KR100487071B1 - Data Hiding and Extraction System Using JND and CSF - Google Patents

Abstract

The present invention invisibly conceals a large amount of data in a digital image and compares it with the original to generate an image that cannot be distinguished by the human eye, and then stores or transmits it, and then re-extracts the hidden data. A data concealment and extraction system that can be used for purposes.

JND: Just a study of the human visual system of modern biology, in order to conceal more data under the constraint that the image that conceals the data must be indistinguishable from the original image and the human eye. Two results are used: Noticeable Difference (CSF) and Contrast Sensitivity Function (CSF).

J & D said, "In order to make the eyes of ordinary people recognize that the intensity of Pixel (Picture Cell), which is the basic unit of an image, is different from that of other pixels of attention, the brightness between two pixels should be at least. The difference in brightness varies depending on the brightness of the pixel that is used as the reference for comparison. ”The" contrast difference detection power of the human eye is to display the brightness of the pixel as a signal in the spatial domain. " Depends on the frequency of the signal. "

In the present invention, the number of bits concealable for each pixel of the image is determined using the J & D property, and the data concealment order is determined using the CS property, and in the case of a color image, the color is R (red) ), G (green), and B (blue) components, and then R and B components are relatively invisible to the human eye compared to the G component. By concealing, it maintains the same invisibility compared to conventional data concealment and extraction systems, allowing more data to be concealed when concealing data, in other words, much more invisible when concealing the same amount of data in the image. We propose a method to hide while increasing visibility.

Description

Data Hiding and Extraction System Using JND and CSF}

The technical field to which the present invention belongs is to conceal a large amount of data invisibly to the original digital image, generate an image that cannot be distinguished from the human eye by comparing it with the original, and then store or transmit it and then conceal the hidden image. It is a field of data concealment and extraction technology to re-extract data and use it for various purposes. Recently, it is a technology field which is widely applied to the storage, transmission, copyright protection, original proof, secret communication, etc. of multimedia data through the Internet. Also called digital watermarking or steganography. A technology for concealing and extracting multimedia data such as text data, still images, sounds, and moving images, which are called message data, into a digitized original image, and the image with hidden message data is stored and transmitted, and then the message data concealed therefrom according to future use. In this case, the extracted message data must be extracted accurately without loss, and the image where the message data is hidden must be hidden so that it is impossible to distinguish the original image from the human eye. Therefore, the third party except the person who generates and transmits the image where the message data is hidden and does not know whether the message data is hidden should be kept secret.

Conventionally, the simplest and most widely used data concealment technology for digital images is data indicating the intensity of pixels (pixels), which is a basic unit constituting a digital image. Data is used, and in the case of a color image, a message data is concealed in LSB (Least Signigicant bit), which is the least significant bit of R, G, and B color bytes. Van Schyndel, A. Tirkel, and C. Osborne. "A digital watermark" Proc of the IEEE ICIP vol 2, pp. 86-90, 1994]. This method includes a method of concealing message data directly to LSB of a digital image, or converting the message data by first calculating or combining them. [RB Wolfgang and E. J Delp. "A watermark for digital images". Proc of the IEEE ICIP , Lausanne, Switzerland, vol. 3, pp. 219-222, September, 1996].

However, such a method has a limitation in concealing a large amount of message data, and technology is popularized, so it is highly likely that a third party will be able to conceal message data.

The present invention is to solve the problems of the prior art as described above, in consideration of the characteristics of the human visual system (Human Visual System) to adjust the amount of bits concealed for each pixel and the pixel variability and adaptive message to determine the hidden order By providing a data hiding method, it achieves the purpose of concealing a large amount of message data and increasing invisibility than the conventional method.

There are two characteristics of the human visual system called Just Noticeable Difference (JND) and Contrast Sensitivity Function (CSF). J & D said, "The difference in brightness that should be minimum between two pixels in order to make the eyes of ordinary people recognize that the brightness of the pixel, which is the basic unit of the image, is different from the brightness of other pixels of attention. "It depends on the brightness of the pixel," CSF said. "The contrast difference detection power of the human eye depends on the frequency of the signal when displaying the brightness of the pixel as a signal in the spatial domain." It is a study result.

