JP2006270215A - Method for embedding electronic watermark, method for detecting electronic watermark, and method, device, and program for generating composite pattern configuration series - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To embed and detect an electronic watermark that suppresses a "ground glass effect" and dispenses with frame synchronization. <P>SOLUTION: A symbol component contained in the pattern of an electronic watermark to be embedded in each frame of an image is changed for each frame, and information obtained from each frame when the electronic watermark is detected is superimposed for detecting embedded information. <P>COPYRIGHT: (C)2007,JPO&INPIT

Description

  The present invention relates to a digital watermark embedding method, a digital watermark detection method, a synthetic pattern constituent sequence generation method, apparatus, and program, and more particularly to detection of a digital watermark in embedding and detection of a digital watermark for a video signal composed of a plurality of frame images. The present invention relates to a digital watermark embedding method, a digital watermark detection method, a synthetic pattern constituent sequence generation method, an apparatus, and a program capable of detecting a ground glass effect while maintaining accuracy and capable of being detected even when subjected to an attack such as frame thinning.

  Conventionally, by embedding digital watermarks in digital content, the copyright information of the digital content is protected, the copyright information related to the digital content is referred to, or the digital content is transmitted via an analog medium such as a printed matter as an advertisement. Services such as capturing information related to advertisements by taking pictures with digital cameras and reading digital watermarks have come to be realized.

  As a method for embedding a digital watermark, for example, when embedding a digital watermark, spectrum spreading is performed to change the real and imaginary components of the complex matrix independently based on the embedding target component position information, and independently of the input image. There is a method of generating an embedded image by generating a watermark pattern and adding the pattern to an actual image. When detecting a digital watermark, for example, based on detection target component position information, a detection target sequence is generated, offset information is extracted, and after correction to the detection target sequence, spectrum despreading is performed, and the extracted pixel block There is a method for detecting a digital watermark embedded in the document (for example, see Patent Document 1).

  Also in video signals, video signals are generally recorded as a series of frame images that are still images. Therefore, it is possible to embed digital watermarks by applying a digital watermarking method for still images. is there.

For example, a digital watermark for a video signal can be realized by embedding a common digital watermark in each frame image of the video signal using the above-described digital watermark method.
Japanese Patent Laid-Open No. 2003-219148 “Digital watermark embedding method, digital watermark embedding device, digital watermark detection device, and storage medium storing digital watermark embedding program”

  However, in a digital watermark embedding method in which the same fixed pattern is superimposed on each frame of a video as a digital watermark, if the strength of the digital watermark embedding increases, the embedded digital watermark pattern becomes independent of the motion of the video. It looked like it was sticking, so it was easy to perceive as if it were a frosted glass, and the watermark embedding strength could not be increased. Hereinafter, such a phenomenon is referred to as “frosted glass effect”.

  By changing the digital watermark pattern by shifting the position of each frame, the “frosted glass effect” can be prevented. However, in this method, attacks such as frame thinning and frame cutout are performed on the video. In order to detect the digital watermark from the modified video, it is detected how the pattern is changed in each frame or which frame is thinned out. It is necessary to synchronize with the pattern that should have been embedded, and there is a problem that such synchronization is difficult.

  When each frame image is regarded as a still image and a digital watermark is embedded, if the digital watermark detection is performed on each frame independently, there is no need to synchronize as described above. There is a problem that it is necessary to embed a digital watermark as strong as can be detected, and as a result, the image quality of the video signal is greatly deteriorated.

  The present invention has been made in view of the above points, and provides a digital watermark embedding method, a digital watermark detection method, a synthetic pattern constituent sequence generation method, an apparatus, and a program that suppress frame distortion while suppressing the “frosted glass effect” The purpose is to do.

  FIG. 1 is a diagram for explaining the principle of the present invention.

The present invention (Claim 1) is a digital watermark that superimposes a predetermined pattern configured based on embedded information on each frame image of the video signal so as not to be perceived by human perception. An electronic watermark embedding method in an embedding device, comprising:
An embedded information dividing step (step 1), which takes embedded information as an input and divides the embedded information into a plurality of symbols using embedded information dividing means;
A diffusion pattern generation step (step 2) for generating a diffusion pattern used for embedding for each symbol using the diffusion pattern generation means;
A diffusion pattern synthesis step (step 3) for selecting and synthesizing one or more diffusion patterns from the diffusion patterns for each symbol using the diffusion pattern synthesis means, and generating a plurality of resultant diffusion patterns.
A diffusion pattern switching step (step 4) for switching a plurality of synthesized diffusion patterns for each frame to be embedded using the diffusion pattern switching means;
A digital watermark embedding pattern superimposing step for embedding a digital watermark by superimposing a digital watermark embedding pattern image corresponding to the switched composite diffusion pattern on the frame image using a frame image as an input and embedding pattern superimposing means (step 5).

Further, according to the present invention (Claim 2), in the diffusion pattern synthesis step (Step 3) of the digital watermark embedding method according to Claim 1,
A combined diffusion pattern is generated so that the number of consecutive symbols, which is an index of the degree to which a common symbol is included in adjacent or close patterns in the combined diffusion pattern column, is minimized.

Further, according to the present invention (Claim 3), in the diffusion pattern synthesis step (Step 3) of the digital watermark embedding method according to Claim 1 or 2,
A combined pattern configuration sequence indicating which symbols are to be included in the generated combined diffusion pattern is stored in the pattern configuration sequence storage means.

According to a fourth aspect of the present invention, there is provided a composite pattern constituent sequence generation apparatus for generating a composite pattern constituent sequence that indicates which symbols are included in each of the composite diffusion patterns in the digital watermark embedding method according to the first to third aspects. A synthetic pattern composition series generation method,
An initial composite pattern constituent series determining step of acquiring a composite pattern constituent series from the storage means and determining a _first composite pattern constituent series r ₁ ;
An initial change target element selection step of selecting an arbitrary element from among the elements that are ON values of the _first composite pattern constituting series r ₁ ;
A vector obtained by inverting the ON value / OFF value of each element of the 2c-1 th (where c is a counter) composite pattern constituent series r _2c-1 is generated, and is defined as the 2 _c th composite pattern constituent series r _2c . A pattern inversion step;
A next change target element selection step of selecting an element number _ic that has not been selected so far among the elements that are the ON values of the 2c-th composite pattern constituent series r _2c ;
In the 2c-1th composite pattern constituent series r _2c-1 , the _ic-1 th element is changed to an OFF value, the _ic th element is changed to an ON value, and this is changed to a 2c + 1th composite pattern constituent series r _{2c + 1.} The next pattern generation step is performed.

  Further, according to the present invention (Claim 5), any one of the initial composite pattern constituent series determining step, the initial change target element selecting step, and the next change target element selecting step according to Claim 4 may randomly generate a composite pattern constituent series or Select an element.

According to a digital watermark embedding method of the present invention (Claim 6), in the diffusion pattern synthesis step of the digital watermark embedding method according to Claims 1 to 3,
Each step of the synthetic pattern composition series generation method according to claim 4 or 5 is performed.

According to the present invention (Claim 7), a predetermined pattern configured based on embedded information is preliminarily embedded in each frame image of a video so as not to be easily perceived by human perception, and embedded as a digital watermark. A digital watermark detection method in a digital watermark detection apparatus for detecting embedded information from a video signal placed in
A filtering step (step 11) that takes each frame image of the video signal in which the digital watermark is embedded as an input, filters the frame image using the filter means, and outputs the filtered image;
A block calculation step (step 12) for dividing the frame image subjected to the filter processing into blocks according to the embedded pattern size of the digital watermark using a block calculation unit;
A block addition step (step 13) of acquiring a plurality of block images respectively obtained in the block calculation step using a block addition unit for a plurality of frame images, and adding information for each of the plurality of block images. And do.

  According to the present invention (Claim 8), in the block addition step of the digital watermark detection method according to Claim 7, a plurality of block images themselves are used as information for each block image.

  According to the present invention (Claim 9), in the block addition step of the digital watermark detection method according to Claim 7, the detection obtained from a plurality of block images based on the detection target component position information as information for each block image Use the target series.

