JP2000358154A - Electric watermark capable of adjusting visibility of watermark - Google Patents

Abstract

PROBLEM TO BE SOLVED: To prepare picture data with a watermark expressing a color picture with a watermark by converting color picture data of a first color system to a second color system and inverting color picture data of the obtained second color system to the first color system through the use of an inversion matrix. SOLUTION: In embedding a watermark to an original picture P, a mark picture M including a watermark to be embedded is prepared in addition to this picture P. This picture M is a binary picture and is converted from an RGB color system to a YCbCr color system first. A pixel at a position equivalent to the effective area of the watermark is converted by using an embedding conversion matrix obtained by adding an original conversion matrix and a watermark matrix. At the time of executing it concerning all the pixels, YCbCr picture data F with a watermark embedded with a digital watermark is obtained. At the time of operating an inversion matrix with respect to this YCbCr picture data F with watermark, RGB picture data P, with a watermark are obtained.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

[0001] 1. Field of the Invention [0002] The present invention relates to a digital watermark technique for electronically embedding a watermark mark in a color image.

[0002]

2. Description of the Related Art With the development of computer networks such as the Internet, digitization of information has progressed, and many users can easily access necessary information. On the other hand, digital content in which copyright is generated in digital information is becoming an environment where data can be easily copied without asking the author, and the problem of copyright infringement due to unauthorized copying has been attracting attention. ing. Therefore, digital watermark technology for embedding watermark information such as copyright information in color image data has been receiving attention for the purpose of preventing copyright infringement of a color image, which is the main information of digital content.

[0003] Regarding this type of conventional digital watermarking, many technical proposals have been made for grayscale images. Considering the luminance component (Y component) of a color image, it can be handled in the same way as a gray image, and a method of embedding an electronic watermark in a luminance component corresponding to the gray image has been used.

[0004]

However, there have been almost no proposals to embed an electronic watermark using a signal or data unique to a color image. "Deep information recording of color images" using the color space of a color printer as a digital watermark embedding technology (IEEJ, Annual Conference Proceedings, 7, 20, pp. 47-48), skillfully using visual characteristics Model construction type “a method of color watermarking considering visual characteristics” (IMPS 97, I-3.14,
pp. 45-46 (1997)), "A Technique for Embedding Digital Watermarks in Digital Images" (SITA 97, V
ol. 2, pp. 541-544 (1997)). These techniques have the disadvantage that the model is complicated and the watermark embedding process is not easy. It has also been pointed out that the method of embedding a watermark in the lower bits of a color image degrades the image quality of the color image.

SUMMARY OF THE INVENTION It is an object of the present invention to provide a digital watermark technique for embedding a watermark mark using a characteristic unique to a color image.

[0006]

In the present invention, color image data of a first color system is converted into color image data of a second color system different from the first color system. A transformation matrix to be transformed and an inverse transformation matrix for performing an inverse transformation of the transformation are prepared. Further, an additional matrix is set in which real values x and -x are used as components and the sum of the components in each row and each column is substantially zero. Then, a conversion matrix for embedding a watermark is created by adding the additional matrix to the conversion matrix. Next, a watermark mark to be embedded in the color image is prepared. At a pixel corresponding to a position in the watermark mark, the color image data of the first color system is converted into the second color image by using the conversion matrix for watermark embedding. Convert to system. The color image data of the second color system obtained by this conversion is inversely transformed to the first color system by using an inverse transformation matrix, so that a watermark representing a watermarked color image in which a watermark mark is embedded. Create attached image data. Further, at the time of setting the additional matrix, by adjusting at least one of the real value x and the array of each component in the additional matrix,
Adjust the visibility of the watermark mark in the watermarked color image.

According to the present invention, a watermark mark can be embedded in a color image by performing conversion and inverse conversion using a conversion matrix and an inverse conversion matrix between two color systems and an additional matrix. When the value of the real value x defining the components of the additional matrix changes, the color of the watermark mark changes. Also,
The color of the watermark mark also changes by changing the arrangement of each component of the additional matrix. Therefore, the visibility of the watermark mark can be adjusted by adjusting at least one of the value of the real value x of the additional matrix and the arrangement of each component in the additional matrix.

Here, as the first or second color system, an RGB color system using three primary colors of light as a parameter and a YCbCr color system using a luminance signal and a color difference signal as parameters can be appropriately adopted. . These color systems are often used when dealing with image data,
And the mutual transformation matrix is well known. Therefore, there are a number of retouching software and the like that handle image data in the future, and digital watermarks can be easily embedded.

Note that the watermark mark may include a plurality of mark groups spatially separated from each other. At this time, in at least some of the plurality of mark groups, the visibility can be adjusted by setting the real value x to a value different from the other mark groups. This makes it possible to embed a plurality of mark groups having different visibility in the color image.

Further, the watermark mark may include a mark area composed of spatially continuous pixels, and the mark area may be divided into a plurality of division areas. At this time,
The visibility can be adjusted by setting the real value x to a value different from that of the other divided areas in at least some of the divided areas. By doing so, it is possible to embed a watermark mark in a color image that includes a mark area composed of a plurality of section areas having different visibility.

The visibility can be adjusted by exchanging columns or rows of the additional matrix. By doing so, it is possible to set the watermark mark to a desired visibility simply by exchanging columns or rows.

When the watermark mark in the watermarked color image is visible to the naked eye, the visibility of the watermark mark can be reduced. At this time, first, an invisible matrix is prepared, which is an inverse matrix of a product of an inverse transformation matrix and the additional matrix and a unit matrix having the same degree as the transformation matrix. Then, in a pixel corresponding to a position in the watermark mark, the visibility of the visible watermark mark is reduced by operating the invisible matrix on the watermarked image data.

Further, the image data of the YCbCr color system obtained by the conversion using the conversion matrix for watermark embedding is
When performing the inverse conversion to the RGB color system using the inverse conversion matrix, it is preferable to perform underflow correction and overflow correction so that each pixel value in the image data of the RGB color system becomes an integer value within a predetermined range. This way,
The result of the inverse conversion can be contained within the range that the image data of the RGB color system can take.

The real value x of the additional matrix is preferably set to a value in a range from about 0.01 to about 0.7. By doing so, the rate of occurrence of underflow correction and overflow correction can be kept low, so that substantial degradation in image quality due to embedding of a watermark mark can be reduced.

The present invention is a digital watermarking method and apparatus, a computer program for realizing the functions of the method or apparatus, a recording medium storing the computer program, and embodied in a carrier wave including the computer program. It can be realized in various modes such as a data signal.

[0016]

DETAILED DESCRIPTION OF THE INVENTION Principles of Digital Watermark Embedding Processing and Decoding Processing: Embodiments of the present invention will be described below based on examples. FIG. 1 is an explanatory diagram of the principle of the process of embedding a digital watermark in a digital color image in the embodiment. FIG. 2 is an explanatory diagram showing the contents of the embedding process.

FIG. 2A shows a digital color image which is an original image expressed in the RGB color system. The original image P is defined as an aggregate of a large number of pixels arranged in a matrix, and the size of the original image P is defined by a horizontal resolution Hsize and a vertical resolution Vsize. The unit of the horizontal resolution or the vertical resolution is the number of pixels.

In this specification, image data representing the color of each pixel of the original image P is defined by a matrix P = [r, g, b] ^t (the superscript “ ^t ” indicates a transposed matrix). . here,
r, g, and b are three primary colors of light, red (r), green (g),
The luminance of each color component of blue (b). The luminance of each color is determined by a predetermined resolution (for example, 256 from 0 to 255).
(Gradation). In addition to the RGB color system, a YCbCr color system composed of a luminance signal and two color difference signals is widely known as a color system defining the color of each pixel. In the present specification, image data representing the color of each pixel in the YCbCr color system is represented by a matrix F = [y, cb, cr].
Defined by ^t .

These two color systems are represented by the following (1a),
As shown in the expression (1b), the color signal conversion matrix A and its inverse matrix A ^-1 can mutually reversibly convert each other.

[0020]

(Equation 1)

Here, (i, j) given to the color signals P and F means the position of the pixel.

In this embodiment, as the transformation matrix A and the inverse transformation matrix A- ¹ , those given by the following equations (2a) and (2b) are used.

[0023]

(Equation 2)

In the present embodiment, a watermark mark is embedded when the image data P of the RGB color system is converted into the image data F of the YCbCr color system, and the watermark mark is formed by inversely converting the watermark data to the RGB color system again. Create an image. The transformation matrix used for embedding the watermark mark (referred to as “embedding transformation matrix”) is obtained by adding an additional matrix X given by the following equation (3) to the transformation matrix A (that is, (A +
X)).

[0025]

(Equation 3)

The additional matrix X is a matrix in which real values x and -x are components and the sum of the components in each row and each column is zero. In this specification, the real value x is also referred to as a “watermark value”, and the additional matrix X is also referred to as a “watermark matrix”. The configuration of such a watermark matrix X is determined for the following reasons. First,
Consider embedding the watermark mark so that it cannot be visually recognized by the naked eye. For this reason, it is preferable to prevent image quality degradation due to embedding of a watermark as much as possible. By the way, (2a)
Observing each component of the transformation matrix A given by the expression, it is understood that the sum of the absolute values of the components in each row is all 1. Therefore, it is preferable for the embedding conversion matrix (A + X) to maintain a relationship such that the sum of the absolute values of the components in each row is 1, in order to prevent image quality deterioration.
Since the watermark matrix X in the equation (3) is set so that the sum of the components in each row and each column becomes 0, this requirement is satisfied.

When embedding a watermark mark in the original image P, a mark image M including a watermark mark to be embedded is prepared in addition to the original image P. FIG. 2B shows an example of the mark image M used in the present embodiment. The mark image M is a binary image. The pixel value m in the area of the two characters “I” and “T” forming the watermark mark is 1, and the pixel value m in the other areas is 0.
It is. In the following, an area where the pixel value m is 1 is referred to as a “watermark effective area” or simply “mark area”,
An area where the pixel value m is 0 is referred to as a “watermark ineffective area”. Note that the effective area of the watermark mark may be simply referred to as “watermark”. For example, the phrase “position in the watermark mark” has the same meaning as “position in the effective area of the watermark mark”.

The mark image M according to the present embodiment has the same size as the original image P. However, the mark image M does not need to have the same size as the original image P, and it is only necessary that the position of the watermark mark embedded in the original image M is determined by some method.

When embedding a watermark mark, first, a Y color is converted from the RGB color system in accordance with the following equations (4a) and (4b).
Conversion to the CbCr color system is performed.

[0030]

(Equation 4)

That is, at a pixel (m (i, j) = 1 pixel) at a position corresponding to the effective area of the watermark mark, the embedding conversion matrix (A + X) obtained by adding the original conversion matrix A and the watermark matrix X is added. ) Is used to perform color system conversion. On the other hand, at the pixel at the position corresponding to the non-effective area of the watermark mark (the pixel of m (i, j) = 0), the conversion of the color system is performed using the original conversion matrix A as it is.

[0032] (4a), when executed for all pixels arithmetic processing by Expression (4b), the watermark information is watermarked YCbCr image data F _X embedded obtained (FIG. 1, FIG. 2 (C)) .

Next, when the inverse transformation matrix A ^-1 is operated from the left with respect to the watermarked YCbCr image data F _X ,
RG with watermark given by the following equations (5a) and (5b)
B image data P _X is obtained (FIG. 1).

[0034]

(Equation 5)

As shown in the equation (5a), in the watermarked RGB image data P _X , the noise component (A ^-1 XP) is superimposed on the original RGB image data P in the effective area of the watermark mark. It is a thing. In the non-effective area of the watermark mark, the watermarked RGB image data P _X has the same value as the original RGB image data P.

By the way, as shown in the above equation (3), the sum of the components of each row of the watermark matrix X is 0. Therefore, when the values of the r, g, and b components of the original RGB image data P (i, j) are equal to each other (that is, when the color is achromatic), the noise component (A ^-1 XP) of the equation (5a) is 0. Become. In order to prevent such a problem, it is preferable to position the watermark mark in a non-achromatic region in the original image.

FIG. 2C shows a state where the watermark mark in the watermarked image is seen faintly. However, in practice, the visibility of the watermark mark (easiness of recognition by the naked eye) can be adjusted by adjusting the watermark value x in the watermark matrix X. A method of adjusting the visibility of the watermark mark will be further described later.

By the way, RGB image data is generally composed of integer values, whereas YCbCr image data is usually composed of real values. Therefore, (4a), (4
When converting to the YCbCr color system by the expression (b), realization is performed, and when inversely converting to the RGB color system by the expressions (5a) and (5b), integer conversion is performed. At the time of this conversion into integers, there are cases where the pixel value becomes negative and an underflow occurs, and cases where the overflow exceeds the upper limit of the pixel value. In order to solve this problem, in the present embodiment, underflow correction and overflow correction are performed at the time of conversion into integers. In other words, when the pixel value before the integer conversion becomes negative, underflow correction of 0 is performed, and when the pixel value exceeds the upper limit of the bit expression (for example, 255), the overflow correction adopts the upper limit. I do.

When underflow correction or overflow correction is performed, the substantial image quality of the watermarked image is degraded. Here, “substantial deterioration” means a deterioration in image quality that cannot be recovered by erasing the watermark mark. Since the underflow correction and the overflow correction are caused by adding the watermark matrix X to the original transformation matrix A, the occurrence rate of the underflow correction and the overflow correction depends on the watermark value x which is a component of the watermark matrix X. I do. Therefore, it is preferable to set the watermark value x to a value at which underflow correction and overflow correction do not occur so much.

FIG. 3 is an explanatory diagram showing the relationship between the watermark value x and the occurrence rates of underflow correction and overflow correction. Here, results regarding two images, "girl" and "graphic", which are standard images used for evaluating the characteristics of image quality, are shown. Note that the difference between the case where the watermark value x is 0.1 and the case where the watermark value x is -0.1 is considered to be due to the bias in the distribution of the three color components r, g, and b in the image. . From the results shown in FIG. 3, it can be seen that it is preferable to set the watermark value x to a value of about 0.7 or less in order to keep the occurrence rates of the underflow correction and the overflow correction low. When the watermark value x is set to 0, a watermark mark cannot be embedded. Therefore, it is actually preferable to set the watermark value x to a value of about 0.01 or more. Therefore, the watermark value x is about 0.01 to about 0.5.
It is preferable to set a value in the range of 7.

When the embedded watermark mark cannot be recognized by the naked eye, it is necessary to restore the watermark mark. Watermarks can not be recognized with the naked eye by utilizing a noise component contained in the watermarked RGB image data P _X (A ^-1 XP), it is possible to decode in various ways as described below.

The first method is a method of changing the color of an image so that the luminance values of the components r, g, and b of the watermarked RGB image are increased to near the maximum value. In this case, since the noise component (A ^-1 XP) also becomes large, the watermark mark may be easily recognized by the naked eye.

The second method is a method in which a noise component (A ^-1 XP) contained in the watermarked RGB image data P _X is directly used. That is, the watermark mark (that is, the watermark information) can be decoded by determining a region having the noise component (A ^-1 XP) in the watermarked RGB image data P _X using the original image data P as a key. . Specifically, determine the watermarked RGB image data P _X and the difference between the original RGB image data P (P _X -P), to determine the region this difference (P _X -P) is not 0 as a mark area It is possible. The difference (P _X
−P) is a matrix having three components of r, g, b,
The phrase “difference (P _X −P) is not zero” means that at least one of the three components is not zero.

The noise component (A) in the mark area
^-1 XP), the product (XP) of the watermark matrix X and the original image data P is obtained. If the original image data P is known, the watermark value x of the watermark matrix X can be decoded from the product (XP) even when the watermark value x is unknown.

As described above, the conversion matrix A is added to the watermark matrix X
The watermark mark can be embedded in the color image by using the embedding conversion matrix (A + X) to which is added. It is also possible to restore the embedded watermark mark from the watermarked image.

B. Adjustment of the visibility of the watermark mark: As described above, in the watermarked RGB image P _X , the noise component (A ⁻¹ XP) is superimposed in the mark area. Therefore, the noise component (A ^-1) is adjusted by adjusting the size of the watermark value x.
By changing the value of XP), it is possible to adjust the visibility of the watermark mark.

FIG. 4 is an explanatory diagram showing a method of adjusting the visibility of a watermark mark. 4 (A) and FIG.
The mark image M in (B) is the same as that shown in FIGS. 2A and 2B.

When the watermark value x is set to a relatively large value of about 0.7, the watermark mark in the watermarked image can be clearly recognized with the naked eye, as shown in FIG.
Such a watermark mark is referred to herein as a “visible watermark”. In particular, if a watermark mark exists in an area where there is little change in color, the change in color in the mark area is
Can be more clearly recognized. Therefore, in order to make the watermark mark easily visible, it is preferable to position the watermark mark in an area where there is little change in color.

