US20080150960A1 - Method for mapping colors between imaging sytems and method therefor - Google Patents

Abstract

A method for converting CMYK source device values to CMYK destination device values includes converting the CMYK source device values to a four-color device-independent color space using a source profile with a four-color device-independent color space. The four color device-independent color space is converted to the CMYK destination values using a destination profile with the four color device-independent color space, and three colors of the four color device-independent color space are used to define the color of a combined CMYK values and the remaining device-independent color space color is used to define the color of only a K colorant.

Description

CROSS REFERENCE TO RELATED APPLICATIONS
• This application is a Divisional of U.S. application Ser. No. 09/536,366 filed Mar. 27, 2000, which is a Continuation of U.S. application Ser. No. 08/882,561 filed Jul. 3, 1997, now U.S. Pat. No. 6,088,038, and is a Continuation-In-Part of U.S. application Ser. No 08/884,411 filed Jun. 27, 1997, now U.S. Pat. No. 6,108,442.
FIELD OF THE INVENTION
• The present invention relates to color imaging. More particularly, the present invention relates to mapping colors between color imaging systems.
BACKGROUND OF THE INVENTION
• Color reproduction processes typically involve using color imaging systems to produce colors on various media. These color imaging systems may be used to duplicate a color image from one medium to another medium, e.g., from one printed copy to another or from a display screen to a printed copy. Color reproduction processes are used in various application environments, for example, color proofing applications.
• Some color reproduction processes use approaches known as color management systems (CMSs) to characterize various color imaging systems and to transform color data between the color imaging systems. Characterizing color imaging systems typically involves calculating color response functions using color coordinate systems known as color spaces. One commonly-used color space is Commission Internationale de'Éclairage L*a*b; (CIELAB) space. CMSs attempt to reproduce an original color image on a color imaging system so as to preserve the appearance of colors between the original and the reproduction within the limitations of the color imaging system of the reproduction process.
• Various CMS approaches have been proposed to achieve accurate color reproduction. Many of these approaches involve producing color samples using an output or display device and measuring the color values of the samples using an input device. Such approaches correlate the output colors with the measured color values. This correlation is performed using, for example, forward and reverse transforms between device-specific color spaces and a device-independent color space. These transformation techniques are often supplemented by interpolation between entries in a multidimensional lookup table. These techniques exhibit inaccurate color conversion between similar devices, potentially resulting in undesirable contamination of colors. Furthermore, accurate color conversion of dark colors has often been particularly difficult because of inadequate processing of black channel data in many applications.
• CMSs often perform gamut mapping to correlate the range or gamut of colors that can be realized by a device with regions of a color space. Because many devices are incapable of realizing the complete range of colors in a color space, gamut mapping typically involves compressing or scaling regions of the color space. The device can then approximate colors outside its gamut using the compressed regions of the color space. For many CMSs, gamut mapping is potentially inconsistent under certain circumstances, such as when using profiles generated by software from different vendors. In addition, many CMSs exhibit inconsistencies when performing forward and reverse transformations between imaging systems. For example, color shifting often occurs with repeated forward and reverse transformations.
• Many CMS techniques exhibit other limitations in addition to the lack of accuracy in converting colors. For example, many CMS techniques are relatively inflexible with respect to changes in illumination and observer conditions, gamut mapping, and choice of color space. Certain techniques lack forward compatibility with future color standards.
BRIEF DESCRIPTION OF THE DRAWINGS
• Other aspects and advantages of the present invention will become apparent upon reading the following detailed description and upon reference to the drawings in which:
• FIG. 1 is a block diagram illustrating an example color mapping system, according to an embodiment of the present invention;
• FIG. 2 is a block diagram illustrating an example arrangement implementing part of the color mapping system ofFIG. 1, according to an embodiment of the present invention;
• FIG. 3 is a block diagram illustrating another example arrangement implementing part of the color mapping system ofFIG. 1, according to an embodiment of the present invention;
• FIG. 4 is a block diagram illustrating yet another example arrangement implementing part of the color mapping system ofFIG. 1, according to an embodiment of the present invention;
• FIG. 5 is a block diagram illustrating still another example arrangement implementing part of the color mapping system ofFIG. 1, according to an embodiment of the present invention; and
• FIG. 6 is a flow chart illustrating an example color mapping method, according to another embodiment of the present invention.
