KR100485554B1 - Image processing apparatus, image processing method, and recording medium - Google Patents

Abstract

The character with the highest occurrence frequency is decomposed into individual copies, and the reference value is obtained. The reference value here is represented by the relative distance of the coordinates of four steps of a character image. Then, as the second or later characters selected in step S412, the relative positions of the copies of the characters are changed in accordance with the embedding information in consideration of the reference value for the characters corresponding to the predetermined bits of the document access control information to be embedded. .

Description

An image processing apparatus, an image processing method, and a recording medium {IMAGE PROCESSING APPARATUS, IMAGE PROCESSING METHOD, AND RECORDING MEDIUM}

The present invention relates to an image processing apparatus for embedding watermark information in a document image in which one character includes a character consisting of a plurality of character parts, and one character comprising a character composed of a plurality of character parts. An image processing apparatus, an image processing method, a program, and a storage medium for extracting watermark information embedded in a document image.

In digital image forming apparatuses, such as printers and copiers, in recent years, the improvement of the image quality is remarkable, and high quality printed matter can be easily obtained. That is, anyone can obtain the printed matter requested by the high performance scanner, printer, copier, and computer image processing. For this reason, problems such as illegal copying and alteration of documents have occurred, and the trend of embedding access control information in the printed matter itself has recently been active (watermark information) in order to prevent or suppress them.

As a method of realizing such a request, embedding access control information in a printed matter invisibly, embedding a bitmap pattern corresponding to the access control information in the margin of a document, applying a scrambled password to a document image, and the like. This is common at present. Among them, embedding access control information invisibly is a general realization method. The type of embedding information by controlling the amount of space between alphabetic strings, the type of embedding information by controlling the amount of rotation of characters, Background Art A type of embedding information by controlling an enlargement / reduction amount, a type of embedding information by transforming characters, and the like have been proposed.

9 is a view for explaining a method of embedding information by controlling the amount of space between English character strings. Here, 801 to 804 are called spaces. In addition, the space | interval of space 801 is p, and the space | interval of space 802 is s. In this state, if the bit of embedded information is 0, the intervals p and s of spaces 801 and 802 are p ← (1 + p) (p + s) / 2 and s ← (1-p) (p + s) / If it is changed to 2 and the bit of the embedding information is 1, p ← (1-p) (p + s) / 2 and s ← (1 + p) (p + s) / 2. This is similarly applicable to the spaces 803 and 804.

It is a figure explaining the method of embedding information by controlling the rotation amount of a character. Here, the left side of FIG. 10 shows a state before rotation, and the right side of FIG. 10 shows a state after rotation. 901 indicates the rotation angle of the character. This rotation angle is changed in accordance with the bits of the embedding information, similarly to the method shown in FIG.

11 is a view for explaining a method of embedding information by controlling the amount of enlargement / reduction of a character. 1001 represents the original size. 1002 represents the size after enlargement. This enlargement amount is changed in accordance with the bits of the embedding information, similarly to the method shown in FIG. The same applies to reduction.

However, the method of embedding the access control information invisibly is advantageous in order to maintain confidentiality. However, since the information is embedded in a document image (usually a two-value image) having a small redundancy in image, the information is embedded in characters and spaces. A feeling of incongruity occurs, and the deterioration of original quality becomes easy to be outstanding. In addition, such an image is also generally poor in print resistance (information holding force after outputting from paper).

This invention is made | formed in view of the above problem, Comprising: It aims at ensuring the information embedding precision and information embedding amount more than a fixed minimum, suppressing deterioration of a font.

In order to achieve the object of the present invention, for example, the image processing apparatus of the present invention includes the following configuration.

That is, an image processing apparatus which embeds watermark information into a document image in which one character includes a character consisting of a plurality of character parts, comprising: extracting means for extracting characters from the document image, and the extracting means; Selecting means for selecting a character having a predetermined configuration among the characters extracted by the character; and changing the position of the character portion of the character selected by the selection means in accordance with the document access control information, so that the character And embedding means for embedding the document access control information as watermark information.

In order to achieve the object of the present invention, for example, the image processing apparatus of the present invention includes the following configuration.

