Contemplate instead using cumulants as discussed by Barbara Frisken for best long-time convergence.

Thanks

An easier place to start would be the regularized inverse laplace transform function (rilt.m) in the Mathworks online library. Iari-Gabriel Marion has shared his work with the rest of us. It is a port of Provencher's CONTIN fortran code to matlab. http://www.mathworks.com/matlabcentral/fileexchange/6523-rilt

G. D. J. Phillies, “Upon the Application of Cumulant Analysis to the Interpretation of

Quasielastic Light Scattering Spectra”, J. Chem. Phys. 89 , 91-99 (1988)

The important issue here is that the number of parameters that are present in a simple relaxation spectrum, even with very high signal-to-noise ratios, say 1000, and observation over a very long range of times, is very limited, say to a half-a-dozen. With a light scattering spectrum extending over six orders of magnitude in time, and an S/N (ratio of the signal at earliest time to the scatter in the relaxation spectrum around the fitting curve) of 1000 or so, Kiril Streletzky and I were pushing the limit of the technique to extract three stretched-exponential relaxations. If you think your fits are working, try checking the reproducibility.

Can anyone suggest to me, how can I plot T2 relaxation time from NMR image using CONTIN S. Provencher.

Thanks

I have made a matlab optimization which incorporates matlab function rilt and it is working fine. It is easy but slightly time consuming.