JP2005328424A - Digital watermark extractor - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a digital watermark extractor capable of extracting digital watermark data accurately from a superposed video signal even when the degree of deterioration in a recorded video signal is high. <P>SOLUTION: The digital watermark extractor comprises an image quality deterioration model storing section 12, analog/digital converters 3, 5 and 6 for converting a video signal superposed with digital watermark data into digital data, a characteristics correction circuit 7 performing correction processing of the converted digital data based on an image quality deterioration model, and a digital watermark extraction circuit 8 for extracting the digital watermark data from the digital data subjected to correction processing by the characteristics correction circuit 7. <P>COPYRIGHT: (C)2006,JPO&NCIPI

Description

  The present invention relates to a digital watermark extraction apparatus that extracts a digital watermark or a digital watermark candidate from a superimposed video signal obtained by superimposing data such as copyright data on a video signal as a digital watermark.

  In the digital watermark technology, for example, metadata such as copyright data can be superimposed (embedded) in a video signal as minute data (digital watermark data) that cannot be perceived by humans. In the following description, a video signal in which digital watermark data is embedded is referred to as a “superimposed video signal”. Video content is stored after being compressed by MPEG (Motion Picture Experts Group) in the case of DVD (Digital Versatile Disc), and is transmitted after being compressed by MPEG in the case of digital broadcasting. When the compression rate is high in the MPEG system, the decoded video is distorted as compared with the original video, so that the superimposed digital watermark data cannot be held and tends to be lost. Further, when a video signal is dubbed by an analog VTR (video tape recorder) or the like, the distortion of the video is further increased, so that the superimposed digital watermark data tends to be lost more easily.

  As an improvement measure, there is a proposal for setting the signal level when embedding copyright data based on the pixel values of the pixels constituting the frame image (see, for example, Patent Document 1). When recording a video signal with an analog VTR, a reference signal (for example, a signal representing the white peak level of the luminance signal) is recorded by being inserted into the overscan area of the video signal, and the video is based on the reference signal. There is also a proposal for extracting digital watermark data after correcting the luminance level of a signal (see, for example, Patent Document 2).

JP 2004-15626 A (paragraph 0027) JP 11-146342 A (paragraphs 0015-0016)

  However, in the apparatus described in Patent Document 1, when the degree of deterioration of the recorded video signal is large as in the case where the video signal is recorded by the analog VTR, due to distortion of the video signal itself, There is a problem that the embedded digital watermark data cannot be held, and the digital watermark data cannot be extracted accurately.

  Further, in the apparatus described in Patent Document 2, the reference signal embedding area is an extremely narrow area and only a small amount of information can be used for correction. However, when the degree of deterioration of the recorded video signal is large, there is a problem that the embedded data cannot be extracted accurately. Further, the correction of the luminance level in the apparatus described in Patent Document 2 is based on the premise that the reference signal is recorded in the overscan area, and the video in which the reference signal is not recorded in the overscan area. There is a problem that it cannot be applied to signals.

  Furthermore, a measure of avoiding the disappearance of the digital watermark data by increasing the amplitude of the signal superimposed on the video signal is also conceivable. In this case, however, the image quality degradation caused by the digital watermark data becomes significant. There's a problem.

  In addition, in the devices described in Patent Documents 1 and 2, when the digital watermark data is not correctly extracted, only the result that “the extraction could not be performed” is obtained, and there is a slight error in the extraction result. Even in such a case, there is a problem that the extraction result is not utilized and the convenience is lacking.

  Accordingly, the present invention has been made to solve the above-described problems of the prior art, and the object thereof is to convert an electronic image from a superimposed video signal even when the degree of degradation of the recorded superimposed video signal is large. An object of the present invention is to provide an electronic watermark extracting apparatus capable of accurately extracting watermark data.

  Another object of the present invention is to extract approximate data that can be candidates for digital watermark data from a database even when digital watermark data is not correctly extracted from a recorded superimposed video signal. It is an object of the present invention to provide a digital watermark extraction apparatus capable of performing the above.

  An electronic watermark extraction apparatus according to the present invention includes an image quality deterioration model storage unit that holds data relating to image quality deterioration that occurs when a video signal is recorded by each model of the video recording apparatus as an image quality deterioration model for each model of the video recording apparatus; An analog-to-digital conversion unit that converts a superimposed video signal obtained by superimposing digital watermark data on a video signal into digital data, a characteristic correction unit that performs correction processing on the digital data based on the image quality degradation model, and the characteristic correction unit And a digital watermark extracting means for extracting digital watermark data from the digital data subjected to the correction process.

  In addition, another digital watermark extracting apparatus according to the present invention is an image quality degradation model that holds data relating to image quality degradation that occurs when a video signal is recorded by each model of the video recording apparatus as an image quality degradation model for each model of the video recording apparatus. A storage means, an analog-digital conversion means for converting a superimposed video signal obtained by superimposing digital watermark data on a video signal into digital data, and a threshold used when extracting the digital watermark data are set to values based on the image quality degradation model. And a digital watermark extracting means for extracting digital watermark data from the digital data using the threshold set by the threshold control means.

  According to the present invention, the threshold used after digital data is corrected using the image quality degradation model for each model of the video recording apparatus or when digital watermark data is extracted is based on the image quality degradation model. Since the digital watermark data extraction process is executed after setting the value, the digital watermark data, which is a minute signal, can be accurately extracted even when the degree of deterioration of the superimposed video signal is large. Can be obtained.

  In addition, according to the present invention, since it is not necessary to embed digital watermark data with an unnecessarily large amplitude in a video signal, it is possible to obtain an effect that image quality degradation of the video signal due to embedding of the digital watermark data can be suppressed. it can.

