JP5743498B2 - Image correction apparatus and image correction method - Google Patents

Description

  The present invention relates to an image correction apparatus.

  In recent years, printer engine units such as an ink jet system and an electrophotographic system have been printed with higher density in order to perform higher quality printing. As a result, printers with a print density of 600 dpi or higher have been put on the market. However, as the print density increases, the resolution improves, but the reproducibility of fine lines of the image tends to become unstable. Therefore, a printer capable of high-density printing requires an image correction process for improving the reproducibility of fine lines.

  According to the method described in Patent Document 1, it is possible to reproduce the thin line to a user's desired thickness by determining the thickness of the thin line and performing gradation correction according to the thickness of the thin line.

  There have also been proposed several methods for quantitatively evaluating image quality by reading an image output by an output device such as a printer with an image acquisition device such as a CCD camera.

  According to the method described in Patent Document 2, the line width and the line density can be obtained from a line image read and acquired by a CCD camera.

  According to the method described in Patent Document 3, it is possible to measure the degree of unevenness of an edge (hereinafter referred to as jaggedness) and the line density from a line image read and acquired by a CCD camera.

  It is possible to correct the fine line image based on the evaluation value calculated by these measurement methods.

JP2004-153747 JP-A-10-28383 JP 10-288507 A

  However, by any of these methods, it is impossible to comprehensively evaluate and correct the six factors that reduce the reproducibility of thin lines: line width, line density, jaggedness, blur, color shift, and discontinuity. There are challenges.

  Specifically, “Line width” in Patent Document 1, “Line width” and “Line darkness” in Patent Document 2, and “Jaggedness” and “Line darkness” are considered in the evaluation method of Patent Document 3. However, other factors are not considered. In particular, even if the reproducibility of thin lines is deteriorated due to discontinuity, blur, or color shift, they are not taken into consideration. Therefore, the fluctuation of these factors could not be corrected comprehensively.

  Therefore, the object of the present invention is to comprehensively evaluate factors related to the reproducibility of thin lines such as “line width”, “line density”, “jaggedness”, “blur”, “color shift”, and “discontinuity”. An object of the present invention is to provide an image correction apparatus for correcting.

  In order to solve the above problems, an image correction apparatus according to the present invention includes a spectrum calculation unit that calculates a spectrum in a frequency space of an input image, and a reference information acquisition unit that acquires a reference spectrum in a frequency space corresponding to the input image. An evaluation unit that compares the spectrum of the input image with the reference spectrum and evaluates the input image; and a correction unit that corrects the input image according to a result of the evaluation unit; The means divides a region where the amplitude appears in the reference spectrum and a region where the amplitude does not appear, and compares the spectrum of the input image with the reference spectrum.

  According to the present invention, it is possible to comprehensively evaluate and correct a plurality of factors related to fine line reproducibility.

1 is a block diagram of an image correction apparatus according to a first embodiment. Block diagram of the line reproducibility evaluation unit 110 in FIG. Thin line reproducibility evaluation process flowchart Interference calculation flowchart Fidelity calculation flowchart DC difference calculation flowchart Amplitude spectrum representing relationship between conventional evaluation items and evaluation values of the present invention Graph of spectrum (cross-sectional view) showing relationship between conventional evaluation items and evaluation values of the present invention Flowchart of fine line correction method of embodiment 2 Diagram showing faithful and disturbing areas Block diagram of the correction processing unit 109 of FIG.

  Hereinafter, the present invention will be described in detail based on preferred embodiments with reference to the accompanying drawings. The configurations shown in the following embodiments are merely examples, and the present invention is not limited to the illustrated configurations.

  An apparatus that comprehensively evaluates the reproducibility of fine lines for image data including fine lines acquired from a reading device such as a scanner and corrects items that have a large influence on the deterioration of the fine line reproducibility will be described. The thin line reproducibility is calculated based on the difference in frequency characteristics (spectrum) between the line image to be evaluated and the line image to be referenced.

Thin Line Correction Device FIG. 1 is a block diagram showing the configuration of a thin line correction device according to this embodiment.

  The image input unit 101 acquires image data whose fine line reproducibility is to be evaluated as input image data. Here, the color space of the input image data is RGB composed of red (R), green (G), and blue (B), and the input image data is R (x, y), G (x, y), It is defined as B (x, y). Here, x and y are image coordinates.

  The image preprocessing unit 102 performs a process of cutting out (trimming) the periphery of the line so that the line in the input image data passes through the center of the image.