That is, an image processing apparatus for extracting watermark information embedded in a document image containing a character in which one character includes a plurality of character parts, comprising: character extracting means for extracting a character from the document image, and the character extracting means; Selecting means for selecting a character whose character portion has a predetermined configuration among the characters extracted by; and extracting the bit string embedded in the character based on the position of the character portion of the character selected by the selection means; And watermark information extracting means for restoring the watermark information as document access control information based on the column.

In order to achieve the object of the present invention, for example, the image processing method of the present invention includes the following configuration.

That is, an image processing method of embedding watermark information in a document image in which one character includes a character consisting of a plurality of character parts, comprising: an extraction step of extracting characters from the document image, and extracting from the extraction step; A selection step of selecting a character in which a character part has a predetermined configuration among the characters, and changing the position of the character part of the character selected in the selection step according to the document access control information, thereby giving the character the document access control information. And a embedding process for embedding as watermark information.

In order to achieve the object of the present invention, for example, the image processing method of the present invention includes the following configuration.

That is, an image processing method for extracting watermark information embedded in a document image containing a character in which one character includes a plurality of character parts, comprising: a character extraction step of extracting a character from the document image, and the character extraction step; A selection process of selecting a character having a predetermined configuration from the characters extracted in the step S, and extracting a bit string embedded in the character based on the position of the character portion of the character selected in the selection process, And a watermark information extraction step of restoring the watermark information as document access control information based on the above.

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

[First Embodiment]

Fig. 1 is a block diagram showing the basic configuration of a watermark information embedding apparatus for embedding watermark information in a document image in this embodiment.

In Fig. 1, reference numeral 2 denotes an input unit composed of a scanner, a camera, a file reading device or the like for inputting a target document embedding watermark information, and reference numeral 4 denotes a processor for performing various processes, 6) is a keyboard for inputting a command to the processor 4, 8 is a disc for storing embedded information or a read document image, and 10 is a reference numeral for various processes performed by the processor 4; A memory for temporarily storing data or storing a document image read by the input unit 2, reference numeral 12 denotes a command input to the processor 4 and a display indicating a processing state, and reference numeral 14 It is an output part comprised by the printer which outputs the document image in which the access control information was embedded, and a network interface, such as internet and LAN.

2 is a diagram illustrating a basic configuration of a watermark information extraction apparatus for extracting watermark information from a document image in which watermark information is embedded by the watermark information embedding apparatus shown in FIG.

In Fig. 2, reference numeral 22 denotes an input unit composed of a scanner, a camera, a file reading device, a network interface, or the like for inputting a document in which watermark information is embedded, and reference numeral 24 denotes a processor for performing various processes; Reference numeral 26 denotes a keyboard for inputting a command to the processor 24, reference numeral 28 denotes a disk for storing the read document image, or searching for an original file of the read document, and reference numeral 30 denotes a processor 24. Memory for temporarily storing data or the like or accumulating document images read by the input unit 22, reference numeral 32 denotes an instruction input to the processor 24, and a processing state. The display and reference numerals 34 and 36 are a network interface and a printer for utilizing the read document access control information, respectively.

In the present embodiment, the watermark information embedding apparatus and the watermark information extracting apparatus are treated as individual apparatuses, but the present invention is not limited thereto, and one of these apparatuses (watermark information embedding apparatus and watermark information extracting apparatus) is used. May be used as a watermark information embedding unit and a watermark information extracting unit in the apparatus.

Next, an approximate flow of the watermark information embedding process will be described. In response to a command input from the keyboard 6, first, an embedded document image which has been electronicized from the input unit 2 is acquired and expanded into the memory 10. Furthermore, embedding information (document access control information) is input from the keyboard 6 or the disk 8, and the processor 4 embeds the information into the document image developed on the memory 10. FIG. The document image in which the predetermined document access control information is embedded is output by the output unit 14 as a watermark information embedding end document.

Next, the general flow of the process of extracting watermark information from the document in which the watermark information embedding output by the output unit 14 is completed will be described. In accordance with the command input from the keyboard 26, first, the watermark information embedding end document is input from the input unit 22, and expanded into the memory 30. Next, the processor 24 reads the embedded document access control information from the document image developed on the memory 30, and performs a predetermined process based on the instruction. The predetermined processing is, for example, to notify the outside when a false reading is detected, to retrieve the original document to the internal disk 28 or the outside, or to print attribute information. Network I / F 34 and printer 36 are used for this purpose.