  Furthermore, according to the present invention, even if the digital watermark data cannot be correctly extracted, approximate data that is a digital watermark candidate is extracted from the database, so that the user knows the digital watermark candidate. Therefore, the effect that the convenience of the apparatus is improved can be obtained.

Embodiment 1 FIG.
FIG. 1 is a block diagram schematically showing a configuration of a digital watermark extracting apparatus according to Embodiment 1 of the present invention.

  As shown in FIG. 1, the digital watermark extraction apparatus according to Embodiment 1 includes a Y signal input terminal 1 to which a luminance (Y) signal separated from a superimposed video signal on which digital watermark data is superimposed, A C signal input terminal 2 to which a color (C) signal separated from a superimposed video signal on which digital watermark data is superimposed is input, and an A / D (analog) that digitizes the Y signal input from the Y signal input terminal 1 / Digital) converter 3. Also, the digital watermark extraction apparatus according to Embodiment 1 performs quadrature AM (amplitude) demodulation on the C signal input from the C signal input terminal 2, and converts the C signal into an RY signal (red difference signal) and B- Demodulator (Dem) 4 for converting to Y signal (blue difference signal), A / D converter 5 for digitizing RY signal output from demodulator 4, and B− output from demodulator 4 And an A / D converter 6 for digitizing the Y signal.

  Furthermore, the digital watermark extraction apparatus according to the first embodiment records video signals (that is, Y signals, RY signals, BY signals) converted into components by an analog recording device such as a VTR. A characteristic correction circuit 7 that performs characteristic correction (for example, correction of frequency characteristics, linearity, timing, etc.) based on the image quality degradation model when performing the correction, and the Y signal, the RY signal, and the BY signal that have undergone characteristic correction A digital watermark extraction circuit 8 for extracting digital watermark data, a video signal data output terminal 9 for outputting video signal data, and a digital watermark data output terminal 10 for outputting digital watermark data are provided. Furthermore, the digital watermark extracting apparatus according to the first embodiment has image quality that is generated when a video signal is recorded by the control circuit 11 that controls the operation of the entire apparatus and each model of the video recording apparatus (for example, an analog VTR). And a storage unit 12 that holds data relating to deterioration as an image quality deterioration model for each model of the video recording apparatus. The image quality deterioration model held in the storage unit 12 and used by the characteristic correction circuit 7 includes a frequency characteristic deterioration model related to frequency characteristic deterioration caused when a video signal is dubbed to an analog recording apparatus such as a VTR, and a linearity characteristic deterioration of the video signal. At least one of the linearity degradation model related to the above and the timing shift model related to the time shift characteristic generated when the video signal is magnetically recorded.

  FIG. 2 is a block diagram schematically showing the configuration of the characteristic correction circuit 7 shown in FIG. As shown in FIG. 2, the characteristic correction circuit 7 performs a frequency characteristic correction on the input video signal (Y signal, RY signal, BY signal) based on the frequency characteristic deterioration model. A linearity correction circuit 72 that performs linearity correction based on the linearity degradation model, and a timing correction circuit 73 that performs timing correction based on the timing shift model.

  FIG. 3 is a block diagram schematically showing the configuration of the digital watermark extraction circuit 8 shown in FIG. As shown in FIG. 3, the digital watermark extraction circuit 8 includes a digital watermark data embedding position identification circuit 81 for specifying a location where the digital watermark data is embedded in the video signal, and the digital watermark data exceeds the threshold value. A threshold value processing circuit 82 for determining whether or not the digital watermark data is a positive signal or whether the digital watermark data is below a threshold value (if the digital watermark data is a negative signal); A representative value for determining whether the position specified by the embedded position identification circuit 81 is 1 (when the digital watermark data is a positive signal) or 0 (when the digital watermark data is a negative signal). And a sampling circuit 63.

FIG. 4 is a diagram for explaining the frequency correction characteristic of the frequency characteristic correction circuit 71 shown in FIG. The frequency characteristics of the recorded video signal differ between when a certain video signal is dubbed with a VTR of model A and when dubbed with a VTR of model B (model different from model A). In the first embodiment, the frequency response of each model of the analog recording apparatus is stored in the storage unit 12 as a frequency characteristic deterioration model. For example, when the frequency response of a signal recorded on a VTR (actual machine) of a certain model is a characteristic that changes according to the frequency as shown by a solid line (curve F ₁ ) in FIG. The curve (curve F ₁ ) is corrected with a dashed characteristic correction curve (curve C ₁ ) (here, the value of the frequency characteristic curve (curve F ₁ ) and the value of the characteristic correction curve (curve C ₁ ) are multiplied) The frequency characteristic is corrected to a flat characteristic as shown by a one-dot chain line (curve F ₂ ) in FIG. Such a correction is performed by the frequency characteristic correction circuit 71 shown in FIG.

FIG. 5 is a diagram for explaining the input / output characteristics of the linearity correction circuit 72 shown in FIG. The linearity degradation model is, for example, when a video signal having a 8-bit representation and having a range from 0 to 255 is dubbed to an analog recording device such as a VTR and digitized with 10 bits. relationship between the output becomes nonlinearly as shown by the solid line L ₁ in FIG. If it is the relationship shown by the solid line L ₁ in FIG. 5, as the linearity degradation model, the level response of the input and output, a model having a solid characteristic. To compensate for this, the linearity correction circuit 72 shown in FIG. 2, to correct the so-level response shown by the broken line L ₂ in FIG. 5, is converted to 8 bits.

