JP4038956B2 - Image generation system and image generation method - Google Patents

Description

[0001]
BACKGROUND OF THE INVENTION
The present invention relates to a technique for concealing an image including information to be recorded, and in particular, can be suitably used for concealing the image by dispersing the image.
In this specification, an image including information to be recorded is referred to as an “embedded image”, an image in which information is embedded is referred to as an “embedded image”, and a plurality of images in which the “embedded image” is distributed is referred to as a “share pattern”. An image in which the “share pattern” is embedded in the “embedded image” is referred to as a “share image”, and a recorded matter obtained based on the “share image” is referred to as a “share”. Further, a group of “share patterns” is called a “share pattern group”, a group of “share images” is called a “share image group”, and a group of “shares” is called a “share group”. In addition, an operation for making an embedded image visible to a human is expressed as “play”.
[0002]
[Prior art]
Conventionally, a technique for concealing an image representing information will be described below with reference to a document describing it.
[0003]
(1) Taku Kato, Hideki Imai: “Extended configuration method of the visual decryption secret sharing method”: IEICE Transactions A Vol. J79-A No. 8 pp. 1344-1351
According to this technique, a plurality of black and white random images are formed, and when these are superimposed, a significant image is reproduced. It is a technique that can disseminate secrets by combining not only two images but also three or more images. If secrets are distributed among multiple images, it is impossible to know any information about other images even if only one of the multiple images is obtained. Therefore, only those that can obtain a plurality of sheets at the same time can reproduce the information, and the secret can be safely distributed and stored.
[0004]
(2) Japanese Patent Laid-Open No. 9-248935
In this technique, another additional image is superimposed and recorded on a certain main image, and the additional image is reproduced by overlapping the image with the pattern image. It is characterized in that information can be recorded in a color image in a form that is difficult for humans to perceive.
[0005]
(3) Koko Matsui: “Digital Watermark”: Journal of the Institute of Image Electronics Engineers of Japan, Vol. 26, No. 3 (1997)
In this technique, a digital watermark is recorded in digital data, and the digital watermark is read by performing a predetermined decoding process. In particular, as a method of putting a digital watermark in a still image, methods such as “pixel space use type”, “frequency domain use type”, and “statistics use type” are known.
[0006]
[Problems to be solved by the invention]
In the technique (1), it is composed only of a black and white random pattern, and a significant image cannot be recorded / reproduced by being superimposed on another image. For example, a significant image in which characters are drawn cannot be superimposed on another landscape image. For this reason, there were problems such as poor expressiveness and appearance.
[0007]
The technique (2) is based on the premise that the main image is mainly a high-redundancy image such as a natural image, and it has been difficult to adapt to a low-redundancy image such as a pattern image in which small images are arranged. In addition, this technology is based on the premise that only two pairs of images are used, and it has not been possible to perform complex reproduction by combining three or more images.
[0008]
In addition, when the technologies (1) and (2) are copied with a copy machine, the same effect as the original one can be obtained. Another common problem is that the recording medium needs to be transparent, and the applicable media is limited.
[0009]
Both the techniques (1) and (2) have the great feature that decoding can be easily performed, but are not actually used commercially because of poor representation and limited adaptive media.
[0010]
The technique (3) is intended for digital data, and requires an arithmetic device such as a computer for encoding / decoding, and information cannot be easily decoded. In addition, since an arithmetic device is required, it has been difficult to reduce the cost.
[0011]
An object of the present invention is to eliminate the drawbacks of the related art and to record various information in a distributed manner and to easily reproduce the information.
[0012]
[Means for Solving the Problems]
In order to achieve the above object in the present invention, first, in the invention of claim 1, A means for inputting an embedded image and an embedded image as a color pattern image as image data, and generating one or a plurality of auxiliary pixels for each pixel of the embedded image, thereby generating a share pattern in which the embedded image is distributed. And means for generating an embedded share image in the embedded image by moving the pixel of the embedded image based on one or more auxiliary pixels corresponding to the share pattern. Image generation system characterized by It is what.
[ 0013 ]
And claims 2 In the invention of The image generation system according to claim 1, The present invention is an image generation system comprising means for embedding additional information in an embedded image.
[ 0014 ]
And claims 3 In the invention, the embedded image is dispersed by inputting the embedded image and the embedded image which is a color pattern image as image data, and by making one or a plurality of auxiliary pixels correspond to each pixel of the embedded image. Generating a share pattern; and generating a share pattern by embedding the share pattern in the embedded image by moving the pixels of the embedded image based on one or more auxiliary pixels corresponding to the share pattern; The image generating method is characterized by including the above.
