JP4213337B2 - Image processing method and apparatus, and recording medium - Google Patents

Description

[0001]
BACKGROUND OF THE INVENTION
The present invention relates to an image processing method and apparatus for performing non-sharp mask processing (so-called blur mask processing) for emphasizing a high-frequency component of the image signal, and a computer recording a program for causing the computer to execute the image processing method. The present invention relates to a readable recording medium.
[0002]
[Prior art]
In various fields, an image signal representing an image is obtained, and an appropriate image process is performed on the image signal, and then the image is reproduced and displayed. For example, in order to improve the diagnostic performance of radiographic images, a method of applying frequency enhancement processing to an image signal using a non-sharp mask (blur mask) has been proposed by the present applicant (Japanese Patent Laid-Open No. 55-163472). JP, 55-87953, etc.). In this frequency enhancement process, a signal related to the high frequency component of the original image is obtained by subtracting the unsharp mask image signal Sus from the original image signal Sorg to the read original image signal Sorg, and the enhancement coefficient β is added to the signal related to the high frequency component. Is applied to the original image signal Sorg, thereby enhancing a predetermined spatial frequency component in the image. This is expressed by the following equation (1).
[0003]
Sproc = Sorg + β × (Sorg−Sus) (1)
(Sproc: frequency enhanced signal, Sorg: original image signal,
Sus: unsharp mask image signal, β: enhancement factor)
Here, the non-sharp mask image signal Sus is, for example, about the original image signal Sorg within an M × N range around each pixel for every other pixel constituting the image.
Sus = ΣSorg / (M × N) (2)
Is obtained by performing the following calculation.
[0004]
Japanese Patent Laid-Open No. 10-75395 discloses a method for preventing the occurrence of artifacts in a frequency-enhanced signal by adjusting the frequency response characteristics of a signal related to a high frequency component added to an original image signal. Proposed. In this method, first, a plurality of non-sharp mask image signals having different sharpness, that is, frequency response characteristics are created, and the difference between the two signals in the non-sharp mask image signal and the original image signal is taken. After creating a plurality of band limited image signals representing the frequency components of a certain limited frequency band of the original image signal, and further converting the band limited image signals to each desired size by different conversion functions The signal relating to the high-frequency component is created by integrating the plurality of suppressed band-limited image signals. This process can be expressed by, for example, the following formula (3).
[0005]
Sproc = Sorg + β (Sorg) × Fusm (Sorg, Sus1, Sus2,... SusN)
Fusm (Sorg, Sus1, Sus2, ... SusN)
= F1 (Sorg-Sus1) + f2 (Sus1-Sus2) + ...
+ Fk (Susk-1-Susk) + ... + fn (SusN-1-SusN) (3)
(However, Sproc: processed image signal
Sorg: Original image signal
Susk (k = 1 to N): Unsharp mask image signal
fk (k = 1 to N): conversion function for converting the band-limited image signal
β (Sorg): enhancement coefficient determined based on the original image signal)
The unsharp mask image signal used for the processing described in the above-mentioned Japanese Patent Application Laid-Open No. 10-75395 is first subjected to a predetermined filtering process at predetermined intervals with respect to the pixels of the original image signal, thereby reducing the number of pixels in the vertical and horizontal directions. The number of pixels is reduced to 1/2 by repeating the same filtering process on the image signal obtained in this manner.^2kA plurality of reduced image signals that are reduced step by step are created, and each reduced image signal is created by performing an interpolation process so that the number of pixels is the same as the original image by a predetermined interpolation method. The Therefore, the non-sharp mask image signal has the same number of pixels as the original image signal but is an image signal representing an image having a sharpness lower than that of the original image signal.
[0006]
Further, the band limited image signal is created, for example, by obtaining a difference between unsharp mask image signals in adjacent frequency bands, or by obtaining a difference between the original image signal and each unsharp mask image signal. Therefore, the band limited image signal has the same number of pixels as that of the original image, and represents a frequency response characteristic for each frequency band of the original image signal.
[0007]
[Problems to be solved by the invention]
The band limited image signal is 1/2 of the original image.^2NTherefore, the frequency band of each band limited image signal is 1/2 of the Nyquist frequency of the original image signal.^NIt represents the frequency band of the interval. For this reason, 1/2 the Nyquist frequency^NIt is possible to perform frequency processing of the original image at a frequency interval of. On the other hand, half the Nyquist frequency^NIt is desired to perform processing by dividing the frequency band into a frequency band smaller than the frequency band.
[0008]
The present invention has been made in view of the above circumstances, and provides an image processing method and apparatus capable of obtaining a band-limited image signal divided into finer frequency bands and performing image processing in a finer frequency band, and It is an object of the present invention to provide a computer-readable recording medium on which a program for causing a computer to execute an image processing method is recorded.
[0009]
[Means for Solving the Problems]
An image processing method of the present invention is an image processing method for enhancing a high frequency component of an original image by adding a signal related to the high frequency component of the original image to an original image signal representing the original image.
A process based on square checkered reduction processing and checkered square reduction processing is performed on the original image signal to create a plurality of non-sharp mask image signals having different frequency response characteristics from each other,
Based on the original image signal and the plurality of non-sharp mask image signals, or the plurality of non-sharp mask image signals, create a plurality of band limited image signals representing images of the plurality of frequency bands of the original image,
A predetermined conversion process is performed on at least one band-limited image signal among the plurality of band-limited image signals to create at least one converted image signal,
A signal relating to the high frequency component to be added to the original image signal is obtained based on the converted image signal.
[0010]
“Square checkered reduction processing” refers to a method of performing a filtering process based on pixel values of pixels adjacent to each other in the upper, lower, left, and right directions alternately with respect to an image signal of a square pixel array. This is a process of changing the pixel arrangement from square to checkered while reducing the number of pixels. By this square checkered reduction process, the processed image signal becomes a checkered pixel array, and the number of pixels becomes approximately ½ of the image signal of the square pixel array before the process.
[0011]
The “checkered square reduction process” refers to a checkered pixel array image signal in which pixels in adjacent pixel rows are alternately thinned out or a filtering process based on pixel values of pixels adjacent in a diagonal direction is performed. Thus, the pixel array is changed from checkered to square while decreasing the pixel value. By this checkered square reduction process, the processed image signal becomes a square array, and the number of pixels becomes approximately ½ of the checkered array image signal before the process.
[0012]
Here, the “square pixel arrangement” means that pixels in the image represented by the image signal are arranged with the closest pixels arranged at equal intervals in the vertical and horizontal directions as shown in FIG.
[0013]
As shown in FIG. 48, the “checkered pixel arrangement” is such that each pixel of the image represented by the image signal is arranged in the diagonal direction at equal intervals, and the pixels are arranged between pixels in a certain pixel column. An arrangement in which each pixel in a pixel column adjacent to the pixel column is positioned.
[0014]
“Processing based on square checkered reduction processing and checkered square reduction processing” refers to the following processing. First, when the original image signal is a square pixel array, a square checkered reduction process is performed on the original image signal to create a reduced image signal of the checkered pixel array, and further, a checkered square reduction is performed on the reduced image signal. Processing is performed to create a reduced image signal having a square pixel array. Then, a square checkered reduction process and a checkered square reduction process are alternately performed on the obtained reduced image signal, thereby obtaining a plurality of reduced image signals with the number of pixels reduced stepwise. On the other hand, when the original image signal is a checkered pixel array, a checkered square reduction process is performed on the original image signal to create a reduced image signal of the checkered pixel array, and a square checkered reduction is further performed on the reduced image signal. Processing is performed to create a reduced image signal of a checkered pixel array. Then, a checkered square reduction process and a square checkered reduction process are alternately performed on the obtained reduced image signal, thereby obtaining a plurality of reduced image signals in which the number of pixels is reduced stepwise. The reduced image signal is subjected to an interpolation operation so that the number of pixels is the same as that of the original image by an interpolation method such as a spline interpolation operation.
[0015]
Note that when performing the interpolation calculation, an interpolation filter in which interpolation coefficients are arranged according to the pixel arrangement of the reduced image signal is used. For example, when the reduced image signal is a square pixel array, an interpolation filter in which interpolation coefficients are arranged in a square shape is used, and when the reduced image signal is a checkered pixel array, the interpolation coefficients are in a checkered pattern. Use an interpolation filter with. As the interpolation filter, for example, an interpolation filter that performs B-spline interpolation calculation can be used. In this case, as the interpolation filter in which the interpolation coefficients are arranged in a checkered pattern, a filter having an interpolation coefficient obtained by rotating the interpolation coefficient of the interpolation filter in which the interpolation coefficients are arranged in a square shape by 45 degrees can be used.
[0016]
Then, a plurality of non-sharp mask image signals having different frequency response characteristics are generated by the “process based on the square checkered reduction process and the checkered square reduction process”.
[0017]
“Acquiring a signal relating to a high frequency component to be added to the original image signal based on the converted image signal” means that when a predetermined conversion process is performed only on one band-limited image signal, only one converted image signal is obtained. Since it is obtained, this one converted image signal is referred to as a signal related to a high frequency component. On the other hand, when a predetermined conversion process is performed on two or more band-limited image signals, two or more converted image signals are obtained. By adding these converted image signals, a signal related to a high-frequency component is obtained. Say to get.
[0018]
In the image processing method according to the present invention, it is preferable that the predetermined conversion process is a process of reducing at least a part of a signal value of the band limited image signal.
[0019]
Further, the predetermined conversion process may be a process of decreasing the absolute value of the signal value when the absolute value of the band limited image signal is larger than a predetermined threshold.
[0020]
At this time, the absolute value of the signal value of the band limited image signal may be reduced as the absolute value of the signal value of the band limited image signal is smaller than another threshold value smaller than the predetermined threshold value.
[0021]
The absolute value of the signal value of the band limited image signal may be changed according to the frequency band of the band limited image signal.
[0022]
In addition, the predetermined conversion processing is performed based on the absolute value of the signal value of the band limited image signal based on a function that is different depending on the frequency band of the band limited image signal. The conversion may be performed so that the value is equal to or less than the value. When converting a plurality of band limited image signals, a different function is used for each frequency band of each band limited image signal.
[0023]
At this time, if the absolute value of the signal value of the band-limited image signal is larger than a predetermined value, the function converts the band-limited image signal so that the signal value of the converted image signal becomes a substantially constant value. The predetermined value is preferably a smaller value as the band limited image signal is in a higher frequency band.
[0024]
Further, the function is obtained by converting the signal value of the converted image signal obtained by converting the signal value within a predetermined range in which the absolute value in the band limited image signal is close to 0 as the band limited image signal is in a lower frequency band. The function is preferably a function for converting the band-limited image signal so that the absolute value becomes a smaller value.
[0025]
The creation of the band limited image signal, the creation of the converted image signal, the creation of a signal related to the high frequency component, and the addition of the signal related to the high frequency component to the original image signal are specifically expressed by the following equations: Therefore, it may be performed.
[0026]
Sproc = Sorg + β (Sorg) · Fusm (Sorg, Sus1, Sus2, ... SusN)
Fusm (Sorg, Sus1, Sus2, ... SusN)
= {F1 (Sorg-Sus1) + f2 (Sus1-Sus2) +
+ Fk (Susk-1−Susk) +... + FN (SusN−1−SusN)}
(However, Sproc: Image signal in which high frequency components are emphasized.
Sorg: Original image signal
Susk (k = 1 to N): Unsharp mask image signal
fk (k = 1 to N): function for performing a predetermined conversion process
β (Sorg): enhancement coefficient determined based on the original image signal)
Or you may carry out according to the following formula.
[0027]
Sproc = Sorg + β (Sorg) · Fusm (Sorg, Sus1, Sus2, ... SusN)
Fusm (Sorg, Sus1, Sus2, ... SusN)
= (1 / N) · {f1 (Sorg−Sus1) + f2 (Sorg−Sus2) +
+ Fk (Sorg-Susk) + ... + fN (Sorg-SusN)}
(However, Sproc: Image signal in which high frequency components are emphasized.
Sorg: Original image signal
Susk (k = 1 to N): Unsharp mask image signal
fk (k = 1 to N): function for performing a predetermined conversion process
β (Sorg): enhancement coefficient determined based on the original image signal)
When the original image signal is a square pixel array, an unsharp mask image signal with an odd number k is obtained by performing an interpolation operation on a reduced image signal with a checkered pixel array, and an unsharp mask image signal with an even number k. Is obtained by performing an interpolation operation on the reduced image signal of the square pixel array. On the other hand, when the original image signal is a checkered pixel array, an unsharp mask image signal with an odd number k is obtained by performing an interpolation operation on a reduced image signal with a square pixel array, and an unsharp mask image signal with an even number k. Is obtained by performing an interpolation operation on the reduced image signal of the checkered pixel array.
[0028]
Further, the creation of the band limited image signal, the creation of the converted image signal, the creation of a signal related to the high frequency component, and the addition of the signal related to the high frequency component to the original image signal may be performed according to the following equations: .
[0029]
Sproc = Sorg + β (Sorg) · Fusm (Sus1, Sus2, ... SusN)
Fusm (Sus1, Sus2, ... SusN)
= {F2 (Sus1-Sus2) + f3 (Sus2-Sus3) + ...
+ Fk (Susk-1−Susk) +... + FN (SusN−1−SusN)}
(However, Sproc: Image signal in which high frequency components are emphasized.
Sorg: Original image signal
Susk (k = 1 to N): Unsharp mask image signal
fk (k = 1 to N): function for performing a predetermined conversion process
β (Sorg): enhancement coefficient determined based on the original image signal)
Or you may carry out according to the following formula.
[0030]
Sproc = Sorg + β (Sorg) · Fusm (Sus1, Sus2, ... SusN)
Fusm (Sus1, Sus2, ... SusN)
= (1 / N) · {f2 (Sus1-Sus2) + f3 (Sus1-Sus3) +
+ Fk (Sus1-Susk) + ... + fN (Sus1-SusN)}
(However, Sproc: Image signal in which high frequency components are emphasized.
Sorg: Original image signal
Susk (k = 1 to N): Unsharp mask image signal
fk (k = 1 to N): function for performing a predetermined conversion process
β (Sorg): enhancement coefficient determined based on the original image signal)
Furthermore, the predetermined conversion process in each of the above equations may be a conversion process corresponding to the enhancement coefficient. At this time, the conversion process according to the enhancement coefficient is a conversion process that differs depending on the frequency band of the band-limited image signal, and the band limit image signal becomes lower as the frequency band of the band-limited image signal is lower. It is a process of greatly suppressing the image signal, and it is desirable that the larger the enhancement coefficient, the larger the difference between the degree of suppression for the band limited image signal in the high frequency band and the degree of suppression for the band limited image signal in the low frequency band.
[0031]
Further, the predetermined conversion processing creates a suppressed image signal by converting the band limited image signal so as to be a value equal to or less than the absolute value determined based on the absolute value of the signal value of the band limited image signal. And
Based on the original image signal and the plurality of non-sharp mask image signals, or the plurality of non-sharp mask image signals, including a signal in a frequency band lower than the band-limited image signal used to create the suppression image signal Create an auxiliary image signal,
The auxiliary image signal is converted so that the smaller the absolute value of the signal value of the auxiliary image signal is, the closer the value is to 1, and the larger the value is, the closer the value is to 0, thereby creating a magnification signal corresponding to each of the suppressed image signals. ,
Processing for obtaining the converted image signal by multiplying the suppression image signal by the magnification signal corresponding to the suppression image signal may be performed.
[0032]
Specifically, the band-limited image signal, the converted image signal, the signal related to the high-frequency component, and the addition of the signal related to the high-frequency component to the original image signal are expressed by the following equations: You may go according to.