• While the invention is amenable to various modifications and alternative forms, specifics thereof have been shown by way of example in the drawings and will be described in detail. It should be understood, however, that the intention is not to limit the invention to the particular embodiments described. On the contrary, the intention is to cover all modifications, equivalents, and alternatives falling within the spirit and scope of the invention as defined by the appended claims.
DETAILED DESCRIPTION OF THE VARIOUS EMBODIMENTS
• The present invention is believed to be applicable to a variety of systems and arrangements that characterize color imaging systems. The invention has been found to be particularly advantageous for transforming colors between different color imaging systems. An appreciation of various aspects of the invention is best gained through a discussion of these particular application examples.
• According to one aspect of the present invention, a color mapping technique may be applied to a variety of color imaging systems to generate a color map that can be used to transform the color response of one color imaging system, referred to as a source color imaging system, to match the color response of another color imaging system, referred to as a destination color imaging system. The color mapping technique projects color coordinates in the color space used by the source color imaging system into, for example, a device-independent color space. optimal color coordinates in the color space used by the destination color imaging system are determined that realize a relatively close match between the projections into the device-independent color space of the color coordinates in the color space used by the source color imaging system and the optimal color coordinates. The color mapping technique then generates a color map based on the optimal color coordinates for a number of color coordinates in the color space used by the source color imaging system.
• FIG. 1 illustrates anexample system100 according to the present invention configured to transform colors between imaging systems. Thesystem100 includes an appropriately-programmedcomputer arrangement102. Thecomputer arrangement102 may be implemented using any of a variety of conventional resources, for example, a personal computer and CD-ROM based software. Other computer-based designs may be used as well. For example, thecomputer arrangement102 may be implemented using a microprocessor that acts as a read-only memory (ROM) into which a software application program is loaded. The software application program may be incorporated, for example, in a color management software package.
• Acolor mapping system104 includes acolor management system106. Thecolor management system106 receives asource device profile108 and adestination device profile110. These device profiles describe mappings from device-dependent color coordinate systems used by respective color imaging systems to device-independent color coordinate systems.
• Thecolor management system106 processes thesource device profile108 and thedestination device profile110 to generate acolor map114. Thecolor map114 describes a relationship between the color imaging systems used by the source and destination devices. Amemory116 stores thecolor map114. Subsequently, thecolor management system106 uses thecolor map114 to transform a set of source coordinates118 in a device-dependent source device color space into a set of destination coordinates120 in a device-dependent destination device color space.
• FIG. 2 illustrates an examplecolor management system200 for transforming colors between imaging systems according to the present invention. A sourcedevice profile interpreter202 receives asource device profile206. Thesource device profile206 is used to map coordinates in the source device color space to some form of color data, such as spectral or XYZ tristimulus values. For example, if the source device is a halftone color printer, thesource device profile206 may map CMYK color values to a XYZ color space. The sourcedevice profile interpreter202 interprets thesource device profile206 and converts coordinates in the source device color space to a device-independent color space known as a profile connecting space (PCS). The PCS is used for converting the coordinates in the source device color space to the destination device color space. The PCS may be, for example, the CIELAB color space. Another example PCS is described in co-pending U.S. patent application Ser. No. 08/884,411, now U.S. Pat. No. 6,108,442, entitled “Characterization of Color Imaging System” (Christopher Edge et al.), filed on Jun. 27, 1997, and incorporated herein by reference.
• A destinationdevice profile interpreter208 receives a destination device profile210. The destination device profile210 is used to map color coordinates in a destination device color space used by adestination device212 to some form of color data, such as spectral or XYZ tristimulus values. For example, if thedestination device212 is a cathode ray tube (CRT) monitor, the destination device profile210 may map color coordinates in a red-green-blue (RGB) color space to XYZ tristimulus values. The destinationdevice profile interpreter208 interprets the destination device profile210 and converts color coordinates in the destination device color space to the PCs.
• The source and destinationdevice profile interpreters202 and208 may be implemented using any of a variety of hardware and software arrangements and are configurable for a variety of application environments. For example, if the source and destination device profiles206 and210 are International Color Consortium (ICC) device profiles, the source and destinationdevice profile interpreters202 and208 are optionally configured to include white- and black-point parameters to account for color variations between media and colorants used by different color display devices. The source and destinationdevice profile interpreters202 and208 can also be configured to include pleasing color corrections, such as L′ rescaling and a′b′ hue adjustments. Alternatively, the pleasing color corrections can be incorporated into thecolor transformer214. In certain other application environments, the source and destinationdevice profile interpreters202 and208 are further configurable to include, for example, illuminant and observer functions.
