JP3684178B2 - Image processing apparatus and image processing method - Google Patents

Description

[0001]
BACKGROUND OF THE INVENTION
The present invention relates to an image processing apparatus and an image processing method, and in particular, in image information, information other than the image information, for example, voice information, text document information, various information about an image, completely different image information, and the like. As additional information, the present invention relates to image processing that is embedded in a visually inconspicuous manner.
[0002]
[Prior art]
Conventionally, research on multiplexing other information related to an image in image information has been actively conducted. In recent years, it has been referred to as digital watermarking technology. In 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 them, and networks such as the Internet The technology distributed through is being standardized.
[0003]
Another application field is output from images output on paper to prevent unauthorized counterfeiting of banknotes, stamps, securities, etc. as image quality of image output devices such as copiers and printers increases. There is a technique for embedding additional information in an image in order to specify a device and its machine number.
[0004]
For example, Japanese Patent Application 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 with low visual sensitivity.
[0005]
However, the technique described above has the following problems. FIG. 14 is a diagram showing embedding of general additional information in the digital watermark technique. Image information A and additional information B are multiplexed through an adder 1401 and changed to multiplexed information C. FIG. 14 shows an example in which additional information is multiplexed in the real space area of image information. If it is possible to distribute this multiplexed information C without performing various image processing such as filtering or encoding such as lossy compression, it is possible to decode the additional information B from the multiplexed information C even in the prior art. Easy. Image information distributed on the Internet can be decoded through a digital filter for improving image quality such as edge enhancement and smoothing if it has some noise tolerance.
[0006]
However, it is assumed that the multiplexed image is printed by an output device such as a printer, and additional information is extracted from the printed matter. In addition, a printer output is assumed in which the printer to be used has an expression capability of about 2 to several gradations per single color. In recent years, inkjet printers have been launched on the market that can display several gradations per single color by using ink with a low dye concentration or variably controlling the output dot diameter. Unless tone is used, the gradation of a photographic image cannot be expressed.
[0007]
That is, in the above-described assumption that the multiplexing method using the digital watermark technique of FIG. 14 is output to the printer, the multiplexed information C is changed to the quantized information D by the pseudo gradation processing 1501 as shown in FIG. After that, by printing on the paper with the printer output 1502, the information changes to the on-paper information (printed matter) of E, which is very degraded. Therefore, decoding additional information from information on paper for the purpose of preventing forgery described above means decoding additional information B from on-paper information E after a series of processing in FIG. The amount of change in information by both processes 1501 and 1502 is very large, and it is very difficult to multiplex additional information so that it cannot be visually discriminated, and to correctly decode the multiplexed additional information from the paper. become.
[0008]
FIG. 16 shows an example of a conventional digital watermark technique in which image information is converted into a frequency domain using a Fourier transform or the like, and then synthesized into a high frequency area or the like instead of the real space domain. In FIG. 16, image information is converted into a frequency domain by an orthogonal transformation process 1601, and additional information is added by a adder 1602 to a specific frequency that is difficult to visually distinguish. After returning to the real space region again by 1603 inverse orthogonal transform processing, this corresponds to passing through a filter with large changes such as pseudo gradation processing and printer output, as in the example of FIG.
[0009]
FIG. 17 shows a procedure for separating additional information from paper. That is, information on the printed matter is input via the image reading device 1701 such as a scanner. Since the input information is an image whose gradation is expressed by pseudo gradation processing, restoration processing 1702 which is reverse pseudo gradation processing is performed. The restoration process generally uses an LPF (low pass filter). After the restored information is subjected to orthogonal transform processing by 1703, in the separation processing 1704, the additional information embedded from the power of a specific frequency is separated.
[0010]
As is clear from FIGS. 16 and 17, it can be seen that a large number of complicated processing steps are passed from the multiplexing of the additional information to the separation. In the case of a color image, this series of processing steps includes a color conversion process for converting to a color unique to the printer. In order to achieve good separation even in such a complicated processing step, a very strong signal must be input. It is difficult to input a strong signal while maintaining good image quality. In addition, the large number of processing steps and the complexity make the processing time required for multiplexing and separation very long.
[0011]
Further, in the above-mentioned Japanese Patent Laid-Open No. 7-123244, information is added to the high frequency region. However, when the error diffusion method is performed in the pseudo gradation processing in the subsequent stage, the high-pass filter characteristic unique to the error diffusion method is used. The additional information band is buried in the texture band generated by error diffusion, and there is a possibility that decoding may fail. In addition, a highly accurate scanner device is required for decoding.
[0012]
In other words, it is understood that the methods shown in FIGS. 15 and 16 are not suitable when pseudo gradation processing is assumed. In other words, a method of multiplexing additional information that greatly utilizes the characteristics of pseudo gradation processing is required.
