JP2019004313A - Electronic watermark device and method - Google Patents

Abstract

To provide an electronic watermarking device and method that extract watermark information from an image obtained by reading a print image or a display image by a scanner or a camera, can increase the amount of information to be embedded, and can stably extract information with high reliability.SOLUTION: An image is divided into blocks, and a dot pattern showing a characteristic that the spectral intensity is reduced in the low frequency region of the spatial frequency is embedded in each of the blocks as watermark information. At this time, smoothing processing is performed on the image, and the spectrum of the image is confined in the low frequency region and localized, thereby obtaining highly reliable watermark information.SELECTED DRAWING: Figure 1

Description

The present invention relates to an invisible digital watermarking apparatus and method for embedding information in digital image data. In particular, the present invention relates to a watermarking apparatus and method capable of data hiding and annotation capable of extracting watermark information from a printed image and a display image.

In today's digital information society, it has become possible for many people to share information by copying 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, due to the high image quality of digital cameras and printers in recent years, it has become possible to easily obtain copies that are not inconsistent with the original image. In addition to illegal copies that infringe copyright, counterfeits such as banknotes and securities. This is a result of promoting a malicious criminal act.

Under such circumstances, digital watermark technology for protecting 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 beginning to be widely used as a protection and security measure.

As in the case of digital watermarking, there is a data hiding technique as a technique for embedding some sub information in an image in an unperceived form. Although it is common in the point of invisibility to embed information so that neither data hiding nor digital watermark is perceived, the purpose of use is different. Even if the amount of digital watermarks to be embedded is small, they must be highly resistant to attacks. On the other hand, since data hiding embeds auxiliary sub-information for the original image in the image rather than copyright protection, the amount of information to be embedded is better, and on the other hand, resistance to attacks is not considered much.

On the other hand, there is a technology called steganography. Steganography is the same in that information is embedded so as not to be perceived, but the information is secret information, and a secret key is required to extract a watermark. In addition, there are methods and technologies called Fingerprint1 (embedding information about users to identify unauthorized distributors), Integrity Mark (detecting whether an image has been tampered with), Annotation (embedding descriptions about images), etc. Conditions such as strength, capacity, and system required for digital watermarking are different.

As an application such as data hiding and Annotation, fusion of images and the net can be considered. Various information is provided by using information embedded in an image as sub information or meta information. It is advantageous to have a large amount of information to be embedded, but since there is a limit to the amount of embedding for invisible embedding, an amount of embedding that can be guided to the information provider's home page or a simple explanation is required. .

As a method of guiding a user to the net or explaining a product, there are a barcode and a two-dimensional barcode. The user can navigate to a related home page or obtain various information by reading the coded mark with a smartphone or the like. These methods can embed a large amount of information in the code pattern. However, in these cases, a tasteless dry code pattern is placed beside an image or character information, and the content cannot be inferred for a general user only by this pattern, and it is not interesting.

For this reason, technologies such as data hiding and Annotation that embed information in related images have been attracting attention in recent years because they embed information in meaningful images and do not require extra space like code patterns. It is a digital watermark application technology. However, in order to embed as much information as possible and to read the watermark information by reading it with a scanner or camera, strong tolerance and accuracy are required.

Until now, many methods have been proposed as a digital watermark technique having resistance to image data. One example is the patchwork method. Information embedding by the patchwork method is an approach that uses statistics, so it shows strong tolerance. This method gives a slight deviation to a large number of pixel pairs in real space (pixel space) and embeds them as watermark information. First, a pixel pair is selected at random, and one luminance value is increased by δ, and the other luminance value is decreased by δ. By repeating this operation, the expected value of the distribution is biased. The watermark information can be extracted by statistically obtaining this bias. However, since this method causes statistical bias, the amount of information that can be embedded is extremely small. Also, it is vulnerable to uneven illumination during reading and uneven density during printing.

Another method is to embed watermark information in the frequency space of an image. 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. However, in this method, if the amount of information to be embedded is increased, it is difficult to separate the spatial frequency of the image and the frequency of the embedded spectrum. In addition, if the embedding strength is increased, a periodic pattern is generated, the image is easily noticeable in the flat part of the image, and moire fringes with the image are generated, resulting in deterioration in image quality, and the accuracy of extracting watermark information is not good. . There is not much information that can be embedded.

