JP2005229508A - Electronic watermarking system and method, and data structure of an electronic watermark detection evaluation table used for this system - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To shorten a processing time in electronic watermark information detection. <P>SOLUTION: After electronic watermark information is embedded in an input image, that electronic watermark information is detected block by block, and detection evaluation data indicating a level of easiness of reading are generated for each block. Furthermore, the generated detection evaluation data are stored and compared with a preset threshold, and electronic watermark information is detected in an order starting from a block having the highest level of the detection evaluation data among extracted blocks. <P>COPYRIGHT: (C)2005,JPO&NCIPI

Description

  The present invention relates to a digital watermark embedding system and method that divides an image into blocks of an arbitrary size in a grid pattern and embeds digital watermark information in units of the blocks, and a data structure of a digital watermark detection evaluation table used in the system .

There is an “electronic watermark” technology that can conceal other information in digital content such as images and sounds.
In other words, “digital watermark” is a technique for embedding other information in digital content such as image data and audio data, and hiding it in an invisible or inaudible state. As a feature, other information (digital watermark information) can be embedded in the content, and the digital watermark information cannot be erased even if the content is processed (compressed, clipped, enlarged, or reduced). In addition, it is difficult for a third party to remove the digital watermark information while maintaining the quality of the content. For example, if the content containing the digital watermark information is illegally released on the Web (World Wide Web), the patrol system Detected and warned.

On the other hand, there is a print-compatible digital watermark technology that applies the above-described “digital watermark” technology to printed matter. FIGS. 11 (a) and 11 (b) show a printed matter production method with electronic watermark and its detection method, respectively.
That is, the image information processed in the digital plate making process is changed in such a way that it cannot be recognized by the naked eye when it becomes a printed matter, and identification information for digital watermark is embedded. There is almost no effect on the finish of the printed matter such as the pattern and color, and the human eye cannot recognize the difference. When this printed matter is digitized again using a scanner or the like and input to dedicated software, the embedded digital watermark information can be detected (see, for example, Patent Document 1 and Patent Document 2).
JP 2002-64700 A JP 2003-14388 A

  By the way, in the above-described digital watermark, as shown in FIG. 12, the image is divided into blocks of the same shape (here, blocks 1 to 18), and appropriate digital watermark information is embedded therein. When detecting digital watermark information from an image in which the digital watermark information is embedded, it is not possible to identify and detect a position where the digital watermark is easy to detect. Therefore, it takes a considerable time to detect all blocks.

  The present invention has been made in view of the above circumstances, estimates a block from which digital watermark information can be detected from a digital watermark-embedded image, detects digital watermark information from the estimated block, and performs processing in digital watermark information detection An object of the present invention is to provide a digital watermark embedding system and method, and a data structure of a digital watermark detection evaluation table used in the system, with a reduction in time.

  In order to solve the above-described problems, the present invention is a digital watermark embedding system that divides an image into blocks of arbitrary size in a lattice shape and embeds digital watermark information in units of the blocks, and embeds with a specified embedding strength. Based on the electronic watermark information, an electronic watermark embedding means for embedding the electronic watermark information in the image, and after embedding the electronic watermark information in the image, the electronic watermark information is detected in units of blocks and can be easily read for each block. And a detection evaluation data generation unit that generates detection evaluation data indicating the level.

  Further, the present invention is characterized by comprising: digital watermark detection means for detecting the digital watermark information in the order of the blocks having the highest detection evaluation data level.

Further, in the present invention, a digital watermark detection means for detecting the digital watermark information in the order of blocks with a high level of the detection evaluation data in a block extracted by comparison with a preset threshold value,
It is provided with.

  In the present invention, if the digital watermark information cannot be read, the digital watermark detection unit embeds the digital watermark information by changing the embedding strength, or changes the threshold level and detects the digital watermark information again. It is characterized by performing.

  In order to solve the above-described problems, the present invention is an electronic watermark embedding system that divides an image into blocks of an arbitrary size on a lattice and embeds electronic watermark information in units of the blocks. An electronic watermark embedding unit for embedding in the image, noise is added to the image embedded with the electronic watermark information to degrade the image, and the electronic watermark information is detected in units of blocks for each intensity to which the noise is applied. And a detection evaluation data generating means for generating detection evaluation data indicating the level of readability.

