JP2005020775A - Method of judging electronic watermark information - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To judge embedded information highly precisely without increasing a strength of embedding information. <P>SOLUTION: Embedded electronic watermark information isjudged from image data by revising a data value of at least one piece of pixel data among a plurality of pieces of pixel data of the image data. First, a data value of at least one piece of pixel data is estimated by using at least two pieces of pixel data placed in the vicinity of the pixel data through interpolation. Then, the estimated data value of the pixel data is compared with an actual data value to detect the amount of revision applied to the pixel data. Then, the electronic watermark information is judged based on the amount of revision. <P>COPYRIGHT: (C)2005,JPO&NCIPI

Description

  The present invention relates to a digital watermark technique for digital data, particularly image data.

  In recent years, digital watermark technology has been attracting attention from the viewpoint of copyright protection for digital data such as image data. The digital watermark technique is a technique for embedding predetermined information in digital data according to a predetermined rule so that the predetermined information cannot be extracted from the digital data without using at least the predetermined rule. For example, in accordance with a predetermined rule, when information relating to a purchaser of image data is embedded in an invisible form in the image data itself and illegally copied, the illegal copy is made according to the predetermined rule. By extracting the embedded information from the data, the person who made the illegal copy (that is, the purchaser) is specified.

  FIG. 11 is a diagram for explaining the principle of information embedding / extraction processing for image data using a conventional digital watermark technique.

As shown in the figure, in the information embedding process, the pixel data at a predetermined position of the image data for the bits b _i (1 ≦ i ≦ 50) among the bits b _{1 to} b ₅₀ constituting the information to be embedded. Is increased by U if the bit is 1, and decreased by U if the bit is 0. The information is embedded in the image data by performing the processing for all the bits b _{1 to} b ₅₀ while changing the position where the processing is embedded.

On the other hand, there is a possibility that the image value in which the information is embedded in this way has been subjected to various image processing such as enlargement, reduction, compression, expansion, and copying, and the data value has been changed. Therefore, in the process of extracting the embedded information, the pixel at the predetermined position of the image data for the bit b _i (1 ≦ i ≦ 50) among the bits b _{1 to} b ₅₀ constituting the embedded information. The total luminance S of the data and the total luminance R (Reference Value) of the pixel data adjacent to the predetermined position are detected, and S−R> T (T is a threshold value, For example, if T ≧ U × (the number of pixel data in which b _i is embedded)), it is determined that b _i = 1, and if S−R <−T, b _i. If it is determined that = 0, and -T ≦ S−R ≦ T, it is determined that no information is embedded in the pixel data at the predetermined position. By performing this processing on all of the bits b _{1 to} b ₅₀ , information embedded in the image data is extracted.

  Regarding digital watermark extraction technology, Seiji Kobayashi, Koichi Kamijo, Shuichi Shimizu: “Data hiding using the properties of neighboring pixels-Statistical properties of neighboring pixels”, IPSJ 56th (1998) National Convention It is described in detail on pages 3-37-38 of the proceedings.

As described above, in the information extraction process using the conventional digital watermark technique, the sum S of the luminances of the pixel data considered to be embedded with the bits b _i to be extracted is adjacent to each of the plurality of pixel data. Whether the bit b _i is 1 or 0 is determined based on whether or not the sum R of the luminance of the pixel data has a statistically significant size. If it is not statistically significant, it is determined that information (bits b _i ) is not embedded.

  By the way, the image data may have a remarkable luminance difference between adjacent pixel data due to luminance unevenness (for example, contour portion) of the image itself. For this reason, even when information is not embedded, it is conceivable that the above S-R is larger than T or smaller than -T. In this case, although information is not originally embedded, it is erroneously detected as being embedded.

  On the other hand, even when information is embedded, it is conceivable that the SR is -T≤SR≤T. In this case, although information is originally embedded, it is erroneously detected as not embedded.

  In order to eliminate such erroneous detection, the information embedding strength U may be increased and the criterion (threshold) T used for information extraction may be increased. However, this increases the effect on image quality.

  When the embedding strength is expressed by an equation, U = Δ × √ (V / N). Here, V is an S-R variance, N is the number of pixel data for each bit, and Δ is a constant depending on the required detection accuracy. Specifically, when the S-R fluctuation (variance V) caused by the luminance unevenness of the image itself is increased by n times, the information embedding strength U is set to maintain the same false detection probability. It must be root n times. On the other hand, if the S-R fluctuation (variance V) is reduced, the same detection accuracy can be maintained even if the embedding amount U is suppressed.