According to a preferred embodiment of the present invention, a data concealment system for concealing message data in an original image, the data concealment system comprising: a concealment coefficient generator for generating a concealment coefficient using a J & D value for each pixel of the original image; A concealed order generator for generating a concealed order by calculating a variance or difference value with surrounding pixels to consider the CS from the original image; A comparator comparing the total amount of concealable data calculated from the concealment coefficient of the original image with the amount of message data and outputting a result thereof; And a message data concealer for concealing the message data based on the concealment coefficient and the concealment order when the amount of concealable data of the original image is greater than or equal to the amount of concealed data to be concealed by the comparator. A system is provided.

Preferably, if the amount of data concealable in the original image is smaller than the amount of message data to be concealed, the comparison result of the comparator is terminated.

Also preferably, the message data concealer is:

A hidden bit initializer for determining the number of bits to be hidden for each pixel according to the order of concealment when the amount of data concealable in the original image is greater than or equal to the amount of message data to be concealed;

A message data bit concealer for concealing one bit in the initialized pixel; And an extra bit calculator for determining whether to end the concealment process. If the result of the extra bit calculator corresponds to the end of the concealment process, the concealment process is terminated.

According to another aspect of the present invention, in the data extraction system for extracting the message data from the original image where the message data is hidden,

A hidden coefficient generator for generating a hidden coefficient from an image in which message data is hidden;

A concealed order generator for generating a concealed order from an image in which message data is concealed; And

Provided is a data extraction system comprising a message data extractor for extracting message data based on the concealment order generator and the concealment order generator.

Preferably, the message data extractor is:

A message data amount calculator for calculating a message data amount from an image in which the message data is hidden;

A message data bit extractor for extracting message data by reading a position of a pixel to be extracted from a concealment order and a concealment coefficient value and a bit position in the pixel; And

And an extra bit calculator for determining whether to end the message data extraction process.

Also preferably, the data concealment system may conceal the concealment order from the high frequency band by using the characteristics of the CS in order to improve the invisibility.

In addition, preferably, when obtaining the concealment order, the dispersion order is calculated using only the G component, instead of the dispersion of the R, G, and B components, respectively.

In addition, preferably, the block unit variance is obtained by dividing the image into block units to reduce the time required and to use less memory.

In addition, preferably, the image is divided into blocks to obtain a difference from neighboring pixels in blocks in order to reduce the time required and to use less memory.

Also, preferably, the message data concealment order is performed by dividing the message data bit concealer into four processes to improve invisibility. In the first process, among the pixels having a hidden coefficient of 4, the message data is hidden one bit in the order of pixels in the high frequency band from the high frequency band to the low frequency band. In the second process, among the pixels with the hidden coefficients 4 and 3, the position of the hidden sequence is not overlapped with the first process in the order of the highest to the lowest pixel, that is, if the hidden coefficient is 4, one of the hidden positions in the previous process is LSB. In the higher bits, since the concealment coefficient is 3, it is the first concealment. In the third process, when the hidden coefficients are 4, 3, and 2, that is, the message data is concealed by one bit according to the concealment order so that all pixel values do not overlap with the concealed position in the previous process. In the last process, the concealment position and order are the same as in the third process, and only the bit positions to be concealed are concealed in higher bits than the third process. During this process, if there is no more message data to conceal, the message data bit concealer terminates.

Preferably, when concealing bits of the message data by the number of bits corresponding to the concealment coefficient for each pixel during message data concealment, one bit is concealed for each message data bit concealment process for each pixel, instead of being concealed at once. If all of the message data is concealed in the pixels corresponding to the concealment coefficient to be concealed in the following process, when concealing according to the concealment procedure in the following process, it is characterized by concealing one bit again in the concealment order.

In addition, preferably, the invisibility of the color image is enhanced by using the characteristics of the color image. In the concealment process, the relation between the R, B component and the G component is obtained. It is characterized in that it is pixel adaptive to allocate and conceal fewer bits than the concealment coefficient of.

Also, preferably, when concealing the same amount of message data, in order to prevent the low-frequency component from being much less invisible than the high-frequency component, it is more than the pre-calculated concealment coefficient for pixels having high priority according to the concealment order. Many bits are allocated, and low priority pixels are assigned fewer bits than hidden coefficients.

