JP4154252B2 - Image processing apparatus and method - Google Patents

Description

[0001]
BACKGROUND OF THE INVENTION
The present invention relates to an image processing apparatus and method, for example, image processing of a digital document or the like.
[0002]
[Prior art]
In recent years, the concept of ISO 15408 as a standard for security of office documents in general has spread worldwide, and from this point of view, it is becoming a very important technical field. Under such circumstances, various so-called digital watermark techniques have been devised and used as one of document information security management methods.
[0003]
There are various security management purposes such as prevention of unauthorized copying of data, prevention of leakage or falsification of important information, copyright protection of document information, or charging for the use of image data, etc. Various digital watermarking methods have been proposed. For example, as a technique for embedding watermark information into digital image data so that it cannot be perceived by humans, the image data disclosed in Japanese Patent Application No. 10-278629 is subjected to wavelet transform to reduce redundancy in frequency space. There are known methods such as embedding watermark information.
[0004]
A binary image such as a document image has little redundancy and it is difficult to realize a digital watermark technique. However, there are several digital watermark methods (hereinafter referred to as “document watermark”) using characteristics unique to a document image. Are known. For example, a method of moving the baseline of a line (Japanese Patent No. 3160661), a method of operating a space length between words (US Pat. No. 6,086,706, Japanese Patent Laid-Open No. 9-186603), a method of operating a space length between characters (King Mongkut University "Electronic document data hiding technique using inter-character space" The 1998 IEEE Asia-Pacific Conf. On Circuits and Systems, 1998, pp. 419-422) 11-234502).
[0005]
The above method is characterized in that the user cannot determine that watermark information is embedded in an image (hereinafter referred to as “invisible watermark”). Conversely, a method of embedding watermark information by clearly indicating to the user that the watermark information is embedded (hereinafter referred to as “visible watermark”) has also been proposed. For example, in Japanese Patent Application No. 10-352619, the result of performing a reversible operation between the original image and the embedded sequence by comparing the pixel position of the original image and the shape of the watermark image to be embedded is transmitted to the user. A method of embedding information in a visible form is disclosed.
[0006]
[Patent Literature]
Japanese Patent Application No. 10-278629
Japanese Patent No. 3160661
U.S. Patent No. 6,086,706
JP-A-9-186603
Japanese Patent Laid-Open No. 11-234502
Japanese Patent Application No. 10-352619
[0007]
[Non-patent literature]
King Mongkut University "Electronic document data hiding technique using inter-character space" The 1998 IEEE Asia-Pacific Conf. On Circuits and Systems, 1998, pp. 419-422
[0008]
[Problems to be solved by the invention]
The digital watermarking method is basically intended to embed some additional information in the image data itself, and uses the embedded additional information to prevent unauthorized use, copyright protection, data tampering, etc. This is intended to protect the original image. In other words, the purpose of prohibiting viewing the original image itself or permitting copying only to a user having a predetermined authority is not assumed.
[0009]
In addition, the protection of the original image is applied to the entire image. Therefore, there is a problem that even an unprotected image included in the protected image cannot be seen or copied.
[0010]
The present invention provides image information text In the area Adaptively according to the display / non-display selected by the user The purpose is to control the process.
[0011]
Also, text region about Another purpose is to protect the image.
[0012]
[Means for Solving the Problems]
The present invention has the following configuration as one means for achieving the above object.
[0013]
The present invention is an image processing method performed by an image processing apparatus,
The input means of the image processing apparatus inputs image information,
A recognition unit included in the image processing apparatus recognizes a plurality of text regions included in the input image information;
The generation unit included in the image processing apparatus generates authentication information for controlling at least one of display, printing, copying, and transmission processing for the text area for each text area,
The selection means included in the image processing apparatus selects at least one of display, printing, duplication, and transmission of the text area and then displaying or not displaying the text in the text area,
When the display unit is selected in the selection, the embedding unit included in the image processing apparatus embeds the authentication information as a digital watermark in a state where the text in the text region is visible in the text region, When non-display is selected in the selection, the authentication information is embedded as a digital watermark with the text in the text area invisible in the text area.
It is characterized by that.
[0014]
An input means for inputting image information;
Recognition means for recognizing a plurality of text regions included in the input image information;
Generating means for generating authentication information for controlling at least one of display, printing, copying, and transmission processing for the text area for each text area;
A selection means for selecting either display or non-display of the text in the text area after performing at least one of display, printing, duplication, and transmission processing on the text area;
When display is selected by the selection means, the text in the text area is visible, and the authentication information is embedded as a digital watermark for each text area,
An embedding unit that embeds the authentication information as a digital watermark in a state where the text in the text area is invisible when non-display is selected by the selection unit;
It is characterized by having.
[0017]
DETAILED DESCRIPTION OF THE INVENTION
Hereinafter, embodiments of the present invention will be described in detail with reference to the drawings.
[0018]
[First Embodiment]
[Constitution]
FIG. 1 is a block diagram illustrating a configuration example of an image processing system according to an embodiment.
[0019]
This image processing system is realized in an environment in which offices (such as a plurality of sections) 10 and 20 are connected by a WAN 104 such as the Internet.
[0020]
A LAN 107 built in the office 10 includes a multi-function processor (MFP) 100, a management PC 101 that controls the MFP 100, a client PC 102, a document management server 106, and a database managed by the document management server. 105 etc. are connected. The office 20 has substantially the same configuration as that of the office 10, but a LAN 108 built in the office 20 is connected to at least a document management server 106, a database 105 managed by the document management server, and the like. The LAN 107 of the office 10 and the LAN 108 of the office 20 are connected to each other via a proxy server 103 connected to the LAN 107, a WAN 104, and a proxy server 103 connected to the LAN 108.
[0021]
The MFP 100 is in charge of a part of image processing for reading an image of a paper document and processing the read image. An image signal output from the MFP 100 is input to the management PC 101 via the communication line 109. The management PC 101 is a normal personal computer (PC), and has a memory such as a hard disk for storing images, an image processing unit composed of hardware or software, a monitor such as a CRT or LCD, and an input unit such as a mouse and keyboard. However, some of them are integrated into the MFP 100.
[0022]
FIG. 2 is a block diagram illustrating a configuration example of the MFP 100.
[0023]
The image reading unit 110 including an auto document feeder (ADF) irradiates an image of each of one or a plurality of originals with a light source, and forms a reflection image from the originals on a solid-state image sensor with a lens. An image read signal in raster order (for example, 600 dpi) is obtained from the solid-state imaging device. When copying a document, the image reading signal is converted into a recording signal by the data processing unit 115. When copying to a plurality of recording papers, the recording signal for one page is temporarily stored in the recording unit 111. By repeatedly outputting the recording signal to the recording unit 112, images are formed on a plurality of recording sheets.
[0024]
On the other hand, the print data output from the client PC 102 is input to the network interface (I / F) 114 via the LAN 107, converted into raster data that can be recorded by the data processing unit 115, and then the recording unit. By 112, an image is formed on the recording paper.
[0025]
An operator's instruction to the MFP 100 is performed by a key operation unit provided in the MFP 100 and an input unit 113 including a keyboard and a mouse of the management PC 101. Display of operation input, display of image processing status, and the like are performed by the display unit 116.
[0026]
The operation of the MFP 100 is controlled by a control unit (not shown) in the data processing unit 115.
[0027]
The storage unit 111 can also be controlled from the management PC 101. Data exchange and control between the MFP 100 and the management PC 101 are performed via a network I / F 117 and a signal line 109 directly connecting both.
[0028]
[processing]
3 to 6 are flowcharts for explaining the outline of the processing by the image processing system.
