how does the function findpeaks work in matlab?

I use the function findpeaks to detect peak in a vector,it worked as i expected,but I want to understand how does it work,and what the principle it use.

n=0:1:10;
x = [25 8 15 5 6 10 10 3 1 20 7];
[k,v]=findpeaks(x);
plot(n,x,'b');grid;hold on
plot(n(k),v,'r.');
Can anyone can help me?I will be very appreciate for your kindness.

Answer 1

[Rik]

Excerpt from the documentation:

pks = findpeaks(data) returns a vector with the local maxima (peaks) of the input signal vector, data. A local peak is a data sample that is either larger than its two neighboring samples or is equal to Inf. Non- Inf signal endpoints are excluded. If a peak is flat, the function returns only the point with the lowest index.

So that's how it works.

[yawen kang]

first,thanks for your answer.but I already know how to use it.I want to know what algorithm it use to find the peak.I want to understand what the code in erery line mean,and what principle it use to find the peak.

n=0:1:10; x = [25 8 15 5 6 10 10 3 1 20 7]; nx=length(x); x=x(:); dx=x(2:end)-x(1:end-1); r=find(dx>0); f=find(dx<0); if length(r)>0 & length(f)>0 % we must have at least one rise and one fall dr=r; dr(2:end)=r(2:end)-r(1:end-1); rc=repmat(1,nx,1); rc(r+1)=1-dr; rc(1)=0; rs=cumsum(rc); % = time since the last rise df=f; df(2:end)=f(2:end)-f(1:end-1); fc=repmat(1,nx,1); fc(f+1)=1-df; fc(1)=0; fs=cumsum(fc); % = time since the last fall rp=repmat(-1,nx,1); rp([1; r+1])=[dr-1; nx-r(end)-1]; rq=cumsum(rp); % = time to the next rise fp=repmat(-1,nx,1); fp([1; f+1])=[df-1; nx-f(end)-1]; fq=cumsum(fp); % = time to the next fall k=find((rs<fs) & (fq<rq) & (floor((fq-rs)/2)==0)); % the final term centres peaks within a plateau v=x(k); else k=[]; v=[]; end

[Jan]

@yawen: It seems like you noticed, that the code is not readable. Please edit your comment and use the "{} Code" button for a proper formatting. Thanks.

Answer 2

[yawen kang]

first,thanks for your answer.but I already know how to use it.I want to know what algorithm it use to find the peak.I want to understand what the code in erery line mean,and what principle it use to find the peak.

n=0:1:10;
x = [25 8 15 5 6 10 10 3 1 20 7];
nx=length(x);
x=x(:);
dx=x(2:end)-x(1:end-1);
r=find(dx>0);
f=find(dx<0);

if length(r)>0 & length(f)>0    % we must have at least one rise and one fall
    dr=r;
    dr(2:end)=r(2:end)-r(1:end-1);
    rc=repmat(1,nx,1);
    rc(r+1)=1-dr;
    rc(1)=0;
    rs=cumsum(rc); % = time since the last rise

      df=f;
      df(2:end)=f(2:end)-f(1:end-1);
      fc=repmat(1,nx,1);
      fc(f+1)=1-df;
      fc(1)=0;
      fs=cumsum(fc); % = time since the last fall

      rp=repmat(-1,nx,1);
      rp([1; r+1])=[dr-1; nx-r(end)-1];
      rq=cumsum(rp);  % = time to the next rise

      fp=repmat(-1,nx,1);
      fp([1; f+1])=[df-1; nx-f(end)-1];
      fq=cumsum(fp); % = time to the next fall

      k=find((rs<fs) & (fq<rq) & (floor((fq-rs)/2)==0));   % the final term centres peaks within a plateau
      v=x(k);

else
    k=[];
    v=[];
end

[Rik]

If you wan't to look into the function itself, type edit findpeaks and you will see the code Matlab uses.

BTW, as this isn't an answer to your question, but a follow up question based on my answer, you shouldn't use the answer option, but comment instead on my answer.

[Adam]

What is the relevance of that code anyway? It doesn't even include a call to findpeaks!

[Rik]

Skimming this code, it looks like it aims to replicate findpeaks, although this version doesn't return the first index of a plateau, but the middle.