JP2018207247A - Digital watermarking apparatus and method - Google Patents

Abstract

To provide an apparatus and method which can keep high image quality, embed digital watermarks with strong resistance and extract watermarks, and safely remove and re-embed watermark information.SOLUTION: An image is divided into blocks and a dot pattern showing the blue noise characteristic reduced in a low frequency range of a spatial frequency is embedded into each of the blocks as watermark information. In the extraction, highly reliable watermark information can be obtained when the block embedded with the highly reliable watermark information is selected. In addition, the watermarks can be removed and re-embedded by a key.SELECTED DRAWING: Figure 2

Description

The present invention relates to a digital watermark method for embedding copyright information and the like in digital image data, and particularly to an invisible digital watermark apparatus and method with high image quality.

In today's digital information society, it has become possible for many people to share information by duplicating information, and societies have developed greatly. However, due to its convenience, illegal copying and distribution of personal works has caused incidents such as copyright infringement. In recent years, with high-quality digital cameras and printers, it has become possible to easily reproduce images that are not different from the original images. This is a result of promoting a malicious criminal act.

Under such circumstances, digital watermark technology that protects copyright by embedding other information such as author information in image information has been developed. This digital watermark technology embeds the name, contact information, and handling information of the copyright holder in the copyrighted work, not only alerts the user, but also enables tracking of unauthorized use. It is starting to be widely used as a protection and security measure. In order to ensure security and reliability, digital watermarking technology is required to have strong resistance to attacks, assuming attacks for deletion from third parties.

Digital watermarks as digital data do not deteriorate or disappear as long as they are copied as they are. However, when image editing, processing, or image compression is performed, the embedded information is easily lost because various calculations and conversions are performed on the image data. Therefore, in order to withstand the editing process, the strength of the embedded data must be increased and embedded in the deep layer to improve the durability.

In general, there is a trade-off between robustness and image quality. If the embedding tolerance is increased strongly, the image quality deteriorates. Conversely, if the image quality is prioritized, the tolerance is weak. Therefore, it is very difficult to satisfy the tolerance and the image quality at the same time.

So far, many methods have been proposed as digital watermark technology that is resistant to image data. One example is the patchwork method. Information embedding by the patchwork method is an approach that uses statistics, so it is highly resistant to local attacks.
This method gives a small deviation to many pixel pairs in real space (pixel space) and embeds them as watermark information. First, a pixel pair is selected at random, and the brightness value of one is increased by δ, and the brightness value of the other is decreased by δ. By repeating this operation, the expected value of the distribution is biased. It is possible to extract watermark information by statistically obtaining this bias.
However, this method needs to increase the embedded data value δ in order to increase the tolerance, and since it is embedded in real space (pixel space), the embedded region (patch) becomes visually noticeable. The amount of information that can be embedded is extremely small.

Another method is to embed watermark information in frequency space. In this method, the entire image is subjected to DCT (Discrete cosine transformation), Fourier transform, and Wavelet transform and embedded in a specific intermediate frequency band. The embedded information is distributed in the entire image in real space, so it is not visually noticeable. However, when the embedding strength is increased, a periodic pattern is generated, and it becomes easy to see in the flat part of the image, and moire fringes with the image are generated.

On the other hand, Patent Document 1 (Japanese Unexamined Patent Application Publication No. 2016-163197) proposed a method using a dot profile showing the characteristics of the green noise noise method. This invention realizes a digital watermarking method with printing durability by using a dot profile with green noise characteristics that lower frequency response in high and low frequencies. Compared to the above-mentioned digital watermarking method, the printing tolerance is high, and even in the printed image, as long as it is observed at a clear distance from the low-pass effect due to human visual characteristics, the image quality is not perceived. However, when observing on a display, etc., the image is often enlarged, and the image quality is not sufficient because the spatial frequency of the high frequency is reduced.

JP 2016-163197

W. Blender, D. Gruhl, N. Morimoto; “Techniques for data hiding”, Proceedings of SPIE, Vol.2020, pp.2420-2440 (1995). Minoru Mizumoto, Koko Matsui; “Examination of digital watermark printing tolerance using DCT”, IEICE Journal A, Vol J85-A, No.4, pp.451-459 (2002)

For this reason, when displaying an image with a watermark on a display, it is a problem to obtain a digital watermark method with little moiré and image quality degradation and little image quality degradation even when enlarged. And it is necessary to be resistant to various attacks that try to remove watermark information from outside.

