System and method for the generation of correlationbased digital watermarks using uniformrosette color halftoning

0Associated
Cases 
0Associated
Defendants 
0Accused
Products 
7Forward
Citations 
0
Petitions 
1
Assignment
First Claim
1. An image forming method for digitally reproducing a moiré
 free color;
halftone image having an embedded correlationbased digital watermark using an enhanced halftone screen set consisting of a halftone screen for each of N colorants forming N color separations (where N≧
3), the method comprising;
providing a color input image to be watermarked;
providing a watermark image to be embedded in the Ncolor image;
generating N different uniform rosette halftone screen configurations each meeting uniform rosette halftone screen conditions;
generating N threedimensional threshold arrays each having a phase shift value as an input; and
halftoning the color input image by halftoning N different color separations using a different one of the halftone screen configurations and threedimensional threshold arrays for each color separation to produce a moiré
free color output image having the watermark image embedded therein.
1 Assignment
0 Petitions
Accused Products
Abstract
Systems and method for digitally reproducing a moiréfree color halftone output image having an embedded correlationbased digital watermark are provided using an enhanced halftone screen set consisting of a halftone screen for each of N colorants forming N color separations (where N≧3). The N different uniform rosette halftone screen configurations are generated such that each meets uniform rosette halftone screen conditions. A corresponding threedimensional threshold array is generated for each halftone screen configuration having a phase shift value as an input. The color input image is halftoned, halftoning each one of the N different color separations using a corresponding halftone screen configuration and threedimensional threshold array to form a moiréfree color output image having the correlationbased digital watermark embedded therein. A single key can be used to produce a full color image of the output image having the watermark image visible therein.
23 Citations
View as Search Results
APPARATUS FOR APPRAISING THE GENUINENESS OF PERSONAL IDENTIFICATION DOCUMENTS  
Patent #
US 20110033089A1
Filed 08/06/2010

Current Assignee
Toshiba Solutions Corporation, Toshiba Corporation

Sponsoring Entity
Toshiba Solutions Corporation, Toshiba Corporation

METHOD FOR INFRARED IMAGE CORRECTION AND ENHANCEMENT  
Patent #
US 20090073262A1
Filed 09/13/2007

Current Assignee
Monument Peak Ventures LLC

Sponsoring Entity
Eastman Kodak Company

Method for infrared image correction and enhancement  
Patent #
US 8,159,555 B2
Filed 09/13/2007

Current Assignee
Monument Peak Ventures LLC

Sponsoring Entity
Eastman Kodak Company

Stochastic temporal dithering for color display devices  
Patent #
US 9,811,923 B2
Filed 09/24/2015

Current Assignee
Snaptrack Incorporated

Sponsoring Entity
Snaptrack Incorporated

Predicting detectability and grading prior to printing  
Patent #
US 10,382,645 B2
Filed 03/12/2018

Current Assignee
Digimarc Corporation

Sponsoring Entity
Digimarc Corporation

Converting at least a portion of a 3D object into a format suitable for printing  
Patent #
US 10,452,055 B2
Filed 10/29/2014

Current Assignee
HewlettPackard Development Company L.P.

Sponsoring Entity
HewlettPackard Development Company L.P.

Encoded signal systems and methods to ensure minimal robustness  
Patent #
US 10,506,128 B1
Filed 06/18/2018

Current Assignee
Digimarc Corporation

Sponsoring Entity
Digimarc Corporation

Ncolor printing with hexagonal rosettes  
Patent #
US 7,679,787 B2
Filed 11/30/2006

Current Assignee
Xerox Corporation

Sponsoring Entity
Xerox Corporation

Moirefree color halftone configuration employing common frequency vectors  
Patent #
US 7,675,651 B2
Filed 11/30/2006

Current Assignee
Xerox Corporation

Sponsoring Entity
Xerox Corporation

Halftone patterns  
Patent #
RE41604E1
Filed 03/16/2000

Current Assignee
Megadot Systems Limited

Sponsoring Entity
Megadot Systems Limited

Moiréfree color halftone configuration  
Patent #
US 7,480,076 B2
Filed 12/08/2005

Current Assignee
Xerox Corporation

Sponsoring Entity
Xerox Corporation

MOIREFREE COLOR HALFTONE CONFIGURATION  
Patent #
US 20090091795A1
Filed 12/12/2008

Current Assignee
Xerox Corporation

Sponsoring Entity
Xerox Corporation

Image forming apparatus and method thereof  
Patent #
US 7,639,391 B2
Filed 09/23/2003

Current Assignee
Fuji Xerox Company Limited

Sponsoring Entity
Fuji Xerox Company Limited

METHOD FOR COHERENT WATERMARK INSERTION AND DETECTION IN COLOR HALFTONE IMAGES  
Patent #
US 20080151311A1
Filed 12/22/2006

Current Assignee
Xerox Corporation

Sponsoring Entity
Xerox Corporation

Halftone patterns  
Patent #
US 7,268,920 B1
Filed 03/16/2000

Current Assignee
Megadot Systems Limited

Sponsoring Entity
Megadot Systems Limited

Method for fourcolor zeromoire halftoning using nonorthogonal halftone cells  
Patent #
US 6,985,262 B2
Filed 07/13/2001

Current Assignee
Xerox Corporation

Sponsoring Entity
Xerox Corporation

Moirefree color halftone configuration  
Patent #
US 20060170975A1
Filed 12/08/2005