  The present invention has been made in view of the above circumstances, and an object of the present invention is a digital watermark technique capable of accurately determining and extracting embedded information without causing an increase in information embedding strength. Is to provide.

  Another object of the present invention is to provide an information embedding technique capable of accurately determining and extracting embedded information while reducing the influence on image quality, specifically, suppressing deterioration in image quality. It is in.

  In order to solve the above-mentioned problem, the present invention is to determine digital watermark information, wherein digital watermark information to be embedded by changing a data value is determined for at least one element data among a plurality of element data constituting digital data. A first step of estimating a data value of the determination target element data using data values of at least two adjacent element data located in the vicinity of the element data to be determined; and the determination target A second step of detecting a change amount applied to the determination target element data by comparing the estimated data value with an actual data value for the element data, and a digital watermark based on the change amount And a third step of determining information.

  Here, the element data corresponds to pixel data when the digital data is image data, for example. In this case, the data value corresponds to, for example, luminance data.

  In the third step, for example, the digital watermark information may be determined by comparing the change amount detected in the second step with a predetermined threshold value.

  In recent years, there has been a remarkable improvement in techniques for interpolating digital data such as image data, that is, techniques for estimating missing element data from peripheral element data when element data constituting the digital data is missing. In the present invention, the data value of the at least one element data is estimated using at least two element data located in the vicinity of at least one element data considered to be embedded with information. Then, information embedding is determined by comparing the estimated data value as a reference value with an actual data value of the at least one element data. That is, it is determined whether or not information is embedded, and if it is embedded, the value is determined.

  By doing so, the estimated data value is substantially equal to the data value before the information of the at least one element data is embedded, so that it is possible to perform the determination more accurately than in the conventional technique. Become.

  As described above, according to the present invention, it is possible to accurately determine whether to embed information.

  It is also possible to prevent quality degradation due to information embedding.

  First, prior to specific description of a digital watermark information determination apparatus to which one embodiment of the present invention is applied, the principle of information embedding / determination processing in image data according to the present embodiment will be described.

  FIG. 1 is a diagram for explaining the principle of information embedding / determination processing in image data according to an embodiment of the present invention.

As shown in the drawing, in the information embedding process, the bits b _i (1 ≦ i ≦ 50) of the bits b _{1 to} b ₅₀ constituting the information to be embedded are stored in predetermined locations 1 to 100 of the image data. The luminance of 2 × 2 pixel data located at each position is changed so as to increase by U if the bit b _i is 1 and decrease by U if the bit b _i is 0. By performing this processing on all of the bits b _{1 to} b ₅₀ while changing the place to embed, the information of bits b _{1 to} b ₅₀ is embedded in the image data.

On the other hand, in the determination process of the information embedded in this way, among the bits b _{1 to} b ₅₀ , the bits b _i (1 ≦ i ≦ 50) are stored in the predetermined locations 1 to 100 of the image data. The luminance values of the 2 × 2 pixel data located at each position are read out, and the total sum of the read out luminance values (in this case, 2 × 2 × 100 = 400 pixels) is obtained. Next, for 2 × 2 pixel data respectively located in the predetermined locations 1 to 100 of the image data, at least two pieces of pixel data (hereinafter referred to as peripheral pixel data) located in the vicinity of the 2 × 2 pixel data. In other words, the luminance value of each 2 × 2 pixel data is estimated using an interpolation technique. Then, a total sum (in this case, 2 × 2 × 100 = 400 pixels) R of the estimated luminance values is obtained. Then, the difference between the sum S and the sum R is obtained, and S−R> T (where T varies depending on the required bit error rate, but, for example, T ≧ U × (the number of pixel data in which b _i is embedded, that is, 400 )), It is determined that b _i = 1. If S−R <−T, it is determined that b _i = 0, and if −T ≦ S−R ≦ T, the predetermined value is determined. It is determined that the information of the bits b _i is not embedded in the 2 × 2 pixel data respectively located at the locations 1 to 100. By performing this processing on all of b _{1 to} b ₅₀ , it is determined whether or not information of bits b _{1 to} b ₅₀ is embedded in the image data.

  Examples of the interpolation technique include the nearest neighbor method, the bi-linear method, and the bi-cubic method. The nearest neighbor method is a method that employs the luminance value of a pixel located closest to the interpolation target pixel. The bi-linear method is a method of determining a luminance value by linear approximation using four points around the interpolation target pixel. The bi-cubic method is a method for determining a luminance value by performing three-dimensional approximate interpolation using 16 points around an interpolation target pixel.