Furthermore, in order to maintain image quality, it is also necessary to remove the watermark once using a key for image editing and processing. After editing, it is also necessary to write new watermark information, such as second copyright information, using this key. At this time, depending on the application, it may be necessary to embed a watermark that is resistant to printing at the time of distribution, and it is possible to use another digital watermark method after removal.

On the other hand, it is also a problem to secure the security of the key so that the damage does not spread over a wide range against the loss of the key. Furthermore, it is also a problem to satisfy that the security is maintained even if the algorithm is disclosed in order to spread.

In addition, there are many cases where the original image (image data before watermarking) is edited and cut out and used, which makes it difficult to extract the watermark. In order to perform watermark extraction with high accuracy, it is possible to improve the extraction accuracy by obtaining at least information about how the original image was edited. In other words, the size and direction of the original image can be estimated from the embedded image data.

In order to solve the above problems, the present invention provides an apparatus for realizing a high-quality but durable digital watermark by superimposing and embedding a random dot pattern with a band limitation in a frequency space on an original image, and It provides a method.

The random dot pattern used for embedding shows a spectral characteristic with reduced intensity at low frequencies. Because it is random, there is no moiré with the original image. It is also characterized by a very fine dot pattern and difficult to recognize from human visual characteristics. For this reason, even if it is enlarged on the display, it does not feel image quality degradation. Moreover, it is possible to increase printing durability by embedding it strongly.

In addition, it is highly resistant to external attacks and has a key that can remove the watermark if necessary. The key includes a dot pattern of watermark information to be embedded, and such a pattern is made up of a large number of random dots, making reverse analysis difficult. Furthermore, by changing the pattern configuration for each image data, Even if the watermark information of is decoded, it is almost impossible to remove the watermark of other images with this information.

In order to solve this problem, the present invention embeds the original image using a dot pattern exhibiting blue noise characteristics. When this blue noise pattern is viewed in the spatial frequency space, there is no spectrum near 0 frequency. On the other hand, since the image data is distributed near the 0 frequency, there is little overlap between the spectra. For this reason, when extracting embedded information by separating each spectrum, there is a characteristic that it can be extracted well. In addition, the dot pattern is distributed, low graininess, and visually inconspicuous, so it is possible to maintain the high quality of the original image. To put it the other way around, even if the embedding strength (gain) is increased, the image quality is not deteriorated, so the tolerance is improved.

To embed watermark information, the digital image data is divided into R pixel x R pixel blocks, and the Fourier spectrum has a low frequency range corresponding to the bit information (0 or 1) of the watermark information embedded in each block. Watermark embedding is performed using several different dot patterns that exhibit reduced blue noise characteristics. In the extraction of watermark information, the received or read image data is corrected, then divided into blocks, the Fourier spectrum distribution of each block is obtained, the contrast of the spectrum distribution is improved, and the embedding information is recognized by the distribution pattern recognition. Is extracted.

Furthermore, when extracting watermark information, it is possible to extract watermark information with high accuracy by calculating reliability and selecting watermark information with high reliability.

According to the present invention, it is possible to maintain a high image quality and to embed a strong watermark. In addition, the Affine transform coefficients (enlargement / reduction ratio and rotation angle) applied in image editing and processing can be detected from this dot pattern. And, by obtaining the corrected image from this coefficient, it is possible to extract the watermark with high accuracy.

Furthermore, it is possible to return to the original image (original image) by obtaining the key for the image with embedded watermark. The key is different for each distribution image, and it is powerless for other images. It is possible to embed new secondary copyright information as a digital watermark using the key.