[0029]
The original is scanned by the image reading unit 110 to obtain an image signal of 600 dpi and 8 bits (image information input process, S1201). The data processing unit 115 performs preprocessing such as trimming, skew correction (including direction correction), noise removal, and the like on the image signal (S1202), and generates a binary image by binarization processing (S1203). Then, one page of image data (multi-value and binary image data) is stored in the storage unit 111.
[0030]
The CPU of the management PC 101 executes block selection on the image data stored in the storage unit 111 to identify the character / line drawing part, the gradation image part, and the background part where no character / line drawing or image exists ( S1204). Further, the character / line drawing section is divided into, for example, a paragraph unit area and other structures (tables and line drawings having ruled lines), and these are segmented (text area). On the other hand, the gradation image portion and the background portion are segmented (picture region) into independent objects for each unit that can be divided, such as a rectangular region (S1205). Then, based on the position information of the separated text area and the position information of the picture area, the text area cuts out a binary image and the picture area cuts out a multi-valued image from the image data stored in the storage unit 111 (S1206). In the following description, the clipped image area may be referred to as a “block”.
[0031]
The following processing is performed for each block. If the block to be processed is a text area, it is determined whether or not watermark information is embedded in the block by a document watermark detection process, and if it is a picture area, it is determined by a background watermark detection process (S1207). If it is determined that the watermark information is embedded, the display flag of the area is set to OFF (S1210). If it is determined that the watermark information is not embedded, the display flag of the area is set to ON (S1209). . Then, it is determined whether or not the same processing has been performed for all blocks (S1211), and the processing from step S1207 to S1210 is repeated until the setting of the display flag for all the blocks is completed.
[0032]
Subsequently, a block to be processed is selected (S1212), and whether or not watermark information is embedded in the selected block is determined by a display flag (S1213). If not embedded, the process proceeds to `` Processing A '' described later. To do. On the other hand, if watermark information is embedded, the user is prompted to enter a password (S1214). As will be described later, this password is used to control the display of the block and to authenticate other control functions such as printing and transmission.
[0033]
When the password is input, the validity is determined (S1215). If the password is invalid, the process proceeds to “Process B” described later. If the password is correct, it is determined whether or not the password is for display (S1216). If the password is for display, it is further determined whether or not the block is a background part (S1217). If it is a gradation image portion), the display flag of the block is turned ON (S1221).
[0034]
If it is determined in step S1217 that it is a background part, that is, if it is a background part in which watermark information is embedded, the image does not exist, so the watermark information embedded in the background (hereinafter referred to as “background watermark”) Pointer information indicating a data storage location is extracted (S1218), and original data is acquired from the document management server 106 or the like (S1219). It should be noted that when the watermark information is not embedded in the original data, it is necessary to inherit the watermark information. If the watermark information is not inherited, thereafter, various controls of the block will be disabled. Alternatively, the watermark information may be newly input instead of inheriting the watermark information. After the watermark data is inherited in the original data of the block (that is, the background watermark information is embedded in the image as an invisible watermark) or new watermark information is embedded (S1220), and the display image embedded with the watermark information is prepared Then, the display flag of the block is turned ON (S1221).
[0035]
On the other hand, if it is determined in step S1216 that the password is not for display, it is determined whether or not the block is a text area (S1222). If not, the process proceeds to step S1225. In the case of a text area, the binary image data of the block is sent to the document management server 106 and saved (S1223), and watermark information (pointer information indicating the save destination of the image data, various passwords, various control information, etc.) The block is masked by embedding a background watermark (S1224). In step S1225, the display flag of the block is turned OFF.
[0036]
Next, other control information of the block (whether printing, copying, transmission, etc.) is extracted from the watermark information (S1226), and other control flags of the block are turned ON or OFF according to the control information (S1227). Subsequently, it is determined whether or not the processing of all the blocks has been completed (S1228). If not completed, the process returns to step S1212, and if completed, various controls are performed according to the control flag (S1229). Note that there are a print flag, a duplication flag, and a transmission flag corresponding to control information such as printing, duplication, and transmission. If they are on, the image data of the block is printed, duplicated or transmitted, and they are off. If so, the image data of the block is not printed, copied or transmitted.
[0037]
Next, “Processing A” when it is determined in step S1213 that there is no watermark information will be described.
[0038]
First, it is determined whether or not the block is a text area (S1241). If it is not a text area, it is not a control target block, and the process advances to step S1228. In the case of a text area, the watermark embedding mode is set, and the user selects whether to embed a document watermark that can read the text (display mode) or to embed a background watermark and mask the block (non-display mode). (S1242). When the display mode is selected, various passwords are set (S1246) and embedded as a document watermark as watermark information including these passwords (S1247). If the non-display mode is selected, various passwords are set (S1243), the binary image data of the block is sent to the document management server 106 and the like (S1244), pointer information, various passwords, A background watermark including various control information and the like is embedded, and the block is masked (S1245).
[0039]
Then, the image of the block (the image or background after embedding watermark information) is displayed again (S1248), and the process proceeds to step S1228.
[0040]
Next, “Process B” when it is determined in step 1215 that the password is invalid will be described. ,
[0041]
First, it is determined whether or not the block is a text area (S1251). If it is not a text area (it is a masked area, there is no problem with display-related maintenance), all control flags are turned OFF to perform control-related maintenance. (S1255), the process proceeds to step S1228. In addition, in order to hide the text area, the binary image data of the block is sent to the document management server 106 or the like for storage (S1252), and includes destination pointer information, various passwords, various control information, etc. The background watermark is embedded, the block is masked (S1253), the block is redisplayed (S1254), all control flags are turned off (S1255), and the process proceeds to step S1228.
[0042]
Printing restrictions and transmission restrictions will be described as an example of various controls as follows.
・ When there is a print instruction
Block with print flag off: Prints the background image with the background watermark embedded
Block with print flag on: Prints an image with embedded document watermark or original data
・ When there is a transmission instruction
Block with send flag off: Sends the background image with the background watermark embedded
Block with send flag turned on: Sends an image with embedded document watermark or original data
[0043]
By performing such control, it is possible to freely manage security (for example, viewing restriction, copy restriction, transmission restriction, printing restriction, etc.) for each object of the document. Also, when a document is printed, the document watermark and invisible watermark are embedded in the text area and the picture area, respectively, which makes it possible to manage the security of the object read from the printed image and greatly improve the security of the document. be able to.
[0044]
Below, the detail is demonstrated about main processes.
[0045]
[Block selection]
First, block selection in steps S1204 and S1205 will be described.
[0046]
Block selection recognizes the image of one page shown in Fig. 7 as a collection of objects, and the attributes of each object are text (TEXT), drawing (PICTURE), photo (PHOTO), line (LINE), table (TABLE) ) And is divided into areas (blocks) having different attributes. Next, a specific example of block selection will be described.
[0047]
First, an image to be processed is binarized into a black and white image, and a block of pixels surrounded by black pixels is extracted by contour tracking. For a black pixel block having a large area, outline tracing is performed for white pixels inside the block to extract a white pixel block. Further, the extraction of the black pixel block and the white pixel block is recursively repeated such that the black pixel block inside the white pixel block having a predetermined area or more is extracted.
[0048]
The pixel blocks obtained in this way are classified by size and shape, and are classified into regions having different attributes. For example, if a pixel block with an aspect ratio close to 1 and a size within a specified range is used as a pixel block for character attributes, and if pixel blocks with adjacent character attributes are aligned and can be grouped, they can be grouped into a character area And In addition, a flat pixel block having a small aspect ratio is defined as a line region, and a range occupied by a black pixel block having a shape larger than a predetermined size and close to a rectangle and including aligned white pixel blocks is defined as a table region. A region where irregular pixel clusters are scattered is classified as a photographic region, and a pixel block of other arbitrary shape is classified as a graphic region.