[0013]
Special registration 2640939 and special registration 2777800 are examples in which multiplexing of additional information and redundancy of pseudo gradation processing are combined.
[0014]
In the former, when binarization is performed by the systematic dither method, data is mixed in the image signal by selecting one of the dither matrices representing the same gradation.
[0015]
However, with the systematic dither method, it is difficult to output a photographic image with high image quality unless the printer has high resolution and excellent mechanical accuracy. This is because a slight deviation in mechanical accuracy occurs as low-frequency noise such as horizontal stripes and is easily visible on paper. In addition, when the dither matrix is periodically changed, a band of a specific frequency generated by the regularly arranged dither is disturbed, which adversely affects image quality. Also, the gradation expression ability varies greatly depending on the type of dither matrix. Particularly on paper, since the change in area ratio due to dot overlap or the like varies depending on the dither matrix, it may be possible to cause a change in density by switching the dither matrix even in a region having a uniform density on the signal.
[0016]
In addition, for the decoding (separation) side, a decoding method that estimates what dither matrix is used for binarization while the pixel value of the image information that is the original signal is unknown may cause erroneous decoding. Very big.
[0017]
The latter is a method of multiplexing additional information according to the arrangement using a color dither pattern method. In this method, as in the former case, image quality deterioration cannot be avoided by switching. 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. Also, it is expected that decoding on paper will become even more difficult.
[0018]
In any case, both methods of changing the dither matrix have a problem that decoding is difficult although image quality deterioration is large.
[0019]
Therefore, the applicant of the present invention first uses the texture generated by the error diffusion method to artificially create a combination of quantized values that cannot be generated by normal pseudo gradation processing, thereby embedding codes. Proposed method to do.
[0020]
In this method, the texture shape only slightly changes microscopically, so that the image quality is not visually deteriorated. In addition, if a method for changing the quantization threshold of the error diffusion method is used, the density value of the area gradation is visually maintained, so that multiplexing of different signals can be realized very easily.
[0021]
However, according to the above proposal, the decoding side must determine whether the texture is artificial or not. In the printed matter output on the paper, the texture may not be reproduced well due to a deviation from a desired landing point position such as a warp of dots.
[0022]
In color images, the method of multiplexing to the color component with the lowest visual sensitivity is the mainstream, but the texture discrimination in the real space region is easily influenced by other color components, and the multiplexing information Separation becomes difficult.
[0023]
In addition, in order to solve the above-described problems, the present applicant performs amplitude modulation on the quantization threshold of the error diffusion method with a predetermined periodicity, and controls a plurality of types of the periodicity of the threshold modulation for each region. Proposed a method of embedding codes based on this periodicity by controlling the generation probability of the quantization value of pseudo gradation processing. In this method, since the relative power information in a plurality of predetermined frequency bands is an important decoding factor rather than the phase information forming the code, compared to the method of determining the position and shape of the texture described above, Good decoding can be realized even on paper.
[0024]
[Problems to be solved by the invention]
However, the above proposal has the following problems.
[0025]
That is, when considering a decoding unit that separates additional information from a printed matter, deterioration of the printed matter due to changes over time becomes a serious problem. In the case of electronic files, such deterioration factors do not need to be considered at all due to the long-lasting nature of digital data. However, in the printed matter, especially on paper output by an inkjet printer, immediately after printing and after printing. After several days, months, and years, fading and discoloration occur due to the characteristics and fastness of the ink and the recorded printing medium, and the degree of change varies.
[0026]
Until now, since the decoding method was single and the printing situation could not be understood, there was no choice but to use the same decoding method for both the printed matter immediately after printing and the printed matter after several years after fading. It was. Further, since there was no means for knowing not only the date and time of creation of the printed matter but also the environment such as temperature and humidity at the time of creation, the same decoding method could only be used for printed matter printed in any environment.
[0027]
That is, it has not been possible to propose a decoding system that can realize optimal design of detection accuracy and decoding processing time in decoding.
[0028]
The present invention has been made to solve the above-described problems, and provides an image processing apparatus and an image processing method capable of optimizing extraction accuracy and extraction time when extracting predetermined information from an image. The purpose is to do.
[0029]
[Means for Solving the Problems]
  In order to achieve the above object, an image processing apparatus of the present invention includes an image input unit that reads a recorded material in which an image in which predetermined information is embedded is formed and inputs an image corresponding to the recorded material, and the recording Related information input means for inputting related information relating to the image formation status of the object, and the predetermined information from the input imageExtractBased on the extraction means and the input related informationExtraction accuracy by the extraction meansSwitching means for switching betweenIt is characterized by that.