The first color space conversion unit 103 uses the color conversion table unit 104 to convert the color space of the input image data into an XYZ color space. This is to make the color space suitable for calculating the average value of the input image data in the paper white area (area other than the line).
X = X_LUT_3D (R, G, B)
Y = Y_LUT — 3D (R, G, B) (1)
Z = Z_LUT — 3D (R, G, B)

  The line area extraction processing unit 105 separates the input image data into a line area and a paper white (other than the line) area in order to reduce the influence of the area other than the line on the evaluation value. For extracting the line area, for example, the method by Otsu (Otsu, automatic threshold selection method based on discrimination and least square criterion, IEICE Transactions, Vol. J63-D, No. 4, pp. 349-356). , 1980) may be used.

  Also, the input image data in the paper white area is replaced with constant values X_wh, Y_wh, Z_wh (hereinafter referred to as paper white representative values). As the paper white representative value, for example, an average value of the input image data in the paper white area is used. In an area other than the line, the input image data of each pixel is not necessarily a constant value. Therefore, by using the average value of the input image data of each pixel in the paper white area as the paper white value, it is possible to suppress the influence of the paper white area on the evaluation of the fine line reproducibility.

  The second color space conversion unit 106 converts the XYZ color space to the L * a * b * color space.

At this time, the paper white representative values X_wh, Y_wh, and Z_wh are also converted into the L * a * b * color space and are set as L_wh, a_wh, and b_wh.
L = L_LUT — 3D (X, Y, Z)
a = a_LUT — 3D (X, Y, Z) (2)
b = b_LUT — 3D (X, Y, Z)

  The fine line reproducibility evaluation unit 107 calculates an evaluation value for evaluating the fine line reproducibility. The reproducibility is evaluated by comparing with information related to the fine line in the reference image stored in the reference fine line table unit 108.

  In the present embodiment, three kinds of evaluation values of “interference level”, “fidelity”, and “DC difference” are calculated for each of L, a, and b (hereinafter, L_interference level, a_disturbance level, b_disturbance level, L_fidelity, a_fidelity, b_fidelity, L_DC difference, a_DC difference, b_DC difference).

  Furthermore, a “reproducibility evaluation value” (hereinafter referred to as a reproducibility evaluation value LN), which is a value obtained by combining these nine evaluation values, is calculated, and in this embodiment, an example of calculating the above ten evaluation values is shown. . The number of evaluation values is not limited to this, and not all of them may be used. Details of the fine line reproducibility evaluation unit 107 will be described later.

  The reference thin line table unit 108 stores a thin line direct current component (hereinafter referred to as reference L_DC, reference a_DC, and reference b_DC) and a fine line amplitude spectrum necessary for the reference image in the thin line reproducibility evaluation unit 107. It should be noted that the amplitude spectrum of the reference line is in the form of a sinc function as indicated by 801 in FIG. As can be seen from 701 and 801, since the amplitude in the frequency band other than on the u-axis is zero in the reference image line, only the amplitude spectrum on the u-axis may be stored in the reference thin line table unit 108. In the present embodiment, the amplitude spectrum stored in the reference thin line table unit 108 is an amplitude spectrum on the u axis (hereinafter referred to as reference F_L faithfulness, reference F_a faithfulness, and reference F_b faithfulness). Details of reference L_DC, reference a_DC, reference b_DC, reference F_L faithfulness, reference F_a faithfulness, and reference F_b faithfulness will be described later.

  The correction processing unit 109 performs correction processing for a factor that deteriorates the reproducibility of the thin line according to the evaluation result obtained from the thin line reproducibility evaluation unit 107. Details of the correction processing will be described later.

  The image output device 110 represents a printer, a display, a projector, or the like, and outputs the corrected image data obtained from the correction processing unit.

  In the configuration of the present embodiment, there may be various components other than the above, but the description thereof is omitted because it is not the main point of the present invention.

Thin Line Reproducibility Evaluation Unit Hereinafter, the thin line reproducibility evaluation unit 107 will be described with reference to the block diagram of FIG. 2 and the flowchart of FIG.

  First, an amplitude spectrum is obtained in order to calculate the disturbance degree and the fidelity.

  In step S <b> 301, the second color space conversion unit 106 calculates input image data (L, a, b) and paper white representative values L_wh, a_wh, b_wh, and inputs them to the fine line reproducibility evaluation unit 107.