Current Assignee
Xerox Corporation

Sponsoring Entity
Xerox Corporation

Method and apparatus for multicolor printing using a rosette or diamond halftone screen for one or more of the colors  
Patent #
US 20050243340A1
Filed 04/30/2004

Current Assignee
Eastman Kodak Company

Sponsoring Entity
Eastman Kodak Company

Method for moirefree color halftoning using nonorthogonal cluster screens  
Patent #
US 6,798,539 B1
Filed 10/30/2000

Current Assignee
Xerox Corporation

Sponsoring Entity
Xerox Corporation

Digital watermarking using phaseshifted stoclustic screens  
Patent #
US 6,252,971 B1
Filed 04/29/1998

Current Assignee
Xerox Corporation

Sponsoring Entity
Xerox Corporation

Color image processing apparatus capable of suppressing moire  
Patent #
US 5,572,600 A
Filed 06/22/1994

Current Assignee
NEC Corporation

Sponsoring Entity
NEC Corporation

Electronic halftone generator  
Patent #
US 4,196,451 A
Filed 07/28/1978

Current Assignee
Xerox Corporation

Sponsoring Entity
Xerox Corporation

Electronic halftone generator  
Patent #
US 4,149,183 A
Filed 05/21/1976

Current Assignee
Xerox Corporation

Sponsoring Entity
Xerox Corporation

16 Claims
 1. An image forming method for digitally reproducing a moiré
 free color;
halftone image having an embedded correlationbased digital watermark using an enhanced halftone screen set consisting of a halftone screen for each of N colorants forming N color separations (where N≧
3), the method comprising;providing a color input image to be watermarked; providing a watermark image to be embedded in the Ncolor image; generating N different uniform rosette halftone screen configurations each meeting uniform rosette halftone screen conditions; generating N threedimensional threshold arrays each having a phase shift value as an input; and halftoning the color input image by halftoning N different color separations using a different one of the halftone screen configurations and threedimensional threshold arrays for each color separation to produce a moiré
free color output image having the watermark image embedded therein.  View Dependent Claims (2, 3, 4, 5)
 free color;
 6. A system for digitally reproducing an N (where N≧
 3) color channel moiré
free color halftone output image having an embedded correlationbased digital watermark, comprising;an input image source providing a color input image; image memory for storing the input image to be watermarked; watermark memory for storing the watermark image to be embedded in the color input image; and an image processor for generating N (where N≧
3) different uniform rosette halftone screen configurations each meeting uniform rosette halftone screen conditions, generating N threedimensional threshold arrays each having a phase shift value as an input, and halftoning the color input image by halftoning N different color separations using a different one of the halftone screen configurations and threedimensional threshold arrays for each color separation to produce a moiré
free color output image having the watermark image embedded therein.  View Dependent Claims (7, 8, 9, 10, 11, 12, 13, 14)
 3) color channel moiré
 15. A method of retrieving a correlationbased digital watermark embedded in an Ncolor channel color image using N halftone screens each meeting uniform rosette halftone screen conditions comprising:
overlying a key having a periodic halftone structure meeting a uniform rosette halftone screen condition on top of the color image.  View Dependent Claims (16)
1 Specification
Attention is directed to copending applications: U.S. application Ser. No. 11/897,772 filed Aug. 31, 2007, entitled “SYSTEM AND METHOD FOR THE GENERATION OF CORRELATIONBASED DIGITAL WATERMARKS”; U.S. application Ser. No. 11/897,826 filed Aug. 31, 2007, entitled “SYSTEM AND METHOD FOR THE GENERATION OF MULTILAYER CORRELATIONBASED DIGITAL WATERMARKS”; U.S. application Ser. No. 11/848,908 filed Aug. 31, 2007, entitled “SYSTEM AND METHOD FOR THE GENERATION OF MULTIPLE ANGLE CORRELATIONBASED DIGITAL WATERMARKS”; U.S. application Ser. No. 11/565,470 filed Nov. 30, 2006 entitled “NCOLOR PRINTING WITH HEXAGONAL ROSETTES”; U.S. application Ser. No. 11/565,455 filed Nov. 30, 2006 entitled “ROSETTE PRINTING WITH UP TO FIVE COLORS”; U.S. application Ser. No. 11/565,434 filed Nov. 30, 2006 entitled “MOIRÉFREE COLOR HALFTONE CONFIGURATION EMPLOYING COMMON FREQUENCY VECTORS” the disclosure found in these copending applications are hereby incorporated herein by reference in its entirety. The appropriate components and processes of the above copending applications may be selected for the teaching and support of the present application in embodiments thereof.
Disclosed in embodiments herein are methods and systems for generation of a color watermarked image having a watermark embedded in different color channels of the image using halftone structures that meet a uniformrosette condition, also referred to as minimumrosette halftone screens. The color watermark can be retrieved, or viewed, using an overlaid reference key constructed with halftone structures meeting similar uniformrosette conditions.
With the advent of inexpensive digital color printers, methods and systems of color digital halftoning have become increasingly important in the reproduction of printed or displayed images possessing continuous color tones. It is well understood that most digital color printers operate in a binary mode, i.e., for each color separation, a corresponding color spot is either printed or not printed at a specified image location or pixel. Digital color halftoning controls the printing of color spots, where the spatial averaging of the printed color spots by either a human visual system or a viewing instrument, provides the illusion of the required continuous color tones.
