JP4534530B2 - Digital watermark detection method and digital watermark detection apparatus - Google Patents

Description

  The present invention relates to a digital watermark technique for concealing other information (watermark information such as a signature) in video data for the purpose of, for example, video copyright management, and more particularly to a watermark information embedding method and detection method.

  As a conventional watermark information embedding technique, there is a method of embedding information by spectrum-spreading a low-frequency component obtained by wavelet transform of an image and then quantizing an average value of the block (Non-patent Document 1). .

  Further, after frequency-converting the image, information is embedded by manipulating the value of the amplitude component, and at the time of detection, there is a method of determining the embedded information by calculating the correlation with the embedded information for the difference image from the original image ( Non-patent document 2).

  Furthermore, assuming that pixel values are randomly distributed, information is embedded using the fact that the sum of the differences converges to 0 when two points are selected randomly and the difference is taken. There is a method (Non-patent Document 3).

Onishi et al., "Digital Watermarking Using Multi-resolution Analysis and PN Sequence", IEICE Transactions D-II Vol.J80-D-II No.11 pp.3020-3028, 1997 I. J. Cox et.al., "Secure spread spectrum watermarking for multimedia", NEC Research Institute, TR95-10, 1995. W. Bender et.al, "Techniques for data hiding", IBM systems journal, vol.35, no.3-4, pp.313-336, 1996.

  In the conventional watermark information embedding technique as described in Non-Patent Document 1, since information is embedded by quantization, for example, the same value is added to all pixel values by analog processing or the like. However, there is a problem that the embedded information changes and cannot be detected. Further, since the pixel average value for each block is manipulated when information is embedded, there is a problem that block noise is likely to occur.

  In addition, the conventional watermark information embedding technique as described in Non-Patent Document 2 requires an original image when detecting embedded information, so that detection is performed when the original image corresponding to the detection target image does not exist or cannot be obtained. The problem is that it may not be possible, or after searching / discovering the “correct” original image corresponding to the watermark detection target image from the vast library, it is necessary to set “correct” in the watermark detector, which may require a lot of work. There was a point.

  Furthermore, in the conventional watermark information embedding technique such as Non-Patent Document 3 described above, it is assumed that pixel values are randomly distributed. However, depending on the image, pixel values are not necessarily randomly distributed. However, a large number of pixels are required to converge the difference sum to zero. In other words, in order to stably detect information correctly, it is necessary to calculate for a large number of pixels or to reduce the amount of embedded information.

  The present invention has been made to solve the above-described problems, and even if there is a change in pixel value such as analog processing performed on the video signal, the digital watermark information is hardly deteriorated, and the image quality of the video information is deteriorated. An object of the present invention is to realize digital watermark embedding in which the image is suppressed. It is another object of the present invention to realize stable digital watermark detection that does not require an original image.

The present invention generates a time direction spreading code string composed of +1 and −1 set for each video frame constituting a video, and sets a time direction spreading code corresponding to each video frame to each pixel value of the corresponding video frame. An integration step for obtaining an integrated time direction spread frame, an integration step for obtaining a frame integration image obtained by integrating the adjacent frames by a specified number of frames for the time direction diffusion frame obtained in this integration step, and frame integration in this integration step. If the positive and negative code numbers of the time-direction spreading code sequences generated by the integration step that have been added to the specified number of time-direction spread frames do not match, the specification is made until the positive and negative code numbers match. In the change step for changing the number of frames and the frame integration image obtained in the integration step, the above A pixel position determining step for determining a pixel position on the frame integral image for detecting digital watermark information embedded in an image, and a statistical value of a pixel value set at the pixel position determined in the pixel position determining step; And a detection step of detecting digital watermark information .

  According to the present invention, since the digital watermark information is diffused and embedded in the video information, the digital watermark information deteriorates even if the video signal is subjected to processing such as encoding or intentional processing such as analog processing. It is possible to perform difficult digital watermark embedding.

  In addition, the digital watermark information embedded in the video is detected by despreading the digital watermark information based on the diffusion pattern consisting of the diffusion code set for each frame that is the same as when the digital watermark information is embedded in the video. Thus, stable digital watermark information can be detected without using an original image.

Embodiment 1 FIG.
FIG. 1 is a block diagram showing a digital watermark embedding apparatus according to an embodiment of the present invention.
In the figure, 11 is a digital watermark embedding device. The digital watermark embedding device 11 includes a time direction spreading code generation unit 12, an address generation unit 13, a blocking unit 14, an information embedding unit 15, a bit generation unit 16, a block release unit 17, a first integration unit 100, a second integration unit 100, The integrating means 101 is comprised.

Next, the operation will be described.
In the following description, it is assumed that the digital watermark information embedded in the video is digital information composed of a bit string having a predetermined number of bits, and here is the number of bits N (N is a positive integer).
First, an input video to be embedded with digital watermark information is input to the digital watermark embedding device 11. Note that the input video may be a low-frequency component signal for each frame extracted from the original video.

