JPH10187964A - Method and device for image filtering, and image outline emphasis processor - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a processor which can satisfactorily emphasize the image outlines. SOLUTION: The ID value is outputted at a mask ID extraction part 10 based on the main signal S of a marked pixel. The weighting coefficient of a smoothing filter is selected at a weighting coefficient set selection part 20 according to the ID value among the constitutions of plural smoothing filters which are stored in a storage part of a memory, etc. The selected weighting coefficient is outputted to a smoothing process part 30 where an unsharp signal U is produced based on the received weighting coefficient of every smoothing filter. Then an unsharp mask signal (S-U) is produced. The signal S is also inputted to a 1st correction part 50 where an intensity coefficient P is outputted and an outline emphasis signal P (S-U) is produced. At a 2nd correction part 60, a corrected outline emphasis signal Q is produced and synthesized with the signal S for production of an outline emphasized signal E.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

[0001] 1. Field of the Invention [0002] The present invention relates to an image filtering method, an image filtering apparatus, and an image edge enhancement processing apparatus in image processing of a digital image obtained from a scanner, a digital camera, or another device.

[0002]

2. Description of the Related Art Conventionally, image filters used for filtering an image include a smoothing filter and a sharpening filter. The smoothing filter is a filter for blurring the image by removing high-frequency components of the image, and the sharpening filter is a filter for enhancing the outline of the image by enhancing the high-frequency components of the image.

[0003] When the contour of an image is enhanced, there are a method using the above-mentioned smoothing filter and a method using the sharpening filter.

[0004] In a method using a sharpening filter, if an image is filtered using the filter, the obtained image is an edge-enhanced image.

On the other hand, in the method using a smoothing filter, a blurred image is first generated by filtering with a smoothing filter. Then, a difference between the blurred image and the original image is obtained to generate a difference image. By adding the difference image to the original image, an image whose contour is enhanced is obtained.

FIG. 16 is a block diagram showing a conventional image enhancement processing apparatus using a smoothing filter. The main signal S enters the smoothing filter 70, where the image is smoothed. The output of the smoothing filter is an unsharp signal U obtained by performing a weighted average of the main signal S of the target pixel and the density values of the peripheral pixels. Then, the difference between the main signal S and the unsharp signal U is obtained to obtain an unsharp mask signal (SU). Two types of correction have been performed on the obtained unsharp mask signal (SU).

The first correction is to change the degree of contour emphasis according to the density value of the main signal S of the target pixel by the first correction unit 50 in the figure. The output of the first correction unit 50 is an intensity coefficient P. And the unsharp mask signal (SU)
And the intensity coefficient P are multiplied to obtain an outline emphasis signal P (S−
U) is generated. For example, in a highlight portion where noise is conspicuous or a shadow portion where there is a lot of electrical noise, the value of the intensity coefficient P is reduced to reduce noise.

Next, the second correction is performed by the second correction unit 60. The intensity of the contour emphasizing signal is corrected according to the level of the contour emphasizing signal P (SU). For example, in a portion where the level of the contour emphasis signal P (SU) is small, this correction is performed for the purpose of further reducing the level and suppressing the graininess of the image.

Then, a contour-enhanced signal E is generated by adding the finally corrected contour-enhanced signal and the main signal S.

[0010]

Since the conventional image filter operates in the same manner for all pixels, the smoothing filter performs smoothing even on a portion of the image that is not to be blurred, such as an intermediate density range, and provides a sharp image. There is a disadvantage that the noise reduction filter also emphasizes a noise component which is often present in a high density region of the image.

There is also a demand for enhancing the outline of an image to enhance the gradation difference (contrast) between fine lines in a highlight portion of the image. In order to emphasize the gradation difference between the fine lines in the highlight part, it is necessary to increase the effective mask size in the highlight part. However, such a method using a sharpening filter or a method using a smoothing filter cannot meet such a demand.

That is, in the method of enhancing the outline of an image using a sharpening filter, the entire original image is processed with the same filter characteristics, and therefore, it is not possible to perform another processing only on a highlight portion. It is possible, and the method of enhancing the outline of an image using a smoothing filter cannot enhance the gradation difference of a thin line in a highlight portion.

Therefore, conventionally, there is a problem that, for example, "the cloth texture of white clothes is inconspicuous".

