JP3715821B2 - Image processing apparatus and method - Google Patents

Description

[0001]
BACKGROUND OF THE INVENTION
The present invention relates to an image processing apparatus and method, and more particularly to an image processing apparatus and method for multiplexing additional information on image information.
[0002]
[Prior art]
Conventionally, research on multiplexing other information related to an image in image information has been actively conducted.
[0003]
In recent years, it has been referred to as digital watermarking technology. In the image information such as photographs and paintings, additional information such as the author's name and permission to use is multiplexed so that it is difficult to visually distinguish it, etc. The technology to distribute through the network is being standardized.
[0004]
As another application field, with the purpose of preventing unauthorized counterfeiting of securities such as banknotes and stamps as the image quality of image forming apparatuses such as copiers and printers increases, images output from recording paper are used. A technique is known in which additional information is embedded in an image in order to specify the output device type and its machine number.
[0005]
For example, Japanese Patent Laid-Open No. 7-123244 proposes a technique for multiplexing information by embedding additional information in a high-frequency region of a color difference component and a saturation component that are visually low in sensitivity.
[0006]
[Problems to be solved by the invention]
However, the conventional multiplexing technique has the following problems.
[0007]
FIG. 12 is a diagram schematically illustrating a conventional method for embedding additional information. According to the figure, the multiplexed information C is generated by multiplexing the image information A and the additional information B via the adder 1201. The adder 1201 may perform addition in the real space of the image information A, or after the image information A is converted into the frequency domain using Fourier transform or the like, the additional information B is synthesized into the high frequency domain or the like. Also good.
[0008]
If the multiplexed information C generated in this way can be distributed on the network without being subjected to image processing such as various types of filtering or encoding such as lossy compression, additional information from the multiplexed information C can be obtained. Decoding B can be easily performed in the past. For example, image information distributed on the Internet can be decoded through a digital filter for improving image quality, such as edge enhancement and smoothing, as long as it has some noise resistance.
[0009]
Here, it is assumed that the image forming apparatus has only an expression capability of 2 to several gradations per single color. In recent years, inkjet printers have been known to have several gradations per color by having ink with a low dye concentration or variably controlling the output dot diameter. Unless gradation processing is used, an image cannot be expressed in photographic style.
[0010]
FIG. 3 These are figures which show typically the example which performs a pseudo gradation process in the case of the multiplexing of the conventional additional information. That is, FIG. 2 In addition to the configuration shown in FIG. 3 01, the multiplexed information C is converted into quantized information D, and then the printer output 1 3 By printing out on the recording paper in step 02, the on-paper information E that is very deteriorated is obtained.
[0011]
Therefore, for the purpose of preventing forgery described above, decoding the additional information from the information on the recording paper means that FIG. 3 The additional information B is decoded from the on-paper information E after the series of processing shown in FIG. However, pseudo gradation processing 1 3 01 and printer output 1 3 The amount of change in the image information by both processes 02 is very large. Therefore, it is very difficult to multiplex the additional information so that it cannot be visually discerned and to correctly decode the multiplexed additional information from the recording paper.
[0012]
In the multiplexing technique described in JP-A-7-123244 described above, information is added to the high frequency region of an image. However, when the error diffusion method is performed in the pseudo gradation processing in the subsequent stage, the band of the additional information is buried in the texture band generated by the error diffusion due to the characteristic of the high-pass filter unique to the error diffusion method. There is probably a risk of difficulty. In addition, a highly accurate scanner device is required for accurate decoding.
[0013]
That is, when pseudo gradation processing is a premise, FIG. 3 It can be seen that the multiplexing scheme shown in FIG. In other words, there is a need for a method for multiplexing additional information that greatly utilizes the characteristics of pseudo gradation processing.
[0014]
Japanese Patent Nos. 2640939 and 2777800 are examples in which multiplexing of additional information is combined with redundancy of pseudo gradation processing.
[0015]
In the former, when binarization is performed using systematic dither, additional data is mixed in an image signal by selecting one from dither matrices representing the same gradation. However, with a systematic dither method, it is difficult to output a photographic image with high quality unless the printer has high resolution and excellent mechanical accuracy. This is because a slight difference in mechanical accuracy is generated as low-frequency noise such as horizontal stripes and is easily seen on paper. Further, when the dither matrix is periodically changed, a band of a specific frequency generated by a dither arranged regularly is disturbed, which adversely affects image quality. On the decoding side, the pixel value of the image information, which is the original signal, is unknown, and it must be decoded by guessing what dither matrix is used for binarization, and accurate decoding is difficult. .
[0016]
The latter shows a method of multiplexing additional information by the arrangement using a color dither pattern method. In this method, as in the former case, image quality deterioration due to switching of the dither matrix is inevitable. In addition, as compared with the former, more additional information can be multiplexed, but color change is caused by changing the arrangement of the color components, and image quality deterioration is particularly large in a flat portion. Decoding is also expected to become more difficult.