  When estimating the luminance value of each of the 2 × 2 pixel data, the direction in which the luminance variation is the smallest in the vertical and horizontal diagonal directions (number of directions is not limited) as shown in FIG. By selecting and performing interpolation along that direction, interpolation can be performed more accurately.

  Next, a digital watermark information determination apparatus that performs the above-described digital watermark information determination process will be described. FIG. 2 is a functional configuration diagram of a digital watermark information determination apparatus to which an embodiment of the present invention is applied.

  As illustrated, the digital watermark information determination apparatus according to the present embodiment includes a processing unit 501 and a storage unit 507.

  The processing unit 501 performs overall control of the image data to be determined for the digital watermark information, the input / output unit 502 responsible for input / output of the digital watermark information with respect to the image data, and each unit of the digital watermark information determination device. Unit 503, luminance value reading unit 504, luminance value estimating unit 505, and information extracting unit 506.

The storage unit 507 includes information insertion image holding unit 508 that holds image data for determining digital watermark information input via the input / output unit 502, and digital watermark information received via the input / output unit 502. For each of the bits b _{1 to} b ₅₀ to be performed, an information insertion position holding unit 509 that holds information for specifying a plurality of locations on the image data that is an embedding location of the bit, a read luminance holding unit 510, and an estimated luminance holding unit 511 and an extraction information holding unit 512.

The luminance reading unit 504 applies the bits b _{1 to} b ₅₀ constituting the digital watermark information to the image data held in the information insertion image holding unit 508 according to the information held in the information insertion position holding unit 509. For each bit, the luminance value of 2 × 2 pixel data located at a plurality of places where the bit is considered to be embedded is read out and summed (in this case, 2 × 2 × 100 = per bit) S), which is the sum of 400 pixels. The total luminance S obtained for each bit b _{1 to} b ₅₀ is read and held in the luminance holding unit 510.

The luminance value estimating unit 505 holds the image data held in the information insertion image holding unit 508 in the information insertion position holding unit 509 for each bit b _{1 to} b ₅₀ constituting the digital watermark information. The brightness values of at least two peripheral pixel data in the vicinity of 2 × 2 pixel data, which are respectively located at a plurality of places where the bit specified by the information that is specified is embedded, are read. Then, the luminance value of the corresponding 2 × 2 pixel data is estimated from the luminance values of the at least two neighboring pixel data using an interpolation technique, and the sum (in this case, 2 × 2 × 100 per bit). R = the sum of 400 pixels). Then, the estimated luminance holding unit 511 holds the total sum R obtained for each bit of the bits b _{1 to} b ₅₀ .

The information extraction unit 506 calculates the difference between the sum S held in the read luminance holding unit 510 and the total R held in the estimated luminance holding unit 511 for each bit b _{1 to} b ₅₀ constituting the digital watermark information. The bit value of the corresponding bit is determined by obtaining S−R and comparing the difference with a predetermined threshold T. Specifically, for bit b _i , if S−R> T, it is determined that b _i = 1, if S−R <−T, it is determined that b _i = 0, and −T ≦ S−. If R ≦ T, it is determined that the bits b _i are not embedded.

  In this embodiment, the number of bits constituting the digital watermark information is 50, the embedding location of each bit is a 2 × 2 pixel data area, and the position and number of each embedding area are known. Is assumed. Such information may be notified in advance by a person who embeds the digital watermark information.

  Next, a hardware configuration of the digital watermark information determination apparatus to which the present embodiment is applied will be described. FIG. 3 is a diagram illustrating an example of a hardware configuration of the digital watermark information determination apparatus illustrated in FIG.

  As shown in the figure, a digital watermark information determination apparatus according to the present embodiment includes a CPU 601, a memory 602, an external storage device 603 such as a hard disk device and other external storage devices 604, an input device 605 such as a keyboard, a display, and the like. The output device 606 and an interface 607 with an external storage device or input / output device can be constructed on an information processing device having a general configuration. Here, each unit of the processing unit 501 illustrated in FIG. 2 is realized as a process embodied on the information processing apparatus when the CPU 601 executes a program loaded on the memory 602. In this case, the memory 602 and the external storage devices 603 and 604 are used as the storage unit 507 shown in FIG.