  FIG. 2 is a principle configuration diagram of the present invention.

According to the present invention (Claim 10), a digital watermark for superimposing a predetermined pattern configured on the basis of embedded information on each frame image of the video signal so as not to be perceived by human perception. An embedding device,
Embedded information dividing means 101 for dividing the inputted embedded information into a plurality of symbols;
A diffusion pattern generation means 102 for acquiring the divided symbols and generating a diffusion pattern used for embedding for each symbol;
A diffusion pattern synthesizing unit 103 that acquires a diffusion pattern for each symbol, selects one or more diffusion patterns, combines them, and generates a plurality of resultant diffusion patterns;
A diffusion pattern switching unit 104 that acquires a plurality of generated composite diffusion patterns and switches each frame to be embedded;
A digital watermark in which a frame image is input, a switched composite diffusion pattern is obtained, a digital watermark embedding pattern image corresponding to the switched composite diffusion pattern is superimposed on the frame image, and output to embed a digital watermark Embedded pattern superimposing means 107.

The present invention (invention 11) is characterized in that in the diffusion pattern synthesis means 103 of claim 10,
Means for generating a composite diffusion pattern so that the number of consecutive symbols, which is an index of the degree to which common symbols are included in adjacent or close patterns in the composite diffusion pattern column, is included.

Further, the present invention (claim 12) is the diffusion pattern synthesis means 103 according to claim 10 or 11,
Pattern configuration sequence storage means for storing a composite pattern configuration sequence that indicates which symbols are included in each of the generated composite diffusion patterns.

The present invention (Claim 13) is a composite pattern constituent series generation apparatus for generating a composite pattern constituent series representing which symbols each of the composite diffusion patterns includes in the digital watermark embedding apparatus according to any of claims 10 to 12. There,
Initial composite pattern constituent series determining means for acquiring a composite pattern constituent series from the storage means and determining a _first composite pattern constituent series r ₁ ;
An initial change target element selecting means for selecting an arbitrary element from elements that are ON values of the _first composite pattern constituting series r ₁ ;
A vector obtained by inverting the ON value / OFF value of each element of the 2c-1 th (where c is a counter) composite pattern constituent series r _2c-1 is generated, and is defined as the 2 _c th composite pattern constituent series r _2c . Pattern inversion means;
A next change target element selecting means for selecting an element number _ic that has not been selected so far among the elements that are the ON values of the 2c-th composite pattern constituent series r _2c ;
In the 2c-1th composite pattern constituent series r _2c-1 , the _ic-1 th element is changed to an OFF value, the _ic th element is changed to an ON value, and this is changed to a 2c + 1th composite pattern constituent series r _{2c + 1.} And next pattern generation means.

  According to the present invention (Claim 14), any one of the initial composite pattern constituent series determining means, the initial change target element selecting means, and the next change target element selecting means of the composite pattern constituent series generating device according to Claim 13 is randomly synthesized. Generate pattern composition series or select elements.

  According to the present invention (Claim 15), each of the spread pattern synthesizing means of the digital watermark embedding apparatus according to Claims 10 to 12 includes each means of the synthesized pattern constituent series generation device according to Claim 13 or 14.

The present invention (Claim 16) embeds a predetermined pattern configured based on embedded information on each frame image of the video in advance so as not to be perceived by human perception, and embeds it as a digital watermark in the video signal. An electronic watermark detection apparatus for detecting embedded information from a video signal placed in
Filtering means 401 for receiving each frame image of the video signal in which the digital watermark is embedded as input and performing filtering processing on each frame image;
A block calculation unit 402 that divides the frame image subjected to the filter processing into blocks corresponding to the embedded pattern size of the digital watermark;
Block addition means 404 for adding the acquired information for each of the plurality of block images to the plurality of frame images;

  According to the present invention (Claim 17), in the block addition means of the digital watermark detection apparatus according to Claim 16, a plurality of block images themselves are used as information for each block image.

  Further, according to the present invention (Claim 18), in the block adding means of the digital watermark detection apparatus according to Claim 16, the detection obtained from a plurality of block images based on the detection target component position information as information for each block image Use the target series.

According to the present invention (Claim 19), a digital watermark for superimposing a predetermined pattern configured on the basis of embedded information on each frame image of the video signal so as not to be perceived by human perception. An embedded program,
A program for causing a computer to execute processing for realizing the digital watermark embedding method according to claim 1.

The present invention (Claim 20) is a composite pattern constituent sequence generation program for generating a composite pattern constituent sequence that indicates which symbols each includes a composite diffusion pattern in the digital watermark embedding method,
A program for causing a computer to execute processing for realizing the synthetic pattern composition series generation method according to claim 4 or 5.

The present invention (Claim 21) embeds a predetermined pattern configured based on embedded information on each frame image of the video in advance so as to be hardly perceived by human perception and embedded as a digital watermark in the video signal. An electronic watermark detection program for detecting embedded information from a video signal placed in
A program for causing a computer to execute a method for realizing the digital watermark detection method according to claim 7.

  The present invention changes the symbol component included in the digital watermark pattern embedded in each frame for each frame, and simultaneously detects the embedded information after superimposing information obtained from each frame when detecting the digital watermark. .

  Furthermore, the “frosted glass effect” is suppressed by selecting the digital watermark pattern so as to minimize the component that the digital watermark pattern includes in common between frames.

  As described above, according to the digital watermark embedding method (apparatus / program) of claims 1, 10 and 19 of the present invention and the electronic watermark detection method (apparatus / program) of claims 7, 16, and 21, embedded images While suppressing the “frosting glass effect” by changing from frame to frame, even when a frame thinning attack is performed, digital watermark detection can be performed without the need for frame synchronization. Even if the embedding strength of is increased, it becomes difficult for humans to perceive, and the resistance of the digital watermark can be increased.

  Furthermore, according to the digital watermark detection method (apparatus / program) of claims 7, 8, 9, 16, 17, 18, and 21 of the present invention, a watermark embedding pattern while removing original image components having high correlation between frames. Can be enhanced, and the detection accuracy can be improved. As a result, the resistance to signal deformation and noise can be increased. Conversely, when the same level of tolerance is required, the digital watermark can be detected even if the digital watermark embedding strength is reduced. As a result, it is possible to suppress deterioration in image quality due to the digital watermark.

  Furthermore, according to the digital watermark embedding method (apparatus / program) of claims 2, 11 and 19 of the present invention, the “frosted glass effect” appearing in the watermark embedded video is minimized while maintaining detection accuracy as much as possible. Can do.

  In addition, according to the digital watermark embedding method (apparatus / program) according to claims 3, 12, and 10 of the present invention, it is possible to omit the calculation of a composite pattern constituent sequence that requires time when embedding a digital watermark, and to perform high-speed processing. Digital watermark embedding can be performed.

  Further, according to the synthetic pattern constituent sequence generation method (apparatus / program) of claims 4, 13, and 20 of the present invention, the composite spread pattern constituent sequence that minimizes the number of consecutive symbols can be calculated at high speed. Digital watermark embedding can be performed.

  Further, according to the composite pattern constituent sequence generation method (apparatus / program) of claims 5, 14, and 20 of the present invention, the composite pattern constituent sequence can be given randomness, and the resistance to the thinning-out attack can be increased. it can.

  Hereinafter, embodiments of the present invention will be described with reference to the drawings.

[First Embodiment]
In this embodiment, a digital watermark embedding device will be described.

  FIG. 3 shows a configuration example of the digital watermark embedding apparatus according to the first embodiment of the present invention.

  The digital watermark embedding device 100 includes an embedded information dividing unit 101, a diffusion pattern generating unit 102, a diffusion pattern combining unit 103, a diffusion pattern switching unit 104, an embedded pattern generating unit 105, a frame dividing unit 106, an embedded pattern superimposing unit 107, and a frame connection. Part 108.

  The digital watermark embedding apparatus 100 receives a pre-embedding signal 110 that is a video signal before embedding a digital watermark and embedding information 109 that is information to be embedded, and embeds the embedding information 100 in the pre-embedding signal 110. The embedded signal 111 obtained as a result is output.

