JP2003115991A - Image processing apparatus and image processing method - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide an image processing apparatus and an image processing method that can prevent occurrence of overflow without including header information. SOLUTION: The image processing apparatus 10 includes; an embedded error calculation means 110 for calculating an embedded error of a luminance of an image caused by embedding information to the image; a conversion method setting means 120 for setting a conversion method of the luminance of the image on the basis of the embedded error; a luminance conversion means 130 for converting the luminance of the image by the set luminance conversion method; a frequency conversion means 140 for applying frequency-conversion to the image whose luminance is converted; an information embedding means 150 for embedding information to the image subjected to frequency conversion; and a frequency inverse conversion means 160 for applying frequency inverse conversion to the image to which the information is embedded. Converting the luminance of the image before the information is embedded to the image can prevent occurrence of overflow without the need for including header information.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an image processing apparatus and an image processing method for embedding predetermined information in digital data (hereinafter referred to as an image) such as a frame of a still image or a moving image.

[0002]

2. Description of the Related Art Various digital watermarking technologies have been developed as image processing apparatuses for protecting the copyright of images and detecting tampering. In addition, various methods have been proposed as a method of embedding data in the digital watermark technology. One method is to convert the image to the frequency domain, and the conversion coefficient is used as electronic watermark information (hereinafter simply referred to as information).
There is a way to embed. For example, DCT (Discrete Cosine Tra) is used for the frequency domain conversion.
(nsform) and DWT (Discrete Wave)
Let Transform).

However, in the above technique, information may not be accurately restored from an image in which a digital watermark is embedded. This is because the embedding of the digital watermark causes a so-called overflow in which the brightness value (pixel value) after the digital watermark is embedded exceeds a possible value as the brightness value.

A technique for solving this problem is disclosed in the document "Japanese Patent Application Laid-Open No. 2001-24878: Image Processing Apparatus and Method Therefor". In this method, the image is DCT
Information is embedded in the converted coefficient (hereinafter referred to as the conversion coefficient). Information is not embedded in the position where the overflow has occurred, and the position information where the overflow has occurred is held as the header information of the image. To restore the information, the header information is referred to and the position where the information is embedded is specified to correctly restore the information.

[0005]

By the way, the above-mentioned conventional method has a drawback that the amount of information to be held and transmitted increases because header information is required in addition to the image.
Further, since the header information is included, it is easily inferred that some information is embedded in the image, and the image is targeted for decoding. From this point of view, there is a demand for a digital watermark technique that can prevent overflow without including header information.

From the above viewpoints, the object of the present invention is to
An object of the present invention is to provide a new and improved image processing apparatus and image processing method capable of preventing overflow from occurring without including header information.

[0007]

To solve the above problems, according to a first aspect of the present invention, there is provided an image processing apparatus for embedding predetermined information in an image. The image processing device (10) of the present invention includes an image input means (100) for inputting and storing an image to be embedded with information, and an error (embedding error) in the brightness value of the image caused by embedding the information. An embedding error calculating means (110) for calculating the brightness value, a brightness converting means (130) for converting the brightness value of the image, a frequency converting means (140) for frequency converting the image having the brightness value converted, and the frequency. Information embedding means for embedding the information in the converted image (1
50), frequency inverse conversion means (160) for performing frequency inverse conversion on the image in which the information is embedded, and image output means (170) for outputting the image obtained by performing the frequency inverse conversion. Characterize.

Alternatively, the image processing apparatus of the present invention can be configured as follows. That is, the image processing device (10) detects an error (embedding error) in the brightness value of the image caused by embedding the information, and the image input unit (100) that receives and stores the image to be embedded with the information. Embedding error calculating means (110) for calculating
A conversion method setting means (120) for setting a conversion method of the brightness value of the image based on the embedding error, and a brightness conversion means (120) for converting the brightness value of the image by the conversion method of the set brightness value ( 130) and frequency conversion means (140) for frequency-converting the image whose brightness value has been converted.
An information embedding unit (150) for embedding the information in the frequency-converted image, and a frequency inverse transforming unit (160) for inversely transforming the image in which the information is embedded.
And an image output means (170) for outputting an image obtained by performing the frequency inverse conversion.

According to this structure, by converting the brightness value of the image so that the overflow does not occur before the information is embedded, the overflow due to the embedding of the information does not occur, and the embedded information can be correctly restored. Is. Since it is not necessary to retain header information or the like resulting from embedding a digital watermark in the image, it is possible to make it difficult for a malicious tamperer to guess that some information is embedded in the image. Is.

Further, a brightness distribution measuring means (112) for measuring the frequency distribution of the brightness values of the image is further provided, and the conversion method setting means (120) is based on the embedding error and the frequency distribution of the brightness values. Therefore, it is preferable to set the conversion method of the brightness value of the image. With this configuration, it is possible to select a method for converting the brightness value of the image from a plurality of methods based on the measured brightness distribution. Therefore, it is possible to convert the luminance value of an image according to the characteristics of the image in such a manner that deterioration of the image is minimized.

Further, it is preferable to further include overflow control means (165) for judging and controlling whether or not an overflow has occurred in the image obtained by the frequency inverse transform. According to such a configuration, first, the brightness value of the image is converted by a method that reduces the deterioration of the image, and the overflow control means determines whether or not an overflow has occurred in the image obtained from the frequency inverse conversion means. If overflow occurs, the process can be returned to the conversion method setting means. In this way, it is possible to convert the brightness value of the image according to the characteristics of the image in such a manner that deterioration of the image is minimized.