In addition, preferably, in order to prevent the concealment of the message data sequentially, when obtaining the concealment order, a uniform random number is applied to the same variance to add an arbitrary property of the concealment order. .

In another aspect of the present invention, a hidden coefficient is generated from an original image;

Generate a concealment order from the original image;

Comparing the amount of message data to be concealed with the amount of concealable data of the original image to determine whether the amount of concealable data of the original image is greater than or equal to the amount of message data to be concealed;

If the amount of concealable data of the original image is greater than or equal to the amount of message data to be concealed, the data may be subsequently concealed based on the message data concealment sequence according to the concealment sequence corresponding to the concealment coefficient. A concealment method is provided.

Also preferably, the message hiding process is:

If the amount of concealable data of the original image is greater than or equal to the amount of message data to be concealed, find the number of bits to be concealed, and then find a position to conceal;

Initialize a bit of a location to conceal;

Conceal one bit in the initialized concealment position;

Thereafter, it is determined whether the concealment process ends, and if it corresponds to the end of the concealment process, the concealment process is ended.

According to another aspect of the present invention, there is provided a method for reading an image in which message data is hidden;

Generating a concealment coefficient of an image in which the message data is concealed;

Generate a hiding sequence of an image in which the message data is hidden;

A data concealment method is provided, wherein data is extracted from a message by reading coordinate values and bits to be extracted from the generated concealment order and concealment coefficient values.

Comparing the advantages of the method proposed by the present invention with the conventional method, first, message data is hidden according to pixel values without uniformly hiding the message data of LSB, so that the amount of message data concealment can be increased, and surrounding pixel values. By using the statistical properties with respect to the invisibility deterioration due to the increase of the amount of concealment can be prevented.

In the present invention, first, the amount of data concealment of each pixel of the original image is determined using J & D, and the message data is concealed based on the statistical message data concealment order considering invisibility. The concealment coefficient is defined as an amount of concealment of data of each of the R, G, and B components of the pixel.

First, as to the concealment coefficient generator design, the implementation of the concealment coefficient generator in the present invention obtains each region according to the contrast of the image based on Weber's law and allocates the concealment coefficient to each region [MK Kundu] , SK Pal, "Thresholding for edge detection using human psychovisual phenomena", Pattern Recognition Letters 4, pp. 433-441, 1986]. Where I is the brightness of the pixel, Is a maximum variable value at which the pixel value does not cause a visual difference with the surrounding background. The image is divided according to the contrast as follows.

, 0 = I <{x} _ {1}

, {x} _ {1} = I <{x} _ {2}

, {x} _ {2} = I <{x} _ {3}

, {x} _ {3} = I (1)

Where beta is a Weber constant of 0.04, Is 2.05. , Where Is the maximum value that I can have and {2} ^ {8} -1 = 255 when the image is represented by 8 bits. {a} _ {1}, {a} _ {2}, {a} _ {3} are 0.1, 0.3, and 0.9, respectively. Thus, the image is divided into four areas according to the contrast, and the hidden coefficient for each area is determined according to the value of J & D as follows [WN Lie, LC Chang, "Data Hiding In Images with adaptive Numbers of Least Significant Bits. Based On The Human Visual System, " Proc of the IEEE ICIP, vol. 1, pp. 286-290, 1999].

(2)

The hidden coefficient of 8-bit black and white image is as follows.

(3)

Where {k} _ {1}, {k} _ {2}, {k} _ {3} are 0, 86, 183.

In Equation (1), it can be seen that the first region can be modified to 2 bits with a value of 3.62. Therefore, the concealment coefficient can be simplified as follows.

(4)

here Is the final calculated concealment coefficient and {k_2} ^ {'}, {k_3} ^ {'} are 88 and 192, respectively.

In the present invention, in order to increase the invisibility in the message data concealment process, concealment is sequentially performed from the high frequency region to the low frequency region. The hidden order generator is implemented based on CSF. 11, it can be seen that the CS value decreases as the frequency increases. This means that the high frequency band is not as sensitive to human vision to change in pixel value as compared to the low frequency band. In the present invention, a local "dispersion" or "difference from the surrounding pixel" of each pixel is used as a number representing the value of the frequency for each pixel. The difference D between the variance {sigma} ^ {2} and the surrounding pixels is obtained as follows.

