JP2001016480A - Image processor, image processing method and medium - Google Patents

Abstract

PROBLEM TO BE SOLVED: To execute a fixed quantity of high-frequency boosting for any kind of image without being in proportion to the high frequency component of the image. SOLUTION: A discrete Fourier transforming part 4 individually executes a discrete Fourier transformation processing to image data stored in an image memory 2 in a horizontal direction and a vertical direction respectively. A line set mean calculating part 6 and a spectrum set mean calculating part 7 calculate a spectral mean value within a prescribed range of frequency in a plurality of lines. A divider 8 divides a constant number T by the calculated mean value and supplies the value to coefficient registers 10-1 and 10-2 as a coefficient. An adder 31 and a divider 32 calculate the mean value of the coefficients which are stored in the registers 10-1 and 10-2. The reciprocal of the calculated mean value is used as gain to be multiplied by an output from a high-pass filter which is constituted of LPF 34 of a high-frequency boosting filter 33, a delay part 35-1 and a subtracter 36.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an image processing apparatus, an image processing method, and a medium, and more particularly, to a gain of a filter for contour correction processing according to a result obtained by analyzing a high-frequency energy component of image data. The present invention relates to an image processing apparatus and method for changing image quality, and a medium.

[0002]

2. Description of the Related Art When performing contour correction on a predetermined image, the image data of the image is passed through a predetermined filter so that the energy is increased in proportion to the amount of high frequency components of the image. Was. At that time, the gain of the filter was fixed. 1 and 2 show the results of performing contour correction on different images (spectrum when DFT (Discrete Fourier Transform) is performed for each horizontal line). The image used in FIG. 1 is an image in which only the subject is in focus (an image with a small high-frequency component).
Is an image in which the focus from the near view to the distant view is sharply focused (an image having many high frequency components). However, FIG. 1 shows 1024 horizontal pixels × 661 vertical pixels, and FIG.
Used an RGB bitmap image of 1024 × 686 pixels.

In FIGS. 1A and 2A, the horizontal axis (X axis) indicates the sampling frequency, and the vertical axis (Y axis)
Indicates a spectrum at each sampling frequency. A frequency of 0 Hz on the X axis indicates direct current. In the case of a digital image image, as a characteristic, in most cases, a DC component having a frequency of 0 Hz tends to show the highest spectrum. Frequency 2π / 8 (π / 4) H
z is の of the sampling frequency (Fs) of the image,
That is, it is Fs / 2, which is a frequency called the Nyquist frequency.

When the above-described bitmap image image (hereinafter, appropriately referred to as a source image) is created, sampling is performed from an original image using a predetermined frequency. When performing the contour correction process from the source image,
Often lost. However, in digital image processing, it is possible to perform subsequent processing, for example, processing such as contour correction, without requiring a specific value of the sampling frequency Fs.

In FIG. 1A and FIG. 2A, the dark graphs are the spectrums of the source image, and the light graphs are through a high-pass filter and further DFT.
9 shows a frequency spectrum when processing is performed (when contour correction processing is performed). As the high-pass filter, a filter having the frequency characteristics shown in FIG. 3 was used.

FIGS. 1B and 2B are graphs obtained by subtracting a dark graph value from a light graph value in the graphs shown in FIGS. 1A and 2A, respectively. The result is shown on a graph having the same X axis. From these graphs, it is possible to know at which frequencies the results were enhanced and by how much. That is, it can be seen that the enhancement is centered on the high frequency component (2π / 8) of the source image.

[0007]

FIG. 1B and FIG. 2
By comparing with (B), when the frequency characteristic and the gain of the high-pass filter are constant, the high-frequency component as the output result of the high-pass filter is proportional to the energy depending on the strength of the high-frequency component energy included in the source image. You can see that In other words, since the characteristics of the high-pass filter are fixed, the larger the value before passing through the filter, the larger the value after passing through the filter. Will be done.

However, in the case of performing contour correction, the image after passing through the filter is fixed at a fixed magnification for any source image, in other words, for an image having many high frequency components or an image having few high frequency components. It is desirable that a certain amount of high frequency enhancement be performed instead of high frequency enhancement. That is, FIG. 1 (B) and FIG.
It is desirable that the graphs shown in FIG.

