Hi C Prasad,
To extract features using the Wavelet Packet Transform (WPT), you can go through the following steps along with the attached code snippets:
1. Decompose the signal using the ‘wpdec’ function and loop over each column of your signal to cover different channels and extract the coefficients from each sub-band using the ‘wpcoef’ function.
for i = 1:3
wpt{i} = wpdec(emg_notch(:, i), level, wname, crit, critv);
for j = 0:7
WP_cfs{j+1, i} = wpcoef(wpt{i}, [3, j]);
end
end
2. After getting the coefficients, you can extract various features accordingly.
features(j, 1, i) = mean(coeffs);
features(j, 2, i) = std(coeffs);
features(j, 3, i) = sum(coeffs.^2);
features(j, 4, i) = entropy(coeffs);
features(j, 5, i) = max(coeffs);
To know more about the ‘wpdec’ and ‘wpcoef’ function in MATLAB, you can refer to the following documentations:
- https://www.mathworks.com/help/wavelet/ref/wpdec.html
- https://www.mathworks.com/help/wavelet/ref/wptree.wpcoef.html
Hope this helps!