In step S302, the paper white inversion processing unit 200 subtracts the paper white representative value from the input image data so that the paper white representative values of the input image data L, a, and b become zero, and inverts the difference. To do.
L ′ (x, y) = L_wh−L (x, y)
a ′ (x, y) = a_wh−a (x, y) (3)
b ′ (x, y) = b_wh−b (x, y)

In step S303, the two-dimensional windowing processing unit 201 applies a two-dimensional window function win () to the input image data L ′ (x, y), a ′ (x, y), b ′ (x, y). x, y). By this two-dimensional windowing process, an extra high frequency is removed and a signal is cut into a finite time length for Fourier transform. A Hamming window or the like may be used for the window function.
Lw (x, y) = L ′ (x, y) win (x, y)
aw (x, y) = a ′ (x, y) win (x, y) (4)
bw (x, y) = b ′ (x, y) win (x, y)

  Next, in step S304, the DC component calculation unit 202 calculates DC components L_DC, a_DC, and b_DC of the line image. In order to calculate the direct current component, for example, the sum of the pixel values of the image after window calculated by the two-dimensional windowing processing unit 201 may be divided by the sum of the values of the window function.

Next, in step 305, the Fourier transform unit 203 creates an amplitude spectrum by Fourier transform. Note that the Fourier transform is performed after subtracting the direct current components (L_DC, a_DC, b_DC) from the values (Lw, aw, bw) subjected to the two-dimensional windowing process before the Fourier transform.
F_L (u, v) = Fourier {Lw (x, y) −L_DC × w (x, y)}
F_a (u, v) = Fourier {aw (x, y) −a_DC × w (x, y)} (6)
F_b (u, v) = Fourier {bw (x, y) −b_DC × w (x, y)}
U and v represent frequencies. Here, when u, v = 0 (the origin of the uv coordinate), a direct current component is indicated, and u, v <0 indicates a negative frequency.

  Next, in step S306, the spectrum region dividing unit 204 separates the amplitude spectrum output from the Fourier transform unit 203 as shown in FIG. Specifically, the reference line image is divided into a region where the amplitude spectrum appears (hereinafter referred to as a faithful region) and a region where the amplitude spectrum does not appear (hereinafter referred to as a disturbing region). In the case of the input image data 1005, the faithful region of the amplitude spectrum 1006 is on the u axis that is perpendicular to the direction of the line, and the disturbing region is other than on the u axis. In the case of the input image data 1007, the faithful region of the amplitude spectrum 1008 is the direction perpendicular to the thin line, and the disturbing region is other than that.

In this case, the boundary lines 1001 to 1004 that divide the faithful area and the disturbing area in FIG. 10 are c [rad] for the inclination of the line of the image, e [rad] for the tolerance of the line inclination estimation, and d [ Frequency] can be expressed by the following formula.
1001: v = tan (c−e + π / 2) u + d / 2 (sin (c) −cos (c) tan (c−e + π / 2))
1002: v = tan (c + e + π / 2) u + d / 2 (sin (c) −cos (c) tan (c + e + π / 2))
1003: v = tan (c + e + π / 2) u−d / 2 (sin (c) −cos (c) tan (c + e + π / 2))
1004: v = tan (c−e + π / 2) u−d / 2 (sin (c) −cos (c) tan (c−e + π / 2)) (7)
At this time, it can be said that the area surrounded by the boundary lines 1001 to 1004 is the faithful area, and the other area is the disturbing area. Hereinafter, the amplitude spectrum of the interference region is described as F_L interference, F_a interference, and F_b interference, and the amplitude spectrum of the faithful region is described as F_L faithful, F_a faithful, and F_b faithful.

  Next, in step S307, the interference level calculation unit 205 calculates the L_interference level, a_interference level, and b_interference level for evaluating the amplitude spectrum F_L interference, F_a interference, and F_b interference in the interference area. Details will be described later.

  Next, in step S308, the fidelity calculation unit 206 calculates L_fidelity, a_fidelity, and b_fidelity for evaluating the amplitude spectra F_L fidelity, F_a fidelity, and F_b fidelity of the fidelity region. Details of the processing will be described later.

  Next, in step S309, the DC difference calculation unit 207 obtains a DC difference for evaluating the direct current component of the input image data. Differences between the direct current components L_DC, a_DC, and b_DC of the input image data calculated in S304 and the direct current component references L_DC, reference a_DC, and reference b_DC of the line in the reference image stored in the reference thin line table unit 108 are obtained. Then, the L_DC difference, a_DC difference, and b_DC difference are calculated. Details will be described later.

