JP2019208708A - Image creation method and ocular fundus image processing device - Google Patents

Abstract

To provide an image creation method and an ocular fundus image processing device capable of accurately specifying a part corresponding to a macular area in an ocular fundus image.SOLUTION: An image creation method includes: a frequency dimension conversion step (S200) for subjecting images of each color component of a color ocular fundus image to frequency dimension conversion, and converting them to frequency component data composed of complex-number components for each color component; a frequency component modification step (S300) for modifying the frequency component data for each color component by setting a modification frequency range of component R narrower than modification frequency ranges of component G and component B and attenuating the frequency component with a predetermined frequency range as a modification frequency range for each color component in a low frequency region lower than a predetermined frequency for the complex-number components for each color component of the frequency component data; and a space dimension conversion step (S400) for subjecting the modified frequency component data for each color component to frequency dimension inverse conversion, and creating a correction color ocular fundus image composed of three color components of RGB.SELECTED DRAWING: Figure 6

Description

  The present disclosure relates to a fundus image processing technique, and more particularly, to a technique for identifying a predetermined target region in a fundus image and supporting diagnosis of a disease.

  Conventionally, various studies have been made on eyeball treatment. The macular region in the eyeball is the center of the optical sensor of the eyeball, and when the macular region is damaged, it sometimes leads to blindness. As popular diseases in the macular region, there are diabetic retinopathy (macular edema) and age-related macular degeneration. Along with the glaucoma of the optic nerve disease and the cataract of the lens disease, early detection is screamed in the ophthalmic field. However, contrast on the fundus image is low in contrast to the optic nerve region, and it is often difficult to identify the fundus of a healthy person with the naked eye, and it is difficult to find early macular lesions by fundus examination.

In the diagnosis of glaucoma using the fundus image, the macular position is also specified in accurately specifying the optic nerve region (see Patent Document 1). However, the technique of Patent Document 1 does not assume that the macular region is analyzed or diagnosed. Fortunately, the central fovea, which is the center of the macular region in the retina, has a peculiar dent. Therefore, it is relatively easy to identify the macular region by performing tomographic imaging of the retina by OCT. Therefore, in order to automatically diagnose diabetic retinopathy (macular edema), a method using both fundus image and OCT has also been proposed, and the macular region is identified and mapped on the fundus image based on the three-dimensional shape by OCT. A method has been proposed (see Patent Document 2). That is, since it is difficult to diagnose a macular region only by a fundus examination, a method for analyzing and diagnosing the macular region using OCT has become the current mainstream.

JP-A-9-313447 Japanese Patent No. 6025311

  However, although the technique described in Patent Document 2 exists, since OCT is expensive, there are few opportunities to be used in medical examinations and the field of view is narrower than fundus images. There is a problem that it is easy to miss.

  Therefore, an object of the present disclosure is to provide an image generation method and a fundus image processing apparatus that can accurately specify a portion corresponding to a macular region in a fundus image.

The present disclosure includes a plurality of means for solving the above problems.
An image creation method for creating a corrected color fundus image in which a macular region is emphasized based on a color fundus image composed of three color components of RGB,
A frequency dimension conversion step of performing frequency dimension conversion on an image for each color component of the color fundus image and converting the frequency component data to complex frequency components for each color component;
In a low frequency range lower than a predetermined frequency, the predetermined frequency range is set as a modified frequency range for each color component, the modified frequency range of the R component is set narrower than the modified frequency range of the G component or the B component, A frequency component modification step for attenuating the value of the modified frequency range in the frequency component data of
A spatial dimension conversion step of performing frequency dimension inverse transformation on the frequency component data for each modified color component to create a corrected color fundus image composed of three color components of RGB;
An image creating method is provided.

In addition, the image creation method of the present disclosure includes:
The frequency component modification step further sets a predetermined frequency range as a modified frequency range in a high frequency range higher than the predetermined frequency with respect to the complex number component of the frequency component data of the R component, and the frequency of the R component Of the component data, the value of the modified frequency range is attenuated.

In addition, the image creation method of the present disclosure includes:
The modified frequency range is defined by a sector region in which the two-dimensional Euclidean distance from the position of the DC component is a predetermined range, and the position of the DC component is the four corners of the frequency component data expressed by a square. It is characterized by being.

In addition, the image creation method of the present disclosure includes:
In the frequency component modification step, a process of increasing the frequency component data value for each color component at a predetermined magnification is further performed.

In addition, the image creation method of the present disclosure includes:
The spatial dimension conversion step further performs a process of subtracting a negative value from the maximum possible values for the values of the three color components of RGB when creating the corrected color fundus image, and performing negative inversion. .

In addition, the image creation method of the present disclosure includes:
After the spatial dimension conversion step,
The method further includes a specific color image extracting step of extracting the R component image of the corrected color fundus image as a specific color image in which a macular region is particularly emphasized.

