Extracting the variance of peaks of s2FunHarmonic

[Ian Chesser]

Hi again, 

I attached a video that shows a density evolving with distinct peaks that broaden with increasing temperature. These are density plots of large sets of unit normal vectors: to get here, I have used the calcDensity command and used the standard plot function on the resulting s2FunHarmonic object. I have two questions: 

1. I would like to automatically count the number of unique peaks (3 for this simple example). I am trying [~,localMax] = max(density,'numlocal',100) and am enforcing some cutoff below which I don't consider a peak to be unique. In practice, the algorithm usually gives more than 3 peaks with the noisy data. Do you have any suggestions for peak finding here? 

2. I would also like to quantify the broadening of the peaks. I see a variety of functions that can be used with a S2FunHarmonic object but do not know what would be appropriate to use here. Any guidance is much appreciated. 

Thank you,

Ian

[Rüdiger Kilian]

Hi,

wrt 1: Maybe it would help to specify the minimum distance between peaks - in combination with a threshold for peak hight to identify the largest local maxima.

Cheers,

Rüdiger

[iwch...@gmail.com]

Hello, 

I am returning to question 2 after some time. I have resolved 1 but would still like to determine whether there is a simple way to quantify the "broadening" of the peaks of the spherical harmonic fit on S^2, as in the previously attached video. If I could automatically extract a variance or average peak width for each peak in the distribution, that would be quite useful. Any thoughts (including why this might be more difficult than I think) would be greatly appreciated!

Best, 

Ian