FIG. 6 is a diagram for explaining the input / output characteristics of the timing correction circuit 73 shown in FIG. As shown in FIG. 6, the video signal generally has a horizontal synchronizing signal, and the synchronizing signal is inserted with a negative polarity once in one horizontal scanning period. The horizontal synchronizing signal HS ₁ is defines a horizontal scanning start timing of the video, in the analog recording apparatus such as a VTR, the timing deviation due to the sliding of the magnetic head and the magnetic tape for magnetic recording occurs. That is, as shown in FIG. 6, when a situation occurs in which the immediately preceding horizontal scanning period is slightly extended due to sliding or the like, in the next horizontal scanning period, the entire video signal is There is a phenomenon of shifting slightly backward (lagging behind in time). This is a timing shift model, and in order to correct this, correction is performed so as to shift the time axis within the horizontal blanking interval with reference to the horizontal synchronization signal. The timing correction circuit 73 shown in FIG. 2 performs this correction. In performing this correction, it is desirable to perform sampling conversion by setting the sampling frequency when digitizing to a higher frequency than when digital watermark data is extracted. Here, the frequency characteristic correction is performed before the linearity correction. However, when the linearity correction simply converts the pixel level, it is desirable to perform the correction in this order. The reason is that the pixel level also changes due to the deterioration of the frequency characteristics of the video signal. Therefore, by correcting the deterioration of the frequency characteristics before the linearity correction, the influence of the deterioration of the frequency characteristics is not affected on the linearity correction. It is because it can be made.

  7A and 7B are waveform diagrams for explaining the digital watermark extraction operation of the digital watermark extraction circuit 8 shown in FIG. 7A and 7B, the horizontal axis represents the time axis, and the vertical axis represents the luminance value of the video signal.

In the digital watermark technology, a very small signal (digital watermark data) is added to a video signal so that it is not easily perceived by human eyes, and the digital watermark data is superimposed on the video. In FIGS. 7 (a) and (b), (in FIGS. 7 (a) and (b), the luminance signal) video signals _{S 1} indicated by the solid line in the flat portion _{S 2} of the values of the value of the video signal Compared with the electronic watermark data having a very small amplitude (in FIG. 7A and FIG. 7B, it is depicted as a projecting convex portion or a projecting concave portion within a range surrounded by a dotted circle). Is shown. Where the digital watermark data is added to the video signal is determined in advance by the digital watermark algorithm. In FIGS. 7A and 7B, the video signal is embedded in a flat portion. It is shown. At this time, as shown in FIG. 7A, it is 1 that the digital watermark data is added as a positive signal to the video signal, and as shown in FIG. Is added as a signal of 0. At this time, it is determined from the flat portion S ₂ of the video signal S ₁ with a certain threshold value TH _P or TH _N, and is 1 when the value of the location where the digital watermark data is added exceeds the threshold value TH _P. electronic watermark data is superimposed, it is determined that the electronic watermark data of 0 if below a certain threshold TH _N is superimposed. Therefore, in order to extract the digital watermark data added to the flat portion S ₂ of the video signal S ₁ , the digital watermark data is extracted by distinguishing 1 or 0 using a certain threshold value. The configuration for executing such extraction processing is the digital watermark data extraction circuit 8 shown in FIG.

  In order to determine in which position the watermark data is embedded according to the watermark algorithm, the embedded position identification circuit 14 shown in FIG. 3 identifies the location where the watermark data is embedded in the video signal. To do. The embedding position identification circuit 14 specifies, for example, what line and what sample are embedded in the horizontal synchronization signal. The embedded position of the digital watermark can be easily identified when it is known that the type of the digital watermark data is embedded in, for example, an area used for the character broadcast in the vertical blanking period. If the type of digital watermark algorithm cannot be known, digital watermark extraction is performed for all embedding positions employed in known digital watermark algorithms. On the other hand, the threshold processing circuit 15 shown in FIG. 3 determines whether the digital watermark data exceeds the threshold (when the digital watermark data is a positive signal) or whether the digital watermark data is below the threshold. (When the digital watermark data is a negative signal), it is possible to determine whether the digital watermark data is 1 or 0, and either the 1 or 0 digital watermark data is superimposed on the specified position. Is determined by the representative value sampling circuit 16. Therefore, the digital watermark data output terminal 10 outputs a binary digital signal sequence composed of 0 and 1 by such a continuous extraction.

Since the electronic watermark extracting apparatus according to the first embodiment is constructed as described above, for example, when recording a superimposed image signal W ₁ as a dotted line in FIG. 8 in the analog recording apparatus such as a VTR, the frequency characteristic deteriorated, there is the waveform W ₂ becomes like the solid line, is the peak value of the superimposed data becomes low. On the other hand, the characteristic correction circuit 7 corrects the frequency characteristic, so that the video signal is returned to the waveform W ₁ like a dotted line, so that the likelihood with respect to the threshold can be maintained. That is, the threshold value TH ₁ of dashed line on FIG. 8, it is possible to correctly extract the digital watermark data in the original superimposed video signal W _1, superimposed image signal becomes a waveform W ₂ as a solid line In such a case, the digital watermark data cannot be correctly extracted with the one-dot chain line threshold TH ₁ . In addition, it may be possible to lower the threshold TH _1, unless, can not only wrong extraction know the criteria for determining that the good by lowering the threshold to what extent. As described above, the digital watermark extraction apparatus according to Embodiment 1 can accurately extract digital watermark data that is a minute signal.

  Also, if there is linearity degradation in the video signal level, correct threshold control does not work, and even if the timing is shifted, the same is true. By correcting for linearity degradation and time shift, a minute signal can be obtained. It is possible to accurately extract the electronic watermark data.

  Which image quality degradation model (for example, frequency characteristics, linearity, timing correction, etc.) of the image quality degradation models stored in the storage unit 12 is used can be specified by the model of the VTR used for video signal dubbing. In this case, the image quality degradation model for that model is selected. For example, when information that can identify the model of the VTR used for dubbing the video signal is embedded in an area such as a teletext broadcast in the vertical blanking period, an appropriate image quality degradation model is determined based on this information. Make a selection. If the VTR model used for video signal dubbing cannot be known, all image quality degradation models are used to correct the characteristics, and for example, errors are detected using error correction codes included in the video signal. Adopt what did not exist.