Diagram of the digital watermark embedding and extracting device of the present invention Diagram showing the delivery of watermarked images over the Internet Diagram showing the processing flow for obtaining the dot profile used for embedding This figure shows the dot pattern and its spectrum. (A) is the one with the elliptical spectral characteristics, (b) is rotated 90 degrees, (c) is for Affine coefficient detection, (d) is For header detection In the actual embedded image, (a) is the embedded image, (b) is the extracted spectral image. Enlarged spectral image Image and enlarged view with different gain embedded Diagram showing watermark embedding flow Diagram showing gain automatic adjustment processing flow when embedding watermark The figure showing the flow which acquires the correction picture at the time of watermark extraction Diagram showing the processing flow of watermark extraction Illustration of mask pattern for watermark extraction The figure which shows the mask processing of watermark extraction The figure which shows reliability improvement by the mask processing of watermark extraction Diagram showing different dot patterns in the same spectrum

This digital watermark is embedded in an image whose copyright is to be protected, and mainly tracks and protects the copyright in online distribution. The copyright holder publishes images with such watermark information on the Internet, but it is possible to track unauthorized use because of the watermark. In addition, the purchaser can release the watermark and rewrite it by using a key, and can embed new author information.

Unlawful duplication and use can be monitored by watermark extraction software. This watermark extraction software has no confidentiality, and can be used by anyone who has made it widely available, for example, it can be freely downloaded from the copyright holder's homepage. This allows third parties to know the owner of the image, copyright information, contact information, etc., and can also be used as a warning of unauthorized use.

In order to implement such digital watermarking, the present invention uses a plurality of different dot patterns to be embedded in the original image. This dot pattern exhibits blue noise characteristics, and the embedding pattern is selected according to 0, 1 of the bit information to be embedded and header information (the beginning of the character string). It also includes patterns for analyzing the coefficients of Affine transforms and patterns for headers.

Such a dot pattern is a blue noise pattern obtained by iterative calculation from a random dot pattern generated from pseudo-random numbers. Using these patterns, for example, the copyright holder's name, email address, URL, and other character strings are converted to bit information and embedded in the image data. It can also be used as part of a “key” to remove embedded information and restore the original image. Such a key consists of dot pattern information, embedding strength (gain) information, embedded character information, and the like. Such key information becomes a different dot pattern by changing the initial random number of the generated dot pattern. For this reason, the key information is different for each dot pattern. However, there is no need to change the watermark extraction software.

In the following, the details will be described along the examples.
FIG. 1 is a block diagram of a digital watermark system for embedding and extracting information according to the present invention. The copyright owner's computer 3 stores copyrighted image data in a data memory 7 such as a hard disk. The image data is processed by the image processing program in the program memory 6 using the CPU 11, ROM 4, RAM 5, etc., and displayed on the monitor 8. A scanner 1 and a printer 2 are connected to the computer 3, and the processed image is displayed on the monitor 8 or output from the printer, and the image can be read from the scanner 1. Since such image processing is expensive, there are cases where processing boards for speeding up such as GPUs are included.

Figure 2 shows the processing procedure for Internet distribution. The image in which the watermark information is embedded is published on the Internet 12 from the computer of the copyright holder 16 (13). The purchase applicant 17 who sees it notifies the distributor of the purchase request (14). After a predetermined procedure, the copyright holder distributes a secret key for embedding and removing the watermark (15). Based on the contract with the copyright holder, the purchaser removes the digital watermark embedded in the sent image and edits the image.

Further, the purchaser can embed the edited image data as the secondary copyright holder with his / her copyright information and URL information using this secret key. Images with embedded watermark information are distributed in the market and used in various ways.

In order to monitor illegal copying and usage, watermark extraction software can extract watermark information from an image and confirm that it is a copyrighted work. This watermark extraction software has no confidentiality, and can be used by anyone who has made it widely available, for example, it can be freely downloaded from the copyright holder's homepage. This allows third parties to know the owner of the image, copyright information, contact information, etc., and can also be used as a warning of unauthorized use.

There is a one-to-one correspondence between the private key and the embedded image. For this reason, the copyright holder uses a different key for each image so that it cannot be misused. For this reason, the watermark of other images cannot be removed using this key. It is also resistant to collusion attacks as described below. At this time, the watermark reading software can be used in common even if the keys are different, as described above.

