[argyllcms] Re: rspl weights

Gerhard Fuernkranz wrote:

Graeme Gill schrieb:

The RSPLs are not related to radial basis functions. The technique is
essentially the fitting of the regular grid to the data.

Thanks for the explanation. I think I was a bit confused, since in the past you used several times the term "thin plate splines" in conjunction with RSPL. However, to the best of my knowledge, true "thin plate splines" are radial basis functions r^2*ln(r). 

I'm not sure that's true. Some descriptions of thin plate splines
such as:

"The Thin plate spline is a physically based 2D interpolation which
 represents a thin metal sheet that is constrained not to move at
 the grid points, and is free from any external force relied upon
 control points, from this sight the bending energy in control points
 should have been minimized."

indicating that it has many similarities to the RSPL. (but
maybe I'm wrong about this.)

> Thus may I conclude, the
RSPLs are not really TPS, but linear splines operating directly on the regular CLUT grid, but with otherwise similar characteristics as TPS, trying to minimize overall bending energy (with a trade-off between error and smoothness). I.e. something like a discrete variant of TPS? 

Something like that :-)

Graeme Gill.


Other related posts: