I have a series (n = 45) of 3D volumes of sparse fluorophores, taken over time, for which I would like to extract the (x, y, z) positions of. This temporal data has varying dynamic range - some images have very low contrast and still others are very high.
The number of spots I expect to detect should be approx ~ 1000 such spots per time point, so it’s a matter of finding the right height “h” (for each of the 45 3-D volumes) in skimage’s
skimage.extrema.h_maxima to call such local maximas. Given that the # of spots I expect to detect should be constant, is there a fast and robust grid search that allows me to find the best height “h” for each volume? That each of my imaging volumes has very different dynamic range - despite histogram matching - makes setting one standard height very difficult! Would appreciate any advice!