Here, the dot profile creation algorithm used in the present invention will be described with reference to FIG.
The size of the desired dot profile is now R × R (R = 2 ^ m, where ^ represents a power), and R ^ 2/2 random dots (initially white noise) using a pseudo-random number generator. Generate p (i, j) so that they do not overlap. At this time, the initial dot profile can be changed by changing the SEED value of the pseudo-random number generator.
STEP1: Perform two-dimensional Fourier transform of the p (i, j) dot profile to obtain P (u, v).
STEP2: P (u, v) is multiplied by filter D (u, v) to obtain a new P '(u, v). Here, D (u, v) is a filter whose low frequency is zero.
STEP3: Perform inverse Fourier transform on P '(u, v) to obtain a multivalued point profile p' (i, j).
STEP4: The error function e (i, j) = p '(i, j) -p (i, j) is obtained, and white and black are reversed in descending order of error at each pixel position.
STEP5: Repeat the above operation until the error is within the allowable range.
Perform the above operation to finally obtain a dot profile that exhibits the desired blue noise characteristics.

Next, filter D (u, v) is explained. Since D (u, v) has anisotropy, a different filter is used in the u-axis direction and the v-axis direction. This can be an ellipse or a rectangle. When the cutoff frequency (fmin) in the U-axis direction is fu and f0 = (1/2) ^ (1/2) · fn,
b = (fu-f0) / fn
Would give a dot profile independent of R. Where fn is the Nyquist frequency. The v-axis direction is β times the u-axis direction.

As an example, Fig. 4 shows the dot profile obtained with an elliptic filter when R = 64, b = -1 / 16, and β = 1.5. This dot profile p0 (i, j) is shown in Fig. 3 (a). β = 1.5 means that the filter D (u, v) is expanded 1.5 times in the y-axis direction. Since the spectral distribution is targeted for the u-axis and v-axis, the imaginary part of p0 (i, j) is 0. The dot pattern p1 (i, j) in Fig. 2 (b) is obtained by rotating p0 (i, j) by 90 °, and the spectrum is also rotated by 90 °. (C) and (d) in the figure are an Affine coefficient detection pattern and a header pattern, which will be described later.

Next, the watermark embedding algorithm for multi-valued image data is explained.
To embed watermark information, color image data is converted to Y, Cb, and Cr, and watermark information is embedded in luminance Y. Ideally, it should be embedded in Blue (yellow when printed). However, accurate color separation is necessary.

Assuming that different spectrum patterns are embedded in the frequency space in block units, let different spectrum patterns be Pi (u, v) (i = 0, ..., n). The spectrum of Pi shows the blue noise characteristics described above. Ideally, the cutoff frequency fmin of blue noise should be selected to be higher than the maximum frequency of the embedded image. This is because the spectrum of the image can be pushed into the zero region (low frequency region) of the blue noise spectrum, so separation is easy.
If the frequency spectrum of the image block is I (u, v), the embedded spectrum W (u, v) is
W (u, v) = I (u, v) + gain ・ Pi (u, v) ---------- (1)
It becomes. Here, in order to be invisible, it must be gain << 1. In order to extract watermark information Pi (u, v) with high accuracy, it is desirable to reduce the overlap of I (u, v) and Pi (u, v) as described above.

The actual watermark information is embedded in real space. Converting the previous expression into real space (pixel space) (represented in lower case)
w (x, y) = i (x, y) + gain ・ pi (x, y) ---------- (2)
It becomes. As watermark information, 1-bit data of 0 and 1 is embedded, and two dot patterns pi (x, y) (i = 1, 2) showing different blue noise characteristics are prepared. In addition, for the bit information (0,1) to be embedded,
When padding bit = 0 → p0 (x, y))
When embedded bit = 1 → p1 (x, y)
Embed as. Here, p0 and p1 are binary values of (1,0), but they are (1/2, -1/2) in order to preserve the average luminance.
If 2-bit multi-value data is embedded, four dot patterns may be prepared.

The dot patterns of p0 and p1 are juxtaposed according to the watermark information, but because they are random dots, the boundary between them is not noticeable. In addition, halftone screens exhibiting blue noise characteristics are often used as FM screens in printers, and they are uniform in dispersive dots, so they are uniform and do not feel grainy. However, if the anisotropy of the dots is increased, it becomes noticeable, so β needs to be within 1.5.

Figures 4 (c) and 4 (d) show the blue noise pattern obtained using a rectangular filter. (C) is a pattern generated from a rectangular filter with b = -3 / 16. (D) is a dot pattern generated from a rectangular band filter with b = 0. The pattern (c) is used to search for the Affine transform coefficient, which will be described later, and the pattern (d) is used at the beginning of the character string, and is used as a repeated delimiter pattern.