(5)

(6)

Where mu is the average of the regions for which variance is to be obtained, and I_ {m, n} is the pixel value. M and N use 3x3 or 5x5.

In the present invention, the variance of all R, G, and B components of the color image is not determined for the color image, and only the variance of the G component is obtained, and then the value is applied to the R and B components. The reason is that when the color represented by the R, G, and B components is converted into Y, U, and V components using equation (7), the G component has the greatest effect on the Y component, and this Y component corresponds to the brightness of the pixel. Because it is a component, only the G component is used to save computation time.

Y = 0.29900 * R {} + {} 0.58700 * G {} + {} 0.11400 * B (7)

1 is a schematic flowchart of a data concealment system according to the present invention, FIG. 2 is a flowchart specifically illustrating a message data concealment process in FIG. 1, and FIG. 3 is a schematic block diagram of a data concealment system according to the present invention. 4 is a detailed block diagram of a message concealer for performing a message data concealment process in FIG.

As shown in Fig. 1, first, an original image and message data to be concealed in the original image are stored in a memory (not shown) of the device. After that, when a signal for requesting concealment of message data is input from the user (S1), the hidden coefficient generator 2 (see FIG. 3) reads the original image stored in the memory (S2). Then, the hidden coefficient generator 2 generates a hidden coefficient of the original image (S3). Then, the concealment order generator 4 is operated to generate the concealment order from the original image. The concealment coefficient and concealment order for the original image thus generated are stored in the memory.

Thereafter, the comparator 6 compares the amount of message data to be concealed with the amount of concealable data of the original image (S6). Specifically, it determines whether the amount of concealable data of the original image is greater than or equal to the amount of message data to be concealed (S7). ). If the amount of concealable data of the original image is greater than or equal to the amount of message data to be concealed, a message data concealment process is then performed (S9), and then an image having concealed message data is output (S10).

As shown in FIG. 2, a detailed message data concealment process after step S7 is performed. If the amount of concealable data of the original image is greater than or equal to the amount of message data to be concealed, the process of finding the number of bits to be concealed proceeds. In operation S20, a process of searching for a location to be concealed is performed (S22), and a bit of the position to be concealed is initialized through the concealment bit initializer 20 (S24). Then, one bit is concealed through the message data bit concealer 22 at the initialized concealed position (S26). Thereafter, it is determined whether or not the concealment process is finished through the extra bit calculator 24 (S28). If it corresponds to the end of the concealment process, the concealment process is terminated, and accordingly, an image in which the message data is concealed is output (S10).

Meanwhile, FIG. 5 is a diagram illustrating a message data extraction process in a message data extractor according to the present invention in detail. FIG. 6 is a schematic block diagram of a data concealment and extraction system using J & D and CSF according to the present invention. 7 is a detailed block diagram of a message data extractor for performing a message data extraction process in FIG. 6.

As shown in the drawing, first, an image in which message data is concealed is stored in a memory (not shown) of the device. Thereafter, when a signal for requesting extraction of the message data of the image where the message data is hidden from the user is input (S30), the hidden coefficient generator 2 (see FIG. 3) reads the image where the message data stored in the memory is hidden. Comes (S32). Then, the hidden coefficient generator 2 generates a hidden coefficient of the image in which the message data is hidden (S34). Then, the concealed order generator 4 is operated to generate a concealed order from the image in which the message data is concealed. The concealment coefficient and concealment order for the image in which the message data thus generated are concealed are stored in a memory.

Thereafter, the message data amount calculator 32 of the message data extractor 30 calculates the message data amount, and then extracts the message data by one bit from the message data bit extractor 34 (S40). Thereafter, it is determined whether or not the message data extraction process is finished through the extra bit calculator 36 (S42). If the message data extraction process is finished, the message data is output (S44).

In other words, in the present invention, the message data is concealed using the concealable data amount and the concealment order generated by the concealment coefficient generator 2 and the concealment order generator 4. As a result of experiments using the coefficients from the hidden coefficient generator (4), this method was developed considering only the contrast of the image.

First, this fact should be taken into account when concealing the message data because the G component is more sensitive to change than the B component when the concealment coefficients are the same.

