JPH10283470A - Image processor, image processing method and recording medium - Google Patents

Abstract

PROBLEM TO BE SOLVED: To perform appropriate gradation correction in accordance with the characteristic of image data by scanning input image data that is expressed in a prescribed gradation number, extracting a representative value that represents the gradation distribution and correcting the gradation distribution of image data in accordance with the extracted representative value. SOLUTION: A gradation number converting part 22 converts the gradation number of image data which is inputted from an image input device 21 into a prescribed gradation number and supplies it to a histogram measuring part 23. The part 23 counts the frequency number in each gradation of an image that is converted in the part 22 and creates a histogram. A representative value setting part 24 sets a representative value from the created histogram, and a correction coefficient setting part 25 sets a correction coefficient based on the representative value. An LUT setting part 26 sets an LUT(look-up table) for gradation correction based on the correction coefficient. A gradation correcting part 27 refers to the LUT which is set by the part 26 and performs gradation correction of an input data.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

[0001] 1. Field of the Invention [0002] The present invention relates to an image processing apparatus, an image processing method, and a recording medium for performing gradation correction according to the characteristics of an image.

[0002]

2. Description of the Related Art In recent years, digital copying machines have been required to have not only faithful reproduction but also higher quality reproduction. For example, when copying a document where the white part of the background is dirty, it is preferable to reproduce it in a clean state rather than copying the dirt as it is, and if the pencil character is faint, it is clearly reproduced sharply Is preferred. Usually, by using the feature amount of the extracted image and external information such as brightness of the image, if there is such information,
The image quality correction is realized automatically or manually.

[0003] However, with the progress of networking and systematization, a situation has arisen in which it is necessary to perform correction processing on an image whose environment is unknown under what environment. In this case, external information such as photographing conditions cannot be reflected in the correction processing. Therefore, a correction process must be performed using information from the image itself as a main feature amount.

As a correction method in such a case, several gradation correction methods using a histogram as a feature have been proposed. Among them, a relatively simple and widely known technique is dynamic range conversion. This method is a linear transformation that linearly extends the gradation distribution indicated by the histogram to a wider density region. FIG. 1 is a conceptual diagram illustrating this dynamic range conversion. FIG.
(A) is a histogram H (i) showing the distribution of the input image, where the minimum value of the gradation is x1 and the maximum value is x2. When the histogram H (i) is extended in the range from gradation 0 to gradation M, L having the characteristic shown in FIG.
Tone conversion is performed using a UT (look-up table). Then, as shown in FIG. 1C, the histogram H (i) of the input image becomes the histogram G indicated by the broken line.
(I). By this gradation conversion, the gradation distribution concentrated in a certain specific range can be extended over a wide range, and the contrast can be enhanced.

[0005] Here, the dynamic range conversion is a linear conversion. When the gradation before correction is x, the gradation after correction is y, and the correction coefficients are a and b, the LUT is generally represented by the following equation.

(Equation 1)

[0006]

However, the dynamic range conversion has the following problems. For example, assume that when correcting an image in which the gradation distribution is concentrated in a very narrow range, the distribution of the original image is extended to the maximum range in which the gradation can be obtained. In such a case, the contrast becomes too strong, the roughness becomes conspicuous, and a false contour or the like occurs.

[0007] Further, it is assumed that when the brightness of an extremely bright image is corrected, that is, when the brightness distribution of the original image is concentrated on the low gradation side, the image is stretched at the same ratio over all gradations. In such a case, although the contrast of the halftone portion is improved by stretching the distribution,
Since the low gradation side is also stretched at the same ratio, even a sufficiently bright portion that does not require correction may be further brightened.

The present invention has been made in view of such a problem, and an object of the present invention is to provide an image processing apparatus, an image processing method, and a recording medium capable of performing appropriate gradation correction according to the characteristics of image data. And

[0009]

In order to solve the above-mentioned problems, the present invention scans input image data represented by a predetermined number of gradations and extracts a representative value representative of the gradation distribution. And a correction unit for correcting the gradation distribution of the image data according to the representative value extracted by the extraction unit.

[0010]

Embodiments of the present invention will be described below with reference to the drawings.