[0033]
Sproc = Sorg + β (Sorg) · Fusm (Sorg, Sus1, Sus2, ... SusN)
Fusm (Sorg, Sus1, Sus2, ... SusN)
= {F1 (Sorg−Sus1) · g (Sus1−Sus2)
+ F2 (Sus1-Sus2) · g (Sus2-Sus3) + ...
+ Fk (Susk-1−Susk) · g (Susk−Susk + 1) +
+ FN (SusN-1−SusN) · g (SusN−SusN + 1)}
(However, Sproc: Image signal in which high frequency components are emphasized.
Sorg: Original image signal
Susk (k = 1 to N + 1): unsharp mask image signal
fk (k = 1 to N): function for creating a suppression image signal
g: a function for converting the auxiliary image signal to create a magnification signal
β (Sorg): enhancement coefficient determined based on the original image signal)
Or you may carry out according to the following formula.
[0034]
Sproc = Sorg + β (Sorg) · Fusm (Sorg, Sus1, Sus2, ... SusN)
Fusm (Sorg, Sus1, Sus2, ... SusN)
= {F1 (Sorg-Sus1) · g (Sorg-Sus2)
+ F2 (Sus1-Sus2) g (Sorg-Sus3) + ...
+ Fk (Susk-1−Susk) · g (Sorg−Susk + 1) +
+ FN (SusN-1−SusN) · g (Sorg−SusN + 1)}
(However, Sproc: Image signal in which high frequency components are emphasized.
Sorg: Original image signal
Susk (k = 1 to N + 1): unsharp mask image signal
fk (k = 1 to N): function for creating a suppression image signal
g: a function for converting the auxiliary image signal to create a magnification signal
β (Sorg): enhancement coefficient determined based on the original image signal)
You may go according to.
[0035]
Further, the predetermined conversion process passes a low frequency side band limited image signal, which is a band limited image signal of a frequency band lower than the converted band limited image signal, which is the at least one band limited image signal, through the origin. An auxiliary image signal of the band-limited image signal to be converted is generated by performing conversion based on a nonlinear function in which the inclination at the origin is substantially 0 and the inclination gradually increases as the value to be processed increases. A composite band limited image signal is created by adding to the converted band limited image signal, and the composite band limited image signal is less than the absolute value determined based on the absolute value of the signal value of the composite band limited image signal. The conversion image signal may be obtained by performing conversion so as to obtain a value.
[0036]
Specifically, creation of the band limited image signal, creation of the converted image signal, creation of a signal related to the high frequency component, and addition of the signal related to the high frequency component to the original image signal are performed according to the following equations: May be.
[0037]
Sproc = Sorg + β (Sorg) · Fusm (Sorg, Sus1, Sus2, ... SusN)
Fusm (Sorg, Sus1, Sus2, ... SusN)
= [F1 {(Sorg-Sus1) + g (Sus1-Sus2)}
+ F2 {(Sus1-Sus2) + g (Sus2-Sus3)} + ...
+ Fk {(Susk-1−Susk) + g (Susk−Susk + 1)} +.
+ FN {(SusN-1−SusN) + g (SusN−SusN + 1)}]
(However, Sproc: Image signal in which high frequency components are emphasized.
Sorg: Original image signal
Susk (k = 1 to N + 1): unsharp mask image signal
fk (k = 1 to N): function for converting the composite band limited image signal
g: a function for generating an auxiliary image signal by converting a band-limited image signal
β (Sorg): enhancement coefficient determined based on the original image signal)
In addition, it is preferable that all the predetermined conversion processes are performed by converting the band-limited image signal in accordance with an imaging part when the original image is captured.
[0038]
An image processing apparatus according to the present invention is an image processing apparatus that emphasizes a high-frequency component of an original image by adding a signal related to the high-frequency component of the original image to an original image signal representing the original image.
Non-sharp mask image signal creating means for performing processing based on square checkered reduction processing and checkered square reduction processing on the original image signal to create a plurality of non-sharp mask image signals having different frequency response characteristics,
Bands for generating a plurality of band-limited image signals representing images of a plurality of frequency bands of the original image based on the original image signal and the plurality of unsharp mask image signals or the plurality of unsharp mask image signals A restricted image signal creating means;
Conversion means for performing a predetermined conversion process on at least one band-limited image signal among the plurality of band-limited image signals to create at least one converted image signal;
Frequency enhancement processing means for obtaining a signal related to the high-frequency component to be added to the original image signal based on the converted image signal.
[0039]
In the image processing apparatus according to the present invention, the predetermined conversion process performed by the conversion means may be the same process as the image processing method according to the present invention.
[0040]
In the image processing apparatus according to the present invention, the conversion unit converts the band limited image signal so that the band limited image signal is equal to or smaller than the absolute value determined based on the absolute value of the signal value of the band limited image signal. A suppressed image signal creating means for creating a suppressed image signal,
Based on the original image signal and the plurality of non-sharp mask image signals, or the plurality of non-sharp mask image signals, including a signal in a frequency band lower than the band-limited image signal used to create the suppression image signal An auxiliary image signal generating means for generating an auxiliary image signal;
The auxiliary image signal is converted so that the smaller the absolute value of the signal value of the auxiliary image signal is, the closer the value is to 1, and the larger the value is, the closer the value is to 0, thereby generating a magnification signal corresponding to each of the suppressed image signals. A magnification signal creating means;
Multiplying means for creating the converted image signal by multiplying the suppressed image signal by the magnification signal corresponding to the suppressed image signal may be provided.
[0041]
Further, in the image processing apparatus according to the present invention, the conversion means is a low frequency side band limited image which is a band limited image signal of a frequency band lower than the converted band limited image signal which is the at least one band limited image signal. The auxiliary image signal of the converted band-limited image signal is created by converting the signal based on a non-linear function that passes through the origin and has a substantially zero slope at the origin and gradually increases as the value to be processed increases. Auxiliary image signal creating means for
A composite band limited image signal creating means for creating a composite band limited image signal by adding the auxiliary image signal to the converted band limited image signal;
Converted image signal generating means for generating the converted image signal by converting the composite band limited image signal so as to have a value equal to or less than the absolute value determined based on the absolute value of the signal value of the composite band limited image signal It is good also as what comprises.
[0042]
The image processing method according to the present invention may be provided by being recorded on a computer-readable recording medium as a program for causing a computer to execute the image processing method.
[0043]
【The invention's effect】
According to the image processing method and apparatus of the present invention, a square checkered reduction process and a process based on the checkered square reduction process are performed on the original image signal to generate a plurality of non-sharp mask image signals having different frequency response characteristics. Then, a plurality of band limited image signals representing images of the plurality of frequency bands of the original image are generated from the unsharp mask image signal. Here, the band limited image signal has a small value in each frequency band in a so-called flat portion where the density change of the original image is relatively small. On the other hand, in the vicinity of the edge portion where the density rapidly changes, the size of the non-sharp mask is relatively large when the band-limited image signal is in a relatively low frequency band, that is, when the non-sharp mask image signal is obtained. In this case, since the edge portion of the pixel near the edge portion is included in the unsharp mask, the band-limited image signal is affected by the edge portion, and the absolute value of the signal value becomes relatively large. Thus, artifacts such as overshoot and undershoot occur in the edge portion of the image obtained by performing image processing, because the portion that is not originally an edge portion is affected by the density value of the edge portion.
[0044]
The present invention has been made in view of this point, and creates a converted image signal by performing a conversion process, such as reducing at least a part of the band limited image signal, on at least one of the band limited image signals. A signal relating to a high frequency component to be added to the original image signal is obtained based on the converted image signal. For this reason, the band-limited image signal having a relatively large absolute value of the signal value has a smaller influence on the signal regarding the high-frequency component to be added to the original image signal, and is substantially the same as the size of the non-sharp mask being reduced. Signal. As a result, since the influence of the signal causing the artifact is weakened even in the vicinity of the edge portion where the density rapidly changes, the image obtained by performing the processing can be made satisfactory without the artifact. .
[0045]
On the other hand, in the process described in Japanese Patent Laid-Open No. 10-75395, the pixel decimation when creating the unsharp mask image signal is halved, so that the band limited image signal is the original image signal. 1/2 of the Nyquist frequency^N~ 1/2^{N + 1}It represents an image in the frequency band. On the other hand, the band limited image signal obtained in the present invention is 1/2 of the Nyquist frequency of the original image signal.^N~ 1/2^kAnd 1/2^k~ 1/2^{N + 1}It represents an image in a frequency band of (N <k <N + 1). For example, if the band-limited image signal in the highest frequency band and the one-step lower frequency band than the highest frequency band is the first and second band-limited image signals, the second band-limited image signal is 1/2 of the original image.²The first band-limited image signal represents a size image, but the first band-limited image signal is 1/2 of the original image and the original image.²It represents an image of an intermediate size with the size image. Therefore, if the Nyquist frequency of the original image signal is fs / 2, the first band limited image signal is fs / 2 to fs / k, and the second band limited image signal is fs / k to fs / 4 (2 <k It represents an image in the frequency band <4). For this reason, according to the present invention, it is possible to obtain a band-limited image signal obtained by dividing the original image signal into finer frequency bands as compared with the conventional method.
[0046]
Therefore, by performing a predetermined conversion process on at least one band-limited image signal, the conversion process can be performed at finer frequency band intervals, so that the image quality can be controlled in more detail. Therefore, a processed image signal capable of reproducing a processed image with higher image quality can be obtained by obtaining a signal related to a high frequency component based on the obtained converted image signal and adding it to the original image signal.
[0047]
Further, the band limited image signal used for the processing described in the above-mentioned Japanese Patent Application Laid-Open No. 10-75395 is obtained by dividing a square pixel array original image signal into square bands to obtain a square array non-sharp mask image signal. Since the band-limited image signal is obtained from this non-sharp mask image signal, particularly when sharp edges are to be suppressed in a predetermined conversion process, the edges in the vertical and horizontal directions, that is, in the direction in which the pixels are arranged in a square shape are favorably edged. Can be suppressed. However, since the band-limited image signal is not band-divided in an oblique direction, that is, a direction in which pixels are arranged in a checkered pattern, the edge cannot be suppressed satisfactorily in the oblique direction. On the other hand, the band limited image signal obtained by the present invention is band-divided into squares and checkers, so that if the band-limited image signal of a checkered pixel array that is band-divided diagonally is used. Thus, the edge can be well suppressed in the oblique direction. Therefore, it is possible to prevent the edge from being excessively emphasized regardless of the direction of the edge.
[0048]
Furthermore, depending on the image acquisition device that acquires the original image signal, the sharpness of the original image may differ between the vertical and horizontal directions and the diagonal direction. For example, in a digital camera, an optical low-pass filter is used to remove noise at the time of shooting, but the reproduction image obtained by reproducing the image signal having the characteristics of the low-pass filter is vertically and horizontally. May differ in direction and diagonal direction. Here, if the characteristic of the image acquisition device is that the vertical and horizontal sharpness of the original image is higher than that in the oblique direction, the band limited image signal is vertically and horizontally in the method described in Japanese Patent Laid-Open No. 10-75395. Since only the band is divided, the sharpness cannot be increased only in the oblique direction. On the other hand, according to the present invention, since the band limited image signal is also band-divided in the diagonal direction, the sharpness can be increased only in the diagonal direction. Therefore, a processed image signal that can reproduce a high-quality image can be obtained regardless of the characteristics of the image acquisition device.
[0049]
As the predetermined conversion process, when the absolute value of the signal value of at least one band limited image signal is larger than a predetermined threshold, the absolute value of the signal value of the band limited image signal is reduced. In particular, the influence of large signal values can be weakened.
[0050]
In addition, if processing for reducing the absolute value of the signal value of the band limited image signal as the absolute value of the signal value of at least one band limited image signal is smaller than another threshold value smaller than a predetermined threshold value, The response of a component with a small absolute value of a signal value that can be regarded as noise can be reduced, thereby reducing the noise of the obtained image.
[0051]
In addition, by changing the magnitude of the absolute value of the signal value of the band limited image signal according to the frequency band of the band limited image signal, it is also possible to perform processing according to the frequency band.
[0052]
Also, regardless of whether it is larger than the predetermined threshold, if the conversion process is performed by a function that differs depending on the frequency band, it becomes possible to perform a more appropriate process. It is also possible to freely control the overall frequency response characteristics of the image signal. This has an effect that not only the above-mentioned artifact but also a step-like artifact generated at the boundary of the frequency band can be suppressed.
[0053]
Alternatively, the stepped artifact is generated by creating a converted image signal by adding or multiplying two types of signals using an image signal having a frequency band lower than that of the band-limited image signal to be subjected to predetermined conversion processing. Can be further suppressed, and a smoother image signal can be created to obtain a good processed image.
[0054]
Furthermore, by performing such conversion according to the imaging region when the original image is obtained, high-frequency components suitable for each imaging region can be emphasized.
[0055]
DETAILED DESCRIPTION OF THE INVENTION
Hereinafter, embodiments of the present invention will be described with reference to the drawings. FIG. 1 is a schematic block diagram showing the configuration of an image processing apparatus according to the first embodiment of the present invention. As shown in FIG. 1, the image processing apparatus according to the first embodiment has different frequency response characteristics by performing processing based on square checkered reduction processing and checkered square reduction processing on the input original image signal Sorg. A blur image signal creating unit 1 for creating a multi-resolution blur image signal (unsharp mask image signal) Susk (k = 1 to N), and a plurality of blur image signals Susk created by the blur image signal creating unit 1 Band-limited image signal generating means 2 for generating a band-limited image signal, and for at least one band-limited image signal among a plurality of band-limited image signals generated by the band-limited image signal generating means 2, Conversion means 3 that obtains a converted image signal by performing conversion processing so as to reduce at least a part thereof, and an integrated signal (high frequency component) by integrating the converted image signal And a frequency enhancement processing unit 5 that obtains a processed image signal Sproc in which the high frequency components of the original image are enhanced by multiplying the integration signal by the enhancement coefficient and adding it to the original image signal Sorg. With.
[0056]
First, processing performed in the blurred image signal creation unit 1 will be described. FIG. 2 is a schematic block diagram showing processing performed in the blurred image signal creation unit 1. Here, the original image represented by the original image signal Sorg is assumed to have pixels arranged in a square shape as shown in FIG. In FIG. 3 and subsequent figures, the position of ◯ represents the position of a pixel having a signal value.
[0057]
First, an interpolation process a which is a square checkered reduction process is performed on the original image signal Sorg. This interpolation process a is performed by filtering each pixel of the original image signal Sorg every other pixel with the filter shown in FIG. This filtering process is performed by shifting the pixels one pixel at a time in adjacent pixel columns. As a result, the filtering process is performed only on the pixel at the position ◯ shown in FIG. 5, and the reduced image signal S1 representing the reduced image in which the pixels are arranged in a checkered pattern having the signal value only at the position ◯ is obtained. Instead of the filtering process, a process of thinning out pixels every other pixel may be performed.
[0058]
Compared with the original image signal Sorg, the reduced image signal S1 has a checkered pixel arrangement and the number of pixels is approximately ½. Note that “substantially ½” means that the interpolation process “a” is a process that reduces the number of pixels to ½, but if the number of pixels in the pixel row is an odd number, the number of pixels is halved. This is because an extra one pixel is generated and it is not completely ½.
[0059]
Further, an interpolation process c, which is a checkered square reduction process, is performed on the reduced image signal S1 obtained by the interpolation process a. This interpolation process c is performed by filtering each pixel column of the reduced image signal S1 every other column using the filter shown in FIG. As a result, the filtering process is performed only on the pixel at the position ◯ shown in FIG. 7, and the reduced image signal S2 representing the reduced image in which the pixels are arranged in a square shape having the signal value only at the position ◯ is obtained. In place of the filtering process, a process of simply thinning out pixels may be performed.
