JP3464006B2 - Digital color image reproduction device - Google Patents

Description

Description: BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a digital color image reproducing apparatus having filtering means for improving image quality. 2. Description of the Related Art Documents handled by image processing apparatuses such as digital copiers and facsimile machines are roughly classified into three types: text documents, photo documents, and halftone print documents. Then, when these originals are reproduced by the image processing apparatus, the image quality evaluation required for each is different, so that optimal processing is performed on each area by the image area separation processing. Various methods have been proposed as such image area separation processing. For example, FIG. 1 of the 1992 Institute of Image Electronics Engineers of Japan Annual Conference 40 pp 183-186
(Paper 1), or “Image Area Separation Method for Mixed Image of Text / Pattern (Dot, Photo)” Proposed by the present applicant, IEICE Transactions on Electronics, Vol. J75-DI1 No. 1 pp39
-47 There is an image area separation method described in January 1992 (Paper 2). However, the adaptive processing using the conventional image area separation processing has the following disadvantages. That is, (1) Characters on a white background are subjected to character processing by image gamut separation, but characters with halftone dots and characters on a color ground are mixed with characters and pictures due to the difficulty of image gamut separation. Made or
Alternatively, there has been a problem that most of the image processing is performed (low resolution). (2) Since image area separation is determined based on local information, some erroneous separation is inevitable. For this reason, there are some areas where the image quality is deteriorated in the photographic original and the halftone print original. (3) For a halftone-dot-printed original with a low number of lines, mixed processing of characters and pictures is performed due to the difficulty in separating the image area from characters. [0007] In order to solve the problem in the image area separation processing as described above, the above-mentioned paper 1 is used.
Then, by performing a filter process indicating edge enhancement on a character region and performing a filter process indicating a smoothing characteristic on a halftone region, the image quality of the character region is improved.
The occurrence of moire in the halftone dot area is suppressed. However, in such a processing method, since the smoothing is performed uniformly, there is a problem that the character image is deteriorated. Another method is disclosed in Japanese Patent Application Laid-Open No. 61-15716.
The technique described in Japanese Patent Application Laid-Open No. 2 (1999) -1995 performs edge enhancement processing on a character area in accordance with edge detection and performs smoothing processing on a halftone dot area. However, in the edge detection by the differential value detection in the above technique, the image quality is deteriorated (that is, the edge is erroneously emphasized due to the low recognition degree of the character and the halftone dot in the halftone print original, and as a result, the moiré increases, The area where image quality is degraded due to frequent changes in emphasis and smoothing) occupies a considerable portion. Further, in an actual digital color copying machine, since a density signal is smoothed, not only a sufficient moiré suppression effect cannot be expected for a dot-printed document, but also the color may change. An object of the present invention is to provide a character on a white background, a halftone dot,
It is an object of the present invention to provide a digital color image reproducing apparatus which improves the image quality of color ground characters and effectively removes moire generated in a halftone print original. In order to achieve the above object, according to the first aspect of the present invention, a document having a linear reflectance is provided .
R, G, and B image signals (hereinafter, referred to as first image signals)
Read out and perform predetermined processing on the first image signal.
Image signal of density y, m, c or density y, m, c, k
Signal (hereinafter referred to as a second image signal)
An image reproducing apparatus, wherein an image of interest of the first image signal is
A first method for calculating the probability that the element is a character edge on a white background
Determining means for determining the calculated character edge on a white background.
The degree of smoothing for the first image signal is determined according to the rate.