[0049]
FIG. 8 is a diagram showing the result of block selection, and FIG. 8 (a) shows block information of each extracted block. FIG. 8 (b) shows the total number of blocks extracted by block selection as input file information. Such information is used when embedding and extracting watermark information.
[0050]
[Embed document watermark]
Next, embedding of a document watermark will be described.
[0051]
A document image 3001 shown in FIG. 9 is a block separated as a text area by block selection. Further, a circumscribed rectangle 3004 for each character element is extracted from the text area by a document image analysis 3002 described later. A character element refers to a rectangular area extracted using projection, and may be a single character or a character component (hen, structure, etc.).
[0052]
Then, a blank length between circumscribed rectangles is calculated from the extracted circumscribed rectangle 3004 information, and each circumscribed rectangle is shifted to the left and right based on an embedding rule described later, thereby embedding 1-bit information between circumscribed rectangles ( An embedding process 3003) generates a document image 3005 in which watermark information 3006 is embedded.
[0053]
The document image analysis 3002 is originally an element technology for character recognition, and is a technology for dividing a document image into graphic regions such as a text region and a graph, and cutting out characters in the text region in character units using projection. As an example, the technique described in JP-A-6-68301 can be cited.
[0054]
[Extract Document Watermark]
Next, a document watermark extraction method will be described.
[0055]
First, as in the case of embedding the document watermark, the circumscribed rectangle 3103 of the character is extracted from the image 3005 shown in FIG. 10 by the block selection and document image analysis 3002, and the extracted circumscribed rectangle 3103 information is used. Calculate the blank length of. Further, in each row, a character for embedding 1-bit information is specified, and embedded watermark information 3105 is extracted based on an embedding rule described later (extraction process 3104).
[0056]
Next, the embedding rule will be described.
[0057]
The space lengths before and after the character in which 1-bit information is embedded are P and S as shown in FIG. The characters that embed 1-bit information are every other character except for the characters at both ends of the line. If (PS) / (P + S) is calculated from the blank length, quantized at an appropriate quantization step, and the remainder is calculated, 1-bit information can be restored. Equation (1) shows this relationship, and the embedded value V ('0' or '1') can be extracted.
V = floor [(P-S) / {α (P + S)}] mod 2… (1)
Where α is the quantization step (0 <α <1)
[0058]
When embedding watermark information, the circumscribed rectangle is shifted by 1 pixel to the left and right, and the shift amount (number of pixels) to the left or right until the value to be embedded ('0' or '1') by Equation (1) Increase.
[0059]
FIG. 12 is a flowchart showing a process for searching for the shift amount. In FIG. 12, variable i is a candidate value of the shift amount, and variables Flag1 and 2 indicate whether or not to touch the adjacent character when shifting the character to be shifted to the right or left by the distance i. Becomes '1'.
[0060]
First, the initial value of the variable is set (S3402), and if the character (or character element) to be shifted is shifted to the right by the distance i, it is determined whether or not it touches the character (or character element) on the right. (S3403) When making contact, Flag1 is set to '1' (S3404). Subsequently, when the character to be shifted is shifted to the left by the distance i, it is determined whether or not the character adjacent to the left is touched (S3405). If touched, Flag1 is set to '1' (S3406).
[0061]
Next, it is determined whether or not the shift of the distance i is possible (S3407). If both flags are “1”, it is determined that the shift is impossible and the shift amount is set to 0 (S3408). In this case, it is impossible to embed information by shifting the character to be shifted.
[0062]
If Flag1 is '0' (S3409), it is determined by Equation (1) whether or not the value V to be embedded is obtained when the character to be shifted is shifted to the right by the distance i. However, if the value V is obtained, the shift amount is set to + i (S3411). Note that the sign of the shift amount means that a positive shift to the right and a negative shift to the left.
[0063]
If Flag1 is '1' or value V is not obtained by right shift and Flag2 is '0' (S3412), if the character to be shifted is shifted left by distance i, an attempt is made to embed Whether or not the value V to be obtained is obtained is determined by the equation (1) (S3413). If the value V is obtained, the shift amount is set to -i (S3414).
[0064]
If the value V cannot be obtained by either the right shift or the left shift, the variable i is incremented (S3415), and the process returns to step S3403.
[0065]
In accordance with the shift amount searched in this way, the character is shifted to embed 1-bit information. The watermark information is embedded in the document image by performing the above processing on each character.
[0066]
[Digital watermark embedding processor]
The digital watermark (digital watermark) described below is also referred to as “invisible digital watermark”, and is a change in original image data that is hardly recognized by human vision. One of the changes or a combination of changes represents some additional information.
[0067]
FIG. 13 is a block diagram showing a configuration of an embedding processing unit (functional unit) for embedding watermark information.
[0068]
The embedding processing unit includes an image input unit 4001, an embedded information input unit 4002, a key information input unit 4003, a digital watermark generation unit 4004, a digital watermark embedding unit 4005, and an image output unit 4006. The digital watermark embedding process may be realized by software having the above-described configuration.
[0069]
The image input unit 4001 inputs image data I of an image in which watermark information is embedded. In the following description, to simplify the description, it is assumed that the image data I represents a monochrome multivalued image. Of course, if watermark information is embedded in image data composed of a plurality of color components such as color image data, each of the plurality of color components, for example, RGB components, luminance, and color difference components, is the same as a monochrome multi-valued image. Watermark information can be embedded in each component. In that case, it is possible to embed watermark information having an information amount about three times that of a monochrome multi-valued image.
[0070]
The embedding information input unit 4002 inputs watermark information to be embedded in the image data I as a binary data string. This binary data string is used as additional information Inf. The additional information Inf is composed of a combination of bits representing either “0” or “1”. The additional information Inf represents authentication information for controlling an area corresponding to the image data I, pointer information to the original data, and the like. Hereinafter, an example in which additional information Inf expressed by n bits is embedded will be described.
[0071]
Note that the additional information Inf may be encrypted so that the additional information Inf is not easily abused, or has been changed (hereinafter referred to as an “attack”) so that the additional information Inf cannot be extracted from the image data I. However, error correction coding may be performed so that the additional information Inf is correctly extracted. There may be unintentional attacks. For example, this is a case where watermark information is removed as a result of irreversible compression, luminance correction, geometric transformation, filtering, and the like of general image processing. Since processes such as encryption and error correction encoding are well known, detailed description thereof will be omitted.
[0072]
Key information input section 4003 inputs key information k necessary for embedding and extracting additional information Inf. The key information k is represented by L bits, and is “01010101” (“85” in decimal notation) or the like if L = 8. The key information k is given as an initial value of a pseudo random number generation process executed by a pseudo random number generation unit 4102 described later. Only when the embedding processing unit and the later-described extraction processing unit use the common key information k, the embedded additional information Inf is correctly extracted. In other words, only the user who owns the key information k can correctly extract the additional information Inf.
[0073]
The digital watermark generation unit 4004 receives additional information Inf from the embedded information input unit 4002 and key information k from the key information input unit 4003, and generates a digital watermark w based on the additional information Inf and the key information k. FIG. 14 is a block diagram showing details of the digital watermark generation unit 4004.
[0074]
The basic matrix generation unit 4101 generates a basic matrix m. The basic matrix m is used to associate the position of each bit constituting the additional information Inf with the pixel position of the image data I in which each bit is embedded. The basic matrix generation unit 4101 can selectively use a plurality of basic matrices, and it is necessary to change which basic matrix is used according to the purpose / situation. (Additional information Inf) can be embedded.
