Jhsmit
commented
2 years ago We should do some kind of smoothing instead of the current biexponential fit See for example in HDFlex: https://www.biorxiv.org/content/10.1101/2021.12.09.471740v1
Where Piecewise cubic Hermite interpolation is used
When using weighted averaging
And the result is hard to interpret (and shouldnt be)
Perhaps t50 interpolation should be used by default (and/or improved)
Some testing will be required to see what the influence of the initial guesses is on the final fit result