Segmetation of a heart sound signal, then after applying a threshold on the segments, concatenating the modified segments

I am working with a heart sound signal. i need to make "silence part" to "zero" leaving "Non-silence-part" as it is. To perform it, I took a signal sample of length 6893. Then I segmented it into 33 segments of 200 sample length. Further I am trying to implement a threshold using “for-loop” and “if-else condition. In next step, I need to recombine these outputs to make it a single signal having “silence part” and “non-silence-part of major sounds”

Here, I am sharing the signal file for your kind processing. Sampling frequency is Fs=5000 Hz.

I tried as following:

hs_ecg_file=input('input HS_ECG file Name : ','s');

y=load(hs_ecg_file); % it takes the file data_file_name as variable

y1=y(:,1);

y1=decimate(y1,5);% to decrese the sampling rate of signal by 5 times

% i.e. from Fs=5000 to Fs=1000.

plot(y1,'r')

Seperated_HS_Signal =reshape(y1(1:6800),200,[]);

% Separated 1st segment of the signal of 200 length

Seperated_HS_Signal_1=Seperated_HS_Signal(:,1);

%--------------------------------

nn=6800/200;

for j=1:nn

Spr_HS=Seperated_HS_Signal(:,j);

forNn=1:length(Spr_HS)

if abs(Spr_HS(Nn))>0.15

Spr_HS_new(Nn)= Spr_HS(Nn);

else

Spr_HS_new(Nn)=0;

end

end

Spr_HS_neww(k,:)= Spr_HS_new(:,j);

end

combined_HSS=reshape(Spr_HS_neww,[],1);

figure

plot(combined_HSS)

Answer 1

[Mathieu NOE]

hello Maddy

I tried first to use your code but got an error message as "k" is undefined

Spr_HS_neww(k,:)= Spr_HS_new(:,j);

As I more or less guessed what your intention was , I tried my own approach based on findpeaks to locate the first major peak of each segment

from there I selected only 100 samples extracts as after 100 samples we got only noise; we can make the segments longer and make the second half zero but what does it bring ?

here my code for you below

NB : Spr_HS_neww array contains all valide segments "time aligned" based on the first major peak

now what you want to do after with

combined_HSS=reshape(Spr_HS_neww,[],1);

is a bit mysterious for me...

clc
clearvars

% hs_ecg_file=input('input HS_ECG file Name : ','s');
hs_ecg_file='HS_ECG_fs5000_p.txt';
y=load(hs_ecg_file); % it takes the file data_file_name as variable
y1=y(:,1);
 
y=decimate(y1,5);% to decrese the sampling rate of signal by 5 times
                  % i.e. from Fs=5000 to Fs=1000.
samples = length(y);
x = 1:samples;

MPH = 0.25;
MPD = 100;
[PKS,LOCS] = findpeaks(y,'MinPeakHeight',MPH,'MinPeakDistance',MPD);

figure(1)
plot(x,y,'r',LOCS,PKS,'db')

 nn = numel(PKS);
newsamples = 100;
pre_trig = 20;

for ci = 1:nn
    ind_start = LOCS(ci)-pre_trig;
    ind_stop = ind_start+newsamples-1;
    if ind_stop<=samples
        Spr_HS_neww(ci,:)= y(ind_start:ind_stop);
    end
    
    
end

figure(2)
plot(Spr_HS_neww')


combined_HSS=reshape(Spr_HS_neww,[],1);
 
figure(3)
plot(combined_HSS)

[Maddy]

Thank you Mathieu NOE for your kind support. Actually i wanted to segment the signal into a fixed length. Then after applying some condition on each segment, i had to concatenate the all modified segments. Anyways, i did it and got my desired output..

I appreciate your endeavours..

Thanks a lot..