[0075]
FIG. 15 is a diagram illustrating an example of the basic matrix m. The matrix 4201 is an example of a basic matrix m used when embedding 16-bit additional information Inf, and numbers from 1 to 16 are assigned to each 4 × 4 element. The value of the element of the basic matrix m is associated with the bit position of the additional information Inf. In other words, the bit position of the additional information Inf is “1” (most significant bit) corresponding to the position of the element value “1” of the basic matrix m, and similarly, the additional information is set to the position of the element value “2”. The bit position of Inf corresponds to “2” (the bit next to the most significant bit).
[0076]
The matrix 4202 is an example of a basic matrix m used when embedding 8-bit additional information Inf. According to the matrix 4202, 8 bits of the additional information Inf are associated with elements having values from “1” to “8” among the elements of the matrix 4201, and the additional information Inf corresponds to elements having no value. There is no. As shown in the matrix 4202, by dispersing the positions corresponding to each bit of the additional information Inf, it is more difficult to recognize the image change (image quality degradation) due to the embedding of the additional information Inf than when the matrix 4201 is used. it can.
[0077]
Similar to the matrix 4202, the matrix 4203 is an example of a basic matrix m used when embedding 8-bit additional information Inf. According to the matrix 4202, 1-bit information is embedded in one pixel, whereas according to the matrix 4203, 1-bit information is embedded in two pixels. In other words, the matrix 4202 uses 50% of all pixels for embedding the additional information Inf, whereas the matrix 4203 uses all pixels (100%) for embedding the additional information Inf. Therefore, if the matrix 4203 is used, the number of times of embedding the additional information Inf increases, and the additional information Inf can be extracted more reliably (has attack resistance) than the matrices 4201 and 4202. The ratio of pixels used for embedding watermark information is hereinafter referred to as “filling rate”. Incidentally, the filling rate of the matrix 4201 is 100%, the filling rate of the matrix 4202 is 50%, and the filling rate of the matrix 4203 is 100%.
[0078]
The matrix 4204 has a filling rate of 100%, but only embeds 4-bit additional information Inf. Therefore, 1-bit information is embedded using four pixels, and the number of times of embedding additional information Inf is further increased and attack resistance is further improved. However, the amount of information that can be embedded is smaller than other matrices. .
[0079]
Thus, the filling rate, the number of pixels used for 1-bit embedding, and the amount of information that can be embedded can be selectively set depending on the configuration of the basic matrix m. The filling rate mainly affects the image quality of an image in which watermark information is embedded, and the number of pixels used for embedding 1 bit mainly affects attack resistance. Therefore, when the filling rate is increased, the image quality is greatly deteriorated, and when the number of pixels used for embedding 1 bit is increased, the attack resistance is increased and the amount of information that can be embedded is reduced. Thus, image quality, attack resistance, and information amount are in a trade-off relationship.
[0080]
In the present embodiment, it is possible to control and set attack resistance, image quality, and information amount by adaptively selecting a plurality of types of basic matrices m.
[0081]
The pseudorandom number generator 4102 generates a pseudorandom number sequence r based on the input key information k. The pseudo random number sequence r is a real number sequence according to a uniform distribution included in the range of {-1, 1}, and the key information k is used as an initial value for generating the pseudo random number r. That is, the pseudo random number sequence r (k1) generated using the key information k1 is different from the pseudo random number sequence r (k2) generated using the key information k2 (≠ k1). Since the method for generating the pseudo-random number sequence r is known, a detailed description thereof will be omitted.
[0082]
The pseudo random number assigning unit 4103 receives the watermark information Inf, the basic matrix m, and the pseudo random number sequence r, assigns each bit of the watermark information Inf to each element of the pseudo random number sequence r based on the basic matrix m, and assigns the digital watermark w. Generate. Specifically, each element of the matrix 4204 is scanned in raster order, the most significant bit is assigned to the element having the value “1”, the next bit is assigned to the element having the value “2”, and the additional information Each bit of Inf is made to correspond to each element of the basic matrix m. When the bit of the additional information Inf is '1', the corresponding element of the pseudo random number sequence r is left as it is, and when it is '0', the corresponding pseudo random number sequence r Multiply the element of by -1. When the above processing is executed for n bits of the additional information Inf, a digital watermark w shown as an example in FIG. 16 is obtained. Note that the digital watermark w shown in FIG. 16 has a basic matrix m as the matrix 4204 shown in FIG. 15, a pseudo-random number sequence with a real number sequence of r = {0.7, -0.6, -0.9, 0.8}, and additional information Inf (4 bits). Is an example of “1001”.
[0083]
In the above description, an example in which a 4 × 4 basic matrix m is used to embed 16-bit, 8-bit, and 4-bit additional information Inf has been described. However, the present invention is not limited to this, and more information is used to embed 1-bit information. A larger basic matrix m can be used. If a larger basic matrix m is used, a longer real number sequence is used as the pseudorandom number sequence r. Actually, the extraction process described later may not function correctly in a random number sequence composed of four elements as used in the description. That is, although the additional information Inf is embedded, there is a possibility that the correlation coefficient between the integrated image c and the digital watermarks w1, w2,. For this reason, for example, in order to embed 64-bit additional information Inf, a basic matrix m of 256 × 256 is used at a filling rate of 50%. In this case, 512 pixels are used for 1-bit embedding.
[0084]
The digital watermark embedding unit 4005 receives the image data I and the digital watermark w, and outputs the image data I ′ in which the digital watermark w is embedded. The digital watermark embedding unit 4005 executes digital watermark embedding processing according to the equation (2).
I ' _{i, j} = I _{i, j} + aw _{i, j} … (2)
Where I ' _{i, j} Is image data with embedded digital watermark
I _{i, j} Is the image data before embedding the digital watermark
w _{i, j} Is watermark
i and j are the x and y coordinate values of the image or watermark
a is a parameter to set the watermark strength
[0085]
For example, a value of about “10” can be selected as a. When a is increased, it is possible to embed a digital watermark having a high attack resistance, but the image quality deterioration is increased. On the other hand, if a is reduced, attack resistance is reduced, but image quality deterioration can also be suppressed. Similar to the configuration of the basic matrix m, it is possible to adjust the balance between attack resistance and image quality by appropriately setting the value of a.
[0086]
FIG. 17 is a diagram specifically illustrating the digital watermark embedding process shown in Expression (2). Reference numeral 4401 corresponds to the image data I ′ in which the electronic watermark is embedded, reference numeral 4402 corresponds to the image data I before the electronic watermark is embedded, and reference numeral 4403 corresponds to the electronic watermark w. As shown in FIG. 17, the calculation of Expression (2) is performed on each element in the matrix.
[0087]
The processing shown in Expression (2) and FIG. 17 is repeatedly performed on the entire image data I. When the image data I is composed of 24 × 24 pixels shown in FIG. 18, the image data I is divided into non-overlapping blocks (macroblocks) each consisting of 4 × 4 pixels, and the expression (2 ) Is executed.
[0088]
By repeatedly executing digital watermark embedding processing for all macroblocks, it is possible to embed watermark information in the entire image as a result. Since the additional information Inf composed of n bits is embedded in one macroblock, the embedded additional information Inf can be extracted if there is at least one macroblock. In other words, the extraction of the additional information Inf does not require the entire image, and it is sufficient if there is a part of the image data I (at least one macro block). The fact that the additional information Inf can be completely extracted from a part of the image data I is referred to as “cut-off resistance”.
[0089]
The image data I ′ in which the additional information Inf generated in this way is embedded as a digital watermark becomes the final output of the embedding processing unit through the image output unit 4006.
[0090]
[Digital Watermark Extraction Processing Unit]
FIG. 19 is a block diagram illustrating a configuration of an extraction processing unit (functional unit) that extracts watermark information embedded in an image.