[0017]
In any case, both methods of changing the dither matrix have a problem that decoding is difficult although image quality deterioration is large.
[0018]
As a method for embedding a large amount of information such as audio information in an image, there is a method described in Japanese Patent No. 2833975, for example. In this method, audio information is converted into a dot code called a so-called two-dimensional bar code, and is printed in a margin portion of the image or inside the image.
[0019]
However, this method does not multiplex the dot code, which is additional information, with respect to the image information, and does not make it difficult to visually recognize the additional information (dot code). As an example that is devised that it is difficult to visually recognize the dot code, an example of embedding the code in the image using a transparent paint is described, but not only a cost increase because a special ink is required, Naturally, the image output on the recording sheet naturally deteriorates in image quality.
[0020]
The present invention has been made to solve the above problems, and an object of the present invention is to provide an image processing apparatus and method capable of multiplexing a large amount of additional information while suppressing image quality deterioration of the image information. And
[0021]
[Means for Solving the Problems]
In order to achieve the above object, an image processing apparatus of the present invention comprises the following arrangement.
[0022]
That is, an image processing apparatus that multiplexes additional information with image information, wherein a target pixel of the image information is detected. Error diffusion The additional information is indicated by quantization means for performing quantization by processing, and quantized values of a plurality of pixels including the target pixel. And a specific pattern that cannot be generated by the error diffusion processing by the quantization means at a specific density. A multiplexing unit for multiplexing the additional information, and the specific pattern formed in the multiplexing unit. Multiplexing interval And control means for controlling according to the amount of information of the additional information.
[0024]
DETAILED DESCRIPTION OF THE INVENTION
DESCRIPTION OF EMBODIMENTS Hereinafter, preferred embodiments according to the present invention will be described in detail with reference to the drawings. The image processing apparatus in each embodiment is efficiently incorporated mainly as printer driver software in a computer that creates image information to be output to the printer engine. It is also effective to incorporate it as hardware and software.
[0025]
<Outline of the invention>
An object of the present invention is to make it possible to multiplex a large amount of additional information while suppressing image quality deterioration of image information.
[0026]
As a method of multiplexing additional information while suppressing image quality degradation, the applicant uses a texture generated by the error diffusion method to artificially create a pattern of quantized values that cannot be generated by normal pseudo gradation processing. A method of multiplexing the code of the additional information by creating it has been found.
[0027]
According to this method, since the texture shape only changes microscopically, no image quality degradation is visually recognized. In addition, by creating the pattern by changing the quantization threshold of the error diffusion method, the density value of the area gradation can be maintained visually, so that multiplexing of different signals can be realized very easily.
[0028]
However, in the multiplexing method using the texture of the error diffusion method, it is assumed that the additional information to be multiplexed is mainly index information such as a model name and a model number. That is, the case where the amount of information to be embedded is very small compared to the image information has been assumed. For this reason, considering the case of multiplexing a large amount of information such as audio information, it may be difficult to multiplex all additional information in an image.
[0029]
Therefore, in the present invention, in order to be able to multiplex a large amount of additional information, a multiplex block size that is an allowable limit of image quality degradation is calculated based on a value calculated experimentally in advance. By multiplexing 1 bit within the block size, a multiplexing process according to the size of the additional information and the image size was realized.
[0030]
Preferred embodiments according to the present invention will be described below.
[0031]
<First Embodiment>
In the first embodiment according to the present invention, an example in which audio information is multiplexed with image information will be described.
[0032]
FIG. 1 is a block diagram showing a schematic configuration of an image processing system in the present embodiment. In the figure, a block 100 surrounded by a broken line indicates a transmission system, and similarly, a block 101 indicates a reception system.
[0033]
In the transmission system 100, reference numeral 102 denotes a multiplexed pseudo gradation processing unit, which is a means for multiplexing input image information and audio information to be embedded in the image. The multiplexed pseudo gradation processing unit 102 implements multiplexing and pseudo gradation processing at the same time, that is, realizes two types of different processes at once. Details of the multiplexed pseudo gradation processing will be described later. The information subjected to the multiplexed pseudo gradation processing is output on a medium such as recording paper by an output device such as a printer in the image output unit 103. Hereinafter, the information output on the recording paper is referred to as on-paper information.
[0034]
This on-paper information is sent to the transmission partner at a remote location, for example. The receiving system 101 that has received the on-paper information receives two types of information, image information and audio information, as images on the recording paper.
[0035]
In the receiving system 101, reference numeral 104 denotes an image input unit, which inputs on-paper information by an image reading device or the like. The image reading apparatus preferably has a resolution that is at least twice the printer resolution used in the image output unit 103 according to the sampling theorem. Reference numeral 105 denotes a multiplex information separation unit, which is a means for separating the multiplex information multiplexed on the input image information and obtaining independent audio information. The separated audio information can be reproduced by a reproduction device (not shown) in the reception system 101.