[0060]
Compared with the reduced image signal S1, the reduced image signal S2 has a square pixel arrangement and the number of pixels is approximately ½. Further, the reduced image signal S2 has half the number of vertical and horizontal pixels, that is, the number of pixels is 1/4, compared to the original image signal Sorg. Therefore, the size of the image represented by the reduced image signal S2 is 1/2 of the original image.²It has become.
[0061]
Further, by performing an interpolation process a which is a square checkered reduction process on the reduced image signal S2 obtained by the interpolation process c, a filtering process is performed only on the pixel at the position “◯” shown in FIG. A reduced image signal S3 representing a reduced image having pixels only and having pixels arranged in a checkered pattern is obtained.
[0062]
By repeating the interpolation processes a and c, N reduced image signals Sk (k = 1 to N) are obtained. Here, when k is an odd number, the reduced image represented by the reduced image signal Sk has a checkered pixel arrangement. On the other hand, when k is an even number, the reduced image represented by the reduced image signal Sk has a square pixel arrangement. When the reduced image signal BN in the lowest frequency band is obtained up to the interpolation process c, the reduced image represented by the reduced image signal SN has a checkered pixel arrangement. On the other hand, when the reduced image signal SN in the lowest frequency band is obtained up to the interpolation process a, the reduced image represented by the reduced image signal SN has a square pixel arrangement.
[0063]
Further, the reduced image signal Sk having an even number k is ½ of the original image.^kIt represents a size image. On the other hand, a reduced image signal with an odd number k is 1/2 of the original image.^2kSize and 1/2^{2 (k + 1)}It represents an image of an intermediate size with respect to the size. Here, FIG. 9 shows frequency response characteristics of the reduced image signal Sk. As shown in FIG. 9, the response of the reduced image signal Sk is such that the higher the k, the higher the frequency component is removed (where k = 1 to 3 in FIG. 9). In FIG. 9, fs / 2 is the Nyquist frequency of the original image signal Sorg.
[0064]
Next, interpolation processes b and d are performed on the reduced image signal Sk obtained in this way. Here, the interpolation process b is a checkered pixel array reduced image signal Sk (k: odd number) obtained by the interpolation process a, while the interpolation process d is a square pixel array reduced image obtained by the interpolation process c. This is applied to the signal Sk (k: even number). First, the interpolation process d will be described for simplicity. This interpolation process d is a process for obtaining a blurred image signal Susk representing a blurred image having the same size as the original image by performing a filtering process by an interpolation filter that performs a B-spline interpolation operation on the reduced image signal Sk having a square pixel array. It is.
[0065]
Hereinafter, the B-spline interpolation calculation process will be described. The B-spline interpolation calculation is an interpolation calculation method for obtaining an interpolation image signal that can reproduce a smooth secondary image with relatively low sharpness. Here, for the sake of simplicity, a one-dimensional B-spline will be described. In this B-spline interpolation calculation, the first-order differential coefficient and the second-order differential coefficient (represented by f ″ (X)) are continuous between the sections instead of passing through the original sample points (pixels). It is necessary to do.
[0066]
That is,
f_k(X) = A_kx^Three+ B_kx²+ C_kx + D_k          (4)
(In formula (4), B_kIs a coefficient used for convenience and is different from an image obtained by filtering processing. ),
f_k′ (X_k) = F_k-1′ (X_k(5)
f_k′ (X_{k + 1}) = F_{k + 1}′ (X_{k + 1}(6)
f_k″ (X_k) = F_k-1″ (X_k(7)
f_k″ (X_{k + 1}) = F_{k + 1}″ (X_{k + 1}(8)
Is a condition. However, pixel X_kIs the first derivative of the pixel X_kX before and after_k-1And X_{k + 1}And these image signals Y_k-1, Y_{k + 1}Slope (Y_{k + 1}-Y_k-1) / (X_{k + 1}-X_k-1) Must satisfy the following formula (9).
[0067]
f_k′ (X_k) = (Y_{k + 1}-Y_k-1) / (X_{k + 1}-X_k-1(9)
Similarly, pixel X_{k + 1}Is the first derivative of the pixel X_{k + 1}X before and after_kAnd X_{k + 2}And these image signals Y_k, Y_{k + 2}Slope (Y_{k + 2}-Y_k) / (X_{k + 2}-X_k) Must satisfy the following formula (10).
[0068]
f_k′ (X_{k + 1}) = (Y_{k + 2}-Y_k) / (X_{k + 2}-X_k(10)
The function f (X) is generally approximated by the following equation (11).
[0069]
f (X) = f (0) + f ′ (0) X + {f ″ (0) / 2} X²          (11)
Where each section X_k-2~ X_k-1, X_k-1~ X_k, X_k~ X_{k + 1}, X_{k + 1}~ X_{k + 2}And the pixel X_kPixel X from_{k + 1}Interpolation point X in the direction_pWhere t (0 ≦ t ≦ 1)
f_k'(0) = C_k= (Y_{k + 1}-Y_k-1) / 2
f_k'(1) = 3A_k+ 2B_k+ C_k= (Y_{k + 2}-Y_k) / 2
f_k″ (0) = Y_{k + 1}-2Y_k+ Y_k-1= 2B
Therefore,
A_k= (Y_{k + 2}-3Y_{k + 1}+ 3Y_k-Y_k-1) / 6
B_k= (Y_{k + 1}-2Y_k+ Y_k-1) / 2
C_k= (Y_{k + 1}-Y_k-1) / 2
Where D_kIs unknown,
D_k= (D₁Y_{k + 2}+ D₂Y_{k + 1}+ D_ThreeY_k+ D_FourY_k-1) / 6
far. The spline interpolation function f_kSince (x) performs variable conversion of X = t as described above,
f_k(X) = f_k(T)
It becomes. Therefore,
f_k(T) = {(Y_{k + 2}-3Y_{k + 1}+ 3Y_k-Y_k-1) / 6} t^Three
+ {(Y_{k + 1}-2Y_k+ Y_k-1) / 2} t²
+ {(Y_{k + 1}-Y_k-1) / 2} t
+ (D₁Y_{k + 2}+ D₂Y_{k + 1}+ D_ThreeY_k+ D_FourY_k-1) / 6
And this is the image signal Y_k-1, Y_k, Y_{k + 1}, Y_{k + 2}Can be expressed by the following formula (12).
[0070]
f_k(T) = {(-t^Three+ 3t²-3t + D_Four) / 6} Y_k-1
+ {(3t^Three-6t²+ D_Three) / 6} Y_k
+ {(-3t^Three+ 3t²+ 3t + D₂) / 6} Y_{k + 1}
+ {(T^Three+ D₁) / 6} Y_{k + 2}
(12)
Here, if t = 1,
f_k(1) = {(D_Four-1) / 6} Y_k-1+ {(D_Three-3) / 6} Y_k
+ {(D₂+3) / 6} Y_{k + 1}+ {(D₁+1) / 6} Y_{k + 2}
Next section X_{k + 1}~ X_{k + 2}Equation (9) for
f_{k + 1}(T) = {(-t^Three+ 3t²-3t + D_Four) / 6} Y_k
+ {(3t^Three-6t²+ D_Three) / 6} Y_{k + 1}
+ {(-3t^Three+ 3t²+ 3t + D₂) / 6} Y_{k + 2}
+ {(T^Three+ D₁) / 6} Y_{k + 3}
(13)
Here, if t = 0,
f_{k + 1}(0) = (D_Four/ 6) Y_k+ (D_Three/ 6) Y_{k + 1}
+ (D₂/ 6) Y_{k + 2}+ (D₁/ 6) Y_{k + 3}
Condition of continuity (f_k(1) = f_{k + 1}(0)), and the condition that the coefficients corresponding to the filtered image signals are equal to each other, D_Four-1 = 0, D_Three-3 = D_Four, D₂+ 3 = D_Three, D₁+ 1 = D₂, D₁= 0, and therefore
D_k= (Y_{k + 1}+ 4Y_k+ Y_k-1) / 6
It becomes. Therefore,
Y_p= F_k(T) = {(-t^Three+ 3t²-3t + 1) / 6} Y_k-1
+ {(3t^Three-6t²+4) / 6} Y_k
+ {(-3t^Three+ 3t²+ 3t + 1) / 6} Y_{k + 1}
+ {T^Three/ 6} Y_{k + 2}                      (14)
Therefore, the filtered image signal Y_k-1, Y_k, Y_{k + 1}, Y_{k + 2}Interpolation coefficient b respectively corresponding to_k-1, B_k, B_{k + 1}, B_{k + 2}Is
b_k-1= (-T^Three+ 3t²-3t + 1) / 6
b_k  = (3t^Three-6t²+4) / 6
b_{k + 1}= (-3t^Three+ 3t²+ 3t + 1) / 6
b_{k + 2}= T^Three/ 6 (15)
Thus, it is obtained as a function of t.
[0071]
Then, a one-dimensional interpolation coefficient is obtained by the above formula (15), an interpolation filter is created by expanding the one-dimensional interpolation coefficient into two dimensions, and the reduced image signal Sk is filtered by this interpolation filter. Thus, the pixel value at the x position can be obtained, and a blurred image signal having the same size as the original image can be obtained.
[0072]
Here, the interpolation coefficient of the one-dimensional B-spline interpolation calculation for obtaining the interpolation signal value at the pixel position where the signal value is known is 1/6, 4/6, and 1/6 because t = 0. Therefore, when this is expanded two-dimensionally, it becomes an interpolation filter composed of a matrix of 3 × 3 interpolation coefficients as shown in FIG. On the other hand, since the interpolation coefficient of the one-dimensional B-spline interpolation calculation for obtaining the interpolation signal value at a position between pixel positions where the signal value is known is t = 1/2, 1/48, 47/48. , 47/48, 1/48, when this is expanded two-dimensionally, an interpolation filter composed of a matrix of 3 × 3 interpolation coefficients as shown in FIG. 11 is obtained.
[0073]
Therefore, the reduced image signal S2 shown in FIG. 7 is filtered by the interpolation filter shown in FIG. 10 at the position ◯ and by the interpolation filter shown in FIG. Both signal values can be obtained, and thereby a blurred image signal Sus2 representing a blurred image having the same number of pixels as the original image can be obtained. Note that when performing the filtering process using the interpolation filters shown in FIGS. 10 and 11, a value of 0 may be inserted between the interpolation coefficients so that the pixel having the signal value is located at a position corresponding to the interpolation coefficient. .
[0074]
On the other hand, the interpolation process b performs an interpolation process on the reduced image signal Sk (k: odd number) having a checkered pixel arrangement. Here, in the square pixel array, the pixels are evenly arranged at the smallest intervals in the vertical and horizontal directions, whereas the checkered pixel array is the pixel evenly arranged at the smallest intervals in the diagonal direction. Therefore, the interpolation filter for obtaining the interpolation signal value at the position ◯ in the interpolation process b is obtained by rotating the interpolation coefficient of the interpolation filter shown in FIG. 10 by 45 degrees as shown in FIG. Further, as shown in FIG. 13, the interpolation filter for obtaining the interpolation signal value at the x position may be obtained by rotating the interpolation coefficient of the interpolation filter shown in FIG. 11 by 45 degrees. When performing the filtering process using the interpolation filter shown in FIGS. 12 and 13, a value of 0 may be inserted between the interpolation coefficients so that the pixel having the signal value is located at the position corresponding to the interpolation coefficient. .
[0075]
As described above, by performing the filtering process by the interpolation filter for performing the B-spline interpolation calculation on each reduced image signal Sk, it is possible to obtain the blurred image signal Susk corresponding to each reduced image signal.
[0076]
FIG. 14 shows the frequency response characteristics of the blurred image signal Susk thus obtained. As shown in FIG. 14, the higher the k value of the blurred image signal Susk, the higher the component of the original image signal Sorg is removed.
[0077]
FIG. 15 is a block diagram showing in detail the configuration of the image processing apparatus according to the first embodiment of the present invention including the blurred image signal creating means 1 of FIG. As shown in FIG. 15, each blurred image signal Susk created by the blurred image signal creating unit 1 is then processed by the band limited image signal creating unit 2 and the converting unit 3. First, a band limited image signal Bk (k = 1 to N) is generated based on the original image signal Sorg and a plurality of blurred image signals Susk generated by the blurred image signal generating means 1, and this band limited image signal Bk. Is obtained by subtracting the blurred image signal Susk between adjacent frequency bands by the subtractor 21. That is, a plurality of band limited image signals Bk are obtained by sequentially calculating Sorg-Sus1, Sus1-Sus2,... SusN-1-SusN. FIG. 16 shows the frequency response characteristics of the band limited image signal Bk. As shown in FIG. 16, the band limited image signal Bk (here, k = 1 to 3) becomes a signal representing the low frequency band of the original image signal Sorg as the value of k increases.
[0078]
Next, the conversion means 3 converts the band limited image signal Bk thus obtained in accordance with the signal value of the band limited image signal Bk. This conversion is performed in the converter 22 using, for example, a function f as shown in FIG. This function f is the function f used in the description of the B-spline interpolation calculation._kDifferent from (X). This function f is a nonlinear function such that the slope is 1 when the absolute value of the signal value of the band limited image signal Bk is smaller than the threshold Th1, and the slope is smaller than 1 when the absolute value is larger than the threshold Th1. . This function f may be the same for all band limited image signals Bk, but may be different for each frequency band of the band limited image signal Bk. In FIG. 15, different functions fk (k = 1 to N) are used for each frequency band of the band limited image signal Bk.
[0079]
The converted signal Bk ′ obtained by converting the band-limited image signal Bk with such a function f is input to the arithmetic unit 23 that includes the integrating means 4 and the frequency enhancement processing means 5 described above. The calculator 23 performs the following processing. First, the converted image signal Bk ′ is integrated to obtain an integrated signal. When this integration signal is obtained, the frequency enhancement processing means 5 multiplies the enhancement coefficient β corresponding to the value of the original image signal Sorg, and further adds the integration signal multiplied by the enhancement coefficient β to the original image signal Sorg. Thus, a processed image signal Sproc is obtained.
[0080]
The processing performed in the above band-limited image signal creation means 2, conversion means 3, integration means 4 and frequency enhancement processing means 5 is shown in the following equation (16).
[0081]
Sproc = Sorg + β (Sorg) · Fusm (Sorg, Sus1, Sus2, ... SusN)
Fusm (Sorg, Sus1, Sus2, ... SusN)
= {F1 (Sorg-Sus1) + f2 (Sus1-Sus2) +
+ Fk (Susk-1-Susk) + ... + fN (SusN-1-SusN)} (16)
(However, Sproc: Image signal in which high frequency components are emphasized.
Sorg: Original image signal
Susk (k = 1 to N): Blur image signal
fk (k = 1 to N): function for converting each band limited image signal
β (Sorg): enhancement coefficient determined based on the original image signal)
The processed image signal Sproc thus obtained has a frequency response characteristic as shown in FIG. 18, for example. That is, the band-limited image signal Bk described above has a small absolute value of the signal value in each frequency band in a so-called flat portion where the density change of the original image is relatively small. On the other hand, in the vicinity of the edge where the density changes rapidly, when the band limited image signal Bk is in a relatively low frequency band, that is, when the mask size for obtaining the blurred image signal Susk is relatively large. As shown in FIG. 49, since the edge portion is included in the mask placed on the pixels near the edge portion, the band-limited image signal Bk is affected by the edge portion and the absolute value of the signal value is relatively low. It will be big. As described above, since the portion that is not the edge portion is affected by the density value of the edge portion, artifacts such as overshoot and undershoot occur in the edge portion of the image obtained by performing the image processing. .