[ 0015 ]
And claims 4 In the invention of The image generation method according to claim 3, wherein The image generation method includes a step of embedding additional information in an embedded image.
[ 0018 ]
DETAILED DESCRIPTION OF THE INVENTION
(Overall common embodiment)
First, a common embodiment relating to an image generation system, an image generation method, and a recorded material will be described.
[ 0019 ]
FIG. 1 shows a physical configuration of an image generation system according to the present embodiment. This image generation system includes a CPU 101, an image memory 102, an image input unit 103, a program memory 104, and an image recording unit 105, all of which are connected by a bus 106. The CPU 101, the image memory 102, the image input unit 103, and the program memory 104 constitute an image processing unit 107.
[ 0020 ]
The operation of this image generation system will be briefly described. First, the embedded image and the embedded image are written in predetermined areas of the image memory 102 through the image input unit 103. Then, based on the algorithm shown below, these images are subjected to calculation processing to create a share image, and this share image is recorded on a recording medium by the image recording unit 105 or recorded in a form such as printing. Output.
[ 0021 ]
All of these series of processes are performed by the CPU according to the program stored in the program memory 104. A dedicated device may be used for the processing, but a general-purpose computer such as a personal computer may be used. In this case, the image memory 102 and the program memory 104 generally use the same memory in divided areas.
[ 0022 ]
A schematic diagram of the flow of these processes is shown in FIG. In FIG. 19, the embedded image is first decomposed into a share pattern, and the share image is generated by superimposing the share pattern on the embedded image. Then, a share is generated by outputting the share image as a recorded material.
[ 0023 ]
Next, the configuration of the embedded image and the embedded image, which are input images of this system, and the contents and meaning of image processing will be described in detail.
[ 0024 ]
The input image is usually expressed as digital information having a defined density on each grid point in the Cartesian coordinate system in the same manner as used for expression in a computer. Here, the two axes of the orthogonal coordinate system are the x-axis and the y-axis, and are expressed as horizontal and vertical for convenience.
[ 0025 ]
In this embodiment, a gray scale image or a monochrome binary image is basically used as an embedded image, but a color image may be used. The density value of the pixel (x, y) is represented as P (x, y). In the case of a monochrome binary image, black pixels have P (x, y) = 1 and white pixels have P (x, y) = 0. In the share pattern, one or a plurality of auxiliary pixels exist for one pixel of the embedded image as shown in FIG.
In the example of FIG. 2, four auxiliary pixels correspond to one pixel. This auxiliary pixel is a monochrome binary image, and the auxiliary pixel corresponding to the embedded image P (x, y) is represented as Q (x, y, r). Here, r is an auxiliary pixel number and is an integer of 1 or more. A color image or a gray scale image is used as the embedded image.
[ 0026 ]
However, in the present embodiment, more generally, it can be expanded to a binary image or a grayscale image. A binary image means that a black and white pair in a monochrome binary image may be changed to a combination of two colors such as blue and white, red and white, and blue and yellow. In this case, the upper P (x, y) means a pixel of the first color when P (x, y) = 1, and the second color when P (x, y) = 0. It means pixel. For example, when using a pair of blue and yellow, P (x, y) = 1 means a blue pixel, and P (x, y) = 0 means a yellow pixel.
[ 0027 ]
Similarly, in the grayscale image, the density value of the first color at the pixel (x, y) is expressed by P (x, y), and the second value at the pixel (x, y) is expressed by 1-P (x, y). Expresses the density value of the color. For example, when using a blue and yellow pair instead of a white and black pair,
P (x, y) represents a blue density value in the pixel (x, y), and 1-P (x, y) represents a yellow density value in the pixel (x, y).
This gray image is a straightforward extension of the grayscale image. This is because in a grayscale image, P (x, y) can be considered to represent the black density value at pixel (x, y), while 1-P (x, y) is the pixel (x, y). This is because it is considered to express the density value of white at.
[ 0028 ]
When a binary image or a grayscale image is used, the auxiliary pixel is a binary image composed of two colors. For example, in the case of a binary image composed of two colors using a pair of blue and yellow or a grayscale image composed of two colors, the auxiliary pixel is also a binary image composed of a pair of blue and yellow.
[ 0029 ]
In the following description, for the sake of simplicity, description will be given mainly in the case where a monochrome binary image or a grayscale image is used. More generally, a monochrome binary image can be expanded to a binary image, and a grayscale image can be expanded to a grayscale image. .
[ 0030 ]
Next, an image processing algorithm in the present embodiment will be described with reference to the flowchart of FIG.
[ 0031 ]
First step (S1); input of embedded image and embedded image
First, an embedded image and an embedded image are input as digital data. If the image is already digital data, it is converted into an appropriate image format as it is. If the image is a hard copy or a silver halide photograph, it is read by a digital scanner and input as digital data.
[ 0032 ]
Second step (S2): Generation of share pattern
Next, based on the embedded image, a share pattern is generated using, for example, a visual decryption secret sharing method. An embedded image can be reproduced by a combination of share patterns. It is also possible to make a plurality of embedded images by recombination of a pair of share patterns. In generating a share pattern, first, a random number is generated. As a means for generating random numbers, pseudo-random numbers such as M series are used. Then, a combination of share patterns is selected according to the random number. Since the method for selecting a combination differs depending on each embodiment, it will be described in the description of each embodiment.
[ 0033 ]
Third step (S2); image superimposition processing
Next, a share pattern is embedded in the embedded image to generate a final share image. This superimposition process differs depending on each embodiment, and will be described in the description of each embodiment.
[ 0034 ]
Fourth step (S3): Shared image correction processing
Next, in a fourth step, the shared image expressed by the RGB color components is converted into an ink amount signal used for printing as the YMCK four-color ink amount in the image recording unit. This conversion has been widely known as a color conversion technique. If processing is performed on a YMCK basis from the beginning, the processing in the fourth step can be omitted.
Further, this process is not necessary when the shared image is recorded by making a hole in the medium instead of the ink amount signal.
[ 0035 ]
In the above description, a series of share image generation processing is realized by software processing, but can also be realized by hardware. FIG. 4 shows the configuration of an image generation system that implements the series of share image generation described above with hardware.
[ 0036 ]
Two image memories for storing an embedded image and an embedded image, a share pattern generating unit for generating a share pattern, an image superimposing unit for superimposing the image, a correction processing unit for correcting the share image, and an image recording unit for recording the image Are combined to form an image generation system.
[ 0037 ]
First, the embedded image is converted into digital data and stored in the image memory 1. Similarly, the embedded image is also converted into digital data and stored in the image memory 2. Next, a random number is generated and input from the image memory 1 in the share pattern generation unit
A share pattern is generated according to a random number based on the embedded image data. The method for generating the share pattern differs depending on each embodiment, and will be described in the detailed description of each embodiment. In the image superimposing unit, an embedded image is input from the image memory 2 and a share pattern is input from the share pattern generating unit, and these images are superimposed to generate a shared image. Then, the shared image flows to the image correction unit, and the RGB data is converted into YMCK data. Finally, the share image is recorded on the medium in the form of a hard copy in the image recording unit, thereby outputting the share.
[ 0038 ]
As described above, the image processing can be easily realized by hardware. When hardware is used, it is possible to perform image processing at high speed, which is particularly effective when a large amount of share needs to be generated.
[ 0039 ]
Next, a reproduction apparatus that reproduces an embedded image from a share in which the share image is recorded on the medium as described above will be specifically described.
[ 0040 ]
FIG. 5 is a diagram illustrating a configuration example of a playback device. Load the share into the playback device and fix it. Thereby, the shares are fixed to a predetermined positional relationship. By observing this fixed share, the embedded image can be synthesized and recognized from the share. Further, by arranging a light source such as a light bulb below the reproduction device, reproduction can be made clearer. Note that the playback device is not limited to the configuration shown in FIG. 5 and may have any structure as long as the share positional relationship can be fixed.
[ 0041 ]
Next, the principle that the embedded image is reproduced by the reproduction apparatus will be described by taking a simple example. As an example, a case will be described in which a monochrome binary image is used as an embedded image and a plain image is used as an embedded image.
[ 0042 ]
When the shared images recorded at the time of reproduction are superimposed, the embedded images are visually seen in the form of overlapping. In this example, the generated share image is also a monochrome binary image. Then, depending on the degree of overlap of the shares of each pixel, the pixel looks white or black to the human eye as shown in FIG. Therefore, by configuring the share pattern so that the black pixels of the embedded image appear black during playback, and configuring the share pattern so that the white pixels of the embedded image appear white during playback, the embedded image is It becomes possible to play with. At this time, it is important that the contrast ratio between the two portions be sufficiently large so that the human eye can easily distinguish between the black portion and the white portion. In this example, a black and white pair has been described, but the same effect can be obtained with other color combinations, for example, blue and white, red and white, and blue and yellow. In the following description, a black and white pair will basically be described, but other color combinations are possible.