1. First Embodiment First, a first embodiment of the present invention will be described. 1-1 Configuration of First Embodiment FIG. 2 is a block diagram illustrating a configuration of the first embodiment. In FIG. 2, an image input device 21 is, for example, a digital copying machine or a digital multi-value image input device having a scanner or an image reading function, and supplies image data to a gradation number conversion unit 22.

The tone number conversion unit 22 converts the number of tones of the input image data into a predetermined number of tones and supplies it to the histogram measuring unit 23. The histogram measuring unit 23 generates a histogram by counting the frequency of each tone of the image converted by the tone number converting unit 22. The representative value setting unit 24
A representative value described later is set from the created histogram. The memory 20 is connected to the histogram measuring unit 23 and the representative value setting unit 24 via a bidirectional bus, and stores the histogram.

The correction coefficient setting section 25 sets a correction coefficient based on the representative value. This correction coefficient will be described later. The LUT setting unit 26 sets an LUT (lookup table) for gradation correction based on the correction coefficient. The gradation correction unit 27 performs gradation correction of the input image with reference to the LUT set in the LUT setting unit 26. The image output device 28 is a device that outputs corrected image data, and is configured by, for example, an image output device such as a printer or a digital copying machine having a printout function.

1-2. Operation of First Embodiment Next, an operation of the image processing apparatus having the above configuration will be described with reference to FIGS. FIG. 3 is a flowchart for explaining the operation of the first embodiment. First, an image is input in the image input device 21 (S
301), and subsequent operations are roughly classified into pre-scanning and main scanning. In brief, in the pre-scanning, the feature amount of the image itself is calculated, and a correction coefficient is set using the feature amount (path indicated by a broken arrow in FIG. 2). Then, in the main scanning, gradation correction is performed using the correction coefficient set in the pre-scanning (path indicated by a solid arrow in FIG. 2).

(1) Pre-scanning First, pre-scanning will be described. The gradation number conversion unit 22
The sampling gradation number of the image data supplied from the image input device 21 is operated (S302). The conversion of the number of gradations is mainly for the purpose of removing noise.
That is, when the supplied image data has high image quality, it is not necessary to change the number of gradations. However, when the image data is low in image quality, it is necessary to remove noise by coarsely sampling and reducing the number of gradations. Then, the histogram measuring unit 23
Creates a histogram of the data supplied from the tone number conversion unit 22 and stores it in the memory 20 (S303).

The representative value setting unit 24 represents a representative value Iall of the entire distribution, a minimum representative value Imin representing a value near the minimum value, and a value near the maximum value from the histogram created by the histogram measuring unit 23. The maximum representative value Imax to be performed is calculated (S304). For example, as the representative value Iall, a gradation value indicating the peak value of the frequency is calculated, and as the minimum-side representative value Imin, the gradation value at which the cumulative frequency number becomes 5% or more at first is calculated. Representative value Im
As ax, a gradation value at which the cumulative frequency number first becomes 95% or more is calculated.

The correction coefficient setting section 25 includes a representative value setting section 24.
By using the representative value set in step (a), it is determined how much the distribution is to be expanded (range after gradation distribution correction), and correction coefficients a and b are set (S305). That is, the range after the gradation distribution correction is automatically set according to the image, and linear conversion is performed. Here, the correction coefficients are a and b in Equation 1.

First, the range after the gradation distribution correction will be described. Here, the minimum representative value after the distribution is stretched is defined as Omin, and the maximum representative value is defined as Omax. The reference value set in advance is THD. This reference value T
HD indicates a so-called standard gradation, and is a value experimentally set depending on the type of a target document such as a photograph or a text document, and the type and accuracy of an input / output device. The minimum representative value Omin and the maximum representative value Omax are first compared with a preset reference value THD and a representative value Iall,
When the representative value Iall <THD is calculated according to the following equation 2, when the representative value Iall ≧ THD is calculated according to the following equation 3.

(Equation 2)

[0019]

(Equation 3)

Here, F (x) is a function of the representative value Iall of the entire distribution, and the reference value T is set in advance.
Equation 4 based on the difference between HD and the representative value Iall of the entire distribution
Given by

(Equation 4)

Here, α and β are values that are experimentally set depending on the type of the target document such as a photograph or a text document and the type and accuracy of the input / output device.