[0091]
The extraction processing unit includes an image input unit 4601, a key information input unit 4602, a digital watermark generation unit 4603, a digital watermark extraction unit 4604, and a digital watermark output unit 4605. The digital watermark extraction process may be realized by software having the above-described configuration.
[0092]
The image input unit 4601 receives image data I ″ that may have watermark information embedded therein. The image data I ″ input to the image input unit 4601 is watermarked by the above-described embedding processing unit. It may be image data I ′ with embedded information, image data I ′ with an attack, or image data I without embedded watermark information.
[0093]
The key information input unit 4602 inputs key information k for extracting watermark information. The key information k input here must be the same as that input to the key information input unit 4003 of the embedding processing unit described above. If different key information is input, additional information cannot be extracted correctly. In other words, only the user having the correct key information k can extract the correct additional information Inf ′.
[0094]
The extraction pattern generation unit 4603 receives the key information k and generates an extraction pattern based on the key information k. FIG. 20 is a diagram illustrating details of processing of the extraction pattern generation unit 4603. The extraction pattern generation unit 4603 includes a basic matrix generation unit 4701, a pseudo random number generation unit 4702, and a pseudo random number assignment unit 4703. The basic matrix generation unit 4701 performs the same operations as the basic matrix generation unit 4101 and the pseudo random number generation unit 4702 described above, and thus detailed description thereof is omitted. However, if the basic matrix m generated by the basic matrix generation unit 4701 and the basic matrix m generated by the basic matrix generation unit 4101 are not the same for the same key information k, the additional information can be correctly extracted. Can not.
[0095]
The pseudorandom number assigning unit 4703 receives the basic matrix m and the pseudorandom number sequence r, and assigns each element of the pseudorandom number sequence r to a predetermined element of the basic matrix m. The difference between the embedding processing unit and the pseudo-random number assigning unit 4103 is that the pseudo-random number assigning unit 4103 outputs one digital watermark w, whereas the pseudo-random number assigning unit 4703 receives the number of bits of the additional information Inf. The extraction pattern wn for (n bits here) is output.
[0096]
Details of assigning each element of the pseudo-random number sequence r to a predetermined element of the basic matrix m will be described with reference to an example using the matrix 4204 shown in FIG. When the matrix 4204 is used, since 4-bit additional information Inf can be embedded, four extraction patterns w1, w2, w3, and w4 are output. Specifically, each element of the matrix 4204 is scanned in raster order, each element of the pseudo random number sequence r is assigned to an element having the value “1”, and each element of the pseudo random number sequence r is assigned to all elements having the value “1”. When the element assignment is completed, a matrix to which the pseudo random number sequence r is assigned is generated as the extraction pattern w1. FIG. 21 is a diagram illustrating an example of an extraction pattern, in which a real number sequence of r = {0.7, −0.6, −0.9, 0.8} is used as the pseudo random number sequence r. The above processing is executed for the elements having the values “2”, “3”, and “4” of the matrix 4204 to generate extraction patterns w2, w3, and w4, respectively. When the extracted patterns w1, w2, w3, and w4 generated in this way are superimposed, it becomes equal to the digital watermark w created by the embedding processing unit.
[0097]
The digital watermark extraction unit 4604 inputs the image data I ″ and the extraction patterns w1, w2,..., Wn, and extracts the additional information Inf ′ from the image data I ″. The additional information Inf ′ extracted here is desired to be equal to the embedded additional information Inf, but does not necessarily match when the image data I ′ is subjected to various attacks.
[0098]
The digital watermark extraction unit 4604 calculates the correlation between the integrated image c generated from the image data I "and the extraction patterns w1, w2, ..., wn. The integrated image c converts the image data I" into a macroblock. It is the image which divided | segmented and calculated the average value of the element value of each macroblock. FIG. 22 is a diagram for explaining an integrated image c when a 4 × 4 pixel extraction pattern and 24 × 24 pixel image data I ″ are input. The image data I ″ shown in FIG. 22 includes 36 macros. An integrated image c is obtained by dividing the block into blocks and calculating the average value of the elements of these 36 macroblocks.
[0099]
Correlations between the integrated image c generated in this way and the extraction patterns w1, w2,. The correlation coefficient is a statistic that measures the degree of similarity between the accumulated image c and the extraction pattern wn, and is expressed by Equation (3).
ρ = c ' ^T ・ W'n / | c ' ^T || w'n |… (3)
Where c 'and w'n are matrices whose elements are the difference between each element and the average value of the elements
c ' ^T Is the transpose of c '
[0100]
The correlation coefficient ρ takes a value from −1 to +1. When the positive correlation between the integrated image c and the extraction pattern wn is strong, ρ approaches +1, and when the negative correlation is strong, ρ approaches -1. “Positive correlation is strong” is the relationship “extracted pattern wn is larger as integrated image c is larger”, and “negative correlation is stronger” is relationship “extracted pattern wn is smaller as integrated image c is larger” is there. If there is no correlation between the integrated image c and the extraction pattern wn, ρ is 0.
[0101]
Based on the correlation thus calculated, whether or not the additional information Inf 'is embedded in the image data I ", and if embedded, whether each bit constituting the additional information Inf' is '1' or '0'. That is, the correlation coefficient between the integrated image c and the extraction patterns w1, w2,..., Wn is calculated, and when the calculated correlation coefficient is close to 0, “additional information is not embedded”, It is determined that “1” is embedded when the relation number is a positive number away from 0, and “0” is embedded when the correlation coefficient is a negative number away from 0.
[0102]
Obtaining the correlation is equivalent to evaluating the similarity between the integrated image c and each of the extraction patterns w1, w2,. That is, when the pattern corresponding to the extraction pattern w1, w2,..., Wn is embedded in the image data I ″ (integrated image c) by the embedding processing unit described above, a correlation value indicating a high similarity is calculated. The
[0103]
FIG. 23 is a diagram illustrating an example of extracting a digital watermark from image data I ″ (integrated image c) embedded with 4-bit additional information using w1, w2, w3, and w4.
[0104]
Correlation values between the integrated image c and the four extraction patterns w1, w2, w3, and w4 are calculated. When the additional information Inf ′ is embedded in the image data I ”(integrated image c), the correlation values are calculated as, for example, 0.9, −0.8, −0.85, and 0.7. From this result, the additional information Inf ′ is“ 1001 ”. It is possible to finally determine that 4-bit additional information Inf ′ is extracted.
[0105]
The n-bit additional information Inf ′ thus extracted is output as an extraction result of the extraction processing unit through the digital watermark output unit 4605. At that time, if the error correction encoding process or the encryption process is performed in the embedding processing unit when the additional information Inf is embedded, the error correction decoding process or the encryption decoding process is executed. The obtained information is finally output as a binary data string (additional information Inf ′).
[0106]
[Modification]
In the above description, an example in which a document watermark and a background watermark are selectively used as a watermark has been described. However, the present invention is not limited to this, and an optimal watermark method may be used for each object.
[0107]
Moreover, although the example which implement | achieves authentication control using a password was demonstrated, it is not restricted to this, You may implement | achieve by key control.
[0108]
Second Embodiment
The image processing apparatus according to the second embodiment of the present invention will be described below. Note that in the second embodiment, identical symbols are assigned to configurations similar to those in the first embodiment and detailed description thereof is omitted.
[0109]
[Constitution]
FIG. 24 is an external view showing a configuration example of a digital copying machine. A reader unit 5l that digitally reads an original image and performs predetermined image processing to generate digital image data, and a copy using the generated digital image data. The printer includes a printer 52 that generates an image.