Also, the brightness value of the image in which the information is embedded is converted by a conversion method corresponding to the inverse conversion of the conversion method of the brightness value set by the conversion method setting means (120), and the converted image It is preferable to further include an image display means (180) for displaying. According to this structure, the brightness value of the image in which the predetermined information is embedded is converted by the conversion method corresponding to the inverse conversion of the conversion used by the brightness conversion unit, and the converted image is used as the image for display. , It is possible to visually confirm the image before embedding the predetermined information and the image for display. Apart from the actual calculation result, it is possible to actually confirm the deterioration of the image with human eyes, and it is possible to select the optimum conversion method of the brightness value according to the characteristics of the image.

In order to solve the above problems, according to a second aspect of the present invention, there is provided an image processing device for embedding predetermined information in an image. The image processing device (50) of the present invention includes an image input means (100) for inputting and storing an image to be embedded with information, and an error in luminance value of the image (information embedding error) caused by embedding the information. ) And an error in the brightness value of the image (additional information embedding error) caused by embedding the additional information indicating the embedding position of the information, and the brightness value of the image is converted. Brightness conversion means (130), and additional information determination means (13) for determining the additional information based on the converted brightness value, the information embedding error, and the additional information embedding error.
5), frequency conversion means (140) for frequency-converting the image having the brightness value converted, and additional information embedding means (1) for embedding the additional information in the frequency-converted image.
45), referring to the additional information, and in accordance with the additional information, an additional information reference / information embedding means (155) for embedding the information in the frequency-converted image, and a frequency of the image in which the information is embedded. It is characterized by comprising a frequency inverse conversion means (160) for inverse conversion and an image output means (170) for outputting an image obtained by the frequency inverse conversion.

According to this structure, before embedding information, the position where overflow may occur by embedding the information is specified, and the information is embedded in the image, so that header information other than the image is retained. It is possible to restore information correctly without the need. Furthermore, an image in which information is embedded does not overflow due to information embedding, and it is possible to specify the position of the embedded information from the additional information and restore the correctly embedded information from the image.

In order to solve the above problems, according to a third aspect of the present invention, there is provided an image processing method for embedding predetermined information in an image. The image processing method of the present invention comprises a first step of calculating an error (embedding error) in the brightness value of an image caused by embedding information, and a first step of setting a conversion method of the brightness value of the image based on the embedding error. Two steps, a third step of converting the brightness value of the image by the set brightness value conversion method, a fourth step of frequency-converting the brightness value-converted image, and the frequency-converted image And a fifth step of embedding the information in the image, and a sixth step of frequency-inverse transforming the image in which the information is embedded.

According to such a method, by converting the luminance value of the image so that the overflow does not occur before the information is embedded, the overflow due to the embedding of the information does not occur, and the embedded information can be correctly restored. Is. Since it is not necessary to retain header information or the like resulting from embedding a digital watermark in the image, it is possible to make it difficult for a malicious tamperer to guess that some information is embedded in the image. Is.

In the first step, the frequency distribution of the brightness values of the image is measured, and in the second step, based on the embedding error and the frequency distribution of the brightness values,
It is preferable to set a conversion method of the brightness value of the image. According to this method, it is possible to select, from a plurality of methods, the conversion method of the brightness value of the image from the measured brightness distribution. Therefore, it is possible to convert the luminance value of an image according to the characteristics of the image in such a manner that deterioration of the image is minimized.

After the sixth step, it is preferable to further include a step of determining whether or not an overflow has occurred in the image obtained by the frequency inverse transform. According to such a method, first, the luminance value of the image is converted by a method that reduces the deterioration of the image, and it is determined whether or not an overflow has occurred in the image after the frequency inverse conversion, and the overflow has occurred. In this case, the process of resetting the brightness value conversion method can be performed again. In this way, it is possible to convert the brightness value of the image according to the characteristics of the image in such a manner that deterioration of the image is minimized.

After the sixth step, the brightness value of the image in which the information is embedded is converted by a conversion method corresponding to the inverse conversion of the conversion method of the brightness value set in the second step, and the conversion is performed. It is preferable to further include a step of displaying a later image. According to this method, the brightness value of the image in which the predetermined information is embedded is converted by the conversion method corresponding to the inverse conversion of the conversion used by the brightness conversion means, and the converted image is used as the image for display. , It is possible to visually confirm the image before embedding the predetermined information and the image for display. Apart from the actual calculation result, it is possible to actually confirm the deterioration of the image with human eyes, and it is possible to select the optimum conversion method of the brightness value according to the characteristics of the image.

In order to solve the above problems, according to a fourth aspect of the present invention, there is provided an image processing method for embedding predetermined information in an image. The image processing method of the present invention comprises a first step of calculating an error (information embedding error) in the luminance value of the image caused by embedding information, and the image caused by embedding additional information indicating the embedding position of the information. Of the brightness value of the image (additional information embedding error), a third step of setting a conversion method of the brightness value of the image based on the additional information embedding error, and A fourth step of converting the brightness value of the image by a conversion method, and a fifth step of determining the additional information based on the converted brightness value, the information embedding error, and the additional information embedding error.
A sixth step of frequency-converting the image whose brightness value has been converted, a seventh step of embedding the additional information in the frequency-converted image, and the additional information referring to the additional information, and the frequency in accordance with the additional information. The method is characterized by including an eighth step of embedding the information in the converted image and a ninth step of inversely transforming the image in which the information is embedded.