SUMMARY OF THE INVENTION The present invention has been made in view of the above problems, and has an image filtering method and an image filtering apparatus for changing a filter characteristic for a specific region of an image, and satisfactorily enhances the contour of a specific region of the image. It is an object of the present invention to provide an image edge enhancement processing device capable of performing image processing.

[0015]

In order to achieve the above object, according to the first aspect of the present invention, an original image is represented by m × n (m, n).
Is an integer of 2 or more), and performs predetermined processing on the original image by scanning with an image filter. In accordance with the density value of the pixel of interest located at the center of the image filter, The size is changed.

According to a second aspect of the present invention, a density value detecting means for detecting a density value of a target pixel to be processed and an image filter specifying means for specifying an image filter having an effective size corresponding to the density value are provided. The density value of the pixel of interest is changed using an image filter obtained by the image filter generation means.

According to a third aspect of the present invention, there is provided an apparatus for performing edge enhancement processing on an original image, wherein the density value detecting means detects a density value of a target pixel to be processed, and the density value detecting means comprises: Image filter specifying means for specifying an image filter having an effective size corresponding to the obtained density value; density value changing means for changing the density value of the pixel of interest using the image filter specified by the image filter specifying means; A contour emphasizing signal generating means for generating an outline emphasizing signal by subtracting the density value changed by the value changing means from the density value of the pixel of interest before the change, and a contour emphasizing signal obtained from the contour emphasizing signal generating means. And a contour emphasizing means for generating a contour emphasized signal of the pixel of interest.

[0018]

BEST MODE FOR CARRYING OUT THE INVENTION

<1. Overview of Image Filtering Process> m × n (m,
A process of smoothing or sharpening an image (enhancing an outline) by scanning a matrix image filter (n is an integer of 2 or more) in an image plane is referred to as image filtering.

FIG. 1 is an explanatory diagram showing an outline of image filtering. With respect to the image I shown in FIG. 1A, the X direction is set as the main scanning direction, and the Y direction is set as the sub-scanning direction. . At the time of filtering, the target pixel OP is positioned at the center of the image filter F as shown in FIG. In the example of FIG. 1B, 5 × 5
(M = n = 5) matrix image filters are used. Then, a value obtained by performing a weighted average of the density value of the target pixel OP and the density values of peripheral pixels based on the weighting coefficient assigned to each component of the image filter is set as the density value of the target pixel OP.

That is, it is _assumed that the image filter F has a matrix shape of (2m + 1) × (2n + 1) as shown in FIG. 2, and each component is assigned a weighting coefficient of k- _mn to kmn. . The image filter using the F coordinates (x, y) the filtered density value S _'xy obtained by performing filtering for the pixel of interest is located in the

[0021]

(Equation 1)

## EQU1 ## Here, S _xy is the density value of the pixel located at the coordinates (x, y) of the original image.

[0023] In the embodiment of the present invention, the weighting coefficient k _-mn to k _mn image filter F in accordance with the main signal is the concentration value before filtering of the pixel of interest
Is to make changes. The case where such an image is smoothed or sharpened (contour enhancement) will be described below.

<2. Smoothing Process> First, the smoothing process will be described. FIG. 3 is a diagram showing an image smoothing filtering device according to an embodiment of the present invention. The main signal S for the target pixel is input, and the mask ID extraction unit 10
And the main signal S input to the smoothing processing unit 30.

The mask ID extracting section 10 outputs an ID value according to the main signal S. The mask ID extracting unit 10 includes a storage unit such as a memory therein, and stores a table in which the relationship between the input and the output as shown in FIG. As shown in FIG. 4, when the main signal S is small and is on the shadow side of the image, the output ID value is “3”. When the main signal S is at the intermediate level, the output ID value is When it becomes "1" and the main signal S is a highlight portion of an image, the output ID value becomes "5". In addition, the ID value becomes “2” or “4” according to the level of the input main signal S. Then, the I output from the mask ID extraction unit 10
The D value is sent to the weighting coefficient set selection unit 20.

[0026] In the weighting coefficient set selection section 20, based on the ID value sent from the mask ID extraction unit 10, the determination of all the weighting coefficient k _-mn to k _mn used for image filter. The weighting coefficient set selector 20 includes:
A storage unit such as a memory is provided in the same manner as the mask ID extraction unit 10, and each of the storage units is provided therein as shown in FIG.
A smoothing filter in which a weighting coefficient corresponding to the D value is set is stored.

