How could I create a sliding window of 300 seconds, with an overlap of 60 seconds?

I have a Ecg signal. I need to process it with consecutive windows of 5 minutes, and a overlap of 1 minute.(it's like first array from 1 to 300, second one from 61 to 361, third from 121 to 421..) My Ecg has a 250 Hz sample frequency. This is my code, but it doesn't work as i would:

N2=length(ECG);

fs=250; % sample freq

smp2=1:1:N2;

tmp2=smp2*(1/fs); % time array in seconds

h=1; %1 min

w=300*fs; %5 mins

for   %%%here is missing something but i don t understand what is it

      tmp_prova=tmp2(h:w); 

      h=h+60*fs;

      w=w+(60*fs);

end

[JAYANTH BHIMAVARAPU]

Hey... if you say first array is from 1 to 300 and second one from 61 to 360 , So you will have an overlap of 300-60 i.e 240. Think about this

Answer 1

[JAYANTH BHIMAVARAPU]

for i = 0:60:length(ECG)-300

temp = ECG(i+1 : 300 + i)

end

[Walter Roberson]

Suppose length(ECG) is 500, then your code would do

for i=0:60:500
  temp = ECG(i+1:300+i)
end

the first time this would be 0+1:300+0; the second time this would be 60+1:300+60; the third 120+1:300+120 = 121:420; the fourth 180+1:300+180 = 181:480; the fifth 240+1:300+240 = 241:540 which would fail.

You need to end i such that 300+i <= length(ECG), so

 for i = 0:60:length(ECG)-300

[JAYANTH BHIMAVARAPU]

Thanks for that ..

[TP]

Thanks for the code and explanation. It worked like a breeze even for 2D array with minor modificaiton.

[ASAD RASHEED]

Prasanth, what do you mean by "minor modification"? Please explain this little bit. Thanks

[Hemen Gogoi]

If i want to execute the same operation for 500 columns then what should be the modification in the code

[Walter Roberson]

https://www.mathworks.com/matlabcentral/answers/408053-how-could-i-create-a-sliding-window-of-300-seconds-with-an-overlap-of-60-seconds#comment_584713 I show the modification for 500

Answer 2

[MUHAMMED IRFAN]

Does this help?

for i = 1 : 60: length(fullsignal)
    current_signal = fullsignal( i : (i + 300-1))

end

Answer 3

[Image Analyst]

If you have the Image Processing Toolbox, you can use blockproc(). Demo attached above.