Guy K. Kloss wrote: > On Sun, 17 Feb 2008 12:03:06 pm Alastair M. Robinson wrote: > >> Unless you especially want to do the programming for this yourself, you >> might find that ArgyllCMS is a better fit for your needs - it has a >> complete workflow for profiling, and with a bit of spreadsheet work it's >> easy enough to use an arbitrary set of data points. >> > > On that matter ... > > Is there somewhere in a publication or on some web site the algorithm > documented on how to go about for creating a LUT based ICC profile from patch > measurement data? >
"The algoritm" does not exist. There are many ways to do it. Basically you need to fit a mathematical model of the device behavior to the data points in order to determine the model parameters, and then you can evaluate this model at each CLUT grid point in order to fill the CLUT. If you have no a priori knowledge that the device behavior can be expected to follow a very particular mathematical model, then non-parametric regression is basically the way to go. And there are also many ways to do a non-parametric regression, e.g. multivariate polynomials, multivariate splines, radial basis function networks, neural networks, etc. Another exercise is building the perceptual/saturation tables and the reverse (B2A) tables, since this additionally involves gamut mapping and the inversion of the fitted model. All in all this involves loads of mathematics, and additionally some art for finding a pleasing gamut mapping... Regards, Gerhard ------------------------------------------------------------------------- This SF.net email is sponsored by: Microsoft Defy all challenges. Microsoft(R) Visual Studio 2008. http://clk.atdmt.com/MRT/go/vse0120000070mrt/direct/01/ _______________________________________________ Lcms-user mailing list Lcms-user@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/lcms-user