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, even if a large amount of information is embedded as much as possible, the image quality is not visually deteriorated, and even if a watermarked image is printed or displayed with a scanner or camera, the watermark information can be accurately obtained. The challenge is to obtain a robust digital watermarking method that can be extracted.

Furthermore, it is necessary that the watermark information can be stably extracted with high reliability regardless of the content of the embedded image. In particular, it is also an object to obtain a digital watermark method that can stably extract watermark information even in the case of an image composed of a fine pattern with a high spatial frequency and a high spectrum in the embedded image.

In order to solve the above problems, the present invention increases the amount of embedded information by superimposing and embedding a small-sized random dot pattern having a reduced spectrum intensity in a low frequency band in the original image, and , Increase the strength of embedding and increase resistance. Even in such a case, a durable digital watermark is realized without visually degrading image quality.

Random dot patterns used for embedding exhibit blue noise characteristics with reduced intensity at low frequencies or green noise characteristics with reduced intensity at high and low frequencies. Such dot patterns are difficult to recognize from human visual characteristics. For this reason, it is possible to further increase the resistance by embedding strongly.

In general, the spectral characteristics of image data are distributed in the vicinity of zero frequency. For this reason, the dot pattern used for embedding has little spectrum overlap with the image data, it is easy to separate each spectrum, and the embedded information can be extracted with high reliability.

To embed watermark information, digital image data is divided into R pixel x R pixel blocks, and dot patterns with different Fourier spectra are created corresponding to the bit information (0 or 1) of the watermark information embedded in each block. Use it to embed watermarks. In the extraction of watermark information, received or read image data is corrected, and then divided into blocks. The Fourier spectrum distribution of each block is obtained, and embedding information is extracted from the spectrum distribution by pattern recognition.

In order to increase the amount of information to be embedded, it is divided into as small blocks as possible. When the block size is reduced, the reliability of the extracted data is generally lowered. For this reason, smoothing processing is performed on the entire image before embedding, and an operation for confining the spectral characteristics of the original image in a low frequency is performed, thereby improving the extraction accuracy.

According to the present invention, it is possible to increase the amount of embedded information and to perform robust watermark embedding. For this reason, embedded watermark information can be extracted with high reliability even in an image obtained by reading a print image or a display image with a scanner or a camera.

1 is a diagram of an electronic watermark embedding and extracting apparatus according to the present invention. FIG. The figure which showed the delivery on the internet of the image with a watermark. The figure which showed 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 characteristic, (b) is the one rotated 90 degrees, (c) is the blue noise characteristic, (d ) Has a circular ring spectrum characteristic, and (e) is for header. In the actual embedded image, (a) is the embedded image and (b) is a partially enlarged view. The figure which shows the flow of watermark embedding. The figure showing the image data with which the watermark was embedded, and its spectral characteristic. The figure showing the processing flow of watermark extraction. The figure of the mask pattern of watermark extraction. The figure which shows the mask process of watermark extraction. The figure which shows reliability improvement by the mask process of watermark extraction. The figure showing the automatic adjustment of smoothing. This is a picture taken with a camera and watermark extraction. (A) is a picture of a watermarked image displayed on a monitor taken with the camera, (b) is a picture cut out after correcting the image part, (c) is the extracted spectral image. The figure which shows the dot pattern when embedding 2-bit multi-value data, and its spectrum.

This digital watermark implements data hiding and annotation in which information is embedded in an image, the image is read by a scanner, a digital camera, a smartphone, etc., and embedded information is extracted. The information to be embedded includes the home page address and URL to be linked, and sub information of the image. Since images of a relatively small size are often used, it is necessary to embed a large amount of information even in a small image.

Furthermore, it is necessary to improve printing durability and visually mitigate image quality degradation due to embedding. For this reason, a dot pattern exhibiting green noise or blue noise is embedded and superimposed on the image. The embedded image is a medium for these sub-information, and since there is no secondary use of the image itself, some degradation can be tolerated, and priority is given to ensuring the amount of embedded information and the accuracy of the extracted information. Is done.

