by plotting the original signal and one of its projections. To better illustrate the alignment, zoom in.

plot(t,wecg,'b')
hold on
plot(t,mraecg(4,:),'-')
hold off
grid on
xlim([4 8])
legend('Signal','Projection','Location','northwest')
xlabel('Time (s)')
ylabel('Amplitude')

Make a similar plot using the MODWT coefficients at the same scale. Features will not be time-aligned. The MODWT is not a zero-phase filtering of the input.

plot(t,wecg,'b')
hold on
plot(t,wtecg(4,:),'-')
hold off
grid on
xlim([4 8])
legend('Signal','Coefficients','Location','northwest')
xlabel('Time (s)')
ylabel('Amplitude')

Input Arguments

w — MODWT transform

matrix

MODWT transform of a signal or multisignal down to level LEV, specified as a matrix or 3-D array, respectively. w is an LEV+1-by-N matrix for the MODWT of an N-point signal, and an LEV+1-by-N-by-NC array for the MODWT of an N-by-NC multisignal. By default, imodwt assumes that you obtained the MODWT using the 'sym4' wavelet with periodic boundary handling.

Data Types: single | double
Complex Number Support: Yes

wname — Synthesis wavelet

'sym4' (default) | character vector | string scalar

Synthesis wavelet, specified as a character vector or string scalar. The synthesis wavelet must be the same wavelet used to obtain the MODWT with the modwt function.

Lo,Hi — Filters

even-length real-valued vectors

Filters, specified as a pair of even-length real-valued vectors. Lo is the scaling filter, and Hi is the wavelet filter. Lo and Hi must be the same filters used in the analysis with modwt. The filters must satisfy the conditions for an orthogonal wavelet. The lengths of Lo and Hi must be equal. See wfilters for additional information. You cannot specify both wname and a filter pair Lo,Hi.

Note

By default, the wfilters function returns two pairs of filters associated with an orthogonal or biorthogonal wavelet you specify. To agree with the usual convention in the implementation of MODWT in numerical packages, when you specify an orthogonal wavelet wname, the modwtmra function internally uses the second pair of filters returned by wfilters. For example,

mra = modwtmra(wt,"db2");

is equivalent to

[~,~,Lo,Hi] = wfilters("db2"); mra = modwtmra(wt,Lo,Hi);

This convention is different from the one followed by most Wavelet Toolbox™ discrete wavelet transform functions when decomposing a signal. Most functions internally use the first pair of filters.

Data Types: single | double

Output Arguments

mra — Multiresolution analysis

matrix | 3-D array

Multiresolution analysis, returned as a matrix or 3-D array. mra is a LEV+1-by-N matrix or LEV+1-by-N-by-NC array where LEV is the level of the MODWT and N is the length of the analyzed signal. The k^th row of mra contains the details for the k^th level. The (LEV+1)^th row of mra contains the LEV^th level smooth.

By default, mra is the same size as the input w. If you specify reflection boundary handling, then mra has one half the size of the column dimension as the input w.

References

[1] Percival, Donald B., and Andrew T. Walden. Wavelet Methods for Time Series Analysis. Cambridge Series in Statistical and Probabilistic Mathematics. Cambridge ; New York: Cambridge University Press, 2000.

[2] Whitcher, Brandon, Peter Guttorp, and Donald B. Percival. “Wavelet Analysis of Covariance with Application to Atmospheric Time Series.” Journal of Geophysical Research: Atmospheres 105, no. D11 (June 16, 2000): 14941–62. https://doi.org/10.1029/2000JD900110.

[3] Mesa, Hector. “Adapted Wavelets for Pattern Detection.” In Progress in Pattern Recognition, Image Analysis and Applications, edited by Alberto Sanfeliu and Manuel Lazo Cortés, 3773:933–44. Berlin, Heidelberg: Springer Berlin Heidelberg, 2005. https://doi.org/10.1007/11578079_96.

Extended Capabilities

C/C++ Code Generation

Generate C and C++ code using MATLAB® Coder™.

Usage notes and limitations:

• The input wname must be constant.

GPU Code Generation

Generate CUDA® code for NVIDIA® GPUs using GPU Coder™.

Usage notes and limitations:

• The input wname must be constant.

GPU Arrays

Accelerate code by running on a graphics processing unit (GPU) using Parallel Computing Toolbox™.

This function fully supports GPU arrays. For more information, see Run MATLAB Functions on a GPU (Parallel Computing Toolbox).

Version History

Introduced in R2015b