The time direction spreading code generator 12 generates a spreading code string (spreading pattern) composed of +1 and -1. As shown in FIG. 2, each one of the spread code strings is associated with each frame constituting the input video in order. In FIG. 2, 21 is each frame constituting the input video, and 22 is a spreading code corresponding to each frame constituting the input video.
For this spreading code, a fixed spreading code string may be used, or in consideration of security, for example, a spreading code string is generated using a secret value as a seed of a random number for outputting +1 and -1. Also good.

Then, in the first integrating means 100, the corresponding spreading code is integrated for each pixel value constituting the frame of the input video information.
For example, for simplicity, the input video is composed of a frame sequence having a size of 2 × 2 pixels, the pixel values of a certain frame are 10, 200, 50, and 35 in the scanning order, and the spreading code corresponding to the frame is −1. Suppose that At this time, pixel values obtained by integrating the frame and the spreading code are −10, −200, −50, and −35 in the scanning order. Hereinafter, the resulting frame is referred to as a “spread frame”. Each spread frame is input to the blocking unit 14.

The address generation unit 13 determines an area corresponding to N blocks corresponding to the number of bits of the digital watermark for each spread frame captured by the blocking unit 14, and positions of pixels included in the area An address indicating is generated. That is, the address generation unit 13 has a function as a pixel position determination unit that determines an area as a pixel position in which digital watermark information is embedded.
Note that the region may be a group of adjacent pixels or a group of randomly extracted pixels.

  Each spread frame is input to the blocking unit 14, and here, based on the information on the pixel position generated by the address generation unit 13, a set of N pixel values corresponding to the number of bits (N) of the digital watermark ( Block).

  The address generation unit 13 described above is configured to generate address information related to pixel positions so that the positions of pixel values included in the pixel value set blocked by the blocking unit 14 are uniformly distributed. Also good. At this time, as long as the pixels are uniformly distributed on the image, the pixels may be arranged randomly and the pixels may be regularly arranged.

  The N blocks extracted by the blocking unit 14 are input to the information embedding unit 15 serving as a pixel value changing unit. On the other hand, the bit generating unit 16 generates a digital value (embedded information) for each bit of the digital watermark information and inputs it to the information embedding unit 15. Then, the information embedding unit 15 changes each pixel value in the corresponding block by adding or subtracting a predetermined displacement in accordance with the digital value for each bit of the digital watermark information.

  The deblocking unit 11 executes processing for returning the information in which the digital watermark is embedded to the image signal, and outputs it as a spread frame sequence in which the digital watermark is embedded.

  Then, the second spreading unit 101 outputs the same spread code string that is output from the deblocking unit 11 as the one accumulated by the first integrating unit 100 in correspondence with each frame constituting the input video. The codes are integrated for each pixel value constituting the spread frame, corresponding to each column in the same order. For example, if the input video is composed of a frame sequence having a size of 2 × 2 pixels for the sake of simplicity, the pixel values of a diffusion frame in which digital watermark information has been embedded are −9, −199, −49, − 34, and the spreading code corresponding to the frame is -1. At this time, pixel values obtained by integrating the spread frame and the spread code are 9, 199, 49, and 34 in the scanning order. The frame sequence as a result of integration is output as a video after information embedding.

  As described above, in this embodiment, a spreading code sequence in the time direction consisting of +1 and −1 is generated, and each of the spreading code sequences is sequentially associated with each frame constituting the input video. Then, the code is added to each pixel value constituting the frame to generate a spread frame, and after embedding the digital watermark information in the spread frame, the same spread code sequence in the time direction as that previously accumulated is again generated. A process of returning the diffusion frame in which the digital watermark is embedded to the original video signal by executing the integration is executed.

  By doing so, the digital video information embedded in the information embedding unit is returned to the signal close to the original video signal with respect to the input video signal that has been converted into the diffusion frame by the integration in the second integration unit 101 in particular. Is diffused by the diffusion processing in the second integration unit 101. In other words, an operation corresponding to the process of diffusing the embedded digital watermark information (integration in the first integration unit 100) is performed on the video signal, the digital watermark information is embedded in the video signal, and then the diffusion is performed. By integrating the code strings again, the video signal is restored and the digital watermark information is diffused.

  Since the digital watermark information is spread spectrum, the digital watermark information itself is less likely to deteriorate even if the video signal is subjected to processing such as encoding or deliberate processing such as analog processing. Digital watermark embedding can be performed.

  Also, since predetermined digital watermark information is embedded over a plurality of frames (diffused in the time direction), it is possible to suppress image quality deterioration of an image after the digital watermark is embedded, and to embed a large amount of bit information as a digital watermark.

  In addition, even if a person who does not know information about the embedding method (embedding area and spreading code string) such as the embedding position tries to erase or tamper with the embedded digital watermark, it is difficult or forcibly erases the image quality. It is possible to execute digital watermark embedding that itself deteriorates significantly. Therefore, it is possible to improve the resistance of the digital watermark information to the operation of the video signal.

  In the above description, the spreading code string is a fixed spreading code string, or a spreading code string may be generated using a secret value as a seed of random numbers for outputting +1 and −1. As another example, in order to eliminate the code bias in the spread code string, for example, a randomly generated spread code string is made to correspond only to an odd frame, and for an even frame, the previous adjacent odd frame It is good also as a code | symbol which changed the code | symbol. An example of this is shown in FIG. Reference numeral 31 denotes a spreading code string in which a randomly generated spreading code string is made to correspond only to odd frames, and an even frame is a code obtained by changing the code of the immediately preceding odd frame.