For this reason, in the present invention, the amount of information and the reliability are ensured by dividing the image into small block sizes and increasing the embedding strength gain. Decreasing the block size and increasing the embedding strength gain generally causes image quality degradation and lowers extraction accuracy. Therefore, a dot pattern showing green noise and blue noise characteristics is used to prevent the image quality from being deteriorated, and at the same time, the extraction accuracy is improved by preprocessing and correction processing at the time of extraction.

In order to implement such a digital watermark technique, in the present invention, a plurality of dot patterns having different small sizes are used for embedding in an original image. Such a dot pattern exhibits green noise and blue noise characteristics, and the graininess and noise of the embedded image are visually low. A plurality of patterns are selected according to the bit information to be embedded.

Such a dot pattern is obtained by repetitive calculation from a random dot pattern generated from a pseudo-random number. Many different patterns are generated by changing the filter shape and initial state, and using these patterns, a pattern is selected according to bit information including a character string such as an e-mail address and a URL, and embedded in image data.

Hereinafter, it demonstrates in detail along an Example.
FIG. 1 is a block diagram of a digital watermark system for embedding and extracting information according to the present invention. In the computer 3 of the information provider who wants to embed a watermark, image data photographed by the digital camera 18 or the like is stored in a data memory 7 such as a hard disk. The image data is subjected to image processing 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 heavily loaded, there may be a processing board for speeding up the GPU or the like.

FIG. 2 shows an Internet distribution processing procedure. The image in which the watermark information is embedded is distributed from the computer system of the information provider 16 via the Internet 12 (13). Or you may distribute as printed matter with the paper medium of a flyer or an advertisement. The user 17 who has received the information reads the watermark information with the scanner 1, digital camera, smart phone 18, etc., extracts the watermark information with watermark extraction software, and accesses the information provider's homepage (14). Then, further detailed information can be received (15).

Here, the dot profile creation algorithm used in the present invention will be described along the processing flow of 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. And p (x, y). At this time, the initial dot profile can be changed by changing the initial value (SEED value) of the pseudorandom number generator.
STEP1: Perform a two-dimensional Fourier transform of the p (x, y) dot profile to obtain P (u, v).
STEP 2: Multiply P (u, v) by filter D (u, v) to obtain a new P ′ (u, v).
STEP 3: Perform inverse Fourier transform on P ′ (u, v) to obtain a multivalued point profile p ′ (x, y).
STEP 4: An error function e (x, y) = p ′ (x, y) −p (x, y) is obtained, and white and black are inverted in descending order of error at each pixel position.
STEP5: Repeat the above operation until the error is within the allowable range.
The target dot profile is finally obtained by performing the above operations.

Next, the filter D (u, v) will be described. D (u, v) is a filter that is axisymmetric in the u-axis direction and the v-axis direction, and cuts low and high frequencies. This may be a circle, an ellipse or a rectangle. When the low-frequency cutoff frequency in the u-axis direction is fmin, the high-frequency cutoff frequency is fmax, and f0 = (1/2) ^ (1/2) · fn,
a = (fmax-f0) / fn
b = (fmin-f0) / fn
Then, a dot profile independent of R is obtained. Here, fn represents the Nyquist frequency. Here, when a is infinite or close to it, the blue noise characteristic cuts only the low range, and when fmax and fmin are finite, the green noise characteristic cuts the low range and the high range.

As an example, a dot profile with a block size R = 32 is obtained. In general, the larger the block size, the higher the accuracy of extracting the watermark. This is because the watermark is extracted from the dot collection state and distribution, and if the block size is small, the number of populations of the collection decreases, statistical fluctuation increases, and accuracy decreases. As a result of various experiments, R = 64 or more is optimal. However, in order to increase the amount of information to be embedded, R = 32 was separately considered as a measure for improving accuracy.

FIG. 3 shows a dot profile for 32 pixels × 32 pixels. FIG. 4A shows a dot profile p0 (i, j) represented by (a, b) = (0, −5 / 16) and its spectral characteristic P0. An ellipticity of 1.5 means that the filter D (u, v) is enlarged 1.5 times in the v-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. 9B is obtained by rotating p0 (i, j) by 90 °, and the spectrum is similarly rotated by 90 °. (C) is a blue noise pattern of (a, b) = (∞, 0), (d) is a pattern by a circular filter of (a, b) = (3/16, -1 / 16), (e ) Is a pattern with a rectangular filter of (a, b) = (0, -1 / 4). All of these patterns are random dot patterns with a reduced spectrum low frequency range.