The above will be further described with reference to FIG. FIG. 4B shows an image composed of a black portion and a white portion as shown in FIG. 4A with an analog waveform. Further, FIG. 4C shows an analog waveform when the image shown in FIG. 4A is subjected to contour correction using a filter. It can be seen that by correcting the contour, the signal level of the black portion is enhanced by the level a and the signal level of the white portion is enhanced by the level b at the boundary between the black portion and the white portion.

Similarly, a black portion as shown in FIG.
FIG. 4E shows an image composed of a gray part and an analog waveform, and FIG. 4F shows a waveform obtained by performing contour correction using a filter. become. From FIG. 4 (F), it can be seen that the signal level of the black part is enhanced by the level a ′ and the signal level of the gray part is enhanced by the level b ′ by performing the contour correction.

Comparing the levels a and a 'and the levels b and b', respectively, shows that the levels a and b are higher than the levels a 'and b', respectively. This is because the waveform shown in FIG. 4E has a smaller signal level than the waveform shown in FIG. 4B, and if the signal level is so small, the level after contour correction is also small, This indicates that when the signal level is high, the level after the contour correction is also high.

[0012] This means that in a photographic image in which the focus is sharp from a distant view to a close view, and the high-frequency component is a high energy, the contour is too emphasized, ringing appears, and jaggies of oblique lines become conspicuous. Or
In many cases, the image after contour correction tends to look unnatural. Therefore, if the gain of the filter is set small in order to prevent such unnaturalness, an expected contour correction effect can be obtained for an image having a small high-frequency component (low contrast). There was a problem that there was no.

The present invention has been made in view of such a situation, and analyzes a high frequency component of an input image, and according to the obtained result, a gain of a filter for contour correction processing. Is changed so that a certain amount of enhancement can be performed even in different images, thereby suppressing the occurrence of ringing and jaggy, and performing an outline correction process without unnaturalness in the image.

[0014]

An image processing apparatus according to claim 1, wherein a storage means for storing image data, and a discrete Fourier transform processing means for performing a discrete Fourier transform process on the image data stored by the storage means Average value calculating means for calculating an average value of a spectrum in a predetermined frequency band of the image data from the image data subjected to the discrete Fourier transform by the discrete Fourier transform processing means; and a predetermined value calculated by the average value calculating means. Division means for dividing by an average value, extraction means for extracting high-frequency components of image data stored in the storage means, and multiplication means for multiplying the image data extracted by the extraction means by a value calculated by the division means And addition means for adding the image data multiplied by the multiplication means and the image data stored by the storage means. .

The discrete Fourier transform processing means may read out the image data stored in the storage means for each predetermined line, and perform the discrete Fourier transform processing for each of the read lines. .

According to a third aspect of the present invention, there is provided an image processing method comprising: a storing step of storing image data; a discrete Fourier transform processing step of performing a discrete Fourier transform process on the image data stored in the storing step; and a discrete Fourier transform processing step An average value calculating step of calculating an average value of a spectrum within a predetermined frequency band included in the image data from a result of the discrete Fourier transform performed in the above, and dividing the predetermined value by the average value calculated in the average value calculating step Dividing step,
An extraction step of extracting high-frequency components of the image data stored in the storage step, a multiplication step of multiplying the image data extracted in the extraction step by a value calculated in the division step, and an image multiplied in the multiplication step An adding step of adding the data and the image data stored in the storing step.

A storage medium for storing image data, a discrete Fourier transformation processing step for performing a discrete Fourier transformation processing on the image data stored in the storage step, and a discrete Fourier transformation processing step An average value calculating step of calculating an average value of a spectrum within a predetermined frequency band included in the image data from a result of the discrete Fourier transform performed in the above, and dividing the predetermined value by the average value calculated in the average value calculating step A dividing step, an extracting step of extracting high frequency components of the image data stored in the storing step, a multiplying step of multiplying the image data extracted in the extracting step by a value calculated in the dividing step, and a multiplying step. An addition step of adding the image data multiplied by and the image data stored in the storage step Characterized in that it comprises.

An image processing apparatus according to claim 1, wherein
In the image processing method described in the above, the image data is stored, the stored image data is subjected to a discrete Fourier transform process, and the average value of the spectrum in a predetermined frequency band is calculated. The calculated and predetermined value is divided by the calculated average value, and the value is multiplied by a high-frequency component extracted from the stored image data, and the multiplied image data and the stored image data are multiplied. Is added.