  In order to solve the above-described problems, the present invention is a digital watermark embedding detection method that divides an image into blocks of arbitrary size in a grid pattern and embeds digital watermark information in units of the blocks, and includes a specified embedding strength and Embedding the electronic watermark information in an image based on the embedded electronic watermark information; and embedding the electronic watermark information in the image; then detecting the electronic watermark information in units of blocks; Generating the detection evaluation data indicating the level; storing the generated detection evaluation data; and comparing the digital watermark with the highest level of the detection evaluation data in the block extracted by comparison with a preset threshold value And a step of detecting information.

  In order to solve the above-described problems, the present invention divides an image into blocks of arbitrary size in a lattice shape, and data of a digital watermark detection evaluation table used in a digital watermark information embedding system that embeds digital watermark information in units of the blocks. Based on the specified embedding strength and embedding digital watermark information, the digital watermark information is sequentially embedded in the input image to be captured, and the digital watermark information is sequentially detected in block units according to the specified embedding strength. The level of readability is calculated for each block in which the digital watermark information is detected, and is written in the digital watermark evaluation table.

According to the present invention, after the digital watermark information is embedded in the input image, the digital watermark information is detected in units of blocks, and detection evaluation data indicating the level of readability for each block is generated. By referring to the detection evaluation data and estimating a block from which digital watermark information can be detected, and detecting the digital watermark information from the estimated block, the processing time for detecting the digital watermark information can be shortened.
Also, by detecting the digital watermark information in the order of the blocks with the highest level of detection evaluation data, the position where the digital watermark information embedded in the image is easy to read can be found, and the electronic watermark information can be detected efficiently from that position. In addition, digital watermark information can be detected.

Furthermore, when digital watermark information is detected, if the digital watermark information cannot be read, the digital watermark information can be embedded by changing the embedding strength, or by changing the threshold level and detecting the digital watermark information again. It can also be used to deal with the strength of prints or to determine the authenticity of printed matter.
When creating the detection evaluation table, in addition to the method of embedding by changing the strength of the digital watermark, the detection evaluation table should be created by changing the strength that gives noise such as blur to the image embedded with the digital watermark. Is also possible. The stronger the noise, the harder it is to remove the digital watermark than the method of changing the digital watermark strength.

FIG. 1 is a diagram quoted for explaining the principle of the digital watermark embedding method of the present invention.
As shown in FIG. 1, the digital watermark embedding device 1 inputs a digital image, gives the strength to embed the digital watermark, and information to be embedded, so that an image block of any size (here, 3 × 6 18 blocks) A digital image with a digital watermark divided into two is output. At this time, whether or not digital watermark information is easily detected for each block is evaluated to generate a detection evaluation table (TBL), which is registered in the digital watermark detection evaluation DB 6 prepared as a database. Here, the numerical value shown in the image block is level data generated by evaluating whether or not digital watermark information is easily detected. Details will be described later.

FIG. 2 is a diagram cited for explaining the principle of the method for detecting the digital watermark information embedded as described above.
As shown in FIG. 2, the digital watermark detection apparatus 2 acquires data of an image block detected from the digital watermark detection evaluation value DB 6 and detects digital watermark information from the image block. In the figure, the hatched block is an image block for detecting digital watermark information. Details will be described later.

FIG. 3 is a block diagram showing the functional expansion of the internal configuration of the digital watermark embedding system of the present invention.
The digital watermark embedding system according to the present invention includes a digital watermark embedding device 1, a digital watermark detection device 2, an input image DB3, a digital watermark embedded image DB4, a detection evaluation TBL generation unit 5, and a digital watermark detection evaluation value DB6. And the authenticity determination unit 7.

The digital watermark embedding device 1 functions as a digital watermark embedding unit that embeds digital watermark information in an input image fetched in block units from the input image DB 3 based on the designated embedding strength and embedded digital watermark information.
The detection evaluation data generation unit 5 detects the digital watermark information in units of blocks after the digital watermark information is embedded in the input image by the digital watermark embedding device 1, and sets the level of readability for each block. The detection evaluation data has a function of generating a detection evaluation TBL, and the detection evaluation data is stored in the digital watermark detection evaluation value DB 6.

On the other hand, the digital watermark detection apparatus 2 compares the detection evaluation data stored in the digital watermark detection evaluation value DB 6 with a preset threshold value, and the detection evaluation data in the image block extracted from the digital watermark embedded image DB 4. It functions as a digital watermark detection means for detecting digital watermark information in the order of higher level blocks.
If the digital watermark information cannot be read, the digital watermark detection device 2 also controls the digital watermark embedding device 1 to embed again by changing the embedding strength, or changes the threshold level and detects the digital watermark information again. It also has a function to perform.
The authenticity determination unit 7 determines whether the printed material is a regular printed material or an illegally copied material based on the digital watermark information detected by the digital watermark detection device 2.