Second, when message data is concealed in the R and B components, in consideration of the relation with the G component, when the R and B components are larger than the G component, the number of bits smaller than the calculated concealment coefficient is allocated.

Third, in the case of a high frequency band having a large variance value, since the sensitivity to change is low, an allocation of more message data than a hidden coefficient is considered considering that the invisibility is not greatly reduced.

Fourth, since a low frequency band having a small variance value has high sensitivity to change, much attention needs to be paid because of low visibility.

The designed data concealment system roughly performs the following process and will be described with reference to FIG. 8.

First, the concealment coefficient is obtained from the original image 100 through the concealment coefficient generator 2. The concealment coefficient is obtained by using J & D for each R, G, and B component in the still image. Arbitrary coefficient values 120 of the original image 100 in which the concealment coefficient is obtained are shown as shown.

In addition, as a process of determining the concealment order from the original image 100 through the concealment order generator 4, for each point of the G component among the R, G, and B components, 3x3, 5x5 block dispersion or differences from surrounding pixels Determine the hiding order by The random order values 122 of the original image 100 in which the concealed order thus obtained is displayed are as shown.

Arbitrary order values 122 of the original image 100 in which the concealment order is displayed indicate a concealment order generated through a concealment sequence generation process as shown in FIG. 9. That is, FIG. 9 is a diagram showing an example of concealment order generation, in which pixel values of 0 and 255 are displayed in arbitrary pixel values 400. For convenience, the results are shown for ease. Arbitrary mean values 410 are shown in which the mean value of the difference value between the surrounding pixels is displayed for the arbitrary pixel values 400. For example, the average of the difference value between pixels between the position A at the arbitrary pixel values 400 and the pixels surrounding it is indicated at (A) at the arbitrary average values 410. This value can be used by obtaining a variance value as necessary. In addition, the random order values 420 represent values in which the hidden order is indicated. The hidden order values 1 to 16 are determined by the average value between the surrounding pixels, and 17 to 49 are uniform random because the average value of the surrounding pixel values is all zero. A number is generated (uniform random number generator not shown) to determine the concealment order.

8, the size of the binarized message data 210 of the message data 200 to be concealed in the original image 100 is the amount of concealable amount of the original image 100 indicated by the concealment coefficient in the comparator 6. That is, the sum of the hidden coefficients). The message data concealer 8 then conceals according to the concealment order as indicated by the concealment order display 220 for each concealment coefficient according to the concealment order generator 4. In other words, the message concealment starts from the position where the first 4 appears. A detailed process will be described later with reference to FIG. 10.

That is, as the message data concealment process, concealment of the message data by one bit in the LS ratio of the points whose concealment coefficient values of the R, G, and B components is 4 according to the concealment order, and the R, G, and B components according to the concealment order The message data is concealed by one bit into the upper and LS ratios of each of the points whose concealment coefficient values are 4 and 3, and the concealment values of the R, G, and B components are 4 and 3 according to the concealment order. Points that are higher bits and 2 than the hidden positions of the previous process conceal message data by 1 bit in LSB, and the hidden coefficient values of R, G, and B components are 4, 3, 2 according to the concealment order. The message data is concealed by one bit in the higher bits than the concealed positions of the previous procedure. The black part in the figure is the message data concealment part. At the bottom, a form in which message data is concealed is presented. In other words,

In this manner, 4, 3, and 2 bits of message data are hidden in each pixel according to the hidden coefficient. The extraction process differs from finding the location to be concealed and extracting the message data at that location in the same way as the concealment process. The basic concealment process is shown above. The message data extraction process will be described with reference to FIG.

First, as in the case of message data concealment through the image 300 where the message data is concealed, a concealment coefficient is generated through the concealment coefficient generator 2. Arbitrary coefficient values 120 are shown which show the hidden coefficients generated through this process. Also, the concealment order generator 4 generates the concealment order. Arbitrary order values 122 are shown indicative of the concealment order generated by this process.

Thereafter, the message data extractor 30 displays the x-coordinate to be extracted, the y-coordinate to be extracted, and the number of bits to be extracted through the random order values 122 having the hidden order and the random coefficient values 120 having the hidden coefficient. As shown in 510, the message data is extracted and stored in a buffer (memory) (not shown), and then the extracted message data 200 is output.