[0082]
Therefore, when the absolute value of the signal value of the band limited image signal Bk is larger than the threshold Th1, the band limited image signal Bk is converted into the converted image signal Bk ′ so that the absolute value of the signal value is reduced by the function f described above. The converted image signal Bk ′ is integrated and further emphasized by a predetermined enhancement coefficient to obtain a signal relating to a high frequency component to be added to the original image signal Sorg.
[0083]
As a result, as shown in FIG. 18, the frequency response characteristic of the processed image signal Sproc is as shown by a solid line in the flat portion where the edge portion does not exist, but in the region near the edge portion, the processed image is processed. The signal Sproc has frequency response characteristics such that the response in a relatively low frequency band is lowered as shown by the broken line in FIG. This has the same effect as the mask for obtaining the blurred image signal (Sus in Expression (1)) smaller than the actual mask in the region near the edge portion.
[0084]
Therefore, the band limited image signal Bk having a relatively large absolute value of the signal value corresponding to the region near the edge portion has less influence on the signal related to the high frequency component to be added to the original image signal Sorg. For this reason, even in the vicinity of the edge portion where the density rapidly changes, the influence of the signal causing the artifact is weakened, and the image obtained by performing the processing can be a good image without the artifact. .
[0085]
FIG. 19 is a diagram for explaining the difference in frequency response characteristics between the band limited image signal used in the processing described in the above-mentioned Japanese Patent Application Laid-Open No. 10-75395 and the band limited image signal Bk obtained by the method of the present embodiment. (A) shows the frequency response characteristics of a band-limited image signal (hereinafter referred to as a conventional band-limited image signal) used for the processing described in JP-A-10-75395, and (b) shows the present embodiment. It is a figure which shows the frequency response characteristic of the band-limited image signal Bk obtained by the method of. In FIG. 19, fs / 2 is the Nyquist frequency of the original image signal Sorg having a square pixel array, and the solid line represents the frequency band division in which each band limited image signal can be reproduced, that is, the maximum boundary of the band limited image signal. Yes.
[0086]
The conventional band-limited image signal is half the original image represented by the original image signal Sorg.^2NTherefore, the frequency band of the original image signal S0 is divided into a rectangular shape, and the frequency response characteristic of each band-limited image signal is rectangular as shown in FIG. 19 (a). Become.
[0087]
On the other hand, the band limited image signal Bk obtained by this embodiment is 1/2 of the original image.^2N1/2 size with a detail image^2NSize and 1/2^{2 (N + 1)}It represents a detail image having a size intermediate to the size. For this reason, the frequency band of the original image signal S0 is alternately divided into a rectangular shape and a diamond shape, and the frequency response characteristics of the rectangular shape and the diamond shape are also alternately changed as shown in FIG. 19B. It will appear in
[0088]
Here, among the band limited image signals Bk obtained by the present embodiment, k is an even number and the maximum frequency band boundary of the band limited image signal Bk having a rectangular frequency response characteristic is the frequency of the conventional band limited image signal. It is the same as the maximum bandwidth boundary. On the other hand, the maximum boundary of the frequency band of the band limited image signal Bk having an odd-numbered and rhombic frequency response characteristic is located between the maximum boundaries of the conventional band limited image signal. Therefore, according to the present embodiment, it is possible to obtain a band limited image signal Bk obtained by dividing the original image signal S0 into finer frequency bands as compared with the conventional band limited image signal. Specifically, the conventional band-limited image signal is 1/2 of the Nyquist frequency fs / 2 of the original image signal Sorg.^N~ 1/2^{N + 1}In this embodiment, in the present embodiment, one frequency band of the conventional band-limited image signal is ½ of the Nyquist frequency fs / 2 of the original image signal Sorg.^N~ 1/2^MAnd 1/2^M~ 1/2^{N + 1}Band-limited image signals of two frequency bands (N <M <N + 1) are included.
[0089]
Therefore, the band-limited image signal Bk obtained according to the present embodiment can be subjected to the conversion process at finer frequency band intervals, so that the image quality can be controlled in more detail. As a result, the processed image signal Sproc obtained by adding the integrated image signal obtained by integrating the processed converted image signal Bk ′ to the original image signal Sorg is converted into a processed image with higher image quality. It can be reproducible.
[0090]
Further, the conventional band-limited image signal is obtained by dividing the original image signal S0 into a rectangular band, and particularly when it is desired to suppress steep edges using the function f as shown in FIG. With respect to the edge in the direction, that is, the direction in which the pixels are arranged in a square shape, the edge can be suppressed well. However, since the conventional band-limited image signal is not band-divided into rhombus shapes, the edge cannot be suppressed well in the oblique direction, and as a result, the edge is excessively emphasized in the oblique direction. . On the other hand, the band-limited image signal Bk obtained by the present embodiment is a band-limited image signal of a checkered arrangement in which the original image is band-divided alternately in a square shape and a checkered shape, and is band-divided in an oblique direction If is used, the edge can be well suppressed in the oblique direction. Therefore, it is possible to prevent the edge from being excessively emphasized regardless of the direction of the edge.
[0091]
In the present embodiment, the processed image signal Sproc is obtained by the above equation (16), but the processed image signal Sproc may be obtained by the following equation (17). In Expression (16) and Expression (17), when obtaining the band limited image signal Bk, subtraction is performed between adjacent frequency bands in Expression (16), but in Expression (17) The difference is that the subtraction processing is performed between the blurred image signal Susk in the frequency band and the original image signal Sorg. FIG. 20 shows the frequency response characteristics of the processed image signal Sproc obtained by Expression (17). As shown in FIG. 20, in the flat portion where the edge portion does not exist, the frequency response characteristic of the processed image signal Sproc is as shown by a solid line, but in the region near the edge portion, the processed image signal Sproc is As shown by the broken line in FIG. 20, the frequency response characteristic is such that the response in the relatively low frequency band is greatly reduced. This has the same effect as the mask for obtaining the blurred image signal Susk smaller than the actual mask in the region near the edge portion.
[0092]
Here, in comparison with FIG. 18, in the case of FIG. 20, the response is reduced over the entire frequency band in the region near the edge portion. For this reason, it is more preferable to obtain the processed image signal Sproc according to the equation (17) because the response of only the region near the edge portion is reduced without reducing the response of the flat portion.
[0093]
Sproc = Sorg + β (Sorg) · Fusm (Sorg, Sus1, Sus2, ... SusN)
Fusm (Sorg, Sus1, Sus2, ... SusN)
= (1 / N) · {f1 (Sorg−Sus1) + f2 (Sorg−Sus2) +
+ Fk (Sorg-Susk) + ... + fN (Sorg-SusN)}
(17)
(However, Sproc: Image signal in which high frequency components are emphasized.
Sorg: Original image signal
Susk (k = 1 to N): Blur image signal
fk (k = 1 to N): function for converting each band limited image signal
β (Sorg): enhancement coefficient determined based on the original image signal)
Furthermore, in the above-described embodiment, as shown in FIG. 17, when the absolute value of the signal value is larger than the threshold value Th1, the band limited image signal Bk is converted using a function that reduces the absolute value. However, as shown in FIG. 21, for example, when the absolute value of the signal value of the band limited image signal Bk is larger than the threshold value Th1, the absolute value is reduced and when the absolute value is smaller than the threshold value Th2, You may make it use the function which makes small.
[0094]
As described above, by converting the band limited image signal Bk so that the absolute value of the band limited image signal Bk is smaller as the absolute value of the band limited image signal Bk is smaller than the threshold Th2 smaller than the threshold Th1. Thus, the response of the component having a small absolute value of the signal value that can be regarded as noise in the image can be reduced, and thus the processed image signal Sproc representing the processed image with reduced noise can be obtained.
[0095]
Further, the processed image signal Sproc may be obtained by the following equation (18). In the equation (17), the blurred image signal Susk is subtracted from the original image signal Sorg in order to obtain the band limited image signal Bk. However, in the equation (18), the original image signal Sorg is used. Without this, the band-limited image signal Bk is created by subtracting the blurred image signal Susk (k = 2 to N) from the blurred image signal Sus1. FIG. 22 shows the frequency response characteristics of the band limited image signal Bk obtained in Expression (18). As shown in FIG. 22, the frequency response characteristic of the band limited image signal Bk obtained in the equation (18) does not use the original image signal Sorg, so that the frequency band corresponding to the highest frequency band is removed. It has become.
[0096]
Since the original image signal Sorg is used in the processing performed by Expression (17), a high-frequency component that can be regarded as noise in the image may be emphasized, so that noise is present in the obtained processed image. It may become noticeable. On the other hand, in the case of performing the processing according to the equation (18), the band limited image signal Bk is not generated using the original image signal Sorg, so the band limited image signal B1 in the highest frequency band that can be regarded as noise is removed. It will be. Therefore, even if the conversion process is performed, a component that can be regarded as noise is not emphasized, and a processed image signal Sproc that can reproduce a processed image with higher image quality can be obtained.
[0097]
Sproc = Sorg + β (Sorg) · Fusm (Sus1, Sus2, ... SusN)
Fusm (Sus1, Sus2, ... SusN)
= (1 / N) · {f2 (Sus1-Sus2) + f3 (Sus1-Sus3) +
+ Fk (Sus1-Susk) + ... + fN (Sus1-SusN)}
(18)
(However, Sproc: Image signal in which high frequency components are emphasized.
Sorg: Original image signal
Susk (k = 1 to N): Blur image signal
fk (k = 2 to N): function for converting each band limited image signal
β (Sorg): enhancement coefficient determined based on the original image signal)
Further, the processing may be performed by the following equation (19). In the above equation (16), the band limited image signal Bk created using the original image signal Sorg is used for the conversion process, whereas in the equation (19), the original image signal Sorg is used. The band limited image signal B1 (Sorg-Sus1) is not used for the conversion process. As a result, as shown in FIG. 22, the band-limited image signal B1 in the highest frequency band is removed, so that the high-frequency component of the processed image signal Sproc is removed as in the case of performing the processing according to the equation (18). Thus, noise is not emphasized, and a processed image signal Sproc that can reproduce a processed image with higher image quality can be obtained.
[0098]
Sproc = Sorg + β (Sorg) · Fusm (Sus1, Sus2, ... SusN)
Fusm (Sus1, Sus2, ... SusN)
= {F2 (Sus1-Sus2) + f3 (Sus2-Sus3) + ...
+ Fk (Susk-1−Susk) +... + FN (SusN−1−SusN)}
(19)
(However, Sproc: Image signal in which high frequency components are emphasized.
Sorg: Original image signal
Susk (k = 1 to N): Blur image signal
fk (k = 2 to N): function for converting each band limited image signal
β (Sorg): enhancement coefficient determined based on the original image signal)
When the original image is a medical radiographic image, the frequency band necessary for diagnosis differs depending on the imaging region. For example, in a double contrast image of the lung and stomach, the lung preferably emphasizes a relatively low frequency component, and the stomach preferably emphasizes a relatively high frequency component in order to observe folds on the stomach wall. On the other hand, in images containing metal such as bones and artificial bones, it is necessary to prevent artifacts due to these edge portions being too emphasized, but for images that do not contain edge portions such as bones such as mammography. As a result, artifacts are unlikely to occur, and if frequency components are emphasized in the same manner as an image including bones, there is also a problem that it becomes difficult to see parts necessary for observation. Therefore, it is desirable to change the function f for performing the conversion process in accordance with the imaging part when obtaining the original image or the frequency band of the band limited image signal Bk.
[0099]
For example, as shown in FIG. 23, in an image including a bone, it is preferable that the function f is set to B so that the high-frequency component of the band limited image signal Bk is largely suppressed so that artifacts are not easily generated at the edge portion. Conversely, in an image that does not include bones such as mammography, the function f is set to A to emphasize higher frequency components than B, and the absolute value of the signal value of the band limited image signal Bk is substantially over the entire frequency band. It is preferable to perform the conversion process so as to increase.
[0100]
Further, in the case of a chest image, it is preferable that the band limited image signal Bk is emphasized over the entire frequency band. Therefore, the function is changed as a function fk shown in FIG. 24 according to the frequency band of the band limited image signal Bk. Is preferred. On the other hand, in the case of a double contrast image of the stomach, it is preferable to emphasize the high-frequency component of the band-limited image signal Bk and suppress the low-frequency component. Therefore, the function shown in FIG. It is preferable to change it as fk. Here, “high” and “low” in the figure respectively indicate a function used for converting the band limited image signal Bk in the high frequency band and a function used for converting the band limited image signal Bk in the low frequency band. Yes.
[0101]
The frequency response characteristic of the processed image signal Sproc obtained by converting the band limited image signal Bk according to the function fk shown in FIG. 24 is converted in FIG. 26, and the band limited image signal Bk is converted in accordance with the function fk shown in FIG. FIG. 27 shows the frequency response characteristics of the processed image signal Sproc obtained by doing so. When the function fk shown in FIG. 24 is used, the response is emphasized over almost the entire frequency band as shown in FIG. 26, whereas when the function fk shown in FIG. 25 is used, the response in the high frequency band. It can be seen that is emphasized more than other frequency bands.
[0102]
As described above, in the case of the chest image, the lung field is more easily observed by performing the conversion process so that the response is emphasized over the entire frequency band. On the other hand, by performing a conversion process so that the response in the high frequency band is emphasized in the double contrast image of the stomach, the folds of the stomach wall can be more easily observed. In this manner, by performing processing so as to change the absolute value of the signal value of the band limited image signal Bk according to the frequency band of the band limited image signal Bk or according to the imaging region, the imaging region or frequency band is obtained. Accordingly, an image suitable for observation can be obtained.
[0103]
In this way, by making the function fk for converting the band limited image signal Bk different for each frequency band of the band limited image signal Bk, the frequency response characteristic of the processed image signal Sproc obtained becomes an arbitrary frequency response characteristic. Can be adjusted. Therefore, it is possible to adjust the frequency response characteristics of the processed image signal Sproc by setting an appropriate combination of functions fk according to the conditions required for the image to be processed.
[0104]
Hereinafter, effects obtained by making the function fk for converting the band limited image signal Bk different for each frequency band will be described in detail. FIG. 28 is a diagram illustrating a problem of a method of converting the band-limited image signal Bk of all frequency bands by the conversion process using the same function with all the functions fk being the same function, and the image density rapidly increases. The process in the vicinity of the changing edge part is shown in steps. (A) shows an original image signal Sorg whose signal value changes stepwise in the vicinity of the edge portion and a blurred image signal Susk created based on the original image signal Sorg, and (b) shows a band limitation corresponding to (a). The image signal Bk, (c) shows the converted image signal Bk ′, and (d) shows the integrated signal obtained by integrating the converted image signal Bk ′.
[0105]
As shown in FIG. 28 (d), the integrated signal obtained by integrating the converted image signal Bk ′ has an unnatural joint at the boundary of the frequency band, which causes streak-like artifacts. In order to prevent this, it is necessary to create the converted image signal Bk ′ in consideration that the boundary portion is connected as naturally as possible. However, when the function fk is uniquely determined, it is not possible to perform arbitrary conversion processing in consideration of the influence of the boundary on each band limited image signal Bk, and as a result, the occurrence of streak-like artifacts is generated. Could not be prevented.
[0106]
Therefore, the function fk used for the conversion process is different for each frequency band of the band-limited image signal Bk, and these functions fk are set in consideration of the boundary part of the frequency band, thereby generating streak-like artifacts. It is to prevent.