Embodiment 2. FIG.
In the first embodiment described above, a spreading code string in the time direction composed of +1 and −1 is generated, and each spreading code string is sequentially associated with each frame constituting the input video to form a frame. By adding the code to each pixel value to generate a spread frame, embedding digital watermark information in the spread frame, and then re-adding the same spread code sequence in the time direction as previously accumulated In this embodiment, processing for returning information embedded with a digital watermark to an image signal is executed. Next, in addition to a spreading code sequence in the time direction, a spreading code sequence in the spatial direction is used.
FIG. 4 is a block diagram showing a digital watermark embedding apparatus in such a case. The configuration described in FIG. 1 is the same as that of the spatial direction spreading code generation unit 41, the third integration unit 400, and the fourth integration unit. 401 is added.

Next, the operation will be described.
Similar to the first embodiment, the time direction spreading code generator 12 generates a spreading code string consisting of +1 and −1, and sequentially applies each of the spreading code strings for each frame constituting the input video. Correspondingly, the sign is integrated for each pixel value constituting the frame.
For example, for simplicity, the input video is composed of a frame sequence having a size of 2 × 2 pixels, the pixel values of a certain frame are 10, 200, 50, and 35 in the scanning order, and the spreading code corresponding to the frame is −1. Suppose that At this time, pixel values obtained by integrating the frame and the spreading code are −10, −200, −50, and −35 in the scanning order. Hereinafter, the frame resulting from the integration is referred to as a “time direction spread frame”.

Next, the spatial direction spreading code generation unit 41 generates a spatial direction spreading code sequence consisting of +1 and −1, and sequentially assigns each spatial direction spreading code sequence to each pixel in each time direction spreading frame. The code is integrated for each pixel value constituting the time-direction spread frame. Here, the spatial direction spreading code sequence generated in the spatial direction spreading code generation unit 41 may be generated by a combination of a plurality of spatial direction spreading code sequences.
FIG. 5 shows an example of a spatial direction spreading code sequence finally generated in the spatial direction spreading code generation unit 41 by a combination of a plurality of spatial direction spreading code sequences. In FIG. 5, 51 is a spatial direction spreading code string 1 in which +1 and −1 are alternately set for each pixel, and 52 is a spatial direction spreading code in which +1 and −1 are alternately set in units of 2 × 2 pixels. Column 2 is a final spatial direction spreading code sequence generated by combining the spatial direction spreading code sequence 1 and the spatial direction spreading code sequence 2.

Further, a fixed spreading code string may be used as the spatial spreading code string, or in consideration of security, for example, a spreading code string is used by using a secret value as a seed of random numbers for outputting +1 and −1. It may be generated. As a seed of random numbers in which one spatial direction spreading code sequence outputs +1 and -1 when finally generated in the spatial direction spreading code generation unit 41 by a combination of two spatial direction spreading code sequences in FIG. An example in the case of a spreading code string generated using a secret value is shown.
In FIG. 6, 61 is a spatial spreading code string 1 in which +1 and -1 are alternately and fixedly set for each pixel, and 62 is +1 and -1 alternately and randomly set in units of 2 × 2 pixels. The spatial direction spreading code sequence 3 is a final spatial direction spreading code sequence 63 generated by combining the spatial direction spreading code sequence 1 and the spatial direction spreading code sequence 3.

  For example, for simplicity, the input video is composed of a frame sequence having a size of 2 × 2 pixels, and the pixel values of a certain time direction diffusion frame are −10, −200, −50, and −35 in the scanning order. Assume that the spatial direction spreading codes corresponding to each pixel are +1, -1, +1, -1, respectively. At this time, pixel values obtained by integrating the frame and the spreading code are −10, +200, −50, and +35 in the scanning order.

  Hereinafter, a frame obtained by integrating the spatial direction spreading codes is referred to as a “space / time direction spreading frame”. Each space / time spread frame is input to the blocking unit 14.

The processing from the blocking unit 14 to the deblocking unit 17 is the same as that in the first embodiment. That is, the address generation unit 13 determines the position of the pixel included in the area corresponding to the N blocks corresponding to the number of bits of the digital watermark for each spread frame captured by the blocking unit 14.
The blocking unit 14 extracts N pixel value sets (blocks) corresponding to the number of bits of the digital watermark based on the information regarding the pixel position input from the address generation unit 13.
Further, the address generation unit 13 described above is configured to generate address information related to pixel positions so that the positions of pixel values included in the pixel value set blocked by the blocking unit 14 are uniformly distributed. Also good. At this time, as long as the pixels are uniformly distributed on the image, the pixels may be arranged randomly and the pixels may be regularly arranged.
In the information embedding unit 15, each pixel value in the corresponding block is added or subtracted by a predetermined displacement according to the digital value (embedding information) for each bit of the digital watermark information input from the bit generation unit 16. . The deblocking unit 11 executes processing for returning the information in which the digital watermark is embedded to the image signal, and outputs the information as a space / time diffusion frame sequence in which the digital watermark is embedded.

