numCoefficients

Number of wavelet scattering coefficients

Since R2019a

Syntax

ncf = numCoefficients(sf)

Description

ncf = numCoefficients(sf) returns the number of scattering coefficients for each scattering path in the wavelet time scattering network sf. The number of scattering coefficients depends on the values of the SignalLength, InvarianceScale, and OversamplingFactor properties of sf.

Examples

Oversample 1-D Wavelet Scattering Transform

This example shows how to oversample a 1-D wavelet scattering transform.

Load an ECG signal sampled at 180 Hz, and create a wavelet time scattering network to process the signal. To perform a critically downsampled wavelet scattering transform, do not change the value of the OversamplingFactor property in sf. Return the number of scattering coefficients for the scattering network.

load wecg
Fs = 180;
sf = waveletScattering('SignalLength',numel(wecg),'SamplingFrequency',Fs);
ncf = numCoefficients(sf)

ncf = 8

Return the 1-D wavelet scattering transform of wecg, and plot the zeroth-order scattering coefficients. Confirm the number of zeroth-order scattering coefficients is equal to ncf.

s = scatteringTransform(sf,wecg);
display(['Number of zeroth-order scattering coefficients: ',...
  num2str(numel(s{1}.signals{1}))])

Number of zeroth-order scattering coefficients: 8

plot(s{1}.signals{1},'x-')
grid on
axis tight
title('Zeroth-Order Scattering Coefficients')

To oversample the scattering coefficients by a factor of 2, set the OversamplingFactor property of sf equal to 1 (because $${\log }_{2}2=1$$). Return the number of scattering coefficients for the edited network. Confirm the number of scattering coefficients has doubled.

sf.OversamplingFactor = 1;
ncf = numCoefficients(sf)

ncf = 16

Return the wavelet scattering transform of wecg using the edited network, and plot the zeroth-order scattering coefficients. Since the number of coefficients in the critically sampled transform is equal to 8, confirm that the number of zeroth-order coefficients in the oversampled transform is equal to 16.

s = scatteringTransform(sf,wecg);
figure
plot(s{1}.signals{1},'x-')
grid on
axis tight
title('Zeroth-Order Scattering Coefficients')

Input Arguments

sf — Wavelet time scattering network

waveletScattering object

Wavelet time scattering network, specified as a waveletScattering object.

Extended Capabilities

C/C++ Code Generation
Generate C and C++ code using MATLAB® Coder™.

Version History

Introduced in R2019a