That is, when the representative value Iall of the entire distribution is lower than the reference value THD, it is determined that the distribution is concentrated on the low gradation side, and the representative value Iall is fixed while the minimum representative value Imin is fixed. Is set so as to approach the reference value THD (that is, to extend only to the high gradation side) (Equation 2). Conversely, if the representative value Iall of the entire distribution is higher than the predetermined reference value THD, it is determined that the distribution is concentrated on the high gradation value, and the representative value Iall is fixed while the maximum representative value Imax is fixed. The setting is made so as to approach the reference value THD (that is, to extend only to the low gradation side) (Equation 3).

Next, the obtained representative values Imin, Ima
Using x, Omin, and Omax, correction coefficients a and b
Is calculated according to Equation 5.

(Equation 5)

The LUT setting section 26 includes a correction coefficient setting section 25
An LUT is created using the correction coefficients a and b set in (S306). The above is the operation in the prescan.

(2) Main Scan Next, the main scan will be described. The tone correction unit 27 calculates L
The gradation distribution of the input image supplied from the image input device 21 is corrected with reference to the LUT set by the UT setting unit 26 (S307). As a result, when the distribution is concentrated on the low gradation side, the correction is performed so that the minimum representative value is fixed and expanded to the high gradation side, and when the distribution is concentrated on the high gradation value, the maximum side representative value is corrected. The correction is performed so that the representative value is fixed and is extended to the low gradation side. Thus, the image output device 2
8 outputs the image corrected by the gradation correction unit 27 (S
308).

2. Second Embodiment Next, a second embodiment will be described. In the present embodiment, a case where the brightness of an input color image is corrected will be described as an example.

2-1. Configuration of Second Embodiment FIG. 4 is a block diagram illustrating a configuration of the second embodiment.
As shown in FIG. 4, the image processing apparatus according to the second embodiment includes a first color space conversion unit 402, a gradation number conversion unit 403, a representative value setting unit 404, a correction coefficient setting unit 405, a gradation correction unit 406, And a second color space conversion unit 407. Note that the input image 401 is an image stored in the memory in advance, and the output image 408 is an image stored in the memory as a corrected image.

2-2. Operation of Second Embodiment First, the input image 401 is an RGB image, and the first color space conversion unit 402 performs color space conversion of the input RGB image into an L ^* a ^* b ^* image. In the present embodiment, after the color space conversion, only the brightness signal L ^* is subjected to the gradation correction processing. That is, L ^*
Are subjected to gradation correction in the gradation correction unit 406 and supplied to the second color space conversion unit 407 as L ^* ′. a ^* and b ^* are left uncorrected as they are in the second color space
7 is supplied.

Hereinafter, the correction processing will be described in detail.
First, the L ^* signal is supplied to the gradation number conversion unit 403, and the gradation number conversion unit 403 calculates the gradation number of the image data in the same manner as in the first embodiment (the gradation number conversion unit 22 shown in FIG. 2). Convert.

The representative value setting section 404 includes a gradation number conversion section 40.
3, the representative value Ial of the entire distribution
l, minimum representative value Imin, and maximum representative value Ima
Find three representative values of x. Here, a gradation average value is calculated as the representative value Iall of the entire distribution. Either arithmetic mean or geometric mean may be used as the gradation average value.
Further, the minimum value of the gradation is set as the minimum representative value Imin, and the maximum value of the gradation is set as the maximum representative value Imax. Therefore,
In the present embodiment, a correction coefficient can be set without creating a histogram as in the first embodiment.

Next, the correction coefficient setting unit 405 uses the representative value set by the representative value setting unit 404 in the same manner as in the first embodiment (the correction coefficient setting unit 25 shown in FIG. 2). And b are set.

The tone correction section 406 is based on the correction coefficients a and b set as described above,
The correction processing is performed on the L ^* signal supplied from O.02 according to Equation 6, and the corrected L ^* ′ signal is output.

(Equation 6)

In equation (6), the L ^* signal is x and the L ^* 'signal is y.