  The simplest method for increasing the detection accuracy rate of digital watermark data without performing such characteristic correction can be avoided by increasing the peak value of the embedded waveform, increasing the amplitude, and embedding. However, if it is embedded in this way, the distortion of the video signal due to the embedding of the digital watermark data will increase, leading to an increase in image quality degradation. By correcting the characteristics based on such an image quality degradation model, there is no need to set a peak value that increases image quality degradation at the time of embedding, so when considering the total image quality of embedding and extraction, high image quality Can be maintained.

In the above description, the case where the digital watermark extraction apparatus according to the first embodiment is realized by an H / W (hardware) configuration has been described. However, the signal processing in the first embodiment is realized by software processing. May be. In that case, the configuration of the digital watermark extracting apparatus is not a circuit but a processing module. For example, in the frequency characteristic correction may be realized digital filter having characteristics correction curve C ₁ in FIG. 4 on software, it is possible to very cheap realization costs.

  Furthermore, the threshold value used in the digital watermark extracting apparatus according to the first embodiment is preferably about ½ of the peak value at the time of embedding, but the noise level of the analog recording device is higher than the peak value. If so, the threshold value may be raised so as not to be buried in the noise level. Further, by devising a digital watermark algorithm and embedding a plurality of digital watermark data, the extraction result of the digital watermark data may be determined by majority determination of the detection results of a plurality of embedded digital watermark data.

  Furthermore, in the first embodiment, the case where the digital watermark data is superimposed on the video baseband signal has been described. However, for example, DCT (discrete cosine transform) is performed to embed the digital watermark data in the frequency conversion data or the like. Even in the case of performing the same, the same effect can be obtained.

  Further, in the first embodiment, the case where the input terminal of the digital watermark extracting apparatus is an S terminal in which the Y signal and the C signal are separated from each other has been described. A similar effect can be obtained even with a configuration having a luminance signal / color signal separation circuit for separating signals.

  In addition, the order of A / D conversion and component signal conversion is not limited to the case of the first embodiment. First, the Y signal and the C signal are digitized by A / D conversion, and then digital processing is performed. , It may be converted into a component signal.

  As a method of using the digital watermark extraction apparatus according to the first embodiment, it is assumed that dubbing is performed using a copy guard canceller (for example, a macrovision canceller) on package media distributed as a DVD or the like. This is a case in which a large amount of illegally copied package media is created and sold in a system as shown in FIG. 9 in which a signal that is controlled so as not to be released is illegally dubbed with a VTR or the like. In such a case, the digital watermark extracting apparatus according to the first embodiment can be used as an apparatus for obtaining such package media and determining whether illegal copying has been performed.

  In addition, according to the digital watermark extraction apparatus according to the first embodiment, the reference signal (for example, the white peak level of the luminance signal) is represented as in the apparatus described in Patent Document 2 (Japanese Patent Laid-Open No. 11-146342). Signal) need not be recorded in the overscan area of the video signal. For this reason, the digital watermark extraction apparatus according to Embodiment 1 is not limited to a special video signal in which a reference signal is recorded by being inserted into an overscan area of the video signal, and can be widely used for general video signals. .

  Note that the digital watermark extraction apparatus according to the first embodiment is assumed to be used in a facility such as a management center that detects illegal copies, and all VTRs sold (distributed) to the world. The digital watermark data is extracted after performing the characteristic correction based on the image quality degradation model of the model.

  In addition, it is assumed that dubbing is performed a plurality of times (for example, twice), and the characteristic deterioration of the combination of the video recording apparatuses that perform the dubbing can be calculated, and after performing the characteristic correction based on the characteristic deterioration of the combination, It is also possible to extract digital watermark data.

Embodiment 2. FIG.
FIG. 10 is a block diagram schematically showing the configuration of the digital watermark extraction apparatus according to Embodiment 2 of the present invention. In FIG. 2, the same or corresponding components as those in FIG. 1 (Embodiment 1) are denoted by the same reference numerals.

  As shown in FIG. 10, the digital watermark extracting apparatus according to Embodiment 2 includes a Y signal input terminal 1 to which a Y signal separated from a superimposed video signal on which digital watermark data is superimposed, and digital watermark data. Are provided with a C signal input terminal 2 to which a C signal separated from a superimposed video signal is input, and an A / D converter 3 for digitizing the Y signal input from the Y signal input terminal 1. . In addition, the digital watermark extracting apparatus according to the second embodiment is a demodulator 4 that performs quadrature AM demodulation on the C signal input from the C signal input terminal 2 and converts the C signal into an RY signal and a BY signal. And an A / D converter 5 for digitizing the RY signal output from the demodulator 4, and an A / D converter 6 for digitizing the BY signal output from the demodulator 4. Yes.

  Furthermore, the digital watermark extraction apparatus according to the second embodiment uses a threshold value used in digital watermark data extraction based on an image quality degradation model (for example, frequency characteristics, linearity, noise, etc.) when analog recording such as VTR is performed. The threshold control circuit 13 to be set, the digital watermark extraction circuit 8 that extracts the digital watermark data from the Y signal, the RY signal, and the BY signal using the threshold set by the threshold control circuit 13, and the video signal data A video signal data output terminal 9 for outputting and a digital watermark data output terminal 10 for outputting digital watermark data are provided. Furthermore, the digital watermark extracting apparatus according to the second embodiment has an image quality generated when a video signal is recorded by the control circuit 14 that controls the operation of the entire apparatus and each model (for example, analog VTR) of the video recording apparatus. And a storage unit 15 that holds data relating to deterioration as an image quality deterioration model for each model of the video recording apparatus.