First filtering means for controlling, and the first image
Probability that the pixel of interest of the signal is a character edge or black
The second probability of calculating the probability of being a black character edge
Determination means, and the probability of the calculated character edge,
Or black probability, or black character edge probability
The degree of edge enhancement for the second image signal is
And a second filtering means for controlling . The image data input to the apparatus is linear reflectance R, G, B data. The first determination circuit calculates the probability that the pixel of interest is a character edge, and the second determination circuit calculates the probability that the pixel of interest is a character edge on a white background. The smoothing circuit controls the degree of smoothing in accordance with the result of the second determination circuit, that is, controls the degree of smoothing to be weaker as the probability of being a character on a white background increases, The emphasis circuit controls the degree of edge emphasis according to the result of the first determination circuit, that is, controls so that the degree of edge emphasis increases as the probability of being a character increases. As a result, since the edge emphasis processing is basically performed on characters on a white background, image quality is improved, and characters on halftone dots and color characters are smoothed with appropriate strength. Next, since the edge enhancement processing is performed, the image quality can be improved, and since the smoothing processing is performed on the linear reflectance data, it is possible to effectively remove the moire generated in the halftone print original. Can be. An embodiment of the present invention will be specifically described below with reference to the drawings. FIG. 1 is a block diagram of the present embodiment. 1 is a determination circuit for calculating the probability that the pixel of interest is a character edge, 2 is a determination circuit for calculating the probability that the pixel of interest is a character edge on a white background, and 3 is a smoothing circuit according to the result of the determination circuit 2. Smoothing circuit for changing the degree of the intensity of
Is an edge emphasizing circuit that changes the degree of edge emphasizing strength according to the result of the determination circuit 1. <Input Image Data> The input image data of this embodiment is linear reflectance data R, G, and B. When a halftone print original is read by a scanner using a CCD, aliasing noise due to sampling that does not exist in the original and moiré due to interference between halftone periods occur, and the image quality deteriorates significantly. In order to prevent this, it is generally effective to perform smoothing by filtering or the like, but it is necessary to apply a filter to linear reflectance data. This is for the following reason. FIG.
(A), (b) is a figure which shows before and after smoothing of a reflectance linear signal and a density linear signal. As shown in FIG. 2B, even if the linear density data is smoothed, a periodic density change still remains in a highlight portion, a dark portion, and the like, so that not only moiré cannot be removed but also the color changes. Problem. On the other hand, as shown in FIG.
When smoothing is performed on the linear reflectance data, the smoothed data becomes substantially flat (that is, there is no moiré). For this reason, in this embodiment, the R, G, and B data of the linear reflectance are input image data. <Smoothing Circuit> The smoothing circuit 3 changes the degree of smoothing according to the result of the determination circuit 2 described later. FIG. 3 is a diagram illustrating the configuration of the smoothing circuit according to the first embodiment, in which a filtering coefficient is determined according to a signal j from the determination circuit 2.
In this embodiment, the filtering size is set to 3 × 3, and four coefficients A (weak) to D (strong) shown in FIG. 4 are prepared. Then, the greater the value of j described later, that is, the greater the probability that the character is a character on a white background, the greater the coefficient selector selects a weaker coefficient, that is, the coefficient A, and performs a filtering process on the input signal. FIG. 5 is a diagram showing a configuration of a smoothing circuit according to a second embodiment. The image data S2 which is not smoothed and the image data S1 after smoothing are converted in accordance with a signal j from the decision circuit 2. Mix in proportions. Here, in the filtering for smoothing, for example, strong smoothing using the coefficient D in FIG. 4 described above is performed. For example, if j is a parameter from 0.0 to 1.0, the mixing unit calculates an output signal according to the following equation. That is, Sr = S1r × (1-j) + S2r × j Sg = S1g × (1-j) + S2g × j Sb = S1b × (1-j) + S2b × j That is, the ratio of the image signal S2 that is not smoothed increases as the probability that the character is a white background character increases. <Edge emphasis circuit> This edge emphasis circuit changes the degree of edge emphasis in accordance with the result of the judgment circuit 1 described later. FIG. 6 shows the configuration of the edge emphasizing circuit of the first embodiment, which is similar to the configuration of the smoothing circuit of FIG. A filtering coefficient is determined according to the signal h from the determination circuit 1. In this embodiment, the filtering size is 3 × 3, and four coefficients a (weak) to d (strong) shown in FIG. 7 are prepared. Then, as h described later is larger, that is, as the probability of being a character is larger, the coefficient selector selects a stronger coefficient, that is, coefficient d, and performs filtering processing on the input signal. FIG. 8 shows a configuration of the edge emphasizing circuit of the second embodiment, which is similar to the configuration of the smoothing circuit of FIG. The image data S2 without edge enhancement and the image data S1 after edge enhancement are mixed at a ratio according to the signal h from the determination circuit 1. The filtering performs strong edge enhancement using the coefficient d in FIG. 7, for example. For example, if h is a parameter from 0.0 to 1.0, the mixing unit calculates an output signal according to the following equation. That is, Sr = S1r × h + S2r × (1-h) Sg = S1g × h + S2g × (1-h) Sb = S1b × h + S2b × (1-h) As a result, the larger h described later is, that is, in characters, The higher the certain probability, the higher the ratio of the edge emphasis signal S1. FIG. 9 shows the configuration of the edge emphasizing circuit according to the third embodiment. According to the signal h from the decision circuit 1, FIG.