• Thedevice profile interpreters202 and208 can be configured using any of a variety of approaches. For example, plug-in software modules can be used to configure thedevice profile interpreters202 and208. Using plug-in software modules obviates the need to use new versions of thecolor management system200 or of the device profiles206 and210 when adding, for example, a newly defined color space, a custom illuminant, such as fluorescent light, or a new gamut mapping technique. These options can be selected, e.g., using a setup window at the operating system level. For example, if the operating system is Apple OS version 7.5, these options can be selected using a control panel interface.
• If the device characterization is non-spectral, thecolor management system200 can use the original spectral data that is saved with the profile to reconstruct the device profiles according to various conditions, such as illuminant functions and color space choices. For example, if one uses an RGB regression to convert scanner RGB values into color space values for a particular combination of color space and illuminant and observer conditions based on a set of spectral data, the regression for a new set of conditions can be generated based on the same spectral data. Accordingly, the device profiles206 and210 can be used to calculate color values for a variety of conditions and color appearance models
• Acolor transformer214 obtains PCS color coordinates from the source and destinationdevice profile interpreters202 and208. Thecolor transformer214 uses these color coordinates to develop acolor map216 that expresses a relationship between the color spaces used by the source anddestination devices204 and212. To generate thecolor map216, thecolor transformer214 may use any of a variety of gamut mapping techniques. One such technique that has been found to yield particularly accurate results involves reducing the color error between the source and destination devices. The color error is defined, for example, by Euclidean distances in the PCS or by weighted sum square errors in a color space that is polar in the chromatic dimensions of the PCS. Defining the color error using weighted sum square errors results in a mapping between color imaging systems that accurately maintains colors in reproduced images. By using error reduction techniques, thecolor transformer214 avoids generating significant cumulative error in performing multiple forward and reverse transformations between color spaces.
• Thecolor transformer214 is implemented using, for example, a software program, and can be configured for a variety of applications. For example, thecolor transformer214 can be configured to perform a100% black point scaling for mapping a printed color image to a monitor display of the image. On the other hand, because newsprint has a relatively weak black point attributable to its ink density and light-transmitting properties, thecolor transformer214 can be configured to perform, for example, a 50% black point scaling when mapping a color image printed on newsprint to the Matchprint™ color imaging system. Thecolor transformer214 is also configurable to use, for example, illuminant and observer functions, which thecolor transformer214 provides to the source and destinationdevice profile interpreters202 and208. Thecolor management system200 receives user preferences from aninput218 to determine how to configure thecolor transformer214.
• After developing thecolor map216, thecolor transformer214 can be used to transform colors between the source anddestination devices204 and212. Thecolor transformer214 receives color coordinates from thesource device204 and transforms them using thecolor map216. This transformation produces a set of color coordinates in the destination device color space. The destination color imaging system then reproduces the color on thedestination device212 using these color coordinates.
• FIG. 3 illustrates an exampledevice profile interpreter300 implementing part of thecolor management system200 ofFIG. 2. Thedevice profile interpreter300 uses adevice profile302 to convert device coordinates received at aninput304 to PCS color coordinates, which thedevice profile interpreter300 provides at anoutput306. Thedevice profile302 describes the relationship between the device coordinates and some form of color data. Additionally, thedevice profile302 optionally stores the raw spectral data used to construct thedevice profile302. The raw spectral data allows subsequent construction of more accurate device profiles302, e.g., if ICC specifications change. This updating can be performed automatically, for example, upon detecting that some component of thedevice profile302 is out of date. Updates can also be performed periodically based on a schedule. To update thedevice profile302, a new profile can be generated using the spectral data. Alternatively, error reduction, such as a one-dimensional correction, can be performed on each channel in the original look-up table for constructing the new profile. This correction can be applied as a separate set of one-dimensional tables or applied directly to the analytical model or multidimensional look-up table. For additional information concerning an example error reduction procedure that can be used in constructing a new profile, reference can be made to U.S. patent application Ser. No. 08/431,614, entitled “Apparatus and Method for Recalibrating a Multi-Color Imaging System,” assigned to the instant assignee and incorporated herein by reference.
