JP2008098949A - Image processor and image processing method - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a technology for extracting embedded information from a document image having been electronically watermarked while deterioration in document quality is suppressed. <P>SOLUTION: A spectrum 1103 is obtained as to a circumscribed rectangle 1101. A spectrum 1104 is obtained as to a circumscribed rectangle 1102 after a character part in the circumscribed rectangle 1101 is filled with pixels having pixel values of pixels constituting a character part included in an original document image. A difference spectrum 1105 between the spectrum 1103 and spectrum 1104 is obtained. Spectra are obtained as to respective patterns corresponding to all values that embedded bit values possibly have. Similarities between the difference spectrum 1105 and the respective spectra are obtained. A bit value corresponding to a pattern having the highest similarity is outputted. <P>COPYRIGHT: (C)2008,JPO&INPIT

Description

  The present invention relates to a digital watermark extraction technique.

  In recent years, digital image forming apparatuses such as printers and copiers have been remarkably improved in image quality, and it has become possible to easily obtain high-quality printed materials. That is, anyone can obtain the required printed matter by image processing using a high-performance scanner, printer, copier, and computer. For this reason, problems such as illegal copying and falsification of documents have occurred, and in order to prevent or prevent them, a movement to embed access control information as watermark information in the printed matter itself has recently become active.

  As such a function, there is an invisible type digital watermark which embeds access control information in a printed matter invisible. As a general realization method, there are a type in which information is embedded by controlling the amount of space in an English character string (see Patent Document 1) and a type in which information is embedded by rotating characters. In addition, there are proposed a type in which information is embedded by enlarging / reducing a character, a type in which information is embedded by deforming a character, and the like.

  FIG. 12 is a diagram illustrating a type in which information is embedded by controlling the amount of space in an English character string. In the figure, reference numeral 1201 denotes a sentence to be embedded with a digital watermark, and reference numeral 1202 denotes a sentence as a result of embedding the digital watermark into the sentence 1201.

  Further, p indicates the amount of space between the word “Modern” and the word “computer” in the sentence 1201. S indicates the amount of space between the word “computer” and the word “networks”.

  Also, p1 indicates the amount of space between the word “Modern” and the word “computer” in the sentence 1202. Further, s1 indicates the amount of space between the word “computer” and the word “networks”.

  Here, when the information bit “0” is embedded in the sentence 1201, p1 and s1 are determined as follows.

p1 = (1 + q) (p + s) / 2
s1 = (1-q) (p + s) / 2
On the other hand, when the information bit “1” is embedded in the sentence 1201, p1 and s1 are determined as follows.

p1 = (1-q) (p + s) / 2
s1 = (1 + q) (p + s) / 2
Here, the range of q is 0 <q <1.

  FIG. 13 is a diagram illustrating a type in which information is embedded by enlarging or reducing the size of a character.

  For example, it is assumed that the embedded information “1” is embedded in the character “image” in the area 1301. In this case, as shown in a region 1303, the size of this character is enlarged. On the other hand, it is assumed that the embedded information “0” is embedded in the character “RE” in the area 1302. In this case, as shown in a region 1304, the size of this character is reduced.

  The character to be embedded may be a continuous character, an interval of several characters, or a character at a predetermined position. In FIG. 13, since the character “image” is enlarged and the character “re” is reduced, the embedded information “10” is embedded in the sentence.

  FIG. 14 is a diagram illustrating a type in which information is embedded by rotating characters. For example, it is assumed that the embedded information “1” is embedded in the character “image” in the area 1401. In this case, as shown in a region 1403, the character is rotated by a predetermined angle clockwise. On the other hand, it is assumed that the embedded information “0” is embedded in the character “composition” in the area 1402. In this case, as shown in a region 1404, the character is rotated by a predetermined angle counterclockwise.

The character to be embedded may be a continuous character, an interval of several characters, or a character at a predetermined position. In FIG. 14, the character “image” is rotated clockwise, and the character “composition” is rotated counterclockwise, so that embedded information “10” is embedded in the sentence. Become.
US Pat. No. 6,086,706

  However, when the embedded information is embedded invisible, especially in a document image, the character and space are uncomfortable, and the deterioration of the document quality is easily noticeable.

  The present invention has been made in view of the above problems, and an object of the present invention is to provide a technique for extracting embedded information from a document image embedded with a digital watermark while suppressing deterioration in document quality.

  In order to achieve the object of the present invention, for example, an image processing apparatus of the present invention comprises the following arrangement.

That is, an acquisition unit that acquires a watermark document image obtained by reconstructing a character part included in a document image with a pattern corresponding to an embedded bit value;
Circumscribed rectangle calculating means for obtaining a circumscribed rectangle for each character portion in the watermark document image;
First calculation means for obtaining a spectrum for the target circumscribed rectangle obtained by the circumscribed rectangle calculating means;
Spectrum acquisition means for acquiring a spectrum for the focused circumscribed rectangle after the character portion in the focused circumscribed rectangle is filled with pixels having pixel values of pixels constituting the character portion included in the document image When,
Second calculation means for obtaining a difference spectrum between the spectrum obtained by the first calculation means and the spectrum obtained by the spectrum acquisition means;
Third calculation means for obtaining a spectrum for each of the patterns corresponding to all possible values of the embedded bit value;
Fourth calculation means for obtaining a similarity between the difference spectrum and each spectrum obtained by the third calculation means;
Output means for specifying a pattern constituting a character portion in the noted circumscribed rectangle based on the respective similarities obtained by the fourth calculation means, and outputting a bit value corresponding to the specified pattern; It is characterized by providing.

  In order to achieve the object of the present invention, for example, an image processing method of the present invention comprises the following arrangement.

That is, an acquisition step of acquiring a watermark document image obtained by reconstructing a character portion included in a document image with a pattern corresponding to an embedded bit value;
A circumscribed rectangle calculation step for obtaining a circumscribed rectangle for each character portion in the watermark document image;
A first calculation step of obtaining a spectrum for the target circumscribed rectangle obtained in the circumscribed rectangle calculation step;
A spectrum acquisition step of acquiring a spectrum for the focused circumscribed rectangle after the character portion in the focused circumscribed rectangle is filled with pixels having pixel values of pixels constituting the character portion included in the document image When,
A second calculation step for obtaining a difference spectrum between the spectrum obtained in the first calculation step and the spectrum obtained in the spectrum acquisition step;
A third calculation step for obtaining a spectrum for each of the patterns corresponding to all possible values of the embedded bit value;
A fourth calculation step for obtaining a similarity between the difference spectrum and each spectrum obtained in the third calculation step;
An output step of identifying a pattern constituting the character portion in the noted circumscribed rectangle based on the respective similarities obtained in the fourth calculation step, and outputting a bit value corresponding to the identified pattern; It is characterized by providing.

  According to the configuration of the present invention, it is possible to extract embedded information from a document image in which electronic watermarks are embedded while suppressing deterioration in document quality.

  Hereinafter, the present invention will be described in detail according to preferred embodiments with reference to the accompanying drawings.

[First Embodiment]
<About embedding embedded information>
FIG. 1 is a block diagram illustrating a functional configuration of an image processing apparatus that embeds a digital watermark into a document image. As shown in the figure, an image processing apparatus 190 according to this embodiment includes an image input unit 101, a document analysis unit 102, an embedding determination unit 103, a watermark information input unit 105, an embedding unit 106, and an image output unit 107. Yes.

  A document image 100 to be embedded with a digital watermark is input into the apparatus by the image input unit 101 and sent to the subsequent document analysis unit 102.

  The document analysis unit 102 performs processing for obtaining a circumscribed rectangle for each character included in the document image 100. The digital watermark is embedded in the character portion in the circumscribed rectangle.

  The embedding determination unit 103 performs a process of determining a circumscribed rectangle capable of embedding a digital watermark from each circumscribed rectangle.

  A watermark information input unit 105 inputs watermark information (embedding information) 104 used for embedding a digital watermark into the apparatus.

  The embedding unit 106 embeds the embedding information 104 input by the watermark information input unit 105 for the character portion in the circumscribed rectangle that the embedding determination unit 103 has determined that the digital watermark can be embedded.

  The image output unit 107 outputs the document image in which the embedding information is embedded by the embedding unit 106 as a watermark embedded image 108.

  FIG. 3 is a flowchart of the digital watermark embedding process for the document image performed by the image processing apparatus 190.

  First, in step S301, the document image 100 is input into the apparatus by the image input unit 101. The input form of the document image 100 to the apparatus is not particularly limited. For example, the document image 100 may be acquired by reading a document with a scanner or the like, or the document image 100 created in advance may be acquired from a memory in the apparatus or an external memory. Further, text data created by application software or the like converted into image data may be used as the document image 100.

  In step S <b> 302, the document analysis unit 102 performs processing for obtaining a circumscribed rectangle for each character included in the document image 100 input into the apparatus by the image input unit 101. The circumscribing rectangle of a character is a rectangle circumscribing the character and is originally information indicating an area where character recognition is performed. In the digital watermark technique, it indicates a character area to be embedded. Examples of processing for obtaining a circumscribed rectangle of a character include the following.

  First, each pixel value of the document image is projected onto the vertical coordinate axis, a blank part (part without a character) is searched, a line is determined, and line division is performed. Thereafter, the document image is projected on the horizontal coordinate axis line by line, and a blank portion is searched and divided into characters. Thereby, each character can be cut out by a circumscribed rectangle.

  In step S303, the watermark information input unit 105 acquires embedded information (watermark information) 104 to be embedded in the document image 100 input in step S301. The embedded information 104 is composed of a bit string that takes a value of “0” or “1”.

  The acquisition form of the embedded information 104 is not particularly limited. For example, the user of the apparatus may input using a keyboard or mouse (not shown), or the pre-created embedded information 104 may be acquired from a memory in the apparatus or an external memory. .

  In step S304, the embedding determination unit 103 selects one circumscribed rectangle that has not yet been selected from the circumscribed rectangle created in step S302 for each character in the document image 100. The selection order is not particularly limited, but any selection order may be used as long as it is known in what selection order the circumscribed rectangle is selected from the document image on the embedding information extraction side described later.

  In step S305, the embedding determination unit 103 obtains the size of the circumscribed rectangle selected in step S304 (for example, the vertical length and the horizontal length of the circumscribed rectangle), and determines whether or not the obtained size is equal to or larger than a predetermined size. Check. In the present embodiment, it is assumed that the embedding information 104 is embedded only in a circumscribed rectangle having a certain size or larger, and the embedding information 104 is not embedded in a circumscribed rectangle having a certain size or larger. That is, a character in a circumscribed rectangle having a small size is easily affected by a slight disturbance, and therefore, the embedded information 104 embedded in the character may not be extracted. Therefore, in the present embodiment, only the circumscribed rectangle having a certain size or larger is selected as the target for embedding the embedded information 104 in order to extract the embedded information 104 from the characters more robustly.

  If it is determined that the circumscribed rectangle selected in step S304 has a certain size or more, the process proceeds to step S306. On the other hand, if the circumscribed rectangle selected in step S304 does not have a certain size or more, the process returns to step S304 to select another circumscribed rectangle that has not yet been selected.

  In step S306, the embedding unit 106 selects one bit from each bit constituting the embedding information 104, and performs a process of embedding the selected one bit into the character portion in the circumscribed rectangle selected in step S304.

  Here, the digital watermark embedding process performed by the embedding unit 106 will be described in more detail. In the present embodiment, patterns corresponding to all values (0, 1) that can be taken by the bits constituting the embedded information 104 are created in advance. When 1 bit is selected from the embedding information 104, a pattern corresponding to the selected 1-bit value is selected. Then, the character portion in the circumscribed rectangle selected in step S304 is reconstructed with the selected pattern.

  FIG. 6 is a diagram for explaining the embedding process. In the figure, reference numeral 604 denotes a pattern to be selected when the 1-bit value selected from the embedded information 104 is “0”, and reference numeral 605 denotes a 1-bit value selected from the embedded information 104. This is the pattern to select. In the figure, reference numeral 601 denotes a character in which the embedded information 104 is embedded.

  Therefore, when the bit value “0” is embedded in the character 601, the character 601 is reconstructed using the pattern 604 corresponding to the bit value “0”. Then, as shown at 602, although the character shape is the same as that of the character 601, a character having the pattern 604 can be obtained.

  On the other hand, when the bit value “1” is embedded in the character 601, the character 601 is reconstructed using the pattern 605 corresponding to the bit value “1”. Then, as shown by 603, although the shape of the character is the same as that of the character 601, a character whose pattern has the pattern 605 can be obtained.

  Note that the size and pattern of the pattern described here are merely examples, and may be appropriately changed according to the purpose.

  FIG. 5 is a flowchart showing details of the embedding process in step S306.

  In step S306a, one bit is selected from each bit constituting the embedded information 104. The selection order is not particularly limited, but here, selection is made bit by bit from the top.

  In step S306b, it is checked whether or not the value of 1 bit selected in step S306a is “1”. If the result of the check is “1”, the process proceeds to step S306c. If the result is “0”, the process proceeds to step S306d.

  In step S306c, the character portion in the circumscribed rectangle selected in step S304 is reconstructed with a pattern (pattern 1) corresponding to the bit value “1”. On the other hand, in step S306d, the character portion in the circumscribed rectangle selected in step S304 is reconstructed with a pattern (pattern 0) corresponding to the bit value “0”.

  Then, the process proceeds to step S307.

  FIG. 4 is a flowchart showing details of the process in step S306c.

  First, in step S306c1, the circumscribed rectangle selected in step S304 is divided into the size of pattern 1. Each divided portion is referred to as a division pattern.

  Next, in step S306c2, one division pattern that has not yet been selected is selected.

  Next, in step S306c3, the pixel values of pixels corresponding in position in the divided pattern selected in step S306c2 and pattern 1 are compared. Here, for the sake of explanation, the pixel value of the pixel constituting the character part is “0” (black dot), and the pixel value of the pixel constituting the non-character part is “255” (white dot). It is not limited to pixel values.

  When the pixel value of the pixel referred to on the divided pattern side is “0” and the pixel value of the pixel corresponding to the referenced pixel on the pattern 1 is “255”, the pixel value is referred to on the divided pattern side. The pixel value of this pixel is “255”. That is, when the division pattern and the pattern 1 are completely overlapped, the pixel value “255” is set for all pixels on the division pattern at the same position as the pixel value “255” on the pattern 1. To do. As a result, the division pattern is reconstructed with the pattern 1.

  In step S306c4c, it is checked whether all division patterns have been selected. If all the division patterns are selected as a result of the check, the process proceeds to step S307. On the other hand, if there is a division pattern that has not yet been selected, the process returns to step S306c2, one division pattern that has not yet been selected is selected, and the subsequent processing is performed using the selected division pattern.

  The process in step S306d in FIG. 5 is the same as step S306c except that pattern 0 is used instead of pattern 1, and thus detailed description thereof is omitted.

  Returning to FIG. 3, if all circumscribed rectangles have been selected in step S304, the process proceeds to step S308 via step S307. On the other hand, if there is a circumscribed rectangle that has not been selected in step S304, the process returns to step S304 via step S307 to select one circumscribed rectangle that has not yet been selected. Process.

  In step S309, the image output unit 107 outputs the watermark embedded image 108 in which the embedding information 104 is embedded by the embedding unit 106. The output form is not particularly limited, and the watermark embedded image 108 may be output to a memory provided inside or outside the apparatus, or the watermark embedded to an apparatus having a printing function such as a printer. The image 108 may be printed out. Further, the watermark embedded image 108 may be transmitted to the outside via a network.

<About extraction of embedded information>
FIG. 7 is a block diagram illustrating a functional configuration of the image processing apparatus 790 that extracts the embedded information 104 embedded in the document image 100 by the image processing apparatus 190 performing the above-described processing. In this embodiment, the image processing apparatus 190 and the image processing apparatus 790 are separate apparatuses, but may be included in the same apparatus.

  In FIG. 7, the image processing apparatus 790 includes an image input unit 701, a document analysis unit 702, an embedding determination unit 703, and a watermark information extraction unit 704.

  The document image 700 as the watermark embedded image 108 is input into the apparatus by the image input unit 101 and sent to the subsequent document analysis unit 702.

  The document analysis unit 702 performs processing for obtaining a circumscribed rectangle for each character included in the document image 700. The digital watermark is extracted for each circumscribed rectangle.

  The embedding determination unit 703 performs a process of determining a circumscribed rectangle in which a digital watermark is to be embedded from among each circumscribed rectangle.

  The watermark information extraction unit 704 performs processing for extracting embedded information (bits) from the circumscribed rectangle determined by the embedding determination unit 703 that the digital watermark will be embedded. Then, the sequentially extracted bits are sequentially connected and output as watermark information 705. This watermark information 705 is the same as the watermark information 104.

  FIG. 8 is a flowchart of processing performed by the image processing apparatus 790 to extract the watermark information 705 from the watermark embedded image 108.

  In step S801, a document image 700 is input into the apparatus by the image input unit 701. The input form of the document image 700 to the apparatus is not particularly limited. For example, a document image may be acquired by reading a document with a scanner or the like, or a document image 700 created in advance may be acquired from a memory in the apparatus or an external memory.

  In step S <b> 802, the document analysis unit 702 performs processing for obtaining a circumscribed rectangle for each character included in the document image 700 input into the apparatus by the image input unit 701. Such processing is performed in the same manner as the processing performed by the document analysis unit 102.

  In step S803, the embedding determination unit 703 selects one circumscribed rectangle that has not yet been selected from the circumscribed rectangle created in step S802 for each character in the document image 700. The selection order is not particularly limited, but it is preferable to match the selection order in step S304.

  In step S804, the embedding determination unit 703 obtains the size of the circumscribed rectangle selected in step S803 (for example, the vertical length and the horizontal length of the circumscribed rectangle), and determines whether or not the obtained size is equal to or larger than a predetermined size. Check. If it is determined that the circumscribed rectangle selected in step S803 has a certain size or more, the process advances to step S805. On the other hand, if the circumscribed rectangle selected in step S803 does not have a certain size or larger, the process returns to step S803 to select another circumscribed rectangle that has not yet been selected.

  In step S805, the watermark information extraction unit 704 performs processing for extracting the bit value embedded in the circumscribed rectangle selected in step S803. FIG. 9 is a flowchart showing details of the processing in step S805.

  First, in step S805a, an amplitude spectrum is obtained for the circumscribed rectangle selected in step S803, and an amplitude spectrum of pattern 0 is obtained. And the similarity g0 between each amplitude spectrum is calculated | required. FIG. 11 is a flowchart showing details of the processing in step S805a. FIG. 10 is a diagram for explaining the processing in step S805a. Hereinafter, the processing in step S805a will be described in detail with reference to FIGS.

  In step S805a11, first, an amplitude spectrum (hereinafter referred to as spectrum 1) is obtained for the circumscribed rectangle selected in step S803. In general, the amplitude spectrum for the two-dimensional image f is obtained as follows. Here, x and y are variables indicating the x-coordinate value and y-coordinate value in the two-dimensional image f, respectively, and f (x, y) is the pixel value of the pixel located at the coordinate (x, y) in the image f. Indicates.

  First, the two-dimensional Fourier transform for the two-dimensional image f is performed based on the following equation.

  It was asked in this way

Is called the spectrum of image f. spectrum

Generally takes complex values. Therefore,

Polar coordinates are expressed using the size and the angle of deviation. Then size

Is the amplitude spectrum.

  In FIG. 10, 1101 indicates the circumscribed rectangle selected in step S803, and 1103 is an amplitude spectrum (spectrum 1) obtained for the circumscribed rectangle 1101.

  In step S805a12, white dots in the character portion in the circumscribed rectangle selected in step S803 are converted into black dots. That is, the character portion in the circumscribed rectangle selected in step S803 is filled with pixels having pixel values of pixels constituting the character portion included in the original document image 100. In this process, for example, white dots are searched while scanning the pixels in the circumscribed rectangle selected in step S803 pixel by pixel, and if there are white dots, they are changed to black dots.

  In FIG. 10, reference numeral 1102 denotes a circumscribed rectangle obtained by performing the processing in step S805a12 on the character portion in the circumscribed rectangle 1101.

  Returning to FIG. 11, next, in step S805a13, an amplitude spectrum (hereinafter referred to as spectrum 2) is obtained for the circumscribed rectangle obtained by the processing in step S805a12. The method for obtaining the spectrum is as described above. In FIG. 10, reference numeral 1104 denotes the spectrum 2 obtained for the circumscribed rectangle 1102.

  The spectrum 2 obtained in this step is obtained by obtaining in advance for each character in the document image 100. Therefore, the spectrum 2 may be obtained in advance for each character, and in this step, the spectrum 2 obtained in advance for the circumscribed rectangle selected in step S803 may be obtained.