[0030]
  The image processing method of the present invention also relates to an image input step of reading a recorded material in which an image in which predetermined information is embedded is imaged and inputting an image corresponding to the recorded material, and an image forming state for the recorded material. A related information input step for inputting related information, and the predetermined information from the input image.ExtractBased on the extraction process and the input related informationExtraction accuracy by the extraction processSwitching step for switching betweenIt is characterized by.
[0031]
DETAILED DESCRIPTION OF THE INVENTION
Hereinafter, an embodiment according to the present invention will be described in detail with reference to the drawings. The image processing apparatus according to the present embodiment is efficiently incorporated mainly as printer driver software or application software in a computer that creates image information to be output to the printer engine. It is also effective to incorporate the printer main body as hardware and software.
[0032]
(First embodiment)
FIG. 1 is a block diagram illustrating a configuration of an image processing system according to the first embodiment.
[0033]
Reference numerals 100, 101, and 102 denote input terminals. From 100, multi-tone image information is input, and from 101, necessary additional information to be embedded in the image information is input. This additional information is information different from the image information input at the input terminal 100, for example, voice information, text document information, copyright relating to the image input at the input terminal 100, shooting date / time, shooting location, shooting. Various applications such as various information on the person and completely different image information can be considered.
[0034]
From 102, information related to the date and time when a printed matter is created from the printer is input. In this embodiment, in addition to the above-described additional information, the print date and time is embedded in the image information as header information. This printing date and time may be input by a user using a keyboard or a mouse, or may be a date and time in a clock managed on a connected computer or other system.
[0035]
Reference numeral 103 denotes an embedded information multiplexing apparatus that embeds date and time information and additional information (hereinafter referred to as embedded information together) in image information so that it is difficult to visually discriminate. The embedded information multiplexing apparatus 103 also controls the quantization of input multi-gradation image information as well as the multiplexing of embedded information.
[0036]
Reference numeral 104 denotes a printer, which outputs information created by the embedded information multiplexing apparatus 103 using a printer engine. The printer is assumed to be a printer that realizes gradation expression by using pseudo gradation processing, such as an inkjet printer or a laser printer.
[0037]
The output printed matter reads information on the printed matter using the scanner 105, and the embedded information separating device 106 separates (extracts) the embedded information embedded in the printed matter and outputs it to the output terminal 107.
[0038]
FIG. 2 is a block diagram showing a configuration of the embedded information multiplexing apparatus 103 in FIG.
[0039]
An error diffusion processing unit 200 converts the input image information into a quantization level smaller than the number of input gradations by performing pseudo gradation processing using an error diffusion method, Tone is expressed in terms of area. Details of the error diffusion processing will be described later.
[0040]
Reference numeral 201 denotes a blocking unit that divides input image information into predetermined area units. This blocking may be rectangular or may be divided into areas other than rectangular.
[0041]
Reference numeral 202 denotes a quantization condition control unit, which changes and controls the quantization condition in units of areas blocked by the blocking unit 201. The quantization condition control unit 202 controls the quantization condition in units of blocks based on the embedded information input from the input terminal 101.
[0042]
A control unit 210 includes a CPU 211, a ROM 212, a RAM 213, and the like. The CPU 211 controls the operation and processing of each component described above according to a control program stored in the ROM 212. The RAM 213 is used as a work area for the CPU 211.
[0043]
FIG. 3 is a block diagram showing details of the error diffusion processing unit 200. General error diffusion processing is described in detail in the document R. Floyd & L. Steinberg: “An Adaptive Alogorithm for Spatial Grayscale”, SID Symposium Digest of Paper pp. 36-37 (1975).
Now, an error diffusion process in which the quantized value is binary will be described as an example. The quantized value is not limited to binary, but may be multivalued, for example, ternary or quaternary.
[0044]
Reference numeral 300 denotes an adder, which adds the target pixel value of the input image information and the distributed quantization error of the peripheral pixels already binarized. The comparison unit 301 compares the quantization threshold value from the quantization condition control unit 202 and the addition result obtained by adding the error. When the comparison value is larger than the predetermined threshold value, “1” is set. Otherwise, “0” is set. Output. For example, when expressing the gradation of a pixel with an accuracy of 8 bits, it is generally expressed by “255” as the maximum value and “0” as the minimum value. Now, it is assumed that when the quantized value is “1”, dots (ink, toner, etc.) are printed on paper.
[0045]
Reference numeral 302 denotes a subtracter that calculates an error between the quantization result and the above-described addition result, and distributes the error to surrounding pixels to be subjected to future quantization processing based on the error distribution calculation unit 303. As the error distribution ratio, an error distribution table 304 experimentally set based on the relative distance to the pixel of interest is previously owned, and the error is distributed based on the distribution ratio described in the distribution table.
[0046]
The distribution table 304 in FIG. 3 shows a distribution table for four surrounding pixels, but is not limited to this.