  The digital watermark embedding process by the digital watermark embedding apparatus 100 is performed in the following procedure.

  FIG. 4 is a flowchart of the digital watermark embedding process according to the first embodiment of this invention.

Step 101) Embedded information division:
First, when the embedded information 109 is input to the embedded information dividing unit 101, the embedded information 109 is divided into a plurality of symbols S ₁ , S ₂ ,..., S _k and stored in storage means (not shown) such as a memory. To do.

  Here, the symbol represents information of a part of the embedded information, and is information that is an actual processing unit for embedding the digital watermark. For example, when the embedded information 109 is given as a 64-bit binary value As shown in FIG. 5, each 12-bit information may be a symbol by dividing the information into 12-bit information. Alternatively, 1-bit information may be represented by one symbol. If the length of the embedded information 109 (here 64 bits) is not divisible by the length of each symbol (12 bits) as in this example, a part of bits of some symbols (here the last symbol) are fixed. You may make it padding with a value (here value 0).

Step 102) Diffusion pattern generation:
In the spread pattern generation unit 102, each symbol obtained by the embedded information division unit 101 and stored in the storage means is acquired, the spectrum spread process is performed on each symbol, and the spread pattern P ₁ corresponding to each symbol is obtained. ,..., P _k are generated and stored in storage means (not shown) such as a memory.

  As a method of spread spectrum, for example, there are the following methods.

For example, when one symbol represents 1-bit information, a total of k PN sequences PN ₁ , each having a value of {1, −1}, for each symbol S ₁ ,..., S _k = (Pn ₁₁ , pn ₁₂ ,...), PN _k = (pn _k1 , pn _k2 ,...), And for the symbol i, if the symbol value represents bit 1, PN _i is -PN _i may be used as the diffusion pattern P _i when the value of represents a bit 0.

For example, when one symbol represents 12-bit information, 4096 PN sequences PN _(1,1) ,..., PN _(1,4096) , PN _{(2, 4096), ...,} prepared _{PN (k, 4096),} the symbol i, when the value of the symbol i represents an integer value x in _{12bit, PN (i, x)} and used as a diffusion pattern _{P i} It may be. For example, an M sequence may be used as the PN sequence. Such a diffusion pattern generation method is also described in Patent Document 1 described above.

  Here, a one-dimensional number sequence is generated as a diffusion pattern. However, a diffusion pattern may be generated as a two-dimensional image pattern based on the number sequence obtained by the same method. For example, a coefficient matrix obtained by increasing / decreasing a component value of a complex matrix based on a one-dimensional number sequence may be subjected to orthogonal transform such as discrete inverse Fourier transform, and the obtained image pattern may be used as a diffusion pattern.

Step 103) Diffusion pattern synthesis:
In diffusion pattern combining section 103, obtained by diffusion pattern generation unit 102, acquires the respective diffusion patterns stored in the storage means, the sum of n multiplexed spread pattern R _1, ..., to generate a R _n, memory, etc. Is stored in a storage means (not shown).

  Here, each combined diffusion pattern is generated by selecting one or more patterns from the k diffusion patterns and adding all the selected patterns.

For example,
R ₁ = P ₁ + P ₃ + P ₄
In this case, R ₁ is referred to as “includes” the symbols 1, 3, and 4, respectively.

  In the above, an example of a synthesis method in which a synthesized diffusion pattern is generated by adding the diffusion patterns has been shown. However, a synthesis that can detect that the synthesized symbol is included in the synthesized synthesis diffusion pattern. As long as it is a method, a method using another calculation may be used. For example, when the original diffusion pattern is represented by a one-dimensional number sequence, the number sequence may be divided into predetermined lengths, and the original diffusion pattern may be switched and connected for each section. An example is shown below.

P ₁ = {1, 2, 3, 4, 5, 6}
P ₂ = {7, 8, 9, 10, 11, 12}
P ₃ = {13, 14, 15, 16, 17, 18}
, By dividing the sequence into two sections of length and selecting each section from P ₁ , P ₂ , and P ₃ , respectively,
R ₁ = P ₁ + P ₂ + P ₃ = { _{1, 2,} 9, 10, 17, 18}
You may make it become.

  When the diffusion pattern is represented by a two-dimensional image pattern, the pattern may be divided into blocks and the diffusion pattern may be switched for each block. It goes without saying that these methods may be used in combination.

  Hereinafter, when the symbol “+” is used for the synthesis of the diffusion pattern, it indicates that other synthesis methods may be used in addition to the addition.

A number obtained by counting the number of symbols including symbols that are common to adjacent patterns (for example, R ₃ and R ₄ ) of the combined diffusion pattern with respect to the entire column of the combined diffusion pattern is referred to as “symbol consecutive number”. . At this time, it is considered that the rows of the synthesized diffusion patterns are arranged in a cyclic manner. That is, the first combined diffusion pattern R ₁ and the last combined diffusion pattern R _n are considered “adjacent”, and the number of consecutive symbols is determined.

Symbol number of continuous, diffusion pattern P _1, ..., multiplexed spread pattern R ₁ from P _k, ..., the calculation formula for obtaining the R _n is determined one, so that the number of continuous one symbol is determined. For example, when k = 3 and n = 3, the number of consecutive symbols is 3 when the combined diffusion pattern is determined as follows.

R ₁ = P ₁ + P ₂
R ₂ = P ₂ + P ₃
R ₃ = P ₁ + P ₃
The diffusion pattern combining unit 103 combines diffusion patterns so that the number of consecutive symbols is as small as possible. In particular, on the condition that m symbols are always “included” in one combined diffusion pattern, the diffusion pattern may be combined so that the number of consecutive symbols is minimized when given k and n are given. Good.

Step 104) Diffusion pattern switching:
In the diffusion pattern switching unit 104, the synthesized diffusion patterns are cyclically switched one by one from the sequence of synthesized diffusion patterns stored in the storage means (not shown) of the diffusion pattern synthesis unit 103, and the digital watermark is selected. A composite diffusion pattern used for embedding is determined and stored in storage means (not shown) such as a memory.

  The composite diffusion pattern is switched so that the composite diffusion pattern is sequentially switched in synchronization with the switching of the frame image of the video signal passed to the embedding pattern superimposing unit 107 by the frame dividing unit 106. That is, for each frame of the video signal, a synthesized diffusion pattern that is different in order for each frame is embedded according to the order output by the diffusion pattern synthesis unit 103.

For example, when the video signal is composed of 10 frames, and the combined diffusion patterns R ₁ , R ₂ , and R ₃ are obtained with n = 3, each frame has the following corresponding to the combined diffusion pattern: The composite diffusion pattern is switched.

Frame 1 ← R ₁
Frame 2 ← R ₂
Frame 3 ← R ₃
Frame 4 ← R ₁
Frame 5 ← R ₂
Frame 6 ← R ₃
Frame 7 ← R ₁
Frame 8 ← R ₂
Frame 9 ← R ₃
Frame 10 ← R ₁
Step 105) Embedding pattern generation:
The embedding pattern generation unit 105 generates an image pattern that is actually superimposed on the video signal based on the combined diffusion pattern selected by the diffusion pattern switching unit 104 and stored in the storage unit, and stores the storage unit such as a memory ( (Not shown).

  This can be performed, for example, by the following method described in Patent Document 1 described above. That is, based on the synthesized diffusion pattern obtained as a one-dimensional number sequence, the watermark coefficient matrix obtained by increasing / decreasing the component value of the complex matrix is subjected to orthogonal transformation such as discrete inverse Fourier transformation, and the resulting image pattern May be used as an embedding pattern.

Further, when the diffusion pattern generation unit 102 obtains diffusion patterns P ₁ ,..., P _k as two-dimensional image patterns, the combined diffusion pattern represented by the sum may be used as an embedded pattern as it is. .

Step 106) Frame image division:
The frame dividing unit 106 divides the pre-embedding signal 110 input to the digital watermark embedding apparatus 100 into each frame image constituting the video and stores it in a storage means (not shown) such as a memory.