According to this method, before embedding information, the position where overflow may occur by embedding the information is specified, and the information is embedded in the image to retain header information other than the image. It is possible to restore information correctly without the need. Furthermore, an image in which information is embedded does not overflow due to information embedding, and it is possible to specify the position of the embedded information from the additional information and restore the correctly embedded information from the image.

[0022]

BEST MODE FOR CARRYING OUT THE INVENTION Referring to the accompanying drawings,
A preferred embodiment of an image processing apparatus and an image processing method according to the present invention will be described in detail. In the present specification and the drawings, components having substantially the same functional configuration are designated by the same reference numerals, and duplicate description will be omitted.

(First Embodiment) A digital watermark embedding apparatus for embedding a digital watermark in an image will be described as an embodiment of an image processing apparatus according to the present invention. Figure 1
Is a digital watermark embedding device 10 according to the present exemplary embodiment.
It is explanatory drawing which shows the structure of.

The digital watermark embedding device 10 according to the present embodiment, as shown in FIG.
Embedded error calculation means 110, conversion method setting means 120, luminance conversion means 130, frequency conversion means 14
0, information embedding means 150, frequency inverse transforming means 1
60 and the image output means 170.
The arrow in the figure indicates that the image is processed in that direction.

(Image Input Means 100) Image Input Means 10
In 0, an image to be embedded with predetermined information (digital watermark information) is input, and the input image is stored.

(Embedding Error Detecting Means 110) The embedding error calculating means 110 includes an error in the brightness value of an image generated by embedding predetermined information in the image stored in the image input means 100 (hereinafter referred to as an embedding error). .)
To calculate. As a result, the maximum embedding error is calculated.

(Conversion Method Setting Unit 120) The conversion method setting unit 120 sets the conversion method of the brightness value to be executed by the brightness conversion unit 130, which will be described later, based on the maximum error of the embedding errors calculated by the embedding error calculating unit 110. To do.

(Brightness conversion means 130) Brightness conversion means 13
0 converts the brightness value of the image to be processed by the brightness value conversion method set by the conversion method setting unit 120.

(Frequency conversion unit 140) The frequency conversion unit 140 performs frequency conversion on the image whose brightness value has been converted by the brightness conversion unit 130.

(Information Embedding Unit 150) The information embedding unit 150 embeds predetermined information in the transform coefficient of the image whose frequency has been converted by the frequency converting unit 140.

(Frequency Inverse Transform Unit 160) The frequency inverse transform unit 160 applies the frequency transform unit 140 to the image in which the predetermined information is embedded by the information embedding unit 150.
Inverse frequency conversion corresponding to is performed.

(Image Output Means 170) Image Output Means 17
0 stores the image obtained by the frequency inverse conversion means 160 as an image in which predetermined information is embedded, and outputs it as needed.

The digital watermark embedding device 10 according to this embodiment is configured as described above. Next, a digital watermark embedding method in the digital watermark embedding device 10 will be specifically described with reference to FIGS.

An image to be processed is input and stored in the image input means 100.

In the embedding error calculating means 110, the luminance of the image generated by embedding the predetermined information is calculated from the error generated by embedding the predetermined information in the conversion coefficient converted by the frequency converting means 140 and the conversion formula used in the frequency inverse conversion. Calculate the maximum error in a value. Specifically, the LSB (Least Significant Bi) of the conversion coefficient is
When the predetermined information is embedded by changing t), the error of the conversion coefficient due to the embedding is −1 to +1.
Becomes Furthermore, IDWT (Inver
se DWT), the maximum error of the luminance value of the image can be calculated from the filter coefficient of the wavelet used in IDWT and the above error. In the present embodiment, for example, the calculated maximum error is -α to + β.
Suppose

The conversion method setting means 120 embeds predetermined information on the basis of the maximum error calculated by the embedding error calculation means 110 so that the brightness value of the image does not overflow even if the maximum error occurs, so that the brightness conversion method To decide. For example, as shown in FIG. 2, with respect to the input luminance value before conversion, the output luminance value after conversion is α to (Ma
x-β) is selected. By doing so, with respect to the luminance value after the luminance conversion, maximum −α to + β
Even if the error of occurs, it does not overflow. However, Max is the maximum value that the input image can take, and when each pixel is represented by a non-negative value of 8 bits, Max = 2
55.

The brightness conversion means 130 converts the brightness value of the image to be processed by the method determined by the conversion method setting means 120. In the present embodiment, the input luminance value of each pixel is converted as shown in FIG. 2 to obtain the converted luminance value.

The frequency conversion means 140 performs frequency conversion on the luminance-converted image. In the present embodiment, the DWT is used to perform the frequency conversion by performing the filtering process to obtain the conversion coefficient.

The information embedding means 150 embeds predetermined information in the LSB of the value at the predetermined position of the conversion coefficient obtained by the frequency converting means 140.

The frequency inverse transform means 160 performs frequency inverse transform on the transform coefficient in which the predetermined information is embedded, thereby performing frequency inverse transform and obtaining an image in which the predetermined information is embedded.

The image output means 170 stores the image obtained by the frequency inverse transformation means 160 as an image in which predetermined information is embedded.

As described above, according to the present embodiment, the error caused by embedding the predetermined information is calculated in advance, and the brightness conversion is performed so that the overflow of the brightness value of the image does not occur even if the predetermined information is embedded. By performing the frequency conversion and embedding the predetermined information in the conversion coefficient after performing the above, even if the predetermined information is embedded, an overflow does not occur in the embedded image, and the predetermined information can be correctly restored. Also, header information other than images is not required.