  The above-described program for implementing the digital watermark information determination apparatus of the present embodiment on the information processing apparatus by being executed by the CPU 601 is stored in advance in the external storage device 603, and is stored in the memory 602 as necessary. It is loaded and executed by the CPU 601. Alternatively, it is loaded from the portable storage medium 608 onto the memory 602 as necessary via a portable storage medium 608, for example, an external storage device 604 that handles a CD-ROM, and executed by the CPU 601. Alternatively, after being installed from the portable storage medium 608 to the external storage device 603 via the external storage device 604, it is loaded from the external storage device 603 onto the memory 602 and executed by the CPU 601 as necessary. The

  Next, the operation of the digital watermark information determination apparatus to which this embodiment is applied will be described.

First, the control unit 503 causes the information insertion image holding unit 508 to hold image data for determining digital watermark information in cooperation with the input / output unit 502. Further, the control unit 503, for each of bits b _{1 to} b ₅₀ constituting the digital watermark information, is a place on the image data where the information of the bit is embedded (in the example shown in FIG. 1, 100 for each bit). It is assumed that information for specifying the location is held in the information insertion position holding unit 509 in advance.

FIG. 4 shows an example of a table in which information is stored to identify the embedding location of each of bits b _{1 to} b ₅₀ constituting the digital watermark information, which is held in the information insertion position holding unit 509. Here, for each bit b _{1 to} b ₅₀ constituting the digital watermark information, the embedding location 1 to 100 of the bit is specified by the X and Y coordinate values (see FIG. 1) of the pixel data located at that location. Like to do. In the present embodiment, as shown in FIG. 1, for each of the locations 1 to 100, 2 × 2 pixel data located at the location is an information embedding location. The pixel data specified by the Y coordinate may be set to any one of the 2 × 2 pixel data. Here, it is assumed that the pixel data located at the upper left of the 2 × 2 pixel data is set.

For each of bits b _{1 to} b ₅₀ , information for specifying the embedding location of the bit on the image data may be stored in advance in external storage devices 603 and 604 in FIG. 3, for example. The image data may be provided from a portable storage medium via the external storage device 604, or may be provided from a communication line via a communication interface (not shown). It may be.

Next, for each of bits b _{1 to} b ₅₀ constituting the digital watermark information, the digital watermark information determination apparatus obtains the sum S _{1 to} S ₅₀ of the actual luminance values of a plurality of pixel data at the bit embedding location. .

FIG. 5 shows a sum S of actual luminance values of a plurality of pieces of pixel data serving as information embedding locations of the bits b _{1 to} b ₅₀ constituting the digital watermark information in the digital watermark information determination apparatus shown in FIG. it is a flowchart for explaining the operation for obtaining the ₁ to S _50.

First, the control unit 503 sets _i for specifying the bit b _i to be subjected to luminance extraction to 1 (step S1001), and then determines the extracted luminance value of the pixel data that is the embedding location of the information of the bit b _i . The sum S _i is set to 0 (step S1002). Then, j (1 ≦ j ≦ 100) specifying the embedding location of the information of the bit b _i is set to 1 (step S1003).

Next, the luminance value reading unit 504 reads the coordinates (k, l) of the pixel data P (i, j) corresponding to the bit b _i and the location j from the information insertion position holding unit 509 (see FIG. 4), and information The luminance value I (k, l) of the pixel data P (i, j) at the coordinates (k, l) is read from the image data held in the inserted image holding unit 508, and this is embedded with the information of bit b _i. This is added to the sum S _i of the extracted luminance values of the pixel data at the location (step S1004). Next, the luminance value I (k + 1, l) of the pixel data at the position (k + 1, l) right next to P (i, j) is read out, and this is used as the information embedding location of the bit b _i. Is added to the sum S _i of the extracted luminance values of the pixel data (step S1005). Next, the luminance value I (k, l + 1) of the pixel data at the position (k, l + 1) immediately below P (i, j) is read out, and this is used as the information embedding location of the bits b _i. Is added to the sum S _i of the extracted luminance values of the existing pixel data (step S1006). Then, the luminance value I (k + 1, l + 1) of the pixel data at the lower right position (k + 1, l + 1) of P (i, j) is read out, and this is embedded in the information of bit b _i This is added to the sum S _i of the extracted luminance values of the pixel data at the place (step S1007). Through the processing in steps S1004 to S1007, the actual luminance values of all the 2 × 2 pixel data that are the embedding locations of the information of the bits b _i located at the location j are read, and the values are summed S _i. Will be added.

Next, the control unit 503 increments the value of j (step S1008), and repeats the processing of steps S1004 to S1008 until the value of j reaches 100 (step S1009). As a result, for the bit b _i , all the actual luminance values of the 2 × 2 pixel data located in the places 1 to 100 where the information of the bit is embedded are read, and the sum S _i is calculated. It will be done.