[0107]
Next, an example of the function fk that differs for each frequency band of the band limited image signal Bk will be described. FIG. 29 is a diagram illustrating an example of a function fk used for creating the converted image signal Bk ′. The horizontal axis represents the signal value of the band limited image signal Bk to be processed, and the vertical axis represents the converted image signal Bk ′. Each signal value is shown. These functions fk are used to convert the band limited image signal Bk so that the band limited image signal Bk is determined based on the absolute value of the signal value of the band limited image signal Bk so as to be equal to or less than the absolute value. The band limited image signal Bk whose absolute value is larger than a predetermined value is converted so that the value of the converted image signal Bk ′ becomes a substantially constant value, and the band limited image signal Bk converted by the function fk is converted. The higher the frequency band is, the smaller the predetermined value is.
[0108]
In other words, each of these functions fk passes through the origin, and the gradient of the function fk is 1 or less regardless of the signal value of the band limited image signal Bk converted by the function fk, and the band converted by the function fk. As the absolute value of the signal value of the limited image signal Bk increases, the band limit when the slope of the function fk becomes 0 or converges to 0 and the slope becomes 0 or a predetermined value close to 0. The absolute value of the signal value of the image signal Bk is smaller as the function fk for converting the high frequency band is smaller.
[0109]
The function fk shown in FIG. 29 performs conversion that suppresses the band-limited image signal Bk having a large amplitude, and the degree of suppression of the band-limited image signal Bk having a high frequency band is set to a band-limited image having a low frequency band. Although it is stronger than the signal Bk, this takes into account that the amplitude of the high frequency component included in the edge of the actual radiographic image is smaller than that of the low frequency component. That is, as shown in FIG. 30, in an actual radiographic image, even a fairly steep edge does not have an accurate step shape as shown in FIG. 30 (a), but as shown in FIG. 30 (b). The higher the frequency component, the smaller the amplitude. For this reason, it is desirable to suppress the band-limited image signal Bk having a higher frequency from a smaller amplitude in accordance with the amplitude of each frequency component, and this can be realized by the function fk shown in FIG.
[0110]
Next, the function shown in FIG. 31 will be described. The function fk shown in FIG. 31 converts the band limited image signal Bk so as to be a value equal to or smaller than the absolute value determined based on the absolute value of the signal value of the band limited image signal Bk. As the frequency band of the limited image signal Bk is lower, the converted image signal Bk ′ obtained by converting the band limited image signal Bk whose absolute value is within a predetermined range near 0 is converted. The absolute value is a small value.
[0111]
In other words, each of these functions fk passes through the origin, and the slope of the function fk is 1 or less regardless of the signal value of the band limited image signal Bk transformed by the function fk, and the slope of the function fk in the vicinity of 0 Is smaller as the function fk is for converting the band-limited image signal Bk in the low frequency band.
[0112]
The function fk shown in FIG. 31 is obtained by adding an integrated signal (FIG. 28 (d)) obtained by integrating the converted image signal Bk ′ to the original image signal Sorg. There is an effect of making the joint, that is, the rise of the signal more natural.
[0113]
Next, the function shown in FIG. 32 will be described. The function fk shown in FIG. 32 has characteristics of both the functions shown in FIG. 29 and FIG. 31, and the effect of both functions can be obtained.
[0114]
As described above, the frequency response characteristic of the processed image signal Sproc can be made arbitrary by changing the function fk for converting the band limited image signal Bk depending on the purpose for each frequency band. It becomes. FIG. 33 is a diagram showing an example of the effect. In FIG. 33, (a-1), (b-1), and (c-1) are converted image signals Bk ′ obtained by converting the band limited image signal Bk with different functions fk depending on the frequency band. The frequency response characteristics for each frequency band are shown, and (a-2), (b-2), and (c-2) are processes corresponding to (a-1), (b-1), and (c-1), respectively. The frequency response characteristic of the finished image signal Sproc is shown. (A-1) and (a-2) are set when the function of slope 1 is set in all frequency bands, (b-1) and (b-2) are set when the slope is set smaller in the lower frequency band, c-1) and (c-2) show frequency response characteristics when the slope of the function is set to 1 only in a specific frequency band and the slope of the function in other frequency bands is made smaller than 1, respectively.
[0115]
As shown in FIGS. 33 (a-1) and (a-2), when the function fk having the gradient 1 is used in all the frequency bands, the frequency response characteristics of the processed image signal Sproc over all the frequency bands. The response is almost constant. On the other hand, as shown in FIGS. 33 (b-1) and (b-2), when the slope of the function fk is set to be lower in the lower frequency band, the frequency response characteristic of the processed image signal Sproc is higher in the frequency band. The response can be increased. Further, as shown in FIGS. 33 (c-1) and (c-2), when the slope of the function fk is set to 1 only in a specific frequency band and the slope of the function fk in other frequency bands is made smaller than 1, processing is performed. The response of a specific frequency band in the completed image signal Sproc can be increased.
[0116]
Further, the conversion process performed by the conversion unit 3 may be different depending on the enhancement coefficient β. Hereinafter, the effect will be described using an example in which different functions fk are used depending on the frequency band of the band limited image signal Bk.
[0117]
Examples of such a function include a function fk as shown in FIG. 34 when the enhancement coefficient β is relatively small, and a function fk as shown in FIG. 35 when the enhancement coefficient β is large. Here, “high” and “low” in the figure respectively indicate a function used for conversion of the band limited image signal in the high frequency band and a function used for conversion of the band limited image signal in the low frequency band.
[0118]
Each of these functions fk suppresses the signal value of the band-limited image signal Bk so as to be smaller than the absolute value of the signal value. As can be seen from a comparison between FIGS. 34 and 35, the enhancement coefficient β Is larger, the difference between the degree of suppression of the signal value by the function fk for converting the band limited image signal Bk in the high frequency band and the degree of suppression of the signal value by the function fk for converting the band limited image signal Bk in the low frequency band The combination of the functions fk is defined so that becomes large. Specifically, the function fk for converting the band limited image signal Bk in the high frequency band is constant regardless of the enhancement coefficient β, and the degree of suppression of the function fk for converting the band limited image signal Bk in the low frequency band is set as the enhancement coefficient β. By increasing the value, the difference in the degree of suppression is widened. Note that the function fk for converting the band-limited image signal Bk in the high frequency band may be relaxed with the function fk for converting the band-limited image signal Bk in the low frequency band being constant, and all the functions fk may be suppressed. You may change so that the difference in a degree may spread.
[0119]
FIG. 36 is a diagram showing the frequency response characteristics of the processed image signal Sproc obtained by performing processing using the function fk shown in FIG. 34, and FIG. 37 is obtained by performing processing using the function fk shown in FIG. It is a figure which shows the frequency response characteristic of the obtained processed image signal Sproc. As shown in FIGS. 36 and 37, when the emphasis coefficient β is increased, the degree of emphasis of the high frequency component is increased, and the degree of emphasis of the low frequency component is not so different from that when the emphasis coefficient β is small. You can see that
[0120]
When determining the combination of the functions fk regardless of the enhancement coefficient β, increasing the enhancement coefficient β increases the degree of enhancement in all frequency bands. In general, the enhancement coefficient β is increased in order to enhance a high-frequency component with a small amount of information. However, in this case, even a low-frequency component is enhanced, and as a result, contrast is added to the processed image. This may cause artifacts. Therefore, as shown in FIG. 35, even if the enhancement coefficient β is increased, the degree of enhancement of the low-frequency band-limited image signal Bk is not changed so much that artifacts can be prevented.
[0121]
In the description of the function fk shown in FIG. 34 and FIG. 35, the combination of the function fk is shown one by one for the case where the enhancement coefficient β is small and large, but this is a combination of two functions fk. It does not mean that there is a thing, but is merely an example to show how the characteristics of the function fk should be changed according to the change of the enhancement coefficient β. Therefore, it is possible to classify the degree of change of the enhancement coefficient β into several levels, and determine the combination of the function fk according to each level. By setting such stages more finely, Needless to say, a higher quality image can be obtained.
[0122]
In the description of the function fk shown in FIGS. 34 and 35, the function fk is determined based on the interpretation that the enhancement coefficient β is increased because only the high-frequency component is desired to be enhanced. The purpose of changing the coefficient β is not limited to this, and various functions can be applied as the function fk according to the purpose.
[0123]
Furthermore, in the description of the function fk shown in FIGS. 34 and 35, the function fk is a set of a plurality of functions fk different for each frequency band. However, depending on the purpose to be achieved, the same function fk for all frequency bands. It is also possible to use.
[0124]
Next, a second embodiment of the present invention will be described. Note that the processing of the blurred image signal creation unit 1, the band limited image signal creation unit 2, the integration unit 4, and the frequency enhancement processing unit 5 in the second embodiment is the same as that in the first embodiment, and will be described here. Will be omitted and only the processing of the conversion means 3 will be described.
[0125]
FIG. 38 is a block diagram showing in detail the configuration of the image processing apparatus according to the second embodiment. As in the first embodiment, the band limited image signal Bk is created based on the original image signal Sorg and the plurality of blurred image signals Susk created by the blurred image signal creating means 1. This band limited image signal Bk is obtained by subtracting two blurred image signals Susk (however, Sorg and Sus1 for the original image signal Sorg) in the adjacent frequency bands by the subtractor 21. That is, by sequentially calculating Sorg-Sus1, Sus1-Sus2,... SusN-1-SusN, a plurality of band limited image signals Bk (k = 1 to N) are obtained. In the second embodiment, for example, the auxiliary image signal corresponding to the band limited image signal Sus1-Sus2 is Sus2-Sus3. Therefore, the means for creating the band limited image signal Bk and the means for creating the auxiliary image signal are substantially the same means. That is, the created band limited image signal Bk is processed as the band limited image signal Bk, and at the same time, is processed as an auxiliary image signal corresponding to the adjacent band limited image signal Bk-1.
[0126]
The band limited image signal Bk obtained as described above is converted by the conversion means 3. This conversion process is performed by the converter 22 and the converter 24 for each band limited image signal Bk. That is, in the converter 22, each band limited image signal Bk is subjected to conversion processing by a function fk that differs for each frequency band to obtain a suppressed image signal Rk (k = 1 to N). Is converted by the function g to obtain a magnification signal Mk (k = 1 to N). Then, the multiplier 25 multiplies the suppression image signal Rk and the magnification signal Mk to obtain a converted image signal Bk ′. In this case, as shown in FIG. 38, multiplication is performed, for example, as a suppression image signal R1 obtained by converting the band limited image signal Sus1-Sus2, and a magnification signal M1 obtained by converting the auxiliary image signal Sus2-Sus3. In addition, it is performed between signals in adjacent frequency bands. Here, the converter 22 that performs the conversion using the function fk corresponds to the suppression image signal generating unit, and the converter 24 that performs the conversion using the function g corresponds to the magnification signal generating unit, and the output signals of the converters 22 and 24 are the same. Multiplier 25 that multiplies is equivalent to multiplication means.
[0127]
The function fk is different for each frequency band, but may be the same function and can be arbitrarily set according to the purpose of image processing. In the second embodiment, as this function fk, for example, as shown in FIG. 29, the band limited image signal Bk is suppressed to be smaller than the absolute value of the signal value of the band limited image signal Bk. A different function fk is used for each frequency band of the image signal Bk.
[0128]
For example, the function g shown in FIG. 39 is used. In the function g shown in FIG. 39, when the auxiliary image signal is converted by this function g, when the absolute value of the signal value of the auxiliary image signal is small, a value close to 1 is obtained as the converted value, and the absolute value is large. It shows that a value closer to 0 is obtained. Note that K represents the minimum value among the values whose converted values are 0.
[0129]
The converted image signal Bk ′ obtained by the converting means 3 is input to the calculator 23. The calculator 23 includes the integrating means 4 and the frequency enhancement processing means 5. In the arithmetic unit 23, a plurality of converted image signals Bk ′ are integrated, and the integrated signal obtained by the integration is multiplied by an enhancement coefficient β determined according to the original image signal Sorg, and further multiplied by the enhancement coefficient β. The integrated signal thus obtained is added to the original image signal Sorg to obtain a processed image signal Sproc. When this process is expressed as an equation together with the other processes described above, the following expression (20) is obtained.
[0130]
Sproc = Sorg + β (Sorg) · Fusm (Sorg, Sus1, Sus2, ... SusN)
Fusm (Sorg, Sus1, Sus2, ... SusN)
= {F1 (Sorg−Sus1) · g (Sus1−Sus2)
+ F2 (Sus1-Sus2) · g (Sus2-Sus3) + ...
+ Fk (Susk-1−Susk) · g (Susk−Susk + 1) +
+ FN (SusN-1−SusN) · g (SusN−SusN + 1)} (20)
(However, Sproc: Image signal in which high frequency components are emphasized.
Sorg: Original image signal
Susk (k = 1 to N + 1): blurred image signal
fk (k = 1 to N): a function for creating a suppression signal by converting each band-limited image signal
g: a function for converting each auxiliary image signal to create a magnification signal
β (Sorg): enhancement coefficient determined based on the original image signal)
FIG. 40 is a diagram for explaining the effect when the signal value in the vicinity of the edge portion where the density of the image changes rapidly in the second embodiment. (1) shows an original image signal Sorg whose signal value changes stepwise in the vicinity of the edge portion and a blurred image signal Susk created based on the original image signal Sorg, and (2) shows a band limitation corresponding to (1). The image signal Bk is processed by the function g, (3) is an auxiliary image signal including a signal in a frequency band one lower than the band limited image signal Bk of (2), and (4) is the auxiliary image signal of (3). (5) shows the converted image signal Bk ′ obtained by multiplying the suppressed image signal Rk obtained by processing the signal (2) with the function fk by the magnification signal Mk (4). Is shown. The value K shown in (3) is the value K shown in FIG. 39, and when the auxiliary image signal in (3) becomes K, the magnification signal Mk in (4) becomes 0. Has been.
[0131]
As shown in FIG. 40, when the signal of (2) is simply converted so that the absolute value becomes small, the peak shape of the converted image signal becomes smooth, but the rising portion remains steep. On the other hand, the rising portion of the converted image signal Bk ′ shown in (5) is smooth. By smoothing the rising portion of the converted image signal Bk ′ in this way, it is possible to prevent artifacts that occur stepwise at the boundary of the frequency band of the integrated signal obtained by integrating the converted image signals Bk ′.
[0132]
Note that the auxiliary image signal processed by the function g does not have to be the same as the band limited image signal Bk as described above, and the difference value between the original image signal Sorg and the blurred image signal Sus (that is, Sorg−Susk). ) May be used as the auxiliary image signal. Processing using such an auxiliary image signal is shown in the following equation (21).
[0133]
Sproc = Sorg + β (Sorg) · Fusm (Sorg, Sus1, Sus2, ... SusN)
Fusm (Sorg, Sus1, Sus2, ... SusN)
= {F1 (Sorg-Sus1) · g (Sorg-Sus2)
+ F2 (Sus1-Sus2) g (Sorg-Sus3) + ...
+ Fk (Susk-1−Susk) · g (Sorg−Susk + 1) +
+ FN (SusN-1−SusN) · g (Sorg−SusN + 1)} (21)
(However, Sproc: Image signal in which high frequency components are emphasized.
Sorg: Original image signal
Susk (k = 1 to N + 1): blurred image signal
fk (k = 1 to N): a function for creating a suppression signal by converting each band-limited image signal
g: a function for converting each auxiliary image signal to create a magnification signal
β (Sorg): enhancement coefficient determined based on the original image signal)
FIG. 41 is a diagram for explaining the effect when the image signal in the vicinity of the edge portion is processed by applying the formula (21). Like FIG. 40, (1) is stepped in the vicinity of the edge portion. An original image signal Sorg whose signal value changes and a blurred image signal Susk created based on the original image signal Sorg, (2) a band limited image signal Bk corresponding to (1), and (3) (2 ) Shows the auxiliary image signal corresponding to the band-limited image signal Bk, (4) shows the magnification signal Mk obtained by processing the auxiliary image signal of (3) with the function g, and (5) shows the magnification signal Mk of (2). The converted image signal Bk ′ obtained by multiplying the suppressed image signal Rk obtained by processing the signal with the function fk by the magnification signal Mk of (4) is shown.