Next, in step S310, the reproducibility evaluation value calculation unit 208 calculates the L_disturbance, a_disturbance, b_disturbance, L_fidelity, a_fidelity, b_fidelity, L_DC difference, a_DC obtained in S307 to S309. The difference and b_DC difference are added to each of the 9 values, and appropriate weights (αL, βL, χL, αa, βa, χa, αb, βb, χb) are added, and finally an index γall is used for γ correction. Multiply and output the value obtained.
Reproducibility evaluation value LN = (αL interference L + βL fidelity L + χLDC difference L
+ Αa Interference a + βa Fidelity a + χaDC difference a
+ Αb Interference b + βb Fidelity b + χbDC difference b) ^ γall (8)
At this time, the values of the weights (αL, βL, χL, αa, βa, χa, αb, βb, χb) and the index γall are acquired from the evaluation parameter table unit 209. It is desirable that the parameters of the evaluation parameter table unit 209 store appropriate settings depending on the type of line image to be evaluated. For example, it is desirable to use an appropriate parameter for each application such as a parameter for an electrophotographic line image and a parameter for an image captured by a display.

Interference Degree Calculation Method FIG. 4 is a flowchart showing the obstruction degree calculation processing procedure of the obstruction degree calculation unit 205 in S307. In the line image referred to as described above, no amplitude appears in the disturbing area. Therefore, it is evaluated whether the amplitude does not appear in the disturbing region of the line image to be evaluated, or how much amplitude is present when it appears. Hereinafter, the operation of the disturbance degree calculation unit 205 will be described with reference to the flowchart of FIG.

First, in S401, the spectrum (F_L interference, F_a interference, F_b interference) of each of L, a, and b interference regions input from the spectrum region dividing unit 204 is weighted with visual characteristic functions G_L, G_a, G_b. Thereby, the disturbance degree corresponding to human visibility can be calculated.
F_L obstruction w (u, v) = F_L obstruction (u, v) × G_L (u, v)
F_a interference w (u, v) = F_a interference (u, v) × G_a (u, v)
F_b interference w (u, v) = F_b interference (u, v) × G_b (u, v) (9)
As the visual characteristic functions G_L, G_a, and G_b, for example, VTF (Visual Transfer Function) or frequency response characteristics obtained from experimental results may be used. Here, VTF is a curve representing the sensitivity to human visual spatial frequency. References to VTF include R.I. Pdolly, “Prediction Brightness Appearance at Edges Using Linear and Non-Linear Visual Describing Functions”, APSE Annual Meeting 1975, and the like.

Next, in step S402, the weighted power (amplitude squared) of the disturbing area is integrated.
P_L interference = Σ F_L interference w (u, v) ^ 2
P_a interference = Σ F_a interference w (u, v) ^ 2
P_b interference = Σ F_b interference w (u, v) ^ 2 (10)
Here, Σ represents summing all u and v in the interference area.

Finally, P_L interference, P_a interference, and P_b interference are normalized using direct current components (L_DC, a_DC, b_DC) in S403. Specifically, a value obtained by integrating the reciprocal of the square of the DC component (L_DC, a_DC, b_DC) to P_L interference, P_a interference, and P_b interference is appropriately γ-corrected, and the interference level (L interference level, a interference level, b) Interference level). Here, γ may be determined so that the evaluation value calculated by the reproducibility evaluation unit 208 can have a good correspondence with the subjectivity.
L Interference = {P_L Interference / (L_DC ^ 2)} ^ γL Interference a Interference == P_a Interference / (a_DC ^ 2)} ^ γa Interference b Interference = {P_b Interference / (b_DC ^ 2)} ^ γb Interference (11)
Here, the reason why the DC component is squared is to match the unit with the power of the molecule (the square of the amplitude). The reason for normalization by the square of DC is to evaluate the shape of the spectrum by aligning the DC components of the spectrum. It can be said that the smaller the value of the interference obtained as described above (the minimum is zero), the better the reproducibility of the amplitude spectrum in the interference region. When the degree of interference is getting worse, there are many breaks and jagged lines on the thin lines.

Fidelity Calculation Method FIG. 5 is a flowchart showing the fidelity calculation processing procedure by the fidelity calculation unit 206 in S308. As described above, in the case of an ideal line image, an amplitude spectrum appears only in the faithful region. Therefore, how much the amplitude spectrum in the faithful region of the line image to be evaluated differs from the line image to be referenced is evaluated. Hereinafter, the operation of the fidelity calculation unit 206 will be described with reference to the flowchart of FIG.

