peak width classification by width

In this plot peak 2,4 have different widths than other peaks . is there is a way to find peaks widths and classify them according to that ?

Answer 1

[Image Analyst]

You could use for loops to start at the peak and then start traveling the signal to the right and left until the peak starts to turn around (get smaller). For example, for peak (valley) 2:

% If the peak is at index 150
% Scan to the right.
for k = 151:length(signal)
    if signal(k) < signal(k-1)
        % Signal is starting to fall.
        rightIndex = k-1;
        break;
    end
end
% Scan to the left.
for k = 149: -1 : 1
    if signal(k) < signal(k+1)
        % Signal is starting to fall.
        leftIndex = k+1;
        break;
    end
end
peakWidth = rightIndex - leftIndex;

[sarmad m]

Thanks fr answering .

1. Is this code will take exactly this peak values for example , or the values is extended until the start of next peak ?

2. is this code ( widths_smoothed) will give the same width result for the code above ?

 [pks_smoothed,locs_smoothed,widths_smoothed,proms_smoothed] = findpeaks(AV,'MinPeakProminence',0.007);