The filtering coefficient is determined according to the following equation. For example, a strong edge enhancement coefficient such as the coefficient d in FIG. 7 is used as the initial coefficient. <Judgment Circuit 2> The judgment circuit 2 of this embodiment is a circuit for calculating the probability that the pixel of interest is a character edge on a white background. Is a control signal generation circuit for performing As shown in FIG. 11, the determination circuit 2 includes two detection blocks, a white background detection unit and a character probability detection unit, and a comprehensive determination unit for them. The white background detection section is a block for determining a white background probability near the pixel of interest. FIG. 12A is a diagram illustrating an example of a white background detection method, and FIG. 12B is a diagram illustrating a configuration of a white background detection unit. That is, for each of R, G, and B data, the number of white pixels after binarization is counted on both sides of the target pixel, and the white background probability is calculated according to the maximum value of each count value. decide. In this embodiment, the output white background probability (S) is represented by a real number from 0 to 1. The character probability detecting section is a block for determining the character edge probability of the pixel of interest. That is, in the 3 × 3 mask, matching is performed for the four patterns shown in FIG. That is, the character degree P indicates the continuity between the target pixel and the surrounding pixels. When Th1>Th2> 0, if (all pixels xi> Th1) then character degree P1 = 2 else if (all pixels xi> Th2) then character degree P1 = 1 else then character degree P1 = 0 Is calculated. The density level of the target pixel is L, and the pixel A is
When the density level of A is A, the contrast is LA
It becomes. Thα>Thβ>Thγ>Thδ>Thε> 0
If (LA> Thα) then character degree P2 = 5 else if (LA> Thβ) then character degree P2 = 4 else if (LA−Thγ) then character degree P2 = 3 else if (LA> Thδ) then character degree P2 = 2 else if (LA> Thε) then character degree P2 = 1 is calculated. Then, the character probability detecting section adds the calculated character degree P1 and character degree P2 (character degree P = P1 +
P2) The maximum value of the character degrees P of the four patterns is set as the character degree P of the target pixel. The input signal for character probability detection uses green data (G). The comprehensive judgment section calculates a control signal for smoothing according to the equation j = S × P. Then, quantization, normalization, and the like are performed on necessary control signals according to the smoothing circuits of the first and second embodiments that are implemented as hardware. For example, for the smoothing circuit of the first embodiment, a 2-bit signal is output as j, and for the smoothing circuit of the second embodiment, j is normalized to [0, 1]. Output. <Judgment Circuit 1> FIG. 14 shows a judgment circuit 1 (character probability detection section) of the first embodiment. This determination circuit is a circuit that calculates the probability that the pixel of interest is a character edge. That is, it is a control signal generation circuit for "performing stronger edge enhancement as the probability of being a character edge is higher". FIG. 15 shows a judgment circuit 1 (black probability detection section) of the second embodiment. This determination circuit is a circuit that calculates the probability that the target pixel is black. In other words, it is a control signal generation circuit for "performing stronger edge enhancement as the probability of being black is higher". FIG. 16 shows a decision circuit 1 according to the third embodiment. This determination circuit includes a black probability detection unit, a character probability detection unit, and a comprehensive determination unit, and is a circuit that calculates the probability that the pixel of interest is a black character edge. That is, it is a control signal generation circuit for “the higher the probability of being a black character edge, the stronger the edge is emphasized”. In this embodiment, the character probability detecting section shown in FIG. 11 is used as the judgment circuit constituted by the character probability detecting section. The character probability P is a 3-bit output, and performs quantization, normalization, and the like on a control signal required for the edge enhancement circuit.
For example, for the edge emphasizing circuit of the above-described first embodiment, a signal of higher 2 bits is output as the signal h,
The signal h is normalized to [0, 1] and output to the edge emphasizing circuit of the third embodiment, and the signal h is inverted and output as the signal h to the edge emphasizing circuit of the third embodiment. (For example, when the character degree P is 7, h is set to 0, and strong edge emphasis processing using an initial coefficient is performed). The black probability detecting section is a block for calculating how close the target pixel is to black. FIG. 17 and FIG.
This is a configuration of two embodiments of a black probability detection unit. In the black probability detecting section of FIG. 17, after the linear data of the reflectance is converted into the linear data of the density by the log conversion circuit, the minimum value of each data is obtained by the minimum value circuit.
This is equivalent to calculating the maximum value of the linear reflectance image data. The output black probability K is a real number from 0 to 1. In the black probability detecting section of FIG.
After converting the linear R, G, and B data into linear density data c, m, and y by the conversion circuit, the minimum value a of each data is obtained by the minimum value circuit. Is the same as calculating the maximum value of Further, an operation represented by the following equation is performed on the data c, m, and y by an arithmetic circuit. That is, b = max (| cm |, | m
−y |, | y−c |) Then, after performing the calculation of b−a, the black probability K is normalized.
Is output. In the overall judgment section of FIG. 16, a control signal for edge enhancement is calculated by h = K × P. Then, quantization, normalization, and the like are performed on necessary control signals according to the edge emphasizing circuits of the first, second, and third embodiments that are implemented as hardware. For example, for the edge emphasizing circuit of the first embodiment, a 2-bit signal is output as h, and for the edge emphasizing circuit of the second embodiment, h is normalized to [0, 1]. Output. FIG. 19 shows an example of the configuration of the present embodiment in the case where the judging circuit 1 is configured using FIG. 16 and the judging circuit 2 is configured using FIG. FIG. 20 shows a configuration when the above embodiment is applied to a digital color image reproducing apparatus. The input image data is smoothed according to the determination result of the determination circuit 2, and is converted to density data by a log conversion circuit. UCR
The circuit generates a black component from the minimum value of the signal of the density data c ′, m ′, y ′, and subtracts the black component from the original density data to generate new density data c, m, y. That is, k = min (c ′, m ′, y ′) c = c′−km = m′−ky = y′−k The density data c, m, y, k are determined. Edge emphasis is performed in accordance with the determination result of the circuit 1 to become output image data. FIG. 21 shows the configuration of another embodiment of the digital color image reproducing apparatus. In this embodiment, an edge enhancement circuit α for a color component (c, m, y) and an edge enhancement circuit β for a black component (k) are provided as edge enhancement circuits. FIG. 22 shows filter coefficients for c, m, and y,
FIG. 23 shows filter coefficients for k. And c,
It is assumed that the coefficients of the k filter are stronger than the coefficients of the m and y filters. As a result, only the black component is subjected to stronger edge enhancement than the other color components. As described above, according to the present invention,
Then, the following effects can be obtained. (1) Since the degree of filtering is controlled according to the feature amount of the image, the image quality is further improved as compared with the conventional image area separation type processing. In addition, since the smoothing process is performed on the linear reflectance data, moire generated in the halftone print original can be effectively removed without changing the color. (2) Since smoothing is not performed on characters on a white background and only edge enhancement processing is basically performed, the image quality of characters on a white background can be improved. (3) Since the edge emphasis processing is performed on a portion where the probability of the character is high, the image quality of the character can be improved. In addition, for characters on halftone dots and colored characters, smoothing processing and then edge emphasis processing are performed with appropriate strength, so that image quality can be improved. (4) Since the edge emphasis processing is performed only on the portion where the black component is high, the image quality of the picture is improved and the image quality of the black character is improved. (5) Since edge enhancement processing is performed only on black characters, the image quality of the gray portion of the picture can be improved, and the image quality of black characters can be further improved. (6) Since the degree of edge enhancement of the black part is stronger than the degree of edge enhancement of other color components, the image quality of black characters is greatly improved.

