Extreme points finding from set

d = xlsread('FaceFour_TRAIN.xlsx')
s = d(1:1,2:end);
fa = movstd(s,20,1);
secarray = movstd(fa,20,1)  ;
secidx = find(secarray>.04);
f = secarray(secidx);
[maxvals, maxidx] = maxk(f,14);
sidx = secidx(maxidx);
x = 1:length(s);
key=plot(x, s, 'b-', sidx, s(sidx), 'go')

here the maximum points are extracted and plotted.I need to check and find only the extreme points from this,how can I do this.Please help me.

[Geoff Hayes]

Silpa - please clarify what you mean by extreme. Are these the minimum and maximum points, or something else?

[Silpa K]

The graph is like a wave.I mean the highest points that coming in the wave crest and wave trough.

[Silpa K]

The graph like

[Rik]

I suggested findpeaks in your duplicate post.

[Silpa K]

Sir Iam new to matlab.If I used findpeaks only the peaks are extracting.How can I find both peak points and all the mininimum troughs points.And that points I need to mark in the original graph.Please help me.

Thank you.

[Rik]

There are probably better ways, but if you do this

findpeak(-data)

you will find the local minima.

And once you have x and y, you should be able to plot those points.

[Silpa K]

Iam getting 1×0 empty double row vector when I use this.

[Rik]

Without your data and your code I can do nothing to help you. Have you read the documentation? I suspect it may return an empty vector is none of your peaks satisfy the requirements. Use the optional arguments to relax the requirements.

[Silpa K]

d = xlsread('FaceFour_TRAIN.xlsx')

s = d(1:1,2:end );

fa = movstd(s,20 );

secarray = movstd(fa,20 ) ;

sec = secarray(secarray>.04 );

k=maxk(sec,14)

peak points of k

[darova]

Why those points are not extreme?

[Silpa K]

Sir,It is only a sample graph.I need to find all the extreme points(Present in the crust and trough)

[darova]

HOw you define "extreme" point? Where is it?

[Silpa K]

The maximum peak points present in the crust and trough.If I use findpeaks(d),then iam not getting the answer.

1×0 empty double row vector is the output coming.

Answer 1

[Steven Lord]

I suspect some combination of islocalmin, islocalmax, and/or ischange will identify the points you've called out with + markers in your plot. All three of these functions were introduced as part of MATLAB in release R2017b.

[Silpa K]

sir,Iam not getting.