  FIG. 11 is a block diagram schematically showing the configuration of the threshold control circuit 13 and the digital watermark extraction circuit 8 shown in FIG. As shown in FIG. 11, the threshold control circuit 13 includes a time axis shift information analysis circuit 131, a frequency analysis circuit 132, a noise parameter control circuit 133, and a threshold calculation circuit 134. Also, the digital watermark extraction circuit 8 includes a digital watermark data embedding position identification circuit 81 for specifying a location where the digital watermark data is embedded in the video signal, and whether or not the digital watermark data exceeds a threshold (digital watermark). A threshold value processing circuit 82 for determining whether or not the digital watermark data is below a threshold value (when the digital watermark data is a negative signal) and an embedded position identification circuit 81; A representative value sampling circuit 63 for determining whether the specified position is 1 (when the digital watermark data is a positive signal) or 0 (when the digital watermark data is a negative signal). doing.

  The threshold video control circuit 13 and the digital watermark extraction circuit 8 receive the superimposed video signal decomposed into each component, and the time axis shift information analysis circuit 131 shifts the time axis from the horizontal synchronization signal as shown in FIG. By analyzing the information, for example, analyzing that a signal delay of 20 nsec occurs, the threshold value is also controlled by delaying by 20 nsec, while the representative value sampling circuit determines the digital watermark data by delaying by 20 nsec ( A control instruction is issued to determine whether it is 1 or 0).

The frequency analysis circuit 132 obtains the frequency characteristic from the image quality degradation model. When the frequency characteristic is obtained, the frequency analysis of the video signal is performed, and the result is subjected to the inverse frequency conversion to be on the time axis. Returning to the signal at, it is estimated how the waveform that should have been superimposed is distorted. As a result, if it is estimated that the embedded waveform W ₁ shown by the dotted line in FIG. 12 is as shown by the solid line W ₂ , a control instruction is issued to control the threshold level so that the threshold TH ₁ becomes the lower one-dot chain line TH _2. put out.

  Further, the noise parameter control circuit 133 issues a control instruction to control the threshold level from the amount of noise when the image quality degradation model is recorded by a predetermined analog recording device. In accordance with these threshold control instructions, the threshold value calculation circuit 134 determines the threshold value and gives the threshold value to the threshold value processing circuit 82 of the digital watermark extraction circuit. For example, a delay of 20 nsec is instructed by the time axis shift information analysis circuit 131 and the peak value of the watermark data is estimated to be 1/3 by frequency analysis, and the noise amount is 50 mVp-p by the noise parameter control circuit 133. When a control instruction is received, the threshold value calculation circuit 134 operates so as to generate a threshold value having a peak value of 1/3 or less and 50 mVp-p or more with a delay of 20 nsec.

Since the digital watermark extraction apparatus is configured as described above, for example, when assuming a case in which a superimposed video signal such as a dotted line in FIG. 12 is recorded by an analog recording apparatus such as a VTR, the frequency characteristics are deteriorated. , a waveform shown by the solid line W _2, the peak value of the superimposed data is lowered. On the other hand, since the threshold value operation circuit 134 operates to change the threshold value from the upper one-dot chain line TH ₁ to the lower one-dot chain line TH ₂ , the threshold value for digital watermark extraction can be set correctly. Become. That is, it is possible to correctly extract the digital watermark data in the original superimposed video signal with the threshold value of the one-dot chain line TH ₁ in FIG. 12, but if the waveform becomes a solid line, In one-dot chain line TH ₁ can not be correctly extracted electronic watermark data. However, by performing this threshold calculation, it is possible to accurately extract superimposed data of a digital watermark that is a minute signal.

  Further, since the same is true even if the timing is shifted, by performing correction based on the timing shift, it is possible to accurately extract the superimposition data of the digital watermark that is a minute signal.

  Furthermore, since the noise level is also grasped and the threshold control is performed based on the noise level, the digital watermark data can be correctly extracted without being buried in the noise.

Note that without performing the control of such a threshold, the simplest way to increase the detection accuracy rate of the electronic watermark data by increasing the peak value of the embedded waveform, as shown by a two-dot chain line W ₃ in FIG. 12, the amplitude Although this can be avoided by embedding it, embedding in this way increases the distortion of the video signal caused by embedding the digital watermark data, leading to a greater deterioration in image quality. By performing threshold control based on such an image quality degradation model, there is no need to set a peak value that increases image quality degradation at the time of embedding, so when considering the total image quality of embedding and extraction, high image quality Can be obtained. As in the first embodiment, the accuracy is further improved by performing threshold control that reflects the degradation of linearity in the level of the video signal.

  In the above description, the case where the digital watermark extracting apparatus according to the second embodiment is realized by the H / W configuration has been described. However, the signal processing in the second embodiment may be realized by software processing. In that case, the configuration of the digital watermark extracting apparatus is not a circuit but a processing module. For example, in order to perform the frequency characteristic analysis, it is only necessary to realize FFT (fast Fourier transform), inverse FFT, etc. on software, so that it is possible to realize it with a very low realization cost.

  Further, the order of A / D conversion and conversion into component signals is not limited to the case of the second embodiment. First, the Y signal and the C signal are digitized by A / D conversion, and then by digital processing. , It may be converted into a component signal.

  Further, the threshold value used in the digital watermark extracting apparatus according to the second embodiment is preferably about ½ of the crest value at the time of embedding, but the noise level of the analog recording apparatus is greater than the crest value. If so, the threshold value may be raised so as not to be buried in the noise level. Further, by devising a digital watermark algorithm and embedding a plurality of digital watermark data, the extraction result of the digital watermark data may be determined by majority determination of the detection results of a plurality of embedded digital watermark data.