BRIEF DESCRIPTION OF THE DRAWINGS FIG. 1 is a block diagram of the present embodiment. FIGS. 2A and 2B are diagrams showing before and after smoothing of a reflectance linear signal and a density linear signal. FIG. 3 is a diagram illustrating a configuration of a smoothing circuit according to the first embodiment. FIG. 4 shows different filtering coefficients for smoothing. FIG. 5 is a diagram illustrating a configuration of a smoothing circuit according to a second embodiment. FIG. 6 is a diagram illustrating a configuration of an edge emphasizing circuit according to the first embodiment; FIG. 7 shows different filtering coefficients for edge enhancement. FIG. 8 is a diagram illustrating a configuration of an edge emphasizing circuit according to a second embodiment. FIG. 9 is a diagram illustrating a configuration of an edge enhancement circuit according to a third embodiment; FIG. 10 is an equation for determining a filtering coefficient of the edge emphasizing circuit according to the third embodiment. FIG. 11 is a diagram illustrating a configuration of a determination circuit 2; FIG. 12A illustrates an example of a white background detection method.
FIG. 3B is a diagram illustrating a configuration of a white background detection unit. FIG. 13 is a diagram showing four types of patterns for character probability detection. FIG. 14 is a diagram illustrating a configuration of a determination circuit 1 according to the first embodiment. FIG. 15 is a diagram illustrating a configuration of a determination circuit 1 according to a second embodiment. FIG. 16 is a diagram illustrating a configuration of a determination circuit 1 according to a third embodiment. FIG. 17 is a diagram illustrating a configuration of a black probability detection unit according to the first embodiment. FIG. 18 is a diagram illustrating a configuration of a black probability detection unit according to the second embodiment. FIG. 19 uses FIG. 16 as a determination circuit 1 and a determination circuit 2
11 is a configuration example of the present embodiment in the case of using FIG. FIG. 20 is a diagram illustrating a configuration of a digital color image reproducing apparatus according to the present embodiment. FIG. 21 is a diagram illustrating another configuration of the digital color image reproducing apparatus. FIG. 22 is a diagram showing filter coefficients for c, m, and y. FIG. 23 is a diagram showing filter coefficients for k. [Explanation of Signs] 1, 2 Judgment circuit 3 Smoothing circuit 4 Edge enhancement circuit

Continuation of the front page (56) References JP-A-3-64264 (JP, A) JP-A-2-253380 (JP, A) JP-A-2-301295 (JP, A) JP-A-1-176560 (JP) , A) (58) Field surveyed (Int. Cl. ⁷ , DB name) G06T 1/00-7/60 H04N 1/409 H04N 1/46

Claims (1)

(57) [Claims 1] An R, G, B image with linear reflectance
Read out as a signal (hereinafter, referred to as a first image signal).
One image signal is subjected to predetermined processing to obtain density y, m,
c or an image signal of density y, m, c, k (hereinafter referred to as a second
Digital color image playback device that outputs image signals)
The pixel of interest of the first image signal is a character on a white background.
First determining means for calculating the probability of being an edge;
According to the probability that the character is a character edge on a white background,
A first signal for controlling the degree of smoothing for one image signal.
Filtering means and a pixel of interest of the first image signal
Probability of a character edge, or black, or black
Second determining means for calculating the probability of a character edge;
Probability of the calculated character edge or black character
Rate or the probability that the character is a black character edge.
A second filer for controlling the degree of edge enhancement for the image signal.
A digital color image reproducing apparatus comprising filtering means .