[0110]
The document feeder 510l of the reader unit 5l supplies the documents one by one on the platen glass 5102 sequentially from the last page, and discharges the document on the platen glass 5102 after reading the document image. When the document is conveyed onto the platen glass 5102, the lamp 5103 is turned on, and the movement of the scanner unit 5104 is started, and the document is exposed and scanned. The reflected light from the original at this time is imaged on a CCD image sensor (hereinafter referred to as “CCD”) 5109 by mirrors 5105, 5106, 5107 and a lens 5108. In this way, the scanned original image is read by the CCD 5109, and the image signal output from the CCD 5109 is subjected to image processing such as shading correction and sharpness correction by the image processing unit 5110, and then the printer unit. Transferred to 52.
[0111]
The laser driver 522l of the printer unit 52 drives the laser emission unit 520l in accordance with the image data input from the reader unit 5l. Laser light output from the laser light emitting unit 5201 scans the photosensitive drum 5202 with a polygon mirror to form a latent image on the photosensitive drum 5202. A developer (toner) is attached to the latent image formed on the photosensitive drum 5202 by the developing device 5203.
[0112]
The recording paper supplied from the cassette 5204 or the cassette 5205 is conveyed to the transfer unit 5206 in synchronization with the start of laser beam irradiation, and the developer attached to the photosensitive drum 5202 is transferred. The recording paper to which the developer has been transferred is conveyed to the fixing unit 5207, and the developer is fixed to the recording paper by the heat and pressure of the fixing unit 5207. The recording paper that has passed through the fixing unit 5207 is discharged by a discharge roller 5208. The sorter 5220 sorts the recording paper by storing the discharged recording paper in each bin. The sorter 5220 stores the recording paper in the uppermost bin when the sorting is not set.
[0113]
When double-sided recording is set, the recording sheet is conveyed to the discharge roller 5208 and then guided to the refeed conveyance path by the reversely rotated discharge roller 5208 and flapper 5209. When multiple recording is set, the recording paper is not conveyed to the discharge roller 5208 but is guided to the refeed conveyance path by the flapper 5209. The recording paper guided to the refeed conveyance path is fed to the transfer unit 5206 at the timing described above.
[0114]
[processing]
FIG. 25 is a flowchart showing processing for concealing an area-designated image, which is executed by the image processing unit 5110 of the reader unit 51.
[0115]
When the image signal read from the document is input, the image processing unit 5110 generates digital image data obtained by quantizing the luminance information of a fine pixel unit with an accuracy of usually about 8 bits (S101). The spatial resolution of the pixels is about 42 μm × 42 μm, which is a resolution of about 600 pixels (600 dpi) per inch (25.4 mm). The image processing unit 5110 displays the image represented by the generated image data on the screen of the operation unit shown in FIG.
[0116]
The operation unit is usually composed of a liquid crystal display or the like that is covered with a touch panel, and can perform a desired operation by operating buttons displayed on the screen. In FIG. 6, a button 601 is a button for selecting a device mode. In the “copy” mode, the read original image is copied (output from the printer unit 52). In the “send” mode, the image data of the read original is converted into an electronic file. In the “accumulation” mode, the image data of the read original is accumulated as an electronic file in an auxiliary storage device such as a hard disk built in the apparatus. Here, assuming that the “copy” mode is selected, the button frame is indicated by a bold line.
[0117]
A display unit 602 displays basic operation conditions of the apparatus according to the selected mode, and displays the output paper size and the enlargement / reduction ratio when the copy mode is selected. The preview display unit 603 reduces and displays the entire image read by the reader unit 51. A frame 604 displayed on the preview display unit 603 indicates an area set for the preview-displayed image. An area indicated by a frame 604 (hereinafter referred to as “area”) is enlarged / reduced by a button 605 and is moved up / down / left / right by a button 606. In other words, the size and position of the area 604 on the preview display unit 603 are changed by operating the buttons 605 and 606.
[0118]
A box 607 is a text box for inputting authentication information, which will be described later. For example, a character string of about 4 digits is input using a numeric keypad (not shown), and a symbol such as “*” corresponding to the number of input character strings. Is displayed. The reason why “*” is displayed without directly displaying the input character string is to improve security.
[0119]
The image processing unit 5110 receives designation of the user's area 604 and input of authentication information using the operation unit (S102, S103), and when the designation and input are completed, the image data designated in the area 104 is input from the input image data. Cutting out (S104), discriminating the type of cut out image data (S105), selecting an image compression method according to the discrimination result (S106), and compressing the image data cut out by the selected image compression method (S107) . Next, based on the compressed image data, the identification code indicating the used image compression method, and the input authentication information, code data obtained by combining them is generated (S108), and the generated code data is converted into bits by a method described later. Conversion into map data (S109). Then, the image data in the area 604 is erased from the input image data (S110), and the image data in which the code data obtained in step S109 is inserted into the blank area after erasing is synthesized (S111), The synthesized image data is output (S112).
[0120]
Here, since the copy mode is selected as the apparatus mode, the output image data is sent to the printer unit 52, and a copy image is formed on the recording paper. Similarly, if the transmission mode is selected as the device mode, the output image data is sent to the network communication unit and electronically transferred to a predetermined destination. If the accumulation mode is selected, the output image data is accumulated in the auxiliary storage device in the apparatus.
[0121]
FIG. 27 is a flowchart for explaining in detail the processing from steps S105 to S111.
[0122]
The type of the cut-out image data is determined, and it is determined whether the area-designated image is a continuous tone image such as a photograph or a binary image such as a character / line image (S203). As a discrimination method, a method using a histogram showing the luminance distribution of the target image, a method using the occurrence frequency for each spatial frequency component, or whether or not the probability of being recognized as a “line” by pattern matching is high is used. Various methods have been proposed, and such known methods can be used.
[0123]
If the image is determined to be a character / line drawing, a histogram indicating the luminance distribution of the image is generated (S204), and an optimal threshold value is obtained from this histogram to separate the background from the character / line drawing (S205). The threshold value is used to binarize the image data (S206), and the obtained binary image data is compressed (S207). A known binary image compression method can be applied to this compression processing. Normally, as a binary image compression method, a lossless (reversible) compression method in which no loss of information occurs, for example, one of MMR compression, MR compression, MH compression, JBIG compression, or the like is used. Of course, any of the above methods can be used adaptively so that the code size after compression is minimized.
[0124]
On the other hand, if the image is determined to be a continuous tone image, resolution conversion is performed (S208). The input image data is read at, for example, 600 dpi, but it is usual that a gradation image such as a photograph does not deteriorate even at about 300 dpi. Therefore, for the purpose of reducing the final code size, for example, the image data is converted into image data equivalent to 300 dpi that has been reduced to a half size both vertically and horizontally. Then, the 300-dpi multi-value image data is compressed (S209). As a compression method suitable for a multi-valued image, a well-known JPEG compression method, JPEG2000 compression method, or the like can be used. However, these compression methods are lossy (irreversible) compression methods that are accompanied by deterioration that is normally difficult to visually identify the original image.
[0125]
Code information for identifying the compression method is added to the obtained compressed image data (S210). This is information necessary for designating a decompression method when restoring the output image to the original image. For example, the following identification codes are assigned in advance to the respective compression methods.
JPEG compression → BB
JPEG2000 → CC
MMR compression → DD
MH compression → EE
JBIG compression → FF
[0126]
Next, an authentication information code is added (S211). The authentication information is information necessary to determine whether or not the person who is going to restore has the authority when restoring the output image to the original image. At the time of restoration, restoration processing to the original image is performed only when the authentication information added here is correctly specified.
[0127]
The digital signal sequence of the code data obtained in this way is converted into binary bitmap data as a binary number (S212), and is fitted into the area 604 and synthesized (S213).
[0128]
FIG. 28 schematically shows the above operation. When an area 604 is designated for the input image data 301, the image data in the area 604 is erased and replaced with bit mapped code data.