Next, a watermark embedding algorithm for multilevel image data will be described.
For embedding watermark information, color image data is converted into Y, Cb, and Cr, and watermark information is embedded in luminance Y. When embedding in Blue (yellow when printing), it is visually inconspicuous, but it is necessary to accurately separate colors when reading.

Now, suppose that different spectral patterns are Pi (u, v) (i = 0,..., N) by embedding different types of spectral patterns in block units in the frequency space. Pi shows the characteristic that the spectrum is reduced at the aforementioned low frequency. As the low-frequency cutoff frequency fmin, it is ideal that a frequency higher than the maximum frequency of the embedded image is selected so that the two do not overlap. If the frequency spectrum of the block image 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, gain represents the embedding strength.

Invisible digital watermarks typically reduce the embedding strength and prevent the embedded pattern from being recognized. This is because copyright information is embedded in an image as a watermark for the purpose of protecting the copyright, but the image itself is a product, and degradation of image quality becomes a problem. However, in an application that embeds a URL or the like through an image and leads to a homepage, priority is given to being able to extract a large amount of information accurately. Therefore, in the present invention, it is embedded with a large gain. Nevertheless, significant degradation in image quality does not make the user feel attractive. As a result of various experiments, it was found that when gain is 0.25 or more and 0.75 or less, deterioration is not visually noticeable and accurate embedded information can be extracted.

The actual watermark information embedding work is performed in real space. Converting equation (1) to real space (pixel space) (represented in lower case):
w (x, y) = i (x, y) + gain ・ pi (x, y) ---------- (2)
It becomes. Assume that 1-bit information of watermark information is embedded in one block. For the watermark information, two dot patterns p0 (x, y) and p1 (x, y) showing different green noise characteristics are prepared, and for the embedded bit information (0,1),
When embedded 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 are assumed to be (1/2, -1/2) in order to preserve the average luminance.

Moreover, since p0 and p1 are random dots, the boundary between them is not noticeable. Halftone screens exhibiting blue noise and green noise characteristics are often used as FM screens in printers. They are highly uniform with dispersive dots and are uniform in human vision and do not feel graininess.

(C) of FIG. 4 is (a, b) = (∞, 0), that is, the patterns p2 (x, y) and (d) showing the blue noise characteristics are (a, b) = (3/16, −1). / 16) is a pattern p3 (x, y) indicating the green noise characteristic generated from the circular filter. These filters are used for embedding 2-bit data described later. Further, (e) is a pattern showing the green noise characteristics by the rectangular filter of (a, b) = (0, -1/4), and is used as a header pattern described later.

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)
Given in. 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. When the block size is R = 32, when 512 pixels × 512 pixels, an information amount of N = 256 bits = 32 bytes can be embedded. This is not so much to embed a normal URL or homepage address. In order to embed more information, the image size is increased or multi-value embedding described later is performed.

Usually, an error occurs in the extraction of watermark information. Also in this method, when the original image is very fine and has a high frequency, the spectrum distribution overlaps with the spectrum distribution of the embedding pattern and is not correctly extracted. Therefore, normally, a bit string to be embedded is repeatedly embedded several times, and a majority vote or a highly reliable one is selected to extract embedded information. However, if the number of repetitions is increased, the number of bits that can be embedded further decreases.

Therefore, embedding this watermark applies a slight low-pass filter to the image data in advance to reduce the high frequency. For example, in a low-pass filter with 2 pixels × 2 pixels and the same kernel coefficient (1), the maximum spatial frequency of the image is ½. In the case of 3x3, it becomes 1/3. Therefore, by performing smoothing processing on the image data before embedding the watermark information, the image spectrum can be confined below the low-frequency cut-off frequency of the dot pattern, and the accuracy in watermark extraction can be improved. It becomes possible.