First, in S501, the spectra (F_L faithfulness, F_a faithfulness, F_b faithfulness) of L, a, and b input from the spectral region dividing unit 204 are weighted with visual characteristic functions H_L, H_a, H_b.
F_L faithful w (u, v) = F_L faithful (u, v) H_L (u, v)
F_a faithful w (u, v) = F_a faithful (u, v) H_a (u, v)
F_b faithful w (u, v) = F_b faithful (u, v) H_b (u, v) (12)
The visual characteristic functions H_L, H_a, and H_b may be determined from the VTF and the result of the experiment as well as the disturbance level. This visual characteristic function is not necessarily the same function as the function used for the disturbance degree.

  Next, in S502, the circumferential maximum power (P_L_max, P_a_max, P_b_max) of the weighted faithful region is calculated, and weighted with the inverse of the square of the DC component (L_DC, a_DC, b_DC).

  Here, the reason for dividing by the square of DC is the same reason as the degree of disturbance, and is omitted.

  Next, in S503, the amplitude spectrum (reference F_L faithfulness, reference F_a faithfulness, reference F_b faithfulness) and direct current components (reference L_DC, reference a_DC, reference b_DC) of the reference fine line are acquired from the reference fine line table unit 108.

Next, in S504, the spectrum of the fidelity region of the reference thin line is weighted with the visual characteristic function in the same manner as Expression (12) in S501.
Reference F_L faithful w (u, v) = reference F_L faithful (u, v) × H_L (u, v)
Reference F_a faithful w (u, v) = reference F_a faithful (u, v) × H_a (u, v)
Reference F_b faithful w (u, v) = reference F_b faithful (u, v) × H_b (u, v) (14)

  Next, in S505, as in S502, the maximum value in the circumferential direction of the power of the fidelity region of the reference thin line is calculated and weighted by the reciprocal of the square of the DC component (P_L_faithful 0, P_a_faithful 0, P_b_faithful 0).

In step S506, the absolute value of the difference between the weighted circumferential maximum values of the line image and the reference thin line is integrated.
Dif_L faithfulness = Σ | P_L faithfulness (f) −reference P_L faithfulness (f) |
Dif_a faithful = Σ | P_a faithful (f) −reference P_a faithful (f) |
Dif_b faithful = Σ | P_b faithful (f) −reference P_b faithful (f) | (16)
Here, Σ represents that the sum is obtained for all f.

Finally, in S507, the integral value is multiplied by an appropriate logarithm (γL faithfulness, γa faithfulness, γb faithfulness) and the result of γ correction is output as the fidelity.
L fidelity = {Dif_L faithful} ^ γL fidelity a fidelity = {Dif_a faithful} ^ γa faithful b fidelity = {Dif_b faithful} ^ γb faithful (17)
Here, γL faithfulness, γa faithfulness, and γb faithfulness may be determined so that the evaluation value calculated by the reproducibility evaluation unit 208 can correspond well with the subjectivity. This value does not necessarily have to be the same as the logarithm of interference.

DC Difference Calculation Method FIG. 6 is a flowchart showing a DC difference calculation processing procedure by the DC difference calculation unit 207 in step S309. The direct current component DC represents the darkness of the line, and the density difference is evaluated by the DC difference between the line image to be evaluated and the line image to be referenced. Hereinafter, the operation of the DC difference calculation unit 207 will be described with reference to the flowchart of FIG.

  First, in S601, the direct current components (L_DC0, a_DC0, b_DC0) of the reference thin line are acquired from the reference thin line table unit 111.

Next, in S602, the relative difference between the direct current component (L_DC, a_DC, b_DC) of the line image and the direct current component of the reference thin line (refer to L_DC, a_DC reference, b_DC) is calculated, and a value obtained by appropriately performing γ correction. (ΓLDC difference, γaDC difference, γbDC difference) is defined as a DC difference. The flow chart follows Weber's law,
LDC difference = {| L_DC−reference L_DC | / | L_DC |} ^ γLDC difference aDC difference = {| a_DC−reference a_DC | / | a_DC |} ^ γaDC difference bDC difference = {| b_DC−reference b_DC | / | b_DC | } ^ ΓbDC difference (18)
In this case, the denominator may become zero,
LDC difference = {| L_DC−reference L_DC | / | L_DC + α_L |} ^ γ LDC difference aDC difference = {| a_DC−reference a_DC | / | a_DC + α_a |} ^ γaDC difference bDC difference = {| b_DC−reference b_DC | / | b_DC + α_b | } ^ ΓbDC difference (19)
(Α_L, α_a, α_b may be appropriate real numbers). The logarithm (γLDC difference, γaDC difference, γbDC difference) here may be determined so that the evaluation value calculated by the reproducibility evaluation unit 208 can be in good correspondence with the subjectivity as well as the interference degree and fidelity. It can be said that the smaller the DC difference, the higher the reproducibility.