[0019]

FIG. 5 is a diagram showing a configuration of an embodiment of a contour correction circuit to which the present invention is applied. The contour correction circuit 1 is incorporated in, for example, a printer for a personal computer, and aims to improve the appearance of a printed image by performing contour correction on input image data. The configuration of the contour correction circuit 1 will be described based on a specific example. The image data input to the contour correction circuit 1 is stored in an image memory 2 (this image memory 2 is provided for convenience in describing a specific example, and may be shared with other processing. if there is,
In particular, it is not necessary to prepare only for the contour correction circuit 1). Since the discrete Fourier transform is desirably performed on each of the vertical and horizontal directions of the image, the switch 3 switches the readout direction. For example, when image data is read for each horizontal line, the switch 3 is connected to the terminal A. After reading all the lines, the switch 3 is connected to the terminal B for reading every vertical line.

The discrete Fourier transform unit 4 subjects the image data read from the image memory 2 to one-dimensional discrete Fourier transform for each line, and outputs the result to a line set average calculating unit 6. The line set average calculator 6 adds the frequency spectra in the horizontal direction or the vertical direction one by one, and calculates the average value of the frequency spectrum of the entire screen in each direction. For this reason, every time the result of the discrete Fourier transform for one line is input to the line set average calculation unit 6, the adder 5 feeds back the accumulated result in the line set average calculation unit 6 and adds the results. The sum of the spectrum in the line is taken,
Finally, division is performed by the predetermined number. The set average of the horizontal or vertical one-dimensional discrete Fourier transform result obtained in this way is output to the spectrum set average calculation unit 7. In the spectrum set calculation unit 7, in order to estimate how much high-frequency energy the source image contains, the specified frequency is defined as a frequency band, and the sum of individual spectra included therein is calculated. The sum of the spectra thus obtained is then divided by a constant T input to the divider 8, that is, the reciprocal of the spectrum sum is multiplied by the constant T. The value is output to the switch 9. If the value is calculated by reading the image data in the horizontal direction (when the switch 3 is connected to the terminal A), the switch 9 is connected to the terminal A. The switch 9 is connected to the terminal B if it is calculated by reading the image data in the vertical direction (when the switch 3 is connected to the terminal B).

When the switch 9 is connected to the terminal A, the divided value is stored in the coefficient register 10-2. When the switch 9 is connected to the terminal B, the divided value is stored in the coefficient register 10-.
1 is stored.

The image data stored in the image memory 2 is also supplied to a horizontal high-frequency enhancement filter 11 in order to obtain a desired image after contour correction. The image data output from the image memory 2 is applied to a horizontal high-frequency enhancement filter 11.
It is supplied to an internal component LPF (Low Pass Filter) 12, a delay unit 13-1, and a delay unit 13-2. LP
As a result, an HPF (HighPass Filter) is configured by the F12, the delay unit 13-1, and the delay unit 13-2. In the subtractor 14, the output data of the LPF 12 is subtracted from the data output from the delay unit 13-1, and the LPF 1
Only the high-frequency components of the source image that have been blocked from passing through 2 remain and are output. This output is input to the multiplier 15, multiplied by the coefficient stored in the coefficient register 10-1, and input to the adder 16. The adder 16 adds the image data after high-pass filtering output from the multiplier 15 and the source image that has passed through the delay unit 13-2 to enhance the high-frequency components of the image, and outputs the result to the image memory 17.

The vertical high-frequency enhancement filter 18 has the same configuration as the horizontal high-frequency enhancement filter 11, and includes an LPF 19, delay units 20-1 and 20-2, a subtracter 21, a multiplier 22, and an adder 23. It is composed of However, the coefficient stored in the coefficient register 10-2 is supplied to the multiplier 22.

Next, the operation of the contour correction circuit 1 will be described. Image data supplied from another device, for example, a personal computer, is stored in the image memory 2.
The image data stored in the image memory 2 is
For example, when the resolution is 2048 × 1536 dots, data is read out every 16 lines. In this case, 128 lines (128 × 16 = 1048) in the vertical direction and 96 lines (96 × 16 = 1536) in the horizontal direction are read, respectively. When the switch 3 is connected to the terminal A, the data is read in the horizontal direction. When the switch 3 is connected to the terminal B, the image data for 96 lines is read in the vertical direction. Are read out, respectively.

As described above, when reading is performed for each predetermined line, the time required for the processing described later can be reduced, and the effect obtained by appropriately setting the lines to be read can be reduced. There is no reduction. Of course, all lines may be read out.

The image data read from the image memory 2 is input to the discrete Fourier transform unit 4. The discrete Fourier transform unit 4 converts the input image data on a line-by-line basis.
A discrete Fourier transform process based on the following equation (1) is performed.

[0027]

(Equation 1) In equation (1), X (n) indicates that the input data in each line is N discrete values of X ₁ , X ₂ , X ₃ ,..., X _n−1 .

Since the calculation result based on equation (1) includes a complex number, its real part is X (k_rl) and its imaginary part is X
Assuming that (k_im), the amplitude component R (k) is represented by the following equation (2).

[0029]

(Equation 2) The line set average calculation unit 6 uses the adder 5 for a predetermined number of times,
When the values calculated based on the above-described equations (1) and (2) are added (cumulative), the accumulated value is divided by the predetermined number of times, so that the vertical or horizontal Calculate the average value of the amplitude components. For example, switch 3
Is connected to the terminal A side, and when the image data is read in the horizontal direction, the line set average calculating unit 6 sets the number of lines subjected to the discrete Fourier transform processing by the discrete Fourier transform unit 4 to 1 to When the counted value reaches 96, the accumulated value output from the adder 5 is incremented by 9
By dividing by 6, the average value of the spectrum obtained in the horizontal reading is calculated.

The calculated average value of the spectrum is output to the spectrum set average calculating section 7. The spectrum set average calculation unit 7 calculates an average value of data in a predetermined frequency range from the input data. When a graph such as that shown in FIG. 6 is obtained as a graph from the data output from the line set average value calculation unit 6, for example, the average value of the spectrum between P1 and P2 is calculated as the spectrum set average. The calculation unit 7 calculates.

Specifically, the value of P1 and P2 is such that P1 is (1
/ 8) π and P2 are (3/8) π. This value is set in order to obtain an average value in a section of (前後) π before and after (（) π which is a half of the Nyquist frequency. The values of P1 and P2 are not limited to the values described above, but the values described above are set as the most effective values based on the results of evaluation using many source images.

The processing performed by the adder 5, the line set average calculator 6, and the spectrum set average calculator 7 is as follows.
This is performed based on the following equation. The value calculated by reading in the horizontal direction is defined as the average value Ah, and the value calculated by reading in the vertical direction is defined as the average value Av.

[0033]

(Equation 3)

(Equation 4) H in equation (3) indicates the number of read lines in horizontal reading, and V in equation (4) indicates the number of read lines in vertical reading. Also, P
1, P2 are the values described above.

The average value Ah or the average value Av thus calculated is supplied to the divider 8. Divider 8
Is supplied (or held), and the constant T is divided by the average value Ah or the average value Av supplied from the spectrum set average calculation unit 7. Average value Ah
Is supplied to the coefficient register 10-2 and stored therein when the switch 9 is connected to the terminal B. The value calculated by the average value Av is supplied to the coefficient register 10-1 and stored therein when the switch 9 is connected to the terminal A.

The value stored in the coefficient register 10-1 is
LPF 12 and delay unit 13 in horizontal high-frequency enhancement filter 11
The value output from the high-pass filter composed of −1 and the subtractor 14 is used as a value that is multiplied by the multiplier 15, that is, a gain. Similarly, the value stored in the coefficient register 10-2 is stored in the LPF 19, the delay unit 20-1, and the subtracter 21 in the vertical high-frequency enhancement filter 18.
To the value output from the high-pass filter composed of
It is used as a value to be multiplied by the multiplier 22. Hereinafter, the value stored in the coefficient register 10-1 is referred to as a gain Kh, and the value stored in the coefficient register 10-2 is referred to as a gain Kv.

The image data (source image) stored in the image memory 2 is also supplied to the horizontal high-frequency enhancement filter 11. The LPF 12 of the lateral high-pass filter 11 is
Of the input source image, only the components at or below a predetermined frequency are output to the subtractor 14. The delay unit 13-1 has an LPF
The input image data is delayed for the time required to complete the processing by the step 12 and output to the subtractor 14. Note that LPF
The gain of 12 is 1.