Each step of the concealment system adaptable to color image is described in detail as follows.

First, the concealment coefficient generation will be described. The concealment coefficient generation uses Equation (4). {K_2} ^ {'}, {k_3} ^ {'} apply 88 and 192, respectively. This is expressed in binary, which is 10110000, 11000000. This is to prevent the hidden coefficient value from being changed by the message data concealed in the still image during the detection process after concealing the message data. Since the lower 4 bits of all the above binary numbers are '0', the coefficient value does not change even after concealment for the concealment coefficients 4, 3, and 2 proposed in the present invention. The concealment coefficients 4, 3, and 2 are respectively set to 2 when the above-mentioned {k_2} ^ {'}, {k_3} ^ {'} values of 192 or more are 4, 88 or more, and 3 or 88 or less.

Second, the calculation of the variance value will be described. In order to generate the concealment order, the variance as a reference is obtained. Since the low frequency band and the high frequency band can be distinguished by the variance value, the message data can be concealed according to the characteristics of CSF. The variance also uses only the top four bits in the calculation rather than all the bits of the G component. This is also to prevent the variance value from being modified by the hidden message data when seeking the variance in the detection process after the message data concealment. As the still image increases, the time for obtaining the variance value also becomes very large. Therefore, it is possible to use a difference from the surrounding pixels to reduce the time.

Third, as to the generation of the concealment order, the variance value is used to preferentially conceal the message data in the high frequency band. This is to protect the invisibility of the still image to the maximum even in hiding many message data. The variance values of all G components are sorted in descending order, and uniform random numbers are applied to the same variance values to ensure randomness of message data concealment. The sorting method used to determine the concealment order may be time consuming or memory intensive. Both of these methods cause problems when concealing message data in large still images such as 5000x5000. In the case of using a large still image or using a low-memory computer with low memory, the method of finding and sorting the variance value can be performed in units of blocks. The still image is divided into 2x2 or 3x3, and each block has one dispersion value for neighboring pixels, and is aligned using the dispersion value corresponding to each block.

Fourth, the message data concealment will be described. As shown in Fig. 10, the message data concealment is performed by using the result from the concealment order generator 2. If the concealment coefficient is 4 among the R, G, and B components of all the pixels, the message data is concealed by 1 bit in the LSBI of the component. If there are no R, G, or B components with a hidden coefficient of 4 anymore, the pixel with 4 or 3 R, G, and B components is found. At this time, if the concealment coefficient is 4, the message data is concealed 1 bit in the LSB upper bits of the corresponding component, and if 3, the message data is concealed 1 bit. If there are no more spaces for concealing the message data in the components having the concealment coefficients 4 and 3, the message data is concealed from the beginning again according to the concealment order. At this time, if the concealment coefficient is 4, the second higher order bits of the LSB, 3, the LS non-order upper bits, and 2 if the concealment is hidden by one bit. If the message data remains after hiding until this process, the message data is added to the 3rd high order bit of LSB if the concealment coefficient is 4, the 2nd high order bit of LSB if 3, and the 2nd high order bit of LSB according to the concealment order from the beginning. Conceal one bit In this manner, the component having the hidden coefficient of 4 conceals 4 bits, the component having 3 has 3 bits, and the component having 2 has 2 bits.

Sixth, if message data concealment with adaptability of R, G, and B components is described, two cases of invisibility become low. This is the case of concealing a lot of message data in the low frequency band and concealing without considering the value of the G component. In the latter case, even if the R and B components are constant, invisibility differences occur according to the G components. In consideration of this point, G component should be considered when concealing the message data according to the concealment coefficient. When the message data is concealed, the pixel values of the R and B components are compared with the pixel values of the G component. If the values of the R and B components are larger than the G components, fewer bits than the hidden coefficients can be allocated to prevent the invisibility from being lowered.