Thin Line Correction Unit The line image correction unit 109 will be described below with reference to the block diagram of FIG. 11 and the flowchart of FIG. In this embodiment, a line image obtained by halftoning the original data by screen processing is set as an evaluation / correction target.

  First, in step S901, the line image correction unit 109 calculates the ten evaluation values calculated by the thin line reproducibility evaluation unit 107 (L disturbance level, a disturbance level, b disturbance level, L fidelity, a fidelity, b faithfulness). Degree, LDC difference, aDC difference, bDC difference, reproducibility evaluation value LN).

  Next, in step S <b> 902, the reproducibility determination unit 1102 determines whether or not the reproducibility evaluation value LN is greater than or equal to the reproducibility evaluation value threshold stored in the threshold storage unit 1101. Note that the threshold storage unit 1101 includes 10 pieces of L obstruction, a obstruction, b obstruction, L fidelity, a fidelity, b fidelity, LDC difference, aDC difference, bDC difference, and reproducibility evaluation value LN. A threshold value corresponding to the evaluation value is stored. These threshold values may be set, for example, as the threshold values of the respective evaluation values when it is difficult to tolerate line changes. These threshold values may be set in advance or may be adjusted later by the user. Here, if the reproducibility evaluation value LN is equal to or smaller than the threshold value, the process is terminated as no correction is necessary.

  Next, in step S903, the interference level determination unit 1103 determines whether the interference level stored in the threshold storage unit 1101 exceeds the threshold value. If the line to be evaluated has a discontinuity or jaggedness that does not exist in the referenced line, the value of the disturbance level increases. Therefore, by this determination, it is determined whether to correct the jaggedness or discontinuity of the line.

  First, the total disturbing degree is calculated by adding up the L disturbing degree, the a disturbing degree, and the b disturbing degree. If the total disturbance level does not exceed the threshold value stored in the threshold value storage unit 1101, the process proceeds to the next process without correcting the jaggedness and discontinuity.

  If the interference level is equal to or greater than the threshold value, the line jaggedness and discontinuity are corrected in step S904. When jaggedness or discontinuity occurs in a line, the screen used for screen processing often interferes with the line image. Therefore, when the original data can be corrected, the screen used for the screen processing is changed. The screen to be changed here is preferably a line screen orthogonal to the direction of the line.

  On the other hand, when the original data cannot be corrected, correction processing such as forcing dots is performed in order to correct jaggedness and discontinuity. Note that whether or not the original data can be corrected is determined by, for example, whether or not the user is allowed to change the image according to user settings, whether or not the original image that can change the screen is a multi-valued image, and the like. Good.

  In step S <b> 905, the fidelity determination unit 1105 determines whether the fidelity exceeds the fidelity threshold stored in the threshold storage unit 1101. Similar to the interference determination unit 1103, the total fidelity is calculated by adding the L fidelity, the a fidelity, and the b fidelity, and is compared with a threshold value. There are three factors that increase fidelity. Color shift, line width, and blur. From here, one of the three factors that most deteriorates fidelity is corrected.

  First, in step S906, the DC difference determination unit 1107 determines whether the aDC difference and the bDC difference exceed a threshold value. When the aDC difference and the bDC difference are large, the fidelity is often deteriorated due to the occurrence of color misregistration. Therefore, if the aDC difference and the bDC difference are equal to or larger than the threshold values, the fidelity correction unit 1106 performs color misregistration correction processing on the image output apparatus 110 in S908, and the processing ends. In the color misregistration correction process, for example, a line may be replaced with a single color, or correction may be performed by a registration function of an image output device.

  If it is less than the threshold value in S906, the DC difference determination unit 1107 determines in S907 whether the L_DC difference is greater than or equal to the threshold value. When the fidelity and the L_DC difference are large, the line width reproducibility is often poor. In step S <b> 913, the fidelity correction unit 1106 performs line width correction processing on the image output apparatus 110 and ends the processing. For the line width correction, for example, if the original data can be corrected, the line width of the original data may be adjusted.