[0036]
Next, the multiplexed pseudo gradation processing unit 102 that is a feature of the present embodiment will be described in detail.
[0037]
FIG. 2 is a block diagram illustrating a detailed configuration of the multiplexed pseudo gradation processing unit 102. In the figure, reference numeral 10 denotes a control unit including a CPU 10a, a ROM 10b, a RAM 10c, and the like, and controls operations and processes of each component described later. In particular, the CPU 10a executes multiplexing and control of related processing, which will be described later, using the RAM 10c as a work memory in accordance with a program stored in advance in the ROM 10b.
[0038]
A multiplexing control unit 201 inputs information to be multiplexed into image information (hereinafter referred to as additional information) and controls multiplexing. A noise removing unit 202 removes a protruding noise component in the image information. This noise removal method can be realized by a known filtering technique. It is also effective to use an LPF (low pass filter) that removes not only noise components but also high frequency components of an image. A quantization unit 203 quantizes the image information in which the additional information is multiplexed into the number of gradations that can be output by the printer engine of the subsequent image output unit 103. Note that the quantization is performed by pseudo gradation processing. A multiplexing unit 204 multiplexes the code of the additional information by changing the quantization condition with reference to the quantization value pattern that has already been quantized.
[0039]
Next, the multiplexing control unit 201 will be described in detail. FIG. 3 is a flowchart illustrating a control processing procedure in the multiplexing control unit 201.
[0040]
In step S301, the number of pixels HEIGHT in the vertical direction of the image information, the number of horizontal pixels WIDTH, and the amount of information (number of bits) N of additional information such as audio information are input. Subsequently, in step S302, in addition to the additional information, the information amount α necessary for multiplexing is added to N and calculated as N2. Here, as α, header information such as a marker code indicating a multiplexed place and encoding conditions, information provided with redundancy such as an error correction code, and the like can be considered. Next, in step S303, variables w and h indicating the size of the multiplexing target block described later are calculated based on the following equations.
[0041]
w = WIDTH x N2 ^ (-1/2) Equation 1
b = HEIGHT x N2 ^ (-1/2) Equation 2
(Here ^ represents power, for example, a ^ b represents a to the bth power)
Subsequently, in step S304, it is determined whether or not both the calculated w and h are larger than a preset threshold th. On the other hand, if it is less than or equal to th, the amount of additional information is too large compared to the image size, so an error is displayed in step S305, and the process ends. This th may be determined by experimentally calculating a critical point that greatly deteriorates in image quality due to multiplexing of additional information, or of course, a value at which multiplexing cannot actually be performed is set. Also good. Needless to say, the value of th is greatly influenced by engine characteristics of a printer or the like.
[0042]
On the other hand, if w and h are both greater than th in step S304, the values of w and h are output to the multiplexing unit 203 in step S306, and the process is terminated.
[0043]
Next, the quantization unit 203 will be described in detail. FIG. 4 is a block diagram illustrating a detailed configuration of the quantization unit 203. The quantization unit 203 performs pseudo gradation processing using error diffusion processing. Details of general error diffusion processing are described in, for example, the document “R. Floyd & L. Steinberg:“ An Adaptive Alogorithm for Spatial Grayscale ”, SID Symposium Digest of Paper pp. 36-37 (1975)”. Detailed description is omitted here.
[0044]
Hereinafter, an error diffusion process in which the quantization value is binary will be described as an example.
[0045]
In FIG. 4, reference numeral 400 denotes an adder, which adds a target pixel value of input image information and a quantization error distributed from already binarized peripheral pixels. The comparison result is compared with the threshold value TH sent from the multiplexing unit 204 by the comparison unit 401. When the addition result is larger than the threshold value TH, “1” is set. Otherwise, “0” is set. Is quantized by outputting. For example, when expressing the gradation of a pixel with 8-bit accuracy, it is common to express it with “255” as the maximum value and “0” as the minimum value. It is assumed that dots (ink, toner, etc.) are printed on the recording paper when the quantization value is “1”.
[0046]
Reference numeral 402 denotes a subtracter, which calculates an error between the quantization result output from the comparison unit 401 and the addition result output from the adder 400 and outputs the error to the error distribution calculation unit 403. The error distribution calculation unit 403 distributes errors to surrounding pixels that will be subjected to future quantization processing. As the error distribution ratio, for example, an error distribution table 404 that is experimentally set based on the relative distance to the pixel of interest is owned in advance, and the error distribution ratio is based on the distribution ratio described in the distribution table 404. Distribute the error to each pixel.
[0047]
As described above, the quantization unit 203 performs a general error diffusion process, but the quantization result can be controlled by changing the threshold value TH input to the comparison unit 401. For example, when the quantization result is to be forcibly set to “1”, the threshold value TH may be set sufficiently small so that the output from the comparison unit 401 becomes “1”.