[ 0043 ]
In addition, as a device that facilitates playback, a mark for positioning as shown in FIG. 7 is added to the medium to facilitate positioning during playback and weak adhesion to one side of the share. It is conceivable to apply the material in advance so that the position is not easily shifted after being overlapped.
In addition, as shown in FIG. 8, the conventional arrangement in which the pixels are connected to each other is arranged so that the pixels are separated from each other so that they are not easily affected by camera shake or the like. It is valid.
[ 0044 ]
Next, specific processing will be described according to each embodiment.
[ 0045 ]
(First embodiment)
First, the first embodiment will be described. In this embodiment, a gray scale image is used as the embedded image. A color image is used as the embedded image. In the present embodiment, since a color image is used as an embedded image, in the example of FIG. 6, during reproduction, the black portion of FIG. 6 is colored and the white portion is colorless or the same color. Visually, a colorless or identical color portion is recognized as distinguished from other portions, and an embedded image can be recognized.
[ 0046 ]
Next, an image processing algorithm in the present embodiment will be described in detail along the flowchart of FIG.
[ 0047 ]
First step (S1); (input of embedded image and embedded image)
An embedded image and an embedded image are input as digital data. In the present embodiment, the embedded image is input as a gray scale image and the embedded image is input as a color image. The embedded image is decomposed into auxiliary pixels simultaneously with the input. Generally, the number of auxiliary pixels increases as the gradation becomes finer. By increasing the number of gradations, it is possible to express a smoother image, but as the number of auxiliary pixels increases, the size of the auxiliary pixels becomes smaller and alignment during reproduction becomes difficult. Therefore, it is important to reduce the number of gradations to such an extent that the quality of the image is not deteriorated and to balance the size of the auxiliary pixels. As a gradation adjustment method, a method such as “error diffusion method” is known.
[ 0048 ]
Second step (S2); generation of share pattern
Next, based on the embedded image whose tone has been corrected, a share pattern is generated using, for example, a visual decryption type secret sharing method. An embedded image can be reproduced by a combination of share patterns. It is also possible to make a plurality of embedded images by recombination of a pair of share patterns. In generating a share pattern, first, a random number is generated. Then, a combination of share patterns is selected according to the random number. Hereinafter, the combination of share patterns is referred to as “set”. As in the example of FIG. 9, when there are four combinations of share patterns, random numbers 1 to 4 are generated, and a set is selected according to the generated random numbers. As shown in FIG. 9, if two patterns are included in one set, the embedded image is distributed into two share patterns. The set is selected from a white set when the pixel of the corresponding embedded image is white, and from a black set when the pixel is black. This is performed for each pixel, and a share pattern is generated for the entire screen. In the present embodiment, the embedded image is a grayscale image that has been subjected to gradation correction, and the number of auxiliary pixels is determined according to the number of gradations. Of FIG.
An example is the case of white and black, where the number of gradations is 2, and the number of sets increases as the number of gradations increases. In this step, the embedded image is distributed into a plurality of share patterns.
[ 0049 ]
Third step (S3); image superimposition processing
Next, in the image superimposing process, a share pattern is embedded in the embedded image to generate a final share image. The embedding is performed by leaving the pixels of the embedded image corresponding to the black portions of the auxiliary pixels of the share pattern and deleting the pixels corresponding to the white portions of the auxiliary pixels of the share pattern to make it transparent. At this time, if more embedded images remain, the final share looks better. Therefore, if the share pattern is devised so that the black portion of the pattern increases as much as possible in the second step, the appearance is improved. In addition, if a specific color is applied instead of the pixel corresponding to the white portion of the share pattern, the embedded image is reproduced more clearly, although the appearance of the share is somewhat impaired.
[ 0050 ]
Fourth step (S4); share image correction processing
Next, in a fourth step, the shared image expressed by the RGB color components is converted into YMCK four-color ink amounts by the image recording unit.
[ 0051 ]
The above is a series of image processing in the present embodiment, and after this processing, the image recording unit outputs the share by recording the share image as a recorded material such as a hard copy in accordance with the ink amount signal.
[ 0052 ]
The reproduction is performed by using the reproduction apparatus shown in FIG. 5 or by combining each share with a human hand.
[0053]
(Second Embodiment)
Next, the image generation system according to claim 1 And the image generation method of claim 3 A second embodiment will be described. In the present embodiment, a pattern image in which small images and characters are arranged is used as an embedded image.
[ 0054 ]
In the present embodiment, a monochrome binary image is used as an embedded image, and a pattern image in which small images and characters are arranged is used as an embedded image. Since the pattern image has low image redundancy, it is difficult to clearly reproduce the embedded image by the method of the first embodiment. Hereinafter, small images and characters are referred to as small patterns.