Next, the same spatial direction spreading code string that has been integrated corresponding to each time direction spreading frame before being input to the blocking unit 14 is embedded with the digital watermark embedded space / time output from the deblocking unit 11. Corresponding to the spread frame sequence in the same order, the code is added to each pixel value constituting the space / time spread frame embedded with the digital watermark.
For example, if the input video is composed of a frame sequence of 2 × 2 pixels for the sake of simplicity, the pixel values of a certain space / time-spread frame embedded with digital watermark information are −9, +201, −49 in the scanning order. , +36, and the spatial direction spreading codes corresponding to the respective pixels of the digital watermark embedded space / time frame are +1, -1, +1, -1, respectively. At this time, pixel values obtained by integrating the digital watermark embedded space / time spread frame and the space spread code are −9, −201, −49, and −36 in the scanning order. The frame sequence obtained as a result of integration is referred to as a “time-spread frame sequence after embedding a digital watermark”.

Then, the same time-direction spreading code sequence as that correspondingly accumulated for each frame constituting the input video is made to correspond to each of the time-spreading frame sequences after embedding the digital watermark in the same order, and the digital watermark embedding is performed. The sign is integrated for each pixel value constituting the post-time spread frame.
In the above example, the pixel values of a time-spread frame after embedding a digital watermark are −9, −201, −49, and −36 in the scanning order, and integration is performed for each frame constituting the corresponding input video. Since the same time direction spreading code is −1, the pixel values obtained by integrating the time spreading frame after embedding the digital watermark and the time direction spreading code are 9, 201, 49, and 36 in the scanning order.

  As described above, in this embodiment, a time-direction spreading code string composed of +1 and −1 is generated, and each time-direction spreading code string is sequentially arranged for each frame constituting the input video. Corresponding to each pixel value constituting the frame to generate a time spreading frame, and then generate a spatial direction spreading code string consisting of +1 and −1. Each space direction spreading code string is associated with each pixel in turn, and the code is added to each pixel value constituting the time direction spreading frame to generate a space / time spreading frame.・ Embedded digital watermark information by embedding digital watermark information in a time-spread frame and then re-adding the same spatial direction spreading code sequence and time direction spreading code sequence as previously accumulated It is obtained so as to perform a process for returning the image signal.

  In this way, the digital watermark information is diffused by the diffusion processing in the second integration unit 101 and the fourth integration unit 401, and processing such as encoding is performed on the video signal. Even if a deliberate process such as an analog process is performed, the digital watermark information itself is not easily deteriorated, so that robust digital watermark embedding can be executed.

  Also, since predetermined digital watermark information is embedded over a plurality of frames (diffused in the time direction), it is possible to suppress image quality deterioration of an image after the digital watermark is embedded, and to embed a large amount of bit information as a digital watermark.

  Also, even if a person who does not know information about the embedding method (embedding area, spreading code string, etc.) such as the embedding position tries to erase or tamper with the embedded digital watermark, it is difficult or forcibly to erase it. It is possible to execute digital watermark embedding in which the image quality itself is significantly deteriorated. Therefore, it is possible to improve the resistance of the digital watermark information to the operation of the video signal.

  Further, as described in the first embodiment, the spreading code string may be generated by using a fixed spreading code string or a secret value as a seed of random numbers for outputting +1 and −1. As shown in FIG. 3, in order to eliminate the code bias in the spread code string, for example, a randomly generated spread code string is made to correspond only to the odd frame, and the even frame is adjacent to the previous one. The code of the odd frame may be changed.

  In any of the embodiments described above, by spreading the video once, embedding the digital watermark information, and accumulating the diffusion pattern again, the digital watermark information embedded when returning the video information from the spread frame is spread. Although the digital watermark information is spread and embedded in the video information, the video signal is processed even if the video signal is subjected to processing such as encoding or intentional processing such as analog processing. It is possible to embed a digital watermark in which watermark information is unlikely to deteriorate.

In the above description, the electronic watermark embedding apparatus has been described. However, the electronic watermark embedding apparatus may be embodied by a program for executing the electronic watermark embedding method according to the present invention using, for example, a general-purpose computer.
That is, the electronic watermark embedding according to the present invention can be executed by causing the computer to execute the electronic watermark embedding program and function as each component shown in FIGS.

Embodiment 3. FIG.
Next, detection of a digital watermark will be described. FIG. 7 is a block diagram showing a digital watermark detection apparatus according to an embodiment of the present invention.
In FIG. 7, reference numeral 71 denotes a digital watermark detection apparatus. The digital watermark detection apparatus 71 includes a time direction spread code generation unit 72, an address generation unit 73, a frame integration unit 74, a blocking unit 75, a detection unit 76, and a fifth integration unit 700.

Next, the operation will be described.
In the following description, it is assumed that the digital watermark information to be detected is digital information including a bit string having a predetermined number of bits, and here, the number of bits is N (N is a positive integer).
First, an input video that is a detection target of digital watermark information is input to the digital watermark detection apparatus 71. Note that, when the digital watermark information is embedded in the low-frequency component video extracted from the original video, the low-frequency component video may be used as the input video.