FIG. 4 shows a detailed procedure for embedding digital watermark information by the digital watermark embedding apparatus 3.
In FIG. 4, reference numeral 34 denotes original image data. By performing Fourier transform (35) on the original image data 34, amplitude information 38 and phase information 39 can be obtained through block division and color space conversion. Here, the amplitude information 38 depends on the above-described embedding strength. Reference numeral 31 denotes digital watermark information embedded in the original image data 534. Spread digital watermark information is obtained by spectrum spreading (43) of the digital watermark information 31. Next, the spread digital watermark information is combined with the amplitude information 38 (45). Further, the inverse Fourier transform (44) is executed using the phase information 39, whereby image data obtained by synthesizing digital watermark information is obtained.

On the other hand, the original image data 34 is wavelet transformed (40) in order to reduce the noise by matching the noise generated by the synthesis of the digital watermark information with the local features of the image, and the original image data 34 is decomposed into subband images. Are represented by partial images orthogonal to each other. Based on this partial image, energy map data having the same size as that of the original image 34 is obtained (41: high-frequency energy calculation).
Next, a difference between the image obtained by combining the digital watermark information obtained by the inverse Fourier transform (44) and the original image data 34 is obtained (47). Further, the above-described energy map data weight mask creation (42) is performed, and the digital watermark information is again synthesized with the original image data 34 (48, 49). In this way, digital watermark embedded image data is obtained and stored in the digital watermark embedded image DB 4.

When detecting the watermark information embedded as described above, the watermark detection apparatus 2 detects the watermark information according to the following procedure. That is, the digital watermark embedded image data to be detected by the digital watermark is, for example, digital data obtained by scanning a printed matter with a scanner or the like. By performing discrete Fourier transform (not shown) on the digital watermark-embedded image data, amplitude information of the image data can be obtained.
On the other hand, a symbol value detection sequence is obtained based on the pseudo-random number sequence. Next, based on the amplitude information and the symbol value detection series, a process of calculating a symbol response value is performed to detect a symbol value. Radix conversion is performed on the symbol value thus obtained to obtain detected digital watermark information.

FIG. 5 is a diagram cited for explaining the generation process of the digital watermark information detection evaluation value by the detection evaluation TBL generation unit 5.
The detection / evaluation TBL generation unit 5 sequentially embeds digital watermark information under the control of the digital watermark embedding device 1 in the input image captured from the input image DB 3 based on the designated embedding strength and the embedded digital watermark information. Thereafter, the digital watermark information is detected in units of blocks according to the sequentially specified embedding strength (FIGS. 5a, b, and c), and the level of readability is calculated for each block in which the digital watermark information is detected. The evaluation value (digital watermark evaluation TBL) obtained as a result is registered in the digital watermark detection evaluation value DB 6 (FIG. 5d).

Specifically, as shown in FIG. 6, first, digital watermark information is embedded with an embedding strength “100”, and an image block in which this is detected is evaluated as “1”, and an image block that is not detected is evaluated as “0”. . Next, the digital watermark information is embedded with the embedding strength “100”, and the image block in which the digital watermark information is detected is evaluated as “1”, and the image block not detected is evaluated as “0”. Further, the digital watermark information is embedded with the embedding strength “300”, and the image block in which the digital watermark information is detected is evaluated as “1”, and the image block not detected is evaluated as “0”.
Finally, an image block in which digital watermark information is detected at each intensity described above is evaluated, and an average value for each image block is calculated to create a digital watermark evaluation TBL. The digital watermark evaluation TBL created here Is registered in the digital watermark detection evaluation value DB 6. Note that the level of readability may be a numerical value or symbol that can be easily read, for example, other than the average level of readability, for example, the total value of the embedding strength of the digital watermark, or You may make it apply the level decided by the intensity | strength.

  FIG. 8 is a flowchart showing the flow of embedding digital watermark information. As shown in FIG. 8, the digital watermark embedding apparatus 1 first takes an input image from the input image DB 3 (S81, S82), and embeds the digital watermark information according to the digital watermark information and the embedding strength set as parameters ( S83: Digital watermark embedding process). The digital image in which the digital watermark information is embedded is registered in the digital watermark embedded image DB 4 (S89, S85). The above is the conventional flow.