Next, the luminance value reading unit 504 stores the calculated sum S _i in the reading luminance holding unit 510 in association with the bit b _i . Thereafter, the control unit 503 increments the value of i (step S1010), and repeats the processing of steps S1002 to S1010 until the value of i reaches 50 (step S1011). As a result, for each of bits b _{1 to} b ₅₀ constituting the digital watermark information, as shown in FIG. 6, the total sum S ₁ to the actual luminance values of the plurality of pixel data that are the places where the information of the bits is embedded. S ₅₀ is required.

Next, for each of bits b _{1 to} b ₅₀ constituting the digital watermark information, the digital watermark information determination apparatus has at least two peripherals located in the vicinity of each of a plurality of pixel data that is an embedding place of the information of the bit The luminance value of the pixel data is estimated from the pixel data, and the sums R _{1 to} R ₅₀ are obtained.

FIG. 7 shows positions of the bits b _{1 to} b ₅₀ constituting the digital watermark information in the vicinity of each of a plurality of pixel data where the information of the bits is embedded in the digital watermark information determination apparatus shown in FIG. to estimate the luminance value of the pixel data from at least two peripheral pixel data is a flowchart for explaining the operation for obtaining the sum R ₁ ~R _50.

First, the control unit 503 sets _i for specifying the bit b _i to be subjected to luminance estimation to 1 (step S2001), and then determines the estimated luminance value of the pixel data that is the embedding location of the information of the bit b _i . the sum R _i is set to 0 (step S2002). Then, j (1 ≦ j ≦ 100) that specifies the embedding location of the information of the bits b _i is set to 1 (step S2003).

Next, the luminance value estimation unit 505 reads the coordinates (k, l) of the pixel data P (i, j) corresponding to the bit b _i and the location j from the information insertion position holding unit 509 (see FIG. 4). Luminance values of at least two peripheral pixel data located in the vicinity of 2 × 2 pixel data with the position at the upper left (for example, luminance values I (8 of pixel data at positions adjacent to the 2 × 2 pixel data) k, l-1), I (k + 1, l-1), I (k-1, l), I (k-1, l + 1), I (k, l + 2), I (k + 1, l + 2), I (k + 2, l), I (k + 2, l + 1)) are read out from the image data held in the information insertion image holding unit 508 (step S2004).

Next, the luminance value estimation unit 505 estimates the luminance value of the pixel data P (i, j) from the pixel position and luminance value of each of the at least two neighboring pixel data using an interpolation technique. For example, the luminance values of the surrounding pixel data at the top, bottom, left and right of the pixel data P (i, j), that is, I (k, l-1), I (k, l + 2), I (k-1, l) , I (k + 2, l) using the luminance values of the four pixel data, the luminance value I (k, l) of the pixel data P (i, j) at the coordinates (k, l) by the bilinear method. Is estimated. Then, the estimated luminance value is added to the sum R _i of the estimated luminance values of the pixel data that is the place where the information of the bits b _i is embedded (step S2005). Next, in the same manner, the luminance value I (k + 1, l) of the pixel data at the right position (k + 1, l) of P (i, j) is estimated, and this is _calculated as information on the bits b _i . Is added to the sum R _i of the estimated luminance values of the pixel data that is the embedding location (step S2006). Next, similarly, the luminance value I (k, l + 1) of the pixel data at the position (k, l + 1) immediately below P (i, j) is estimated, and this is used as the information of the bit b _i information. This is added to the sum R _i of the estimated luminance values of the pixel data that is the embedding location (step S2007). Similarly, the luminance value I (k + 1, l + 1) of the pixel data at the lower right position (k + 1, l + 1) of P (i, j) is estimated, and this is represented by bit b _This is added to the total sum R _i of the estimated luminance values of the pixel data that is the place where the information of _i is embedded (step S2008). Through the processing in steps S2005 to S2008, the luminance values of all the 2 × 2 pixel data that are the embedding locations of the information of the bits b _i located at the location j are estimated, and the values are added to the total R _i. That's right.

Next, the control unit 503 increments the value of j (step S2010), and repeats the processes of steps S2005 to S2009 until the value of j reaches 100 (step S2010). As a result, for the bit b _i , all the luminance values of 2 × 2 pixel data located in the places 1 to 100 where the information of the bit is embedded are estimated, and the sum R _i is calculated. become.