Seventh, the message data concealment with statistical adaptability is explained. In the two cases described in the sixth description, the former determines the concealment order, so that the high-priority pixel in the concealment sequence when concealing the message data has a high frequency band and has priority. Low ranking pixels are in the low frequency band. In this case, when the hidden message data is equal to the amount of concealment of the still image, the same amount of message data is concealed for pixels having the same hidden coefficient regardless of the low frequency band and the high frequency band. For this reason, the invisibility must be prevented from being lowered by the hidden message data in the low frequency band. In the case of high frequency band, invisibility is not greatly lowered even if it is concealed more than 4 bits. Therefore, a hidden bit is allocated to the pixels of the high priority part in the hidden order, and the number of hidden bits is assigned to the low priority part. Allocate less.

As described above, according to the data concealment system and method according to the present invention, it is possible to conceal more message data than the method of using LSB, which is a message data concealment method that has been used previously.

In the existing methods of LSB method, the maximum size of the message data that can be hidden is up to 12.5% of the still image to be concealed, but the method proposed in the present invention can conceal up to 50% and an average of 30% of the message data can be concealed. There is.

In addition, in the case of concealing 12.5% of the message data of the still image, the method using LSB conceals one bit for each R, G, and B of all pixels, whereas the message data concealment method according to the present invention is statistically adaptive. Since the first concealment of pixels with high concealment coefficients and concealment in the high frequency region by the concealment order, it is possible to obtain better invisibility than the LSB concealment method.

1 is a schematic flowchart of a data hiding system according to the present invention.

FIG. 2 is a flowchart specifically illustrating a message data concealment process in FIG. 1.

3 is a schematic block diagram of a data hiding system according to the present invention.

FIG. 4 is a detailed block diagram of a message concealer for performing a message data concealment process in FIG. 3.

5 is a schematic flowchart of a data extraction system according to the present invention.

6 is a schematic block diagram of a data extraction system according to the present invention.

FIG. 7 is a detailed block diagram illustrating a message extractor performing the message data extraction process of FIG. 6.

8 is a diagram illustrating an example of a message data concealment process in the data concealment system of the present invention.

FIG. 9 is a diagram illustrating an example for explaining a process of generating a message data concealment order in FIG. 8.

FIG. 10 is a diagram illustrating an example for explaining a message data concealment procedure in FIG. 8.

11 is a view showing an example of the message data extraction process in the data extraction system of the present invention.

12 is a graph illustrating CSF (Constrast Sensitivity Function).

Explanation of symbols on the main parts of the drawings

2: concealment coefficient generator 4: concealed sequence generator

6: Comparator 8: Message data concealer

20: hidden bit initializer 22: message data bit concealer

24, 36: extra bit calculator 30: message data extractor

32: message data amount calculator 34: message data bit extractor

100: original image 200: message data

300: Image where message data is hidden

Claims (20)

