fmi-faim/faim-ipa

Blob detection finds fewer blobs

Closed this issue · 0 comments

Since the latest release fewer blobs are detected for the same scale factor.

I think this is related to wrong normalization here:

* (np.mean(sigma) / np.mean(sigmas[0])) ** 2

np.mean(sigmas[0]) should be np.mean([axial_sigma, lateral_sigma, lateral_sigma])