The most common halftone technique is screening, which compares the required continuous color tone level of each pixel for each color separation with one or more predetermined threshold levels. The predetermined threshold levels are typically defined for a rectangular cell that is tiled to fill the plane of an image, thereby forming a halftone screen of threshold values. At a given pixel if the required color tone level is darker than the threshold halftone level, a color spot is printed at that specified pixel. Otherwise the color spot is not printed. The output of the screening process is a binary pattern of multiple small “dots”, which are regularly spaced as determined by the size, shape, and tiling of the halftone cell. Conventional screening outputs can be considered as twodimensional repeated patterns, possessing two fundamental spatial frequencies, which are completely defined by the geometry of the halftone screens.
Halftoning images can also provide significant desirable visual effects. If two similar cluster halftone image patterns are superimposed on each other, the output can appear significantly different depending on the relative positions of the dot patterns as defined by their phase shifts.
Prior patents, such as U.S. Pat. No. 6,252,931 for “Digital watermarking using phaseshift stoclustic screens,” by S. Wang, hereby incorporated herein by reference in its entirety, describe a method to embed correlationbased phaseshift digital watermarks, also referred to a correlation marks, into halftone screens. By overlaying a transparency on the prints generated by the special halftone screen, for example as a public key, an invisible watermark embedded in the image can be retrieved.
For example, the two checkerboard patterns, 100 depicted in
Referring now to
Several different approaches may be taken to extend the method for embedding phaseshift correlationmarks for the black/white images, describe above, to color halftoning.
First, it can be achieved by conducting dotondot screening, i.e., applying the embedding to all color channels exactly the same way as for the black/white case. However, it is well known in color printing world that a dotondot color halftoning configuration suffers from misregistration between different color channels. Other drawbacks of the dotondot configuration includes reduced color gamut compared with other configurations. Therefore, the dotondot approach for correlationmarks is somewhat limited and many highquality color printers use separate halftone screens with different angles and/or frequencies for different channels.
The second possible approach for embedding correlationmarks into color images is embedding the watermarks into different channels independently based on the geometries of halftone screens being used for each color. However, different keys, or overlay transparencies, are required to retrieve the watermarks for each of the individual colors. Also, this approach provides satisfactory results for instances when correlationmarks are embedded into areas where one of the primary colors is dominating and the transparency for watermark retrieval can be selected correspondingly. For areas with mixed colors, the contrast of the retrieved watermarks will be severely reduced if the images are halftoned by using different rotated screens. Thus, this second approach also has limited applications.
Furthermore, the superposition of halftone screens for color printing can create interference patterns, known as moiré, which can be seen in the image, thus detracting from the visual appearance of the halftoned image. Significant efforts have been undertaken to reduce the undesirable effects of moiré in color halftoning.
An aspect of the disclosed system and method provided extends use and formation of correlation marks by enabling mulitcolor watermarked images to be formed by embedding a phase shifted (correlationbased) digital watermark into the color halftone used for each color separation, or channel, by halftoning the color input image with different one of a plurality of threedimensional threshold arrays where at least one input thereto is a phase shift value using a different halftone structure for each channel. Each halftone structure is generated meeting uniform rosette conditions, also known as minimum rosette conditions. The color separations are combined to form an output color image having the digital watermark embedded therein but not visible to the unaided eye. A single common key, in the form of a transparency having a matching periodic structure, can be used to retrieve the watermark from the output color image producing a full color image of the output image having the watermark image visible therein.
Disclosed in embodiments herein is a method for digitally reproducing a moiréfree color halftone image having an embedded correlationbased digital watermark using an enhanced halftone screen set consisting of a halftone screen for each of N colorants forming N color separations (where N≧3), including providing a color input image to be watermarked; providing a watermark image to be embedded in the Ncolor image; generating N different uniform rosette halftone screen configurations each meeting uniform rosette halftone screen conditions; generating N threedimensional threshold arrays each having a phase shift value as an input; and halftoning the color input image by halftoning N different color separations using a different one of the halftone screen configurations and threedimensional threshold arrays for each color separation to produce a moiréfree color output image having the watermark image embedded therein.