Next, the process of embedding the watermark information in one document image will be described in detail. The flowchart of the process is shown in FIG.

In step S200, a document is acquired from the input unit 2 and developed into the memory 10 as electronic image data. In this step, preprocessing such as the direction of the read document, the inclination correction, and the like are also performed. In step S202, area identification is performed on the document image developed in the memory 10 in step S200, and all character blocks (text) in the image are extracted. This can be realized by applying, for example, a block selection technique described in Japanese Patent Laid-Open No. 6-068301. In step S204, character recognition is performed on the characters included in all the character blocks extracted in step S202, and a character code as a result of character recognition is generated.

In step S206, a target character embedding document access control information is extracted from the characters included in the character block extracted in step S202. The target character to be extracted is a character of a predetermined font size. Details of the processing in this step will be described later. In step S208, document access control information embedded in the characters extracted in step S206 is input. Here, the document access control information is, for example, copy restriction information, alteration prevention information, original document management information, and the like.

In step S210, the document access control information input in step S208 is embedded in the characters extracted in step S206. Details of the processing in this step will be described later. In step S212, the document image embedding the document access control information is output in step S210.

Next, the process of the processor 24 extracting the watermark information from the document image in which one piece of watermark information embedding is completed will be described in detail. The flowchart of the process is shown in FIG.

In step S300, the document in which the watermark information embedding is completed is read from the input unit 22, and transferred to the memory 30 as electronic image data. The processing in this step also includes preprocessing such as the direction of the read document, the inclination correction, and the like, in step S200.

In step S302, region identification is performed on the document image in which the watermark information embedding completed in the memory 30 is completed in step S300, and all the character blocks in the document image are extracted. The processing in this step is performed in the same manner as the processing in step S202. In step S304, character recognition is performed on all the character blocks extracted in step S302. The processing in this step is performed in the same manner as the processing in step S204.

In step S306, only characters in which the document access control information is embedded are extracted from the characters included in the character block extracted in step S302. Details of the processing in this step will be described later. In step S308, document access control information is read from the characters extracted in step S306. Details of the processing in this step will be described later.

In step S310, predetermined control is performed in accordance with the document access control information read in step S308, for example, a copy prohibition process, a document search process, and the like.

5 is a flowchart for explaining the details of the processing in steps S206 to S210, and S306. In step S400, the character code based on the character recognition result is transferred to the character extraction working memory in the memory 10. In step S402, it is determined whether or not all character codes included in the document have been transferred to the character extraction working memory. If all transfers have been completed, the process proceeds to step S404, and if not finished, the process proceeds to step S400.

In step S404, the character set in advance is counted for each character using the character code transferred to the character extraction working memory. Here, it is set in advance that a Chinese character having a font size of 10 points, for example, has a complex number of sub-compositions in advance, such as "three or more kanji". will be. That is, in step S404, the number of character codes identical to the character code of the character set in advance among the character codes transferred to the character extraction working memory is counted. By making such a setting, it becomes possible to more reliably bury a certain amount or more of information. This will be described later in detail.

In step S406, the character counted in step S404 is sorted by the count number. In step S408, it is determined whether the number of counts is greater than or equal to, i.e., a character having a high appearance frequency or more in the document. This is a measure for taking a certain number or more of the target characters and repeatedly embedding the same information for the same characters in order to ensure the embedding precision of the watermark information. This is also to ensure the extraction precision of the watermark information. The more the fixed number here is, the higher the precision becomes, but may be about twice, for example. Details on this will be described later.

Here, if it is determined that the number of the target characters is not longer than the predetermined number, it is determined that embedding / extraction of the predetermined amount of information is impossible, and the process proceeds to step S414, and if determined to be possible, the process proceeds to step S410.

In step S410, among the documents to be embedded / extracted, a character having the highest appearance frequency among documents is selected to calculate a reference value for the embedding / extraction operation. Details of this reference value will be described later.

In step S412, in addition to the character from which the reference value is obtained, that is, from the result of sorting the character after the second occurrence frequency above in step S406, the document access control information embedding / extraction operation is performed. A concrete method is mentioned later.