On the other hand, when the watermark value x is set to a relatively small value of about 0.05, as shown in FIG. 4E, the watermark mark in the watermarked image cannot be recognized by the naked eye. Such a watermark mark is referred to herein as an “invisible watermark”. In FIG. 4E, the outline of the mark area is indicated by a dotted line to indicate the position of the watermark mark. However, in practice, the watermark mark can be made invisible to such an extent that it cannot be recognized by the naked eye. It is. In particular, there is a tendency that it is difficult to visually recognize the watermark mark in a low brightness area or a chromatic color area close to an achromatic color. Therefore, in order to make it difficult to visually recognize the watermark mark, it is preferable to position the watermark mark in an area with low brightness or a chromatic area close to achromatic.

When the watermark value x is set to an intermediate value between a visible watermark and an invisible watermark, a translucent watermark (called a “semi-visible watermark”) can be generated. For example, when the watermark value x is set to a value of about 0.3, FIG.
As shown in (1), the watermark mark in the watermarked image becomes translucent.

As described above, by setting the watermark value x to an appropriate value, it is possible to set the visibility of the watermark mark to a desired degree.

By the way, instead of adjusting the watermark value x as described above, it is possible to adjust the color of the watermark mark by adjusting the arrangement of the watermark matrix X.
That is, instead of the watermark matrix X shown in Expression (3),
If the above-described embedding process is performed using the watermark matrices X _1,2 and X _1,3 as exemplified in the following equations (6) and (7), the color of the watermark mark is changed.

[0053]

(Equation 6)

[0054]

(Equation 7)

The watermark matrix X _{1,2 in the} equation (6) is obtained by exchanging the first and second rows of the watermark matrix X in the equation (3). Further, the watermark matrix X _{1,3 in the} expression (7) is given by (3)
The first and third rows of the watermark matrix X in the equation are interchanged. As can be seen from these examples, it is generally possible to change the color of the watermark mark by exchanging the columns or rows of the watermark matrix X. When the color of the watermark mark is changed, the easiness of recognition by the naked eye (that is, visibility) often changes. Therefore,
In this specification, changing the color of the watermark mark by exchanging the columns or rows of the watermark matrix in this way also corresponds to “adjustment of visibility”.

C. Visual watermark invisible processing: The visible watermark can be further invisible by the following processing. As can be understood from the above equation (5a),
The watermarks can recognize with the naked eye is because noise components of the watermarked image data P _X (A ^-1 XP) is relatively large. Therefore, it is possible to make the visible watermark invisible by executing the process of reducing the noise component (A ^-1 XP). In the invisibility processing, an invisibility matrix K given by the following equation (8) is used.

[0057]

(Equation 8)

Here, I is a 3 × 3 unit matrix. That is, the invisible matrix K is the sum (I + A ^-1 X) of the product (A ^-1 X) of the inverse transform matrix A ^-1 and the watermark matrix X and the unit matrix I having the same degree as the transform matrix A. Is the inverse matrix.

[0059] The invisible matrix K, as shown in FIG. 5, by operating the watermarked RGB image data P _X of (5a) equation, the multiplication result is given by the following equation (9).

[0060]

(Equation 9)

Here, the right side of the second line of the equation (9) is P ′
(I, j) means that the original image data P (i, j) may not be completely returned due to an arithmetic error (rounding error) during the embedding process described above. ing. That is, RG from the YCbCr signal
Since the conversion to the B signal is performed by integer conversion, the invisible image data P ′ (i, j) is converted to the original image data P ′.
It may not be possible to obtain exactly the same value as (i, j). However, since the error is a small amount after the decimal point, the image after invisibility looks like the original image to the naked eye. However, when the error due to the overflow correction or the underflow correction is large, the color cannot be restored even if the visualization is performed, and the trace of the watermark mark may be visually recognized. Therefore, in order to make invisibility as completely as possible, it is preferable to use a small watermark value x that does not cause much overflow correction or underflow correction.

Since the invisible image data P ′ includes an error in the mark area, the watermark mark can be decoded from the invisible image data P ′. For example, a difference (P'-P) between the image data P 'after invisibility and the original image data P is obtained, and an area where the difference (P'-P) is not 0 can be determined as a mark area. .