(Second Embodiment) In the first embodiment described above, the input image is converted by a certain conversion formula according to the maximum error of the luminance value of the image caused by embedding the predetermined information. The brightness value of was converted. On the contrary,
In the second embodiment of the present invention described below, the brightness value histogram of the input image to be processed is first measured before setting the brightness value conversion method. The optimum conversion method is selected from among a plurality of conversion methods according to the maximum error of the brightness value of the image generated by embedding the measurement result and the predetermined information, and the brightness value of the input image is converted to a predetermined value. Embed information.

FIG. 3 is an explanatory diagram showing the configuration of the digital watermark embedding device 20 according to the present embodiment. As shown in FIG. 3, the digital watermark embedding device 20 according to the present exemplary embodiment has, in addition to the components of the digital watermark embedding device 10 according to the first exemplary embodiment, a luminance before the conversion method setting means 120. It is characterized in that the distribution measuring means 112 is provided.

(Brightness Distribution Measuring Unit 112) The brightness distribution measuring unit 112 measures the frequency distribution of the brightness values of the image to be processed. Then, the conversion method setting unit 120 performs the brightness conversion performed by the brightness conversion unit 130 based on the maximum error of the embedding error calculated by the embedding error calculating unit 110 and the frequency distribution of the brightness values measured by the brightness distribution measuring unit 112. Set the method.

Regarding the other components, the above-mentioned first
Since it is the same as the embodiment described above, duplicate description will be omitted.

The digital watermark embedding device 20 according to this embodiment is configured as described above. Next, a digital watermark embedding method in the digital watermark embedding device 20 will be specifically described with reference to FIGS.

The operations of the image input means 100 and the embedding error calculation means 110 are the same as those in the first embodiment.

The brightness distribution measuring means 112 obtains a frequency distribution of brightness values for all pixels of the input image to be processed. For example, assume that the frequency distribution shown in FIG. 4 is obtained.

The conversion method setting means 120 sets the brightness conversion method executed by the brightness conversion means 130 based on the maximum error calculated by the embedding error calculation means 110 and the frequency distribution of the brightness values obtained by the brightness distribution measurement means 112. . Specifically, when the ratio of the distribution of brightness values existing in α to (Max-β) is equal to or more than a predetermined threshold value, it is determined that the distribution of brightness values is biased near the center, Instead of the conversion method shown in FIG. 2 in the first embodiment,
The conversion method shown in is used.

According to the conversion method shown in FIG. 5, α to (M
For pixels having a luminance value of ax−β), since it is not necessary to convert the luminance value, the average error between the input luminance value and the output luminance value caused by the luminance conversion can be reduced.

The operation subsequent to the brightness converting means 130 is the same as that of the first embodiment. In this way, an image in which a predetermined digital watermark is embedded can be obtained.

As described above, according to the present embodiment, the luminance distribution of the input image to be processed is measured, and the luminance conversion is performed from the frequency distribution and the error caused by embedding the predetermined information. The luminance conversion method that minimizes the average error between the input luminance value and the output luminance value is selected. In this way, the error due to the luminance conversion can be reduced, and finally the image quality of the image in which the predetermined information is embedded can be improved.

(Third Embodiment) In the above-described first embodiment, the image in which the predetermined information is embedded does not overflow in accordance with the maximum error of the luminance value of the image caused by embedding the predetermined information. Such brightness conversion is performed to embed predetermined information. On the other hand, in the third embodiment of the present invention described below, the maximum error of the brightness value of the image caused by embedding the predetermined information is obtained, and the error caused by actually embedding the predetermined information is It is assumed that the calculated maximum error is smaller than that, and luminance conversion is performed based on the assumed error. After that, frequency conversion, embedding predetermined information, frequency inverse conversion,
An image in which predetermined information is embedded is obtained. At that time, if overflow occurs in the obtained image,
The recursive method is to change the error caused by embedding the predetermined information to a value larger than the assumed error, reset the brightness conversion method, and perform the brightness conversion and the process of embedding the predetermined information again. By processing, predetermined information is embedded.

FIG. 6 is an explanatory diagram showing the configuration of the digital watermark embedding device 30 according to the present embodiment. As shown in FIG. 6, the digital watermark embedding device 30 according to the present exemplary embodiment, in addition to the constituent elements of the digital watermark embedding device 10 according to the first exemplary embodiment, further has an overflow control in front of the image output means 170. It is characterized by including means 165.

In the present embodiment, the conversion method setting means 120 assumes that the error caused by actually embedding the predetermined information is smaller than the maximum error calculated by the embedding error calculating means 110, and the assumption is made. Based on the value, the brightness conversion method implemented by the brightness conversion means 130 is set.

(Overflow Control Means 165) The overflow control means 165 applies to the image obtained by the frequency inverse transform by the frequency inverse transform means 160,
Determine whether an overflow has occurred and control as follows. That is, if no overflow occurs in the image, the process is transferred to the image output means 170, and the obtained image is taken as an image in which predetermined information is embedded. On the other hand, if the image has overflowed, the conversion method setting means 120
Then, the processing is moved to, and the error caused by embedding the predetermined information is changed to a value larger than the previously assumed error, and the luminance conversion method is set.

Regarding the other components, the above-mentioned first
Since it is the same as the embodiment described above, duplicate description will be omitted.