In step S414, when it is determined in step S408 that there are few embedded / extracting target characters and embedding / extraction is impossible, this is a routine for performing a predetermined process. The predetermined process is, for example, a process such as displaying a warning of no embedding / extraction on the display 12 or the display 32.

FIG. 6 is a diagram for explaining the principle of embedding processing of document access control information performed in step S412, and FIG. 7 is a diagram for explaining processing for embedding 9-bit information into a document. Fig. 6 is a view for explaining a method of obtaining a reference value and a method of embedding three bits of information (eight kinds of information), respectively. In the following description, the document access control information is 9 bits of information, but is not limited thereto. 6 and 7, for example, the number of copies (letter portions) is three, for example, a small number of copies, such as a pattern of the character " type " A pattern such as a dog is described as an example. In addition, the drawings are shown somewhat enlarged for convenience of description.

First, the character image extracted from the character block in step S206 (an image of a Chinese character of "type" in FIG. 6) is decomposed into individual copies, and the reference value is obtained. Since the method for decomposing the characters into individual copies is not particularly limited, the methods generally disclosed may be used. The reference value is the most important value when the document access control information is embedded in the document in an invisible form. The reference value here is represented by the relative distances, K, P, M, and N of the four coordinates of the character image, as defined in FIG. 6.

Although it is a detailed method of embedding information, it is considered to embed 3 bits of information per character using the four reference values K, P, M, and N defined immediately above. In step S410, reference values K, P, M, and N of the most reliable letters, that is, the most frequent letters are obtained (equivalent to the third step in FIG. 7). On the other hand, the document access control information (9 bits) to be embedded is prepared and any one of the processes shown in Fig. 6 (for each of the number of copies of the characters) for the second and subsequent characters selected in step S412 for each three bits. Processing to change the relative position) (equivalent to the fourth step in FIG. 7). Specifically, when embedding information for every three bits for the second and subsequent characters, for example, the first three bits (as a result of sorting) are the second character, the next three bits are the third character, and the last. Three bits of are embedded in the fourth character. However, it is not limited to this order, ie the first three bits are embedded in the fourth character, the next three bits are embedded in the third character, and the last three bits are embedded in the second character. You may also do it.

In any case, the memory 10 stores information (the embedding information) of which character the information for every three bits is embedded in. An example of this embedding information is shown in FIG. In Fig. 12, the embedding information 1201 is stored in the memory 10, and the embedding information 1201 shows the sort order of the characters embedded in the first 3 bits, the sort order of the characters embedded in the second 3 bits, It consists of the information of the sort order of the character which embedded the 3rd 3 bits.

By referring to this embedding information in the watermark information extraction process, when the information of each three bits is extracted from each character, the information of each of these three bits is rearranged in any order to obtain the original 9-bit document access control information. This is because you can specify whether to restore. Details of the extraction process of the watermark information will be described later.

When embedding information for every three bits in each character, as described above, the position of the character is changed according to the embedding information, but the pattern of change is the information (000, 001, 010, 011) shown in FIG. , 100, 101, 110, and 111.

Further, as can be seen from the above description, by adjusting the number of the target characters (lowest appearance frequency), it is possible to maintain the balance between the embedding / extraction precision of the watermark information and the embedding bit number.

In FIG. 6, K ', P', M ', and N' are the relative distance of 4 steps | pieces after changing a position.

In order to prevent deterioration of the characters, the largest copy, that is, the copy below will not move. Arbitrary information can be embedded in the procedure described above.

On the other hand, at the time of extracting the watermark information, a reference value is obtained in the same manner as in the above-described method, and the bit array embedded in each character is extracted by comparing the relative position of each number of characters having the second appearance frequency with the reference value. . At this time, since embedding information is stored in the memory 10 as described above, with reference to the embedding information, the number of characters embedded in the extracted bit array is specified, and the original document is accessed. Restore control information.

According to the above description, after the area identification and the character recognition are performed by the image processing apparatus and the image processing method in this embodiment, the relative positional change for each number of copies of a character having a complicated copy configuration is skillfully utilized. While minimizing deterioration, it is possible to secure a certain amount of information embedding precision and information embedding amount (controllable by sorting by appearance frequency). In addition, even when the watermark information is extracted, it is possible to realize the watermark information with high noise immunity. In addition, since there is no dependence on the font size in principle, it is obvious that it is an effective method even for an original having a small number of characters.