Also disclosed in embodiments herein is a system for digitally reproducing a moiréfree color halftone image having an embedded correlationbased digital watermark using an enhanced halftone screen set consisting of a halftone screen for each of N colorants forming N color separations (where N≧3), including an input image source providing a color input image; image memory for storing the input image to be watermarked; watermark memory for storing the watermark image to be embedded in the color input image; and an image processor for generating N (where N≧3) different uniform rosette halftone screen configurations each meeting uniform rosette halftone screen conditions, generating N threedimensional threshold arrays each having a phase shift value as an input, and halftoning the color input image by halftoning N different color separations using a different one of the halftone screen configurations and threedimensional threshold arrays for each color separation to produce a moiréfree color output image having the watermark image embedded therein. Wherein, as disclosed herein, the uniform rosette halftone screen conditions include: defining rosette fundamental frequency vectors V_{R1}, V_{R2 }that satisfy a length and sum requirement to meet visual acceptability standards according to V_{R1}>f_{min}, V_{R21}>f_{min}, and V_{R1}±V_{R2}>f_{min}; defining N halftone screens for color separation i=1, N, respectively (where N≧3), possessing first and second frequency vectors (V_{i1}, V_{i2}), where no two screens possess identical fundamental frequency vector pairs; and selecting fundamental frequency vectors for the N halftone screens according to (V_{i1}, V_{i2})=(m_{i1}V_{R1}+m_{i2}V_{R2}, n_{i1}V_{R1}+n_{i2}V_{R2}) for integer m'"'"'s and n'"'"'s, where for each color separation i, at least one fundamental frequency vector or its conjugate must also satisfy the following inequality: V_{ik}>max [V_{R1}, V_{R2}, min[V_{R1}+V_{R2}, V_{R1}−V_{R2}]], k=1 or 2.
The various embodiments described herein are not intended to limit the invention to those embodiments described. On the contrary, the intent is to cover all alternatives, modifications, and equivalents as may be included within the spirit and scope of the invention as defined by the appended claims.
As more particularly set forth below, the disclosed system and methods are directed to the generation of color images having digital watermark, including generating color watermarked images using uniformrosette halftone structures corresponding to different color channels such that the watermarks may be retrieved, or viewed, using the same public key overlaid atop the images.
Referring again to
The halftone output with the embedded watermark pattern looks like the halftone pattern 300 shown in
For example,
Extending this method for embedding phaseshift correlationmarks to color halftoning presents the issues with moiré, as described above, created by the superimposed color channel halftone screens. Moiré can be eliminated using uniformrosette, also referred to as minimum rosette, halftone screens as disclosed in copending application U.S. application Ser. No. 11/565,470 filed Nov. 30, 2006 entitled “NCOLOR PRINTING WITH HEXAGONAL ROSETTES” previously incorporated by reference herein in its entirety. The disclosures provided therein supply teachings to systems and methods for digitally reproducing a moiréfree color halftone image using an enhanced halftone screen set consisting of a nonorthogonal halftone screen for each of N colorants (color channels), with N being 3 or greater, which meets a uniformrosette condition.
When a uniformrosette condition is satisfied, the combination of multicolor channel halftones appears as repeated rosettes with a uniform hexagonal shape 500 as shown
Clustereddot halftoning methods produce halftoned images that possess strong periodic structures defined by the halftone screens. For the purposes of moiré analysis, these images can be well described in the frequency domain by discrete components without regard to amplitude and phase. In the frequency domain representation, a twodimensional halftone for a single colorant x forms a frequency lattice that can be described by two generator vectors V_{x1}, V_{x2 }The generator vectors correspond to the two fundamental frequencies of the halftone and can be referred to as fundamental frequency vectors. Linear combinations of these fundamental frequency vectors form all points (harmonics) in the lattice and no frequencies are formed that are lower than the fundamentals.
Moiré analysis for conventional color halftoning typically focuses on Cyan, Magenta, and Black. The fundamental vector frequencies can be denoted V_{c1}, V_{c2}, V_{m1}, V_{m2}, V_{k1}, and V_{k2}, respectively, where the sub 1 and sub 2 refer to vectors that are above (0° to 90°) or below (270° to 360°, or equivalently −90° to 0°) the 0° axis, respectively. This notation is used within the present disclosure to represent the two quadrants. Unless otherwise noted, the subscripts c, m, and k, are used to aid in teaching the presently described halftoning processes due to the common practice of threecolor printing with cyan, magenta, and black however, the concepts are general in that other colorants may be used. For example, the notation V_{m1 }is referred to as a frequency vector for the magenta screen, it is to be understood that it can be intended generally to imply a frequency vector in the first quadrant for some available colorant. Further, colorants for particular screen geometries can be interchangeable. For example, cyan is halftoned with a screen of a first geometry, and black halftoned with a screen of a second geometry, but it is practical and reasonable to assume that the screens may be interchanged and cyan may be halftoned with the screen of the second geometry and black the first.
In color printing, the superposition of halftone screens creates more frequency components than exist in the single separations of the various individual process colors. The result caused by superposition of two different colors can be expressed as their frequencyvector difference, e.g., V_{cm}=V_{c}±V_{m}, where V_{c }and V_{m }are two frequency components from C and M, and V_{cm }is the difference vector. Since each Fourier component has its conjugate, i.e., there is always a frequency vector −V_{c }that represents the conjugate component of V_{c}, the sign definition of frequency vectors is somewhat arbitrary. For each halftone screen, there are two fundamental frequency vectors, therefore, the color mixing of two screens for two different colors yields eight difference vectors by the fundamental frequencies alone. Considering other harmonics of the halftone frequencies, the combinations can yield a large number of difference vectors.
Referring now to