The digital watermark embedding device 30 according to the present embodiment is configured as described above. Next, a digital watermark embedding method in the digital watermark embedding device 30 will be specifically described with reference to FIG.

The operations of the image input means 100 and the embedding error calculation means 110 are the same as those in the first embodiment.

The conversion method setting means 120 estimates the prediction error caused by embedding the predetermined information from the maximum error calculated by the embedding error calculating means 110, and sets the brightness value conversion method. In the present embodiment, for example, the error is set to half of the maximum error calculated by the embedded error calculating means 110. When the maximum error is −α to + β, the prediction error is −α / 2 to + β / 2. Then, based on this prediction error, a conversion method is selected so that the output luminance value becomes α / 2 to (Max−β / 2) with respect to the input luminance value, as in the first embodiment.

The operation from the brightness converting means 130 to the frequency inverse converting means 160 is the same as that of the first embodiment.

The overflow control means 165 determines whether or not the image in which the predetermined information obtained by the frequency inverse conversion means 160 is embedded overflows. If it has not overflowed, the image output means 170
The processing is transferred to, and the obtained image is set as an image in which predetermined information is embedded. If an overflow has occurred, the processing is moved to the conversion method setting means 120, and the prediction error caused by embedding the predetermined information is changed so as to be larger than the prediction error previously predicted, and the brightness conversion is performed. Set the method. For example, the prediction error is -α / 2-1
Reset to + β / 2 + 1.

The processing from the luminance converting means 130 to the frequency inverse converting means 160 is the same as that of the first embodiment, and the image in which the predetermined information obtained by the frequency inverse converting means 160 by the overflow control means 165 is embedded is embedded in the image. Until it does not overflow, the conversion method setting means 120 updates the prediction error and recursively performs the processing.

As described above, according to the present embodiment, the luminance value conversion method is updated and recursively processed until the image obtained by embedding the predetermined information does not overflow. The image quality of the image in which the predetermined information is embedded can be improved as compared with the case where the luminance conversion of the image is performed with the maximum error generated by embedding the predetermined information and the predetermined information is embedded.

(Fourth Embodiment) A fourth embodiment of the present invention will be described. In the present embodiment, when displaying an image in which predetermined information is embedded, in order to display an image having an image quality closer to that of the original image to be processed, the reverse of the luminance conversion performed to embed the predetermined information. It is characterized in that brightness conversion corresponding to conversion is performed on an image in which predetermined information is embedded and the image is displayed.

FIG. 7 is an explanatory diagram showing the configuration of the digital watermark embedding device 40 according to the present embodiment. As shown in FIG. 7, the digital watermark embedding device 20 according to the present exemplary embodiment, in addition to the components of the digital watermark embedding device 10 according to the first exemplary embodiment, further displays an image at a subsequent stage of the image output means 170. It is characterized by including means 180.

(Image display means 180) Image display means 18
0 converts the brightness value of the image in which the information is embedded by a conversion method corresponding to the inverse conversion of the conversion method of the brightness value set by the conversion method setting unit 120, and displays the converted image on the display device. To do. The image display unit 180 itself may be a display device such as a monitor, or the image display unit 18
0 is a display device external to the digital watermark embedding device 40,
You may make it provide the signal for displaying an image.

Regarding the other components, the above-mentioned first
Since it is the same as the embodiment described above, duplicate description will be omitted.

The digital watermark embedding device 40 according to this embodiment is configured as described above. Next, a digital watermark embedding method in the digital watermark embedding device 40 will be specifically described with reference to FIGS. 2 and 7 to 8.

The operation up to the image output means 170 is the same as in the first embodiment.

In the image display means 180, the image output means 1
The image in which the predetermined information obtained from 70 is embedded is subjected to the brightness conversion corresponding to the inverse conversion of the brightness conversion of the image performed by the brightness conversion unit 120, and the image is displayed on the display device.
For example, when the conversion method shown in FIG. 2 is used for the brightness conversion of the image carried out by the brightness conversion means 120, the conversion method shown in FIG. The brightness value of the obtained image in which the predetermined information is embedded is converted. The converted image is displayed on the display device.

As described above, according to the present embodiment, when the image in which the predetermined information is embedded is displayed, the inverse conversion of the brightness conversion performed on the input image to be processed is performed by the predetermined information. By performing this on the image in which is embedded, the error from the original image is reduced, and it is possible to display an image with a quality closer to the original image.

(Fifth Embodiment) In the above-described first embodiment, the image in which the predetermined information is embedded does not overflow in accordance with the maximum error of the brightness value of the image caused by embedding the predetermined information. Such brightness conversion is performed to embed predetermined information. On the other hand, in the fifth embodiment of the present invention described below, additional information indicating position information for embedding predetermined information is obtained and predetermined information is embedded so that overflow does not occur even if predetermined information is embedded. The additional information is embedded in the image before.
Therefore, first, the brightness conversion is performed according to the maximum error caused by embedding the additional information, and then the converted brightness value and
From the maximum error of the brightness value of the image caused by embedding the predetermined information and the maximum error of the brightness value of the image caused by embedding the additional information, the conversion coefficient that may cause overflow by embedding the predetermined information is determined. Ask. When embedding the predetermined information, the additional information is obtained based on the position information of the conversion coefficient so that the predetermined information is not embedded in the conversion coefficient. Before embedding the specified information, the additional information is embedded in the conversion coefficient in advance, and when embedding the specified information, the additional information embedded earlier is referred to and the overflow can occur by embedding the specified information. Predetermined information is not embedded in the effective conversion coefficient, and predetermined information is embedded in the other conversion coefficients as in the first embodiment.

