Gaussian Fitting under peaks

Is there a specific matlab function that would allow my program to estimate a guassian distrubution for each of the peaks found (in this case 5 peaks).

Answer 1

[Image Analyst]

Yes. I'm attaching a demo that can fit any number of Gaussians. It's setup for 6 in the demo but you can adapt it to 5 by changing the proper variable. Attach your data if you can't figure it out.

[1804Hz]

Thanks a lot for your help, is it fine if I can ask you a follow up question?

[Image Analyst]

You just did. You can ask another one if you want.

[1804Hz]

If I wanted to find the start and end points of the peaks is there a built-in function i can use or do I need to make a for loop to identify it?

[Image Analyst]

As long as there are definite valleys between the peaks (and it's not just like a skewed hump) then you can use findpeaks(). Just negate the signal to find valleys instead of peaks

[valleyValues, indexesOfValleys] = findpeaks(-signal);
valleyValues = -valleyValues;

findpeaks() has lots of parameters to control how big or small the peaks is so you might have to tweak some of those.

Answer 2

[the cyclist]

If you have the underlying data the fitgmdist function in the Statistics and Machine Learning Toolbox does this sort of fit. Here is some code I wrote, that fits an example (with just two peaks):

MU1    =  1;
SIGMA1 =  1;
MU2    = -3;
SIGMA2 =  1;

X = [mvnrnd(MU1,SIGMA1,1000);mvnrnd(MU2,SIGMA2,1000)];

figure
histogram(X,51)

options = statset('Display','final');
obj = fitgmdist(X,2,'Options',options)

26 iterations, log-likelihood = -4090.22

obj = 

Gaussian mixture distribution with 2 components in 1 dimensions
Component 1:
Mixing proportion: 0.499743
Mean:    1.0487

Component 2:
Mixing proportion: 0.500257
Mean:   -3.0288

The output obj has more stats in it.