The amount of data that can be embedded in the original image (maximum number of bits) N is the image size for image data of W pixels x H pixels.
N = int (W / R) × int (H / R) ---------- (3)
Is given by. Here, int () represents rounding down to the nearest whole number. The number of characters that can be embedded is int (N / 8) for ASCII characters. Usually, the target image for copyright protection is an image of HD size (1980 pixels x 1024 pixels) or larger. In HD images, N = 480 bits = 60 bytes, which is sufficient for embedding copyright information. However, if a small size image requires a sufficient amount of information, if R = 32 is embedded, for example, if 512 pixels × 512 pixels, an information amount of N = 256 bits = 32 bytes can be embedded. is there.

Figure 5 shows an image of 1024 pixels by 576 pixels embedded with a dot pattern of R = 64.
Therefore, this image is
N = int (1024/64) x int (576/64) = 16 x 9 = 144 bits = 18 bytes, and character information of up to 18 bytes (in the case of ASCII) can be embedded. Figure (a) shows the embedded image. The gain of embedding is 0.0625. In other words, a dot pattern of ± 8 is synthesized with the original image. In the figure (b), the correct answer is obtained from the extraction results described later. Fig. 6 is an enlarged view of a part of Fig. 5 (b). The spectrum of the image is included in the elliptical spectrum. Therefore, it is easy to separate the two spectra.

It is possible to increase the printing durability of embedded images by increasing the gain. However, increasing gain makes the dot pattern visible. Figure 7 shows the image when the gain is changed and embedded. In the case of gain = 0.125 in Fig. 5 (a), the embedding is hardly visible visually, but the dot structure becomes noticeable in the enlarged image as the gain increases. However, due to the blue noise characteristics, it is hardly noticeable due to human visual characteristics unless the magnification ratio is extremely large.

Figure 8 shows the watermark embedding process flow. First, a watermark information bit string wm (i, j) for embedding is prepared (20). Where i is the i-th character to be embedded and j is the j-th bit (MSB is the 0th bit). Since ASCII characters are 8 bits / character, j = 0,1, ... 7. I is 1 ≦ i ≦ N / 8. N is the maximum number of bits embedded in equation (3). Subsequently, the image data is divided into R × R blocks (21), and a dot pattern is synthesized in units of blocks. Whether or not it is the head of the character string (22). Embed (pHeadder). Further, whether or not the MSB (Most Significant Bit) is a character (23). In the case of the MSB, an Affine detection pattern (pAffine) described later is embedded. Otherwise, the pattern of p0 is embedded when the embedding bit is 0, and the pattern of p1 is embedded when the embedding bit is 1 (24). The process ends when all blocks have been embedded. However, the Affine detection pattern is not necessarily required when the header pattern is also used.

The embedded watermark information may not be extracted because the gain of the embedding is small. This happens when the spatial frequency of the corresponding block in the original image is high. FIG. 9 shows a flow for performing automatic gain adjustment. First, watermark information wm (i, j) is embedded in all blocks of an image with gain = g (31). After all blocks are embedded, the watermark is read (32). Here, the extracted watermark information wm ′ (i, j) is compared with the watermark information wm (i, j) before embedding to determine whether or not they match. At the same time, it is determined whether the reliability wmrel ′ (i, j) of the extracted bit, which will be described later, is greater than a certain threshold (33). If the watermark information does not match, or if the watermark information matches but the reliability is below the threshold, add a constant Δg to gain (34) and embed the watermark again. Since the reliability can be quantified, embedding with the optimal watermark strength can be performed automatically as described above.

The watermark information is extracted by correcting the size and inclination of the scanned image and then obtaining the spectral image in blocks. Fourier transform of equation (2)
F {w (x, y)} = F {i (x, y)} + gain ・ F {pi (x, y)} ---------- (4)
It becomes. Where F {} represents the Fourier transform. Usually, F {i (x, y)} is the spectrum of the original image, localized near the 0 frequency, and F {pi (x, y)} is the spectrum surrounding it. For this reason, there is little overlap between the two, and it is easy to separate them.