FIG. 9 is an explanatory diagram showing the configuration of the digital watermark embedding device 50 according to the present embodiment. As shown in FIG. 9, the digital watermark embedding device 50 according to the present exemplary embodiment includes an image input unit 100, an embedding error calculating unit 110, a conversion method setting unit 120, a luminance conversion unit 130, and additional information determination. Means 135, frequency conversion means 140, additional information embedding means 145, additional information reference / information embedding means 155, frequency inverse conversion means 1
60 and the image output means 170.

As described above, the digital watermark embedding device 50 according to the present embodiment has the components of the digital watermark embedding device 10 according to the first embodiment, and further,
The additional information deciding means 135 is provided after the luminance converting means 130, the additional information embedding means 145 is provided after the frequency converting means 140, and the information embedding means 150 of the digital watermark embedding device 10 according to the first exemplary embodiment is further provided.
It is characterized in that an additional information reference / information embedding means 155 is provided instead of the above.

(Image Input Means 100) Image Input Means 10
In 0, the image to be embedded with the digital watermark is input, and the input image is stored.

(Embedding Error Detecting Unit 110) The embedding error calculating unit 110 calculates an error (embedding error) in the brightness value of the image generated by embedding a digital watermark in the image stored in the image input unit 100. As a result, the maximum embedding error is calculated. Further, in the present embodiment, the embedding error calculating means 110
Calculates the error in the brightness value of the image (hereinafter referred to as the additional information embedding error) caused by embedding the additional information indicating the information embedding position. As a result, the maximum error of the additional information embedding error is calculated.

Here, the additional information is a flag given based on the position information of the conversion coefficient so that predetermined information is not embedded in the conversion coefficient which may overflow. The additional information will be described later.

(Conversion method setting means 120) The conversion method setting means 120 performs a brightness value conversion method executed by the brightness conversion means 130, which will be described later, based on the maximum error of the additional information embedding error calculated by the embedding error calculation means 110. To set.

(Brightness conversion means 130) Brightness conversion means 13
0 converts the brightness value of the image to be processed by the brightness value conversion method set by the conversion method setting unit 120.

(Additional Information Determining Unit 135) The additional information determining unit 135 determines the maximum error between the embedding error calculated by the embedding error calculating unit 110 and the additional information embedding error for the image subjected to the brightness conversion by the brightness converting unit 130. From
The coordinates of the transform coefficient that may cause overflow by embedding predetermined information are obtained. Then, the additional information is determined from the coordinate information.

(Frequency Conversion Unit 140) The frequency conversion unit 140 performs frequency conversion on the image whose brightness value has been converted by the brightness conversion unit 130.

(Additional information embedding means 145) The additional information embedding means 145 embeds the additional information determined by the additional information determining means 135 in the conversion coefficient of the frequency-converted image.

(Additional information reference / information embedding means 15)
5) The additional information reference / information embedding unit 155 refers to the additional information previously embedded by the additional information embedding unit 145, and in accordance with the additional information, predetermined information for the conversion coefficient of the frequency-converted image. Embed.

(Frequency Inverse Transform Unit 160) The frequency inverse transform unit 160 includes the additional information reference / information embedding unit 155.
Thus, the frequency inverse conversion corresponding to the frequency conversion means 140 is performed on the image in which the predetermined information is embedded.

(Image Output Means 170) Image Output Means 17
0 stores the image obtained by the frequency inverse conversion means 160 as an image in which predetermined information is embedded, and outputs it as needed.

The digital watermark embedding device 50 according to the present embodiment is configured as described above. Next, a digital watermark embedding method in the digital watermark embedding device 50 will be specifically described with reference to FIGS. 9 to 10.

In the image input means 100, an image to be processed is input and stored.

In the embedding error calculating means 110, the luminance value of the image generated by embedding the predetermined information is calculated from the error generated by embedding the predetermined information in the conversion coefficient converted by the frequency converting means 140 and the conversion formula used in the frequency inverse conversion. Compute the maximum error of. Similarly, regarding additional information indicating whether predetermined information can be embedded in a certain transform coefficient, by embedding the additional information from the error generated by embedding the additional information and the conversion formula used in the frequency inverse transform, Calculate the maximum error in the resulting image brightness values.

In this embodiment, it is assumed that information is embedded in the transform coefficient after the wavelet transform as shown in FIG. 10, and 1HL, 1LH, and 1HH (i,
Predetermined information is embedded in j) and additional information is embedded in (i, j) of 1LL.

The conversion method setting unit 120 embeds the additional information based on the maximum error generated by embedding the additional information calculated by the embedding error calculating unit 110, so that the brightness value of the image overflows even if the maximum error occurs. The brightness conversion method is determined so that there is no such change. For example, although the values of α and β are different, the same brightness conversion method as in the first embodiment can be used. The brightness conversion unit 130 converts the brightness value of the image to be processed by the method determined by the conversion method setting unit 120.

Further, the additional information determining means 135 embeds the maximum error and additional information generated by embedding the predetermined information calculated by the embedding error calculating means 110 in the image subjected to the brightness conversion by the brightness converting means 130. From the maximum error caused by this, the pixel in which overflow may occur by embedding predetermined information is determined. Specifically, by adding or subtracting the maximum error caused by embedding the above-mentioned predetermined information in the converted luminance value, the minimum value (α) that the converted luminance value can take
Alternatively, a pixel exceeding the maximum value (Max-β) is determined as the pixel.