• Adevice profile processor308 receives the device coordinates from theinput304 and thedevice profile302. Thedevice profile302 may be, for example, an ICC profile. If thedevice profile302 exists in this format, the device profile processor receives the forward portion of the profile, i.e., the portion used for converting device coordinates to PCS color coordinates. Alternatively, thedevice profile302 can be stored in another format. Thedevice profile processor308 processes the device coordinates using thedevice profile302 and outputs certain data based on thedevice profile302. For example, if thedevice profile302 is an ICC profile, thedevice profile processor308 outputs XYZ tristimulus values for a particular set of observer conditions (e.g., illuminant and observer functions). if thedevice profile302 is based on spectral data, thedevice profile processor308 outputs spectral data. Thedevice profile processor308 can be configured for a variety of applications. For example, a user can select between absolute and relative colorimetrics and can configure observer, e.g., illuminant, conditions.
• APCS processor310 receives the data output from thedevice profile processor308 and a set of PCS parameters from aninput312. The PCS parameters may include, for example, XYZ tristimulus values for the media white, the illuminant white, and the black point, as well as blackpoint scaling from a perfect black to the media black. ThePCS processor310 generates the PCS values as a function of the data received from thedevice profile processor308 and the PCS parameters.
• FIG. 4 illustrates anexample color transformer400 implementing part of thecolor management system200 ofFIG. 2. Adevice link generator402 receives as input at least one source profile and one destination profile. WhileFIG. 4 illustrates thedevice link generator402 receiving a source device profile from asource profile interpreter404 and a destination device profile from adestination profile interpreter406, it should be understood that thedevice link generator402 may also receive one or more device profiles that are intermediate between the source and destination device profiles. For example, a device profile characterizing an RGB monitor can be intermediate between a source device profile characterizing an RGB scanner and a destination device profile characterizing a CMYK printer. The source and destination device profiles are forward transforms and optionally include configurable observer conditions and PCS parameters. Thedevice link generator402 also receives a series ofPCS parameters408 to improve linking of different device types (e.g., CRT monitors and printers). Thegamut mapping parameters410 improve mapping of out of gamut colors between device types.
• Thedevice link generator402 generates a color map ordevice profile link412 that maps colors between two devices, e.g., from an RGB device to a CMYK device or between two CMYK devices. Thedevice profile link412 is, for example, a mathematical expression or a look-up table. Thecolor transformer400 optionally stores thedevice profile link412 in a memory, such as a random access memory (RAM), or saves it as a file for multiple transformations between the source and destination device color spaces.
• Adevice link calculator414 receives source device coordinates from aninput416 and processes them using thedevice profile link412. Thedevice link calculator414 uses a single forward calculation to transform the source device coordinates to a set of destination device coordinates for presentation at anoutput418. Because thedevice link calculator414 uses a single forward calculation, interpolation is relatively simple and easily optimized and the transformation process is relatively fast. If thedevice profile link412 is a lookup table, thedevice link calculator414 optionally uses linear interpolation to refine the destination device coordinates. Thedevice link calculator414 can be implemented, e.g., using a conventional multidimensional linear interpolator.
• FIG. 5 illustrates an exampledevice link generator500 that implements part of thecolor transformer400 ofFIG. 4. Thedevice link generator500 includes a devicelink table builder502 that creates a look-up table to enable rapid interpolation of destination device coordinates from source device coordinates. It should be understood that if the device profile link is a mathematical expression rather than as a look-up table, an analogous transformation generator replaces the devicelink table builder502. Such a transformation builder may, for example, generate coefficients for use in the mathematical expression. To facilitate the discussion, however, thedevice link generator500 is assumed to include a devicelink table builder502. The devicelink table builder502 generates the look-up table by generating a series of source device coordinates as input value entries and determining the optimal destination device coordinates as output values corresponding to the input values. The devicelink table builder502 generates all combinations of source device coordinates using, for example, a series of nested loops, one loop for each dimension of the source device color coordinate space.
• To reduce the computational and memory requirements for constructing and storing the look-up table, the look-up table typically contains a relatively small number of entries along each dimension. With a relatively small table, interpolation is used to convert source coordinates to destination coordinates. The total number of entries in the look-up table can be expressed as D_dN₈d, where d is the dimensionality of the source device color space, D_dis the dimensionality of the destination device color space, and N9 is the number of entries along each dimension of the look-up table. For example, a look-up table that is used to transform color coordinates between two CMYK (i.e., four dimensional) color spaces can contain 4×17⁴, or 334,084 entries.