  Returning to FIG. 11, next, in step S805a14, the process of enlarging the size of the pattern 0 to the size of the circumscribed rectangle selected in step S803 is performed. In FIG. 10, reference numeral 1106 denotes a pattern 0 having an enlarged size.

  In step S805a15, an amplitude spectrum (hereinafter referred to as spectrum 0) is obtained for pattern 0 whose size has been increased. The method for obtaining the spectrum is as described above. In FIG. 10, reference numeral 1107 denotes a spectrum 0 obtained for the pattern 0 whose size is enlarged.

  Next, in step S805a16, first, a result obtained by subtracting spectrum 2 from spectrum 1 is obtained as a difference spectrum. In FIG. 10, reference numeral 1105 denotes a differential spectrum obtained by subtracting spectrum 2 indicated by 1104 from spectrum 1 indicated by 1103. Further, in this step, a cross-correlation g0 between this difference spectrum and spectrum 0 is obtained.

  Next, the process proceeds to step S805b in FIG. In step S805b, similarly to step S805a, the amplitude spectrum of the circumscribed rectangle selected in step S803 is obtained, and the amplitude spectrum of pattern 1 is obtained. And the similarity g1 between each amplitude spectrum is calculated | required. That is, the process in step S805b is the same as the process in step S805a, except that pattern 1 is used instead of pattern 0. In FIG. 10, reference numeral 1108 denotes a pattern 1 with an enlarged size. In FIG. 10, reference numeral 1109 denotes a spectrum obtained for the pattern 1 whose size is enlarged.

  Next, in step S805c, the magnitudes of the similarities g0 and g1 are compared. If g1 ≧ g0 as a result of the magnitude comparison, the process proceeds to step S805d. If g1 <g0, the process proceeds to step S805e.

  In step S805d, the pattern constituting the character portion in the circumscribed rectangle selected in step S803 is assumed to be pattern 1, and “1” is output as the bit value embedded in the circumscribed rectangle.

  On the other hand, in step S805e, the pattern constituting the character portion in the circumscribed rectangle selected in step S803 is assumed to be pattern 0, and “0” is output as the bit value embedded in the circumscribed rectangle.

  After any of the processes in steps S805d and S805e, the process proceeds to step S806 in FIG.

  If all circumscribed rectangles have been selected in step S803, the process proceeds to step S807 via step S806. On the other hand, if there is a circumscribed rectangle that has not yet been selected in step S803, the process returns to step S803 via step S806 to select one circumscribed rectangle that has not yet been selected. Process.

  In step S807, the watermark information extraction unit 704 outputs the bit values extracted for each circumscribed rectangle in the order in which the circumscribed rectangle was selected in step S803. As a result, watermark information 705 is output from the watermark information extraction unit 704. The output form is not particularly limited, and the watermark information 705 may be output to a memory provided inside or outside the apparatus, or the watermark information 705 may be output to the outside via a network. You may send it.

  As described above, according to the present embodiment, the embedded information can be extracted from the document image in which the digital watermark is embedded while suppressing deterioration of the document quality.

  Moreover, according to this embodiment, the pattern which comprises a character can be correctly extracted by calculating | requiring the Fourier plane which does not have the influence of the shape of a character.

  In this embodiment, one bit of “0” and “1” is embedded in one character part. However, a plurality of bits such as “010” and “110” may be embedded in one character area. In this case, the pattern is not two types of pattern 0 and pattern 1, but more types of patterns are prepared in advance. For example, when 3 bits are embedded in one character area, a pattern corresponding to the embedded bit string is used among the 8 types of patterns.

  On the extraction side, when calculating the similarity, the difference spectrum is similar to each pattern (pattern 0, pattern 1,..., Pattern n (n = 8 when 3 bits are embedded for one character area). Degrees g0, g1,... Gn are obtained, and if gj (0 ≦ j ≦ 7) is the largest among all the similarities obtained, it is determined that pattern j is embedded, and pattern j The value corresponding to is output.

[Second Embodiment]
The components constituting the image processing apparatus 190 shown in FIG. 1 and the components constituting the image processing apparatus 790 shown in FIG. 7 may be implemented by hardware, but may be implemented by software. good. In this case, the software is installed in a computer such as a general PC (personal computer), and the CPU of the computer executes the software, thereby realizing the functions of the units shown in FIGS. . In this case, the computer functions as the image processing apparatus 190 and also functions as the image processing apparatus 790.