Second Embodiment

In the first embodiment, the copy configuration employed as the embedding target is a single pattern as shown in FIG. 6, but is not limited to this, and includes a plurality of kanji patterns having a copy number greater than or equal to that shown in FIG. 8. It may be set at the same time. In this case, what is necessary is just to employ | adopt the method used in 1st Example with respect to the character which has each additional structure, and the embedding information quantity further increases.

Third Embodiment

In the first embodiment, although 3 bits are embedded in one character, the present invention is not limited to this, and may be freely set as long as the number of bits in the combination can take the additional movement pattern. However, if the embedded bit number is increased, the degree of deformation of the character increases.

[Example 4]

In the first embodiment, although the watermark information is embedded in the kanji, it is not limited to this, and it is also possible to embed it in Korean, Thai, and the like as long as it is a character composed of a plurality of components (letter portions). Lan

[Example 5]

For example, when embedding a bit string of "000" in a kanji using "reference" as shown in Fig. 6, N '= N, but noise is superimposed on the document image embedded in the bit string. In the case of being affected by disturbance, it is difficult to extract the bit string "000" from the Chinese character "type" in which the bit string "000" is embedded. In other words, even if N 'is obtained in the extraction process, N' = N may not be strictly corrected. As a result, the bit string "000" may not be extracted.

Therefore, you may change so that a width | variety may be made by comparison with N 'and N in an extraction process. That is, if | N'-N1 | <ε is satisfied, it is determined that N '= N. This process can also be applied to other reference values. In addition, if | M'-M | <ε is satisfied, it may be determined that M '= M.

[Example 6]

In the first embodiment, a character having the same font size is used as the character to embed the document access control information. This makes the number of shifts (e.g., a change from K to K ') to be shifted by embedding document access control information in each target character is constant, and the balance of characters after the shift of the number of copies is largely between each target character. To keep it constant.

However, even if the font sizes of the respective target characters are not the same, if the amount of shift of the number of shifts for each font size is determined in advance, the balance of the characters after the shift of the number of shifts can be made substantially constant between the target characters. Further, the movement amount of the copies for each font size may be obtained at the time of embedding. In that case, for example, if the moving amount of the number of copies is set to 10 characters, c can be obtained by calculating (c × (character size of 12 points) / (character size of 10 points)) at 12 points.

[Example 7]

In the first embodiment, although the character with the highest frequency of appearance is employed as a character used to obtain a reference value, the character is not limited to this. For example, on the basis of the number of strokes, the number of copies, etc. in advance in each copy pattern, Even if it moves the number of copies, it is divided into the group which is not visually outstanding and the group which is visually outstanding, and it sets as a embedding target character and a reference character, respectively, and the character in the several reference character group which actually appeared may be used as a reference character. . The reference value in that case is obtained by, for example, obtaining the average value of the reference value for each character included in the visually noticeable group when the number of copies moves.

[Example 8]

In the first embodiment, a character having the highest appearance frequency is employed as a character used to obtain a reference value. However, the character is not limited to this and may be, for example, the first character of a document or a character block. In this case, when the first character does not correspond to the predetermined number of copies, control is performed with the next character or the like.

[Example 9]

In the first embodiment, although the character with the highest appearance frequency is adopted as the character used to obtain the reference value, the character is not limited to this, for example, the character from the second appearance frequency to the predetermined rank is used as the reference character. You may also do it. In this case, the reference value is obtained by obtaining the average value of the reference values for each reference character.

[Example 10]

In the first embodiment, the reference values K, P, M, and N are represented by the relative distances, K, P, M, and N, of the coordinates of the four stages of the character image, as defined in FIG. However, it is not limited to this, but the other example of a reference value is shown in FIG.

FIG. 13 is a diagram for explaining the principle of embedding document access control information using reference values obtained by a method different from the first embodiment. The reference value in the present embodiment is expressed by the ratio of the width and height of each number of parts separated to the absolute height and width of the character image. For further explanation, each of the copies is referred to here, but only the top two copies are described, and the ratio of the size is not defined for the bottom copies. If this is defined as all copies, the relative reference value changes, and further, the deformation of the letters is increased, resulting in a noticeable deterioration. Thus, one of the most important copies is left, in which case only two copies are used.