[ 0055 ]
The properties of the shared image recorded in this embodiment will be described. The generated share image is a pattern image in which small images and characters are arranged, and is observed differently by human eyes as in the case of the logo in FIG. When the small pattern fits perfectly, the part looks whitish, and when the small pattern is shifted and joined, the part looks dark. Padding
The embedded image can be observed by configuring the white portion of the embedded image so that the pattern is exactly the same and the black portion so that the pattern is shifted.
[ 0056 ]
Next, the image processing algorithm in this embodiment will be described in detail with reference to the flowchart of FIG.
[ 0057 ]
First step (S1); input of embedded image and embedded image
An embedded image and an embedded image are input as digital data. In this embodiment, the embedded image is input as a monochrome binary image, and a small pattern is actually input as a color image as an embedded image. The embedded image is interpreted by arranging the small patterns vertically and horizontally.
[ 0058 ]
Second step (S2); generation of share pattern
Similar to the first embodiment, a share pattern is generated based on the embedded image. In the example of the logo of FIG. 10, the number of auxiliary pixels is one for each pixel of the embedded image, and the share pattern is expressed by two types of logos on the right side or the left side. Therefore, the share pattern is represented by binary values of black representing the right side and white representing the left side. Of course, there are four types of logos, upper left, upper right, lower left, and lower right, and the number of auxiliary pixels can be set to two.
[ 0059 ]
Third step (S3); image superimposition processing
Next, a small pattern is arranged in accordance with the share pattern to generate a final share image. That is, when there are N sets of share pattern combinations, one set of N sets is written to each share according to the share pattern. This is performed for each pixel, and the entire image screen is processed to generate a shared image. In the case of the logo in FIG. 10, the share patterns are two sets of black representing the right side and white representing the left side. A small pattern is arranged on the right side in a black portion of the share pattern, and a small pattern is arranged on the left side in a portion where the share pattern is white.
[ 0060 ]
Fourth step (S4); share image correction processing
As in the first embodiment, RGB data is converted into YMCK ink amounts.
[ 0061 ]
The above is a series of image processing in the present embodiment. After this processing, the share is output by recording the share image as a recorded material such as a hard copy in the image recording unit as in the first embodiment.
[ 0062 ]
The reproduction method is performed in the same manner as in the first embodiment.
[ 0063 ]
(Third embodiment)
Next, a third embodiment will be described. In this embodiment, instead of recording the shared image as a hard copy, a hole is made in the medium.
[ 0064 ]
In this embodiment, a monochrome binary image is used as an embedded image, and an embedded image is not used. However, it is possible to improve the appearance by printing an arbitrary image on the medium for punching holes. In addition, it is possible to use a grayscale image as an embedded image, but since the number of auxiliary pixels increases, the size of the auxiliary pixels becomes smaller, and it becomes difficult to maintain the strength when a hole is made, Problems such as it becomes difficult to match the shares during playback. In the embodiments described above, a transparent sheet is mainly used as a medium. However, in the case of this embodiment, an opaque sheet can be used.
[ 0065 ]
In the present embodiment, the share generated and recorded from the share pattern is perforated along the pattern. Patterns are observed by the human eye depending on the degree of overlapping of the holes in each pattern. When the pattern fits perfectly, the part appears to have a hole, and when the pattern fits out of alignment, the part appears to be filled with a hole. By embedding the white portion of the embedded image so that the pattern exactly matches the black portion and shifting the pattern, the embedded image can be observed.
[ 0066 ]
Next, an image processing algorithm in the present embodiment will be described in detail with reference to the flowchart of FIG.
[ 0067 ]
First step (S1); input of embedded image
Read an embedded image. The embedded image is input as a monochrome binary image. The embedded image is not used in this embodiment.
[ 0068 ]
Second step (S2); generation of share pattern
Similar to the first embodiment, a share pattern is generated based on an embedded image using, for example, a visual decryption secret sharing method.
[ 0069 ]
In the present embodiment, the third step (image superimposition processing) and the fourth step (share image correction processing) are unnecessary.
[ 0070 ]
The above is a series of image processing in the present embodiment, and after this processing, the image recording unit outputs holes by punching holes in the medium according to the share pattern and recording.