Therefore, the image data of the same source image is input to the subtractor 14, but one of them is image data of a low frequency band (data in which image data of a high frequency band is suppressed). The other is the image data of the source image stored in the image memory 2 itself. Subtractor 1
4 is a LP from the image data input from the delay unit 13-1.
By subtracting the image data input from F12, the image data from which the high-frequency component has been extracted is multiplied by the multiplier 1
5 is output. That is, the LPF 12, the delay unit 13-1,
The subtractor 14 forms a high-pass filter.

The multiplier 15 multiplies the output of the high-pass filter by the coefficient supplied from the coefficient register 10-1. The coefficient supplied from the coefficient register 10-1 is obtained by quantifying the frequency component representing the contour of the image from the average value of the entire screen as a result of the discrete Fourier transform of the source image and obtaining a value proportional to the reciprocal thereof. Has become. For this reason, when the high frequency component included in the source image is large,
When the coefficient value is small and the high frequency component is small, the coefficient value is supplied to the multiplier 15 as a large coefficient value. As a result, the high frequency component is enhanced while adapting to the source image.

The image data of only high-frequency components appropriately enhanced by the multiplier 15 is output to the adder 16. The adder 16 receives the image data input from the multiplier 15 and the image data supplied from the delay unit 13-2 (the image data of the source image supplied from the image memory 2). The adder 16 adds these two image data to generate image data in which the high-frequency component is appropriately enhanced as compared with the source image, and outputs the image data to the image memory 17.

The coefficients stored in the coefficient registers 10-1 and 10-2 are numerical values between 0.0 and 1.0. Therefore, the image data output from the adder 16 has a maximum of 2 high-frequency components compared to the image data of the source image.
The image data is doubled.

The image data whose horizontal high-frequency component has been enhanced in this way is temporarily stored in an image memory 17 and output to a vertical high-frequency enhancement filter 18 in order to enhance the vertical high-frequency component. You. The processing performed in the vertical high-frequency enhancement filter 18 is basically the same as the processing performed in the horizontal high-frequency enhancement filter 11, and a description thereof will be omitted. However, the coefficient supplied to the multiplier 22 is a value stored in the coefficient register 10-2.

FIG. 7 shows a spectrum when the coefficient for enhancing the high frequency component is changed for each source image. FIG. 7A is a spectrum obtained when the source image obtained from the graph of FIG. 2A is corrected by the contour correction circuit 1 of FIG. As in FIG. 2A,
7A, the dark graph shows the spectrum of the source image, and the light graph shows the spectrum of the source image in FIG.
3 shows a spectrum after the processing by the contour correction circuit 1 shown in FIG. FIG. 7B shows the result of calculating the difference between the dark graph and the light graph.

For the graph shown in FIG. 7, the source image used to obtain the graph shown in FIG. 2 was used, and the result (FIG. 7 (B)) is shown in the graph shown in FIG. 2 (B). than,
This is similar to the graph shown in FIG. The source image used to obtain the graph of FIG. 1 was an image with a small amount of high-frequency components, and the source image used to obtain the graph of FIG. 2 (FIG. 7) was an image with a large amount of high-frequency components. As described above, even in a source image having different high-frequency components, it is possible to enhance the high-frequency components by a fixed amount without being proportional to the high-frequency components of the source image. FIG. 8 is a graph showing frequency characteristics of the high-pass filter unit in the contour correction circuit 1 shown in FIG. The characteristic of this high-pass filter is that the gain changes according to the source image,
FIG. 8 shows characteristics when the source image obtained from the graph of FIG. 7 is processed.

FIG. 9 is a graph obtained when two different source images are processed using the contour correction circuit 1 (filter) shown in FIG. A source image composed of a black portion and a white portion as shown in FIG.
FIG. 9B shows an analog waveform. Further, FIG. 9C shows an analog waveform when the contour of the source image as shown in FIG. 9A is corrected using a filter. It can be seen that, by correcting the contour at the boundary between the black part and the white part, the signal level of the black part is emphasized by the level a and the signal level of the white part is emphasized by the level b.

Similarly, a black portion as shown in FIG.
FIG. 9E shows a source image composed of a gray portion and an analog waveform, and FIG. 9F shows a waveform obtained by performing contour correction using a filter. Become like From FIG. 9F, it can be seen that the signal level of the black portion is enhanced by the level a ′ and the signal level of the gray portion is enhanced by the level b ′ by performing the contour correction.