  Furthermore, in the first embodiment, the case where the digital watermark data is superimposed on the video baseband signal has been described. However, even when the digital watermark data is embedded in the frequency conversion data by performing DCT, for example. The same effect can be obtained.

  Further, in the first embodiment, the case where the input terminal of the digital watermark extracting apparatus is an S terminal in which the Y signal and the C signal are separated from each other has been described. A similar effect can be obtained even with a configuration having a luminance signal / color signal separation circuit for separating signals.

  In the second embodiment, points other than those described above are the same as those in the first embodiment.

  Moreover, the structure which combined some of Embodiment 1 and 2 and Embodiment 3 thru | or 5 demonstrated below is also employable. For example, both the video signal correction by the characteristic correction circuit 7 of the first embodiment and the threshold setting by the threshold control circuit 13 of the second embodiment may be executed, or one of them may be selectively executed.

Embodiment 3 FIG.
FIG. 13 is a diagram illustrating an example of a reference signal used for acquiring parameters of an image quality degradation model in the digital watermark extraction apparatus according to the third embodiment. The digital watermark extraction apparatus according to the third embodiment acquires parameters of the image quality degradation model using the reference signal inserted in the video effective area of the video signal, and corrects the characteristics using the parameters (first embodiment). ) Or threshold setting (in the case of the second embodiment) is different from the digital watermark extracting apparatus according to the first or second embodiment. Except for this point, the digital watermark extracting apparatus according to the third embodiment is the same as the digital watermark extracting apparatus according to the first or second embodiment. Therefore, FIG. 1 (Embodiment 1) or FIG. 10 (Embodiment 2) is referred to in the description of Embodiment 3.

  The digital watermark extracting apparatus of the present invention is assumed to be used in a facility such as a management center for copyright protection that detects illegal copying of video content. Therefore, normally, the digital watermark extraction apparatus according to Embodiments 1 and 2 is premised on holding in advance the image quality degradation models of all models of VTRs sold in the world. However, it is expected that it will be difficult to cover all VTR models released worldwide, considering the continuity of the system. For example, it is assumed that a model Z, which is a new VTR, is released and illegal copying is performed using the model Z. In this case, it is necessary to upgrade the digital watermark extraction device and acquire a new model Z image quality deterioration model, but the new model Z image quality deterioration model cannot be acquired in time, and illegal copying can be performed accurately. There may be cases where it cannot be detected. For this reason, it may be necessary to obtain an image quality degradation model using a digital watermark extraction device that is in circulation for models that must be investigated immediately. The digital watermark extraction apparatus according to the third embodiment inserts a reference signal in the video effective area of the video signal, acquires a parameter of the image quality degradation model using the reference signal, and uses the parameter to Correction or threshold setting is performed, and then digital watermark data is extracted.

  As shown in FIG. 9, a reproduction signal from a DVD player when a DVD package medium is reproduced is input to a copy guard canceller (for example, a macrovision canceller), where a dubbing prevention signal included in the reproduction signal is input. In the case of releasing and dubbing to an analog recording device such as an analog VTR, a DVD disc is prepared so that a reference signal as shown in FIG. 13 can be output as a video signal. When this disc is reproduced by a DVD player, a menu screen is generally reproduced and a screen as shown in FIG. 14 is obtained. This is because the menu button highlights and the viewer jumps to the scene they want to see and enjoys the playback video, or jumps to the special video and sees special video that cannot be seen in the theater etc. It is a button. This menu button is created in the authoring process for producing a DVD disc. Authoring is performed so that a hidden button is prepared so that a normal viewer does not notice when the menu button is changed. When the hidden button is selected, a reference signal as shown in FIG. 13 for constructing an image quality degradation model is output as a video signal. Since DVD uses MPEG compression, a reference signal cannot be input in a non-video period such as a synchronization signal period. For this reason, a reference signal is entered in the video area. In addition, in order to extract frequency characteristics, extract linearity finely, and extract noise levels, a very large number of video areas are required stably, and a reference signal cannot be inserted into a non-video area. Even when the MPEG system is not used, it is desirable to insert the reference signal into the video area.

  In the example of the video signal in FIG. 13, a reference signal for constructing an image quality degradation model related to frequency characteristics is recorded in the upper third of the screen. The reason for using a wide area occupying a third of the screen is that it is necessary to obtain stable characteristics as a video signal. Further, in the next third region of the screen, a reference signal for constructing an image quality degradation model related to the linearity of the video signal is recorded. Further, in the remaining one-third area of the screen, a reference signal for constructing an image quality degradation model related to noise is recorded. The reason for the flat video signal is that an image quality degradation model is constructed with the amount of noise when the flat signal is dubbed.

  By preparing such a reference signal, an image quality deterioration model for a real machine such as a new VTR is constructed at a place where a digital watermark extraction apparatus is installed, parameters are extracted, and the characteristic correction or implementation in the first embodiment is performed. Threshold control in form 2 is performed. Since the digital watermarking apparatus according to Embodiment 3 can acquire an image quality degradation model for an actual machine that does not hold the image quality degradation model in the storage unit, an extremely flexible (highly flexible) digital watermark extraction apparatus is configured. Is possible.

Embodiment 4 FIG.
FIG. 15 is a block diagram schematically showing the configuration of the digital watermark extraction apparatus according to Embodiment 4 of the present invention. In FIG. 15, the same or corresponding components as those in FIG. 1 (Embodiment 1) are denoted by the same reference numerals.

  The digital watermark extraction apparatus according to the fourth embodiment performs error detection of digital watermark data, detects an error detection position where the extracted data is extracted, an approximate code detection circuit 23, copyright data, and the like. The electronic device according to the first embodiment is characterized in that it has a metadata DB storage unit 24 that is a metadata database (DB) and a data output terminal 25 that outputs digital watermark approximation data that can be a candidate for digital watermark data. This is different from the watermark extraction apparatus.