The smoothing filter shown in FIG. 5 shows an example where the filter size is 5 × 5. When the ID value is “1” (see FIG. 5A), the size of the corresponding smoothing filter is substantially a 3 × 3 image filter. That is, when the weighting coefficient k _ij (i and j are arbitrary integers) is “0”, the density value of the pixel corresponding to the position (i, j) does not affect the filtering at all. Then, the number of weighting coefficients whose weight value is “0” is changed corresponding to the ID value.
In other words, the effective size of the smoothing filter is changed according to the ID value. Therefore, the effective size of the smoothing filter is changed by the mask ID extracting unit 10 and the weighting coefficient set selecting unit 20 according to the density value of the main signal S of the target pixel.

The smoothing filter corresponding to each ID value shown in FIGS. 5A to 5E has a frequency component that attenuates as the ID value changes from “1” to “5” to a lower frequency. Extend. When the ID value is “1” (see FIG. 5A)
Is a smoothing filter with an effective size of 3 × 3,
In the case of the value “5” (see FIG. 5E), the target pixel and the peripheral pixels have the same weight in the 5 × 5 region, and the complete averaging is performed.

Accordingly, when the relationship between the main signal S and the ID value as shown in FIG.
The frequency component attenuated for the intermediate level of the image is only the high frequency, and the frequency component attenuated for the shadow side of the image is lower than the intermediate level range. Further, on the highlight side of the image, even the lowest frequency components are attenuated.

The smoothing processing unit 30 performs the operation shown in Expression 1 using the weighting coefficient of the smoothing filter obtained from the weighting coefficient set selection unit 20 to generate the unsharp signal U. The unsharp signal U is a blur signal obtained by performing a smoothing process on the main signal S.

As described above, when the smoothing processing shown in this embodiment is performed, according to the signal level of the main signal S,
It becomes possible to adjust the frequency component to be attenuated. For example, it is also possible to lower the frequency components that attenuate the highlight part of the image compared to other parts.

<3. Sharpening Process> Next, the sharpening process will be described. FIG. 6 is a diagram showing an image sharpening filtering device according to an embodiment of the present invention. In addition,
The sharpening process refers to a process of extracting a high-frequency component of an image as in the case of the outline emphasis process. Therefore, the signal obtained as a result of the sharpening process is a contour-enhanced signal E in which the contour of the image is enhanced.
Becomes The device configuration is almost the same as in the case of the smoothing filter. Then, the main signal S for the pixel of interest is input, and the main signal S input to the mask ID extraction unit 10 and the sharpening processing unit 40 is sent.

The mask ID extracting section 10 has the same contents as those described in the case of the smoothing process. The ID value corresponding to the main signal S can be output by referring to the table that stores the relationship between the input and the output stored in the internal storage unit.

Also, the weighting coefficient set selection section 20
The same operation as in the case of the smoothing process is performed, but the weighting coefficient of the filter stored in the storage unit is different. In the smoothing filter in the case of the smoothing process, the value of each weighting coefficient is a positive number. However, in the sharpening filter in performing the edge enhancement, the weighting coefficient corresponding to the pixel of interest is generally a positive number. , The weighting coefficients corresponding to the other peripheral pixels are negative numbers. FIG. 7 is a diagram illustrating an example of a sharpening filter used in the sharpening process (contour enhancement process). FIG. 7 (a)
As shown in (e), when the ID value is “1”, the effective size is the smallest, and as the ID value increases, the effective size of the filter increases. As for the weighting coefficient, only the weight value corresponding to the pixel of interest is a positive number, and the weight values around it are negative numbers. Since each filter performs a weighted average of the pixel of interest and its surrounding pixels in accordance with the respective weighting coefficients, the filter shown in FIG.
When the D value is “1”, only the highest frequency range of the high frequency component can be extracted. When the ID value shown in FIG. Can be extracted.

The sharpening processing unit 40 performs an operation shown in Expression 1 by using the weighting coefficients of the sharpening filters to generate an edge-enhanced signal E. The contour-enhanced signal E is a signal obtained by subjecting the main signal S to contour enhancement.

As described above, when the sharpening process shown in this embodiment is performed, the sharpening process is performed according to the signal level of the main signal S.
It is possible to adjust the frequency component whose contour is emphasized. For example, it is also possible to lower the frequency component whose outline is emphasized in the highlight part of the image as compared with other parts.

