Back to results
US 6,636,628 B1Granted
Iteratively clustered interpolation for geometrical interpolation of an irregularly spaced multidimensional color space
AssigneeXerox Corporation
InventorsYao Rong Wang, Lingappa K. Mestha, Daniel Edgardo Viassolo, Sohail A. Dianat
Life of this patent
Estimating the term…
Abstract
This invention provides an iteratively clustered interpolation (ICI) algorithm for use with irregularly sampled color data to develop a structured inverse. This algorithm improves device independent color across devices, such as, for example, printers, scanners and displays.
Claims15
A method for interpolating irregularly spaced multidimensional color data to provide an improved color device inverse, comprising:
- estimating an initial estimate point of a device inverse for a given set of input color values;
- refining the initial estimate point of the device inverse corresponding to the input color values;
- determining an inverse of the initial estimate point of the device inverse;
- comparing the device color to the input color values;
- changing the initial estimate point of the device inverse to a corresponding output color space of a color image forming device if the inversion of a particular color does not yield an acceptable inverse after refining the initial estimate point of the device inverse;
- and
- repeating the refining, determining, comparing, and changing steps using the changed initial estimate point until the inversion of a particular color yields an acceptable inverse.
4 more elements
A system for determining the accuracy of a determined irregularly spaced multidimensional color look-up table, comprising:
- an inverse printer model defined through a uniformly sampled look-up table that converts a color input Labx in L*a*b* color space into output color values Laby including color values not in the look-up table;
- and
- a forward printer model defined through a uniformly sampled look-up table that converts input color values Laby into corresponding output color values Labz.
Dependencies are inferred from the claim text.