Thus, the second color space conversion unit 407 outputs the corrected image L ^* ′ signal supplied from the gradation correction unit 406,
The a ^* and b ^* signals supplied from the first color space conversion unit 402 are color space converted into RGB images, and the image signals R′G′B ′
Output as This image signal R'G'B 'is stored in the memory as an output image 408.

3. Third Embodiment Next, a third embodiment will be described. In the present embodiment, a case where the saturation of an input color image is corrected will be described as an example.

3-1. Configuration of Third Embodiment FIG. 5 is a block diagram illustrating a configuration of the third embodiment.
As shown in FIG. 5, the image processing apparatus according to the third embodiment includes a memory 500, a first color space conversion unit 502, a first gradation number conversion unit 503, a saturation creation unit 504, and a second gradation number conversion unit. 505, a representative value setting unit 506, a correction coefficient setting unit 507,
It comprises a tone correction unit 508 and a second color space conversion unit 509. The memory 500 is connected to the saturation creation unit 504 via a bidirectional bus, and stores information necessary for calculating saturation from the a ^* signal and the b ^* signal. Note that the input image 501 is an image stored in the memory in advance, and the output image 510 is an image stored in the memory as a corrected image.

3-2. Operation of Third Embodiment The input image 501 is an RGB image, and the first color space conversion unit 502 performs color space conversion of the input RGB image to L ^* a ^* b ^* . In the present embodiment, after color space conversion, tone correction processing is performed only on a ^* and b ^* . That is, a ^* and b ^* are subjected to gradation distribution correction in the gradation correction unit 508, and are referred to as a ^* ″ and b ^* ″ as the second color space conversion unit 5.
09. L ^* is supplied to the second color space conversion unit 509 without correction.

Hereinafter, the correction processing will be described in detail.
First, a ^* and b ^* are supplied to the first gradation number conversion unit 503. The first tone number conversion unit 503 converts the gradation number of a ^* and b ^*, and supplies the converted saturation creating unit 504 a ^* 'and b ^*'. The saturation creation unit 504 calculates saturation from a ^* 'and b ^* '. The saturation may be created using a saturation table preset in the memory 500 or by a general saturation calculation formula.

Next, the saturation calculated by the saturation creation section 504 is further subjected to gradation number conversion in a second gradation number conversion section 505 and supplied to a representative value setting section 506. The representative value setting unit 506 calculates a representative value Iall, a minimum-side representative value Imin, and a maximum-side representative value Imax of the entire distribution.
The setting of each representative value is the same as in the second embodiment (representative value setting unit 404 shown in FIG. 4).

The correction coefficient setting unit 507 calculates the representative value Ial
Using l, Imin, and Imax, it is determined how much the distribution is stretched, and correction coefficients a and b for linear conversion are set.

The minimum representative value Omin and the maximum representative value Omax are first compared with a preset reference value THD and a representative value Iall, and when the representative value Iall <THD is calculated according to the following equation (7). Representative value Iall ≧ T
In the case of HD, it is calculated according to Equation 8. Where F
(X) is the same as Expression 4 in the second embodiment.

(Equation 7)

[0042]

(Equation 8)

That is, when the representative value Iall of the entire distribution is lower than the predetermined reference value THD, it is determined that the distribution is concentrated on the low gradation value, and the higher representative value Imin is fixed and the lower representative value Imin is fixed. Only the key side is extended (Equation 7).

On the other hand, if the representative value Iall of the entire distribution is higher than the predetermined reference value THD, it is determined that the distribution is concentrated on high gradation values, and in this case, the distribution is not changed (Equation 8). That is, it is generally preferable to increase the saturation rather than decrease the value. Therefore, gradation correction for decreasing the saturation is not performed, and the same saturation as that of the original image is maintained. After the correction coefficient is set, gradation correction of a ^* and b ^* is performed. The tone correction unit 508 calculates a ^* ″ and b ^* ″ after tone correction from the correction coefficients a and b, as in the second embodiment (the tone correction unit 406 shown in FIG. 4). It is supplied to the second color space conversion unit 509.

