Volume error propagation should take the error in determining the cell's area into account

[tlnagy]

Currently, the only error that is propagated is the error associated with estimating the median signal of the locality, which I calculate here:

FluorescenceExclusion.jl/src/particles.jl

Line 70 in a08f8e9

push!(medbkgs, median(bkg) ± (1.253 * std(bkg) / sqrt(length(bkg))))

The other substantial source of error could be from how correctly I call the "area" of the cell. Perhaps the best approach there is to resample the edge pixels and see how much the signal varies.