<4. Overall Configuration of Apparatus> Next, the overall configuration of the apparatus to which the embodiment of the present invention is applied will be described. FIG. 8 is a diagram showing an overall configuration of a device to which an embodiment of the present invention is applied. The image input unit 100 generates digital data of an original image such as an input scanner. Then, the generated image data is transferred to the image processing unit 200. The image data is stored in the image processing unit 2
After a predetermined process is performed in 00, the image is output to an image output unit 300 such as an output scanner. The image input unit 100, the image processing unit 200, and the image output unit 300
An operation input unit 201 and an information display unit 202 are provided to allow the operator to perform a desired operation.

The image processing section 200 is provided with an outline emphasis processing section 250. This contour emphasis processing section 250
May be an image sharpening filtering device described above, or may be an edge enhancement processing device using a smoothing filter described later.

<5. Enhancement of Tone Difference in Highlight Part> Here, a concept of enhancing the tone difference (contrast) of a thin line in a highlight part of an image in image outline enhancement will be described. In order to emphasize the gradation difference between the fine lines in the highlight part, it is necessary to emphasize the low-frequency component in the highlight part, which will be described.

FIG. 9 is an explanatory view showing the concept of enhancing the gradation difference between fine lines in a highlight portion of an image. FIG. 9B shows a case where an image having a density value distribution as shown in FIG. 9A is subjected to general edge enhancement processing. Here, the correction performed by the first correction unit 50 or the second correction unit 60 illustrated in FIG. 16 described in the related art is a correction for increasing or decreasing the density difference h. However, the thin line in the highlight portion of the image has blurred edges due to light flare and the like, and the density difference h is unlikely to occur. Even if such a density difference h is increased, it is difficult to visually distinguish the contrast C between FIGS. 9A and 9B. Further, noise, which is a high frequency component, is emphasized more than the edge.

Therefore, in the highlight portion of the image, as shown in FIG. 9C, by increasing the edge width w, the gradation difference of the fine line can be enhanced. Then, in order to make the outline thicker in the highlight portion of the image, an image filter having a large effective size may be used when filtering the target pixel in the highlight portion. By doing so, it becomes possible to emphasize the edge of the thin line which is a low frequency component. In the case of enhancing the outline of an image, in both the method using a smoothing filter and the method using a sharpening filter, the effectiveness of the highlight portion of the image can be improved by applying the filtering device described above. The size can be changed.

For example, in the image sharpening filtering device shown in FIG. 6, the relationship between the input (main signal S) and the output (ID value) in the mask ID extracting section 10 may be as shown in FIG. As shown in FIG. 4, the ID value is “5” in the highlight portion of the image, and the weighting coefficient set selection unit 20 changes the ID value to “5” from the sharpening filters shown in FIGS. Select the corresponding sharpening filter (FIG. 7 (e)). If the filtering process is performed using the sharpening filter shown in FIG. 7E, the edge width w shown in FIG. 9C becomes large because the frequency range is lower than other filters. As a result, the gradation difference (contrast) of the fine line can be enhanced in the highlight portion of the image.

<6. Image Edge Enhancement Processing Apparatus> Next, processing for enhancing an image outline using a smoothing filter will be described. FIG. 10 is a configuration diagram of an image edge enhancement processing apparatus according to the embodiment of the present invention. First, the mask ID extraction unit 10 outputs an ID value based on the main signal S of the pixel of interest. And a weighting coefficient set selection unit 2
At 0, a weighting coefficient for each component of the smoothing filter corresponding to the ID value is selected from among a plurality of smoothing filter configurations stored in a storage unit such as a memory, and output to the smoothing processing unit 30. The smoothing processing unit 30 generates the unsharp signal U by performing the operation shown in Expression 1 based on each of the weighting coefficients k- _mn to _kmn of the smoothing filter sent from the weighting coefficient set selection unit 20. . The processing so far is the filtering processing by the smoothing filter described above. For example, the main signal S as shown in FIG.
Is input, the unsharp signal U becomes as shown in FIG.
The signal becomes as shown in FIG.

Next, the unsharp signal U is subtracted from the main signal S to generate an unsharp mask signal (SU). This unsharp mask signal (SU) is shown in FIG.
The signal becomes as shown in FIG.

On the other hand, the main signal S is also input to the first correction unit 50, and the intensity coefficient P is output. In the first correction unit 50, as shown in FIG. 12, in order to reduce noise components that are conspicuous in a highlight portion and a shadow portion of an image,
The strength coefficient P is small. Then, by multiplying the unsharp mask signal (SU) by the intensity coefficient P, the contour emphasizing signal P (SU) as shown in FIG.
Is generated.