When using classical angles and frequencies, such as the traditional configuration for the 170 cpi halftone screen set, the rosette pattern 700 is not a simple repeated pattern. Rather, it has been shown that these rosette patterns 700 are theoretically nonperiodic in that the same shape and configuration does not repeat on the same page. (J. A. C. Yule, “Principles of Color Reproduction”)
Most halftone screens used in color reproduction are orthogonal screens, or screens in rectangular shapes including squares. However, for most digital printing devices, the size and shape of halftone screens are constrained by the “digital grid”, which defines the locations of printed dots, and therefore, an exact 15° or 75° rotation of a cluster screen is impossible. Although there are many alternative approaches for moiréfree color halftoning, most of them only provide approximate solutions and/or have a tendency to generate additional artifacts associated with halftone outputs. The difficulty to achieve moiréfree color halftoning is greatly relieved by using nonorthogonal halftone screens, i.e., screens in general parallelogram shapes.
The common strategy to avoid objectionable twocolor moiré is to select frequency vectors that ensure that no twocolor difference vector of the fundamental halftone frequency vectors is sufficiently small, or short in length, to be perceived as a noticeably low frequency. The twocolor moiréfree condition can be summarized by
V_{c}±V_{m}>V_{high}, (1)
where V_{c }represents any one of V_{c1}, −V_{c1}, V_{c2}, −V_{c2}; V_{m }represents any one of V_{m1}, −V_{m1}, V_{m2}, −V_{m2}; and V_{high }is a frequency limit set at somewhere between 5070 cyclesperinch for just noticeable moiré.
It is well known that a troublesome moiré is the threecolor moiré, which can appear in cyanmagentablack prints produced by CMYK fourcolor printers. As an extension of the twocolor case, one aspect of the threecolor moiréfree condition can be summarized by
V_{c}±V_{m}±V_{k}>V_{high}, (2)
where V_{k }represents any one of V_{k1}, −V_{k1}, V_{k2}, −V_{k2}; and V_{high }is set similar to the twocolor case. Since there are altogether thirtytwo unique combinations of different color components for the left side of the inequality of Equation (2), it stands as a matter of practicality that to make all threecolor difference vectors as well as all twocolor difference vectors large enough to avoid any color moiré is very difficult, unless the halftone screens have very high frequency fundamentals, say higher than 200 lineperinch. Another aspect of the moiréfree condition is to make two of the threecolor difference vectors null while keeping the rest large. Given that both the signs and the indices of frequency vectors are defined somewhat arbitrarily, without losing the generality, the threecolor moiréfree condition can be specified by the following vector equation:
V_{c1}−V_{m1}+V_{k2}=0, (3a)
or, equivalently due to the conventional screen configuration,
V_{c2}−V_{m2}−V_{k1}=0. (3b)
The Equations (3a) and (3b), are two of all possible frequency combinations of the three colors. In most practical applications, the rest of the combinations satisfy the inequality of Equation (2) for V_{high }as large as min[V_{c}, V_{m}, V_{k}] and are not specially specified, and the combination of halftone outputs produce a rosette appearance rather than objectionable moiré.
Most conventional halftone screens use squareshape halftone cells for tiling. Therefore, the two fundamental frequency vectors of each screen are not independent to each other. Once one of the two equations, either (3a) or (3b) is satisfied, the other one is automatically held. Recently, Wang et al. has taught halftone methods (U.S. Pat. No. 6,798,539, Wang et al.) using nonorthogonal halftone cells to construct halftone screens, or general parallelogramshape halftone cells, for moiréfree color halftoning, in which case the two fundamental frequency vectors of each parallelogramshapebased screen are independent to each other and thus satisfying both Equations (3a) and (3b) is required for the threecolor moiréfree condition. We note that the term “nonorthogonal” as used in the present specification here refers to “not necessarily square,” which is less restrictive than “strictly not orthogonal.” Such terminology follows convention used in mathematics, where terms such as “nonlinear” refers to “not necessarily linear.”
Further concerning moiréfree nonorthogonal halftone configurations, Wang, in US Publication No. 2006/0170975 A1, disclosed a moiréfree color halftone configuration for clustered dots. Unlike conventional methods, the disclosed method produces periodic hexagon rosettes of identical shapes. These exemplary hexagon rosettes have three fundamental spatial frequencies exactly equal to half of the fundamental frequency of the three halftone screens. The resultant halftone outputs are truly moiré free, as all the fundamentals and harmonic frequencies are multiples of and thus higher in frequency than the rosette fundamental frequency.
An example of a hexagonal rosette halftone configuration is easy to understand through extension of the previous classical configuration illustrated in
As shown in
Further, as shown in
The halftoning method, and resulting configuration taught herein is based on defining rosette fundamental frequency vectors, of sufficiently high frequency and angle separation, that can be used to generate a hexagonal lattice of rosette harmonics. The lattice is generated by linear combinations of the rosette fundamental frequency vectors. Angles and frequencies for individual halftone screens are chosen from the rosette lattice points. A screen set selected in such a manner is moiré free because no combination of frequency lattice points can produce a beat lower than the two rosette frequency vectors used to generate the lattice. The lattice structure defined by the rosette makes it possible to choose pairs of frequency vectors for an almost arbitrary number of colorants without introducing any moiré in an Ncolor combination. Practical frequency lattices can be realized through the use of nonorthogonal screens.
To better understand this rosette vector concept, consider the example of
V_{c1}=2V_{R1}−V_{R2 }
V_{c2}=2V_{R2 }
V_{m1}=V_{R1}+V_{R2 }
V_{m2}=−2V_{R1}+2V_{R2 }
V_{k1}=2V_{R1 }