The time direction spreading code generation unit 72 generates a time direction spreading code string (spreading pattern) composed of +1 and -1. As shown in FIG. 2, each one of the spread code strings is associated with each frame constituting the input video in order. In FIG. 2, 21 is each frame constituting the input video, and 22 is a spreading code corresponding to each frame constituting the input video.
For this spreading code, a fixed spreading code string may be used, or in consideration of security, for example, a spreading code string is generated using a secret value as a seed of a random number for outputting +1 and -1. Also good.

Then, in the fifth accumulating unit 700, each of the spread code sequences is sequentially associated with each frame constituting the input video, and the code is accumulated with respect to each pixel value constituting the frame.
For example, for simplicity, the input video is composed of a frame sequence having a size of 2 × 2 pixels, the pixel values of a certain frame are 10, 200, 50, and 35 in the scanning order, and the spreading code corresponding to the frame is −1. Suppose that At this time, pixel values obtained by integrating the frame and the spreading code are −10, −200, −50, and −35 in the scanning order. Hereinafter, the frame sequence resulting from the integration is referred to as a “spread frame sequence”. The spread frame sequence is input to the frame integration unit 74.

The frame integration unit 74 extracts a specified number of adjacent frames from the input spread frame sequence, adds the values of pixels at the same position in all the extracted spread frames (frame integration), Generate a frame integral image.
Here, when the number of positive and negative codes of the spreading pattern corresponding to the spreading frame that is the subject of frame integration does not match during frame integration, the number of frame integrations may be dynamically increased until the number of codes matches. Good.
At this time, the upper limit of the number of frame integrations to be increased may be determined, and the frame integration may be terminated when the upper limit is reached.
In addition, when the number of positive and negative codes of the spreading pattern corresponding to the spreading frame that is subject to frame integration does not match during frame integration, the number of frame integration may be dynamically decreased until the number of codes matches. .
FIG. 8 shows an embodiment of frame integration when a time direction spreading code string is used. In FIG. 8, 81 is a frame sequence of the input video input to the digital watermark detection device 71, 82 is a spreading code sequence generated by the time direction spreading code generating unit 72, and 83 is for each frame constituting the input video. A spread frame sequence obtained as a result of accumulating the codes for each pixel value constituting the frame by associating each of the spread code sequences in order, and 84 is the value of the pixel at the same position in the spread frame It is the frame integral image obtained as a result of adding together.

  The blocking unit 75 extracts N pixel value sets (blocks) corresponding to the number of bits of the digital watermark based on the information regarding the same pixel position as that at the time of embedding input from the address generation unit 73.

The detection unit 76 examines the statistical value of the pixel value in the corresponding block, and determines the type of the embedded bit. As a feature of the distribution to be examined, for example, paying attention to the average value of the pixel values in the block, if the average value is a certain value δ (> 0) or more, it is determined that the bit “1” is embedded. It is determined that 0 "is embedded.
Here, the reliability of the detection information is calculated from the statistical property of the distribution of the pixel value set in the block. If the reliability is low, the frame integration number is increased until the predetermined reliability is reached. The integration unit 74 may specify the increased frame integration number anew and return the processing.
At this time, the upper limit of the number of frame integrations to be increased may be determined, and the frame integration may be terminated when the upper limit is reached.
In addition, the reliability of the detection information is calculated from the statistical properties of the distribution of the pixel value set in the block. If the reliability is low, the frame integration is performed to reduce the number of frame integrations until a predetermined reliability is reached. The process may be returned by designating the reduced number of frame integrations in the unit 74 again.
As the reliability, for example, dispersion of pixel values can be used. That is, if the variance is smaller than a predetermined value, it is determined that the reliability necessary for detection is satisfied. If the variance is greater than a predetermined value, the reliability necessary for detection is not satisfied. to decide. Standard deviation may be used instead of variance. The detection unit 76 outputs the detection result.

  As described above, in the third embodiment, when detecting N-bit digital watermark information concealed in the input video, a spreading code string consisting of +1 and −1 is generated, and each frame constituting the input video is generated. Each of the spreading code sequences is associated with each of the spreading code sequences in turn, and the code is added to each pixel value constituting the frame to generate a spreading frame sequence, and only the number specified from the spreading frame sequence Extract adjacent frames, add the values of the pixels at the same position in all the extracted diffusion frames to generate one frame integrated image, and use the frame integrated image as information on the same pixel position as at the time of embedding. Based on this, N pixel value sets (blocks) corresponding to the number of bits of the digital watermark are extracted, the statistical values of the pixel values in the extracted block are examined, and the embedded video values are checked. It is obtained so as to determine the door of the kind.

  As a result, even if the video signal is subjected to processing such as encoding or intentional processing such as analog processing, the same digital watermark information as the embedded digital watermark can be detected more accurately. In addition, a large amount of bit information can be detected more accurately.

  Even if there is no original image sequence, it is possible to more accurately detect the same digital watermark information as the embedded digital watermark.

  In addition, even if a person who does not know information about an embedding method such as an embedding position tries to detect a digital watermark improperly, it is possible to perform safe digital watermark detection that makes it difficult.

