is there a way to obtain peak width after peak detection ?

I have used the function findpeaks to detect spikes in a signal

however I have initially used the halfprominence as width-reference

is there a way to obtain widths using half-height after such peak detection, having the peak timestamps and their amplitude values ?

Answer 1

[Star Strider]

Not a stupid question at all.

You simply need to ask findpeaks to return them as outputs:

[pks,locs,widths,proms] = findpeaks(PeakSig,x);

See the documentation section on Peak Prominences (link) for an illustration and discussion. The ‘widths’ are defined as the full width at half of the maximum amplitude (FWHM).

[LO]

thanks, I should have probably posted at the beginning the command I was using.

I know you get them as outputs. I was just wondering whether the width values (in heights) can be calculated from those values obtained using half prominence ([pks,locs,widths,proms])

or whether a new peak detection has to be made using half height.

I guess this now is the stupid question :)

[Star Strider]

I am not certain what you are asking.

The ‘height’ (and related ‘width’) and ‘prominence’ values are different, and only loosely related, at least as I read the documentation. The Prominence (link) property seems difficult to define. I would just use the FWHM value to characterise the peaks.

In other applications, I have used both ‘widths’ and ‘proms’ to select particular peaks, defining a desired peak by ‘(width > wv) & (prom > pv)’, where ‘wv’ and ‘pv’ were values determined elsewhere in the code. Perhaps you can combine them that way.

Answer 2

[Adam Danz]

Use the "halfheight" width reference. There's an example here: https://www.mathworks.com/help/signal/ref/findpeaks.html#buhd6xj

[pks,locs,w,p] = findpeaks(PeakSig,x,'Annotate','extents','WidthReference','halfheight')

[Adam Danz]

LO's comment moved here:

"I know that, thanks. What I wanted to know is whether one could get those values without doing the peak detection again"

Why would you need to run it twice? The widths are returned in the 3rd output.

[pks,locs,w,p] = findpeaks(PeakSig,x,'Annotate','extents','WidthReference','halfheight')

[Carlos]

Hi,

I hope this question makes sense (let me know if I need to phrase it better), but I was wondering if there are options other than "halfheight". For example, is there a way to find the width from the bottom (in this case it would be a line going along the x-axis.) Essentially, can you customize the height that you find the width from? I would greatly appreciate any input.

[Star Strider]

@Carlos — See Finding max peaks and full width/half max for both curves in the given plot for one example of determining the FWHM with respect to a signal baseline.

[Carlos]

Thank you so much

[Star Strider]

My pleasure!

Answer 3

[LO]

I know that, thanks.

what I wanted to know is whether one could get those values without doing the peak detection again

but I guess there is no way. stupid question most likely