Next, the luminance value estimation unit 505 stores the calculated sum R _i in the estimated luminance holding unit 511 in association with the bits b _i . Thereafter, the control unit 503 increments the value of i (step S2011), and repeats the processes of steps S2002 to S2011 until the value of i reaches 50 (step S2012). As a result, for each of bits b _{1 to} b ₅₀ constituting the digital watermark information, as shown in FIG. 8, the sum total R _{1 to} R of the estimated luminance values of a plurality of pixel data that are the embedding places of the information of the bits. ₅₀ is required.

Next, the electronic watermark information determination unit, a sum S ₁ to S ₅₀ extraction luminance value of the pixel data obtained by the flow shown in FIG. 5, the sum R of the estimated luminance value of the pixel data obtained by the flow shown in FIG. 7 _{1 to} R ₅₀ are compared, and the bit values of bits b _{1 to} b ₅₀ constituting the digital watermark information are determined.

Figure 9 compares the sum R ₁ to R ₅₀ of the estimated luminance value of the pixel data obtained by the flow shown in sum S ₁ to S ₅₀ and 7 of the extraction luminance value of the pixel data obtained by the flow shown in FIG. 5 FIG. 7 is a flowchart for explaining an operation when determining each bit value of bits b _{1 to} b ₅₀ constituting digital watermark information.

First, the control unit 503 sets _i, which identifies the bit b _i to be a bit value determination target, to 1 (step S3001). Then, the information extracting unit 506, the sum R _i of the sum S _i and the estimated luminance value of the actual luminance values corresponding to the bit b _i, reads read luminance holder 510 respectively from the estimated luminance holding unit 511 (step S3002 ).

Next, the information extraction unit 506 determines whether or not the difference S _i −R _i between the read S _i and R _i is larger than a predetermined threshold T (step S3003). If it is larger, it is determined that the bit value of bit b _i is 1, and the extracted information holding unit 512 holds the result (step S3004). On the other hand, if it is smaller, it is determined whether or not the difference S _i −R _i is smaller than −T (step S3005). If it is smaller, it is determined that the bit value of bit b _i is 0, and the result is held in the extraction information holding unit 512 (step S3006).

If NO in any of the above steps S3003 and 3005, that is, if −T ≦ S _i −R _i ≦ T, it is determined that the information of bit b _i is not stored, and the result is extracted information. It is held by the holding unit 512 (step S3007).

Next, the control unit 503 increments i by 1 (step S3008), and repeats the processes of steps S3002 to S3008 until i becomes 1 to 50 (step S3009). As a result, all the determination results of the bits b _{1 to} b ₅₀ constituting the digital watermark information are held in the extraction information holding unit 512. This result is output via the input / output unit 502 and output to an output device such as a display.

  Hereinabove, one embodiment of the present invention has been described.

  According to the present embodiment, the actual luminance value of the pixel data that is considered to be embedded with the digital watermark information and the interpolation technique of the image that has been remarkably improved in recent years, at least located in the vicinity of the pixel data. The brightness value of the pixel data that is approximately equal to the brightness value in the state where the information of the pixel data is not embedded, estimated from the two pixel data, is compared.

  This makes it possible to accurately determine information embedding without increasing the determination criterion (threshold value) T and the information embedding strength U as compared to the information determination process using the conventional digital watermark technology. Is possible.

  In addition, since it is not necessary to increase the information embedding strength U, it is possible to prevent image quality deterioration of image data.

  In addition, this invention is not limited to said embodiment, A various deformation | transformation is possible within the range of the summary.

For example, in the above embodiment, as shown in FIG. 1, for each of bits b _{1 to} b ₅₀ constituting the digital watermark information, the embedding location of information representing the bit value of the bit is determined on the predetermined image data. The case where the pixel data is 2 × 2 pixel data located at the respective locations 1 to 100 has been described as an example, but is not limited thereto. As shown in FIG. 10, the place may be one piece of pixel data that is located at places 1 to 100 on predetermined image data. In this case, steps S1005 to S1007 in FIG. 5 are not necessary. In FIG. 7, steps S2006 to S2008 are not necessary.

As a modification of the processing flow shown in FIG. 9, it may be determined whether −T ≦ S _i −R _i ≦ T is true or false in the previous step of steps S3003 and 3005. In this way, it is possible to first determine whether information is embedded.

  Needless to say, the number of bits constituting the digital watermark information, the size of the image data area in which each bit is embedded, and the position and number of the embedded area are not limited to the above embodiment.

  Further, it is assumed that the number of bits constituting the digital watermark information, the size of the image data area in which each bit is embedded, and the position and number of the embedded area are known, but the present invention is not limited to this. . By applying the processing flow shown in FIG. 9 or its modification to each pixel of the image data, it is possible to search for an information embedding location. You may make it identify the said information embedding location using other methods together.