[0134]
As shown in FIG. 41 (5), when Sorg-Susk is used as a value processed by the function g, the signal value of the converted image signal Bk ′ is small for an edge with a large contrast, and the edge is converted for an edge with a small contrast. The signal value of the image signal Bk ′ has a magnitude close to that of the original band-limited image signal Bk.
[0135]
The converted image signal Bk ′ is integrated and then added to the original image signal Sorg. In this case, in the processed image signal Sproc, an edge having a large contrast is hardly emphasized, whereas an edge having a small contrast is relatively strongly emphasized compared to an edge having a large contrast.
[0136]
Although the two types of equations (20) and (21) have been described above, various changes can be made to the functions fk and g or the method of creating the band limited image signal Bk processed by the function g.
[0137]
Next, a third embodiment of the present invention will be described. However, the third embodiment is the same as the first embodiment because the processing of the blurred image signal creation unit 1, the band limited image signal creation unit 2, the integration unit 4, and the frequency enhancement processing unit 5 is the same as the first embodiment. Only the processing of the converting means 3 will be described.
[0138]
FIG. 42 is a block diagram showing in detail the configuration of the image processing apparatus according to the third embodiment of the present invention. Similar to the first and second embodiments, the band limited image signal Bk is created based on the original image signal Sorg and the plurality of blurred image signals Susk created by the blurred image signal creating means 1. This band limited image signal Bk is obtained by subtracting two blurred image signals Susk (however, Sorg and Sus1 for the original image signal Sorg) in the adjacent frequency bands by the subtractor 21. That is, a plurality of band limited image signals Bk are obtained by sequentially calculating Sorg-Sus1, Sus1-Sus2,... SusN-1-SusN.
[0139]
The band limited image signal Bk obtained as described above is converted by the conversion means 3. In this conversion process, as shown in FIG. 42, the band limited image signal of the frequency band to be processed is the converted band limited image signal Bk, and the frequency band is one lower than the frequency band of the converted band limited image signal Bk. Is converted by the converter 24 using the function g, and the auxiliary image signal Ak is obtained, and the auxiliary image signal Ak and the converted band limited image signal Bk are added to obtain the composite band limited image signal Ck. Further, the composite band limited image signal Ck is converted by the converter 22 using the function fk to obtain a converted image signal Bk ′. Here, the converter 24 that performs conversion using the function g corresponds to auxiliary image signal generation means, and the adder 26 corresponds to composite band limited image signal generation means.
[0140]
As the function g used in the third embodiment, for example, the function shown in FIG. 43 is used. The function g shown in FIG. 43 is a function that passes through the origin and has an inclination of about 0 at the origin, the inclination gradually increases as the signal value to be processed increases, and finally the inclination becomes about 1. That is, it is strongly suppressed when the signal value is small, and the degree of suppression is relaxed as the signal value increases. In the conversion process described above, the portion of the function g where the slope gradually increases from the origin affects the waveform of the rising portion of the auxiliary image signal Ak. That is, by performing the conversion process with this function g, a steep rising portion can be smoothed. Here, since the size of the actual band-limited image signal Bk is limited, the function g may be any function as long as the gradient gradually increases from 0 near the origin. 44 may be used. In the third embodiment, since the signal is not amplified, the gradient is set to 1 at maximum. However, the effect of the third embodiment is obtained by gradually increasing the gradient of the function near the origin. If this condition is satisfied, it is not always necessary to limit the inclination to 1 at the maximum.
[0141]
The functions fk may all be the same function or different functions, and can be arbitrarily set according to the purpose of image processing. In the third embodiment, for example, a function fk as shown in FIG. 29 is used as the function fk.
[0142]
The converted image signal Bk ′ obtained by the converting means 3 is input to the calculator 23. The calculator 23 includes the integrating means 4 and the frequency enhancement processing means 5. In the arithmetic unit 23, a plurality of converted image signals Bk ′ are integrated, and the integrated signal obtained by the integration is multiplied by an enhancement coefficient β determined according to the original image signal Sorg, and further multiplied by the enhancement coefficient β. The integrated signal thus obtained is added to the original image signal Sorg to obtain a processed image signal Sproc. When this process is expressed as an equation together with the other processes described above, the following expression (22) is obtained.
[0143]
Sproc = Sorg + β (Sorg) · Fusm (Sorg, Sus1, Sus2, ... SusN)
Fusm (Sorg, Sus1, Sus2, ... SusN)
= [F1 {(Sorg-Sus1) + g (Sus1-Sus2)}
+ F2 {(Sus1-Sus2) + g (Sus2-Sus3)} + ...
+ Fk {(Susk-1−Susk) + g (Susk−Susk + 1)} +.
+ FN {(SusN-1−SusN) + g (SusN−SusN + 1)}]
(22)
(However, Sproc: Image signal in which high frequency components are emphasized.
Sorg: Original image signal
Susk (k = 1 to N + 1): blurred image signal
fk (k = 1 to N): function for converting each composite band limited image signal
g: a function for converting each band-limited image signal to create an auxiliary image signal
β (Sorg): enhancement coefficient determined based on the original image signal)
FIG. 45 is a diagram for explaining an effect when a signal value in the vicinity of an edge portion in which the image density is rapidly changed in the third embodiment is processed. (1) is an original image signal Sorg whose signal value changes stepwise in the vicinity of the edge portion and a blurred image signal Susk created based on the original image signal Sorg. (2) and (3) are those of (1). The band-limited image signal Bk related to the signal, and (3) is the low-frequency side band-limited image signal when (2) is the converted band-limited image signal. (4) shows the auxiliary image signal Ak obtained by converting the low frequency side band limited image signal of (3) by the function g, and (5) shows the converted band limited image signal Bk of (2) and (4 ) Shows the composite band limited image signal Ck obtained by adding the auxiliary image signal Ak of (6), and (6) shows the converted image signal Bk obtained by converting the composite band limited image signal Ck of (5) by the function fk. ′ Is shown. In FIG. 45, when the converted band limited image signal Bk in (2) is simply converted so that the absolute value becomes small, the peak shape of the converted image signal becomes smooth, but the rising portion remains steep. Become. On the other hand, the rising portion of the converted image signal Bk ′ shown in (5) is smooth. By smoothing the rising portion of the converted image signal Bk ′ in this way, it is possible to prevent step-like artifacts that occur at the boundary of the frequency band of the integrated signal obtained by integrating the converted image signal Bk ′.
[0144]
The processing of Expression (22) has been described above, but various changes can be made to the functions fk and g in the third embodiment, for example.
[0145]
In the first embodiment, the effect obtained by performing different conversion processing for each imaging region on the band limited image signal Bk has been described. However, the idea of changing the processing for each imaging region is all the above. It can be applied to the embodiment.
[0146]
In the first to third embodiments, the blurred image signal Susk and the band limited image signal Bk are generated from the original image signal Sorg having a square pixel array to obtain the processed image signal Sproc. It is also possible to perform similar processing on the original image signal Sorg having a pixel array. In this case, as shown in FIG. 46, the blurred image signal generating means 1 performs an interpolation process c, which is a checkered square reduction process, on the original image signal Sorg to reduce the image representing the reduced image of the square pixel array. An image signal S1 is obtained, and the reduced image signal S2 is further subjected to interpolation processing a, which is a square checkered reduction process, to obtain a reduced image signal S2 representing a reduced image of a checkered pixel array.
[0147]
By repeating the interpolation processes c and a, N reduced image signals Sk (k = 1 to N) are obtained. Here, when k is an odd number, the reduced image represented by the reduced image signal Sk has a square pixel arrangement. On the other hand, when k is an even number, the reduced image represented by the reduced image signal Sk has a checkered pixel arrangement. When the reduced image signal BN in the lowest frequency band is obtained up to the interpolation process a, the reduced image represented by the reduced image signal SN has a square pixel arrangement. On the other hand, when the reduced image signal SN in the lowest frequency band is obtained up to the interpolation process c, the reduced image represented by the reduced image signal SN has a checkered pixel arrangement.
[0148]
Then, the reduced image signal Sk thus obtained is subjected to interpolation processing d (k: odd number) and interpolation processing b (k: even number) to represent a blurred image representing a blurred image having the same size as the original image. A signal Susk is obtained. Hereinafter, the processed image signal Sproc can be obtained by performing the same processing as in the above embodiments.
[0149]
In each of the above-described embodiments, the band-limited image signal Bk in all frequency bands is converted by the function fk. However, the conversion process is performed only on the band-limited image signal Bk in any one or more frequency bands. May be applied. In this case, only the converted image signal Bk ′ obtained by the conversion process is integrated and further added to the original image signal Sorg to obtain a processed image signal Sproc.
[Brief description of the drawings]
FIG. 1 is a schematic block diagram showing the configuration of an image processing apparatus according to an embodiment of the present invention.
FIG. 2 is a schematic block diagram showing processing performed in a blurred image signal creation unit
FIG. 3 is a diagram showing a pixel arrangement of an original image signal
FIG. 4 is a diagram showing a filter used in interpolation processing a.
FIG. 5 is a diagram showing a pixel array of reduced image signals obtained by interpolation processing a.
FIG. 6 is a diagram showing a filter used in the interpolation process c.
FIG. 7 is a diagram showing a pixel array of reduced image signals obtained by interpolation processing c.
FIG. 8 is a diagram showing a pixel arrangement of a reduced image signal obtained by further performing an interpolation process a on the reduced image signal obtained by the interpolation process c.
FIG. 9 is a diagram illustrating frequency response characteristics of a reduced image signal.
FIG. 10 is a diagram showing a filter used in interpolation processing d (part 1);
FIG. 11 is a diagram showing a filter used in the interpolation process d (part 2);
FIG. 12 is a diagram showing a filter used in the interpolation process b (part 1);
FIG. 13 is a diagram illustrating a filter used in the interpolation process b (part 2).
FIG. 14 is a diagram illustrating frequency response characteristics of a blurred image signal.
FIG. 15 is a block diagram showing in detail the configuration of the image processing apparatus according to the first embodiment of the present invention;
FIG. 16 is a diagram showing frequency response characteristics of a band-limited image signal.
FIG. 17 is a diagram illustrating an example of a function for converting a band-limited image signal in a conversion unit.
FIG. 18 is a diagram illustrating an example of frequency response characteristics of a processed image signal
FIG. 19 is a diagram for explaining a difference in frequency response characteristics between a band limited image signal obtained by a conventional method and a band limited image signal obtained by the present embodiment;
FIG. 20 is a diagram illustrating another example of frequency response characteristics of a processed image signal.
FIG. 21 is a diagram showing another example of a function for converting a band-limited image signal in the conversion means.
FIG. 22 is a diagram illustrating another example of the frequency response characteristics of the band-limited image signal.
FIG. 23 is a diagram illustrating an example of a function for converting a band-limited image signal according to an imaging region.
FIG. 24 is a diagram illustrating an example of a function for converting a band-limited image signal according to a frequency band (part 1);
FIG. 25 is a diagram illustrating an example of a function for converting a band-limited image signal according to a frequency band (part 2);
FIG. 26 is a diagram showing the frequency response characteristics of the processed image signal obtained by the function shown in FIG.
27 is a diagram showing frequency response characteristics of a processed image signal obtained by the function shown in FIG.
FIG. 28 is a diagram showing a problem when all band-limited image signals are converted by one type of function.
FIG. 29 is a diagram illustrating an example of a function for converting a band limited image signal according to a frequency band (part 3);
FIG. 30 is a diagram showing signals of an ideal edge and an actual edge.
FIG. 31 is a diagram illustrating an example of a function for converting a band limited image signal according to a frequency band (part 4);
FIG. 32 is a diagram showing an example of a function having the characteristics of the functions of FIG. 29 and FIG. 31;
FIG. 33 is a diagram showing frequency response characteristics for each frequency band and overall frequency response characteristics;
FIG. 34 is a diagram illustrating an example of a function for converting each band-limited image signal when the enhancement coefficient is small.
FIG. 35 is a diagram illustrating an example of a function for converting each band limited image signal when the enhancement coefficient is large.
FIG. 36 is a diagram showing frequency response characteristics of a processed image signal obtained by using the function shown in FIG. 34 for conversion;
FIG. 37 is a diagram showing frequency response characteristics of a processed image signal obtained by using the function shown in FIG. 35 for conversion;
FIG. 38 is a block diagram showing in detail the configuration of the image processing apparatus according to the second embodiment of the present invention.
FIG. 39 is a diagram showing an example of a function used in the magnification signal creating means
FIG. 40 is a diagram for explaining the effect of the second embodiment (part 1);
FIG. 41 is a diagram for explaining the effect of the second embodiment (part 2);
FIG. 42 is a block diagram showing in detail the configuration of an image processing apparatus according to a third embodiment of the present invention.
FIG. 43 is a diagram illustrating an example of a function for generating an auxiliary image signal by converting each band-limited image signal (part 1);
FIG. 44 is a diagram illustrating an example of a function for converting each band-limited image signal to create an auxiliary image signal (part 2);
FIG. 45 is a diagram for explaining the effect of the third embodiment;
FIG. 46 is a diagram for explaining creation of a blurred image signal from an original image signal having a checkered pixel array;
FIG. 47 is a diagram showing a square pixel array;
FIG. 48 is a diagram showing a checkered pixel array;
FIG. 49 is a diagram showing a blur image signal creation process in the vicinity of an edge portion;
[Explanation of symbols]
1 Blurred image signal creation means
2 Band-limited image signal creation means
3 Conversion means
4 Accumulation means
5 Frequency enhancement processing means
10 Filtering means
11 Interpolation calculation processing means
21 Subtractor
22 Converter
23 Calculator
24 Converter
25 Multiplication means
26 Adder

Claims (53)

In an image processing method for enhancing a high frequency component of an original image by adding a signal related to the high frequency component of the original image to an original image signal representing the original image,
A square checkered reduction process is performed on the original image signal to create a reduced image signal, a checkered square reduction process is performed on the reduced image signal to create a further reduced image signal, and the further reduced image signal A plurality of reduced image signals are created by alternately repeating a square checkered reduction process and the checkered square reduction process for the reduced image signal created thereby, and an interpolation process is performed on the reduced image signals Accordingly, to create a plurality of unsharp mask image signal frequency response characteristics are different from each other,
Based on the original image signal and the plurality of non-sharp mask image signals, or the plurality of non-sharp mask image signals, create a plurality of band limited image signals representing images of the plurality of frequency bands of the original image,
A predetermined conversion process is performed on at least one band-limited image signal among the plurality of band-limited image signals to create at least one converted image signal,
An image processing method comprising: obtaining a signal related to the high frequency component to be added to the original image signal based on the converted image signal.   The image processing method according to claim 1, wherein the predetermined conversion process is a process of reducing at least a part of a signal value of the band limited image signal.   3. The predetermined conversion process is a process of reducing an absolute value of the signal value when the absolute value of the signal value of the band-limited image signal is larger than a predetermined threshold value. Image processing method.   The predetermined conversion process is a process of further reducing the absolute value of the signal value as the absolute value of the signal value of the band limited image signal is smaller than another threshold value smaller than the predetermined threshold value. The image processing method according to claim 3.   5. The predetermined conversion process is a process of changing a magnitude of an absolute value of a signal value of the band limited image signal according to a frequency band of the band limited image signal. Image processing method.   The predetermined conversion processing is performed using the band-limited image signal equal to or less than the absolute value determined based on an absolute value of a signal value of the band-limited image signal based on a function that differs depending on a frequency band of the band-limited image signal. The image processing method according to claim 1, wherein the image processing method is a process of converting to a value of.   The function is a function for converting the band limited image signal so that the signal value of the converted image signal becomes a substantially constant value when the absolute value of the signal value of the band limited image signal is larger than a predetermined value. 7. The image processing method according to claim 6, wherein the predetermined value is smaller as the band-limited image signal is in a higher frequency band.   The absolute value of the converted image signal obtained by converting the signal value within a predetermined range in which the absolute value in the band limited image signal is near zero as the band limited image signal is in a lower frequency band. The image processing method according to claim 6 or 7, wherein the image processing method is a function for converting the band-limited image signal so that becomes a smaller value. The creation of the band limited image signal, the creation of the converted image signal, the creation of the signal related to the high frequency component, and the addition of the signal related to the high frequency component to the original image signal are expressed by the following equation: Sproc = Sorg + β (Sorg) · Fusm (Sorg, Sus1, Sus2, ... SusN)
Fusm (Sorg, Sus1, Sus2, ... SusN)
= {F1 (Sorg-Sus1) + f2 (Sus1-Sus2) +
+ Fk (Susk-1−Susk) +... + FN (SusN−1−SusN)}
(However, Sproc: Image signal in which high frequency components are emphasized.
Sorg: Original image signal
Susk (k = 1 to N): Unsharp mask image signal
fk (k = 1 to N): function for performing a predetermined conversion process
β (Sorg): enhancement coefficient determined based on the original image signal)
The image processing method according to claim 1, wherein the image processing method is performed according to claim 1. The creation of the band limited image signal, the creation of the converted image signal, the creation of the signal related to the high frequency component, and the addition of the signal related to the high frequency component to the original image signal are expressed by the following equation: Sproc = Sorg + β (Sorg) · Fusm (Sorg, Sus1, Sus2, ... SusN)
Fusm (Sorg, Sus1, Sus2, ... SusN)
= (1 / N) · {f1 (Sorg−Sus1) + f2 (Sorg−Sus2) +
+ Fk (Sorg-Susk) + ... + fN (Sorg-SusN)}
(However, Sproc: Image signal in which high frequency components are emphasized.
Sorg: Original image signal
Susk (k = 1 to N): Unsharp mask image signal
fk (k = 1 to N): function for performing a predetermined conversion process
β (Sorg): enhancement coefficient determined based on the original image signal)
The image processing method according to claim 1, wherein the image processing method is performed according to claim 1. The creation of the band limited image signal, the creation of the converted image signal, the creation of the signal related to the high frequency component, and the addition of the signal related to the high frequency component to the original image signal are expressed by the following equation: Sproc = Sorg + β (Sorg) · Fusm (Sus1, Sus2, ... SusN)
Fusm (Sus1, Sus2, ... SusN)
= {F2 (Sus1-Sus2) + f3 (Sus2-Sus3) + ...
+ Fk (Susk-1−Susk) +... + FN (SusN−1−SusN)}
(However, Sproc: Image signal in which high frequency components are emphasized.
Sorg: Original image signal
Susk (k = 1 to N): Unsharp mask image signal
fk (k = 1 to N): function for performing a predetermined conversion process
β (Sorg): enhancement coefficient determined based on the original image signal)
The image processing method according to claim 1, wherein the image processing method is performed according to claim 1. The creation of the band limited image signal, the creation of the converted image signal, the creation of the signal related to the high frequency component, and the addition of the signal related to the high frequency component to the original image signal are expressed by the following equation: Sproc = Sorg + β (Sorg) · Fusm (Sus1, Sus2, ... SusN)
Fusm (Sus1, Sus2, ... SusN)
= (1 / N). {F2 (Sus1-Sus2) + f3 (Sus1-Sus3) +
+ Fk (Sus1-Susk) + ... + fN (Sus1-SusN)}
(However, Sproc: Image signal in which high frequency components are emphasized.
Sorg: Original image signal
Susk (k = 1 to N): Unsharp mask image signal
fk (k = 1 to N): function for performing a predetermined conversion process
β (Sorg): enhancement coefficient determined based on the original image signal)
The image processing method according to claim 1, wherein the image processing method is performed according to claim 1.   The image processing method according to claim 9, wherein the predetermined conversion process is a conversion process corresponding to the enhancement coefficient. The conversion process according to the enhancement coefficient is a different conversion process depending on the frequency band of the band-limited image signal, and the lower the frequency band of the band-limited image signal, the lower the band-limited image signal. Is a process that greatly suppresses,
14. The image processing method according to claim 13, wherein the larger the enhancement coefficient is, the larger the difference between the degree of suppression of the band limited image signal in the high frequency band and the degree of suppression of the band limited image signal in the low frequency band is. . The predetermined conversion process creates the suppressed image signal by converting the band limited image signal so that the band limited image signal has a value equal to or smaller than the absolute value determined based on the absolute value of the signal value of the band limited image signal.
Based on the original image signal and the plurality of non-sharp mask image signals, or the plurality of non-sharp mask image signals, including a signal in a frequency band lower than the band-limited image signal used to create the suppression image signal Create an auxiliary image signal,
The auxiliary image signal is converted so that the smaller the absolute value of the signal value of the auxiliary image signal is, the closer the value is to 1, and the larger the value is, the closer the value is to 0, thereby creating a magnification signal corresponding to each of the suppressed image signals. ,
The image processing method according to claim 1, wherein the converted image signal is generated by multiplying the suppressed image signal by the magnification signal corresponding to the suppressed image signal. The creation of the band limited image signal, the creation of the converted image signal, the creation of the signal related to the high frequency component, and the addition of the signal related to the high frequency component to the original image signal are expressed by the following equation: Sproc = Sorg + β (Sorg) · Fusm (Sorg, Sus1, Sus2, ... SusN)
Fusm (Sorg, Sus1, Sus2, ... SusN)
= {F1 (Sorg−Sus1) · g (Sus1−Sus2)
+ F2 (Sus1-Sus2) · g (Sus2-Sus3) + ...
+ Fk (Susk-1−Susk) · g (Susk−Susk + 1) +
+ FN (SusN-1−SusN) · g (SusN−SusN + 1)}
(However, Sproc: Image signal in which high frequency components are emphasized.
Sorg: Original image signal
Susk (k = 1 to N + 1): unsharp mask image signal
fk (k = 1 to N): function for creating a suppression image signal
g: a function for converting the auxiliary image signal to create a magnification signal
β (Sorg): enhancement coefficient determined based on the original image signal)
The image processing method according to claim 15, wherein the image processing method is performed according to: The creation of the band limited image signal, the creation of the converted image signal, the creation of the signal related to the high frequency component, and the addition of the signal related to the high frequency component to the original image signal are expressed by the following equation: Sproc = Sorg + β (Sorg) · Fusm (Sorg, Sus1, Sus2, ... SusN)
Fusm (Sorg, Sus1, Sus2, ... SusN)
= {F1 (Sorg-Sus1) · g (Sorg-Sus2)
+ F2 (Sus1-Sus2) g (Sorg-Sus3) + ...
+ Fk (Susk-1−Susk) · g (Sorg−Susk + 1) +
+ FN (SusN-1−SusN) · g (Sorg−SusN + 1)}
(However, Sproc: Image signal in which high frequency components are emphasized.
Sorg: Original image signal
Susk (k = 1 to N + 1): unsharp mask image signal
fk (k = 1 to N): function for creating a suppression image signal
g: a function for converting the auxiliary image signal to create a magnification signal
β (Sorg): enhancement coefficient determined based on the original image signal)
The image processing method according to claim 15, wherein the image processing method is performed according to: The predetermined conversion process passes a low frequency side band limited image signal that is a band limited image signal of a frequency band lower than the converted band limited image signal that is the at least one band limited image signal, through the origin at the origin. Creating an auxiliary image signal of the converted band-limited image signal by transforming based on a non-linear function having a slope of approximately 0 and gradually increasing as the value to be processed increases;
Creating a composite band limited image signal by adding the auxiliary image signal to the converted band limited image signal;
The composite band limited image signal is a process of creating the converted image signal by converting the composite band limited image signal so as to be a value equal to or less than the absolute value determined based on the absolute value of the signal value of the composite band limited image signal. The image processing method according to claim 1, wherein the image processing method is an image processing method. The creation of the band limited image signal, the creation of the converted image signal, the creation of the signal related to the high frequency component, and the addition of the signal related to the high frequency component to the original image signal are expressed by the following equation: Sproc = Sorg + β (Sorg) · Fusm (Sorg, Sus1, Sus2, ... SusN)
Fusm (Sorg, Sus1, Sus2, ... SusN)
= [F1 {(Sorg-Sus1) + g (Sus1-Sus2)}
+ F2 {(Sus1-Sus2) + g (Sus2-Sus3)} + ...
+ Fk {(Susk-1−Susk) + g (Susk−Susk + 1)} +.
+ FN {(SusN-1−SusN) + g (SusN−SusN + 1)}]
(However, Sproc: Image signal in which high frequency components are emphasized.
Sorg: Original image signal
Susk (k = 1 to N + 1): unsharp mask image signal
fk (k = 1 to N): function for converting the composite band limited image signal
g: Function for creating the auxiliary image signal by converting the band-limited image signal
β (Sorg): enhancement coefficient determined based on the original image signal)
The image processing method according to claim 18, wherein the image processing method is performed according to:   The image according to any one of claims 1 to 19, wherein the predetermined conversion process is a process of converting the band-limited image signal in accordance with an imaging region when the original image is acquired. Processing method. In an image processing apparatus that emphasizes a high frequency component of an original image by adding a signal related to the high frequency component of the original image to an original image signal representing the original image,
A square checkered reduction process is performed on the original image signal to generate a reduced image signal, a checkered square reduction process is performed on the reduced image signal to generate a further reduced image signal, and the further reduced image signal is A plurality of reduced image signals are generated by alternately repeating a square checkered reduction process and the checkered square reduction process for the reduced image signal generated thereby, and an interpolation process is performed on the reduced image signals. By the non-sharp mask image signal creating means for creating a plurality of non-sharp mask image signals having different frequency response characteristics from each other,
Bands for generating a plurality of band-limited image signals representing images of a plurality of frequency bands of the original image based on the original image signal and the plurality of unsharp mask image signals or the plurality of unsharp mask image signals A restricted image signal creating means;
Conversion means for performing a predetermined conversion process on at least one band-limited image signal among the plurality of band-limited image signals to create at least one converted image signal;
An image processing apparatus comprising frequency enhancement processing means for obtaining a signal related to the high frequency component to be added to the original image signal based on the converted image signal.   The image processing apparatus according to claim 21, wherein the predetermined conversion process is a process of reducing at least a part of a signal value of the band limited image signal.   23. The process according to claim 21, wherein the predetermined conversion process is a process of reducing the absolute value of the signal value when the absolute value of the signal value of the band limited image signal is larger than a predetermined threshold value. Image processing apparatus.   The predetermined conversion process is a process of further reducing the absolute value of the signal value as the absolute value of the signal value of the band limited image signal is smaller than another threshold value smaller than the predetermined threshold value. The image processing apparatus according to claim 23.   25. The predetermined conversion process is a process of changing a magnitude of an absolute value of a signal value of the band limited image signal according to a frequency band of the band limited image signal. Image processing apparatus.   The predetermined conversion processing is performed using the band-limited image signal equal to or less than the absolute value determined based on an absolute value of a signal value of the band-limited image signal based on a function that differs depending on a frequency band of the band-limited image signal. 23. The image processing apparatus according to claim 21, wherein the image processing apparatus performs conversion so as to obtain a value of.   The function is a function for converting the band limited image signal so that the signal value of the converted image signal becomes a substantially constant value when the absolute value of the signal value of the band limited image signal is larger than a predetermined value. 27. The image processing apparatus according to claim 26, wherein the predetermined value is smaller as the band-limited image signal is in a higher frequency band.   The absolute value of the converted image signal obtained by converting the signal value within a predetermined range in which the absolute value in the band limited image signal is near zero as the band limited image signal is in a lower frequency band. 28. The image processing apparatus according to claim 26 or 27, wherein the image processing apparatus is a function for converting the band-limited image signal so that becomes a smaller value. The creation of the band limited image signal, the creation of the converted image signal, the creation of the signal related to the high frequency component, and the addition of the signal related to the high frequency component to the original image signal are expressed by the following equation: Sproc = Sorg + β (Sorg) · Fusm (Sorg, Sus1, Sus2, ... SusN)
Fusm (Sorg, Sus1, Sus2, ... SusN)
= {F1 (Sorg-Sus1) + f2 (Sus1-Sus2) +
+ Fk (Susk-1−Susk) +... + FN (SusN−1−SusN)}
(However, Sproc: Image signal in which high frequency components are emphasized.
Sorg: Original image signal
Susk (k = 1 to N): Unsharp mask image signal
fk (k = 1 to N): function for performing a predetermined conversion process
β (Sorg): enhancement coefficient determined based on the original image signal)
The image processing apparatus according to any one of claims 21 to 28, wherein The creation of the band limited image signal, the creation of the converted image signal, the creation of the signal related to the high frequency component, and the addition of the signal related to the high frequency component to the original image signal are expressed by the following equation: Sproc = Sorg + β (Sorg) · Fusm (Sorg, Sus1, Sus2, ... SusN)
Fusm (Sorg, Sus1, Sus2, ... SusN)
= (1 / N) · {f1 (Sorg−Sus1) + f2 (Sorg−Sus2) +
+ Fk (Sorg-Susk) + ... + fN (Sorg-SusN)}
(However, Sproc: Image signal in which high frequency components are emphasized.
Sorg: Original image signal
Susk (k = 1 to N): Unsharp mask image signal
fk (k = 1 to N): function for performing a predetermined conversion process
β (Sorg): enhancement coefficient determined based on the original image signal)
The image processing apparatus according to any one of claims 21 to 28, wherein The creation of the band limited image signal, the creation of the converted image signal, the creation of the signal related to the high frequency component, and the addition of the signal related to the high frequency component to the original image signal are expressed by the following equation: Sproc = Sorg + β (Sorg) · Fusm (Sus1, Sus2, ... SusN)
Fusm (Sus1, Sus2, ... SusN)
= {F2 (Sus1-Sus2) + f3 (Sus2-Sus3) + ...
+ Fk (Susk-1−Susk) +... + FN (SusN−1−SusN)}
(However, Sproc: Image signal in which high frequency components are emphasized.
Sorg: Original image signal
Susk (k = 1 to N): Unsharp mask image signal
fk (k = 1 to N): function for performing a predetermined conversion process
β (Sorg): enhancement coefficient determined based on the original image signal)
The image processing apparatus according to any one of claims 21 to 28, wherein The creation of the band limited image signal, the creation of the converted image signal, the creation of the signal related to the high frequency component, and the addition of the signal related to the high frequency component to the original image signal are expressed by the following equation: Sproc = Sorg + β (Sorg) · Fusm (Sus1, Sus2, ... SusN)
Fusm (Sus1, Sus2, ... SusN)
= (1 / N). {F2 (Sus1-Sus2) + f3 (Sus1-Sus3) +
+ Fk (Sus1-Susk) + ... + fN (Sus1-SusN)}
(However, Sproc: Image signal in which high frequency components are emphasized.
Sorg: Original image signal
Susk (k = 1 to N): Unsharp mask image signal
fk (k = 1 to N): function for performing a predetermined conversion process
β (Sorg): enhancement coefficient determined based on the original image signal)
The image processing apparatus according to any one of claims 21 to 28, wherein   33. The image processing apparatus according to claim 29, wherein the predetermined conversion process is a conversion process corresponding to the enhancement coefficient. The conversion process according to the enhancement coefficient is a different conversion process depending on the frequency band of the band-limited image signal, and the lower the frequency band of the band-limited image signal, the lower the band-limited image signal. Is a process that greatly suppresses,
34. The image processing apparatus according to claim 33, wherein the difference between the degree of suppression of the band limited image signal in the high frequency band and the degree of suppression of the band limited image signal in the low frequency band increases as the enhancement factor increases. . Suppressed image signal for generating a suppressed image signal by converting the band-limited image signal so that the conversion means has a value equal to or less than the absolute value determined based on the absolute value of the signal value of the band-limited image signal. Creating means;
Based on the original image signal and the plurality of non-sharp mask image signals, or the plurality of non-sharp mask image signals, including a signal in a frequency band lower than the band-limited image signal used to create the suppression image signal An auxiliary image signal generating means for generating an auxiliary image signal;
The auxiliary image signal is converted so that the smaller the absolute value of the signal value of the auxiliary image signal is, the closer the value is to 1, and the larger the value is, the closer the value is to 0, thereby generating a magnification signal corresponding to each of the suppressed image signals. A magnification signal creating means;
23. The image processing apparatus according to claim 21, further comprising multiplication means for creating the converted image signal by multiplying the suppression image signal by the magnification signal corresponding to the suppression image signal. The creation of the band limited image signal, the creation of the converted image signal, the creation of the signal related to the high frequency component, and the addition of the signal related to the high frequency component to the original image signal are expressed by the following equation: Sproc = Sorg + β (Sorg) · Fusm (Sorg, Sus1, Sus2, ... SusN)
Fusm (Sorg, Sus1, Sus2, ... SusN)
= {F1 (Sorg−Sus1) · g (Sus1−Sus2)
+ F2 (Sus1-Sus2) · g (Sus2-Sus3) + ...
+ Fk (Susk-1−Susk) · g (Susk−Susk + 1) +
+ FN (SusN-1−SusN) · g (SusN−SusN + 1)}
(However, Sproc: Image signal in which high frequency components are emphasized.
Sorg: Original image signal
Susk (k = 1 to N + 1): unsharp mask image signal
fk (k = 1 to N): function for creating a suppression image signal
g: a function for converting the auxiliary image signal to create a magnification signal
β (Sorg): enhancement coefficient determined based on the original image signal)
36. The image processing apparatus according to claim 35, wherein The creation of the band limited image signal, the creation of the converted image signal, the creation of the signal related to the high frequency component, and the addition of the signal related to the high frequency component to the original image signal are expressed by the following equation: Sproc = Sorg + β (Sorg) · Fusm (Sorg, Sus1, Sus2, ... SusN)
Fusm (Sorg, Sus1, Sus2, ... SusN)
= {F1 (Sorg-Sus1) · g (Sorg-Sus2)
+ F2 (Sus1-Sus2) g (Sorg-Sus3) + ...
+ Fk (Susk-1−Susk) · g (Sorg−Susk + 1) +
+ FN (SusN-1−SusN) · g (Sorg−SusN + 1)}
(However, Sproc: Image signal in which high frequency components are emphasized.
Sorg: Original image signal
Susk (k = 1 to N + 1): unsharp mask image signal
fk (k = 1 to N): function for creating a suppression image signal
g: a function for converting the auxiliary image signal to create a magnification signal
β (Sorg): enhancement coefficient determined based on the original image signal)
36. The image processing apparatus according to claim 35, wherein The converting means passes a low frequency side band limited image signal, which is a band limited image signal of a frequency band lower than the converted band limited image signal, which is the at least one band limited image signal, through the origin, and has a slope at the origin. Auxiliary image signal generating means for generating an auxiliary image signal of the converted band-limited image signal by performing conversion based on a non-linear function in which the gradient gradually increases as the value to be processed increases substantially at 0.
A composite band limited image signal creating means for creating a composite band limited image signal by adding the auxiliary image signal to the converted band limited image signal;
A converted image signal generating means for generating the converted image signal by converting the composite band limited image signal so as to have a value equal to or less than the absolute value determined based on the absolute value of the signal value of the composite band limited image signal. The image processing apparatus according to claim 21 or 22, further comprising: The creation of the band limited image signal, the creation of the converted image signal, the creation of the signal related to the high frequency component, and the addition of the signal related to the high frequency component to the original image signal are expressed by the following equation: Sproc = Sorg + β (Sorg) · Fusm (Sorg, Sus1, Sus2, ... SusN)
Fusm (Sorg, Sus1, Sus2, ... SusN)
= [F1 {(Sorg-Sus1) + g (Sus1-Sus2)}
+ F2 {(Sus1-Sus2) + g (Sus2-Sus3)} + ...
+ Fk {(Susk-1−Susk) + g (Susk−Susk + 1)} +.
+ FN {(SusN-1−SusN) + g (SusN−SusN + 1)}]
(However, Sproc: Image signal in which high frequency components are emphasized.
Sorg: Original image signal
Susk (k = 1 to N + 1): unsharp mask image signal
fk (k = 1 to N): function for converting the composite band limited image signal
g: Function for creating the auxiliary image signal by converting the band-limited image signal
β (Sorg): enhancement coefficient determined based on the original image signal)
The image processing apparatus according to claim 38, wherein   The image according to any one of claims 21 to 39, wherein the predetermined conversion process is a process of converting the band-limited image signal in accordance with an imaging part when the original image is acquired. Processing equipment. A computer-readable recording of a program for causing a computer to execute an image processing method for enhancing a high-frequency component of the original image by adding a signal related to the high-frequency component of the original image to an original image signal representing the original image In the recording medium,
The program performs a square checkered reduction process on the original image signal to create a reduced image signal, performs a checkered square reduction process on the reduced image signal to generate a further reduced image signal, and further reduces the reduced image signal. By alternately repeating the square checkered reduction process for the image signal and the checkered square reduction process for the reduced image signal created thereby, a plurality of reduced image signals are created and interpolated for the reduced image signals by performing processing, the procedure for creating a plurality of unsharp mask image signal frequency response characteristics are different from each other,
A procedure for creating a plurality of band limited image signals representing images of the plurality of frequency bands of the original image based on the original image signal and the plurality of unsharp mask image signals or the plurality of unsharp mask image signals. When,
A procedure of performing predetermined conversion processing on at least one band limited image signal among the plurality of band limited image signals to create at least one converted image signal;
A computer-readable recording medium comprising: a procedure for obtaining a signal related to the high-frequency component to be added to the original image signal based on the converted image signal.   42. The computer-readable recording medium according to claim 41, wherein the predetermined conversion process is a process of reducing at least a part of a signal value of the band limited image signal. A procedure for creating the band-limited image signal, a procedure for creating the converted image signal, a procedure for creating a signal for the high-frequency component, and a procedure for adding the signal for the high-frequency component to the original image signal are expressed by the following equation Sproc = Sorg + β (Sorg) · Fusm (Sorg, Sus1, Sus2, ... SusN)
Fusm (Sorg, Sus1, Sus2, ... SusN)
= {F1 (Sorg-Sus1) + f2 (Sus1-Sus2) +
+ Fk (Susk-1−Susk) +... + FN (SusN−1−SusN)}
(However, Sproc: Image signal in which high frequency components are emphasized.
Sorg: Original image signal
Susk (k = 1 to N): Unsharp mask image signal
fk (k = 1 to N): function for performing a predetermined conversion process
β (Sorg): enhancement coefficient determined based on the original image signal)
43. The computer-readable recording medium according to claim 41, wherein the recording medium is a computer-readable recording medium. A procedure for creating the band-limited image signal, a procedure for creating the converted image signal, a procedure for creating a signal for the high-frequency component, and a procedure for adding the signal for the high-frequency component to the original image signal are expressed by the following equation Sproc = Sorg + β (Sorg) · Fusm (Sorg, Sus1, Sus2, ... SusN)
Fusm (Sorg, Sus1, Sus2, ... SusN)
= (1 / N) · {f1 (Sorg−Sus1) + f2 (Sorg−Sus2) +
+ Fk (Sorg-Susk) + ... + fN (Sorg-SusN)}
(However, Sproc: Image signal in which high frequency components are emphasized.
Sorg: Original image signal
Susk (k = 1 to N): Unsharp mask image signal
fk (k = 1 to N): function for performing a predetermined conversion process
β (Sorg): enhancement coefficient determined based on the original image signal)
43. The computer-readable recording medium according to claim 41, wherein the recording medium is a computer-readable recording medium. A procedure for creating the band-limited image signal, a procedure for creating the converted image signal, a procedure for creating a signal for the high-frequency component, and a procedure for adding the signal for the high-frequency component to the original image signal are expressed by the following equation Sproc = Sorg + β (Sorg) · Fusm (Sus1, Sus2, ... SusN)
Fusm (Sus1, Sus2, ... SusN)
= {F2 (Sus1-Sus2) + f3 (Sus2-Sus3) + ...
+ Fk (Susk-1−Susk) +... + FN (SusN−1−SusN)}
(However, Sproc: Image signal in which high frequency components are emphasized.
Sorg: Original image signal
Susk (k = 1 to N): Unsharp mask image signal
fk (k = 1 to N): function for performing a predetermined conversion process
β (Sorg): enhancement coefficient determined based on the original image signal)
43. The computer-readable recording medium according to claim 41, wherein the recording medium is a computer-readable recording medium. A procedure for creating the band-limited image signal, a procedure for creating the converted image signal, a procedure for creating a signal for the high-frequency component, and a procedure for adding the signal for the high-frequency component to the original image signal are expressed by the following equation Sproc = Sorg + β (Sorg) · Fusm (Sus1, Sus2, ... SusN)
Fusm (Sus1, Sus2, ... SusN)
= (1 / N). {F2 (Sus1-Sus2) + f3 (Sus1-Sus3) +
+ Fk (Sus1-Susk) + ... + fN (Sus1-SusN)}
(However, Sproc: Image signal in which high frequency components are emphasized.
Sorg: Original image signal
Susk (k = 1 to N): Unsharp mask image signal
fk (k = 1 to N): function for performing a predetermined conversion process
β (Sorg): enhancement coefficient determined based on the original image signal)
43. The computer-readable recording medium according to claim 41, wherein the recording medium is a computer-readable recording medium.   47. The computer-readable recording medium according to claim 43, wherein the predetermined conversion process includes a procedure for performing a conversion process according to the enhancement coefficient. The predetermined conversion processing is a procedure for creating a suppressed image signal by converting the band limited image signal so that the band limited image signal has a value equal to or less than the absolute value determined based on the absolute value of the signal value of the band limited image signal. When,
Based on the original image signal and the plurality of non-sharp mask image signals, or the plurality of non-sharp mask image signals, including a signal in a frequency band lower than the band-limited image signal used to create the suppression image signal A procedure to create an auxiliary image signal;
The auxiliary image signal is converted so that the smaller the absolute value of the signal value of the auxiliary image signal is, the closer the value is to 1, and the larger the value is, the closer the value is to 0, thereby generating a magnification signal corresponding to each of the suppressed image signals. Procedure and
43. The computer-readable recording medium according to claim 41, further comprising a step of creating the converted image signal by multiplying the suppression image signal by the magnification signal corresponding to the suppression image signal. A procedure for creating the band-limited image signal, a procedure for creating the converted image signal, a procedure for creating a signal for the high-frequency component, and a procedure for adding the signal for the high-frequency component to the original image signal are expressed by the following equation Sproc = Sorg + β (Sorg) · Fusm (Sorg, Sus1, Sus2, ... SusN)
Fusm (Sorg, Sus1, Sus2, ... SusN)
= {F1 (Sorg−Sus1) · g (Sus1−Sus2)
+ F2 (Sus1-Sus2) · g (Sus2-Sus3) + ...
+ Fk (Susk-1−Susk) · g (Susk−Susk + 1) +
+ FN (SusN-1−SusN) · g (SusN−SusN + 1)}
(However, Sproc: Image signal in which high frequency components are emphasized.
Sorg: Original image signal
Susk (k = 1 to N + 1): unsharp mask image signal
fk (k = 1 to N): function for creating a suppression image signal
g: a function for converting the auxiliary image signal to create a magnification signal
β (Sorg): enhancement coefficient determined based on the original image signal)
49. The computer-readable recording medium according to claim 48, wherein the recording is performed according to the following. A procedure for creating the band-limited image signal, a procedure for creating the converted image signal, a procedure for creating a signal for the high-frequency component, and a procedure for adding the signal for the high-frequency component to the original image signal are expressed by the following equation Sproc = Sorg + β (Sorg) · Fusm (Sorg, Sus1, Sus2, ... SusN)
Fusm (Sorg, Sus1, Sus2, ... SusN)
= {F1 (Sorg-Sus1) · g (Sorg-Sus2)
+ F2 (Sus1-Sus2) g (Sorg-Sus3) + ...
+ Fk (Susk-1−Susk) · g (Sorg−Susk + 1) +
+ FN (SusN-1−SusN) · g (Sorg−SusN + 1)}
(However, Sproc: Image signal in which high frequency components are emphasized.
Sorg: Original image signal
Susk (k = 1 to N + 1): unsharp mask image signal
fk (k = 1 to N): function for creating a suppression image signal
g: a function for converting the auxiliary image signal to create a magnification signal
β (Sorg): enhancement coefficient determined based on the original image signal)
49. The computer-readable recording medium according to claim 48, wherein the recording is performed according to the following. The predetermined conversion process passes a low-frequency side band-limited image signal, which is a band-limited image signal of a frequency band lower than the converted band-limited image signal, which is the at least one band-limited image signal, through the origin at the origin. A procedure for creating an auxiliary image signal of the converted band-limited image signal by performing transformation based on a nonlinear function having a slope of approximately 0 and gradually increasing as the value to be processed increases;
Creating a composite band limited image signal by adding the auxiliary image signal to the converted band limited image signal;
And a step of creating the converted image signal by converting the composite band limited image signal so as to be a value equal to or less than the absolute value determined based on the absolute value of the signal value of the composite band limited image signal. 43. A computer-readable recording medium according to claim 41 or 42. A procedure for creating the band-limited image signal, a procedure for creating the converted image signal, a procedure for creating a signal for the high-frequency component, and a procedure for adding the signal for the high-frequency component to the original image signal are expressed by the following equation Sproc = Sorg + β (Sorg) · Fusm (Sorg, Sus1, Sus2, ... SusN)
Fusm (Sorg, Sus1, Sus2, ... SusN)
= [F1 {(Sorg-Sus1) + g (Sus1-Sus2)}
+ F2 {(Sus1-Sus2) + g (Sus2-Sus3)} + ...
+ Fk {(Susk-1−Susk) + g (Susk−Susk + 1)} +.
+ FN {(SusN-1−SusN) + g (SusN−SusN + 1)}]
(However, Sproc: Image signal in which high frequency components are emphasized.
Sorg: Original image signal
Susk (k = 1 to N + 1): unsharp mask image signal
fk (k = 1 to N): function for converting the composite band limited image signal
g: Function for creating the auxiliary image signal by converting the band-limited image signal
β (Sorg): enhancement coefficient determined based on the original image signal)
52. The computer-readable recording medium according to claim 51, wherein the recording is performed according to the following.   53. The computer according to claim 41, wherein the predetermined conversion process includes a procedure for converting the band-limited image signal in accordance with an imaging part when the original image is imaged. A readable recording medium.

Images