[0047]
Next, the entire operation procedure including the quantization condition control unit 202 will be described with reference to the flowchart of FIG. Now, an example in which the quantized value is binary will be described. The quantized value is not limited to binary, but may be multivalued, for example, ternary or quaternary.
[0048]
S401 indicates initialization of the variable i. The variable i is a variable for counting vertical addresses.
[0049]
S402 indicates initialization of the variable j. The variable j is a variable for counting horizontal addresses.
[0050]
Subsequently, S403 is a determination step based on the address values of i and j, and it is determined whether or not the coordinates of i and j, which are the current processing addresses, belong to an area where the multiplexing process is to be executed.
[0051]
The multiplexed area will be described with reference to FIG. FIG. 5 shows one image having a horizontal pixel number of WIDTH and a vertical pixel number of HEIGHT.
[0052]
Assume that the embedded information is multiplexed in this image. The upper left of the image is set as the origin, and blocking is performed with N horizontal pixels and M vertical pixels. In the present embodiment, the origin is set as a reference point, but the block may be set as a reference point. When the maximum information is multiplexed in this image, N × M blocks are arranged from the reference point. That is, when the number of blocks that can be arranged in the horizontal direction is W and the number of blocks that can be arranged in the vertical direction is H, the following relationship is established.
W = INT (WIDTH / N) ... Formula 1
H = INT (HEIGHT / M) ... Formula 2
However, INT () indicates an integer part in ().
[0053]
The number of remaining pixels that cannot be divided in Equations 1 and 2 corresponds to an end when a plurality of N × M blocks are arranged, and is outside the code multiplexing region.
[0054]
In FIG. 4, if it is determined in S403 that the pixel of interest currently being processed is outside the multiplexing region, the quantization condition C is set in S404.
[0055]
On the other hand, if it is determined that it is within the multiplexed area, the embedded information to be multiplexed is read. For ease of explanation, it is assumed that the embedded information is expressed bit by bit using an array called code []. For example, assuming that the printing date / time related information is 32 bits and the additional information is 200 bits, the embedded information is 232 bits obtained by adding two types, and the array code [] is changed from code [0] to code [231]. Up to this point, each bit is stored.
[0056]
In S405, the variable bit is substituted with information in the array code [] as follows.
bit = code [INT (i / M) x W + INT (j / N)] ... Equation 3
Subsequently, it is determined whether or not the variable bit substituted in S406 is “1”. As described above, since information in the array code [] is stored one bit at a time, the value of the variable bit indicates either “0” or “1”.
[0057]
If “0” is determined in S406, the quantization condition A is set in S407, and if “1” is determined, the quantization condition B is set in S408.
[0058]
Subsequently, in S409, a quantization process is performed based on the set quantization condition. This quantization process corresponds to the error diffusion method described in FIG.
[0059]
Subsequently, in S410, the horizontal variable j is counted up, and in S411, it is determined whether or not it is less than WIDTH which is the number of horizontal pixels of the image, and the above-described processing is repeated until the number of processed pixels becomes WIDTH. When the horizontal processing is completed for the number of WIDTH pixels, the vertical variable i is counted up in S412, and it is determined in S413 whether or not it is less than HEIGHT which is the vertical pixel number of the image. Repeat the above process until. With the above operation procedure, the quantization condition can be changed in units of blocks each consisting of N × M pixels.
[0060]
Subsequently, examples of the quantization conditions A, B, and C will be described.
[0061]
Although there are various factors in the quantization condition in the error diffusion method, in this embodiment, the quantization condition is a quantization threshold. Since the use of the quantization condition C is outside the multiplexing region, any quantization threshold value may be used. As described above, when one pixel is represented by 8 bits of gradation and the quantization level is binary, the maximum value “255” and the minimum value “0” are the quantization representative value. However, “128” which is the intermediate value is often set as the quantization threshold. That is,
In the quantization condition C, the quantization threshold is fixed to “128”.
[0062]
Since the use of the quantization condition A and the quantization condition B is a block in the multiplexing region, a difference in image quality due to a difference in the quantization condition must be generated. However, the difference in image quality must be expressed so that it is difficult to distinguish visually, and it must be easily identifiable from the paper.
[0063]
FIG. 6 is an example showing the quantization conditions A and B. FIG. 6A is a diagram showing a cycle of change of the quantization threshold in the quantization condition A. In the figure, one square is assumed to be one pixel, a white square is a fixed threshold, and a gray square is a variation threshold.
[0064]
In other words, in the example of FIG. 6A, a matrix of 8 horizontal pixels and 4 vertical pixels is assembled, and a value that protrudes only from the gray square threshold is set as the threshold.