Thus, the second color space conversion unit 509 outputs the a ^* ″ signal and b ^* ″ signal supplied from the gradation correction unit 508 and the L ^* signal supplied from the first color space conversion unit 502. Is converted into an RGB image in a color space and output as an image signal R′G′B ′. This signal R'G'B 'is stored in the memory as the image output image 510.

4. Fourth Embodiment Next, a fourth embodiment will be described. In the present embodiment, a case will be described as an example in which the brightness and the saturation of the input color image are simultaneously corrected.

4-1. Configuration of Fourth Embodiment FIG. 6 is a block diagram illustrating a configuration of the fourth embodiment.
As shown in FIG. 6, the image processing device according to the fourth embodiment includes an image input device 601 and a shading correction unit 60.
2, the first color space conversion unit 603, the gradation correction unit 604, the second
It comprises a color space conversion unit 605, a filter processing unit 606, a gamma correction unit 607, an image output device 608, and a correction coefficient matrix setting unit 609.

Here, the image input device 601 is, for example,
It is a digital multi-value image input device such as a scanner or a digital copying machine having an image reading function. The shading correction unit 602 corrects variations in sensitivity of each pixel of the sensor in the image input device 601 and uneven illumination. The filter processing unit 606 removes a noise component after image processing, and the gamma correction unit 6
Reference numeral 07 denotes fine adjustment of the density in accordance with the characteristics of the image output device 608.

4-2. Operation of Fourth Embodiment An RGB image input by the image input device 601 is corrected by the shading correction unit 602, and the first color space conversion unit 6
03 is converted to an L ^* a ^* b ^* image. The color space-converted L ^* a ^* b ^* is first supplied to a correction coefficient matrix setting unit 609 to set a correction coefficient. The correction coefficient matrix setting unit 609 calculates a matrix in which correction coefficients for linear conversion of each of lightness and chroma are integrated in order to simultaneously correct lightness and chroma.

Here, assuming that the correction coefficient matrices set in the correction coefficient matrix setting section 609 are M and m,
L ^* 'a ^* ' b ^* '(gradation-corrected L ^* a ^* b ^* ) is calculated according to the following equation (9).

(Equation 9)

However, the brightness correction coefficients are al and b
The calculation method is the same as in the second embodiment (Equation 6). The saturation correction coefficients are ac and bc, and the calculation method is the same as in the third embodiment (Equations 7 and 8).

When the correction coefficient matrix is set in this way, the gradation correction unit 604 uses the L ^* a ^* b ^* supplied from the first color space conversion unit 603 and the correction coefficient matrices M and m to perform gradation correction. L ^* 'a ^* ' b ^* 'of the second color space conversion unit 6
05.

The L ^* 'a ^* ' b ^* 'signal after the gradation correction is converted into the YMCK color space by the second color space conversion unit 605, and then passes through the filter processing unit 606 and the gamma correction unit 607. Are output from the image output device 608.

5. Modifications The present invention is not limited to the above-described embodiments, and various modifications as described below are possible. For example, in the first embodiment, the peak value is set as a representative value of the entire distribution, but may be set as an average value such as an arithmetic mean or a geometric mean or a median value. Further, as the minimum-side representative value Imin and the maximum-side representative value Imax, the minimum value and the maximum value may be employed instead of the values near the minimum value and the maximum value, respectively. In this case, similar to the second to fourth embodiments,
The correction coefficient can be set without creating a histogram.

On the other hand, in the second to fourth embodiments, similarly to the first embodiment, a histogram may be created and a correction coefficient may be set.

In each of the embodiments, image data recorded on a medium such as a memory or a hard disk is used as an input image, or a digital multi-valued image input device such as a scanner or a digital copying machine having an image reading function is used. It is possible to arbitrarily select whether to use the image data to input.

Also, it is arbitrarily selected whether the output image is stored in a medium such as a memory or a hard disk, or is output to an image output device such as a printer or a digital copying machine having a printout function. It is possible.

In the embodiment, gradation correction by linear conversion is performed. However, the present invention is not limited to this, and the present invention is also applicable to gradation correction by non-linear conversion.

In the embodiments, the present invention is embodied as an image processing apparatus. However, the technical idea of the present invention can be understood as an image processing method implemented in such an apparatus. Further, the present invention can be configured as a control program executed by a microcomputer that controls the image processing apparatus, and can be stored in a recording medium such as a floppy disk.