On the other hand, according to the present invention, the digital watermark embedding device 1 takes an input image from the input image DB 3 (S81, S82), and follows the digital watermark information and embedding strength range (for example, 100 to 300) set as parameters. The digital watermark information is embedded (S88: digital watermark embedding process). Next, digital watermark information is sequentially detected from the digital watermark-embedded image in accordance with the specified embedding strength (S89, S90), and the level of readability is evaluated. Here, the average value of the readability level of the digital watermark information detected in the range of the embedding strength of 100 to 300 is calculated to create the digital watermark detection evaluation TBL (S91).
The above operations are sequentially repeated within the parameter set range (S87 to S92). Here, the detection evaluation TBL sequentially generated when the specified range is exceeded is registered in the digital watermark information detection evaluation value DB 6 (S93, S94).

FIG. 7 is a diagram showing a flow of a method for detecting digital watermark information. Here, the digital watermark detection apparatus 2 acquires data of an image block to be detected from the digital watermark information detection evaluation value DB 6 and detects digital watermark information embedded in the image block.
FIG. 7A shows an image in which digital watermark information is embedded with an embedding strength of 100, and FIG. 7B shows a detection block selected at that time and detecting the digital watermark information in a hatched manner. Yes. As shown in FIG. 7B, an image block having a specific value (0.6: threshold value) or more is selected as a detection block from the detection evaluation TBL evaluated and generated as described above, and the embedded electron Watermark information is detected.

FIG. 9 is a flowchart showing the flow of digital watermark information detection. In FIG. 9, first, a target image is taken from the digital watermark embedded image DB 4 (S101, S102). Then, the detection evaluation TBL is acquired from the digital watermark information detection evaluation value DB 6 (S103), and an image block having an evaluation value equal to or greater than a specific value is selected (S104, S105). Here, an image block having an evaluation value of 0.6 or more is selected, and digital watermark information is detected from the image block having a high evaluation value (S106, S107).
Then, based on the detection result, authentication is performed by the authenticity determination unit 7 and whether or not the electronic watermark information at the time of printing product shipment is included in the printed material is checked (S108). If the digital watermark information is detected (S109) or if the digital watermark information is not detected from all the designated image blocks, the detection process is terminated (S110).

Here, the processing can be terminated when the digital watermark information cannot be obtained, but the threshold value specified in the selection of the detected image block is lowered by embedding the digital watermark information at the time of embedding. It is also possible to reset and perform re-detection.
On the other hand, in the case of copying using a regular printed material, there is a possibility that the electronic watermark information is detected from the copied product for an image block having a high detection evaluation TBL evaluation value. However, in an image block having an evaluation value near the middle, digital watermark information is detected from a regular printed material, and in a copied product, noise is added, making it difficult to detect the digital watermark information. This can be used to determine authenticity.

In addition, according to the above-described embodiment of the present invention, only the method of embedding by changing the strength of the digital watermark when creating the detection evaluation table has been described, but in addition, noise such as blurring is added to the image in which the digital watermark is embedded. It is also conceivable that detection is performed by changing the intensity to give a detection evaluation table. Noise is added to the image after embedding the digital watermark. The stronger the noise, the harder it is to remove the digital watermark.
Further, according to the above-described embodiment of the present invention, an image is divided into blocks of arbitrary size in a lattice shape, and digital watermark information is embedded in units of the divided blocks. For example, FIG. As shown in FIG. 5, the same applies to the case where the lattices overlap. FIG. 10A shows an example in which the grid is divided into blocks of any size without overlapping.

As described above, according to the present invention, after digital watermark information is embedded in an input image, the digital watermark information is detected in units of blocks, and detection evaluation data indicating the level of readability for each block is generated. Thus, by referring to the detection evaluation data, the block that can detect the digital watermark information is estimated, and the digital watermark information is detected from the estimated block, thereby shortening the processing time in the digital watermark information detection and improving the performance. It is a measure.
Also, by detecting the digital watermark information in the order of the blocks with the highest level of detection evaluation data, the position where the digital watermark information embedded in the image is easy to read can be found, and the electronic watermark information can be detected efficiently from that position. In addition, digital watermark information can be detected. Furthermore, when digital watermark information is detected, if the digital watermark information cannot be read, the digital watermark information is embedded at a different embedding strength, or the threshold level is changed and the digital watermark information is detected again, thereby being resistant to noise. It can also be used to deal with the strength of prints or to determine the authenticity of printed matter.