[0065]
Similarly, FIG. 6B is a diagram illustrating the period of change of the quantization threshold in the quantization condition B. In the example of FIG. 6B, unlike FIG. 6A, a matrix of 4 horizontal pixels and 8 vertical pixels is assembled, and a value that protrudes only from a gray square threshold is set as the threshold.
[0066]
As described above, when one pixel has an 8-bit gradation value as described above, for example, “128” is set as the fixed threshold value, and “48” is set as the protruding threshold value. When the quantization threshold is lowered, the quantization value of the target pixel tends to be “1” (quantization representative value “255”). That is, in both FIGS. 6A and 6B, the quantized value “1” is likely to occur due to the arrangement of gray cells in the figure. In other words, for each N × M pixel block, there are a block in which dots are generated in the arrangement of gray squares in FIG. 6A and a block in which dots are generated in the arrangement of gray squares in FIG. 6B. Will be mixed. Naturally, the matrix shown in FIG. 6A or 6B is repeated in the same block of N × M pixels.
[0067]
A slight change in the quantization threshold in the error diffusion method does not have a significant effect on image quality. In the systematic dither method, the image quality of gradation expression greatly depends on the dither pattern used. However, in the error diffusion method in which the quantization threshold is regularly changed as described above, the gradation expression that determines the image quality is the error diffusion method, so the dot arrangement may change slightly or the texture may change. The image quality of the gradation expression is hardly affected, such as occurrence changes.
[0068]
That is, even when the quantization threshold changes, the error that is the difference between the signal value and the quantized value is diffused to surrounding pixels, so that the input signal value is stored in a macro manner. In other words, the redundancy is very large with respect to dot arrangement and texture generation in the error diffusion method.
[0069]
In the above-described example, the switching is simply performed as the quantization condition A when the value of the variable bit is “0”, and as the quantization condition B when the value of the variable bit is “1”. However, the present invention is not limited to this. It is also possible to express the variable bit by a combination of quantization conditions. For example, as shown in FIG. 7, the N × M pixel block is further divided into four small blocks. When the value of the variable bit is “0”, the arrangement of FIG. It is also possible to make a difference by quantizing using the arrangement of FIG.
[0070]
Next, the embedded information separator 106 will be described.
[0071]
FIG. 8 is a block diagram illustrating a configuration of the embedded information separation device 106.
[0072]
Reference numeral 800 denotes an input terminal to which image information read by the scanner is input. The resolution of the scanner to be used is preferably equal to or higher than the resolution of the printer that creates the printed matter. Naturally, in order to accurately read dot dot information of a printed matter, the scanner side needs to have a resolution that is at least twice that of the printer side according to the sampling theorem. However, if it is equal to or greater than that, it is possible to determine that dots are scattered to some extent even if it is not accurate.
[0073]
In the present embodiment, for ease of explanation, it is assumed that the printer resolution and the scanner resolution are the same.
[0074]
Reference numeral 801 denotes a printed matter creation related information decoding unit, and among the embedded additional information, first, the printing date and time information which is header information is decoded. The printed material creation related information decoding unit 801 preferably uses an additional information decoding unit B described later, but is not limited here. First, the header information is decoded.
[0075]
Reference numeral 802 denotes a selection unit, which calculates the following variable D from the decoded print date / time information.
D = (Decoding date / time) − (Printing date / time) Expression 4
That is, the variable D is for calculating how much time has passed since the printed matter was created. The decoding date and time in this case may be input by the user using a keyboard or mouse as in the case of the printing date and time in the clock managed by the connected computer and other systems. .
[0076]
Reference numerals 803 and 804 denote an additional information decoding unit A and an additional information decoding unit B, respectively, and one is selected based on the value of D calculated by the selection unit 802.
[0077]
Reference numeral 805 denotes a switch, and the following selection is made.
1) When D <TH: Select additional information decoding unit A
2) Otherwise: Select additional information decoding unit B
Here, TH is a threshold value obtained experimentally and empirically in advance. That is, if it is determined that the month and day have not passed since the last printing, the additional information decoding unit A is selected. Otherwise, the decoding unit B is selected.
[0078]
FIG. 9 is a block diagram illustrating a configuration of the decoding unit A.
[0079]
Reference numeral 901 denotes a blocking unit, which performs blocking in units of P × Q pixels. This block must be smaller than the N × M pixels that were blocked during multiplexing. That is,
P ≦ N and Q ≦ M Formula 5
The relationship holds.
[0080]
Further, the block formation in units of P × Q pixels is performed by skipping at certain intervals. That is, the block is formed so that one block of P × Q pixel unit is included in an area assumed to be a block of N × M pixels at the time of multiplexing. The number of skip pixels is basically N horizontal pixels and M vertical pixels.