  In the digital watermark extracting apparatus of the first and second embodiments, when the extracted digital watermark data is wrong, only wrong data can be output (or only an output indicating that the data cannot be detected). Can't output data), and some kind of countermeasure is taken against the inconvenient situation where the data such as copyright data is wrongly obtained (or the information about the data such as copyright data cannot be obtained at all). It is not provided. On the other hand, when trying to estimate the copyright data, humans have no choice but to find the copyright data close to it, and in this respect, the apparatus has not been sufficiently fulfilled by the user's request, and is a low convenience device. That is, in Embodiments 1 and 2 described above, it is basically possible to perform accurate extraction by performing characteristic correction or threshold control, but if it is detected by mistake (or If the digital watermark cannot be detected, the copyright data candidate is not searched.

  The digital watermark extraction apparatus according to the fourth embodiment can embed digital watermark data by adding an error correction code to an error in data, but it still cannot be corrected. In many cases, an incorrect position can be detected even if correction is not possible, and an attempt is made to extract an electronic watermark candidate by specifying the incorrect position. For this purpose, the error detection circuit 22 detects an erroneously extracted position. FIG. 16 shows an example of an ID code related to a copyright whose wrong position is known. In FIG. 16, data “0021000012746” is extracted, but when the error detection circuit 22 finds that the part “7” is wrong, the part corresponding to this “7” is assumed. It fills with possible data 0 to F (hexadecimal number) and inquires of the database of the metadata DB storage unit 24 whether or not this code exists, and if it exists, the approximate code detection circuit 23 outputs an approximate code list to the output terminal 25. Output from.

  In the fourth embodiment, even if the digital watermark extraction result is wrong, the candidate list can be obtained, so that it is easy to estimate copyright data, and the convenience of reducing the burden on the user's operation is improved. An electronic watermark extraction apparatus can be obtained.

  In the description of the fourth embodiment, the description of the error detection circuit has been made by handling the one with the error correction code attached. However, as a simpler method, the digital watermark algorithm is devised to convert the digital watermark data. A method of determining the digital watermark data by majority embedding the detection result of the plurality of digital watermark data embedded may be adopted. In this case, when the result of majority decision for determining the digital watermark data is a small difference, the digital watermark candidates may be extracted for each of the plurality of digital watermark data from which higher-order results are obtained. .

  In addition, it is assumed that the input of this digital watermark extracting apparatus is like the S terminal where the Y signal and the C signal are different paths, but the luminance signal that separates the luminance signal and the color signal by inputting the composite signal A similar effect can be obtained even with a configuration having a color signal dividing circuit.

  The fourth embodiment is the same as the first embodiment except for the points described above.

Embodiment 5 FIG.
FIG. 17 is a block diagram schematically showing a configuration of the digital watermark extracting apparatus according to the fifth embodiment of the present invention. In FIG. 17, the same or corresponding components as those in FIG. 2 (Embodiment 2) are denoted by the same reference numerals.

  The digital watermark extracting apparatus according to the fifth embodiment detects an error in digital watermark data, detects an error extraction position of the extracted data, an approximate code detection circuit 23, copyright data, and the like. The electronic device according to the second embodiment is characterized in that it has a metadata DB storage unit 24 that is a metadata database (DB) and a data output terminal 25 that outputs digital watermark approximation data that can be candidates for digital watermark data. This is different from the watermark extraction apparatus.

  In the digital watermark extracting apparatus of the first and second embodiments, when the extracted digital watermark data is wrong, only wrong data can be output (or only an output indicating that the data cannot be detected). Can't output data), and some kind of countermeasure is taken against the inconvenient situation where the data such as copyright data is wrongly obtained (or the information about the data such as copyright data cannot be obtained at all). It is not provided. On the other hand, when trying to estimate the copyright data, humans have no choice but to find the copyright data close to it, and in this respect, the apparatus has not been sufficiently fulfilled by the user's request, and is a low convenience device. That is, in Embodiments 1 and 2 described above, it is basically possible to perform accurate extraction by performing characteristic correction or threshold control, but if it is detected by mistake (or If the digital watermark cannot be detected, the copyright data candidate is not searched.

  In the digital watermark extracting apparatus according to the fifth embodiment, it is generally considered to embed digital watermark data with an error correction code for data errors. In many cases, an incorrect position can be detected even if correction is not possible, and an attempt is made to extract an electronic watermark candidate by specifying the incorrect position. For this purpose, the error detection circuit 22 detects an erroneously extracted position. FIG. 16 shows an example of an ID code related to a copyright whose wrong position is known. In FIG. 16, data “0021000012746” is extracted, but when the error detection circuit 22 finds that the part “7” is wrong, the part corresponding to this “7” is assumed. It fills with possible data 0 to F (hexadecimal number) and inquires of the database of the metadata DB storage unit 24 whether or not this code exists, and if it exists, the approximate code detection circuit 23 outputs an approximate code list to the output terminal 25. Output from.

  In the fifth embodiment, even if the digital watermark extraction result is wrong, the candidate list can be obtained, so that it is easy to estimate copyright data, and the convenience of reducing the burden on the user's operation is improved. An electronic watermark extraction apparatus can be obtained.

  In the description of the fifth embodiment, the description of the error detection circuit has been made by handling the one with the error correction code attached. However, as a simpler method, the digital watermark algorithm is devised to convert the digital watermark data. A method of determining the digital watermark data by majority embedding the detection result of the plurality of digital watermark data embedded may be adopted. In this case, when the result of majority decision for determining the digital watermark data is a small difference, the digital watermark candidates may be extracted for each of the plurality of digital watermark data from which higher-order results are obtained. .