The levels a and b shown in FIG.
The level a ′ and the level b ′ shown in FIG. 9F have substantially the same level value. As described above, even in different source images, a certain amount of enhancement can be performed without being proportional to the high frequency component of the source image.

In the above description, the contour is corrected by independently controlling the vertical and horizontal directions (using different coefficients).
The same coefficient may be used. FIG. 10 shows a configuration of the contour correction circuit 1 that performs contour correction using the same coefficient. Portions performing the same processes as those of the contour correction circuit 1 shown in FIG. 5 are denoted by the same reference numerals, and description thereof will be omitted as appropriate. The high-frequency enhancement filter 33 shown in FIG. 10 has the same configuration as the horizontal high-frequency enhancement filter 11 and the vertical high-frequency enhancement filter 18 shown in FIG.

Coefficient register 10-1 and coefficient register 10
The coefficients stored in -2 are respectively supplied to the adder 31. The adder 31 adds the two supplied coefficients,
Output to the divider 32. The divider 32 divides the input value by 2, and outputs the result to the multiplier 37 of the high-frequency enhancement filter 33. As described above, the high-frequency enhancement filter 33 performs contour correction of the image data of the input source image using the average value of the coefficients calculated from the horizontal direction and the vertical direction.

As described above, even when the contour correction is performed using the average value of the coefficients obtained from the vertical direction and the horizontal direction, since the coefficients are calculated for each source image, a different source is used for the image. A certain amount of enhancement (contour correction) can be performed on high-frequency components.

The series of processes described above can be executed by hardware, but can also be executed by software. When a series of processing is executed by software, a program constituting the software is configured by a contour correction circuit 1 as dedicated hardware.
It is installed in, for example, a general-purpose personal computer or the like, which can execute various functions by installing a computer incorporated in the PC or various programs.

Next, referring to FIG. 11, a program for executing the above-described series of processing is installed in a computer, and a computer used for making the computer executable is a general-purpose personal computer. The following is an example of the case.

As shown in FIG. 11A, the program can be provided to the user in a state where the program is installed in a hard disk 52 or a semiconductor memory 53 as a recording medium built in the personal computer 51 in advance.

Alternatively, the program is executed as shown in FIG.
As shown in (B), floppy disk 61, CD-ROM
(Compact Disk-Read Only Memory) 62, MO (Magne
to-Optical) 63, DVD (Digital Versatile Disk) 6
4. It can be temporarily or permanently stored in a recording medium such as a magnetic disk 65 or a semiconductor memory 66 and provided as package software.

Further, as shown in FIG. 11C, the program is wirelessly transferred from a download site 71 to a personal computer 51 via an artificial satellite 72 for digital satellite broadcasting, or a local area network or the Internet. The data can be transferred by wire to the personal computer 51 via the network 81 and stored in the personal computer 51 in a built-in hard disk or the like.

The medium in the present specification has a broad concept including all these media.

The personal computer 51 is, for example,
As shown in FIG. 12, a CPU (Central Processing Unit)
92 is built in. An input / output interface 95 is connected to the CPU 92 via a bus 91.
From the user via the input / output interface 95
When a command is input from an input unit 97 including a keyboard, a mouse, and the like, a program stored in a ROM (Read Only Memory) 93 corresponding to the semiconductor memory 53 in FIG. . Alternatively,
The CPU 92 is a program stored in the hard disk 52 in advance, transferred from the satellite 72 or the network 81, received by the communication unit 98, and further installed in the hard disk 52 or the floppy disk 61 attached to the drive 99. CD-R
The program read from the OM 62, the MO disk 63, the DVD 64, or the magnetic disk 65 and installed on the hard disk 52 is stored in a RAM (Random Access Memory).
y) Load to 94 and execute. Further, the CPU 92 outputs the processing result to a display section 96 such as an LCD (Liquid Crystal Display) via the input / output interface 95 as necessary.

In this specification, the steps of describing a program provided by a medium may be performed in a chronological order according to the described order. Alternatively, it also includes individually executed processing.