Furthermore, by increasing the embedding strength gain as described above, it is possible to improve tolerance and improve extraction accuracy. By setting the gain to 0.25 or more and 0.75 or less, it is possible to embed it strongly without causing much deterioration in image quality. Gain = 0.25 is 8-bit image data, and the lower 5 bits of the image data change by superimposing a dot pattern with a value of ± 32. FIG. 5A shows an image obtained by embedding a 512 pixel × 512 pixel image with gain = 0.375, and a dot pattern of ± 48 is superimposed on the original image. FIG. 2B is an enlarged view of a part of which a noise due to a dot pattern is remarkable. However, due to green noise and blue noise characteristics, noise is not so conspicuous in the normal size due to human visual characteristics. When gain = 0.75, the value of ± 96 is superimposed, and the seventh bit of the image data changes.
As described above, by performing smoothing processing on image data and embedding with strong gain, it is possible to enhance tolerance and to perform accurate and highly reliable extraction by watermark extraction.

FIG. 6 shows a watermark embedding process flow. First, the original image is uniformly smoothed by a 2 × 2 filter (20). Subsequently, a watermark information bit string wm (i, j) to be embedded is prepared (21). Here, i represents the i-th character to be embedded, and j represents 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. However, N is the maximum number of embedded bits shown in Equation (3). Subsequently, the image data is divided into R × R blocks (22), and a dot pattern is synthesized in units of blocks. Whether or not it is the head of the character string (23). Embed (pH). After that, if the embedded bit is 0, the pattern of p0 is embedded, and if the embedded bit is 1, the pattern of p1 is embedded (24). The process ends when all blocks have been embedded.

FIG. 7 shows an image of 512 pixels × 512 pixels in which 26 letters of alphabet “a, b, c,... Z” are embedded in a dot pattern with a block size R = 32 and an embedding strength gain = 0.375. This image can embed information of up to 32 bytes (32 ASCII characters) from Equation (3).
FIG. 4A shows an embedded image. FIG. 5B shows the extraction result described later. In the figure, a vertically long ellipse represents a bit “0”, and a horizontally long ellipse represents “1”. A square pattern is a header pattern at the beginning of a character string, and is used as a delimiter for repeated patterns. Since gain is large, embedded information is accurately extracted from data before printing. In anticipation of deterioration in image quality due to printing or reading from a display image, there is a margin in reliability.

In the extraction of watermark information, the spectrum image is obtained in units of blocks after correcting the size and inclination of the read image. Fourier transform of equation (2)
F {w (x, y)} = F {i (x, y)} + gain ・ F {pi (x, y)} ---------- (4)
It becomes. However, F {} represents a Fourier transform. Usually, F {i (x, y)} is the spectrum of the original image, localized near 0 frequency, and F {pi (x, y)} is the spectrum P0, P1 of the dot pattern used for embedding and elliptical The image spectrum is confined in a ring shape. For this reason, there is little overlap of both, and it is easy to isolate | separate both.

FIG. 8 shows a watermark extraction processing flow. First, a watermarked image is taken with a scanner or camera (36), and the image is subjected to projective transformation or Affine transformation to obtain a rectangular corrected image (37). When shooting with a digital camera or the like, trapezoidal distortion or the like occurs when shooting from an oblique direction, so that projection conversion is performed and correction is performed, and an image cut into a rectangle is resized so as to have a predetermined resolution. Subsequently, an edge enhanced image is created (38). The edge emphasis process makes the outline of the dot clear and improves the extraction accuracy. After that, the image is divided into R × R blocks and FFT is performed for each block (39). Subsequently, histogram equalization is applied to the obtained spectral image (40). This is for improving the contrast of the spectral pattern and at the same time standardizing by quantifying the reliability described later. Subsequently, pattern recognition by mask processing described later is performed (41), and data of watermark information wm ′ (i, j) and reliability wmrel ′ (i, j) is obtained (42). The above operation is performed for each block, and ends when all the blocks are completed.