  The divided frame images stored in the storage means are sequentially transferred to the embedding pattern superimposing unit 107. As described above, the switching of the frame image passed to the embedding pattern superimposing unit 107 is performed in synchronization with the switching of the composite diffusion pattern in the diffusion pattern switching unit 104.

Step 107) Embedding pattern superimposition:
The embedding pattern superimposing unit 107 reads the frame image divided by the frame dividing unit 106 and stored in the storage unit, adds the embedding patterns generated by the embedding pattern generation unit 105, and superimposes the digital images. An embedded frame image is generated and stored in storage means (not shown). When superimposing the embedding pattern, if the size (view angle) of the frame image is larger than the size (view angle) of the embedding pattern, the embedding pattern is repeatedly tiled vertically and horizontally as shown in FIG. to add. Although FIG. 6 shows an example in which the embedding pattern has a square shape, it goes without saying that the embedding pattern may have another shape. For example, the embedding pattern may be a parallelogram.

Further, when the embedding pattern is added, embedding may be performed with emphasis by a predetermined intensity parameter α. That is, the frame image is represented by {I _{x, y} } (0 ≦ x <the horizontal size of the frame image, 0 ≦ y <the vertical size of the frame image represents the position of the pixel on the image), and the embedded pattern is represented by {W _{i, j} } (0 ≦ i, j <N), the frame image {I ′ _{x, y} } after embedding is
I ′ _{x, y} = I _{x, y} + α · W _{x% N, y% N}
Seek like. (However, a% b represents the remainder when a is divided by b)
The intensity parameter α may be configured to change according to the feature amount of the entire frame image, the pixel to be subjected to calculation of the frame image, or the surrounding pixels. For example, a frame image and its portion where the added embedding pattern is not noticeable are strong (that is, α takes a large value), and a frame image and its portion that are conspicuous are weak (ie, α takes a small value). ) It may be configured to be embedded.

Step 108) Frame image connection:
In the frame concatenation unit 108, by concatenating the frame image embedded with the digital watermark read from the storage unit (not shown) of the embedding pattern superimposing unit 107, the frame image is concatenated reversely to that performed in the frame dividing unit 106. An embedded signal 111 is obtained. This becomes a video signal in which a digital watermark is embedded, and is output from the digital watermark embedding apparatus 100.

  The following points are mentioned as features of the present embodiment.

  First, the embedding pattern is generated by switching the diffusion pattern for each frame and is superimposed on the frame image, so that the digital watermark pattern existing in the video in the embedded signal 111 changes every moment. Phenomenon that the pattern of the digital watermark generated by embedding the pattern is stationary and easily looks like a frosted glass. Embedding can be realized, and as a result, the resistance of the digital watermark can also be increased.

  In addition, since the components corresponding to the common symbols are also common in the embedding pattern for each frame, if the common symbols are embedded in adjacent frames when differentiating the diffusion pattern for each frame, the above-mentioned Although the “frosted glass effect” is likely to occur, in the synthesis of the diffusion pattern, the “ground glass effect” can be made as difficult as possible by performing the synthesis so that the number of consecutive symbols is reduced.

For example, in an extreme case, if k = n and diffusion patterns are combined so that R ₁ = P ₁ , R ₂ = P ₂ ,..., R _n = P _k , the number of consecutive symbols is 0 and the same This symbol appears only every k frames, so that the phenomenon of being easily seen as a ground glass can be greatly suppressed.

  Next, the definition of the number of consecutive symbols in step 108 will be described.

In the present embodiment, the number of consecutive symbols is defined as the number of symbols including the symbols in which adjacent patterns of the combined diffusion pattern are common to each other, but this is further expanded, You may make it evaluate the congestion degree about each symbol in the row | line | column of a series of synthetic | combination spreading | diffusion patterns. For example, for a symbol S _i , the distance between any two combined diffusion patterns that both contain S _i (ie, the number of other combined diffusion patterns present between the two combined diffusion patterns in the combined diffusion pattern sequence + 1). The sum of the reciprocal of the smaller value (which takes into account the circulation of the combined diffusion pattern) is used as the number of consecutive symbols for symbol S _i , and the sum of all symbols is used as the total number of consecutive symbols. May be. For example, when n = 6 and S _i is included in R ₁ , R ₂ , R ₄ , the number of consecutive symbols for S _i is

となる。

It becomes.

[Second Embodiment]
In the present embodiment, a case will be described in which the digital watermark embedded in the first embodiment is detected.

  FIG. 7 shows a configuration example of the digital watermark detection apparatus in the second embodiment of the present invention.

  The digital watermark detection apparatus 400 includes a plurality of filter units 401, a plurality of block calculation units 402, a plurality of delay units 403, a block addition unit 404, and an embedded information detection unit 405.

  In the digital watermark detection apparatus 400, an embedded signal 111 output from the digital watermark embedding apparatus 100, or a signal that has been compression-encoded / decoded, modified, frame thinned out, tampered with the embedded signal 111, and the like are detected signals. Information input as 406 and embedded in the detection target signal 406 is detected, and detection information 407 is output.

  The digital watermark detection process by the digital watermark detection apparatus 400 is performed according to the following procedure shown in FIG. FIG. 8 is a flowchart of digital watermark detection according to the second embodiment of the present invention.

Step 501) Dividing the frame image:
By applying each of the plurality of delay units 403 stepwise to the input detection target signal 406, the frame image F ₀ = F (T) at a certain time T of the detection target signal 406, T−ΔT The frame image F ₁ = F (T−ΔT), and the frame image F ₂ = F (T−ΔT) in T−2ΔT is divided into a plurality of frame images. ΔT represents a delay time by the delay unit 403.

  Most simply, the delay time ΔT is made to coincide with the frame rate of the detection target signal 406, and all input frame images are used for detection of the digital watermark.

  Further, by configuring the delay time ΔT to be different from the frame rate of the detection target signal 406, or by configuring each delay unit 403 to have a different delay time ΔT, the resistance to frame thinning attacks is slightly increased. The following effects can be obtained instead of weakening. That is, the digital watermark is detected using a wide range of frame images on the time axis of the detection target signal 406, and the correlation between the frames used for the digital watermark detection for the original image components other than the digital watermark is reduced. As a result, resistance to deformation, coding noise, tampering, etc. can be increased.

  In addition, here, a method for dividing the detection target signal 406 into frame images using the delay unit 403 has been described. However, the frame images are cut out in order from the input detection target signal 406 and stored in the buffer. It is also possible to perform division into

Step 502) Apply a filter to each segmented frame image:
The filter unit 401 performs filtering on each of the frame image F0 at a certain point in time and the frame images F1, F2,... Obtained by being delayed by the delay unit 403. The filter processing applied by the filter unit 401 applies a filter according to the frequency characteristics of the embedding pattern. For example, when the embedded information is embedded in a high frequency region of the signal, a high frequency pass filter may be applied and further clipping processing may be performed. As the high frequency pass filter, for example, the Laplacian filter of FIG. 9 may be used.

Step 503) Divide the filtered frame image into block images:
In the block calculation unit 402, storage means (not shown) such as a memory as a block image obtained by dividing each filtered image obtained by the filter unit 401 into small images having a block size that matches the size of the digital watermark embedding pattern To store.

Step 504) Add block images divided in each frame:
The block adder 404 acquires the above block images from the storage means of the block calculator 402 for each of the plurality of frame images, adds them, and outputs the added block image to the embedded information detector 405.

Step 505) Detect embedded information:
The embedded information detection unit 405 detects the embedded information embedded in the embedded signal 406 using the added block image obtained by the block adding unit 404 and outputs it as detection information 407.

  For example, as described in Patent Document 1 described above, embedded information detection is performed by embedding a digital watermark corresponding to a detection target coefficient matrix obtained by performing a discrete Fourier transform on an addition block image and a symbol of the embedded information. Even if the correlation value with the diffusion pattern used for embedding in the apparatus 100 is obtained, it is determined that the symbol having a high correlation value has been embedded, and detection information is constituted and output from these symbols. Good.

  Next, a case where block addition is performed after taking out a one-dimensional series in step 504 will be described.