  In the above-described embodiment, the bit information constituting the digital watermark information is described as being embedded in the image data by changing the luminance value of the pixel data. However, the information can be estimated by an interpolation technique. If there is, a change is made to a feature value other than the luminance value of the pixel data (for example, color difference, YCrCb value, RGB value, CMYK value, etc.), so that the bit information constituting the digital watermark information is embedded in the image data. Also good.

  In the above embodiment, the method using the bilinear method has been described as the interpolation method for estimating the luminance value of the pixel data at the pixel position P (i, j). However, the pixel position of each of at least two pieces of peripheral pixel data is described. If the technique is to estimate the luminance value of the pixel data at the pixel position P (i, j) from the luminance value, the present invention is not limited to the above embodiment. For example, a method using the nearest neighbor method or the bi-cubic method, a method using linear interpolation along the minimum dispersion direction, or the like can be considered.

  The linear interpolation in the minimum dispersion direction will be described using the processing flow shown in FIG.

  First, in a pixel to be interpolated, fluctuations (variances) of a plurality of luminance values in a predetermined region along a plurality of different directions are obtained (S4001). After obtaining the dispersion in the predetermined direction, the direction with the smallest luminance fluctuation (referred to as the minimum dispersion direction) is obtained (S4003). Next, linear interpolation is performed along the determined minimum dispersion direction (S4004).

A specific example is shown below. Let the luminance value of the pixel data P (i, j) at the coordinates (k, l) be I (k, l). Luminance values of peripheral pixel data on the top, bottom, left and right of P (i, j), that is, I (k, l-1), I (k, l + 2), I (k-1, l), I (k + 2, l) using a first vertical variance _{V v = | I (k,} l-1) -I (k, l + 2) | and horizontal variance V _{h =} | I (k -1, l) -I (k + 2, l) | Then, by comparing the _{V v} and _{V h,} obtaining the small direction of the dispersion. Linear interpolation is performed along the obtained direction. When V _v is small, I _R (k, l) = (2 × I (k, l−1) + I (k, l + 2)) / 3, and when V _h is small, I _R (k, l, Let l) = (2 × I (k−1, l) + I (k + 2, l)) / 3 be the estimated luminance value I _R (k, l) of the pixel data at the coordinates (k, l).

  Although the dispersion values in the vertical and horizontal directions are compared here, it is also possible to use the dispersion values obtained for more directions including the licking direction such as four directions and eight directions. In this example, the variance value is obtained from the two neighboring pixel data. However, it is also desirable to obtain a variance value from a larger number of neighboring pixel data by expanding the area used as the neighboring pixel data.

  In this way, the luminance value of the pixel data can be estimated more accurately when the interpolation value varies greatly depending on the orientation, such as when the information embedding location corresponds to the contour portion of the image.

  The present invention can be used as a method for determining the embedding strength of information because the accuracy of determination of embedded information can be increased without increasing the strength of embedding information.

For example, for the original image data in which no information is embedded in the image data, for the pixel data at each embedding location of the information b _i , the actual value sum S _i and the estimated value sum R _i are calculated from the above. What is necessary is just to obtain | require each difference using each Example, obtain | require difference S _i -R _{i, and} to determine the intensity U. By doing so, it is possible to prevent image quality deterioration due to information embedding while ensuring accuracy of determination of embedded information.

  Furthermore, in the above embodiment, image data has been described as an example. However, according to the present invention, the value of digital data in a place where digital watermark information is embedded is located around the digital data by an interpolation technique. As long as it can be accurately estimated from digital data, it can be widely applied not only to image data.