FIG. 9 shows mask patterns M0 and M1 as discriminators for extracting watermark information by pattern recognition from the spectrum distribution after Fourier transform. M0 is a vertically long mask and M1 is a horizontally long elliptical mask corresponding to the spectral characteristics P0 and P1 of the dot pattern for embedding. In the figure, the black region has a value of 1, and the white portion has a value of 0. This mask is overlaid on the watermarked spectrum, and it is determined whether the bit is 0 or 1 from the difference between the integrated luminance values Q0 and Q1 of the overlapping part. FIG. 10 shows this processing flow, and outputs Q0 and Q1 of the following integrated luminance values are obtained for the block spectral pattern W (i, j) (40).
Q0 = M0 ◎ W = 1 / ZΣM0 (i, j) ・ W (i, j)
Q1 = M1 ◎ W = 1 / ZΣM1 (i, j) ・ W (i, j)
Here, ◎ represents a mask calculation for obtaining the integrated luminance value of the overlapping portion, and Z represents the number of pixels having a mask value of 1. Therefore, it represents the luminance value per pixel in the mask area. 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 higher the reliability, the higher the reliability of the extracted bit.

FIG. 11 illustrates measures for improving the accuracy and reliability of watermark information. The maximum number N of embeddable bits is given from Equation (3) depending on the size of the image. If the number of bits to be embedded is n and n <N, the bit strings can be embedded in duplicate. For this reason, first, watermark information wm ′ and reliability wmrel ′ up to N (N is the number that can be embedded) are obtained without considering repetition of character strings in the extraction of watermark information. After that, considering the repetition, the highly reliable one of the continuous character strings is selected and the watermark information up to N is inspected.

This is expressed by an expression, where i is a character number (0 <i ≦ N), j is a bit number (0 ≦ j ≦ 7), and the remainder k is represented by k = i mod n modulo n. (50), (i, j) th reliability wmrel 'is
When wmrel '(i, j)> wmrel (k, j)
wm (k, j) = wm '(i, j)
wmrel (k, j) = wmrel '(i, j)
(52), the watermark information is replaced with one having high reliability. This is performed until i is N, that is, until the total number of embedded bits is completed.

Through 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 has a smaller extraction error than the method determined by majority vote.

The watermark is extracted as described above. FIG. 7B shows the spectral distribution during extraction. Embedding is an ASCII character in which 26 alphabetic letters “abcd... Xyz” are embedded. Each character is expanded into bits, and as described in FIG. 3, a pattern of p0 is embedded for “0” and a pattern of p1 is embedded for “1”. Since all the first bits (MSB) of ASCII characters are 0, this is used as a header pattern for character string separation. This can also be used as an Affine transform coefficient detection pattern by measuring the rectangle and the size direction. Since the number of characters that can be embedded is 32 bytes, the six characters “abcdef” are embedded in duplicate. Therefore, the accuracy of the extraction result can be improved by selecting the above-described one with high reliability.

When the embedded watermark information has a fine original image and a high spatial frequency, extraction may not be possible or reliability may be low. Therefore, a process for increasing the reliability at the time of embedding is performed. FIG. 12 shows a flow for performing this automatic adjustment. First, the image is smoothed by M × M, that is, 2 pixels × 2 pixels, with M = 2 and gain = 0.25 (30). Thereafter, the watermark information wm (i, j) is embedded (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 bits is equal to or greater than a predetermined threshold (33). If the watermark information does not match, or if the watermark information matches but the reliability is equal to or less than the threshold, the gain is increased by a predetermined amount Δ and embedding is performed again. When the gain reaches the maximum value of 0.75 (34), and when it reaches, the smoothing size M is incremented by 1, and the watermark is embedded again. Eventually, embedding satisfying a predetermined reliability becomes possible. The embedding can be automatically performed so that the watermark information embedded as described above can be read with a sufficient margin.

FIG. 13 is a diagram when the watermark is embedded by photographing the image in which the watermark is embedded with a digital camera or a smartphone. In the figure (a), an image of 512 pixels × 512 pixels is smoothed by 2 pixels, and an image embedded with gain = 0.375 is taken with a digital camera. The embedded data is 26 characters of alphabet “a, b, c, d,..., X, y, z”. (B) is obtained by cutting out 512 pixels × 512 pixels after correction. (C) is read by the watermark extraction software and is correctly extracted. Since this method extracts information from the spatial distribution of dots, it is not easily affected by illumination unevenness or density unevenness when reading an image. The automatic adjustment described above makes it possible to accurately capture images with any spatial frequency. It becomes possible to read.

