A while ago you asked about measuring cell growth in live organisms and I suggested labelling with EdU and DAPI.
You thought the growth model would be something like [v(t1) - v(t2)] / delta(t). That it should be that simple! The problem is, to get the labels into cells you have to kill them so you can only obtain values for one time point.
You have to set up a differential system for subpopulations of cells (based on mitotic status) and solve it as a boundary value problem. The BVP the straight forward part.
The complicated part is quantifying the amount of fluorescent label in each nucleus. The tissue surrounding the cell and solution surrounding the sample absorb, refract and convolve the observed fluorescence intensities. Beer Lambert works only for cuvettes. Additionally, the EdU and the DAPI fluorescence affect each other and it's difficult to discern signals from spatially close nuclei. Quantification requires a Fourier system.
After that, using the fluorescence subpopulations to define the mitotic subpopulations is not straight forward either as cells go through successive rounds of commitment and division. In my paper, I validated my model (as well as my computational procedure) by decomposing these "fast dividing" and "slow dividing" subpopulations and proving that they do actually correspond to different cell types in my model organism, which required another decomposition (deconvolution, in this case, not Fourier) technique.
As a footnote, note that you have to simultaneously measure the apoptotic rate as well as the mitotic and commitment rates to make inferences about overall growth of the organism.
That I might have left you with a wrong impression has been buggering me so here's the full answer. In a nut shell.