[0048]
This threshold value TH is set in the multiplexing unit 204. Hereinafter, the operation in the multiplexing unit 204 will be described in detail. The multiplexing unit 204 realizes multiplexing of the code of the additional information by referring to the quantization value pattern that has already been quantized and controlling the quantization threshold TH to be transmitted to the quantization unit 203. To do.
[0049]
FIG. 5 is a flowchart showing a control procedure of the quantization threshold TH in the multiplexing unit 204. Here, an example in which 1-bit additional information is embedded in a block composed of w × h pixels in image information will be described. It is assumed that the quantized value is binary.
[0050]
First, in step S501, a distribution state of pixel values in a block composed of w × h pixels is detected. This scans all the pixels in the block or a plurality of subsampled pixels to detect what pixel value distribution is in the block.
[0051]
Next, in step S502, a density range in which additional information is multiplexed is determined based on the detected pixel value distribution. Here, it is desirable to assign priorities for multiplexing additional information for each concentration range classified in advance. For example, when 8-bit (0 to 255) gradation expression is performed, the density areas are classified into groups of about 16 steps in advance, and the low density, high density, and medium density are arranged in order of each group. Priorities are assigned in advance. In this example, the number of groups is also 16. Therefore, the priority of each group of ranks 1 to 16 is set in the LUT (lookup table) or the like. This ranking may be set experimentally in consideration of the printer engine characteristics and the like. Naturally, since it is more difficult to separate the additional information in the high density area due to ink bleeding on the recording paper, it is preferable to multiplex the additional information with the low density area as the first priority.
[0052]
As described above, in step S502, the block is scanned, and the density area with the highest priority is determined as the density area to be multiplexed.
[0053]
In step S503, it is determined whether the code to be multiplexed is “1” or “0”. The code multiplexed in the block is 1 bit. If the code to be multiplexed is “0”, the process proceeds to step S504, and if it is “1”, the process proceeds to step S505.
[0054]
In steps S504 and S505, a window matching pattern that refers to a plurality of pixels that have already been binarized around the pixel of interest is set. That is, the density area pattern determined in step S502 (hereinafter referred to as a target pattern) is set from among a plurality of pixel value patterns registered in advance for each multiplexed density area. Since there are two types of patterns registered in advance for each density range group (corresponding to the codes “1” / “0”), patterns for the number of groups × 2 types are stored in advance in a table or the like. Is retained. Accordingly, in step S504, the pattern A corresponding to the code “0” registered in advance for the density range is set as the target pattern, and in step S505, the pattern B corresponding to the code “1” is set as the target pattern. To do.
[0055]
Here, an example of a window referred to in steps S504 and S505 is shown in FIG. In the figure, the mark * indicates the pixel of interest, and the other pixels indicated by a, b, c, d, e, f, g, and h are pixels that have already finished final binarization. Since 8 pixels are referenced in this window, there are combinations of 8 bits as patterns that can be taken by the reference pixels.
[0056]
Similarly, an example of patterns A and B set as target patterns is shown in FIG. In the figure, the values “0” and “1” for each pixel indicate quantized values, and the value of the pixel at the position indicated by “−” is unquestioned. Note that the pattern example shown in FIG. 7 is for the low density group. FIG. 7A shows an example of pattern A, that is, a case where the combination of peripheral pixels of the target pixel (corresponding to the above window) is arranged as shown in FIG. A pattern. Similarly, FIG. 7B shows an example of the pattern B, which is the target pattern when the code is “1”.
[0057]
As is clear from FIG. 7, in the pattern A, the quantization value is “1” in the pixel immediately before the horizontal direction, and in the pattern B, the quantization value is “1” in the pixel immediately before the vertical direction. It is characterized by that. Therefore, in this embodiment, although details will be described later, if the pixel array of the pattern A is detected in the block, a horizontal quantization pattern is formed, and if the pixel array of the pattern B is detected, the vertical direction is formed. A continuous quantization pattern is formed.
[0058]
Returning to FIG. 5, in step S506, the variables i, j, and a are all initialized to “0”. Note that i is a variable that counts the address in the vertical direction, j is a variable that counts the address in the horizontal direction, and a is a flag indicating whether or not a threshold is set for multiplexing.
[0059]
In step S507, the process branches depending on the flag a. If the flag a is “0”, the threshold for multiplexing has not been set, and the process proceeds to step S508. In step S508, it is determined whether the value of the target pixel is within the multiplexing target density range determined in step S502. If the pixel of interest is within the density range, the process proceeds to step S509, and whether or not the pattern in the window referring to the plurality of pixels that have already been binarized matches the target pattern set in step S504 or S505. judge.