Fig. 10 shows the processing flow of the corrected image. First, the target image is cut out from the image (36), and the portion corresponding to the R × R window is subjected to FFT to search for a rectangular pattern (37). Since the printed image is reduced or enlarged, in some cases, the window size may be R / 2 ^ n (where n is an integer). If a rectangular pattern is found, the Affine coefficient can be determined (38). After that, inverse Affine transformation is performed to restore the original image size and direction (39). Subsequently, edge enhancement is applied to the image (40) to obtain a corrected image.

The Affine coefficient is obtained as follows. Now, assuming that the embedded image w (x, y) is enlarged a times, the Fourier transform is P (fx, fy), where the Fourier spectrum before scaling is P (fx, fy)
F {w (x / a, y / a)} = F {i (x / a, y / a)} + gain ・ F {p (x / a, y / a)}
= F {i (x / a, y / a)} + gain ・ | a | ・ P (au, av)
And is reduced to 1 / a in the spectral space. Also, if the image is rotated by θ, it is also rotated by θ in the Fourier spectrum. Therefore, the Affine coefficient applied by measuring the spectrum distribution of the embedded image can be obtained. The detection window size can be small. For example, when extraction is attempted with a window size of 40 pixels × 40 pixels without considering registration, detection is possible even if a margin is provided and 64 pixels × 64 pixels are applied and FFT (Fast Fourier Transform) is performed. .

Figure 11 shows the watermark extraction process. First, the corrected image is divided into R × R blocks (41), and FFT is performed on the first block (42). Subsequently, histogram equalization is applied to the obtained spectral image (43). This is to improve the contrast of the spectral pattern and to standardize the reliability by quantifying the reliability described later. Subsequently, a mask process described later is performed (44) to obtain watermark information wm ′ (i, j) and reliability wmrel ′ (i, j). It is judged whether or not all the blocks are finished (46). If not finished, the process goes to the next block and the same processing is performed again.

Figure 12 shows a mask pattern as a discriminator that extracts watermark information from the spectral distribution after Fourier transform by pattern recognition. Vertical and horizontal elliptical masks that correspond to the spectral characteristics of embedding dot patterns. In the figure, the black area has a value of 0, and the white area has a value of 1. This mask is overlaid on the watermarked spectrum, and it is judged whether the bit is 0 or 1 from the difference of the integrated luminance value of the overlapping part. That is, in FIG. 13, the following integrated luminance value outputs Q0 and Q1 are obtained for the spectrum pattern W (i, j) of the block (40).
Q0 = M0 ◎ W = (1 / Z) ΣM0 (i, j) ・ W (i, j)
Q1 = M1 ◎ W = (1 / Z) ΣM1 (i, j) ・ W (i, j) ---------- (5)
Where ◎ represents the mask operation, and Z represents the number of pixels with a mask value of 1. From this output,
When Q0> Q1, extraction bit = wm '(i, j) = 0
When Q1> Q0, extraction bit = wm '(i, j) = 1
(41). At the same time, the reliability is obtained as follows (42).
wmrel '(i, j) = | Q0-Q1 |
The greater the reliability, the higher the reliability of the extracted bits.

Fig. 14 explains how to improve the accuracy and reliability of watermark information. The maximum number of embedded bits N is given by equation (3) depending on the size of the image. If the number of bits to be embedded is n and n <N, the bit string can be embedded in duplicate. In the extraction of the watermark information in FIG. 10, continuous watermark information wm ′ and reliability wmrel ′ up to N were obtained regardless of the number of bits (without considering the repetition of the character string). From this continuous character string, a highly reliable one is selected to obtain watermark information up to n.

First, when the i-th character number (0 <i ≦ N) of the character string in which i is embedded and j is the bit number (0 ≦ j ≦ 7), the remainder k is k = i mod (50), the (i, j) th reliability wmrel '
The reliability is high when wmrel '(i, j)> wmrel (k, j).
wm (k, j) = wm '(i, j)
wmrel (k, j) = wmrel '(i, j) ---------- (6)
(52) to replace the watermark information. This is performed for all bits and all blocks until i is N, that is, until the number of embedded bits N is complete.

By the above operation, the watermark information with the highest reliability of the embedded bit string is selected, and the reliability of the embedded character is increased and extracted correctly. This method is more accurate than the method of deciding by majority vote.