E. Security improvement measure: In the above description, one watermark value x is used for the entire watermark mark. If the watermark value x is known to a third party, a fake watermark may be created or the watermark mark may be easily changed. Therefore, in order to enhance the security of the digital watermark, it is preferable to apply a plurality of watermark values x in the mark image.

FIG. 6 shows a case where a plurality of watermark values x are included in a mark image.
It is explanatory drawing which shows the example which applied. In FIG. 6A, different watermark values x ₁ and x ₂ are applied to the mark areas of two characters “I” and “T” forming the watermark mark. In FIG. 6B, two sets of spatially separated marks are arranged in the mark image, and different watermark values x ₁ and x ₂ are assigned to these two marks.
Has been applied. In the case of FIG. 6A, the characters “I” and “T” can be considered as different mark groups. That is, generally, different watermark values x can be set for a plurality of mark groups spatially separated from each other. As described above, since the visibility of the watermark mark changes according to the watermark value x, by setting different watermark values x to a plurality of mark groups in the mark image, the mark groups of various visibility can be changed to the watermarked image. It is possible to embed inside.

FIG. 7 is an explanatory diagram showing an example of division of a mark area for a multiple watermark. In the example of FIG. 7, a mark area composed of one character “T” is divided into a plurality of divided areas, and different watermark values x are assigned to adjacent divided areas.
Has been applied. As a method of dividing the mark area, a method of dividing by a parallel line as shown in FIG.
Various methods are conceivable, such as a method of dividing by contour lines as shown in FIG. Note that "one mark area" means an effective area composed of spatially continuous pixels.

In such a multiple watermarking method, since the watermark value x applied to each section area can be kept secret, it is possible to prevent a third party from easily discovering the watermark value x.

F. FIG. 8 shows the overall configuration and processing procedure of the device.
FIG. 2 is a block diagram illustrating a configuration of a digital watermark processing apparatus that performs the above-described digital watermark processing according to the present embodiment. This digital watermark processing device includes a CPU 22, a ROM and a RAM.
, A frame memory 26, a keyboard 30, a mouse 32, a display device 34, a hard disk 36, a modem 38, a scanner 39 for reading an image, and a bus 40 for connecting these components. , A computer comprising: In FIG. 8, various interface circuits are omitted. The modem 38 is connected to a computer network via a communication line (not shown). A server (not shown) of the computer network has a function as a program supply device that supplies a computer program to the digital watermark processing device via a communication line.

The main memory 24 stores a computer program for realizing the functions of the digital watermark embedding unit 42, the digital watermark decoding unit 44, and the watermark visualizing unit 46. The functions of the digital watermark embedding unit 42, the digital watermark decoding unit 44, and the watermark visualizing unit 46 are as already described in detail.

A computer program for realizing the functions of these units 42, 44, 46 is provided in a form recorded on a computer-readable recording medium such as a flexible disk or a CD-ROM. The computer reads the computer program from the recording medium and transfers it to an internal storage device or an external storage device. Alternatively, a computer program may be supplied to a computer via a communication path. When implementing the functions of the computer program, the computer program stored in the internal storage device is executed by the microprocessor of the computer. Further, a computer may read a computer program recorded on a recording medium and directly execute the computer program.

In this specification, a computer is
The concept includes a hardware device and an operation system, and means a hardware device that operates under the control of the operation system. In the case where an operation system is unnecessary and a hardware device is operated by an application program alone, the hardware device itself corresponds to a computer. The hardware device includes at least a microprocessor such as a CPU and means for reading a computer program recorded on a recording medium. The computer program includes a program code that causes such a computer to realize the functions of the above-described units. Some of the functions described above may be realized by an operation system instead of the application program. Further, the program for performing the digital watermark embedding process, the decryption process, and the watermark invisibility process may be added in a plug-in format to the image processing program.

The “recording medium” in the present invention includes a flexible disk, a CD-ROM, a magneto-optical disk, an IC card, a ROM cartridge, a punched card, a printed matter on which a code such as a bar code is printed, and a computer internal storage. Device (RAM, ROM, etc.)
Various computer-readable media such as an external storage device and the like can be used.

