Browse Prior Art Database

ITERATIVE MODEL BASED GRAY BALANCE ALGORITHM

IP.com Disclosure Number: IPCOM000027813D
Original Publication Date: 1999-Dec-31
Included in the Prior Art Database: 2004-Apr-09
Document File: 4 page(s) / 188K

Publishing Venue

Xerox Disclosure Journal

Abstract

Disclosed is an iterative model based gray balance algorithm that simplifies the color printer calibration and characterization process in electrophotographic printers by helping to set Cyan, Magenta, and Yellow, Tone Reproduction Curves (TRC) so that patches with C=M=Y are gray. Such TRC's are called gray balanced. Finding the CMY combinations that give gray involve searching a three-dimensional space.

This text was extracted from a PDF file.
At least one non-text object (such as an image or picture) has been suppressed.
This is the abbreviated version, containing approximately 48% of the total text.

Page 1 of 4

XEROX DISCLOSURE JOURNAL

ITERATIVE MODEL BASED GRAY BALANCE ALGORITHM
Martin S. Maltz
Judith E. Stinehour
Shen-ge Wang

Proposed Classification
U. S. C1. 3581075 Int. C1. H04n 01/46

74

Near 70 CMY M / 76

Pseudo 0 L*a*b*

Diagonal

Pick CMY L*a* b*

L CMY t Grays

Sampling v /78 Parameters CMY

Grid Patch

Sampling

      r Parameters CMY

CMY

D

~ Printer Sweep E

Pseudo

Printer

MODEL

I L*a*b*

74

Pseudo

Printer

PASTE -3

76

1

1999 November/December 6 No. 24, Vol. - Xerox Disclosure Journal 259

Norm

\22

I

[This page contains 1 picture or other non-text object]

Page 2 of 4

ITERATIVE MODEL BASED GRAY BALANCE ALGORITHM (Cont'd)

Disclosed is an iterative model based gray balance algorithm that simplifies the color printer calibration and characterization process in electrophotographic printers by helping to set Cyan, Magenta, and Yellow, Tone Reproduction Curves (TRC) so that patches with C=M=Y are gray. Such TRC's are called gray balanced. Finding the CMY combinations that give gray involve searching a three-dimensional space.

The disclosed algorithm employs a two-step process. In the first step, patches are printed and measured, and the parameters for a suitable printer model are determined. The printer model calculates the L*a*b* output of the print engine given CMY where, L* is a measure of lightness, a* is a measure of magenta or green, and b* is an opponent color of yellow or blue.

In the second step, patches are printed with the printer model and the gray balance TRC's are derived, as shown in accompanying figure. Since the CMY to L*a*b* transformation produced by the model is inherently smooth and continuous, a two-dimensional analog of Newton's method is used to find the CMY combinations that give a*=b*=O.

Referring to the accompanying figure, the near diagonal sweep module 10 makes a list of patches that sample the region of CMY space in the vicinity of the C=M=Y axis where a* and b* are expected to be zero. The sampling of this region of CMY space is controlled by sampling parameters 12. A pseudo printer 14 contains the part of the image path doing mappings considered to be part of the printer (ink reduction, etc.) Note that the same pseudo printer 14 must be used throughout the calibration. The printer model calculates L*a*b* values for the corresponding patches, and Pick Grays 16 chooses the patches that are close to neutral.

Next, Grid Patch 18 is used to produce arrays of patches with CMY values in the vicinity of these near neutral CMY values. For one set of patches, cyan is held constant and magenta and yellow are varied. The space between the MY values is controlled by the sampling parameters 13 (note these are not the sampling parameters 12 used with near diagonal sweep 10). A 2x2 matrix is fitted to the a*b* vs. MY data. Solving for the MY that produces a*=b*=O yields a better estimate of the gray balanced CMY values. The same is done holding magenta constant, and varying cyan and yellow. This process is iterated. Each iteration uses...