  Furthermore, when the DC difference determination unit 1107 determines that any DC difference is equal to or less than the threshold value in S906 and S907, the reproducibility often deteriorates due to the occurrence of blur. Therefore, the fidelity correction unit 1106 performs blur correction processing on the image output apparatus 110. In the blur correction process, for example, an unsharp mask may be applied to the input image, or the laser output may be changed for electrophotography.

  If the fidelity is equal to or less than the threshold value in S905, the DC difference determination unit 1007 determines whether or not the DC difference exceeds the DC difference threshold value stored in the threshold value storage unit 1101 in S910. If the DC difference is less than or equal to the threshold value, no correction process is performed. On the other hand, if the DC difference is greater than or equal to the threshold, the DC difference correction unit 1008 performs density correction processing on the image output apparatus 110. For example, if the image is thin, the density correction process can increase the amount of toner delivered in the case of electrophotography, increase the laser output, or increase the density or line width of the original data if the original data can be corrected. Just do it. On the other hand, if it is dark, it is sufficient to reduce the amount of toner delivered in the case of electrophotography, reduce the output of the laser, or reduce the density and line width of the original data if the original data can be corrected. This completes all the correction processing.

  In this embodiment, correction is performed with priority in the order of interference level → fidelity → DC difference, but correction may be performed with priority given to the item having the largest value (deterioration degree) among these values.

  It is also useful to appropriately change the correction procedure according to the printer. Specifically, by changing the threshold depending on the ease of correction, correction processing according to the performance of the printer can be appropriately performed. If a low threshold is given if correction is easy, and a high threshold is given if correction is difficult, correction processing is performed with priority given to factors that the printer can easily correct.

  Moreover, although the example which performs all the correction processes was shown in this flowchart, you may perform the process which corrects a part among six factors. Not all correction processes can be performed, such as when there is a time limit. Therefore, if the evaluation value in this embodiment is used, the factor to be corrected can be appropriately selected from the six factors that deteriorate the reproducibility of the thin line.

  As described above, the six factors that hinder the reproducibility of the fine lines have been evaluated independently, but according to the present embodiment, the fine lines can be comprehensively evaluated.

Correspondence with Evaluation Items FIG. 7 shows the correspondence between the evaluation values described above and the evaluation items “line width”, “jaggedness”, and “line density” mentioned in the prior art. FIG. 7 also shows correspondence between evaluation values and “blur”, “discontinuity”, and “color shift”. The six evaluation items “line width”, “jaggedness”, “line density”, “blur”, “discontinuity”, and “color shift” can be evaluated based on the evaluation values described so far. “Line width”, “blur”, and “color shift” in which the difference from the reference thin line is difficult to understand from the two-dimensional amplitude spectrum of FIG. 7 are also shown in a cross-sectional view in FIG.

  Reference numeral 701 denotes a spectrum of the referenced thin line in the frequency space. The power of the disturbing area is zero, and power exists in the faithful area (u axis). The shape of the amplitude is a sinc function like a reference amplitude spectrum 801 in FIG.

  Reference numeral 702 denotes a spectrum in the frequency space when only the line width is changed. Like the reference thin line, the power of the interference area is zero, but the frequency band in which the power of the faithful area appears changes. This is because the width of the sinc function changes when the line width changes as in 802. Therefore, if the line width difference changes, the interference level is zero, but the fidelity value increases.

  Reference numeral 703 denotes a spectrum in the frequency space when the line becomes jagged. Random power is generated in the disturbing area, and the high-frequency component of the power in the faithful area is attenuated as compared with the thin line of the reference image. As a result, both interference and fidelity increase.

  Reference numeral 704 denotes a spectrum in the frequency space when the density changes. The interference area power is zero. In the faithful region, the power form (frequency band) is not different from the reference thin line, but the power magnitude (DC component) is different. Thus, the interference and fidelity are zero, but the DC difference is large.

  Reference numeral 705 denotes a spectrum in the frequency space when the line is blurred. Although the power in the disturbing area is zero, the high-frequency component is attenuated in the power in the faithful area compared to the reference thin line. Comparing the amplitude spectrum 803 when the line is blurred and the reference amplitude spectrum 801, it can be seen that the amplitude of the high frequency component is attenuated. Therefore, the obstruction degree is zero, but the fidelity is increased.

  Reference numeral 706 denotes a spectrum in the frequency space when color misregistration occurs. The interference area power is zero, but the faithful area power appears in a frequency band different from that of the reference thin line. Comparing the amplitude spectrum 804 and the reference amplitude spectrum 801 when color misregistration occurs, it can be seen that the shape of the function is greatly changed. Therefore, the obstruction degree is zero, but the fidelity is increased.