[ 0071 ]
Next, a method for making a hole in the medium will be described. Select either the black or white pattern of the share pattern, and make a hole of an appropriate size in the pattern part. As shown in the medium 1 with a hole in FIG. 12, when making a hole, a hole is provided with a frame. This is because if the hole is made to the full size of the pixel of the pattern without providing a frame, the black pixel in the middle will fall out in the case of the pattern in the medium 2 with the hole in FIG. . Further, if the corner portion is left as shown in the medium 3 with the holes shown in FIG. 12, the strength of the frame can be sufficiently maintained. A device for making a hole in a medium can also make a fine hole by making a hole with a puncher or using an etching technique.
[ 0072 ]
Also in this embodiment, the embedded information is reproduced by the same method as in the first embodiment.
[ 0073 ]
Conventionally, only a method by hard copy to a medium has been used, but according to the present embodiment, it is possible to adapt to an opaque medium and to expand the application range.
[0074]
(Fourth embodiment)
Next, the image system according to claim 2 And the image generation method of claim 4 A related fourth embodiment will be described. This method makes it possible to embed arbitrary digital information in addition to information related to an embedded image in a share pattern. Hereinafter, digital information to be embedded separately from information related to the embedded image is referred to as “additional information”.
[ 0075 ]
This embodiment can be roughly classified into two types. There are two types: one that embeds additional information as a black and white two-dimensional code and one that embeds additional information as a digital watermark. When embedding additional information as a two-dimensional code, there are two types: one that uses a redundant share pattern and one that further divides auxiliary pixels and embeds a two-dimensional code.
[ 0076 ]
(When 4-a additional information is embedded as a two-dimensional code)
When the additional information is embedded as a black and white two-dimensional code, a monochrome binary image is used as an embedded image, and an embedded image is not used. In this case, it is possible to use an arbitrary image instead of the monochrome binary image as the embedded image, but it becomes difficult to reproduce the additional information by the apparatus.
[ 0077 ]
First, an image processing algorithm when a redundant share pattern is used will be described in detail with reference to the flowchart of FIG.
[ 0078 ]
First step (S1); input of embedded image
Enter an embedded image. Read an embedded image. The embedded image is input as a monochrome binary image. The embedded image is not used in this embodiment. On the other hand, the additional information is digitized and recorded in a memory separately from the image.
[ 0079 ]
Second step (S2); share pattern generation and image superimposition processing
When a redundant share pattern is used, share pattern generation and image superimposition processing are performed simultaneously. That is, the share pattern becomes the final share image as it is. The generation of the share pattern using the visual decryption secret sharing method is performed in two stages: preprocessing of additional information and selection of a set. However, when a redundant share pattern is used, there are two types of sets, which are called “set” and “small set”, respectively. FIG. 14 shows an example in which the number of sets is 3 and the number of small sets is 2. In the first embodiment, etc., only one set of the set of this embodiment is used, but in this embodiment, additional information can be embedded by preparing three sets of redundancy. It is possible. First, in the preprocessing of additional information, information is converted in accordance with the redundancy of the share pattern. In the example of FIG. 14, the number of sets is 3, and three types can be distinguished per pixel. Digital information is generally recorded with a radix of 2. In this case, additional information is converted into data with a radix of 3. Next, superimposition with additional information is performed simultaneously with selection of a small set by random numbers. A set number to be selected is acquired from the preprocessed additional information, and a small set in the set is selected by generating a random number. Thereby, the share pattern of each pixel is uniquely determined. A series of processing for each pixel is performed for all pixels.
[ 0080 ]
In this embodiment, the shared image is a monochrome binary image, and image conversion processing such as the fourth step (sharing image correction processing) of the first embodiment is not necessary.
[ 0081 ]
The above is a series of image processing in the present embodiment. After this processing, the share is output by recording the share image as a recorded material such as a hard copy in the image recording unit as in the first embodiment.
[ 0082 ]
Next, an image processing algorithm when the auxiliary pixel is further divided will be described with reference to the flowchart of FIG. The difference from the case of using a redundant share pattern is in the second step and the third step.
[ 0083 ]
First step (S1); input of embedded image
As in the case of using a redundant share pattern, the embedded image is input as a monochrome binary image, and the embedded image is not used.
[ 0084 ]
Second step (S2): Generation of share pattern
Next, a share pattern is generated based on the embedded image using a visual decryption secret sharing method. As in the first embodiment, a set is selected by generating a random number. Unlike the case where redundant share patterns are used, the share pattern is generated without dividing the set and the small set. In addition to the generation of the share pattern, the additional information is used as a two-dimensional code and the information is converted into a white and black two-dimensional pattern. Then, this two-dimensional pattern is divided into the number of black auxiliary pixels of the share pattern.