  In the above example, the block images divided in each frame are added in step 504. For example, when the digital watermark embedding / detecting method described in Patent Document 1 is used, Instead of adding block images, a one-dimensional detection target sequence obtained from the block images may be added.

  That is, the block image divided in each frame is subjected to discrete Fourier transform, the detection target sequence is obtained based on the detection target component position information, and then the detection target sequence obtained from each block image is added. In step 505, the embedded information detection unit 405 may detect the embedded information by obtaining a correlation value between the added detection target sequence and the diffusion pattern corresponding to the symbol of the embedded information.

  In addition, as an intermediate method, it goes without saying that the detection target sequences of a plurality of frames may be added after the block images are added in units of frames to obtain the detection target sequences.

  Next, effects of the digital watermark detection apparatus 400 will be described.

  According to the configuration of the digital watermark detection apparatus 400 as described above, each frame image includes only some symbols by performing digital watermark detection after adding information obtained from a plurality of frame images. Therefore, even when a frame thinning attack is performed, all symbols can be detected without having to synchronize the frames.

  Further, according to the configuration of the digital watermark detection apparatus 400 as described above, the accuracy of detection of the digital watermark can be improved by the following effects, and as a result, the resistance to signal deformation and noise is increased. be able to. Conversely, when the same level of tolerance is required, the digital watermark can be detected even if the strength parameter in the embedding pattern superimposing unit 107 of the digital watermark embedding apparatus 4 is reduced. These effects are also effective when the diffusion pattern is not switched in the embedding of the digital watermark (that is, when n = 1 in the digital watermark embedding apparatus).

(1) Effect of inter-frame signal addition by the block adder 404:
By adding block images obtained from a plurality of frames by the block adding unit 404, the embedding pattern components embedded in the signal are emphasized, and for example, the correlation value obtained by the embedded information detecting unit 405 becomes larger. , Detection accuracy is improved.

(2) Effect of filtering by the filter unit 401:
The detection accuracy is improved by performing detection in the embedded information detection unit 405 after the digital watermark component is further emphasized by the filter unit 401.

  In addition, in a video signal, since the correlation between consecutive frame images is relatively high, when the block image obtained from a plurality of frames is added by the block adder 404 without the filter unit 401, the result is a digital watermark. In addition to the above components, the original image components are also emphasized, and the effect of adding blocks may be reduced. By emphasizing the high-frequency region by the filter unit 401, it is possible to remove low-frequency image components included in the frame image and leave high-frequency noise components. Since the high-frequency noise component of the image is considered to have a low correlation between frame images as compared with the low-frequency image component, the effect of adding blocks is not reduced. Therefore, by leaving this, the detection accuracy is further improved.

(3) Effects of block division by the block calculation unit 402:
As shown in FIG. 6, the embedding pattern 301 is repeatedly tiled in advance with respect to the frame image, and the block calculating unit 402 further divides the image of one frame into blocks, and then adds them in the block adding unit 404. This contributes to increasing the resistance of the digital watermark to signal deformation and noise in the following points.

  That is, each block image divided into a plurality of blocks within one frame image often has a lower correlation value than the correlation value between images at the same position in adjacent frames. Even if a specific block image contains many components that cancel out the embedding pattern embedded as a digital watermark (that is, a pattern having a negative correlation with the embedded diffusion pattern), While the above components are not strengthened by the addition of, the components of the embedding pattern are emphasized by the addition, and as a result, the accuracy of detection is improved and the resistance to deformation and noise can be increased.

(4) Effect of addition after detection target sequence calculation:
As described above, after obtaining the detection target sequence from the block image divided in each frame, the detection target sequence obtained from each block image is added to remove as much information as possible not related to the detection of the digital watermark. As a result, the digital watermark component is relatively emphasized, and the detection accuracy is improved.

  Next, variations of the frame range in which the block addition unit 404 performs block addition will be described.

  In the digital watermark detection apparatus 400 described above, an example is shown in which block addition is performed for the number of stages of the delay unit 403 plus one frame to detect embedded information. The delay unit 403 and the block addition unit 404 may be configured so as to target various ranges as frame ranges to be subjected to block addition at a time. An example is shown in FIG.

  FIG. 10A shows an example in which block addition is performed on the six frames from the beginning of the detection target signal, and then block addition is performed on the six frames while shifting the first frame one frame at a time.

  FIG. 10B shows an example in which block addition is performed on the six frames from the beginning of the detection target signal, and then block addition is performed on the six frames while shifting the first frame by four frames. In this example, compared to the example of FIG. 10A, the number of times that the embedded information detection unit 405 actually detects the embedded information is reduced with respect to the entire detection target signal, and the digital watermark is more efficiently Detection can be performed.

  FIG. 10C shows an example in which the number of frames for which block addition is performed from the head of the detection target signal is increased by 1 sequentially from 1. In the case of this example, the digital watermark detection operation may be continued until the embedded information is detected by the embedded information detection unit 405, and the operation may be terminated when the embedded information is detected. When it is unclear how difficult it is to detect the digital watermark, it is possible to efficiently detect the digital watermark. Needless to say, the number of frames to be increased may be a plurality of frames instead of one frame.

[Third Embodiment]
In the present embodiment, a digital watermark embedding apparatus that generates a combined diffusion pattern so as to minimize the number of consecutive symbols for minimizing the “frosted glass effect” appearing in a watermark-embedded video will be described.

In the example of R ₁ = P ₁ , R ₂ = P ₂ ,..., R _n = P _k shown as an example in the first embodiment, as described above, as described above, However, the detection accuracy of each symbol is lowered when detecting a digital watermark. In other words, in order to increase the resistance against attacks such as frame thinning when detecting a digital watermark, the embedded information is detected after adding blocks obtained from each frame so that detection can be performed without synchronization between frames. When a certain symbol is viewed, since the frame in which the symbol is embedded is once in k frames, adding blocks obtained from other frames is a noise component for the digital watermark. As a result, the accuracy of symbol detection was reduced.

  In order to prevent this, it is necessary to increase the number m of symbols embedded in one frame. On the other hand, if the number of consecutive symbols is increased as a result of increasing m, the phenomenon that the above-mentioned frosted glass can be easily observed occurs. Therefore, the diffusion pattern is reduced so as to reduce the number of consecutive symbols while increasing m as much as possible. It is necessary to perform synthesis.

  The digital watermark embedding apparatus according to the present embodiment has the same configuration as the digital watermark embedding apparatus according to the first embodiment.

  However, in the digital watermark embedding apparatus of this embodiment, the diffusion pattern synthesis unit 103 is configured as shown in FIG. The spreading pattern synthesis unit 800 shown in the figure includes a synthesis pattern configuration series calculation unit 801 and a spreading pattern addition unit 802.

  The diffusion pattern synthesis in the diffusion pattern synthesis unit 800 is performed in the following procedure.

1) In the composite pattern configuration series calculation unit 801, when obtaining a composite diffusion pattern from the diffusion pattern 803, which symbol is included in which composite diffusion pattern with respect to a predetermined value of k, n, m. A composite pattern constituent series to be expressed is obtained by calculation. The composite pattern configuration series 805 includes, for example, a vector r _i composed of k elements having the following values 0 and 1.

r _i = (1, 0, 0, 1, ..., 1, 0)
Here, the j-th element of r _i is 1, that is, ON means that the i-th combined diffusion pattern R _i includes the j-th symbol S _j, and the j-th element of r _i is 0, that is, OFF indicates that the i-th combined diffusion pattern R _i does not include the j-th symbol S _j .

For example, when k = 6,
r ₁ = ( _1, 0, 0, _{1, 1,} 0)
If so, the combined diffusion pattern R ₁ is
R ₁ = P ₁ + P ₄ + P ₅
It represents that it is calculated by.

2) The diffusion pattern addition unit 802 calculates a composite diffusion pattern 804 from the diffusion pattern 803 according to the composite pattern configuration sequence 805 obtained by the composite pattern configuration series calculation unit 801. Therefore, the combined diffusion pattern R _i is obtained by the following equation.

R _i = r _i · (P ₁ , P ₂ ,..., P _k )
_{= R i, 1 P 1 +} r i, 2 P 2 + ... + r i, k P k
However, “·” represents an inner product, and r _{i, j} represents the j-th element of r _i .

  Next, calculation of the composite pattern constituent series of the composite pattern constituent series calculator 801 will be described in detail.

As the simplest configuration example of the composite pattern constituent sequence calculation unit 801, for a given k and n, the number of consecutive symbol patterns is obtained for all possible composite pattern constituent sequences, and the composite is such that it is minimized. This can be realized by selecting a pattern composition series. However, the number of all synthetic pattern constituent sequences that can be considered for a given k, n is 2 ^{k × n} . For example, even when m is constant, ( _k C _m ) ⁿ Each pattern can be considered, and the calculation takes a very long processing time. Therefore, it is necessary to configure the composite pattern composition series calculation unit 801 so that the composite pattern composition series can be calculated at high speed.

  Hereinafter, a configuration example of a specific procedure in the combined pattern configuration series calculation unit 801 will be described.

  Here, when k = n = 2m, a procedure for calculating a composite pattern constituent sequence so that the number of consecutive symbols is minimized will be described with reference to FIG. As a specific example, a case where k = n = 6 and m = 3 will be described as an example.

Step 901) r ₁ as an initial pattern is arbitrarily determined within a range that satisfies given conditions of k and m. For example,
r ₁ = ( _{1, 1, 1, 0,} 0, 0)
Step 902) The counter c is initialized to 1, and the set I storing the element numbers once selected is initialized to an empty set.

Step 903) One element number i ₀ is arbitrarily selected from elements which are ₁ in r ₁ . For example, since r ₁ = (1, 1, 1, _{0, 0, 0} ), i ₀ = 1 is selected.

  The following steps 904 to 908 are repeated steps using c as a counter.

Step 904) Let _{r2c be} a vector obtained by inverting 0/1 of each element of r2c _-1 . That is,

をＡの要素の０／１を反転したベクトルを表すとすると、

Represents a vector obtained by inverting 0/1 of an element of A,

例えばｒ_１＝（１，１，１，０，０，０）であることから

For example, r ₁ = (1,1,1,0,0,0)

となる。

It becomes.

  Step 905) It is determined whether 2c is smaller than n. If it is smaller, the process proceeds to Step 906. If 2c has reached n, the process repeats and the process proceeds to Step 909.

Step 906) Among the elements that are 1 in r _2c , an element number _ic that is not included in I is selected. Further, the selected _ic is added to I so that it will not be selected in the subsequent steps. For example, r ₂ = (0, 0, 0, 1, 1, 1)
i ₁ = 6
age,
I = {6}
It becomes.

Step _{907) r 2c-1, i} to _c-1 th element thinking vector obtained by changing the 0, _{i c} th element 1, which is referred to as _{r 2c + 1.} That is,

ただし、
ｅ_１＝（1，0，0，…，0）
ｅ_２＝（0，1，0，…，0）
・
・
・
ｅ_ｉ＝（0，…，0，1，0，…，0） （ｉ番目の要素のみ１）
・
・
・
ｅ_ｋ＝（0，0，0，…，1）
例えば、ｒ_１＝(1，1，1，0，0，0)，ｉ_０＝１，ｉ_１＝６であることから、
ｒ_３ ＝ｒ_１−ｅ_１＋ｅ_６
＝（1，1，1，0，0，0）−（1，0，0，0，0，0）＋（0，0，0，0，0，1）
＝（0，1，1，0，0，1）
ステップ９０８） カウンタｃをｃ←ｃ＋1として１進め、ステップ９０４に戻る。

However,
e ₁ = (1, 0, 0, ..., 0)
e ₂ = (0, 1, 0, ..., 0)
・
・
・
e _i = (0, ..., 0,1,0, ..., 0) (only the i-th element is 1)
・
・
・
e _k = (0, 0, 0, ..., 1)
For example, since r ₁ = (1, 1, 1, ₀ , ₀ , ₀ ), i ₀ = 1, i ₁ = 6,
r ₃ = r ₁ −e ₁ + e ₆
= (1, 1, 1, 0, 0, 0)-(1, 0, 0, 0, 0, 0) + (0, 0, 0, 0, 0, 1)
= (0, 1, 1, 0, 0, 1)
Step 908) The counter c is incremented by 1 with c ← c + 1, and the process returns to Step 904.

In this repetition, for example, r _i is determined as follows.

r ₁ = ( _{1, 1, 1, 0, 0, 0} ) i ₀ = 1 I = {}
r ₂ = (0, 0, 0, 1, 1, 1)
r ₃ = (0, 1, 1, 0, 0, 1) i ₁ = 6 I = {6}
r ₄ = (1, 0, 0, 1, 1, 0)
r ₅ = (0, 1, 1, 0, 1, 0) i ₂ = 5 I = {5, 6}
r ₆ = (1, 0, 0, 1, 0, 1)
If 2c in step 905 described above has reached n, the synthesized pattern configuration sequence r ₁ in step 909 exits the _repeat, ..., and terminates with an r _n.

  As a result, when k = n = 2m, it is possible to generate a composite pattern configuration sequence that minimizes the number of consecutive symbols. As a result, it is possible to configure a digital watermark embedding apparatus that suppresses the “frosted glass effect”. .

  In the above example, a composite pattern constituent series having a symbol continuation number of 3 is obtained, and this has the minimum number of symbol continuations under the conditions of k = n = 6 and m = 3.

In the above example, instead of storing the element number ic selected in step 906 in the storage area I, it is adopted a method which does not use the storage area I by selecting the i _c have it order on a predetermined rule Good. For example, in step 901 above, r ₁ is

のように決定し、ｉ_１＝ｍ＋１，ｉ_２＝ｍ＋２，…，ｉ_ｍ−１＝ｋ−１のように順に選択してもよい。例えば、ｋ＝ｎ＝８，ｍ＝４，ｉ_０＝１の場合に次のような合成パターン構成系列になる。

And i ₁ = m + 1, i ₂ = m + 2,..., I _m−1 = k−1 may be selected in order. For example, when k = n = 8, m = 4, and i ₀ = 1, the following composite pattern configuration sequence is obtained.

r ₁ = (1, 1, 1, 1, _{0, 0, 0, 0} ) i ₀ = 1
r ₂ = (0, 0, 0, 0, 1, 1, 1, 1)
r ₃ = (0, _{1, 1, 1, 1,} 0, 0, 0) i ₁ = 5
r ₄ = (1, 0, 0, 0, 0, 1, 1, 1)
r ₅ = (0, 1, 1, 1, 0, _1, 0, 0) i ₂ = 6
r ₆ = (1, 0, 0, 0, 1, 0, 1, 1)
r ₇ = (0, 1, 1, 1, 0, 0, 1, 0) i ₃ = 7
r ₈ = (1, 0, 0, 0, 1, 1, 0, 1)
This also r ₁ obtained by described method _until, ..., matrix for r _n

の各列を入れ替えた行列を

A matrix with each column of

のように定義したとき、ｒ’_１，…，ｒ’_ｎもまた、シンボル連続数が最小となるため、これを合成パターン構成系列としてもよい。

R ′ ₁ ,..., R ′ _n also have the smallest number of consecutive symbols.

[Fourth Embodiment]
In the present embodiment, an example will be described in which the diffusion pattern synthesis unit gives randomness to the determination of the synthesis pattern constituent series.

  In the digital watermark embedding device described in the third embodiment, when the synthesized pattern configuration sequence obtained by the synthesized pattern configuration sequence calculation unit 800 is configured to have a fixed value, the synthesized pattern configuration is given to the attacker. If the sequence is known, there is a vulnerability that makes it difficult to detect an electronic watermark. That is, by intentionally thinning out a frame in which a composite diffusion pattern including a specific symbol is embedded, an attack that deletes the embedded information of the symbol or makes it difficult to detect is possible.

  In the digital watermark embedding apparatus according to the present embodiment, resistance against such an attack is increased by adding randomness to the determination of the composite pattern constituent series.

  The configuration of the digital watermark embedding apparatus of this embodiment is the same as that of the digital watermark embedding apparatus shown in the third embodiment except for the following points.