FIG. 9 is a flowchart showing the procedure of the embedding process performed by the digital watermark embedding unit 42. In step S1, an original RGB image P and a mark image M are prepared. In step S2, a conversion matrix A used for color system conversion and its inverse conversion matrix A- ¹ are prepared. In step S3, a watermark matrix X and a watermark value x are set. In step S3, the adjustment of the visibility based on the value of the watermark value x, the adjustment of the visibility by exchanging the rows or columns of the watermark matrix X, the setting of a plurality of watermark values x in the mark image, and the like. The user can make various settings including the following.

In step S3, the above-described embedding process is executed using the embedding transformation matrix (A + X) and the inverse transformation matrix A- ¹ . In step S5, the watermarked image is displayed on the display 34. Alternatively, the watermarked image may be printed by a printer (not shown). In step S6, the user observes the watermarked image, and if the result is satisfactory, terminates the embedding process. If the result is not satisfactory, the process returns to step S3, changes the settings of the watermark matrix X and the watermark value x, and executes steps S4 and S5 again. Thus, by repeating steps S3 to S6, it is possible to create a watermarked image in which a watermark mark of a desired visibility is embedded.

The decoding of the watermark mark by the watermark decoding unit 44 (FIG. 8) and the invisibility of the visible watermark by the watermark invisibility unit 46 are performed on the watermarked image in which the watermark mark is embedded as necessary. Be executed. The description of these processing procedures is omitted.

It should be noted that the present invention is not limited to the above examples and embodiments, but can be implemented in various modes without departing from the gist of the invention.
For example, the following modifications are possible.

(1) In the above embodiment, part of the configuration realized by hardware may be replaced by software, and conversely, part of the configuration realized by software may be replaced by hardware. You may do so. For example, the digital watermark embedding unit 4 realized by a computer program in FIG.
2, the functions of the digital watermark decoding unit 44 and the watermark visualizing unit 46 can be realized by dedicated hardware circuits.

(2) In the above embodiment, the watermark mark is embedded by performing a conversion from the RGB color system to the YCbCr color system and then performing an inverse conversion. On the contrary, the YCbCr color system is used. May be embedded by performing a reverse conversion after converting once from an RGB color system. The color systems that can be used are not limited to these two color systems, and any two color systems that can perform reversible conversion and inverse conversion can be used.

(3) In the above embodiment, at the time of embedding the watermark mark, conversion and inverse conversion are performed on the pixels in the non-effective area of the watermark mark by using the conversion matrix A and the inverse conversion matrix A- ^1. However, when the inverse conversion is performed on the ineffective area, the image data returns to the original image data. Therefore, such conversion or the inverse conversion may not be performed. In this case, the above-described embedding process is performed on the effective area of the watermark mark, and the image data in the watermark mark area obtained in this manner is combined with the original image data P (ie, overwritten), so that the watermarked image data P _X You can get the whole. In other words, at the time of embedding the watermark mark, at least pixels corresponding to positions in the effective area of the watermark mark are converted using the conversion matrix for embedding (A + X) and inversely using the inverse conversion matrix A ^−1. Conversion may be performed.

(4) In the above embodiment, an example was described in which the watermark mark was constituted by a character string. However, any object (natural image, illustration, logo, etc.) other than the character string is used as the watermark mark. It is possible.

[Brief description of the drawings]

FIG. 1 is an explanatory diagram showing the principle of a digital watermark embedding process according to an embodiment of the present invention.

FIG. 2 is an explanatory diagram showing the contents of a digital watermark embedding process.

FIG. 3 is an explanatory diagram showing a relationship between a watermark value x and an occurrence rate of underflow correction and overflow correction.

FIG. 4 is an explanatory diagram showing a method of adjusting the visibility of watermark information.

FIG. 5 is an explanatory diagram showing the principle of a process for making a visible watermark invisible.

FIG. 6 is an explanatory diagram showing an example in which a plurality of watermark values x are applied to a mark image.

FIG. 7 is an explanatory diagram showing an example of division of a mark area for a multiple watermark.

FIG. 8 is a block diagram illustrating a schematic configuration of a digital watermark processing apparatus that performs digital watermark processing according to the embodiment.

FIG. 9 is a flowchart showing the procedure of watermark embedding processing.