  FIG. 2 is a block diagram showing a hardware configuration of a computer that executes software when the functional configurations of the image processing apparatus 190 and the image processing apparatus 790 are realized by software.

  Reference numeral 202 denotes a CPU that controls the entire computer by executing processing using programs and data stored in the main storage device 203 and executes processing according to the flowcharts described above.

  Reference numeral 203 denotes a main storage device represented by a RAM. The main storage device 203 is an area for temporarily storing programs and data loaded from an HDD (hard disk) 204, a CD-ROM drive 209, a DVD-ROM drive 210, and an FDD (floppy (registered trademark) disk drive). Have The main storage device 203 also has a work area that is used by the CPU 202 to execute various processes. Further, the main storage device 203 has an area for temporarily storing document image data input from the scanner 217 via the I / F (interface) 215. That is, the main storage device 203 can provide various areas as appropriate.

  Reference numeral 204 denotes an HDD. In the above description, the OS (operating system), programs and data for causing the CPU 202 to execute processing according to the above flowcharts, document images, and the like may be created in advance. Various types of data are stored. Programs and data stored in the HDD 204 are loaded into the main storage device 203 as appropriate under the control of the CPU 202. The CPU 202 executes various processes using the loaded program and data.

  Reference numeral 205 denotes a video controller which generates an image signal and a character signal in accordance with image data and character data generated as a result of processing by the CPU 202 and sends them to the monitor 206. The monitor 206 includes a CRT, a liquid crystal screen, and the like, and displays images and characters based on image signals and character signals sent from the video controller 205.

  An I / F 208 functions as an interface for connecting the printer 216 to the computer. As is well known, the I / F 208 functions as an interface for the computer to send print data to the printer 216, and functions as an interface for the computer to acquire the status of the printer 216. is there.

  The CD-ROM drive 209 reads a program and data recorded on a CD-ROM as an example of a storage medium, and sends it to the HDD 204 and the main storage device 203.

  The DVD-ROM drive 210 reads a program and data recorded on a DVD-ROM as an example of a storage medium and sends it to the HDD 204 and the main storage device 203.

  The FDD 211 reads a program or data recorded on a floppy (registered trademark) disk as an example of a storage medium, and sends it to the HDD 204 or the main storage device 203.

  For the CD-ROM drive 209, the DVD-ROM drive 210, and the FDD 211, information may be recorded on a corresponding storage medium.

  The I / F 212 functions as an interface for connecting the mouse 213 and the keyboard 214 to the computer. When a user of the computer operates using the mouse 213 or the keyboard 214, an operation signal indicating the operation content is notified to the CPU 202 via the I / F 212. The mouse 213 and the keyboard 214 are used, for example, for selecting the document image 100 in which the digital watermark is to be embedded, instructing and creating the embedded information 104.

  The I / F 215 functions as an interface for connecting the scanner 217 to the computer, and image data read by the scanner 217 is sent to the HDD 204 and the main storage device 203 via the I / F 215. Is done. For example, the scanner 217 functions as a device for reading a sentence and inputting the document image 100 or the document image 700 into the computer.

  Note that the configuration shown in the figure is an example, and is not necessarily limited to using all the configurations, and other devices may be added as necessary.

[Other Embodiments]
Needless to say, the object of the present invention can be achieved as follows. That is, a recording medium (or storage medium) that records a program code of software that implements the functions of the above-described embodiments is supplied to a system or apparatus. Then, the computer (or CPU or MPU) of the system or apparatus reads and executes the program code stored in the recording medium. In this case, the program code itself read from the recording medium realizes the functions of the above-described embodiment, and the recording medium on which the program code is recorded constitutes the present invention.

  Further, by executing the program code read by the computer, an operating system (OS) or the like running on the computer performs part or all of the actual processing based on the instruction of the program code. Needless to say, the process includes the case where the functions of the above-described embodiments are realized.

  Furthermore, it is assumed that the program code read from the recording medium is written in a memory provided in a function expansion card inserted into the computer or a function expansion unit connected to the computer. After that, based on the instruction of the program code, the CPU included in the function expansion card or function expansion unit performs part or all of the actual processing, and the function of the above-described embodiment is realized by the processing. Needless to say.