[0060]
Here, if it matches the target pattern (pattern A or pattern B), the process proceeds to step S510, the threshold value TH is set to D, and the result is sent to the quantization unit 203. Here, as the value of D, a value at which the quantization output in the quantization unit 203 should be “1” is forcibly set. As a result, if a pixel arrangement of pattern A is detected in the block, a horizontal quantization pattern is formed, indicating a code of “0” multiplexed with the quantization pattern. On the other hand, if the pixel arrangement of the pattern B is detected, a quantization pattern continuous in the vertical direction is formed, and indicates a code of “1” multiplexed with the quantization pattern.
[0061]
On the other hand, if it does not match the target pattern, the threshold value TH is similarly set to C in step S511 and output to the quantization unit 203. Also, when it is determined NO in steps S507 and S508, the threshold value TH is set to C in step S511. As the value of C, in order to perform normal quantization, for example, “0.5”, which is an intermediate value between “0” and “1”, is generally set. Further, if the pixel value is represented by 8 bits, “128” corresponds to the threshold value C. Of course, the threshold value is not limited to the intermediate value of each quantized value.
[0062]
Only when the threshold value TH is set to D and output in step S510, the flag a is set to “1” in step S512 to indicate that the threshold setting for multiplexing in the block has been completed.
[0063]
Subsequently, in step S513, the value of the variable j is counted up, and the horizontal address is shifted by one column. In step S514, it is determined whether the variable j is less than w, which is the number of pixels in the horizontal direction of the block. If not, that is, since processing for w pixels in the horizontal direction has been completed, the variable i in the vertical direction is counted up in step S515, and processing for h rows, which is the number of pixels in the vertical direction, in step S516. It is determined whether or not the process has ended. If not, that is, since an unprocessed line remains, the process returns to step S507 to repeat a series of processes.
[0064]
As described above, the multiplexing unit 204 controls the quantization threshold TH output to the quantization unit 203 by a series of processes shown in the flowchart of FIG. This makes it possible to multiplex a 1-bit signal of additional information indicating “1” or “0” in a block of image information.
[0065]
In the flowchart of FIG. 5, for ease of explanation, an example in which output is performed in any case where the threshold value TH is set to C or D has been described. However, “TH = D If "" is overwhelmingly less than the case of "TH = C", it is also effective to set C as the initial value of the threshold value TH in the quantization unit 203. That is, the multiplexing unit 204 generates a switching signal to the quantizing unit 203 only when the pixel is “TH = D”, and the quantizing unit 203 receives D as the threshold TH only when receiving the signal. Quantization processing is executed using, and other pixels may be quantized by C.
[0066]
In the quantization threshold value control process shown in FIG. 5, there is a possibility that the pattern in the window does not match the target pattern and the scanning in the block ends. That is, it may occur that multiplexing by changing the threshold is not realized in the block. In such a case, it is practically preferable to realize the multiplexing by expressing the code by combining the quantized values so that they can be decoded at the end of the block. In addition, various methods such as a method of changing the target density range to a density range group of the next priority and re-scanning the block to try multiplexing, and a method of postponing the code to be multiplexed to the next block, etc. Can be considered.
[0067]
In addition, the method for determining w and h indicating the block size is not limited to the above-described Equations 1 and 2. According to Equations 1 and 2, blocks are formed according to the aspect ratio of the image. For example, in consideration of intra-block scanning in pseudo gradation processing, the blocks are long in the horizontal direction and short in the vertical direction. W and h may be set, and a square block may be formed.
[0068]
The configuration for performing the multiplexing process in the present embodiment has been described above. However, the idea of the present embodiment is to artificially create a texture that cannot normally be generated by the pseudo gradation process by a combination of quantized values, and to create the texture. The point is that a 1-bit code is expressed by the presence or absence of.
[0069]
In general, the error diffusion method is known as a quantization method in which a very unique texture is generated. Conventionally, the generation of textures that are chain-like and visually uncomfortable has been a major problem of the error diffusion method. For this reason, many proposals have been made to prevent the generation of such unpleasant textures.
[0070]
However, when the texture generation factor is analyzed, the texture pattern includes a texture that can be generated and a texture that cannot be generated. For example, assuming the diffusion matrix shown in FIG. 8 as the error distribution table 404, when the pixel immediately before the target pixel indicated by * is quantized to “1”, a negative quantization error occurs, so that the adjacent pixel is adjacent. The target pixel thus made becomes difficult to be quantized to “1”. Therefore, when decoding the code added according to this embodiment, for example, if it can be detected that it is a flat portion in a low density region (highlight), both pixels adjacent in the horizontal direction become “1”. Since a pattern of a quantized value is usually not possible, if this pattern is detected, it can be recognized as a code.
[0071]
Here, in general, the texture formed by continuously printing dots is very inconspicuous even when printed on recording paper with a low-resolution ink jet printer of about 360 dpi, and is completely visually disturbing to the naked eye. It will not be. However, since a high-resolution image reading apparatus can clearly detect that dots are connected and printed, the embedded code can be easily decoded.