Embodiment 4 FIG.
In the third embodiment described above, a spreading code sequence composed of +1 and −1 is generated, and each of the spreading code sequences is sequentially associated with each frame constituting the input video, and each frame constituting the frame is configured. A code is added to the pixel value to generate a spread frame sequence, a specified number of adjacent frames are extracted from the spread frame sequence, and the value of the pixel at the same position in all the extracted spread frames is extracted. Add together to generate one frame integrated image, and extract N pixel value sets (blocks) according to the number of bits of the digital watermark from the frame integrated image based on the information about the same pixel position at the time of embedding Then, the statistical value of the pixel value in the extracted block is examined to determine the type of embedded bit. Next, in addition to the spreading code sequence in the time direction, the spatial direction Also the spreading code sequence showing an embodiment in which use.
FIG. 9 is a block diagram showing a digital watermark detection apparatus in such a case, in which a spatial direction spreading code generation unit 91 and a sixth integration unit 900 are added to the configuration described in FIG.

Next, the operation will be described.
Similar to the third embodiment, the time-direction spreading code generator 72 generates a spreading code string consisting of +1 and −1, and sequentially applies each of the spreading code strings for each frame constituting the input video. Correspondingly, the sign is integrated for each pixel value constituting the frame.
For example, for simplicity, the input video is composed of a frame sequence having a size of 2 × 2 pixels, the pixel values of a certain frame are 10, 200, 50, and 35 in the scanning order, and the spreading code corresponding to the frame is −1. Suppose that At this time, pixel values obtained by integrating the frame and the spreading code are −10, −200, −50, and −35 in the scanning order. Hereinafter, the frame sequence obtained as a result of integration is referred to as a “time-direction spread frame sequence”.

Next, the spatial direction spreading code generation unit 91 generates a spatial direction spreading code sequence consisting of +1 and −1, and sequentially assigns each spatial direction spreading code sequence to each pixel in each time direction spreading frame. The code is integrated for each pixel value constituting the time-direction spread frame. Here, on the assumption that the same spreading code sequence as that at the time of embedding is obtained as the time direction spreading code sequence, the spatial direction spreading code sequence generated in the spatial direction spreading code generation unit 91 is a plurality of spatial direction spreading code sequences. It may be generated by a combination.
FIG. 5 described above shows an example of a spatial direction spreading code sequence finally generated in the spatial direction spreading code generation unit 41 by a combination of a plurality of spatial direction spreading code sequences.
Further, a fixed spreading code string may be used as the spatial spreading code string, or in consideration of security, for example, a spreading code string is used by using a secret value as a seed of random numbers for outputting +1 and −1. It may be generated.
When the spatial direction spreading code generator 41 finally generates a combination of two spatial direction spreading code sequences in FIG. 6 described above, one of the spatial direction spreading code sequences outputs random numbers +1 and −1. An example in the case of a spreading code string generated using a secret value as a seed is shown.

  For example, for simplicity, the input video is composed of a frame sequence having a size of 2 × 2 pixels, and the pixel values of a certain time direction diffusion frame are −10, −200, −50, and −35 in the scanning order. Assume that the spatial direction spreading codes corresponding to each pixel are +1, -1, +1, -1, respectively. At this time, pixel values obtained by integrating the frame and the spreading code are −10, +200, −50, and +35 in the scanning order.

  Hereinafter, a frame sequence obtained by integrating the spatial direction spreading codes is referred to as a “space / time direction spreading frame sequence”. This is input to the space / time-spread frame sequence frame integration unit 74.

The processing after the frame integration unit 74 is the same as that of the third embodiment. That is, the frame integration unit 74 extracts a specified number of adjacent frames from the input space / time direction spread frame sequence, and calculates the values of pixels at the same position in all the extracted space / time direction spread frames. Addition (frame integration) One frame integration image is generated.
Here, if the number of positive and negative codes in the time direction spreading pattern corresponding to the space / time direction spreading frame that is the frame integration target during frame integration does not match, the frame integration number is dynamically changed until the number of codes matches. You may make it increase.
At this time, the upper limit of the number of frame integrations to be increased may be determined, and the frame integration may be terminated when the upper limit is reached. Also, when the number of positive and negative codes in the time direction spreading pattern corresponding to the space / time direction spreading frame that is the frame integration target during frame integration does not match, the number of frame integrations is dynamically reduced until the number of codes matches. You may make it make it.
FIG. 10 shows an example of frame integration when the spatial diffusion pattern is applied. In FIG. 10, 81 is a frame sequence of the input video input to the digital watermark detection apparatus 71, 1001 is a spatial direction spreading code sequence generated by the spatial direction spreading code generator 91, and 82 is a time direction spreading code generator 72. The generated spreading code sequence, 1002, corresponds to each pixel value constituting the frame by corresponding one by one to the time direction spreading code sequence and the spatial direction spreading code sequence for each frame constituting the input video. A spatial / temporal spread frame sequence 1003 obtained as a result of integrating the codes, and a frame integrated image 1003 obtained as a result of adding the values of pixels at the same position in the spatial / temporal spread frame.