[0060]

As described above, according to the present invention,
Image data represented by a predetermined number of gradations can be corrected so as to have an appropriate gradation distribution according to its characteristics.

[Brief description of the drawings]

FIG. 1 is a conceptual diagram illustrating dynamic range conversion.

FIG. 2 is a block diagram showing a configuration of the first embodiment.

FIG. 3 is a flowchart for explaining the operation of the first embodiment.

FIG. 4 is a block diagram illustrating a configuration of a second embodiment.

FIG. 5 is a block diagram showing a configuration of a third embodiment.

FIG. 6 is a block diagram showing a configuration of a fourth embodiment.

[Explanation of symbols]

20 memory, 21 image input device, 22 gradation number conversion unit, 23 histogram measurement unit, 24 representative value setting unit, 25 correction coefficient setting unit, 26 LUT setting unit , 27... Gradation correction unit, 28.
01 input image, 402 first color space conversion unit, 40
3 ··· gradation number conversion unit, 404 ··· representative value setting unit, 405
..Correction coefficient setting section, 406..gradation correction section, 407.
A second color space conversion unit, 408, an output image, 500
Memory, 501 input image, 502 first color space conversion unit, 503 first gradation number conversion unit, 504 saturation creation unit, 505 second gradation number conversion unit, 506 · Representative value setting unit, 507 ··· correction coefficient setting unit, 508 ··· gradation correction unit, 509 ··· second color space conversion unit, 510 ··· output image, 601 ··· image input device, 602 ··· shading correction Unit, 603 first color space conversion unit, 604 gradation correction unit, 605 second color space conversion unit, 606 filter processing unit, 607 gamma correction unit, 608 image output device , 609... Correction coefficient matrix setting unit

Claims (9)

[Claims] 1. An extracting means for scanning input image data represented by a predetermined number of gradations and extracting a representative value representative of the gradation distribution; An image processing apparatus comprising: a correction unit configured to correct a gradation distribution of image data. 2. An extracting means for scanning input image data represented by a predetermined number of gradations to extract a representative value representative of the gradation distribution, and a predetermined value representative of the representative value extracted by the extracting means. A comparison unit that compares the representative value with the reference value, and a correction unit that corrects a gradation distribution of the image data so that the representative value approaches the reference value in accordance with a comparison result by the comparison unit. Image processing apparatus. 3. The method according to claim 1, wherein the extracting unit extracts a representative value representing the gradation distribution and a minimum-side representative value and a maximum-side representative value corresponding to the range of the gradation distribution. Comparing a representative value representative of the grayscale distribution with the predetermined reference value, and when the representative value representing the gradation distribution is smaller than the reference value, the correction unit stores the minimum-side representative value while storing the minimum-side representative value. While correcting the gradation distribution so that the representative value representing the gradation distribution is closer to the reference value, if the representative value representing the gradation distribution is larger than the reference value, the representative value is stored while storing the maximum-side representative value. 3. The tone distribution is corrected so that a representative value representative of the tone distribution approaches the reference value.
The image processing apparatus according to any one of the preceding claims. 4. The image processing apparatus according to claim 1, wherein said input image data is image data of a color image, and corrects a lightness gradation distribution. 5. The image processing apparatus according to claim 1, wherein said input image data is image data of a color image, and corrects a gradation distribution of saturation. 6. The image processing apparatus according to claim 1, wherein the input image data is image data of a color image, and corrects a lightness gradation distribution and a saturation gradation distribution. Image processing device. 7. The image processing apparatus according to claim 5, wherein the correction unit performs only correction to a high gradation side. 8. A first scanning step of extracting a representative value representing a gradation distribution of input image data represented by a predetermined number of gradations, and according to a representative value extracted in the first scanning step. And a second scanning step of correcting the gradation distribution of the image data. 9. A first scanning step for extracting a representative value representing a gradation distribution of input image data represented by a predetermined number of gradations, and according to the representative value extracted in the first scanning step. And a second scanning step of correcting a gradation distribution of the image data by a computer controlling an image processing apparatus.