In the data hiding system for hiding message data in the original video, A hidden coefficient generator for generating a hidden coefficient from an original image; A hidden order generator for generating a hidden order from an original image; A comparator for comparing the amount of concealable amount calculated by the concealment coefficient of the original image with the amount of message data to be concealed; And A message data concealer for concealing the message data based on the message data concealment order determined according to the concealment order corresponding to the concealment coefficient if the amount of data concealable in the original image is greater than or equal to the amount of message data to be concealed by the comparator; Data hiding and extraction system using J and D and CS characterized in that consisting of. The data concealment and extraction system according to claim 1, wherein the comparison is terminated when the amount of data concealable in the original image is smaller than the amount of message data to be concealed. The method of claim 1, wherein the message data concealer is: A hidden bit initializer for finding the number of bits to be concealed if the amount of data concealable in the original image is greater than or equal to the amount of message data to be concealed, and then finding a position to be concealed and initializing bits of the position to be concealed; A message data bit concealer for concealing one bit in the initialized concealed position; And An extra bit calculator for determining whether to end the concealment process; wherein the concealment process is terminated by the extra bit calculator to terminate the concealment process. . In the data concealment system for extracting the message data from the original image where the message data is hidden, A hidden coefficient generator for generating a hidden coefficient from an image in which message data is hidden; A concealed order generator for generating a concealed order as an image in which message data is concealed; And Data using J & D and CS, comprising: a message data extractor configured to extract message data by reading coordinate values and bits to be extracted from the concealment coefficient generator and the concealment sequence generator; Concealment and extraction system. The method of claim 4, wherein the message data extractor is: A message data amount calculator for calculating a message data amount from data in which the message data is concealed; A message data bit extractor for extracting message data by reading coordinate values and the number of bits to be extracted from the concealment order and concealment coefficient values; And An extra bit calculator for determining whether to end the message data extraction process; and a data hiding and extracting system using J & D and CSF. The data concealment and extraction system using J & D and CS according to claim 1 or 4, wherein the concealment order is concealed from the high frequency band by using the characteristics of the CS to improve the invisibility. The data using J & D and CS according to claim 1 or 4, wherein when the hidden order is obtained, the R, G, and B components are not determined to be dispersed, but the hidden order is calculated using only the G component. Concealment and extraction system. 8. The data concealment and extraction system using J & D and CS according to claim 7, wherein a block-by-block distribution is obtained by blocking an image to reduce the time required and use less memory. 8. The data concealment and extraction system using J & D and CSF of claim 7, wherein the image is blocked to reduce the time required and blocks the image to block the memory. The method of claim 1, wherein the message data concealment order is concealed sequentially from a high frequency band having a high concealment order among pixels having a concealment coefficient of 4 in consideration of the J & D value when concealing the message data in order to improve invisibility. If message data is concealed at all 4 pixels and there is remaining message data to be concealed, then conceal it according to the concealment order precomputed to the pixels with concealment coefficients 4 and 3, and then in order of concealment coefficients 4, 3 and 2. Data concealment and extraction system using J & D and CS characterized in that the concealment.  The method according to claim 10, wherein when the message data is concealed, the message data is not concealed at a time by the number of bits corresponding to the concealment coefficient. And concealing one bit once again in order to conceal the data.  The method of claim 1, wherein the invisibility of the color image is increased by applying the characteristics of the color image. In the concealment process, the relation between the R, B component and the G component is obtained. A system for concealing and extracting data using J & D and CSF, which is pixel-adaptive to allocate and conceal fewer bits than the concealment coefficient of C. The method according to claim 1, wherein when concealing the same amount of message data, more bits than the concealment coefficient are added to pixels having higher priority according to the concealment order in order to prevent the low frequency component from becoming much less invisible than the high frequency component. A system for concealing and extracting data using J & D and CS, which allocates bits having a lower priority than the hidden coefficients. The method of claim 1, wherein in order to prevent the message data from being sequentially hidden, random properties of the hidden order are added by applying a uniform random number to the same variance value when obtaining the hidden order. Data concealment and extraction system using J & D and CSF. Generate a hidden coefficient from the original image; Generate a concealment order from the original image; Comparing the amount of message data to be concealed with the amount of concealable data of the original image to determine whether the amount of concealable data of the original image is greater than or equal to the amount of message data to be concealed; If the amount of concealable data of the original image is greater than or equal to the amount of data of the message to be concealed, the subsequent process of concealing the message data is performed based on the message data concealment sequence according to the concealment sequence corresponding to the concealment coefficient. Data concealment method. 16. The method of claim 15, wherein the message data hiding process is If the amount of concealable data of the original image is greater than or equal to the amount of message data to be concealed, find the number of bits to be concealed, and then find a position to conceal; Initialize a bit of a location to conceal; Conceal one bit in the initialized concealment position; And then determining whether the concealment process is terminated, and if it corresponds to the end of the concealment process, terminating the concealment process. Reading an image in which message data is concealed; Generating a concealment coefficient of an image in which the message data is concealed; Generate a hiding sequence of an image in which the message data is hidden; And extracting message data by reading coordinate values and the number of bits to be extracted from the generated concealment order and concealment coefficient values. 12. The JD and CS of claim 10, wherein when the message data is concealed and extracted, when the J & D value is taken into account, more bits are allocated when the bright pixel has a higher J & D value than the dark pixel. Data concealment and extraction system. 19. The method of claim 10 or 18, wherein in consideration of the J & D value during message data concealment and extraction, a concealment coefficient is assigned to 4, 3, or 2 according to the brightness of each of the R, G, and B components of the image. Data hiding and extraction system using J & D and CS. 7. The data concealment and extraction system using J & D and CS according to claim 6, wherein dispersion and difference with neighboring pixels are used to conceal from the high frequency band using the characteristics of the CSF when the message data is concealed and extracted. .