Then, the second correction section 60 generates a corrected contour emphasis signal Q according to the signal level of the contour emphasis signal P (SU) as shown in FIG. The correction shown in FIG. 13 has an effect of suppressing the graininess of an image that is conspicuous when the value of the edge enhancement signal P (SU) is small (in the case of a shadow portion of the image).

The obtained contour-enhanced signal Q and main signal S are combined to obtain a contour-enhanced signal E. The contour emphasized signal E is the sum of the main signal S and the corrected contour emphasized signal Q as shown in FIG.

In the case of the image edge enhancement processing apparatus shown in this embodiment, the smoothing processing unit 3 according to the main signal S
At 0, the effective size of the smoothing filter used can be changed. Therefore, if a smoothing filter having a relatively small effective size is used for the main signal S as shown in FIG. 14A, an unsharp signal U1 as shown in FIG. When a relatively large smoothing filter is used, an unsharp signal U2 as shown in FIG. 14C is obtained. When the edge enhanced signals E1 and E2 are generated in each case, FIG.
(D) and (e) are obtained. The contour-emphasized signal E1 has been processed by the smoothing filter having a small effective size and thus has a thin outline, and the contour-enhanced signal E2 has been processed with the smoothing filter having a large effective size and has a thick outline. Then, by using a smoothing filter having a large effective size in the highlight portion of the image, the gradation difference of the thin line can be emphasized.

Such an effect can also be obtained when the outline of an image is enhanced using a sharpening filter. In such a case, similarly, the sharpening filter having a large effective size has a thick outline, and conversely, the sharpening filter having a small effective size has a thin outline.

<7. Modification> In the image filtering apparatus for smoothing or sharpening described above, the mask ID extracting unit 10 sets the ID value of the input main signal S to "1, 1" as shown in FIG. 2,3,4,5 "
Can take only one of the following values:

Therefore, in order to further increase the accuracy of filtering, a relationship as shown in FIG. 15 is adopted. That is, an intermediate value below the decimal point, such as an ID value “2.5”, can be taken according to the density value of the main signal S. Then, the ID obtained from the correspondence shown in FIG.
The value is sent to the weighting coefficient set selection unit 20.

In the weighting coefficient set selecting section 20, the ID
The image filter corresponding to the value is selected, but only five types of image filters are stored in the storage unit of the weighting coefficient set selection unit 20. Therefore, for example, when the ID value is “2.5” at the time of the smoothing process, FIG.
An intermediate value of each of the weighting factors of the ID values “2” and “3” shown in (c) is derived, and the obtained value is used as each of the weighting factors of the ID value “2.5”.

Next, the explanation has been given on the assumption that the signal input to the mask ID extracting unit 10 is the main signal S which is the density value of the pixel of interest. However, the amount of noise included in the main signal S differs for each pixel. For example, there is a case where a shadow portion of an image contains a single pixel containing much noise and has a density value of a highlight portion. When that pixel is processed, it is filtered by an image filter used for the highlight portion. In this case, there is an inconvenience that the same processing as in the highlight portion is performed even though the density value is high due to noise.

Therefore, in order to eliminate such inconvenience, an average value of the main signal S for the target pixel and a plurality of pixels located around the target pixel is obtained, and the average value is input to the mask ID extraction unit 10. It is advisable to adopt a configuration that allows for this.
Thus, the noise component included in the main signal S is reduced, and a highly accurate filtering process can be performed.

In this embodiment, the case where the image filter is 5 × 5 has been described as an example.
Generally, an image filter of m × n (m, n is an integer of 2 or more) may be used as long as the target pixel to be processed is located at the center of the image filter.

[0056]

As described above, according to the first aspect of the present invention, in order to change the effective size of the image filter, the filtering is performed in accordance with the density value of the target pixel located at the center of the image filter. It is possible to change the frequency component to be processed. For example, it is possible to lower the frequency range in which the highlight portion of the image is smoothed or sharpened compared to other portions. In addition, it is possible to use an image filter having a different effective size for each pixel, and to reduce the degree of smoothing for a portion that is not to be blurred such as an intermediate density region of the image during the smoothing process, thereby sharpening the image. At the time of processing, the degree of emphasis can be reduced for high-density portions containing much noise in the image.