  When creating the detection evaluation table, in addition to the method of embedding by changing the strength of the digital watermark, detection is performed by changing the strength that gives noise such as blurring to the image embedded with the digital watermark, and the detection evaluation table is created. Also good. In this case, the stronger the noise intensity, the harder it is to remove the digital watermark than the method of changing the digital watermark intensity.

It is the figure quoted in order to demonstrate the principle of the electronic watermark embedding method of this invention. It is the figure quoted in order to demonstrate the principle of the method of detecting electronic watermark information. It is the block diagram which expanded the function and showed the internal structure of the digital watermark embedding system of this invention. It is a figure which shows the detailed procedure of digital watermark information embedding by a digital watermark embedding apparatus. It is the figure quoted in order to demonstrate the production | generation process of the digital watermark information detection evaluation value by a detection evaluation TBL production | generation part. It is the figure quoted in order to demonstrate the production | generation process of the digital watermark information detection evaluation value by a detection evaluation TBL production | generation part. It is a figure which shows the flow of the electronic watermark information detection method. It is a flowchart which shows the flow of an electronic watermark information embedding process. It is a flowchart which shows the flow of a digital watermark information detection process. It is a schematic diagram when an image is divided into blocks of an arbitrary size. It is a figure which shows the flow of the conventional digital watermark information embedding and detection method. It is the figure quoted in order to demonstrate the problem at the time of detecting the conventional digital watermark information.

Explanation of symbols

DESCRIPTION OF SYMBOLS 1 ... Digital watermark embedding apparatus, 2 ... Digital watermark detection apparatus, 3 ... Input image DB, 4 ... Digital watermark embedded image DB, 5 ... Detection evaluation TBL production | generation part, 6 ... Digital watermark detection evaluation value DB, 7 ... Authentication determination Part

Claims (7)

A digital watermark embedding system that divides an image into blocks of an arbitrary size in a grid pattern and embeds digital watermark information in units of the blocks,
A digital watermark embedding means for embedding the digital watermark information in an image based on the specified embedding strength and the embedded digital watermark information;
After the digital watermark information is embedded in the image, detection digital data is detected in units of blocks, and detection evaluation data generation means for generating detection evaluation data indicating the level of readability for each block;
An electronic watermark embedding system comprising: A digital watermark detection means for detecting the digital watermark information in order of the blocks of the detection evaluation data in a high level;
The electronic watermark embedding system according to claim 1, further comprising: A digital watermark detection means for detecting the digital watermark information in the order of the blocks with the highest level of the detection evaluation data in the block extracted by comparison with a preset threshold;
The electronic watermark embedding system according to claim 1, further comprising: The digital watermark detection means
If the digital watermark information could not be read, it is embedded by changing the embedding strength, or the threshold value is changed and the digital watermark information is detected again.
The digital watermark embedding system according to claim 2 or 3. An electronic watermark embedding system that divides an image into blocks of an arbitrary size on a lattice and embeds electronic watermark information in units of the blocks,
Digital watermark embedding means for embedding designated digital watermark information in the image;
Detection and evaluation for indicating the level of readability for each block by adding noise to the image in which the digital watermark information is embedded to degrade the image, detecting the digital watermark information in units of blocks for each intensity of applying the noise Detection evaluation data generation means for generating data;
An electronic watermark embedding system comprising: An electronic watermark embedding detection method that divides an image into blocks of arbitrary size in a lattice shape and embeds digital watermark information in units of the blocks,
Embedding the digital watermark information in an image based on the specified embedding strength and the digital watermark information to be embedded;
After embedding the digital watermark information in the image, detecting the digital watermark information in units of blocks, and generating detection evaluation data indicating a level of readability in the block;
Storing the generated detection evaluation data, and detecting the digital watermark information in the order of the blocks with the highest level of the detection evaluation data in the block extracted by comparison with a preset threshold value;
A method for embedding and detecting a digital watermark, comprising: A data structure of a digital watermark detection evaluation table used in a digital watermark information embedding system that divides an image into blocks of an arbitrary size in a grid and embeds digital watermark information in units of the blocks,
After embedding the digital watermark information in the image based on the specified embedding watermark information to be embedded, the digital watermark information is further detected in units of blocks according to the specified embedding strength, or detected by changing the intensity that gives noise. And a data structure of a digital watermark detection evaluation table in which a level of readability is calculated for each block in which the digital watermark information is detected and written to the digital watermark evaluation table.

Images