The reference values K, P, M, and N thus obtained are used in the same manner as in the first embodiment, and the document access control information embedding process is performed.

Other Examples

In addition, the present invention is not limited to the apparatus and method for realizing the above embodiment, but supplies the program code of the software for realizing the above embodiment to a computer (CPU or MPU) in the system or the apparatus. Even if the computer of the system or apparatus operates the various devices according to the program code to realize the above embodiment, it is included in the scope of the present invention.

In this case, the program code itself of the software realizes the functions of the above embodiment, and the program code itself, and means for supplying the program code to a computer, specifically, a storage medium storing the program code. It is included in the scope of the invention.

As a storage medium storing such a program code, for example, a floppy disk, a hard disk, an optical disk, a magneto-optical disk, a CD-ROM, a magnetic tape, a nonvolatile memory card, a ROM, or the like can be used. .

In addition, by controlling the various devices only in accordance with the program code supplied by the computer, not only the function of the embodiment is realized but also the OS (operating system) or other application software running on the computer. Even if the above embodiment is realized, such a program code is included in the scope of the present invention.

In addition, after the supplied program code is stored in a memory provided in the function expansion board of the computer or the function expansion unit connected to the computer, the CPU is provided in the function expansion board or the function storage unit based on the instruction of the program code. The case where such or the like performs part or all of the actual processing and the above embodiment is realized by the processing is also included in the scope of the present invention.

As described above, according to the present invention, it is possible to secure a predetermined or more information embedding precision and information embedding amount while minimizing deterioration of the font.

BRIEF DESCRIPTION OF THE DRAWINGS Fig. 1 is a block diagram showing the basic configuration of a watermark information embedding apparatus for embedding watermark information (electronic watermark) in a document image in the first embodiment of the present invention.

FIG. 2 is a block diagram showing a basic configuration of a watermark information extracting apparatus for extracting watermark information from a document image in which watermark information is embedded by the watermark information embedding apparatus shown in FIG.

3 is a flowchart of a process in which the processor 4 embeds watermark information in one document image.

4 is a flowchart of a process in which the processor 24 extracts watermark information from a document image in which one watermark information embedding is completed.

5 is a flowchart for explaining the details of the processing in steps S206 and S306;

FIG. 6 is a view for explaining the principle of embedding processing of document access control information performed in step S412; FIG.

Fig. 7 is a diagram explaining a process of actually embedding 9 bits of information in a document.

Fig. 8 is a diagram showing a plurality of Chinese character patterns having a number of copies larger than a given one.

9 is a view for explaining a method of embedding information by controlling the amount of space between English character strings.

10 is a view for explaining a method of embedding information by controlling the amount of rotation of characters.

FIG. 11 is a view for explaining a method of embedding information by controlling an enlargement / reduction amount of a character. FIG.

12 is a diagram illustrating an example of embedding information.

Fig. 13 is a view for explaining the principle of embedding processing of document access control information using reference values obtained by a method different from the first embodiment.

<Explanation of symbols for the main parts of the drawings>

2, 22: input unit

4, 24: processor

6, 26: keyboard

8, 28: disk

10, 30: memory

12, 32: display

Claims (29)