  When the present invention is applied to the recording medium, program code corresponding to the flowchart described above is stored in the recording medium.

It is a block diagram which shows the function structure of the image processing apparatus which embeds a digital watermark with respect to a document image. FIG. 3 is a block diagram illustrating a hardware configuration of a computer that executes software when the functional configurations of the image processing apparatus 190 and the image processing apparatus 790 are realized by software. 10 is a flowchart of processing for embedding a digital watermark into a document image, which is performed by an image processing apparatus 190. It is a flowchart which shows the detail of a process of step S306c. It is a flowchart which shows the detail of the embedding process in step S306. It is a figure explaining an embedding process. It is a block diagram which shows the function structure of the image processing apparatus 790 which extracts the embedding information 104 embedded in the document image 100 by the image processing apparatus 190 performing said process. 10 is a flowchart of processing for extracting watermark information 705 from the watermark embedded image 108, which is performed by the image processing apparatus 790. It is a flowchart which shows the detail of the process in step S805. It is a figure explaining the process in step S805a. It is a flowchart which shows the detail of the process in step S805a. It is a figure explaining the type which embeds information by controlling the amount of spaces of an English character string. It is a figure explaining the type which embeds information by enlarging or reducing the size of a character. It is a figure explaining the type which embeds information by rotating a character.

Claims (6)

Acquisition means for acquiring a watermark document image obtained by reconstructing a character portion included in a document image with a pattern corresponding to an embedded bit value;
Circumscribed rectangle calculating means for obtaining a circumscribed rectangle for each character portion in the watermark document image;
First calculation means for obtaining a spectrum for the target circumscribed rectangle obtained by the circumscribed rectangle calculating means;
Spectrum acquisition means for acquiring a spectrum for the focused circumscribed rectangle after the character portion in the focused circumscribed rectangle is filled with pixels having pixel values of pixels constituting the character portion included in the document image When,
Second calculation means for obtaining a difference spectrum between the spectrum obtained by the first calculation means and the spectrum obtained by the spectrum acquisition means;
Third calculation means for obtaining a spectrum for each of the patterns corresponding to all possible values of the embedded bit value;
Fourth calculation means for obtaining a similarity between the difference spectrum and each spectrum obtained by the third calculation means;
Output means for specifying a pattern constituting a character portion in the noted circumscribed rectangle based on the respective similarities obtained by the fourth calculation means, and outputting a bit value corresponding to the specified pattern; An image processing apparatus comprising: The third calculation means includes:
2. The image according to claim 1, wherein a spectrum is obtained for each of the patterns after expanding the size of each pattern corresponding to all possible values of the embedded bit value to the size of the circumscribed rectangle. Processing equipment. The output means includes
2. The pattern having the highest similarity with the difference spectrum among the similarities obtained by the fourth calculation means is specified, and a bit value corresponding to the specified pattern is output. An image processing apparatus according to 1. An acquisition step of acquiring a watermark document image obtained by reconstructing a character portion included in a document image with a pattern corresponding to an embedded bit value;
A circumscribed rectangle calculation step for obtaining a circumscribed rectangle for each character portion in the watermark document image;
A first calculation step of obtaining a spectrum for the target circumscribed rectangle obtained in the circumscribed rectangle calculation step;
A spectrum acquisition step of acquiring a spectrum for the focused circumscribed rectangle after the character portion in the focused circumscribed rectangle is filled with pixels having pixel values of pixels constituting the character portion included in the document image When,
A second calculation step for obtaining a difference spectrum between the spectrum obtained in the first calculation step and the spectrum obtained in the spectrum acquisition step;
A third calculation step for obtaining a spectrum for each of the patterns corresponding to all possible values of the embedded bit value;
A fourth calculation step for obtaining a similarity between the difference spectrum and each spectrum obtained in the third calculation step;
An output step of identifying a pattern constituting the character portion in the noted circumscribed rectangle based on the respective similarities obtained in the fourth calculation step, and outputting a bit value corresponding to the identified pattern; An image processing method comprising:   A program for causing a computer to execute the image processing method according to claim 4.   A computer-readable storage medium storing the program according to claim 5.

Images