It is a figure for demonstrating the principle of the information embedding and determination process to the image data in one Embodiment of this invention. 1 is a functional configuration diagram of a digital watermark information determination apparatus to which an embodiment of the present invention is applied. It is a figure which shows an example of the hardware constitutions of the digital watermark information determination apparatus shown in FIG. Is held in the information insertion position holding unit 509 shown in FIG. 2 is a diagram illustrating an example of a table containing information for identifying the bit b ₁ ~b ₅₀ each embedding location constituting the digital watermark information. In the digital watermark information determination apparatus shown in FIG. 2, for each of bits b _{1 to} b ₅₀ constituting the digital watermark information, the sum S ₁ of the actual luminance values of a plurality of pixel data that is an embedding place of the information of the bit. it is a flowchart for explaining the operation for obtaining the to S _50. Is held in the read luminance holder 510 shown in FIG. 2, is a diagram illustrating a sum S ₁ to S ₅₀ of the actual luminance values of a plurality of pixel data that is the embedded location information of the bit b ₁ ~b ₅₀ . In the digital watermark information determination apparatus shown in FIG. 2, for each of bits b _{1 to} b ₅₀ constituting the digital watermark information, at least two bits located in the vicinity of each of a plurality of pieces of pixel data that are embedding locations of the information of the bits. estimating the luminance values of the pixel data from the peripheral pixel data is a flowchart for explaining the operation for obtaining the sum R ₁ ~R _50. It is held in estimating the luminance retaining portion 511 shown in FIG. 2 is a diagram showing the sum R ₁ to R ₅₀ of the estimated luminance values of a plurality of pixel data that is the embedded location information of the bit b ₁ ~b _50. By comparing the sum R ₁ to R ₅₀ of the estimated luminance value of the pixel data obtained by the flow shown in sum S ₁ to S ₅₀ and 7 of the extraction luminance value of the pixel data obtained by the flow shown in FIG. 5, the electronic it is a flowchart for explaining the operation when determining the bit b ₁ ~b ₅₀ each bit values constituting the watermark information. It is a figure for demonstrating the principle of the information embedding and determination process to the image data in the modification of one Embodiment of this invention. It is a figure for demonstrating the principle of the information embedding and determination process to the image data by the conventional digital watermark technique. It is a figure for demonstrating the principle of the information embedding and determination process to the image data in the modification of one Embodiment of this invention. It is a flowchart for demonstrating the operation | movement at the time of determining the direction which performs interpolation according to the principle shown in FIG.

Explanation of symbols

501: Processing unit, 502: Input / output unit, 503: Control unit, 504: Luminance value reading unit, 505: Luminance value estimation unit, 506: Information extraction unit, 507: Storage unit, 508: Information insertion image holding unit, 509 : Information insertion position holding unit, 510: read luminance holding unit, 511: estimated luminance holding unit, 512: extracted information holding unit, 601: CPU, 602: memory, 603, 604: external storage device, 605: input device, 606 : Output device, 607: interface, 608: storage medium.

Claims (7)

An electronic watermark information determination method for determining electronic watermark information embedded by changing a data value for at least one element data among a plurality of element data constituting digital data,
A first step of estimating a data value of the determination target element data using data values of at least two adjacent element data located in the vicinity of the element data to be determined;
A second step of detecting the amount of change applied to the determination target element data by comparing the estimated data value and the actual data value for the determination target element data;
And a third step of determining digital watermark information based on the change amount. A method for determining digital watermark information. The method of determining digital watermark information according to claim 1,
The first step uses the data values of the at least two neighboring element data of each of the plurality of determination target element data to estimate the data value of each of the determination target element data,
The second step is added to the plurality of determination target element data by comparing the total sum R of the plurality of estimated data values with the total sum S of actual data values of each of the determination target element data. Detect the total amount of change S-R,
In the third step, the digital watermark information is determined by comparing the sum S-R of the change amounts detected in the second step with a predetermined threshold T. Judgment method. The method of determining digital watermark information according to claim 1,
The first step uses a plurality of sets of data values of the at least two neighboring element data,
For each set, obtain a variance value of the data values of the at least two neighboring element data,
A method for determining digital watermark information, comprising: estimating a data value of determination target element data using the at least two neighboring element data included in a set having a minimum variance value. A method for determining digital watermark information according to any one of claims 1 to 3,
The digital data is image data;
The element data is pixel data. A method of determining digital watermark information. A method for determining digital watermark information according to any one of claims 1 to 4, comprising:
The method of determining digital watermark information, wherein the data value is luminance data of a pixel. A storage medium storing a digital watermark information determination program for determining digital watermark information to be embedded by changing a data value for at least one element data among a plurality of element data constituting digital data ,
The program stores an electronic watermark information determination program characterized by causing the information processing apparatus to execute each step of the electronic watermark information determination method according to claim 1, 2, 3, 4, or 5. Storage medium. An electronic watermark information determination apparatus for determining electronic watermark information embedded by changing a data value for at least one pixel data among a plurality of pixel data constituting image data,
Estimating means for estimating a data value of the determination target element data using data values of at least two adjacent element data located in the vicinity of the element data to be determined;
About the determination target element data, detecting means for detecting the amount of change added to the determination target element data by comparing the estimated data value and the actual data value;
An electronic watermark information determination apparatus comprising: determination means for determining electronic watermark information based on the change amount.

Images