According to the second aspect of the present invention, the density value of the pixel of interest to be processed is detected, and the density value of the pixel of interest is changed using an image filter having an effective size corresponding to the density value. In addition, it is possible to change the frequency component to be filtered according to the density value of the target pixel.

According to the third aspect of the present invention, the density value of the target pixel is changed using an image filter having an effective size corresponding to the detected density value of the target pixel to be processed. In order to generate a contour emphasis signal by subtracting the density value from the density value before the change of the target pixel, and to generate a contour emphasized signal of the target pixel based on the contour emphasis signal,
The contour emphasis processing according to the density range can be realized, and in particular, it is possible to emphasize the gradation difference of the thin line in the highlight portion of the image. Therefore, the "white cloth texture" which is not conspicuous in the conventional processing becomes noticeable.

[Brief description of the drawings]

FIG. 1 is an explanatory diagram showing an outline of image filtering.

FIG. 2 is a diagram illustrating weighting coefficients of respective components of an image filter.

FIG. 3 is a diagram showing an image smoothing filtering device according to an embodiment of the present invention.

FIG. 4 is a diagram illustrating a relationship between an input and an output in a mask ID extraction unit.

FIG. 5 is a diagram illustrating a smoothing filter used in a smoothing process.

FIG. 6 is a diagram showing an image sharpening filtering device according to an embodiment of the present invention.

FIG. 7 is a diagram illustrating a sharpening filter used in a sharpening process.

FIG. 8 is a diagram showing an overall configuration of an apparatus to which an embodiment of the present invention is applied.

FIG. 9 is an explanatory diagram showing a concept of enhancing a tone difference between fine lines in a highlight portion of an image.

FIG. 10 is a configuration diagram of an image edge enhancement processing apparatus showing an embodiment of the present invention.

FIG. 11 is a diagram showing signals in each section of the image edge enhancement processing device.

FIG. 12 is a diagram illustrating a relationship between an input and an output in a first correction unit.

FIG. 13 is a diagram illustrating a relationship between an input and an output in a second correction unit.

FIG. 14 is a diagram illustrating a difference between an image filter having a small effective size and an image filter having a large effective size.

FIG. 15 is a diagram illustrating a modification of the relationship between input and output in the mask ID extraction unit.

FIG. 16 is a configuration diagram showing a conventional image edge enhancement processing device.

[Description of Signs] 10 Mask ID extraction unit 20 Weighting coefficient set selection unit 30 Smoothing processing unit 40 Sharpening processing unit 100 Image input unit 200 Image processing unit 250 Edge enhancement processing unit 300 Image output unit S Main signal U Unsharp signal SU unsharp mask signal P emphasis coefficient P (SU) contour emphasis signal Q corrected contour emphasis signal E contour emphasis signal

────────────────────────────────────────────────── ───

[Procedure amendment]

[Submission date] January 8, 1997

[Procedure amendment 1]

[Document name to be amended] Statement

[Correction target item name] Claim 2

[Correction method] Change

[Correction contents]

[Procedure amendment 2]

[Document name to be amended] Statement

[Correction target item name] 0016

[Correction method] Change

[Correction contents]

According to a second aspect of the present invention, a density value detecting means for detecting a density value of a target pixel to be processed and an image filter specifying means for specifying an image filter having an effective size corresponding to the density value are provided. And changing the density value of the pixel of interest using the image filter obtained by the image filter specifying means.

Claims (3)

[Claims] 1. A method for performing a predetermined process on an original image by scanning the original image with an m × n (m, n is an integer of 2 or more) image filter, comprising: An effective size of the image filter is changed in accordance with a density value of a target pixel located in a portion. 2. An image filter comprising: a density value detecting unit that detects a density value of a target pixel to be processed; and an image filter specifying unit that specifies an image filter having an effective size corresponding to the density value. An image filtering apparatus, wherein a density value of the pixel of interest is changed using an image filter obtained by a generation unit. 3. An apparatus for performing an outline emphasis process on an original image, comprising: a density value detecting means for detecting a density value of a pixel of interest to be processed; Image filter specifying means for specifying an image filter having the effective size, density value changing means for changing the density value of the pixel of interest using the image filter specified by the image filter specifying means, and density value changing means A contour emphasis signal generating means for subtracting the density value changed in the above from the density value of the target pixel before the change to generate a contour emphasis signal, and the contour emphasis signal obtained from the contour emphasis signal generation means. A contour emphasizing means for generating a contour emphasized signal of the pixel of interest.