The watermark is extracted as described above. FIG. 5B shows the spectrum distribution at the time of extraction. Embedding is an ASCII character with 8 characters “kawamura” embedded. Each character is expanded into bits and, as explained in Fig. 3, p0 is embedded for "0" and p1 pattern is embedded for "1". Since the first bit (MSB) of ASCII characters is all 0, we used a pattern for detecting the Affine transform coefficient, and the first bit of the character string was the header pattern for character string delimiter. As can be seen, two orthogonal elliptical spectral patterns and two large and small rectangular patterns are detected. The large rectangular pattern is the header pattern for the string, and the small rectangular pattern is for Affine detection. The number of characters that can be embedded is 18 bytes, and an 8-character string can be embedded twice and 2 characters. Therefore, by selecting a highly reliable one, the accuracy of the extraction result can be improved, and at the same time The reliability of can be quantified.

The watermark information can be removed and the original image can be obtained by knowing the embedding pattern. That is, from equation (2)
i (x, y) = w (x, y)-gain ・ pi (x, y)
The embedding pattern pi (x, y) and gain can be received as a “key” to return to the original image. In order to restore the original image completely, it is necessary to limit the dynamic range in advance to avoid overflow and underflow of the embedded image.
When the image data is 8 bits,
Data '= Data + gain (128-Data) ---------- (7)
With this conversion, the dynamic range is linearly compressed to obtain output data Data ′. After the watermark is removed, the original image can be completely restored by performing the inverse transformation. That is,
Data = (Data'-128 ・ gain) / (1-gain) ---------- (8)
Such conversion can be done automatically by giving gain. Usually, gain << 1, so there is little effect on image quality.

Watermark removal is necessary when editing images. If the watermark is weak, it may be embedded, but if you need a high-quality image, remove it once.
The final data after editing can be embedded with secondary copyright information and watermark information. In that case, watermark embedding software is used and the embedding pattern included in the key is used.

In addition, other digital watermark information may be embedded in the published image. For example, a digital watermark that embeds a green noise pattern with enhanced printing resistance. In this way, it is possible to configure a system according to usability as appropriate.

Next, the attack resistance of this method is explained. Examples of attacks on digital watermarks include signal enhancement (sharpness adjustment), noise addition, filtering (linear and nonlinear), lossy compression (JPEG, MPEG), and transformation (rotation and scaling). These attacks are attacks on luminance information, and embedding with a set of dispersed dots as in this method can be maintained unless there is an attack in the resolution direction (for example, a strong low-pass filter), and is generally robust. I can say that. Furthermore, tolerance can be increased by increasing gain.

In addition, a collusion attack that identifies a watermark pattern from an original image without watermark embedded or multiple embedded images can be avoided by changing the embedded pattern for each image. There are many dot patterns with the same spectral characteristics. For this reason, if the embedding pattern is changed for each image, the watermark data in which another image is embedded cannot be removed with the same key.

FIG. 15 shows a different cluster dot pattern p0 generated by changing the initial value (random SEED value) at the time of dot pattern creation. The spectral distribution is the same and the extraction software is common, but the dot profiles are different. When R = 64, only (4096) C (2048) different patterns are possible (C represents Combination), and the probability of the same pattern is very low, so the safety is high. Since the spectrum distribution is the same, there is no need to change the watermark extraction software. Therefore, collusion attacks can be avoided by changing the dot pattern profile for each image. The copyright holder provides the purchaser with a dot profile obtained from different random numbers as a "secret key". Since this key is different for each purchaser, the watermark cannot be removed from the images of other purchasers.

Furthermore, the security is further enhanced by creating a new dot pattern p1 rotated 90 ° from a different random number pattern. In this case, two key patterns are required, and the watermark can be removed only when both keys are available.

For ordinary digital watermarks, the watermark embedding algorithm can be roughly estimated by comparing it with the original image. For example, if you embed the image in the frequency space, after taking the difference from the original image, you can roughly guess if you try any orthogonal transform. It takes time, but the algorithm is solved. For this reason, the watermarking algorithm is not disclosed in principle for security reasons.
On the other hand, even if this algorithm is disclosed, it takes a lot of effort to decipher watermark information by reverse analysis. Furthermore, in principle, there is only one key for removing a watermark for one image, so even if the key information can be decrypted, the watermark cannot be removed using the key for other images.