[0081]
Reference numerals 902 and 903 denote spatial filters A and B having different characteristics, respectively, and reference numeral 904 denotes a digital filtering unit that calculates a product sum with peripheral pixels. Each coefficient of the spatial filter is created in conformity with the period of the variation threshold of the quantization condition at the time of multiplexing.
[0082]
Now, it is assumed that the quantization condition in the embedded information multiplexing apparatus 106 is changed by multiplexing the additional information by using the two types of periodicity shown in FIGS. 6 (a) and 6 (b).
[0083]
Examples of the spatial filter A 902 and the spatial filter B 903 used in the embedded information separating apparatus 106 at that time are shown in FIGS. 10 (a) and 10 (b). In the figure, the central portion of 5 × 5 pixels is the pixel of interest, and the remaining 24 pixels are peripheral pixels. In the figure, the blank pixel represents that the filter coefficient is “0”. As is apparent from FIG. 10, FIGS. 10A and 10B are edge enhancement filters. In addition, the directionality of the emphasized edge matches the directionality of the variation threshold when multiplexed. That is, FIG. 10A is created to match FIG. 6A, and FIG. 10B is created to match FIG. 6B.
[0084]
Reference numeral 905 denotes a feature amount detection unit, which detects some feature amount based on the converted values after filtering from the filtering unit 904 by the spatial filter A 902 and the spatial filter B 903. Examples of the feature amount to be detected are as follows.
1. Maximum conversion value in the block after digital filter
2. Difference between the maximum and minimum conversion values in the block after digital filtering
3. Variance value of converted value in block after digital filter
In the present embodiment, the variance value shown in 3 above is used as a feature amount.
[0085]
Reference numeral 906 denotes a determination unit that compares the respective variance values and determines that the one with the larger variance value is a code. That is, if the variance value of the filtering by the spatial filter A is large, it is estimated that the quantization is performed under the quantization condition A at the time of printing. On the contrary, if the variance value of the filtering by the spatial filter B is large, the quantization condition B at the time of printing is I guess it was quantized with.
[0086]
Since the quantization condition is linked to the code of additional information (bit in Expression 3), the fact that the quantization condition can be identified corresponds to the ability to specify the multiplexed code. That is, when the quantization condition A is estimated, it can be determined that bit = 0, and when the quantization condition B is estimated, bit = 1.
[0087]
FIG. 11 is a block diagram showing the additional information decoding unit B.
[0088]
FIG. 1101 shows a blocking unit, which is the same as 902 in FIG. 9 and blocks in units of P × Q pixels.
[0089]
Reference numeral 1102 denotes an orthogonal transformation unit, which orthogonally transforms a blocked P × Q pixel. However, when performing two-dimensional orthogonal transformation, it is necessary to block the square block with Q = P. In the present embodiment, DCT (Discrete Cosine Transform) is taken as an example.
[0090]
The conversion coefficient of the two-dimensional DCT of the block consisting of P × P pixels is
However, C (x) = 1 / √2 (x = 0),
C (x) = 1 (x ≠ 0) Expression 6
Given in.
[0091]
Reference numeral 1103 denotes a class classification unit which classifies each orthogonal transform coefficient band.
[0092]
FIG. 12 shows an example of class classification when P = Q = 16. FIG. 12 shows orthogonal transform coefficients F (u, v) in one block, with the DC component on the upper left and the AC component of the remaining 255 components.
[0093]
Now, two classes, class A centering on F (4,8) and class B centering on F (8,4), are created. Two classes are indicated by bold lines in the figure. The class classification unit 1103 does not need to classify all 256 components, and only needs to classify a plurality of classes centering on a desired component. This required number of classes corresponds to the number of conditions that are quantized at the time of multiplexing. That is, the number of classes does not increase beyond the number of conditions subjected to quantization control.
[0094]
Reference numeral 1104 denotes a power comparison unit, which compares the total power of each class. In order to speed up the calculation, the absolute value of the generated conversion coefficient may be used as a substitute for power. The signal of the additional information is determined by comparing the total power of each class.
[0095]
Now, an example in which the quantization conditions A and B of FIGS. 6A and 6B are applied at the time of multiplexing will be described. As described above, in the quantization using the quantization conditions A and B, a texture in which dots are arranged in oblique directions with different angles is likely to occur. That is, in the block quantized under the quantization condition A, when orthogonal transformation processing is performed, large power is generated in class A in FIG.
[0096]
On the other hand, in the block quantized under the quantization condition B, when orthogonal transformation processing is performed, large power is generated in class B in FIG.
[0097]
That is, it is determined whether the quantization condition at the time of multiplexing the corresponding block is the quantization condition A or the quantization condition B by relatively comparing the magnitude relationship between the power of class A and class B. it can. Since the quantization condition is linked to the code of additional information (bit in Expression 3), the fact that the quantization condition can be identified corresponds to the ability to specify the multiplexed code.