[0072]
In this way, if the texture formed on the recording paper can be arbitrarily controlled, it is possible to embed a code visually without a sense of incongruity. This is because the frequency characteristic of the error diffusion method does not have a peak at a certain fixed frequency, but is a broadband high-pass filter.
[0073]
On the other hand, it is difficult to arbitrarily control the texture in the systematic dither method that is regularly repeated in a short period. The systematic dither method has a very large power at a fixed frequency based on the dither period, so when information is embedded in a band other than the frequency having the peak, a visually strange feeling is generated, leading to deterioration in image quality. End up.
[0074]
Although the texture in the low density range has been described above, the texture that cannot be generated in the corresponding density range can be artificially created in the middle and high density ranges other than the low density range. It is possible to embed a code. That is, it is only necessary that the decoding side can recognize and decode a texture pattern that can be discriminated in the density range that cannot be generated in the corresponding density range as a code.
[0075]
Hereinafter, an example in which the additional information according to the present embodiment is actually multiplexed on the image information will be described. FIG. 9 is a diagram illustrating an example in which three types of images are output on a recording sheet. Assume that these three types of images are image A, image B, and image C, and have the same image size. Hereinafter, a case where audio information related to an image is multiplexed on each of these three types of images will be considered. Note that the amounts of audio information multiplexed on the images A, B, and C are 4 kilobytes (kB), 1 kB, and 9 kB, respectively. For easy explanation, an increase in the amount of information due to header information, marker code, error correction code, etc. is ignored.
[0076]
In this case, since the image size is the same and the ratio of the amount of audio information to be multiplexed is 4: 1: 9 in A: B: C, the area of the block in which 1 bit of the audio information code is multiplexed is multiplexed. The ratio is 1/4: 1: 1/9. FIG. 10 shows the area ratio of this block. In FIG. 10, a region surrounded by a square in each image indicates one block in which a 1-bit code is multiplexed in each image. Naturally, the image C having the largest amount of information to be multiplexed has the smallest area of one block, and therefore the image quality deterioration due to code multiplexing becomes the most serious. That is, it can be seen that the possibility of image quality deterioration increases as the amount of information to be multiplexed increases.
[0077]
From this, in the present embodiment, it is understood that it is necessary to provide a limit based on a value (th) obtained experimentally in advance as a critical point of image quality degradation for the size (w × h) of the multiplexed block. .
[0078]
As described above, according to the present embodiment, by expressing a code with a texture that cannot normally be generated, image quality deterioration is suppressed for image information, and additional information is set to the amount of information and the size of image information. Multiplexing can be performed accordingly.
[0079]
Further, since it is possible to easily multiplex a large amount of additional information such as audio information with respect to the image information, the audio information can be distributed on the recording paper while reducing image quality degradation as much as possible. It is possible to easily mix confidential information in an image.
[0080]
Second Embodiment
Hereinafter, a second embodiment according to the present invention will be described.
[0081]
The configuration of the image processing system in the second embodiment is the same as that shown in FIGS. 1 and 2 described in the first embodiment, and only the operation in the multiplexing control unit 201 is different. To do. In the second embodiment, additional information is multiplexed for each of the four components of yellow (Y), magenta (M), cyan (C), and black (K), which are color components at the time of image formation. It is characterized by that. Therefore, it is very effective when applying an apparatus that performs color output as the image output unit 103.
[0082]
FIG. 11 is a flowchart illustrating a control process in the multiplexing control unit 201 according to the second embodiment.
[0083]
In step S1101, the number of pixels HEIGHT in the vertical direction of the image information, the number of horizontal pixels WIDTH, and the amount of information (number of bits) N of additional information such as audio information are input. Subsequently, in step S1102, the information amount α necessary for multiplexing other than the additional information is added to N and calculated as N2. Here, as α, header information such as a marker code indicating a multiplexed place and encoding conditions, information provided with redundancy such as an error correction code, and the like can be considered.
[0084]
Next, in step S1103, a variable M indicating the number of square lattices whose one side that can be taken in the image information is th is calculated based on the following equation.
[0085]
M = INT (WIDHT / th) × INT (HEIGHT / th) Equation 3
In Expression 3, INT represents an integer value, and the fractional part is rounded down.
[0086]
In other words, M indicates the number of square lattice blocks that can be arranged within the image size of WIDTH × HEIGHT, with the value of th as one side. It should be noted that th may be determined by experimentally calculating a critical point that greatly deteriorates in image quality due to multiplexing of additional information, or of course, a value that makes multiplexing impossible in practice. May be. That is, M is the maximum number of blocks, which is the limit of multiplexing, and when 1 bit is multiplexed in one block, only information up to all M bits can be multiplexed.
[0087]
In step S1104, the information amount N2 is compared with M. If N2 is smaller, the following setting for the Y component is performed in step S1105.