  In the digital watermark embedding apparatus according to the present embodiment, the procedure in the combined pattern composition series calculation unit 800 differs in the following points.

In FIG. 12, any one of the determination of the initial pattern in step 901, the selection of element number i _{0 in} step 903, the selection of element number _{ic in} step 906, and the column replacement described at the end of the previous section is randomly performed. By doing so, the resultant composite pattern constituting sequence is randomized, and which symbol is embedded in which frame can be made different depending on the embedding, and the resistance to a thinning-out attack on a specific symbol can be increased.

  Alternatively, the randomness of the composite pattern constituent series may be further increased by repeatedly generating a composite pattern constituent series having a predetermined length n.

  The determination of the random initial pattern, the selection of the element number, and the replacement of the column are, for example, a predetermined key, a PN sequence such as a pseudorandom number calculated based on the characteristics of the video signal to be embedded, an M sequence, It can be performed using a random number using thermal noise of the sensor.

Further, according to the above, the synthetic pattern configuration sequence _r 1, ..., was produced to have a randomness of _{r n,} as _{r n} + ₁ = _r 1, _{further, r} n + 2, _..., a random _{r 2n} same as above It can be generated by selecting an element number and repeated in the same way, so that a composite pattern composition sequence with randomness can be generated continuously. Resistance can be increased.

[Fifth Embodiment]
In the present embodiment, an example will be described in which a fixed synthesis pattern constituent sequence is used in the diffusion pattern synthesis unit of the digital watermark embedding apparatus.

  In the present embodiment, in the diffusion pattern synthesis unit 103 shown in the first embodiment, instead of being sequentially calculated by the synthesis pattern composition series calculation unit 801 as shown in the third embodiment, A fixed composite pattern constituent series is stored.

  The digital watermark embedding apparatus according to the present embodiment is configured in the same manner as the digital watermark embedding apparatus 100 according to the first embodiment.

  In the digital watermark embedding apparatus 100 of the present embodiment, the diffusion pattern synthesis unit 103 is configured as shown in FIG.

  The diffusion pattern synthesis unit 1000 according to the present embodiment includes a synthesis pattern configuration series storage unit 1001 and a diffusion pattern addition unit 1002.

In the combined pattern configuration series storage unit 1001, for example, a combined pattern configuration series 1005 calculated by a method such as the combined pattern configuration series calculation unit 801 in the third embodiment is stored in advance. The configuration of the composite pattern configuration series 1005 is the same as that of the third embodiment described above, and includes, for example, a vector r _i composed of k elements having the following values 0 and 1.

r _i = (1, 0, 0, 1, ..., 1, 0)
Here, the j-th element of r _i being 1 indicates that the i-th combined diffusion pattern R _i includes the j-th symbol S _j, and the j-th element of r _i is 0. Represents that the i-th combined diffusion pattern R _i does not include the j-th symbol S _j .

  The diffusion pattern addition unit 1002 is the same as the diffusion pattern addition unit 802 in the third embodiment.

  Corresponding to the present embodiment, a composite pattern constituent sequence generation apparatus for calculating in advance a composite pattern constituent sequence to be stored in the composite pattern constituent sequence storage unit 1001 is described in the third embodiment of the present invention. It may be configured to have the same function as the combined pattern configuration series calculation unit 801 in the form.

[Sixth Embodiment]
In the present embodiment, an example will be described in which a digital watermark embedding apparatus considers encoding when generating a spread pattern synthesis sequence.

  When a video signal is encoded by an encoding method such as MPEG-2 or Windows (registered trademark) Media Video, for example, an intra coding is generally performed in which a certain frame image is compression-coded using only information in the frame image. Inter-encoded frames (for example, MPEG-) that compress and encode a frame image (for example, an I picture in MPEG-2) and a difference between a frame image and a neighboring frame image by using a motion vector or the like. 2 (P picture and B picture) are used.

  In general, a frame image that has been subjected to intra coding and a frame image that has been subjected to inter coding are encoded at a higher compression rate, so that even when embedding an electronic watermark, The encoded frame image is easier to detect.

  Using this, when it is known in advance that the video is encoded as described above, when generating the spread pattern synthesis sequence, all the symbols can be obtained by collecting the frame images to be subjected to intra encoding. By including a diffusion pattern synthesis sequence so as to be included, detection accuracy can be maintained over all symbols.

  In addition, when detecting a digital watermark from an encoded video signal, the detection accuracy of the digital watermark may be improved by performing block addition on an intra-coded frame image.

  In addition, it is possible to construct | assemble the operation | movement of each apparatus in said 1st-6th embodiment as a program, install it in a computer, to perform, or distribute | circulate through a network.

  In addition, the constructed program can be stored in a portable storage medium such as a hard disk device connected to the computer, a flexible disk, or a CD-ROM, and can be installed or distributed in the computer.

  The present invention is not limited to the above-described embodiment, and various modifications and applications can be made within the scope of the claims.

  The present invention can be applied to a technique of digital watermark embedding and digital watermark detection for a video signal.

It is a figure for demonstrating the principle of this invention. It is a principle block diagram of this invention. It is an example of a structure of the digital watermark embedding apparatus in the 1st Embodiment of this invention. It is a flowchart of the digital watermark embedding process in the 1st Embodiment of this invention. It is a structural example of the symbol in the 1st Embodiment of this invention. It is an example of the tiling of the embedding pattern in the 1st Embodiment of this invention. It is a structural example of the digital watermark detection apparatus in the 2nd Embodiment of this invention. It is a flowchart of the electronic watermark detection in the 2nd Embodiment of this invention. It is an example of the Laplacian filter in the 2nd Embodiment of this invention. It is an example of the frame range which performs block addition in the 2nd Embodiment of this invention. It is a block diagram of the diffusion pattern synthetic | combination part in the 3rd Embodiment of this invention. It is a flowchart of the synthetic | combination pattern structure series calculation part in the 3rd Embodiment of this invention. It is an example of a structure of the diffusion pattern synthetic | combination part in the 5th Embodiment of this invention.

Explanation of symbols

DESCRIPTION OF SYMBOLS 100 Digital watermark embedding apparatus 101 Embedded information division | segmentation means, Embedded information division | segmentation part 102 Diffusion pattern production | generation means, Diffusion pattern production | generation part 103 Diffusion pattern synthesis | combination means, Diffusion pattern synthesis | combination part 104 Diffusion pattern switching means, Diffusion / diffusion pattern switching means 105 Embedded pattern generation Unit 106 frame division unit 107 embedding pattern superimposing unit, embedding pattern superimposing unit 108 frame connecting unit 109 embedding information 110 pre-embedding signal 111 embedding signal 301 embedding pattern 400 digital watermark detection device 401 filter unit, filter unit 402 block calculation unit, block Calculation unit 403 Delay unit 404 Block addition means 405 Embedded information detection unit 406 Detection target signal 407 Detection information 800 Diffusion pattern synthesis unit 801 Synthesis pattern configuration Column calculation unit 802 spreading pattern adding unit 803 diffusion pattern 804 combined spreading pattern 805 combined pattern configuration sequence 1000 diffusion pattern combining section 1001 synthesized pattern configuration sequence storing section 1002 diffusion pattern adding unit 1003 diffusion pattern 1994 synthetic diffusion pattern 1005 synthetic pattern configuration sequence

Claims (21)