[0098]
In the example of the flowchart shown in FIG. 4, since bit = 0 is set to the quantization condition A and bit = 1 is set to the quantization condition B, when the power of class A is larger, bit = 0, If the power of class B is larger, it can be determined that bit = 1.
[0099]
Although two types of decoding units have been described above, switching of the decoding units according to the present embodiment is necessary for optimal design of the decoding detection rate and the decoding time.
[0100]
That is, it is determined that discoloration or fading has not progressed for a printed matter that has not passed the time since printing, and is decoded by the decoding unit A having a fast decoding time. On the other hand, for those that have passed a considerable amount of time since printing, it is determined that discoloration and fading have progressed, and a decoding method with higher accuracy is used with priority given to the decoding detection rate over decoding time.
[0101]
In this way, by using the elapsed time from the print date and time as an evaluation factor, discoloration and fading can be predicted, and a more optimal decoding unit can be selected.
[0102]
In the present embodiment, the decoding unit has been described with two types of A and B, but naturally more than this may be used. Further, the decoding unit is not limited to this.
[0103]
In addition, a method of changing the detection accuracy with the same decoding unit is also conceivable. That is, in a decoding unit that requires higher accuracy, it is effective to perform decoding by repetition with high redundancy.
[0104]
For example, in the above-described method using orthogonal transformation using P × Q pixels (decoding unit B), a block of P × Q pixels is spatially shifted by several pixels, orthogonal transformation is performed a plurality of times, and determination is made through a plurality of class comparisons. A method for improving the accuracy of the above can be considered. In that case, it is also effective to control the elapsed time from the printing date and time so that the number of repetitions gradually increases with the elapsed time.
[0105]
Of course, the determination using multiple orthogonal transforms improves the decoding accuracy, but requires more processing time. The optimization is preferably designed empirically.
[0106]
A method of predicting and correcting the degree of fading is also conceivable. In other words, since the density of the printed matter becomes lighter due to fading, a decoding method including density correction by calculating an empirical correction value using the elapsed time from the printing date as an evaluation factor is also conceivable.
[0107]
In this case, a density correction curve depending on the elapsed time is prepared in advance as a plurality of types of tables and corrected at the time of decoding.
[0108]
Further, when the degree of fading differs depending on the color material (ink), a method of changing the color used for the decoding process is also conceivable. For example, if magenta tends to fade more easily than other colors (cyan, yellow, black), the elapsed time is used as an evaluation factor, and the degree of decoding influence from the G component that is complementary to magenta in RGB is set. It is also effective to change.
[0109]
That is, when the elapsed time has not passed, the determination is made mainly on the decoding of the G component, and the influence of the decoding of the G component is reduced as the elapsed time passes.
[0110]
As described above, according to the first embodiment, the extraction accuracy and extraction at the time of extraction can be changed by switching the extraction method for extracting the predetermined information embedded in the image based on the elapsed time from the printing date and time. Time optimization can be realized.
[0111]
(Second Embodiment)
FIG. 13 shows the configuration of the image processing system of the second embodiment.
[0112]
Since FIG. 13 is partially different from the configuration shown in FIG. 1, only different points will be described. In other words, in this embodiment, printing environment information is embedded as header information from the terminal 1301 instead of the printing date and time information of the first embodiment described above.
[0113]
The printing environment information may be temperature, humidity, atmospheric pressure (altitude), etc. when the printed matter is created. In this embodiment, only the temperature is targeted. The temperature at the time of printing may be input by a user using a keyboard or a mouse, or may be a signal from a temperature detection unit built in the printer. Now, for example, if temperature information is 16 bits and additional information is 200 bits, a total of 216 bits of embedded information is multiplexed.
[0114]
The decoding unit is the same as that in the first embodiment described above. The present embodiment is characterized in that, instead of the variable D in the first embodiment described above, the temperature information is used as a variable T to switch the decoding unit or control the decoding accuracy.
[0115]
The temperature at the time of printing is a factor that changes the dot diameter for an ink jet printer or an electrophotographic printer. That is, a printed matter at a high temperature has a large dot diameter and may be more difficult to decode than at a normal temperature. On the other hand, when printed at a low temperature, the dot diameter is small, and similarly, there is a possibility that decoding is more difficult than at normal temperature. Therefore, the variable T needs to be controlled so as to improve the decoding accuracy as it changes to high temperature or low temperature mainly at normal temperature. As described above, since the decoding accuracy is traded off with the decoding time, it is necessary to set an optimal method experimentally.
[0116]
The printing environment has been described above by taking the temperature as an example, but the same control can be realized with the other environmental information described above.
[0117]
Another possible method is to combine the print date / time information and various print environment information into a header and embed information.