V_{k2}=−V_{R1}+2V_{R2 }
The conjugate fundamental frequency vectors are also shown in
This teaching can be extended to derive new screen configurations with an arbitrary number (N) of halftone screen separations for use in embedding phaseshift correlationmarks with color halftoning.
A mathematical statement of the general principle of Nhalftone latticebased screen configurations can be written by considering a frequency lattice structure defined by rosette vectors V_{R1}, V_{R2}. To avoid objectionable low frequency texture and moiré we require
V_{R1}>f_{min}, V_{R2}>f_{min}, and V_{R1}±V_{R2}>f_{min}. (4)
Typically, to meet visual acceptability standards, f_{min }can be set to be approximately 100 cycles/inch. But, lower quality printed material could use f_{min }as low as 50 cycles/inch, or lower, while high quality printed material might require f_{min}=150 cycles/inch.
For a screen configuration with N halftone screens, let V_{i1}, V_{i2 }respectively denote first and second fundamental frequency vectors for screens i=1, 2, . . . , N, where V_{i1}, V_{i2 }are chosen to satisfy
(V_{i1},V_{i2})=(m_{i1}V_{R1}+m_{i2}V_{R2},n_{i1}V_{R1}+n_{i2}V_{R2}) for integer m'"'"'s and n'"'"'s. (5)
Exemplary halftone configurations are generated by utilizing configurations where halftone fundamental frequency vectors can be selected from any lattice points beyond the firstorder spectral hexagon. To deviate from the fundamental rosette frequencies, for each color separation i, at least one fundamental frequency vector or its conjugate must also satisfy the following inequality: V_{ik}>max [V_{R1}, V_{R2}, min[V_{R1}+V_{R2}, V_{R1}−V_{R2}]], k=1 or 2.
Often, it is desirable to avoid dotoffdot/dotondot configurations due to misregistration sensitivity. That can be achieved by specifying halftone fundamental frequency vectors such that any (V_{i1}, V_{i2}) vector pair is not equal to any other (V_{j1}, V_{j2}) vector pair. That is, no two screens possess identical fundamental frequency vector pairs. On the other hand, if dotoffdot/dotondot screens are desired, say for reasons of increased gamut or reduced texture, then at least one (V_{i1}, V_{i2}) vector pair is specified to equal another (V_{j1}, V_{j2}) vector pair.
It can be desirable to use screens that do not require large m or n, say >6 or 8, because the frequency of those screens might be beyond the resolution of common printing processes, but note that certain high resolution processes (e.g., high resolution proofers) may allow a much higher frequency.
To achieve a balance between minimizing visibility and maximizing stability, it is often desirable to have particular screens at a relatively low frequency, but not as low as the rosette frequency. For example, screens for C, M, and K could be chosen such that
2≦m_{i1}+m_{i2}+n_{i1}+n_{i2}≦4 (6)
Line screens can be generated for a halftone configuration according to the teachings herein by selecting m'"'"'s and n'"'"'s such that one vector of the pair has length 0 (V_{i}=0).
For printing dot screens with 3 or more color separations it is typically desirable to have the rosette vectors V_{R1}, V_{R2 }separated by ≈30° or ≈60° to provide a pleasant hexagonal appearing rosette.
Thus, systems and methods for embedding phase shift digital watermarks with color halftoning, as described herein, include generating a plurality of halftone screen configurations, one for each color channel used in printing the final color image, created from the result of a uniformrosette halftone screen design. An example set of clustered dot halftone screens that satisfies the uniformrosette halftone design is provided below. In the present example, four halftone screens are chosen, one corresponding to each of four separate color channels, or color separations, used in forming the primary color output image 1500. However it should be appreciated that N halftone screen configurations can be used, corresponding to N separate color channels of the color image, where N≧3.
The four (4) halftone screen configurations for this exemplary halftone screen set, can be described by their fundamental frequency vectors (V_{i1}, V_{i2}) for i=1 to 4, or their corresponding spatial vectors (v_{i1}, v_{i2}) for i=1 to 4. The halftone screen configurations meeting the uniformrosette halftone screen design are generated by defining rosette fundamental frequency vectors V_{R1}, V_{R2 }that satisfy a length and sum requirement to meet visual acceptability standards according to V_{R1}>f_{min}, V_{R2}>f_{min}, and V_{R1}±V_{R2}>f_{min}. Four halftone screens for colorants i=1 to 4, respectively, are defined, each possessing first and second frequency vectors (V_{i1}, V_{i2}), where no two screens possess identical fundamental frequency vector pairs. Fundamental frequency vectors are selected for the four halftone screens according to (V_{i1}, V_{i2})=(m_{i1}V_{R1}+m_{i2}V_{R2}, n_{i1}V_{R1}+n_{i2}V_{R2}) for integer m'"'"'s and n'"'"'s, where for each color separation i, at least one fundamental frequency vector or its conjugate must also satisfy the following inequality: V_{ik}>max [V_{R1}, V_{R2}, min[V_{R1}+V_{R2}, V_{R1}−V_{R2}]], k=1 or 2; so that the set of four different halftone screens used in combination produce moiréfree halftone output.
For this example, Color1 (V_{11}, V_{12}) will be used for cyan, Color2 (V_{21}, V_{22}) will be used for magenta, Color3 (V_{31}, V_{32}) will be used for black, and Color4 (V_{41}, V_{42}) will be used for yellow. All frequency components are measured in units of cyclesperinch (cpi), and frequency angles are defined with respect to the horizontal yaxis.
 Rosette: V_{R1}=90.1 cpi @ 56.3°,
 V_{R2}=90.1 cpi, @ −56.3°
 Color1: V_{11}=2V_{R1}+V_{R2}=167.7 cpi @ 26.6°,
 V_{12}=2V_{R2}=180.2 cpi @ −56.3°
 Color2: V_{21}=2V_{R1}=180.2 cpi @ 56.3°,
 V_{22}=V_{R1}+2V_{R2}=167.7 cpi @ −26.6°
 Color3: V_{31}=V_{R1}−V_{R2}=150 cpi @ 90°,
 V_{32}=2V_{R1}+2V_{R2}=200 cpi @ 0°
 Color4: V_{41}=2 V_{R1}=180.2 cpi @ 56.3°,
 V_{42}=2 V_{R2}=180.2 cpi @ −56.3°