[0088]
w_Y = h_Y = th (4)
Here, w_Y and h_Y respectively indicate the number of horizontal pixels and the number of vertical pixels of the block at the time of Y component multiplexing.
[0089]
On the other hand, if N2 is larger than M in step S1104, the process proceeds to step S1106, where N2 is compared with twice M. If N2 is smaller than this, since multiplexing with two colors is possible, the process proceeds to step S1107, and in addition to setting w_Y and h_Y for the Y component, the following settings for the M component are further performed.
[0090]
w_M = h_M = ((WIDTH × HEIGHT) / (N2−M)) ^ (1/2) Equation 5
Here, w_Y and h_Y indicate the number of horizontal pixels and the number of vertical pixels of the block at the time of M component multiplexing, respectively. The setting of w_Y and h_Y is the same as that in step S1105. That is, in step S1107, the maximum number of multiplexed blocks is created with the Y component, and the remaining information amount that cannot be multiplexed is multiplexed with the M component. At this time, the block size (w_M, h_M) of the M component is larger than the block size (w_Y, h_Y) at the time of multiplexing the Y component.
[0091]
On the other hand, if N2 is larger than 2 × M in step S1106, the process proceeds to step S1108, where N2 is compared with three times M. If N2 is smaller, it is possible to multiplex with three colors, so the process proceeds to step S1109, and in addition to the settings of w_Y and h_Y for the Y component, the following settings for the M component and the C component are performed. .
[0092]
w_M = h_M = th Equation 6
w_C = h_C = ((WIDTH × HEIGHT) / (N2-2 × M)) ^ (1/2) Equation 7
Here, w_C and h_C respectively indicate the number of horizontal pixels and the number of vertical pixels of the block at the time of C component multiplexing. The setting of w_Y and h_Y is the same as that in step S1105, and the setting of w_M and h_M is substituted with the value of th similarly to w_Y and h_Y. In step S1109, that is, the minimum block is created for both Y and M components, and multiplexing is performed up to the limit. Information that cannot be multiplexed yet is multiplexed to the C component.
[0093]
On the other hand, if N2 is larger than 3 × M in step S1108, an error is displayed in step S1110, and the process is terminated.
[0094]
As described above, the second embodiment is characterized in that each color component forming an image is prioritized and multiplexed up to the limit of image quality degradation.
[0095]
In general, when a color image is formed with four components Y, M, C, and K, the Y component has the lowest visual sensitivity. As in the first embodiment, the multiplexing method in the second embodiment is a method of mixing information into the texture of the error diffusion method, so there is almost no visual image quality degradation, but still low sensitivity. More preferably, the color material is multiplexed. Therefore, in the second embodiment, multiplexing is executed in order of M and C, with Y as the first priority. Naturally, when the amount of information to be added is smaller than the image size, multiplexing can be realized with only a small amount of color material. However, as the amount of information increases, the number of color materials to be multiplexed increases.
[0096]
In the flowchart shown in FIG. 11, the multiplexing process for the K component is omitted. Of course, the multiplexing may be performed using the K component. Further, the priority order of the color materials is not limited to this. For example, a higher priority may be given to a color material that seems to have a lower visual sensitivity, such as a light ink with a reduced dye concentration used in an inkjet printer.
[0097]
In the second embodiment, a square lattice having the same number of pixels in the vertical direction and the horizontal direction has been described as an example of the multiplexed block. However, the shape of the multiplexed block is not limited to a square.
[0098]
As described above, according to the second embodiment, additional information is multiplexed for each color component up to the limit of image quality degradation for an image composed of a plurality of color components according to a predetermined priority order. Therefore, it is possible to multiplex more additional information while minimizing image quality degradation.
[0099]
In the first and second embodiments described above, examples of multiplexing audio information have been described. However, information to be multiplexed in the present invention is not limited to this.
[0100]
The present invention is characterized in that the control of multiplexing processing is linked to the amount of information added, and is not limited to the control methods described in the above embodiments. In addition to changing the threshold value, various methods are conceivable as a method of multiplexing additional information by changing the shape of the texture by combining the quantized values.
[0101]
In addition, since the multiplexing method of the present invention uses a texture of pseudo gradation processing, it is naturally necessary to refer to a plurality of pixels even when 1-bit multiplexing is performed, and the threshold value of the error diffusion method is changed. Affects the binarization result of a plurality of pixels after the target pixel. Therefore, as described above, it is preferable to calculate an optimum value experimentally so that the minimum number of pixels for forming one block is not visually conspicuous as a texture.
[0102]
In each of the above embodiments, the example in which the multiplexing condition is controlled according to the additional information and the image information has been described. However, it is natural that the multiplexing condition is also multiplexed and transmitted as header information. On the receiving side, the additional information and the image information are separated based on the multiplexing condition described in the header information.
[0103]
In the present invention, an example in which an error is displayed when the amount of additional information exceeds a predetermined value has been described. However, for example, hybrid control in which another multiplexing method is executed when the amount of additional information exceeds a predetermined value is also very high. It is valid.