[0129]
FIG. 29 is a diagram schematically illustrating the flow shown in FIG.
[0130]
An Sx × Sy pixel image in the area 604 is cut out, and since it is determined as a character / line image, it is binarized and lossless compressed. Then, for example, a compression scheme identification code is added to the selection of the compressed code string, authentication information is added to the head, and binary and bitmap conversion is performed. Then, a bit of Sx × Sy pixel of the same size as the area 604 Map data is generated, and the image in area 604 is replaced with this bitmap data. Of course, the addition position of the identification code and the authentication information is arbitrary as long as it is a predetermined position such as the end of the code string, the end, or a predetermined bit number. Furthermore, in order to ensure the extraction of the identification code and the authentication information, it can be repeatedly added to a plurality of positions.
[0131]
[Bitmap of coded data]
FIGS. 30 to 32 are diagrams for explaining a method of converting code data into a bitmap, and show three different methods. Each small rectangle represents one pixel of 600 dpi.
[0132]
In the method shown in FIG. 30, 600 dpi pixels are bitmapped so that 2 × 2 pixels have 1-bit information. If the code data (left side) expressed as a binary number is '1', 4 pixels of 2 x 2 pixels are set to '1' (black), and if the code data is '0', the four pixels are set to '0' (white) ). As a result, binary bitmap data having a resolution (300 dpi) of 1/2 of 600 dpi is generated. It should be noted that 2 × 2 pixels represent 1 bit of information when the original image is restored by scanning a bitmap image printed on the recording paper according to the embodiment with the reader reading accuracy, misalignment, This is to reduce the influence of magnification error and the like and accurately restore the code data from the bitmap image.
[0133]
The method shown in FIG. 31 does not set all 2 × 2 pixels to the same value. When the code data is “1”, the upper left small pixel (equivalent to 600 dpi) is set to “1” (black) among the four pixels, If '0', the lower right small pixel is set to '1' (black). According to such a configuration, the reliability when the printed bit MAP image is scanned to restore the original image is improved.
[0134]
In the method shown in FIG. 32, pixels representing 1 bit are 4 × 2 pixels, and “1” and “0” are represented by the arrangement of monochrome pixels as illustrated. This reduces the amount of data that can be recorded per unit area, but can further improve the reading accuracy when restoring the original image.
[0135]
The bitmapping method is not limited to the above method, and other various methods can be applied.
[0136]
Next, the size of the generated bitmap and the amount of information that can be embedded therein will be described.
[0137]
If the size of the area 604 (Sx × Sy pixel) is 2 inches square (about 5 cm in length and width) on the document, the original image data is 600 dpi, so both Sx and Sy are 1200 pixels. That is, the information amount of the image data in the area 604 is as follows with 8 bits per pixel.
1200 × 1200 × 8 = 11,520,000 bits = 11M bits
[0138]
When the code data is converted into a bitmap by the above-described method and replaced with the image of the area 604, in the method of FIGS. 30 and 31, the amount of information that can be recorded is 1/4 × 1/4 because one bit of information is embedded in four pixels. 8 = 1/32, and the amount of data that can be embedded in the 2-inch square area 604 is as follows.
11M / 32 = 0.34Mbit
[0139]
In other words, 11 Mbit image data must be compressed to 0.34 Mbit of 1/32, which is unrealistic. Therefore, as described above, it is necessary to discriminate the image attribute of the area 604 and adaptively switch between binarization, resolution conversion, and compression method. If the image in the area 604 is a character / line drawing, the image is binarized with the resolution kept at 600 dpi. This reduces the image data volume to 1/8 (11/8 =) 1.38M bits, and 1/4 compression is necessary to make this 0.34M bits. The compression ratio can be easily achieved by the compression method. Of course, since it is necessary to embed an identification code or authentication information of the compression method, a compression ratio higher than 1/4 is required, but it can be achieved relatively easily.
[0140]
On the other hand, in the case of a photo / gradation image, the number of gradations remains 8 bits, and the data volume is reduced to 1/4 by reducing the resolution by half (300 dpi) to (11/4 =) 2.75 Mbits. Become. In order to further reduce this to 0.34 Mbit, compression of 1/8 is required, which is a compression rate that can be achieved very easily with JPEG and JPEG2000 compression methods, suppressing image quality degradation.
[0141]
In addition, if the bit map shown in FIG. 32 is adopted, the amount of information that can be embedded is further halved and the compression rate needs to be further doubled, but this is an unrealistic value as the compression method described above. Do not mean.
[0142]
[Restore original image]
FIG. 33 is a flowchart for explaining a method of restoring the original image from the bitmap, which is executed by the image processing unit 5110 of the reader unit 51.
[0143]
The image processing unit 5110 inputs an image (S801). In the case of print output, the image may be read as a digital image by the reader unit 51, and may be input as it is as a digital image if it is electronically transmitted or stored.
[0144]
Next, the image processing unit 5110 detects an image area concealed from the input image (S802). For this detection, a method is used in which a rectangular area included in the input image is detected, and if the detected rectangular area is periodically switched between black pixels and white pixels, it is determined that the image area is concealed. To do.
[0145]
Next, an array of pixels is read from the detected image data of the concealed image area (S803), a bit map conversion method for the image data is determined, and a binary code string is restored (S804). An identification code indicating a compression method is extracted from the sequence (S805), and authentication information is extracted (S806).
[0146]
Next, the fact that there is an image concealed in the input image is displayed on the screen of the operation unit, etc., and the user is prompted to input authentication information for restoring the image (S807). When authentication information is input, it is determined whether or not the input authentication information matches the extracted authentication information (S808). If the authentication information does not match, the input image is output as it is (S813).
[0147]
If they match, the original image is restored, but the code data of the compressed image excluding the compression method identification code and the authentication information is extracted from the code string (S809), and the extracted code data is added to the extracted identification code. Corresponding compression method decompression processing is performed (S810), the decompressed image is replaced with the detected image of the concealed image area (S811), and the resultant composite image is output (S812). The image output here is an image obtained by restoring the original image before hiding the area-specified image.
[0148]
In this way, the code data obtained by efficiently compressing the area-designated partial image is converted into a bitmap and combined with the original image, thereby replacing the area-designated image with an image that cannot be visually identified. Can be concealed. When there is such an indistinguishable image (a concealed image area), the image in the area is recognized (decoded) as code data, browsed with reference to authentication information set in the code data, etc. For an authorized user, the original image can be restored based on the identification code of the compression method set in the code data.
[0149]
Therefore, a user having a predetermined authority can restore the original image and display, print, copy, transmit and / or store the original image. The authentication information may be set individually for each of display, printing, copying, transmission, and storage image operations, or grouped together, or grouped image operations such as display, printing, copying, and transmission. You may set for every unit.
[0150]
[Modification]
In the above, as shown in FIG. 28 and the like, an example has been described in which one area 604 is specified and the image of the area is concealed. However, the concealment area is not limited to one, and a plurality of areas can be designated. is there. In that case, the processing of steps S102 to S111 may be repeated for each area designation. Further, when the original image is restored from an image having a plurality of concealed image areas, the processes of steps S803 to S811 may be repeated for each detected concealed image area.
[0151]
In the above description, the information concealment, encoding method, and restoration method have been described for a document image read by a digital copying machine, but these can also be applied to documents and drawings on a PC (personal computer). . In this case, when printing of a document or graphic is instructed, a device driver corresponding to the printer to be printed is activated, and image data for print output is generated based on a print code generated by an application on the PC. As shown in FIG. 26, the device driver displays a preview of the generated image data on the user interface screen, and accepts designation of an area 604 that the user desires to hide and input of authentication information. Further, the concealed image area is detected. The subsequent processes are the same as described above, but these processes are realized by a device driver on the PC (specifically, a CPU that executes device driver software).