• It should be understood that the look-up table need not have the same number of entries along each dimension. If the look-up table contains N_kentries along each respective dimension, where k ranges from 1 to d, the total number of entries in the look-up table can be expressed as
• $D_D\prod _{k=1}^dN_k.$
• For example, a look-up table that is used to transform color coordinates between two CMYK color spaces and that is to have fifteen entries along three dimensions and seventeen entries along one dimension contains 4×15×15×15×17, or 229,500 entries.
• In application environments in which it is desirable to further reduce computational and memory requirements, the devicelink table builder502 may select only a subset of the total number of entries in each dimension of the look-up table, perform the method loop calculations using that subset, and perform, for example, a spline interpolation to fill in the remaining entries of the look-up table.
• The devicelink table builder502 provides PCS parameters and source device coordinates to a sourcedevice profile interpreter504. The sourcedevice profile interpreter504 generates source PCS values and provides the source PCS values and the source device coordinates to anerror reducer506. In a specific embodiment, theerror reducer506 is implemented using an error minimization technique. Using the source device coordinates, theerror reducer506 estimates an initial set of destination coordinates that are likely to realize an accurate color match with the source device coordinates. This estimation process may be performed using a relatively simple technique. For example, for estimating destination coordinates in an RGB space corresponding to source coordinates in a CMYK color space, the estimation process may use the following equations:
• 
  C=1−R
• 
  M=1−G
• 
  Y=1−B
• Alternatively, the source coordinates can be used to estimate the destination coordinates if the source and destination imaging systems use similar color coordinate spaces.
• Theerror reducer506 provides the set of estimated destination device coordinates to a destinationdevice profile interpreter508, which also receives the PCS parameters from the devicelink table builder502. The destinationdevice profile interpreter508 then generates a set of destination PCS values as a function of the estimated destination device coordinates and the PCS parameters and provides the destination PCS values to theerror reducer506. If the error between the destination PCS values and the source PCS values is non-zero, theerror reducer506 uses an error reduction (e.g., an error minimization) technique to reduce the error between the source and destination PCS values. In one embodiment, this is implemented by repeatedly querying the destinationdevice profile interpreter508 with selected estimates of destination device coordinates. This process can continue until destination device coordinates are found that satisfy a quality threshold, for example, that yield the minimum error. Theerror reducer506 returns these destination device coordinates to the devicelink table builder502, which enters them in an appropriate location in the look-up table. The devicelink table builder502 then enters the next set of table input entries corresponding to a set of source device color coordinates.
• For colors within the gamut of the destination device, the error can be reduced using any of a variety of reduction techniques. For example, Powell's method can be used to perform the error reduction or error minimization. For additional information regarding Powell's method, reference may be made to William H. Press et al.,Numerical Recipes in C(1992), pp. 309-315, available from Cambridge University Press, incorporated herein by reference, and attached hereto as Appendix A.
• Using this approach, theerror reducer506 generally defines an error function having input parameters that can be varied by the error reduction technique. Theerror reducer506 then determines the optimal values of the input parameters resulting in a minimal error. To determine the values of destination device coordinates using the minimum PCS error between the source and destination PCS values, the variable input parameters are the destination device coordinates. Accordingly, in this specific implementation, theerror reducer506 defines the error function as:
• 
  Error(D)=ΔE(R₈, R_d(D))
• where D is a vector defined by the destination device coordinates, R₈is a vector defined by the source PCS values and R_d(D) is a vector function producing destination PCS values as a function of the destination coordinate vector D, and ΔE is the Euclidean distance error between R₈and R_d(D). The Euclidean distance error may be expressed using the following equation:
• 
  ΔE(R₁, R₂)={square root over (√(L₁−L₂)²+(a₁−a₂)²+(b₁−b₂)²)}{square root over (√(L₁−L₂)²+(a₁−a₂)²+(b₁−b₂)²)}{square root over (√(L₁−L₂)²+(a₁−a₂)²+(b₁−b₂)²)}
• The above equation assumes that the PCS is implemented as the CIELAB color space. It should be understood, however, that other color spaces may be used as a PCS. For example, one color space that is particularly suited for use as a PCS is described in the previously-referenced copending U.S. patent application Ser. No. 08/884,411 now U.S. Pat. No. 6,108,442, entitled “Characterization of Color Imaging Systems.”