  The blocking unit 75 extracts N pixel value sets (blocks) corresponding to the number of bits of the digital watermark based on the information regarding the same pixel position as that at the time of embedding input from the address generation unit 73.

The detection unit 76 examines the statistical value of the pixel value in the corresponding block, and determines the type of the embedded bit. As a feature of the distribution to be examined, for example, paying attention to the average value of the pixel values in the block, if the average value is a certain value δ (> 0) or more, it is determined that the bit “1” is embedded. It is determined that 0 "is embedded. Here, the reliability of the detection information is calculated from the statistical property of the distribution of the pixel value set in the block. If the reliability is low, the frame integration number is increased until the predetermined reliability is reached. The integration unit 74 may specify the increased frame integration number anew and return the processing.
At this time, the upper limit of the number of frame integrations to be increased may be determined, and the frame integration may be terminated when the upper limit is reached. In addition, the reliability of the detection information is calculated from the statistical properties of the distribution of the pixel value set in the block. If the reliability is low, the frame integration is performed to reduce the number of frame integrations until a predetermined reliability is reached. The process may be returned by designating the reduced number of frame integrations in the unit 74 again. As the reliability, for example, dispersion of pixel values can be used. That is, if the variance is smaller than a predetermined value, it is determined that the reliability necessary for detection is satisfied. If the variance is greater than a predetermined value, the reliability necessary for detection is not satisfied. to decide. Standard deviation may be used instead of variance. The detection unit 76 outputs the detection result.

  As described above, in the fourth embodiment, when detecting N-bit digital watermark information concealed in the input video, a time-direction spreading code string composed of +1 and −1 is generated to form the input video. Each of the spreading code sequences is associated with each frame in turn for each frame, and the codes are added to each pixel value constituting the frame to generate a time-direction spreading frame sequence, which consists of +1 and -1. A spatial direction spreading code sequence is generated, and each of the spatial direction spreading code sequences is sequentially associated with each pixel in each time direction spreading frame, and for each pixel value constituting the time direction spreading frame, Codes are integrated to generate a spatial / temporal spread frame sequence, and a specified number of adjacent frames are extracted from the spatial / temporal spread frame sequence, and the same in all extracted spread frames A pixel integrated image is generated by adding the values of the pixels in the position, and N pixels corresponding to the number of bits of the digital watermark are generated from the frame integrated image based on the information regarding the same pixel position as that at the time of embedding. A value set (block) is extracted, a statistical value of pixel values in the extracted block is examined, and the type of embedded bit is determined.

  As a result, even if the video signal is subjected to processing such as encoding or intentional processing such as analog processing, the same digital watermark information as the embedded digital watermark can be detected more accurately. In addition, a large amount of bit information can be detected more accurately.

  Even if there is no original image sequence, it is possible to more accurately detect the same digital watermark information as the embedded digital watermark.

  In addition, even if a person who does not know information about an embedding method such as an embedding position tries to detect a digital watermark improperly, it is possible to perform safe digital watermark detection that makes it difficult.

  As described in the third embodiment, the spreading code string may be generated by using a fixed spreading code string or a secret value as a seed of random numbers for outputting +1 and −1. As shown in FIG. 3, in order to eliminate the code bias in the spread code string, for example, a randomly generated spread code string is made to correspond only to the odd frame, and the even frame is adjacent to the previous one. The code of the odd frame may be changed.

In the above description, the digital watermark detection apparatus has been described. However, the digital watermark detection apparatus may be embodied by a program that executes a digital watermark detection method according to the present invention using, for example, a general-purpose computer.
That is, the electronic watermark embedding according to the present invention can be executed by causing the computer to execute the electronic watermark detection program and function as each component shown in FIGS.

1 is a block diagram of a digital watermark embedding device according to an embodiment of the present invention. It is explanatory drawing which shows a response | compatibility with a spreading code sequence and a video frame sequence. It is explanatory drawing which shows the example which eliminates the code | symbol deviation in a spreading code sequence. 1 is a block diagram of a digital watermark embedding device according to an embodiment of the present invention. It is explanatory drawing which shows the example of a production | generation of a spatial direction spreading code sequence. It is explanatory drawing which shows the example of the spatial direction spreading | diffusion code sequence produced | generated using the secret value. 1 is a block diagram of a digital watermark detection apparatus in an embodiment of the present invention. It is explanatory drawing which shows the mode of the frame integration when a time direction spreading code sequence is used. 1 is a block diagram of a digital watermark detection apparatus in an embodiment of the present invention. It is explanatory drawing which shows the mode of the frame integration when a spatial direction spreading | diffusion pattern is applied.