[0104]
[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.
[0105]
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 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.
[0106]
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.
[0107]
When the present invention is applied to the storage medium, the storage medium stores program codes corresponding to the flowcharts described above.
[0108]
【The invention's effect】
As described above, according to the present invention, the principle of texture generation of the error diffusion method is used for code embedding, and further, the multiplexing condition is controlled according to the information amount and the image size of the additional information. On the other hand, it is possible to multiplex a large amount of additional information while suppressing image quality deterioration.
[Brief description of the drawings]
FIG. 1 is a block diagram showing a schematic configuration of an image processing system according to an embodiment of the present invention;
FIG. 2 is a block diagram showing a detailed configuration of a multiplexed pseudo gradation processing unit in the present embodiment;
FIG. 3 is a flowchart showing an operation procedure of a multiplexing control unit in the present embodiment;
FIG. 4 is a block diagram showing a detailed configuration of a quantization unit in the present embodiment;
FIG. 5 is a flowchart showing an operation procedure of a multiplexing unit in the embodiment;
FIG. 6 is a view showing an example of a binarized information window in the present embodiment;
FIG. 7 is a diagram showing an example of a target pattern in the present embodiment;
FIG. 8 is a diagram showing an example of a diffusion matrix in the error diffusion method;
FIG. 9 is a view showing an example of an image formed according to the embodiment;
FIG. 10 is a diagram showing a state of block division in a formed image;
FIG. 11 is a flowchart showing an operation procedure of the multiplexing control unit in the second embodiment according to the present invention;
FIG. 12 is a block diagram showing an example of conventional multiplexing processing;
FIG. 13 is a block diagram showing an example of a conventional multiplexing process.
[Explanation of symbols]
102 Multiplexed pseudo gradation processing unit
103 Image output unit
104 Image input unit
105 Multiple information separator
201 Multiplexing control unit
202 Noise removal unit
203 Quantizer
204 Multiplexer
400 adder
401 comparison unit
402 Subtractor
403 Error distribution calculator

Claims (11)

An image processing apparatus that multiplexes additional information with image information,
Quantization means for quantizing the pixel of interest of the image information by error diffusion processing;
By forming a specific pattern that can not occur in the error diffusion process shows the additional information by the quantization values of a plurality of pixels, and in particular the concentration by the quantizing means including the target pixel, multiplexing the additional information Multiplexing means to
Control means for controlling the multiplexing interval of the specific pattern formed in the multiplexing means according to the information amount of the additional information;
An image processing apparatus comprising: The image processing apparatus according to claim 1, wherein the control unit controls a block size of the image information for multiplexing the specific pattern according to an information amount of the additional information. The image processing apparatus according to claim 2, wherein the control unit performs control so that a block size of the image information on which the specific pattern is multiplexed is reduced according to an increase in an information amount of the additional information. The image processing apparatus according to claim 1, wherein the image information includes a plurality of color components, and the control unit controls the color component to be multiplexed of the specific pattern according to an information amount of the additional information. . The image processing apparatus according to claim 4, wherein the control unit performs control so as to increase the number of the color components to be multiplexed as the information amount of the additional information increases. The image processing apparatus according to claim 4, wherein the control unit sequentially assigns priorities from color components having low visual sensitivity, and controls the multiplexing target color components based on the priorities. The image processing apparatus according to claim 4, wherein the control unit controls the block size of the image information for multiplexing one specific pattern so that the color component with lower visual sensitivity becomes smaller. The multiplexing means forms the specific pattern by controlling a quantization condition in the quantization means based on the peripheral pixels already quantized by the quantization means and the additional information. The image processing apparatus according to claim 1 . The image processing apparatus according to claim 8, wherein the multiplexing unit controls a quantization threshold in error diffusion processing in the quantization unit. An image processing method for multiplexing additional information with image information,
A quantization step of quantizing the pixel of interest of the image information by error diffusion processing;
By forming a specific pattern indicates the additional information by the quantization values of a plurality of pixels, and that can not occur in the error diffusion processing in the quantization step at specific concentrations that include the pixel of interest, multiplexing the additional information A multiplexing process,
A control step of controlling the multiplexing interval of the specific pattern formed in the multiplexing step according to the information amount of the additional information;
An image processing method comprising: A recording medium recording a program code for image processing for multiplexing additional information with respect to image information, wherein the program code is at least
A quantization process code for quantizing the target pixel of the image information by error diffusion processing;
By forming a specific pattern indicates the additional information by the quantization values of a plurality of pixels, and that can not occur in the error diffusion processing in the quantization step at specific concentrations that include the pixel of interest, multiplexing the additional information The multiplexing process code to be
A control step code for controlling the multiplexing interval of the specific pattern formed in the multiplexing step according to the amount of information of the additional information;
A recording medium comprising:

Images