Color lookup table compression
AssigneeXerox Corporation
InventorR. Victor Klassen
Life of this patent
Estimating the term…
Abstract
A technique to compress multidimensional color look-up tables by computing an inexpensive fit to the contents of the tables is provided. At each node of the table, the difference between the result of evaluating the fit and the original value at that node are completed. In one form, the parameters of the fit and the differences are stored and compressed, possibly losslessly.
Claims25
A method for compressing a multidimensional color look-up table of a color rendering system, the method comprising:
- determining input data to the multidimensional color look-up table of the color rendering system;
- fitting a mathematical model to the input data based on a least squares approach;
- evaluating the mathematical model for each node of the table to obtain a result for each node;
- computing a difference between the result for each node and the input data at each node;
- and,
- compressing all differences.
A method for compressing a multidimensional color look-up table of a color printing system, the method comprising:
- determining input data to the multidimensional color look-up table of the color printing system;
- fitting a tensor product B-spline to the input data using a least squares fitting routine;
- evaluating the tensor product B-spline for each node of the table to obtain a first result for each node;
- rounding the first result for each node to obtain a second result;
- computing a difference between the second result and the input data at each node;
- selecting parameters of the tensor product B-spline;
- compressing all differences;
- and,
- compressing the selected parameters of the tensor product B-spline.
A system for processing a color image comprising:
- an interface operative to receive color input data relating to the color image;
- a memory operative to store the color input data in a plurality of multidimensional color look-up tables;
- and,
- a processor operative to fit a mathematical model to the color input data based on a least squares approach, evaluate the model to obtain a result for each node of the table, compute a difference between the result and the input data at each node, store all differences, and compress the stored differences.
Dependencies are inferred from the claim text.