Explanation of symbols

DESCRIPTION OF SYMBOLS 11 Digital watermark embedding apparatus 12 Time direction spreading code generation part 13 Address generation part 14 Blocking part 15 Information embedding part 16 Bit generation part 17 Block release part 21 Each frame which comprises input video 22 Each frame which comprises input video Spread code 31 Spreading code sequence 41 Spatial direction spreading code generator 51 Spatial direction spreading code sequence 52 Spatial direction spreading code sequence 53 Spatial direction spreading code sequence 61 Spatial direction spreading code sequence 62 Spatial direction spreading code sequence 63 Spatial direction spreading code sequence DESCRIPTION OF SYMBOLS 71 Digital watermark detection apparatus 72 Time direction spreading code generation part 73 Address generation part 74 Frame integration part 75 Blocking part 76 Detection part 81 Frame sequence of input image 82 Spreading code string 83 Spreading frame string 84 Frame integration image 91 Spatial direction spreading code Generation unit 100 First integration unit 101 Second integration unit 400 Third integration unit 401 Fourth integration unit 700 Fifth integration unit 900 Sixth integration unit 1001 Spatial direction spread code sequence 1002 Spatial / time direction spread frame sequence 1003 Frame integrated image

Claims (4)

To generate a time-direction spreading code sequence consisting of +1 and -1 is set for each video frame constituting an image, by integrating the time direction spreading code corresponding to each video frame to each pixel value of the corresponding movies Zofu frame An integration step to obtain a time direction spread frame ;
For the time direction spread frame obtained in this integration step, an integration step for obtaining a frame integration image obtained by integrating the adjacent frames by the specified number of frames, and
The number of positive and negative codes when the number of positive and negative codes of the time-direction spreading code sequence generated by the integration step integrated with the specified number of frames in the time-direction spreading frame integrated in the integration step does not match A change step for changing the number of designated frames until the values match,
In the frame integration image obtained in the integration step, a pixel position determination step for determining a pixel position on the frame integration image for detecting digital watermark information embedded in the video ;
And a detection step of detecting the digital watermark information based on a statistical value of a set of pixel values at the pixel position determined in the pixel position determination step . A time direction in which a time direction spreading code sequence composed of +1 and −1 set for each video frame constituting the video is generated and the time direction spreading code corresponding to each video frame is integrated to each pixel value of the corresponding video frame. A first integrating step to obtain a spread frame;
A spatial direction spreading code string composed of +1 and −1 set for each time direction spreading frame obtained in the first integration step is generated, and a time corresponding to the space direction spreading code corresponding to each time direction spreading frame is generated. A second integration step for obtaining a space / time direction diffusion frame integrated with each pixel value of the direction diffusion frame;
An integration step for obtaining a frame integration image obtained by integrating the adjacent frames by a specified number of frames with respect to the spatial / temporal spread frame obtained in the second integration step;
The positive and negative signs of the time-direction spreading code sequence generated by the first integration step, integrated to the time-direction spreading frame corresponding to the space / time-direction spreading frames of the specified number of frames to be integrated in this integration step. A change step of changing the designated frame number until the positive and negative code numbers match when the code numbers do not match;
In the frame integration image obtained in the integration step, a pixel position determination step for determining a pixel position on the frame integration image for detecting digital watermark information embedded in the video;
A detecting step for detecting the digital watermark information based on a statistical value of a pixel value set at the pixel position determined in the pixel position determining step;
An electronic watermark detection method comprising: To generate a time-direction spreading code sequence consisting of +1 and -1 is set for each video frame constituting an image, the time direction obtained by integrating the pixel values of the corresponding video frame and the time direction spreading code corresponding to each video frame An accumulator for obtaining a spread frame ;
For the time-direction spread frame obtained by this integration unit, an integration unit for obtaining a frame integration image obtained by integrating the adjacent frames by the specified number of frames, and
The number of positive and negative codes when the number of positive and negative codes of the time direction spreading code sequence generated by the integration unit integrated with the specified number of time direction spread frames integrated by the integration unit does not match. A change unit for changing the number of designated frames until the two match,
In the frame integration image obtained by the integration unit, a pixel position determination unit that determines a pixel position on the frame integration image that detects digital watermark information embedded in the video ; and
A digital watermark detection apparatus comprising: a detection unit that detects the digital watermark information based on a statistical value of a pixel value set at a pixel position determined by the pixel position determination unit . A time direction in which a time direction spreading code sequence composed of +1 and −1 set for each video frame constituting the video is generated and the time direction spreading code corresponding to each video frame is integrated to each pixel value of the corresponding video frame. A first integrating unit for obtaining a spread frame;
A time direction spreading code sequence consisting of +1 and −1 set for each time direction spreading frame obtained by the first integration unit is generated, and a time corresponding to the space direction spreading code corresponding to each time direction spreading frame is generated. A second integration unit for obtaining a space / time direction diffusion frame integrated with each pixel value of the direction diffusion frame;
An integration unit that obtains a frame integration image obtained by integrating the adjacent frames by a specified number of frames with respect to the spatial / temporal spread frame obtained by the second integration unit;
The positive and negative signs of the time-direction spreading code sequence generated by the first integration unit integrated with the time-direction spreading frame corresponding to the space / time-direction spreading frame of the specified number of frames integrated by the integration unit. A changing unit that changes the designated frame number until the positive and negative code numbers match when the code numbers do not match;
In the frame integration image obtained by the integration unit, a pixel position determination unit that determines a pixel position on the frame integration image that detects digital watermark information embedded in the video; and
A detection unit for detecting the digital watermark information based on a statistical value of a set of pixel values at the pixel position determined by the pixel position determination unit;
An electronic watermark detection apparatus comprising:

Images