• Using this same approach, a non-zero optimal error indicates that the source device is out of gamut relative to the destination device at that location in the PCS. In such situations, theerror reducer506 optionally uses the destination device coordinates that result in a minimum ΔE value. Alternatively, theerror reducer506 may use these values as an initial estimate and recalculate the optimal destination device coordinates using a new error function that employs weighting factors, polar coordinates in the chromatic plane of the PCS space, or both.
• Theerror reducer506 optionally uses a gamut mapping parameter received from the devicelink table builder502 to decide how to map coordinates that are out of gamut relative to the source device. For example, the gamut mapping parameter may specify modes in which each technique is used for obtaining destination device coordinates. One mode, for example, may use lightness, chroma, and hue values L*, C*, and h* instead of LAB values L*, a*, and b*, where:
• 
  √{square root over ((a^*2+b^*2))}
• 
  h*=C*·ARCTAN b*/a*).
• Another mode uses the above lightness, chroma, and hue values as well as weighting factors:
• $\mathrm{Error}\left(D\right)=\Delta \mathrm{EW}\left(R_s,R_d\left(D\right),W\right)$$\Delta \mathrm{EW}\left(R_1,R_2,W\right)=\sqrt{\begin{array}{c}{W_L\left(L_1-L_2\right)}^2+{W_C\left(C_1-C_2\right)}^2+\\ {W_h\left(h_1-h_2\right)}^2\end{array}}$
• where the PCS vectors R₈, R_d(D) are converted to lightness, chroma, and hue values either before or after passing them to the error function.
• If the weighting factors are one, the above weighted error reduction function ΔEW ( ) is identical to the standard ΔE ( ) error reduction function. It should be noted, however, that weighting factors of W_L=3, W_C,=1, and W_h=1.5 yield particularly accurate visual results. These weighting factors produce an error function that gives priority first to lightness, then to hue, then to chroma. These weighting factors can also be provided to theerror reducer506 as gamut mapping parameters by the devicelink table builder502.
• Creating the device profile link via error reduction of the forward transformations of the devices realizes a number of advantages. For example, errors in color conversion are limited to those attributable to rounding and interpolation. As a result, the cumulative error from repeated forward and reverse transformations between the source and destination device color spaces is substantially reduced. Additionally, the color transformer can select the gamut mapping technique. The color transformer can rely on the forward transform information and realize consistent gamut mapping between device profiles supplied by different vendors. It should be noted that errors due to interpolation of the device profile link decrease as the number of table entries in each dimension of the look-up table increases toward the maximum number of gray levels. This error also decreases if a one-dimensional tone reproduction table is used to transform the color values. For additional information regarding the use of a one-dimensional tone reproduction table, reference is made to U.S. Pat. No. 5,432,906, issued to Gary H. Newman, assigned to Eastman Kodak Company, and incorporated by reference.
• Creating the device profile link using error reduction also allows transformation between CMYK device spaces that maps the tone response of the source and destination black (K) channels while maintaining an accurate match with the L*a*b* data. For transformation from an RGB source device to a CMYK destination device, the RGB color coordinates used by the source device lack K channel information. Some conventional color transformation techniques use a process known as gray component removal (GCR) to define a relationship between K values and CMY values in the reverse transformation (i.e., L*a*b*to CMYK). For example, the reverse transformation may be performed with K initially set to zero. The value of K can then be calculated based on the minimum of the C, M, and Y values. The CMY values can then be recalculated using an algebraic calculation or using the forward model to obtain the closest value of L*a*b* input using the new calculated K value. This process involves a reverse transformation from L*a*b* color values to CMYK color values with a fixed definition of GCR.
• This process, however, loses the K channel information or the CMY channel information during the translation between CMYK color spaces because the source color values are transformed to a three-dimensional intermediary color space during conversion to destination CMYK values. To preserve the K channel information, theerror reducer506 determines optimal K values in the destination color space that correspond to the K values in the source device color space, e.g., values between 0 and 255. These values can be created, for example, by generating a series of source K values ranging from minimum to maximum, fixing the source and destination CMY values at 0, and finding destination K values with minimum AE error relative to each of the source K values. These source and destination K values can be loaded into a lookup table for quick conversion of source K to destination K values. By using error reduction to determine optimal K values in the destination color space, thedevice link generator500 preserves K channel information. This results in improved accuracy of the K channel information when converting colors between CMYK devices.
• After loading the source and destination K values into a lookup table, when theerror reducer506 receives source L*a*b* and CMYK values, theerror reducer506 initially maps the source K channel to the destination K channel. The error reduction procedure is then used for varying the destination CMY values to obtain the best match for the respective L*a*b* values. If ΔE=0, control returns to the devicelink table builder502, which enters the calculated destination CMYK values into the device link table. If ΔE is greater than zero, then the destination CMY values corresponding to the destination K value in question are out of gamut relative to the target L*a*b* values. This may be, for example, because the source CMY values corresponding to K=0 result in a color that is out of gamut with the destination device, or because the destination K value in the particular region of destination CMY color space is either too high or too low, i.e., the mixture of K with CMY is such that the resulting color is too dark or too light relative to the targeted L*a*b* value.