[Explanation of symbols]

 22 CPU 24 Main memory 26 Frame memory 30 Keyboard 32 Mouse 34 Display device 36 Hard disk 38 Modem 39 Scanner 40 Bus 42 Digital watermark embedding unit 44 Digital watermark decoding unit 46 Watermark invisible Department

Claims (9)

[Claims] 1. A digital watermarking method for embedding a watermark mark in a color image, comprising: (a) converting color image data of a first color system into a second color system different from the first color system; Preparing a conversion matrix for converting to system color image data, and an inverse conversion matrix for performing an inverse conversion of the conversion,
(B) setting an additional matrix in which the real values x and −x are components and the sum of the components in each row and each column is substantially 0; and (c) adding the additional matrix to the conversion matrix. Creating a watermark embedding conversion matrix, (d) preparing a watermark mark to be embedded in a color image, and (e) converting the watermark embedding conversion at a pixel corresponding to a position in the watermark mark. The matrix is used to convert the color image data of the first color system into the second color image data.
And (f) inversely converting the color image data of the second color system obtained by the conversion to the first color system by using the inverse transformation matrix. Creating watermarked image data representing a watermarked color image in which the watermark mark is embedded, wherein the step (b) comprises the real number x and an array of each component in the additional matrix. Adjusting the visibility of the watermark mark in the watermarked color image by adjusting at least one of the digital watermarking method and the digital watermarking method. 2. The electronic watermarking method according to claim 1, wherein the watermark mark includes a plurality of mark groups spatially separated from each other, and the adjustment of the visibility is performed among the plurality of mark groups. Is performed by setting the real value x to a value different from other mark groups in at least some of the mark groups.
Digital watermarking method. 3. The digital watermarking method according to claim 1, wherein the watermark mark includes a mark area composed of spatially continuous pixels, and the mark area is divided into a plurality of division areas. Adjustment of the visibility is performed by setting the real value x to a value different from other divided regions in at least some of the divided regions.
Digital watermarking method. 4. The digital watermarking method according to claim 1, wherein the adjustment of the visibility is performed by exchanging columns or rows of the additional matrix. 5. The electronic watermarking method according to claim 1, wherein, when the watermark mark is visible to the naked eye in the watermarked color image, the method further comprises: (g) using the inverse transformation matrix Preparing an invisible matrix that is the inverse of the sum of the product of the additional matrix and a unit matrix having the same degree as the transformation matrix; and (h) in a pixel corresponding to a position in the watermark mark, A digital watermarking method, comprising a step of causing the watermarked image data to operate on the invisibility matrix to reduce the visibility of the visible watermark mark. 6. The electronic watermarking method according to claim 1, wherein one of the first and second color systems represents three primary colors of light.
An electronic color system comprising an RGB color system composed of two color signal components, and the other of the first and second color system being a YCbCr color system composed of a luminance signal and two color difference signal components. Watermark method. 7. The digital watermarking method according to claim 6, wherein in the step (f), the image data of the YCbCr color system obtained by the conversion by the conversion matrix for embedding the watermark is converted into the inverse conversion matrix. When performing the inverse conversion to the RGB color system by using underflow correction and overflow correction so that each pixel value in the image data of the RGB color system becomes an integer value within a predetermined range,
Digital watermarking method. 8. The digital watermarking method according to claim 1, wherein the real value x is a value in a range from about 0.01 to about 0.7. 9. A computer-readable recording medium on which a computer program for embedding a watermark mark in a color image is recorded. A function of setting an additional matrix whose sum is approximately 0; and a function of converting color image data of the first color system into color image data of a second color system different from the first color system. A function of creating a transformation matrix for watermark embedding by adding the transformation matrix prepared in advance and the additional matrix, and using the transformation matrix for watermark embedding, corresponding to a position in a watermark mark prepared in advance. In the pixel, a function of converting the color image data of the first color system into the second color system; and a color image data of the second color system obtained by the conversion, A watermark representing a watermarked color image in which the watermark mark is embedded, by performing an inverse transformation to the first color system using an inverse transformation matrix prepared in advance to perform an inverse transformation of the transformation by the transformation matrix. And a computer program for causing a computer to realize the function of creating attached image data. The additional matrix setting function includes: A computer-readable recording medium having a function of adjusting the visibility of the watermark mark in the watermarked color image by adjusting at least one of the watermark images.