  In addition, it is assumed that the input of this digital watermark extracting apparatus is like the S terminal where the Y signal and the C signal are different paths, but the luminance signal that separates the luminance signal and the color signal by inputting the composite signal A similar effect can be obtained even with a configuration having a color signal dividing circuit.

  In the fifth embodiment, points other than those described above are the same as in the second embodiment.

It is a block diagram which shows roughly the structure of the digital watermark extraction apparatus which concerns on Embodiment 1 of this invention. FIG. 2 is a block diagram schematically showing a configuration of a characteristic correction circuit shown in FIG. 1. FIG. 2 is a block diagram schematically showing a configuration of a digital watermark extraction circuit shown in FIG. 1. FIG. 3 is a diagram for explaining frequency correction characteristics of the frequency characteristic correction circuit shown in FIG. 2. It is a figure for demonstrating the input-output characteristic of the linearity correction circuit shown by FIG. FIG. 3 is a diagram for explaining input / output characteristics of the timing correction circuit shown in FIG. 2. (A) And (b) is a figure for demonstrating the digital watermark extraction operation | movement by the digital watermark extraction circuit shown by FIG. FIG. 3 is a waveform diagram showing a relationship between digital watermark data embedded in a video signal and a threshold value used when detecting the digital watermark data in the first embodiment. It is a block diagram which shows the structure of the system which dubs DVD on which the superimposition video signal was recorded on magnetic tape by VTR. It is a block diagram which shows roughly the structure of the digital watermark extraction apparatus which concerns on Embodiment 2 of this invention. It is a block diagram which shows roughly the structure of the threshold value control circuit and digital watermark extraction circuit which are shown by FIG. In Embodiment 2, it is a wave form diagram which shows the relationship between the digital watermark data embedded in the video signal, and the threshold value used when detecting digital watermark data. It is a figure which shows an example of the reference | standard signal screen and its signal waveform in the digital watermark extraction apparatus which concerns on Embodiment 3 of this invention. FIG. 10 is a diagram illustrating an example of a menu screen for outputting a reference signal in the digital watermark extracting apparatus according to the third embodiment. It is a block diagram which shows roughly the structure of the digital watermark extraction apparatus which concerns on Embodiment 4 of this invention. FIG. 10 is a diagram for explaining an approximate code extraction operation in the digital watermark extraction apparatus according to Embodiments 4 and 5. It is a block diagram which shows roughly the structure of the digital watermark extraction apparatus which concerns on Embodiment 5 of this invention.

Explanation of symbols

  1 Y signal input terminal, 2 C signal input terminal, 3 A / D converter, 4 demodulator, 5, 6 A / D converter, 7 characteristic correction circuit, 8 digital watermark extraction circuit, 9 video signal data output terminal, 10 digital watermark data output terminal, 11, 14 control circuit, 12, 15 storage unit, 13 threshold control circuit, 22 error detection circuit, 23 approximate code detection circuit, 24 metadata DB storage unit, 25 digital watermark approximate data output terminal, 71 Frequency characteristic correction circuit, 72 Linearity correction circuit, 73 Timing correction circuit, 81 Embedding position identification circuit, 82 Threshold processing circuit, 83 Representative value sampling circuit, 131 Time axis shift information analysis circuit, 132 Frequency analysis circuit, 133 Noise parameter control Circuit, 134 threshold value operation circuit.

Claims (6)

Image quality deterioration model storage means for holding data relating to image quality degradation that occurs when a video signal is recorded by each model of the video recording apparatus as an image quality degradation model for each model of the video recording apparatus;
Analog-digital conversion means for converting a superimposed video signal obtained by superimposing digital watermark data on a video signal into digital data;
Characteristic correction means for performing correction processing on the digital data based on the image quality degradation model;
An electronic watermark extraction apparatus comprising: an electronic watermark extraction unit that extracts electronic watermark data from digital data that has been subjected to correction processing by the characteristic correction unit. Image quality deterioration model storage means for holding data relating to image quality degradation that occurs when a video signal is recorded by each model of the video recording apparatus as an image quality degradation model for each model of the video recording apparatus;
Analog-digital conversion means for converting a superimposed video signal obtained by superimposing digital watermark data on a video signal into digital data;
Threshold control means for setting a threshold used when extracting the digital watermark data to a value based on the image quality degradation model;
An electronic watermark extraction apparatus comprising: an electronic watermark extraction unit that extracts electronic watermark data from the digital data using a threshold set by the threshold control unit. Database storage means for holding a database relating to metadata superimposed as digital watermark data on a video signal;
3. The electronic according to claim 1, further comprising: an approximate code detection unit that detects, from the database storage unit, metadata that approximates the digital watermark data extracted by the digital watermark extraction unit. Watermark extraction device. An image quality degradation model extracting means for receiving a video signal obtained by inserting a predetermined reference signal into an effective area of the video and extracting parameters of the image quality degradation model based on the reference signal;
The digital watermark extraction apparatus according to claim 1, wherein the characteristic correction unit performs a correction process on the digital data based on the image quality deterioration model extracted by the image quality deterioration model extraction unit. An image quality degradation model extracting means for receiving a video signal obtained by inserting a predetermined reference signal into an effective area of the video and extracting parameters of the image quality degradation model based on the reference signal;
The electronic device according to claim 2, wherein the threshold control means sets a threshold used when extracting the digital watermark data to a value based on the image quality deterioration model extracted by the image quality deterioration model extraction means. Watermark extraction device.   The image quality degradation model is a model relating to frequency characteristic degradation that occurs when magnetically recording a video signal, a model relating to linearity characteristic degradation of the video signal that occurs when magnetically recording the video signal, and a time shift that occurs when magnetically recording the video signal. 6. The digital watermark extraction according to claim 1, further comprising at least one model selected from a model related to characteristics and a model related to noise characteristic degradation caused when magnetically recording a video signal. apparatus.

Images