  Reference numeral 707 denotes a spectrum in the frequency space when the line is interrupted. The power of the faithful area is the same as that of the reference thin line, but power is generated in the disturbing area due to the change in the vertical shape due to the break. Therefore, the degree of disturbance increases.

  As described above, it has been shown that the six factors that reduce the fine line reproducibility can be comprehensively evaluated by the interference, fidelity, and DC difference.

  As described above, according to the present embodiment, the line image acquired from an input device such as a scanner is comprehensively evaluated for the deterioration of fine line reproducibility caused by a plurality of factors, and appropriately corrected for the cause of the deterioration. Can be performed.

  In addition, as compared with the conventional case where six factors are evaluated and corrected independently, correction of the factors of deterioration in reproducibility can be performed in the order of higher priority and in the order of decreasing line reproducibility.

  Furthermore, when the correction function is limited or when there is a time constraint, an optimal combination of corrections can be selected by using the evaluation method of this embodiment.

(Other examples)
In the thin line correction unit according to the present embodiment, an example in which the evaluation is performed using the total disturbance degree and the total fidelity is shown, but the present invention is not limited thereto. The L interference degree, a interference degree, b interference degree, L fidelity, a fidelity, and b fidelity may all be individually compared with a threshold value. Moreover, you may use one part instead of using all.

  Further, the spectrum calculation method in the frequency space shown in the present embodiment need not be exactly the same.

  The present invention can also be realized by supplying a storage medium storing software program codes for realizing the functions of the above-described embodiments (for example, the functions shown in the above flowchart) to a system or apparatus. In this case, the functions of the above-described embodiments are realized by the computer (or CPU or MPU) of the system or apparatus reading and executing the program code stored in the storage medium so that the computer can read the program code.

Claims (9)

Spectrum calculating means for calculating a spectrum in the frequency space of the input image;
Reference information acquisition means for acquiring a reference spectrum in a frequency space corresponding to the input image;
An evaluation means for evaluating the input image by comparing the spectrum of the input image with the reference spectrum;
Correction means for correcting the input image according to the result of the evaluation means,
The evaluation unit divides a region where amplitude appears in the reference spectrum and a region where amplitude does not appear, and compares the spectrum of the input image with the reference spectrum.   The image correction apparatus according to claim 1, wherein the evaluation unit compares the spectrum of the input image with the reference spectrum in a region where the amplitude does not appear.   The image correction apparatus according to claim 1, wherein the evaluation unit compares the spectrum of the input image with the reference spectrum in a region where the amplitude appears. Furthermore, DC component calculation means for calculating a DC component of the input image,
4. The image correction apparatus according to claim 1, further comprising a normalizing unit that normalizes a spectrum of the input image based on the DC component. 5.   5. The image correction according to claim 1, wherein the evaluation unit calculates an evaluation value based on a result obtained by multiplying the spectrum of the input image by a function corresponding to a visual characteristic. apparatus. The evaluation unit is a fidelity calculation unit that calculates a fidelity based on a difference between the spectrum of the image and the reference spectrum in a region where the amplitude of the reference spectrum appears.
6. Interference degree calculating means for calculating an interference degree based on a difference between the spectrum of the input image and the reference spectrum in a region where the amplitude of the reference spectrum does not appear. The image correction apparatus according to item. The evaluation unit is a fidelity calculation unit that calculates a fidelity based on a difference between the spectrum of the image and the reference spectrum in a region where the amplitude of the reference spectrum appears.
Interference level calculation means for calculating a disturbance level based on a difference between the spectrum of the input image and the reference spectrum in a region where the amplitude of the reference spectrum does not appear,
The image correction apparatus according to claim 4 , wherein the correction unit corrects the input image based on the DC component, the fidelity, and the interference. A spectrum calculation step of calculating a spectrum in the frequency space of the input image;
A reference information acquisition step of acquiring a reference spectrum in a frequency space corresponding to the input image;
An evaluation step of evaluating the input image by comparing the spectrum of the input image with the reference spectrum;
A correction step of correcting the input image according to the result of the evaluation step,
In the image correction method, the evaluation step divides a region where amplitude appears in the reference spectrum and a region where amplitude does not appear, and compares the spectrum of the input image with the reference spectrum.   A computer program for causing a computer to function as each unit of the image correction apparatus according to any one of claims 1 to 7.