How and What features are extracted using Wavelet Packet Transform

Question

C PRASAD 2022-8-22

0
链接

此问题的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/1783005-how-and-what-features-are-extracted-using-wavelet-packet-transform

回答： Suraj Kumar 2024-8-30

I have and Signal of 10000*3 size.I would like to extract the features from Wavelet coefficiensts of last level. Since it is wavalet packet transform,for 3rd level i will be having 8 sub bands.How this help us to extract the Features and What are the Features need to Extract generally

wname='sym5';level=3;keepap=1;
crit='shannon';critv=0;
% sorh='s';
% dwtmode('per'); %signal Extension 
emg_notch=abs(rand(10000,3));
wpt(:,i)=wpdec(emg_notch(:,i),level,wname,crit,critv); %Wpt Decomposition
Index in position 2 is invalid. Array indices must be positive integers or logical values.
%% Level 3 Coefficienst %%
WP_cfs10(:,i)=wpcoef(wpt(:,i),[3 0]);
WP_cfs11(:,i)=wpcoef(wpt(:,i),[3 1]);
WP_cfs12(:,i)=wpcoef(wpt(:,i),[3 2]);
WP_cfs13(:,i)=wpcoef(wpt(:,i),[3 3]);
WP_cfs14(:,i)=wpcoef(wpt(:,i),[3 4]);
WP_cfs15(:,i)=wpcoef(wpt(:,i),[3 5]);
WP_cfs16(:,i)=wpcoef(wpt(:,i),[3 6]);
WP_cfs17(:,i)=wpcoef(wpt(:,i),[3 7]);

0 个评论
显示 -2更早的评论隐藏 -2更早的评论

请先登录，再进行评论。

请先登录，再回答此问题。

Answer 1

Suraj Kumar 2024-8-30

0
链接

此回答的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/1783005-how-and-what-features-are-extracted-using-wavelet-packet-transform#answer_1507984

在 MATLAB Online 中打开

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); 