Images can also be used without removing the watermark. The embedded image is a high-frequency dot pattern with blue noise characteristics, and since the pattern is hardly recognized from the low-pass characteristics of the visual system, high image quality can be maintained. The original watermark information can be retained if it is an editing process related to luminance conversion.

In addition, since information is extracted from the spatial distribution of dots, it is less affected by uneven density and uneven illumination. To increase the amount of information to be embedded, the embedded block size should be reduced, for example, 32 pixels x 32 pixels. At this time, the amount of information to be embedded is four times as large as 64 pixels × 64 pixels. However, if the block size is reduced, the number of included dots is reduced, and the population is reduced.

The embedding of this watermark is a block unit embedding and can be processed at high speed, similar to the binarization by the dither method. Extraction is somewhat complicated and computationally expensive, but because it is independent in units of blocks, parallel processing reduces the computation time in inverse proportion to the number of parallel operations, so it is effective for large images. .

The digital watermarking method of the present invention has been described above. Since this method embeds a dot pattern that exhibits blue noise characteristics in real space, the embedded image has little deterioration in image quality and can maintain high image quality. In addition, it can be restored to the original image by key information and re-embedded. For this reason, it can be presented while maintaining the original image quality in image distribution and disclosure. Also, illegal copying and unauthorized use can be monitored and contribute to copyright protection.

Also, the watermark algorithm and extraction software may be made public. For this reason, it does not become a factor which inhibits development of the system and apparatus using this method.

1 is scanner, 2 is printer, 3 is computer system, 4 is ROM, 5 is RAM, 6 is program memory, 7 is data memory, 8 is monitor, 9 is keyboard, 10 is communication function, 11 is CPU, 12 is Internet, 13 distribution of image data, 14 contact for purchase request, 15 send private key,
Reference numeral 16 denotes a copyright owner's computer, and 17 denotes a purchaser's computer.

Claims (5)

In the digital watermarking method of embedding n-bit information consisting of copyright information etc. into digital image data i (x, y),
To embed watermark information, digital image data is divided into N blocks of R pixels x R pixels (N ≥ n), and bit information (0 or 1) of watermark information embedded in each block , Watermark embedding by repeatedly embedding a watermark bit sequence using a plurality of different dot patterns pi (x, y) (i = 0,1,2, ...) showing blue noise characteristics with reduced Fourier spectrum in low frequency range Generate image w (x, y) = i (x, y) + gain pi (x, y) (where gain is the embedding strength, gain << 1),
In the extraction of watermark information, received or read image data is divided into blocks, embedded bit information and reliability are extracted by mask processing of the Fourier spectrum distribution of each block and an extraction mask, and bits with high reliability are extracted. A digital watermarking system and method characterized by selecting information. The dot pattern pi (x, y) showing the blue noise characteristic is a filter D (u, v) where R ^ 2/2 random dots are zero in the low spatial frequency range (where u, v are 2. The digital watermark system and method according to claim 1, which shows blue noise characteristics obtained by performing a filtering operation based on (two-dimensional spatial frequency). An image before embedding is obtained by obtaining, as a secret key, a dot pattern pi (x, y) at the time of embedding, a gain at the time of embedding, and a set of embedding information for the image w (x, y) in which the watermark is embedded. The i (x, y) = w (x, y) -gain * pi (x, y) can be generated, and new information can be embedded as a digital watermark using the key. Digital watermark system. The digital watermark extraction mask includes masks M0 and M1 corresponding to the spectral characteristics P0 and P1 of the embedded dot pattern, and the mask processing is performed by integrating luminance values Q0 and Q1 of the mask and the extracted spectrum distribution of the block. ,
When Q0> Q1, extraction bit = 0
Extraction bit = 1 when Q1> Q0
And the reliability is reliability = | Q0-Q1 |
4. The digital watermark system and method according to claim 3, wherein the watermark information is extracted by selecting a highly reliable one from the watermark information embedded in duplicate. In embedding the watermark information, the watermark information is first embedded in the image with a predetermined gain value, and then the embedded watermark information is extracted. If all bits are not extracted correctly, or the reliability of each bit is below a certain threshold value. In this case, the gain is increased and embedding is performed again, and the above operation is automatically performed until all bits are correctly extracted and the reliability of each bit is equal to or higher than a certain threshold. 4. The digital watermark system and method according to 4.