[0152]
In the above description, the example in which the code data is converted into a bitmap has been described. However, it is assumed that the original image cannot be accurately restored due to distortion of the print image, dirt on the recording paper, or the like. In order to avoid such a failure, if the error correction code is added to the code data and then converted into a bitmap, the reliability of the data recorded as a bitmap can be improved. Various known methods have been proposed for error correction codes, which can be used. However, since the effective amount of information that can be embedded is reduced, it is necessary to set the image compression rate higher. Of course, in order to improve not only the error correction code but also the robustness against information leakage, it is conceivable that the code data is encrypted and then converted into a bitmap.
[0153]
[Other Embodiments]
Note that the present invention can be applied to a system including a plurality of devices (for example, a host computer, an interface device, a reader, and a printer), and a device (for example, a copying machine and a facsimile device) including a single device. You may apply to.
[0154]
Another object of the present invention is to supply a storage medium (or recording medium) in which a program code of software that realizes the functions of the above-described embodiments is recorded to a system or apparatus, and the computer (or CPU or CPU) of the system or apparatus. Needless to say, this can also be achieved by the MPU) reading and executing the program code stored in the storage medium. In this case, the program code itself read from the storage medium realizes the functions of the above-described embodiments, and the storage medium storing the program code constitutes the present invention. Further, by executing the program code read by the computer, not only the functions of the above-described embodiments are realized, but also an operating system (OS) running on the computer based on the instruction of the program code. It goes without saying that a case where the function of the above-described embodiment is realized by performing part or all of the actual processing and the processing is included.
[0155]
Furthermore, after the program code read from the storage medium is written into a memory provided in a function expansion card inserted into the computer or a function expansion unit connected to the computer, the function is determined based on the instruction of the program code. It goes without saying that the CPU or the like provided in the expansion card or the function expansion unit performs part or all of the actual processing and the functions of the above-described embodiments are realized by the processing.
[0156]
When the present invention is applied to the storage medium, the storage medium stores program codes corresponding to the flowcharts described above.
[0157]
【The invention's effect】
As described above, according to the present invention, image information text In the area Adaptively according to the display / non-display selected by the user The process can be controlled.
[0158]
Also, text region about Images can be protected.
[Brief description of the drawings]
FIG. 1 is a block diagram illustrating a configuration example of an image processing system according to an embodiment;
FIG. 2 is a block diagram illustrating a configuration example of an MFP.
FIG. 3 is a flowchart for explaining an overview of processing by the image processing system;
FIG. 4 is a flowchart for explaining an overview of processing by the image processing system;
FIG. 5 is a flowchart for explaining an overview of processing by the image processing system;
FIG. 6 is a flowchart for explaining an overview of processing by the image processing system;
FIG. 7 is a diagram for explaining block selection;
FIG. 8 is a diagram showing the result of block selection;
FIG. 9 is a diagram for explaining embedding of a document watermark;
FIG. 10 is a diagram for explaining extraction of a document watermark;
FIG. 11 is a diagram for explaining a document watermark embedding rule;
FIG. 12 is a flowchart showing a process for searching for a shift amount;
FIG. 13 is a block diagram showing the configuration of an embedding processing unit (functional unit) that embeds watermark information;
FIG. 14 is a block diagram showing details of a digital watermark generation unit;
FIG. 15 is a diagram showing an example of a basic matrix;
FIG. 16 is a diagram showing an example of a digital watermark w;
FIG. 17 is a diagram showing a digital watermark embedding process;
FIG. 18 is a diagram illustrating a configuration example of image data;
FIG. 19 is a block diagram illustrating a configuration of an extraction processing unit (functional unit) that extracts watermark information embedded in an image;
FIG. 20 is a diagram showing details of processing of the extraction pattern generation unit;
FIG. 21 is a diagram showing an example of an extraction pattern;
FIG. 22 is a diagram for explaining an integrated image;
FIG. 23 is a diagram showing an example of extracting a digital watermark;
FIG. 24 is an external view showing a configuration example of a digital copying machine;
FIG. 25 is a flowchart showing processing for concealing an area-designated image, which is executed by the image processing unit of the reader unit;
FIG. 26 is a diagram showing an outline of the operation unit;
FIG. 27 is a flowchart for explaining in detail the processing from steps S105 to S111 shown in FIG.
FIG. 28 is a diagram schematically showing the flow shown in FIG. 27;
FIG. 29 is a diagram schematically explaining the flow shown in FIG. 27;
FIG. 30 is a diagram for explaining a method of converting code data into a bitmap;
FIG. 31 is a diagram for explaining a method of converting code data into a bitmap;
FIG. 32 is a diagram for explaining a method of converting code data into a bitmap;
FIG. 33 is a flowchart for describing a method of restoring an original image from a bitmap, which is executed by the image processing unit 5110 of the reader unit.

Claims (9)

An image processing method performed by an image processing apparatus,
The input means of the image processing apparatus inputs image information,
A recognition unit included in the image processing apparatus recognizes a plurality of text regions included in the input image information;
The generation unit included in the image processing apparatus generates authentication information for controlling at least one of display, printing, copying, and transmission processing for the text area for each text area,
The selection means included in the image processing apparatus selects at least one of display, printing, duplication, and transmission of the text area and then displaying or not displaying the text in the text area,
When the display unit is selected in the selection, the embedding unit included in the image processing apparatus embeds the authentication information as a digital watermark in a state where the text in the text region is visible in the text region, When the non-display is selected in the selection, the authentication information is embedded as a digital watermark in a state where the text in the text area is invisible in the text area. When the means of the image processing device makes the text invisible, the text information is stored in an external device ,
The image processing method according to claim 1, wherein position information of the stored text information is embedded as a digital watermark. Further, the means included in the image processing apparatus extracts authentication information embedded in the text area,
3. The image processing method according to claim 1, wherein at least one of display, printing, copying, and transmission processing for the text area is controlled based on the extracted authentication information. The accepting means included in the image processing device accepts input of authentication information for controlling at least one of display, printing, duplication, and transmission processing for the text area,
A determination unit included in the image processing apparatus determines whether the received authentication information is correct authentication information;
When the processing means included in the image processing apparatus is determined to be correct in the determination, at least one of display, printing, copying, and transmission processing for the text area is performed,
The image processing apparatus storage means has, when the it is determined not correct in decision saves the text information of the text area to the outside of the device,
The image processing method according to claim 3 , wherein the means included in the image processing apparatus embeds position information of the stored information as a digital watermark in a state where the text information of the text area is invisible. An input means for inputting image information;
Recognition means for recognizing a plurality of text regions included in the input image information;
Generation means for generating authentication information for controlling at least a display, printing, processing of any of replication and transmission to the text area for each of the text area,
A selection means for selecting either display or non-display of the text in the text area after performing at least one of display, printing, duplication, and transmission processing on the text area;
When display is selected by the selection means, the text in the text area is visible, and the authentication information is embedded as a digital watermark for each text area,
An image processing apparatus comprising: an embedding unit that embeds the authentication information as a digital watermark in a state where the text in the text area is invisible when non-display is selected by the selection unit. Furthermore,
Storage means for storing text information in an external device when the text is made invisible;
The image processing apparatus according to claim 5, further comprising means for embedding position information of the stored text information as a digital watermark. Extraction means for extracting authentication information embedded in the text area;
Based on the extracted authentication information, at least a display, printing, copying and image processing apparatus according to claim 5 or 6, characterized in that a control means for controlling the processing of one of transmission to the text area . The program for making a computer perform the image processing method of any one of Claims 1 thru | or 4 . A computer-readable storage medium storing the program according to claim 8 .

Images