An electronic watermark embedding method in an electronic watermark embedding device that superimposes a predetermined pattern configured based on embedding information on each frame image of the video signal so as not to be perceived by human perception. There,
An embedded information dividing step of taking embedded information as input and dividing the embedded information into a plurality of symbols using embedded information dividing means;
A diffusion pattern generation step for generating a diffusion pattern used for embedding for each symbol using a diffusion pattern generation means;
A diffusion pattern synthesizing step of selecting one or more diffusion patterns from the diffusion patterns for each symbol using a diffusion pattern synthesis means and combining them to generate a plurality of resultant diffusion patterns;
A diffusion pattern switching step of switching a plurality of the synthesized diffusion patterns for each frame to be embedded using a diffusion pattern switching means;
A digital watermark embedding pattern superimposing step for embedding a digital watermark by superimposing a digital watermark embedding pattern image corresponding to the switched composite diffusion pattern on the frame image using a frame image as an input and embedding pattern superimposing means; ,
An electronic watermark embedding method comprising: In the diffusion pattern synthesis step,
2. The electron according to claim 1, wherein the composite diffusion pattern is generated so that the number of consecutive symbols, which is an indicator of the degree of common symbols included in adjacent or close patterns in the composite diffusion pattern sequence, is minimized. Watermark embedding method. In the diffusion pattern synthesis step,
3. The digital watermark embedding method according to claim 1 or 2, wherein a composite pattern constituent sequence indicating which symbols are included in each of the generated composite spread patterns is stored in the pattern constituent sequence storage means. A method for generating a composite pattern constituent sequence in a composite pattern constituent sequence generating apparatus for generating a composite pattern constituent sequence that indicates which symbols each includes a composite diffusion pattern in the digital watermark embedding method according to claim 1. And
An initial composite pattern constituent series determining step of acquiring a composite pattern constituent series from the storage means and determining a _first composite pattern constituent series r ₁ ;
An initial change target element selection step of selecting an arbitrary element from elements that are ON values of the _first composite pattern constituent series r ₁ ;
A vector obtained by inverting the ON value / OFF value of each element of the 2c-1 th (where c is a counter) composite pattern constituent series r _2c-1 is generated, and is defined as the 2 _c th composite pattern constituent series r _2c . A pattern inversion step;
A next change target element selection step of selecting an element number _ic that has not been selected so far among the elements that are the ON values of the 2c-th composite pattern constituent series r _2c ;
In the 2c-1th composite pattern constituent series r _2c-1 , the _ic-1 th element is changed to an OFF value, the _ic th element is changed to an ON value, and this is changed to a 2c + 1th composite pattern constituent series r _{2c + 1.} And the next pattern generation step,
A synthetic pattern composition series generation method characterized by performing:   5. The generation of the composite pattern constituent series or the selection of elements is randomly performed in any one of the initial composite pattern constituent series determining step, the initial change target element selecting step, and the next change target element selecting step. The synthetic pattern composition series generation method. In the digital watermark embedding method according to any one of claims 1 to 3, in the diffusion pattern synthesis step,
4. The digital watermark embedding method according to claim 1, wherein each step of the synthetic pattern composition series generation method according to claim 4 or 5 is performed. The video signal is embedded from a video signal embedded in a digital watermark by superimposing a predetermined pattern configured based on the embedded information on each frame image of the video in advance so that it is difficult for human perception. In a digital watermark detection apparatus for detecting embedded information, a digital watermark detection method comprising:
A filtering step of receiving each frame image of the video signal in which the digital watermark is embedded as input, filtering the frame image using a filtering unit, and outputting the filtered image;
A block calculation step of dividing the frame image subjected to the filtering process into blocks according to the embedded pattern size of the digital watermark using a block calculation unit;
A block addition step of obtaining the plurality of block images obtained in the block calculation step using a block addition unit for the plurality of frame images, and adding information for each of the plurality of block images; ,
An electronic watermark detection method comprising: In the block addition step,
The digital watermark detection method according to claim 7, wherein the plurality of block images themselves are used as information for each block image. In the block addition step,
8. The digital watermark detection method according to claim 7, wherein a detection target sequence obtained from the plurality of block images based on detection target component position information is used as the information for each block image. An electronic watermark embedding device that superimposes a predetermined pattern configured based on embedding information on each frame image of the video signal so as not to be perceived by human perception, on the video signal,
Embedded information dividing means for dividing the inputted embedded information into a plurality of symbols;
A diffusion pattern generation means for acquiring the divided symbols and generating a diffusion pattern used for embedding for each symbol;
Obtaining a diffusion pattern for each symbol, selecting one or more diffusion patterns, combining them, and generating a plurality of resultant diffusion patterns;
A diffusion pattern switching means for acquiring a plurality of generated composite diffusion patterns and switching for each frame to be embedded;
An electronic device in which a frame image is input, the switched composite diffusion pattern is acquired, a digital watermark embedding pattern image corresponding to the switched composite diffusion pattern is superimposed on the frame image, and output to embed a digital watermark Watermark embedding pattern superimposing means;
A digital watermark embedding apparatus comprising: The diffusion pattern synthesis means includes
11. The means for generating the composite diffusion pattern so as to minimize the number of consecutive symbols, which is an indicator of the degree to which common symbols are included in adjacent or close patterns in the composite diffusion pattern sequence. The electronic watermark embedding device described. The diffusion pattern synthesis means includes
12. The digital watermark embedding apparatus according to claim 10 or 11, further comprising pattern configuration sequence storage means for storing a composite pattern configuration sequence indicating which symbols are to be included in each generated composite diffusion pattern. 13. A combined pattern configuration sequence generation device that generates a combined pattern configuration sequence that represents which symbols each of the combined diffusion patterns includes in the digital watermark embedding device according to claim 10,
Initial composite pattern constituent series determining means for acquiring a composite pattern constituent series from the storage means and determining a _first composite pattern constituent series r ₁ ;
Initial change target element selection means for selecting an arbitrary element from elements that are ON values of the _first composite pattern constituent series r ₁ ;
A vector obtained by inverting the ON value / OFF value of each element of the 2c-1 th (where c is a counter) composite pattern constituent series r _2c-1 is generated, and is defined as the 2 _c th composite pattern constituent series r _2c . Pattern inversion means;
A next change target element selecting means for selecting an element number _ic that has not been selected so far among the elements that are the ON values of the 2c-th composite pattern constituent series r _2c ;
In the 2c-1th composite pattern constituent series r _2c-1 , the _ic-1 th element is changed to an OFF value, the _ic th element is changed to an ON value, and this is changed to a 2c + 1th composite pattern constituent series r _{2c + 1.} A next pattern generating means, and
A composite pattern composition series generation device characterized by comprising:   14. The initial composite pattern constituent series determining means, the initial change target element selecting means, or the next change target element selecting means randomly generates the composite pattern constituent series or selects an element. The synthetic pattern composition series generation device. In the digital watermark embedding device according to any one of claims 10 to 12, the diffusion pattern synthesis means includes:
13. The digital watermark embedding apparatus according to claim 10, further comprising means of the synthetic pattern composition series generation apparatus according to claim 13 or 14. The video signal is embedded from a video signal embedded in a digital watermark by superimposing a predetermined pattern configured based on the embedded information on each frame image of the video in advance so that it is difficult for human perception. An electronic watermark detection device for detecting embedded information,
Filtering means for inputting each frame image of the video signal in which the digital watermark is embedded, and performing filtering processing on each frame image and outputting the filtered frame image;
A block calculation unit that divides the frame image subjected to the filtering process into blocks according to an embedded pattern size of a digital watermark;
Block addition means for adding the acquired information for each of the plurality of block images to the plurality of frame images;
A digital watermark detection apparatus comprising: The block adding means includes
The digital watermark detection apparatus according to claim 16, wherein the plurality of block images themselves are used as information for each block image. The block adding means includes
The digital watermark detection apparatus according to claim 16, wherein a detection target sequence obtained from the plurality of block images based on detection target component position information is used as the information for each block image. An electronic watermark embedding program for superimposing a predetermined pattern configured based on embedding information on each frame image of the video signal so as not to be perceived by human perception, on the video signal,
A digital watermark embedding program for causing a computer to execute processing for realizing the digital watermark embedding method according to claim 1. A synthesis pattern component sequence generation program for generating a composite pattern component sequence that represents which symbols are included in each of the combined diffusion patterns in the digital watermark embedding method,
6. A composite pattern composition series generation program that causes a computer to execute processing for realizing the synthetic pattern composition series generation method according to claim 4 or 5. The video signal is embedded from a video signal embedded in a digital watermark by superimposing a predetermined pattern configured based on the embedded information on each frame image of the video in advance so that it is difficult for human perception. An electronic watermark detection program for detecting embedded information,
10. A digital watermark detection program for causing a computer to execute a method for realizing the digital watermark detection method according to claim 7.

Images