Match imaging systems from DSV datasets.
This is a static redo of an old Python wrapper around ALGLIB's superb RBF implementation. That version is now a part of camera_match.
Along with RBF, sdfit parameterizes the neural network included in ALGLIB as mlp.
sdfit -h
sdfit v0.2.0
Scattered data fitting for tristimulus lookup tables.
https://github.com/hotgluebanjo
USAGE: sdfit <source> <target> [OPTIONS]
EXAMPLES:
sdfit alexa.csv print-film.csv -d ',' -o alexa_to_print_film.cube
sdfit venice.txt alexa.txt -m rbf -p 6 -f spi -o venice_to_alexa.spi3d
INPUTS:
<source> Plaintext file containing source dataset
<target> Plaintext file containing target dataset
OPTIONS:
-h Help
-m Method to use [mlp | rbf] default: mlp
-o Output path and name default: 'output.cube'
-d Dataset delimiter [' ' | ',' | <tab>] default: ' ' (space)
-p LUT print precision default: 8
-c LUT cube size default: 33
-f LUT format [cube | spi] default: cube
-s RBF basis size default: 5.0
-l RBF layers default: 5
-z RBF smoothing default: 0.0
-L MLP layers default: 5
-r MLP restarts default: 5