• To reduce the ΔE error, K can be varied in a controlled way so as to ensure both optimal L*a*b*color and optimal matching of the K source channel behavior. This can be performed, for example, by alternately fixing the current CMY values while performing error reduction on variable K values and fixing the K value while performing error reduction on variable CMY values. When it is determined that neither varying CMY nor varying K improves the AE error, it can be assumed that the optimal CMYK values have been determined to satisfy both the color matching and K channel accuracy criteria. Control then optionally returns to the devicelink table builder502. While the above discussion assumes that theerror reducer506 performs the mapping between source and destination K values, it should be understood that the devicelink table builder502 can perform the mapping.
• It should be understood that other approaches can be used to improve the accuracy of the K channel information. For example, the PCS can be implemented as a color space having the same number of dimensions, e.g., four, as CMYK space Using a PCS having the same dimensionality as the device space prevents the loss of color channel information. In a specific example embodiment, the first three channels of this PCS are the PCS currently used by the system (e.g., LAB, L*a*b*, or XYZ). The fourth channel indicates a PCS value indicative of the black channel or relating to the black channel (e.g., L* or tristimulus value Y). The process can be performed in a manner similar to that performed by the ICC specification as in, for example, ColorSync 2.1 available from Apple Computer.
• FIG. 6 illustrates an example color transformation method600 according to the present invention. At ablock602, selected source device color coordinates are mapped to a PCS. Destination device color coordinates are then estimated as a function of the source device color coordinates, as depicted at ablock604. These estimated destination device color coordinates are then mapped to the PCS at ablock606.
• At ablock608, an error between the PCS values corresponding to the source and destination device color coordinates is determined. At adecision block610, the method determines whether the error satisfies a quality criterion, such as error minimization. In certain applications, the quality criterion can be defined as reduction of the error below a threshold value. If the error does not satisfy the quality criterion, flow proceeds to ablock612, at which the estimated destination device color coordinates are adjusted to reduce the error. This process repeats until the error is reduced.
• After the error is reduced, flow proceeds to ablock614, at which the optimal destination device color coordinates thus obtained are entered into a color map. Next, the method determines whether the color map is filled, as depicted at adecision block616. If the color map contains empty entries, flow proceeds to ablock618. New source device color coordinates are then selected, and then flow returns to theblock602. This process continues until the color map is filled. The color map can then be stored as, for example, a data file for future reference. The user can specify the desired source, destination, and intermediate profiles and the user preferences used to generate the device profile link. Upon recognizing that a color map has already been developed for a particular combination of these profiles, the system can load the data file. Loading the data file instead of reconstructing the color map saves computation time and other resources.
• The device profile link can be generated each time the user requests a new combination of device profiles. Alternatively, the user can specify in advance a series of source, intermediate, and destination profiles and allow the system to preprocess these lists of profiles into their respective device profile links and store them. When the user requests that a particular transform be performed on image data using a previously defined combination of source, intermediate, and destination profiles, the system retrieves the associated device profile link. Retrieving the device profile link improves the processing speed.
• While the above discussion has assumed that the device profile link describes a conversion between two device profiles, it should be understood that the device profile link can be used to describe a conversion between any number of device profiles. For example, N device profiles can be concatenated using a single device profile link. To concatenate the device profiles, the color conversion is performed using the PCS to convert colors between each device profile to be concatenated. Performing error reduction on the forward transforms between the individual device profiles improves the accuracy of the concatenated device profile link between the first and nth device profiles.
• The various embodiments described above are provided by way of illustration only and should not be construed to limit the invention. Those skilled in the art will readily recognize various modifications and changes that may be made to the present invention without strictly following the example embodiments and applications illustrated and described herein, and without departing from the true spirit and scope of the present invention, which is set forth in the following claims.