An image processing apparatus in which watermark information is embedded in a document image in which one character includes a character consisting of a plurality of character parts. Extraction means for extracting characters from the document image; Selection means for selecting a character in which the character portion configuration of the character has a predetermined configuration among the characters extracted by the extraction means; By changing the position of the predetermined character portion of the character selected by the selection means to a position determined according to the watermark information with respect to another character portion whose position in the character does not change, watermark information is applied to the selected character. Landfill Image processing apparatus comprising a. The method of claim 1, The extraction means, Character block extracting means for extracting a character block from the document image; Character recognition means for performing character recognition on a character included in the character block extracted by said character block extracting means, generating a character code as a recognition result, and extracting an image of said character from said character block; An image processing apparatus further comprising. The method of claim 1, The selecting means includes counting means for counting, for each character, the number of characters among the characters extracted by the extracting means, in which the character partial configuration of the character has a predetermined configuration, And the embedding means embeds the watermark information in the character selected by the selection means when the number of characters by the counting means is equal to or greater than a predetermined number of counts. The method of claim 3, And the counting means counts the number of characters for which the character portion of the character has a predetermined configuration for each character using the character code of the character. The method of claim 3, And when the number of counts by the counting means is smaller than a predetermined number of counts, a warning indicating that embedding of watermark information is impossible on the predetermined display means. The method of claim 1, And calculating means for calculating a reference value to be used when embedding the watermark information. The method of claim 6, And said calculating means calculates said reference value using a predetermined character. The method of claim 6, And said calculating means calculates a circumferential relative distance of circumscribed four-stage coordinates of the character selected for reference value calculation as said reference value. The method of claim 6, And said calculating means calculates the ratio of the width and height of each character portion of the character as the reference value to the width and height of the character selected for the reference value calculation. The method of claim 1, And the embedding means generates information indicating to which character the watermark information is embedded. An image processing apparatus for extracting watermark information embedded in a document image in which one character includes a character consisting of a plurality of character parts. Character extracting means for extracting a character from the document image; Selection means for selecting a character in which the character portion configuration of the character has a predetermined configuration among the characters extracted by the character extraction means; Watermark information extraction means for extracting watermark information determined according to the position of the character portion of the character selected by the selection means Image processing apparatus comprising a. The method of claim 11, The character extracting means, Character block extracting means for extracting a character block from the document image; Character recognition means for performing character recognition on a character included in the character block extracted by said character block extracting means, generating a character code as a recognition result, and extracting an image of said character from said character block; An image processing apparatus further comprising. The method of claim 11, The selecting means includes counting means for counting, for each character, the number of characters in which the character portion of the character has a predetermined configuration among the characters extracted by the character extraction means; And the character extracting means extracts the watermark information from the character selected by the selecting means when the number of characters by the counting means is equal to or greater than a predetermined number of counts. The method of claim 13, And the counting means counts the number of characters for which the character portion of the character has a predetermined configuration for each character using the character code of the character. The method of claim 13, And when the number of counts by said counting means is smaller than a predetermined number of counts, a warning is displayed on predetermined display means that extraction of watermark information is impossible. The method of claim 11, And the watermark information extracting means extracts the watermark information by referring to information indicating which character the watermark information is embedded in. The method of claim 11, And calculating means for calculating a reference value used when extracting the watermark information. The method of claim 17, And said calculating means calculates said reference value using a predetermined character. The method of claim 17 or 18, And said calculating means calculates a circumferential relative distance of circumscribed four-stage coordinates of the character selected for reference value calculation as said reference value. The method of claim 17 or 18, And said calculating means calculates the ratio of the width and height of each character portion of the character as the reference value to the width and height of the character selected for the reference value calculation. The method according to claim 1 or 11, wherein The watermark information is document access control information, and includes copy restriction information, alteration prevention information, and original document management information. The method according to claim 1 or 11, wherein Characters included in the document image include Chinese characters, Korean, and Thai. The method according to claim 1 or 11, wherein And said character portion includes the number of kanji characters. An image processing method in which watermark information is embedded in a document image in which one character includes a character consisting of a plurality of character parts. An extraction step of extracting characters from the document image, A selection step of selecting a character in which the character partial configuration of the character has a predetermined configuration among the characters extracted in the extraction step; By embedding watermark information in the selected character by changing the position of a predetermined character portion of the character selected in the selection process to a position determined according to the watermark information with respect to another character portion whose position in the character does not change. Landfill process Image processing method comprising a. An image processing method for extracting watermark information embedded in a document image including a character in which one character includes a plurality of character portions, A character extraction step of extracting characters from the document image, A selection step of selecting a character in which the character partial configuration of the character has a predetermined configuration among the characters extracted in the character extraction step; A watermark information extraction step of extracting watermark information determined according to the position of the character portion of the character selected in the selection process Image processing method comprising a. A computer-readable recording medium having recorded thereon a program for causing a computer to function in the image processing apparatus according to claim 1. A computer-readable recording medium storing a program for causing the computer to execute the image processing method according to claim 24 or 25. delete delete