The spatial vectors (v_{R1}, v_{R2}) and (v_{i1}, v_{i2}) for i=1 to N, where N=4 corresponding to these frequency vectors (wherein lower case v is used to define the spatial vectors as distinguished from the corresponding frequency vectors defined by upper case V) are:
 Rosette: v_{R1 }(−12, 8),
 v_{R2 }(12, 8)
 Color1: v_{11 }(−3, 6),
 v_{12 }(6, 4)
 Color2: v_{21 }(−6, 4),
 v_{22 }(3, 6)
 Color3: v_{21 }(6, 0),
 v_{32 }(0, 8)
 Color4: v_{41 }(−6, 4),
 v_{42 }(6, 4)
The spatial vectors (v_{R1}, v_{R2}) and (v_{i1}, v_{i2}) for i=1 to N, where N=4, can be derived from the frequency vectors (V_{R1}, V_{R2}) and (V_{i1}, V_{i2}) for i=1 to N, where N=4, as taught by U.S. Pat. No. 6,985,262 to Wang, et al which is hereby incorporated herein by reference in its entirety. For the purposes of illustration, the spatial vectors (v_{i1}, v_{i2}) for i=1 to N, where N=4, each defining a minimum rosette halftone screen configuration to be used for a different color separation are shown at 1000, 1100, 1200 and 1300 in
Next, threedimensional threshold arrays are generated, one for each uniformrosette halftone screen configuration, to provide a smooth phase transition between 0 and π for portions of the halftone screens corresponding to the outer region of the digital watermark image as disclosed by copending applications: U.S. application Ser. No. 11/897,772 filed Aug. 31, 2007 entitled “SYSTEM AND METHOD FOR THE GENERATION OF CORRELATIONBASED DIGITAL WATERMARKS”; U.S. application Ser. No. 11/897,826 filed Aug. 31, 2007 entitled “SYSTEM AND METHOD FOR THE GENERATION OF MULTILAYER CORRELATIONBASED DIGITAL WATERMARKS”; and U.S. application Ser. No. 11/848,908 filed Aug. 31, 2007, entitled “SYSTEM AND METHOD FOR THE GENERATION OF MULTIPLE ANGLE CORRELATIONBASED DIGITAL WATERMARKS” previously incorporated herein by reference in their entirety. This outer region can be considered to correspond to the region 210 of
Unlike conventional halftoning techniques, the threshold value for a particular pixel in a particular color separation (or color channel) is not chosen from a halftone screen based solely on the spatial coordinates, x and y, of the pixel, but rather the digital watermark embedding method uses a separate threedimensional threshold array for each color separation, each having an additional dimension specified by the phase shift, or more particularly the phase shift relative to an initial zeroshift halftone screen. During the halftone process, the threshold value for each pixel in each color separation is chosen from the corresponding threedimensional array (e.g., array 1752 in
For halftoning watermark images specified by 8 bits, or gray levels between 0 to 255, a common design of the twodimensional threshold array with a given cluster geometry can be described mathematically as
T(x,y)=128−127·{ cos [kπ(x·y_{a}+y·x_{a})]+cos [kπ(x·y_{b}+y·x_{b}(]}/2, (7)
where k is a scaling factor constant.
The equation, sometimes referred to as the dot profile, provides rounddot or roundhole shapes for the halftone outputs in the high light or the shadow part of an image, and checkerboardlike patterns for the middle tones. The dot profile T(x, y) in Equation 7 is used as the initial zeroshift halftone screen. The threedimensional threshold array, which is also a function of the phase shift s used for watermark embedding, can be obtained by using a slight modification of Equation 7, and expressed as
T(x,y)=128−127·{ cos [kπ(x·y_{a}+y·x_{a})+s]+cos [kπ(x·y_{b}+y·x_{b})+s]}/2, (8)
where s is the phase shift in radians.
The resolution of the phase shift depends on the application. In general, a higher resolution provides better watermark hiding but requires larger memory space to store the threedimensional array. For the purposes of this example, Ng is chosen to be equal to 255, which provides practical implementation of 8 bit images used in many applications. Ng can be chosen to be larger than 255 to provide greater resolution, or it may be less than 255 to require less memory to implement.
Therefore, it is possible to interpret the graylevels in the watermark image in terms of desired amount of phase shift. To embed a watermark into each color halftone separation, a π shift for all the black areas of the watermark image and no shift for the white background of the watermark image is needed. Consider using 0 for the white and 255 for a complete black, we may interpret the white, or the gray level 0, as a zero phase shift and the complete black, or the gray level 255, as a π phase shift. In other words,
s=g·π/N_{g}, (9)
where g is the gray level, Ng=255 is the total number of gray levels and s is the phase shift. A smooth phase transition can hide seams caused by the imposition of the watermark image, as described above.
Thus, four (4) threedimensional threshold arrays T(x, y) each corresponding to a different color separation used in producing the color output image are generated by using the spatial coordinates x and y as the first two dimensions and the phase shift, represented by the variable s in Eq. (8), as the third dimension. Considering no phase shift as a black (0) and a full πphase shift as a white (255), the phase shift s, varying from 0 to π as shown at 1002, 1102, 1202 and 1302 in
The four (4) threedimensional threshold arrays are then aligned so that for the position x=0 and y=0, the value of the phase shift s is also equal to zero.
The four (4) threedimensional threshold arrays can then be stored in memory of a halftoning unit for use as described below. The threshold values for given x, y and s values can be calculated pixelwise during the embedding process described below using Eq. 8, but storing a precalculated result into memory can reduce the processing time for the halftone process, if so desired.
A primary color image to be halftoned is then selected. The image is a multicolor (Ncolor) image, having N colors where N≧3. In the example provided herein, the image is a fourcolor image (N=4) using four separate halftone screen configurations, one for each color channel or separation to produce the output color image.
A black/white watermark image is selected to be embedded into the primary color image as a correlationbased watermark. For the purposes of example, a black/white Xerox logo image 1400 shown in