Claims (11)

1. A method for converting CMYK source device values to CMYK destination device values, comprising:

converting the CMYK source device values to a four color device-independent color space using a source profile with a four color device-independent color space;

converting the four color device-independent color space to the CMYK destination values using a destination profile with the four color device-independent color space; and

wherein three colors of the four color device-independent color space are used to define the color of a combined CMYK values and the remaining device-independent color space color is used to define the color of only a K colorant.

2. A method as defined inclaim 1, wherein the device-independent color space comprises a profile connecting space (PCS) used for converting the coordinates of the source profile to destination profile.

3. A method as defined inclaim 2, wherein the PCS comprises a Commission Internationale de l'Éclairage L*a*b (CIELAB) color space.

4. A method as defined inclaim 1, further comprising:

generating a color map or device profile link that maps colors between the source device and the destination device.

5. A system for converting CMYK source device values to CMYK destination device values, comprising:

a source device profile interpreter that converts the CMYK source device values to a four color device-independent color space using a source profile with a four color device-independent color space;

a destination device profile interpreter that converts the four color device-independent color space to the CMYK destination values using a destination profile with the four color device-independent color space; and

wherein three colors of the four color device-independent color space are used to define the color of a combined CMYK values and the remaining device-independent color space color is used to define the color of only a K colorant.

6. A system as defined inclaim 5, wherein the device-independent color space comprises a profile connecting space (PCS) used for converting the coordinates of a source profile to a destination profile.

7. A system as defined inclaim 6, wherein the PCS comprises a Commission Internationale de l'Éclairage L*a*b (CIELAB) color space.

8. A system as defined inclaim 7, further comprising:

a color map generator coupled to the source and destination device interpreters for generating a color map or device profile link that maps colors between the source device and the destination device.

9. A method for converting CMYK source device values to CMYK destination device values, comprising:

converting CMYK values to a four color profile connecting space (PCS) using a source profile with a four color PCS;

converting the four color PCS to the CMYK destination values using a destination profile with the four color PCS; and

wherein three colors of the four color PCS are used to define the color of a combined CMYK values and the remaining PCS color is used to define the color of only one particular colorant.

10. A method as defined inclaim 9, wherein the PCS comprises a Commission Internationale de l'Éclairage L*a*b (CIELAB) color space.

11. A method as defined inclaim 9, wherein the one particular colorant comprises a K colorant.

Images