Next, from the coordinates of the pixel and the conversion formula used for frequency conversion, predetermined information is embedded in the conversion coefficient, and the value of the conversion coefficient changes, so that the conversion coefficient of the conversion coefficient that affects the brightness value of the pixel is Find the position. That is, this conversion coefficient is a conversion coefficient that may cause overflow by embedding predetermined information in this conversion coefficient. In the present embodiment, the predetermined information is embedded 1
In HL, 1LH, and 1HH, the above-mentioned conversion coefficient is obtained by using the coordinates of the pixel and the filter coefficient of the wavelet used in frequency conversion.

Furthermore, from the information of this conversion coefficient, 1H
If any one of the conversion coefficients of L, 1LH, and 1HH at the same coordinate is the conversion coefficient obtained above, 1LL corresponding to those conversion coefficients is set in order not to embed predetermined information in those conversion coefficients. The additional information to be embedded in the conversion coefficient of is set to "0". Otherwise, since predetermined information can be embedded in those transform coefficients, the additional information to be embedded in the 1LL transform coefficient corresponding to these transform coefficients is set to "1". Corresponding conversion coefficients are each 1L, as indicated by the correspondence shown by the arrow in FIG.
In L to 1HH, when the upper left coordinate is the origin, it means a pixel existing at the same coordinate.

The frequency conversion means 140 performs frequency conversion on the luminance-converted image. In the present embodiment, the DWT is used to perform the frequency conversion by performing the filtering process to obtain the conversion coefficient. The additional information embedding unit 145 embeds the additional information determined by the additional information determining unit 135 into the value at the predetermined position of the conversion coefficient obtained by the frequency converting unit 140. That is, in the present embodiment, “0” or “1”, which means additional information, is embedded in the conversion coefficient of 1LL, for example, LSB.

The additional information reference / information embedding unit 155 refers to the additional information previously embedded by the additional information embedding unit 145, and if the additional information of the corresponding transform coefficient is "1" If it indicates that the predetermined information may be embedded, the predetermined information is embedded; otherwise, the predetermined information is not embedded in the conversion coefficient.

In this embodiment, in 1HL, 1LH, and 1HH for embedding predetermined information, if the LSB of the conversion coefficient of 1LL corresponding to the conversion information is "1", the predetermined information is embedded, and if "0", it is It is determined that the conversion coefficient is not the conversion coefficient for embedding the predetermined information, and the predetermined information is not embedded. When the predetermined information is embedded, for example, the predetermined information is embedded in the LSB of the value of the conversion coefficient obtained by the frequency conversion means 140 at the predetermined position.

The frequency inverse transforming means 160 performs frequency inverse transform on the transform coefficient in which the additional information and the predetermined information are embedded, thereby performing the frequency inverse transform, and the additional information and the predetermined information are embedded. Get the image.

Finally, the image output means 170 stores the image obtained by the frequency inverse transformation means 160 as an image in which additional information and predetermined information are embedded.

As described above, according to the present embodiment, the conversion coefficient that may cause overflow by embedding the predetermined information is obtained, and the predetermined information is not embedded in the conversion coefficient. Therefore, the additional information indicating whether the predetermined information can be embedded is generated, and the additional information is embedded in advance before the predetermined information is embedded. By embedding the predetermined information while referring to the embedded additional information, the predetermined information is not embedded in the transform coefficient that may cause an overflow even if the predetermined information is embedded. The image does not overflow even when embedded. Further, when the predetermined information is restored, it is sufficient to refer to the additional information in order to know in which conversion coefficient the predetermined information is embedded. Therefore, header information other than the image is not necessary.

Further, even if the predetermined information to be embedded is increased and the maximum error is increased by embedding the predetermined information, only the maximum error due to the additional information is considered in the brightness value conversion before the embedding. Since the luminance conversion may be performed, the image quality of the image in which the predetermined information is embedded can be improved.

The preferred embodiments of the image processing apparatus and the image processing method according to the present invention have been described above with reference to the accompanying drawings, but the present invention is not limited to such examples.
It is obvious to those skilled in the art that various changes or modifications can be conceived within the scope of the technical idea described in the claims, and naturally, these are also within the technical scope of the present invention. It is understood that it belongs.

[0104]

As described above, according to the present invention,
By converting the brightness value of the image so that overflow does not occur before embedding the information, the overflow due to the embedding of the information does not occur, and the embedded information can be correctly restored. Since it is not necessary to retain header information or the like resulting from embedding a digital watermark in the image, it is possible to make it difficult for a malicious tamperer to guess that some information is embedded in the image. Is.

[Brief description of drawings]

FIG. 1 is an explanatory diagram showing a configuration of a digital watermark embedding device according to a first embodiment.

FIG. 2 is an explanatory diagram showing an example of luminance conversion according to the first embodiment.

FIG. 3 is an explanatory diagram showing a configuration of a digital watermark embedding device according to a second exemplary embodiment.

FIG. 4 is an explanatory diagram showing an example of a luminance value frequency distribution according to the second embodiment.

FIG. 5 is an explanatory diagram showing an example of luminance conversion in the second embodiment.

FIG. 6 is an explanatory diagram showing a configuration of a digital watermark embedding device according to a third exemplary embodiment.

FIG. 7 is an explanatory diagram showing a configuration of a digital watermark embedding device according to a fourth exemplary embodiment.

FIG. 8 is an explanatory diagram showing an example of image display luminance conversion in the fourth embodiment.