[0118]
The multiplexing method and the additional information separation method are not limited. In any multiplexing method or separation method, a configuration in which the separation method is controlled based on the elapsed time from the print date or the environment information at the time of printing is effective.
[0119]
As described above, according to the second embodiment, by switching the extraction method for extracting the predetermined information embedded in the image based on the environmental information at the time of printing, the extraction accuracy and the extraction time at the time of extraction can be changed. Optimization can be realized.
[0120]
(Other embodiments)
In addition, the present invention can be applied to a system composed of a plurality of devices (for example, a host computer, interface device, reader, printer, etc.), or a device (for example, a copier, a facsimile device, etc.) composed of a single device. You may apply to.
[0121]
Another object of the present invention is to supply a storage medium (or recording medium) that records software program codes for realizing the functions of the above-described embodiments to a system or apparatus, and to perform a computer (or CPU) of the system or apparatus. Needless to say, this can also be achieved by 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.
[0122]
Further, after the program code read from the storage medium is written in a memory provided in a function expansion card inserted into the computer or a function expansion unit connected to the computer, the function expansion is performed based on the instruction of the program code. It goes without saying that the CPU or the like provided in the 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.
[0123]
【The invention's effect】
As described above, according to the present invention, by switching the extraction method for extracting predetermined information embedded in an image based on the elapsed time from the print date or the environment information at the time of printing, Optimization of extraction accuracy and extraction time can be realized.
[0124]
Further, according to the present invention, additional information can be easily multiplexed with image information, so that a service or application for embedding audio information or confidential information in image information can be provided. In addition, illegal counterfeiting such as banknotes, stamps, and securities can be suppressed, and copyright infringement of image information can be prevented.
[Brief description of the drawings]
FIG. 1 is a principal block diagram showing an image processing system of an embodiment.
2 is a principal block diagram showing the embedded information multiplexing apparatus of FIG. 1; FIG.
3 is a principal block diagram showing the error diffusion processing unit of FIG. 2;
FIG. 4 is a flowchart showing an operation procedure of a multiplexing process including a quantization control unit.
FIG. 5 shows an example of blocking.
FIG. 6 shows an example of quantization threshold change under quantization conditions.
FIG. 7 shows an example of arrangement of combinations of quantization conditions.
8 is a principal block diagram showing the embedded information separating apparatus of FIG. 1. FIG.
9 is a block diagram showing a configuration of an additional information decoding unit A in FIG.
FIG. 10 shows an example of a spatial filter.
11 is a block diagram showing a configuration of an additional information decoding unit B in FIG. 8;
FIG. 12 is an explanatory diagram of a frequency vector in a two-dimensional frequency domain.
FIG. 13 is a principal block diagram showing an embedded information separating apparatus according to a second embodiment;
FIG. 14 is a block diagram illustrating an example of conventional multiplexing.
FIG. 15 is a block diagram illustrating an example of conventional multiplexing.
FIG. 16 is a block diagram illustrating an example of conventional multiplexing;
FIG. 17 is a block diagram showing an example of separation in the conventional method

Claims (11)

Image input means for reading a recorded material in which an image in which predetermined information is embedded is formed and inputting an image corresponding to the recorded material;
Related information input means for inputting related information related to the image formation status of the recorded matter;
Extracting means for extracting the predetermined information from the input image;
An image processing apparatus comprising: a switching unit that switches extraction accuracy by the extraction unit based on the input related information.   The image processing apparatus according to claim 1, wherein the related information is a creation date and time when a recorded matter is created. The switching means estimates an elapsed time of the recorded material based on the creation date and time and an extraction date and time for extracting the predetermined information from the recorded material, and the extraction means extracts based on the estimated elapsed time. The image processing apparatus according to claim 2, wherein the accuracy is switched.   The image processing apparatus according to claim 1, wherein the related information is recorded material creation environment information.   5. The image processing apparatus according to claim 4, wherein the environmental information is temperature information at the time of creating a recorded matter.   The image processing apparatus according to claim 4, wherein the environment information is humidity information at the time of creating a recorded matter.   5. The image processing apparatus according to claim 4, wherein the environmental information is atmospheric pressure information at the time of creating a recorded matter. Said switching means selects the optimum extraction method from a plurality of extraction methods, each extraction accuracy is different, according to claim 1, wherein the switching the extraction accuracy by performing extraction by the selected extracting means Image processing device. An image input step of reading a recorded material in which an image in which predetermined information is embedded is formed and inputting an image corresponding to the recorded material;
A related information input step of inputting related information regarding the image formation status of the recorded matter;
An extraction step of extracting the predetermined information from the input image;
An image processing method comprising: a switching step of switching extraction accuracy by the extraction step based on the input related information. A program that realizes the image processing method according to claim 9 by being executed on a computer. A recording medium on which the program according to claim 10 is recorded.

Images