[0085] ]
Third step (S3); image superimposition processing
Next, in the third step (image superimposition processing), the share pattern generated in the second step and the two-dimensional pattern of additional information generated separately from the share pattern are superimposed to obtain the final share image. Generate. Specifically, the shared image is generated by embedding the two-dimensional pattern divided into the number of black auxiliary pixels of the share pattern in the black portion of the auxiliary pixels of the share pattern in the second step. FIG. 16 shows a state in which additional information is added to the black portion of each auxiliary pixel.
[ 0086 ]
Even when the auxiliary pixels are further divided, the image conversion processing as in the fourth step (sharing image correction processing) of the first embodiment is unnecessary, and after the series of image processing, the same as in the first embodiment. The share is output by recording the share image as a hard copy or other recorded matter in the image recording unit.
[ 0087 ]
(When embedding 4-b additional information as a digital watermark)
When embedding additional information as a digital watermark, a gray scale image is used as an embedded image, and a color image is used as an embedded image.
[ 0088 ]
The image processing algorithm in this embodiment is basically the same as the method described in the first embodiment and the second embodiment, but at the same time as embedding an image to be embedded, a digital watermark is simultaneously embedded in the embedded image. The point is different.
[ 0089 ]
As methods for putting a digital watermark, methods such as “pixel space use type”, “frequency domain use type”, and “statistics use type” are known, and are described in Document (3).
[ 0090 ]
The reproduction of the recorded information is the same whether the additional information is embedded as a two-dimensional code or embedded as a digital watermark. Similar to the first embodiment, the reproduction is divided into a part to be reproduced by human eyes and a reproduction part of additional information using the reproduction apparatus. The reproduction by the human eye is performed by overlapping the shares, and the additional information is reproduced by using a reproduction device.
[ 0091 ]
As shown in FIG. 17, the additional information reproducing apparatus includes an image input unit that reads a share, an image processing unit that processes an input result, and a result output unit that displays the processed result.
[ 0092 ]
For the image input unit, a barcode scanner, a flatbed scanner, a drum scanner, or the like can be used. When a bar code scanner is used, it is possible to input easily, and it is possible to configure a playback device at a relatively low cost. The image processing unit correlates the input image with the information corresponding to the share pattern, and decodes the additional information. The result output unit outputs the result by display on a display, printing by a printer, or the like.
[ 0093 ]
Conventionally, the information embedding method that can be reproduced by human eyes and the information embedding method using a device such as a digital watermark could not be merged, but this embodiment has the advantages of both. It is possible to fuse well.
[ 0094 ]
(Fifth embodiment)
Next, a fifth embodiment will be described. In the embodiments described so far, the embedded image is reproduced based on the color overlap between the shares. However, in this embodiment, the embedded image is reproduced based on the dot overlap instead of the color overlap. As is well known in the art, a moiré phenomenon occurs depending on the angle of a halftone image. The moire phenomenon is caused by adjusting the pattern of the angle of the halftone dots, and the embedded image is reproduced.
[ 0095 ]
In the present embodiment, a monochrome binary image is used as the embedded image. Further, a gray scale image is used as the embedded image. Although it is possible to use a gray scale image as an embedded image, it is difficult to reproduce clearly.
[ 0096 ]
Next, the algorithm of the image processing in this embodiment will be described in detail along the flowchart shown in FIG.
[ 0097 ]
First step (S1); input of embedded image and embedded image
An embedded image and an embedded image are input.
In the present embodiment, the embedded image is input as a monochrome binary image and a grayscale image as an embedded image. The embedded image is processed so that the image becomes dark overall after input. This is because a clear moire phenomenon cannot occur if there are not enough pixel values when the image is halftone in subsequent processing. In order not to impair the contrast of the image as much as possible, the pixel value is raised as a whole and adjusted so that a moire phenomenon that can be sufficiently confirmed by human eyes when the shares are overlapped is made.
[ 0098 ]
Second step (S2): Generation of share pattern
Next, based on the embedded image, a share pattern is generated using, for example, a visual decryption secret sharing method. The set of share patterns is divided into a set that causes moire when the shares are overlapped and a set that does not cause moire, and each is represented by a black pixel and a white pixel as a share pattern. Further, it is possible to express richer by increasing the number of sets according to the intensity of moire and increasing the number of auxiliary pixels of the share pattern accordingly.
[ 0099 ]
Third step (S3); image superimposition processing
Next, the share pattern and the embedded image are superimposed. In practice, the angle of the halftone dot is changed according to the share pattern, and the density of the halftone dot is adjusted according to the density value of the embedded image. For example, the black portion of the share pattern is set to a halftone dot angle of 0, and the white portion is set to a halftone dot angle of 45 degrees. As illustrated in FIG. 18, when the shares are overlapped, the moire phenomenon does not occur when the halftone dot angle is the same, and the moire phenomenon occurs when the halftone dot angle is different, thereby recognizing the embedded image. be able to. If the density of halftone dots is too small, sufficient moire does not occur and reproduction becomes difficult, so it is necessary to correct the image so as to ensure sufficient density in the third step.
[ 0100 ]
Fourth step (S4); share image correction processing
In this step, based on the image created in the third step, it is converted into raster data for outputting a halftone image. Conversion to raster data is performed by a process called rasterization for converting an image into halftone dots.
[ 0101 ]
The above is a series of image processing in the present embodiment, and after this processing, a share is generated by outputting to the medium in the image recording unit. Examples of the output device include a film output device such as an image setter and a halftone dot printer.
[ 0102 ]
Also in this embodiment, the embedded information is reproduced by the same method as in the first embodiment.
[ 0103 ]
In the prior art, it was difficult to generate a share that is difficult to copy with a copy machine or the like, but by using a halftone dot as in this embodiment, it is not possible to make an easy copy with a copy machine or the like. It became possible.
[ 0104 ]
【The invention's effect】
According to the present invention, it is possible to create a share using a pattern image in which an arbitrary image or a small image, which has been impossible in the past, can be created, and the expressiveness is enriched. As a medium, not only a transparent sheet but also an opaque one can be used. In addition, it is possible to embed additional information without hindering share reproduction. By using halftone dots, copying with a copying machine becomes difficult, the effect of preventing forgery is enhanced, and reproduction with better appearance becomes possible.
[Brief description of the drawings]
FIG. 1 is a block diagram showing a physical configuration of an image generation system of the present invention.
FIG. 2 is an explanatory diagram illustrating an example of an auxiliary pixel corresponding to a pixel.
FIG. 3 is a flowchart showing a flow of basic image processing by the image generation system of the present invention.
FIG. 4 is a block diagram showing a configuration when the image generation system of the present invention is realized by dedicated hardware.
FIG. 5 is an explanatory diagram illustrating a configuration example of a playback device.
FIG. 6 is an explanatory diagram showing the principle of how an embedded image is played back.
FIG. 7 is an explanatory diagram illustrating a mark for alignment.
FIG. 8 is an explanatory diagram illustrating disposing pixels at an interval;
FIG. 9 is an explanatory diagram illustrating an example of occurrence of a share pattern.
FIG. 10 is an explanatory diagram showing the principle that an embedded image is reproduced when a small pattern is a logo.
FIG. 11 is a flowchart showing a flow of image processing in the third embodiment.
FIG. 12 is an explanatory diagram showing a method of making a hole in a medium.
FIG. 13 is a flowchart illustrating a flow of image processing when additional information is embedded using redundancy of a share pattern according to the fourth embodiment.
FIG. 14 is an explanatory diagram illustrating a method of embedding additional information using the redundancy of the share pattern.
FIG. 15 is a flowchart when the auxiliary pixel is further divided and a two-dimensional code is embedded in the fourth embodiment.
FIG. 16 is an explanatory diagram illustrating a method of embedding a two-dimensional code in a black portion of an auxiliary pixel in the fourth embodiment.
FIG. 17 is a block diagram showing the configuration of a reproduction apparatus for additional information.
FIG. 18 is an explanatory diagram illustrating the principle that an embedded image is reproduced by moire.
FIG. 19 is an explanatory diagram showing an image processing process by the image generation system of the present invention.
[Explanation of symbols]
101 ... CPU
102: Image memory
103: Image input unit
104 ... Program
105: Image recording unit
106 ... Bus
107: Image processing unit

Claims (4)

Means for inputting an embedded image and an embedded image as a color pattern image as image data;
Means for generating a share pattern in which embedded images are distributed by associating one or more auxiliary pixels with each pixel of the embedded image;
Means for generating a share image by embedding the share pattern in the embedded image by moving the pixels of the embedded image based on one or more auxiliary pixels corresponding to the share pattern;
An image generation system comprising:   2. The image generation system according to claim 1, further comprising means for embedding additional information in an embedded image. A step of inputting an embedded image and an embedded image as a color pattern image as image data;
Generating a share pattern in which embedded images are dispersed by associating one or more auxiliary pixels with each pixel of the embedded image;
Embedding the share pattern in the embedded image by moving the pixel of the embedded image based on one or more auxiliary pixels corresponding to the share pattern; and
An image generation method comprising:   4. The image generation method according to claim 3, further comprising a step of embedding additional information in an embedded image.

Images