FIG. 9 is an explanatory diagram showing a configuration of a digital watermark embedding device according to a fifth embodiment.

FIG. 10 is an explanatory diagram showing a correspondence example of wavelet transforms and transform coefficients in the fifth embodiment.

[Explanation of symbols]

10, 20, 30, 40, 50 Digital watermark embedding device 100 Image input device 110 Embedding error calculating means 112 Luminance distribution measuring means 120 Conversion method setting means 130 Luminance converting means 135 Additional information determining means 140 Frequency converting means 145 Additional information embedding means 150 information embedding means 155 additional information reference / information embedding means 160 frequency inverse conversion means 165 overflow control means 170 image output means 180 image display means

Claims (11)

[Claims] 1. In an image processing apparatus for embedding predetermined information in an image, an image input means for inputting and storing an image to be embeded information, and a brightness value of the image generated by embedding the information. An embedding error calculating means for calculating an error (embedding error), a luminance converting means for converting a luminance value of the image, a frequency converting means for frequency-converting the image having the luminance value converted, and the frequency-converted image. An information embedding unit for embedding the information in the above, a frequency inverse transforming unit for inversely transforming the image in which the information is embedded, and an image outputting unit for outputting an image obtained by the frequency inverse transforming. An image processing device characterized by the above. 2. An image processing device for embedding predetermined information in an image, embedding error calculating means for calculating an error (embedding error) in a luminance value of the image caused by embedding the information, and the embedding error calculating means. A conversion method setting means for setting a conversion method of the brightness value of the image based on the above, a brightness conversion means for converting the brightness value of the image by the conversion method of the set brightness value, and an image in which the brightness value is converted A frequency conversion means for frequency-converting the information, an information embedding means for embedding the information in the frequency-converted image, and a frequency inverse conversion means for inverse-frequency-converting the image in which the information is embedded. Image processing device. 3. A brightness distribution measuring unit for measuring a frequency distribution of brightness values of the image, wherein the conversion method setting unit determines the brightness of the image based on the embedding error and the frequency distribution of the brightness values. The image processing apparatus according to claim 1, wherein a value conversion method is set. 4. The overflow control means according to claim 1, further comprising overflow control means for determining and controlling whether or not an overflow has occurred in the image obtained by performing the frequency inverse transform. The image processing device according to any one of claims. 5. The brightness value of the image in which the information is embedded is converted by a conversion method corresponding to the inverse conversion of the conversion method of the brightness value set by the conversion method setting means, and the converted image is displayed. The image processing apparatus according to claim 1, further comprising image display means. 6. An image processing apparatus for embedding predetermined information in an image, wherein an image input means for inputting and storing an image to be embedded with information, and a luminance value of the image generated by embedding the information. Embedding error calculation means for calculating an error (information embedding error) and an error in the brightness value of the image (additional information embedding error) caused by embedding additional information indicating the embedding position of the information, and a brightness value of the image A luminance conversion means for converting the luminance value, an additional information determining means for determining the additional information based on the converted luminance value, the information embedding error, and the additional information embedding error, and an image in which the luminance value is converted. Frequency conversion means for frequency-converting, the additional information embedding means for embedding the additional information in the frequency-converted image, and the additional information Referring to the additional information, additional information reference / information embedding means for embedding the information in the frequency-converted image according to the additional information, frequency inverse conversion means for inversely frequency-converting the image in which the information is embedded, and the frequency An image processing apparatus, comprising: an image output unit that outputs an image obtained by performing inverse conversion. 7. An image processing method for embedding predetermined information in an image, the first step of calculating an error (embedding error) in a luminance value of an image caused by embedding information, and the step of calculating the error based on the embedding error. A second step of setting a conversion method of the brightness value of the image, a third step of converting the brightness value of the image by the set conversion method of the brightness value, and a frequency conversion of the image of which the brightness value is converted. Fourth
An image processing method comprising: a step, a fifth step of embedding the information in the frequency-converted image, and a sixth step of inversely frequency-converting the information-embedded image. 8. In the first step, the frequency distribution of the brightness values of the image is measured, and in the second step, the brightness value of the image is calculated based on the embedding error and the frequency distribution of the brightness values. Characterized by setting the conversion method,
The image processing method according to claim 7. 9. The method according to claim 7, further comprising the step of, after the sixth step, determining whether or not an overflow has occurred in the image obtained by the frequency inverse transform. The described image processing method. 10. After the sixth step, the brightness value of the image in which the information is embedded is converted by a conversion method corresponding to the inverse conversion of the conversion method of the brightness value set in the second step, and the conversion is performed. 10. The image processing method according to claim 7, further comprising a step of displaying a later image. 11. In an image processing method for embedding predetermined information in an image, a first step of calculating an error (information embedding error) in a luminance value of the image caused by embedding information, and an embedding position of the information. A second step of calculating an error (additional information embedding error) in the luminance value of the image caused by embedding the additional information indicating, and a method of setting a conversion method of the luminance value of the image based on the additional information embedding error. 3 steps, 4th step of converting the brightness value of the image by the conversion method of the set brightness value, based on the converted brightness value, the information embedding error and the additional information embedding error, A fifth step of determining additional information, a sixth step of frequency-converting the image in which the luminance value has been converted, and a step of embedding the additional information in the frequency-converted image. A seventh step, an eighth step of referring to the additional information and embedding the information in the frequency-converted image according to the additional information, and a ninth step of inversely frequency-converting the image in which the information is embedded An image processing method comprising: