The nature of my data is something like this:
As you can see it has a lot of peaks and my data is very noisy as well. Now the peaks can be these wall-like structures running along X or Y axis.
Currently, I am using a very naive approach:
- Find all local maxima
- Find a threshold by taking an average of all local maxima
- Classify the peaks as significant only if they cross the threshold
I want to use a more efficient algorithm for this. Can anybody point me in the right direction?
After much research, I found that persistent homology can be applied to my issue. But again, I don’t have much knowledge on the topic, so any help is appreciated.