In order to increase the amount of information to be embedded, multi-level embedding is possible. In FIG. 14, 2-bit information can be embedded using a plurality of four dot patterns described in FIG. That is, it is possible to embed multiple values (four values in this case). The four patterns p0 (x, y), p1 (x, y), p2 (x, y), and p3 (x, y) having different spectral distributions are 2 bits (00,01,10) of the bit information to be embedded. , 11), each pattern is selected. Extraction is performed using masks M0, M1, M2, and M3 corresponding to the respective spectrum distributions.

Such 2-bit embedding doubles the amount of information that can be embedded. That is, it is possible to embed 64-byte information in an image of 512 pixels × 512 pixels. However, the extraction accuracy decreases as the number of patterns increases. For this reason, it is necessary to further increase the smoothing and embedding strength. Therefore, when it is necessary to embed a large amount of information in a small-sized image, it is effective when the image quality is somewhat sacrificed.

As described above, the digital watermarking method of the present invention is effective as an interface means for connecting an image and the net because the embedded information can be accurately extracted from the image in which the information is embedded by reading it with a scanner or a camera. . For example, it is possible to embed the URL of the seller in the product photo as a watermark and direct the user to the home page.

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 access to homepage and contact for detailed information request, 15 distribution of detailed information, 16 an image distributor's computer, 17 a user's computer, 18 a digital camera or smartphone Represent.

Claims (6)

In a digital watermark method for embedding n-bit information in digital image data i (x, y),
For embedding watermark information, digital image data is divided into N blocks of R pixels x R pixels (R is a power of 2) (N≥n) and corresponds to bit information of watermark information embedded in each block Using a plurality of different dot patterns pi (x, y) (i = 0,1,2, ...) and embedding strength (gain) whose Fourier spectrum is reduced in the low band,
w (x, y) = i (x, y) + gain ・ pi (x, y)
Generate a watermark embedded image w (x, y) as
The watermark information is extracted by performing correction processing on the image in which the watermark is embedded, dividing the image into blocks, and extracting the embedded information from the spectral distribution of each block.
The embedding strength gain is
0.75 ≧ gain ≧ 0.25
A digital watermarking apparatus and method characterized by smoothing image data i (x, y) before embedding. A plurality of different dot patterns in which the Fourier spectrum is reduced in the low frequency range are R ^ 2/2 random dots, and a filter D (u, v) that becomes 0 in the low frequency range (where u, v are The filter D (u, v) is obtained by performing a filtering operation based on (two-dimensional spatial frequency), and in this operation, the filter D (u, v) is axisymmetric with respect to the u axis and the v axis. Electronic watermarking apparatus and method. The plurality of dot patterns exhibit a green noise characteristic whose spectrum is reduced even in a high frequency range, and two elliptical patterns p0 (x, y) and p1 (x, y) whose spectral distributions are orthogonal to each other. The digital watermarking apparatus and method according to claim 2, wherein the watermark information is embedded by selecting a pattern corresponding to the bit information (0, 1) to be embedded. The plurality of dot patterns are composed of four patterns p0 (x, y), p1 (x, y), p2 (x, y), and p3 (x, y) having different spectral distributions. The digital watermarking apparatus and method according to claim 2, wherein a pattern is selected and information is embedded corresponding to 2 bits (00, 01, 10, 11). The extraction of the digital watermark includes masks M0 and M1 corresponding to the spectral characteristics of the two embedded dot patterns. From the integrated luminance values Q0 and Q1 of the mask and the extracted spectral distribution in block units,
When Q0> Q1, extraction bit = 0
Extraction bit = 1 when Q1> Q0
And the reliability is reliability = | Q0-Q1 |
4. The digital watermarking apparatus and method according to claim 3, wherein those having a reliability greater than a certain reference value are extracted. The watermark information is embedded by first smoothing and embedding an image with a minimum size smoothing filter and minimum gain, and then extracting watermark information. If all bits are not extracted correctly, or If the bit reliability is below a certain threshold, increase the size of the gain and smoothing filter, and repeat the above operation until all bits are correctly extracted and the reliability of each bit is above a certain threshold. 6. The watermarking system and method according to claim 5, wherein the watermarking system and method are performed automatically.