A 0 to π phase shift is then determined along all edges of the watermark image 1400 so that, in the image edges, all gray levels between 0 and Ng represent different intermediate phase shift steps between phase zero and phase π.
Referring to
Next, the primary color image is halftoned to embed the watermark image 1410 into the primary color image to produce an output color image having a correlationbased digital watermark shown at 1500 in
Referring to
Briefly, the watermark embedding process can be summarized as the series of steps generally illustrated in accordance with the block diagram of
Similarly, a watermark memory 1734 is employed for storing a black/white watermark image 1730 (1400). System 1710 further includes an image processor 1750 or similar control and processing circuitry, such as a digital front end (DFE) known for use in the processing of digital images for rendering on printing engines and reprographic devices (e.g., Xerox® iGen3™, DocuColor™ and WorkCentre™ systems, etc.). The processor 1750 is employed for embedding the digital watermark image 1730 into the input image 1720 to produce moiréfree digitally watermarked output image 1740, using halftone structures meeting the minimum rosette conditions described above.
As will be appreciated the output images 1500 may be rendered on any image output device such as a suitable marking or printing engine 1770 capable of rendering color output on one or more media.
The processor 1750 generates N different halftone screen configurations, one for each of the N different color separations which are combined to form the output color image 1740. Each halftone screen configuration meeting the minimum rosette configuration conditions which include: defining rosette fundamental frequency vectors V_{R1}, V_{R2 }that satisfy a length and sum requirement to meet visual acceptability standards according to V_{R1}>f_{min}, V_{R2}>f_{min}, and V_{R1}±V_{R2}>f_{min}; defining N halftone screens for color separations i=1, N respectively (where N≧3), possessing first and second frequency vectors (V_{i1}, V_{i2}), where no two screens possess identical fundamental frequency vector pairs; and selecting fundamental frequency vectors for the N halftone screens according to (V_{i1}, V_{i2})=(m_{i1}V_{R1}+m_{i2}V_{R2}, n_{i1}V_{R1}+n_{i2}V_{R2}) for integer m'"'"'s and n'"'"'s, where for each color separation i, at least one fundamental frequency vector or its conjugate must also satisfy the following inequality: V_{ik}>max [V_{R1}, V_{R2}, min[V_{R1}+V_{R2}, V_{R1}−V_{R2}]], k=1 or 2.
System 1710 also generates a plurality of threedimensional (3D) threshold arrays T(x, y) 1752, each having a phase shift as an input value and corresponding to a different on of the N color separations used to form the output color image, as described above, and stores the result into memory. Alternatively, the threshold values for a plurality of given x, y and s values can also be calculated pixelwise while halftoning each color separation using Equation (8) above. Storing the precalculated results into the processor memory, in 3D threshold arrays 1752 is intended for speeding up the halftone process. Accordingly, it will be appreciated that various halftone result generation and storage techniques may be employed in other alternative embodiments.
A lowpass filter 1754 may be applied to smooth out edges of the watermark image and the resultant image is then stored in memory as a multibit gray image (e.g., 8bit), where the different gray levels represent different phase shifts for watermark embedding. If the original watermark pattern does not contain any highfrequency details this step may be omitted, if so desired.
The input image 1720 and the processed watermark image obtained previously are then read in by the image processor 1750 and a pixelwise halftoning operation is conducted for each of the N color separations (color channels) used to form the output color image, such as the four color separations used in the example provided herein. In accordance with the disclosed embodiments, a different one of the threedimensional threshold arrays 1752, stored in memory accessible to the processor 1750, is employed as an input to a threshold operation 1756 for each color separation. In response to input image location coordinates x and y, the input color value for a given color separation of the input image is compared to the threshold value determined by the coordinates x, y, and the phase shift s given by Equation (8), and a resulting gray level g is determined for a plurality of coordinate locations to produce each of the N color separations which are then combined to form the processed color output image 1740 having the correlationbased digital watermark image embedded therein.
As will be appreciated by those familiar with the design of image processing systems, the image processor 1750 further includes timing and control operation block 1758, which controls the flow of data and processing operations within the image processor, including any buffering of the image data as depicted in buffers 1760 and 1762. A wide variety of hardware may be employed to achieve the functionality depicted with regard to the image processor, including dedicated image processing chipsets and conventional computer workstations, as well as combinations thereof or other processing devices.
The watermarked color output image containing the embedded invisible digital watermarks, is then provided as input to the printing engine 1790 for rendering. A single public key 1600, 1610 can be used for retrieving the watermark producing a full color image 1510 of the output image 1500 having the watermark image 1520 visible therein.
It will be appreciated that various of the abovedisclosed embodiments and other features and functions, or alternatives thereof, may be desirably combined into many other different systems or applications. Also, various presently unforeseen or unanticipated alternatives, modifications, variations or improvements therein may be subsequently made by those skilled in the art which are also intended to be encompassed by the following claims.