[0058]

As described above, according to the image processing apparatus according to the first aspect, the image processing method according to the third aspect, and the medium according to the fourth aspect, the image data is stored. Performs a discrete Fourier transform process on the image data, calculates an average value of the spectrum in a predetermined frequency band, divides the predetermined value by the calculated average value, and calculates the value from the stored image data. Since the extracted high-frequency component is multiplied and the multiplied image data and the stored image data are added, it is possible to enhance a certain amount of high-frequency component even for different image data. .

[0059]

[Brief description of the drawings]

FIG. 1 is a diagram showing a graph of a spectrum obtained from an image having a small high-frequency component.

FIG. 2 is a diagram showing a graph of a spectrum obtained from an image having many high-frequency components.

FIG. 3 is a diagram illustrating characteristics of a high-pass filter.

FIG. 4 is a diagram illustrating a result of performing contour correction processing on different source images.

FIG. 5 is a block diagram illustrating a configuration of a contour correction circuit.

FIG. 6 is a diagram for explaining processing in a spectrum set average calculation unit 7;

7 is a diagram showing a graph of a spectrum obtained from an image having many high-frequency components when the contour correction circuit 1 of FIG. 5 is used.

FIG. 8 is a diagram illustrating characteristics of a high-pass filter.

9 is a diagram illustrating a result of performing contour correction processing on different source images when the contour correction circuit 1 of FIG. 5 is used.

FIG. 10 is a block diagram showing another configuration of the contour correction circuit 1;

FIG. 11 is a diagram illustrating a medium.

FIG. 12 is a block diagram showing an internal configuration of the personal computer shown in FIG.

[Explanation of symbols]

Reference Signs List 1 contour correction circuit, 2 image memory, 4 discrete Fourier transform section, 6 line set average calculation section, 7 spectrum set average calculation section, 8 divider, 10 coefficient register, 11 horizontal high band enhancement filter, 17 image memory, 18 Vertical high-pass filter

 ──────────────────────────────────────────────────続 き Continued on the front page F term (reference) 5B057 CA08 CA12 CA16 CB08 CB12 CB16 CE03 CH09 DA16 DB02 DB09 5C021 PA34 PA42 PA54 PA76 PA79 RA02 RB05 SA22 SA25 XB03 ZA01 5C077 LL01 LL19 MP07 PP03 PP49

Claims (4)

[Claims] A storage means for storing image data; a discrete Fourier transform processing means for performing a discrete Fourier transform process on the image data stored by the storage means; and a discrete Fourier transform performed by the discrete Fourier transform processing means. Average value calculating means for calculating an average value of a spectrum in a predetermined frequency band of the image data from the image data, and dividing means for dividing a predetermined value by the average value calculated by the average value calculating means. Extracting means for extracting high-frequency components of the image data stored in the storage means; multiplying means for multiplying the image data extracted by the extracting means by a value calculated by the dividing means; Addition means for adding the image data multiplied by the multiplication means and the image data stored by the storage means; An image processing apparatus comprising: 2. The discrete Fourier transform processing means reads the image data stored by the storage means for each predetermined line, and for each of the read lines,
2. A discrete Fourier transform process is performed.
An image processing apparatus according to claim 1. 3. A storage step of storing image data; a discrete Fourier transform processing step of performing a discrete Fourier transform process on the image data stored in the storage step; and a discrete Fourier transform performed by the discrete Fourier transform step An average value calculating step of calculating an average value of a spectrum in a predetermined frequency band of the image data from the image data; anda dividing step of dividing a predetermined value by the average value calculated in the average value calculating step, An extraction step of extracting a high-frequency component of the image data stored in the storage step; a multiplication step of multiplying the image data extracted in the extraction step by a value calculated in the division step; The image data multiplied in the step and the image data stored in the storing step And an adding step of adding the following. 4. A storage step of storing image data; a discrete Fourier transform processing step of performing a discrete Fourier transform process on the image data stored in the storage step; and a discrete Fourier transform performed by the discrete Fourier transform processing step An average value calculating step of calculating an average value of a spectrum in a predetermined frequency band of the image data from the image data; anda dividing step of dividing a predetermined value by the average value calculated in the average value calculating step, An extraction step of extracting a high-frequency component of the image data stored in the storage step; a multiplication step of multiplying the image data extracted in the extraction step by a value calculated in the division step; The image data multiplied in the step and the image data stored in the storing step And a computer-executable program for causing a computer to execute the program.