In addition, the image creation method of the present disclosure includes:
For the specific color image,
Search for the maximum position that is the coordinates of the pixel with the maximum value,
Set the analysis range of a predetermined size around the searched maximum value position,
Set a threshold value obtained by multiplying the maximum value by a real value less than 1,
Count the number of pixels having a value equal to or greater than the threshold value in the analysis range,
A quantitative evaluation step of calculating a ratio of the counted number of pixels and the total number of pixels in the analysis range as a progress parameter;
Furthermore, it is characterized by having.

In addition, the image creation method of the present disclosure includes:
When the number of pixels of the color fundus image is Xs × Ys, an integer Ns of 2 n (n is a positive integer) that satisfies Ns / 2 <Ys ≦ Xs <Ns is defined,
An image size enlargement step of adding a dummy pixel to the color fundus image and converting it to an enlarged color fundus image of Ns × Ns pixels;
In the frequency dimension conversion step, a two-dimensional fast Fourier transform is performed on a color component image composed of Ns × Ns pixels for each color component of the enlarged color fundus image, and Ns × Ns pixels for each color component. It converts to frequency component data composed of complex number components,
The frequency component modification step modifies the frequency component data composed of complex components of the Ns × Ns pixels,
In the spatial dimension conversion step, two-dimensional fast Fourier inverse transform is performed for each color component of the modified frequency component data of the Ns × Ns pixels, and a corrected enlarged color fundus image of Ns × Ns pixels of three RGB components Then, the pixel corresponding to the dummy pixel is removed from the corrected enlarged color fundus image and converted to the corrected color fundus image of Xs × Ys pixels.

In addition, the image creation method of the present disclosure includes:
In the frequency component modification step, when Ns = 1024 (n = 10), the two-dimensional Euclidean distance is set in the range of 1 to 10 as the modified frequency range in the low frequency range. To do.

In the present disclosure,
A fundus image processing apparatus for creating a corrected color fundus image in which a macular region is emphasized based on a color fundus image composed of RGB three-color components,
Frequency dimensional conversion means for performing frequency dimensional conversion on each color component image of the color fundus image and converting the frequency component data to complex frequency components for each color component;
For the complex number component of each color component of the frequency component data,
In a low frequency range lower than a predetermined frequency, a predetermined frequency range is set as a modified frequency range for each color component, and the modified frequency range of the R component is set to be narrower than the modified frequency ranges of the G component and the B component and attenuated. The frequency component modification means for modifying the frequency component data for each color component,
A spatial dimension conversion means for performing frequency dimension inverse transformation on the frequency component data for each modified color component and creating a corrected color fundus image composed of three color components of RGB;
There is provided a fundus image processing apparatus characterized by comprising:

In the present disclosure,
Computer
A frequency dimension conversion step of performing frequency dimension conversion on the image of each color component of the color fundus image and converting the frequency component data to complex frequency components for each color component;
With respect to the complex number component of each color component of the frequency component data, in a low frequency range lower than a predetermined frequency, the predetermined frequency range for each color component is set as the modified frequency range, and the modified frequency range of the R component is set as the G component or B component. A frequency component modification step for modifying the frequency component data for each color component by setting the component to be narrower than the modified frequency range and attenuating it.
A spatial dimension conversion step of performing a frequency dimension inverse transform on the frequency component data for each modified color component to create a corrected color fundus image composed of three color components of RGB;
Provide a program to function as

In the present disclosure,
Computer
A frequency dimension conversion step of performing frequency dimension conversion on the image of each color component of the color fundus image and converting the frequency component data to complex frequency components for each color component;
With respect to the complex number component of each color component of the frequency component data, in a low frequency range lower than a predetermined frequency, the predetermined frequency range for each color component is set as the modified frequency range, and the modified frequency range of the R component is set as the G component or B component. A frequency component modification step for modifying the frequency component data for each color component by setting the component to be narrower than the modified frequency range and attenuating it.
A spatial dimension conversion step of performing a frequency dimension inverse transform on the frequency component data for each modified color component to create a corrected color fundus image composed of three color components of RGB;
A recording medium on which a program for causing the program to function is recorded.

  According to the present disclosure, it is possible to accurately specify a portion corresponding to a macular region in a fundus image.

It is a figure which shows distribution of the structure | tissue in the eyeball in each of a color fundus image and color space. It is a figure which shows distribution of the structure | tissue in the eyeball in each color space different from a fundus image and FIG. It is a figure which shows the relationship between a color fundus image, a tomographic structure, and a spatial frequency characteristic. 2 is a hardware configuration diagram of a fundus image processing apparatus according to an embodiment of the present disclosure. FIG. 2 is a functional block diagram illustrating a configuration of a fundus image processing apparatus according to an embodiment of the present disclosure. FIG. 6 is a flowchart showing an image creation method as well as an outline of processing of the fundus image processing apparatus of the present embodiment. It is a figure which shows the mode of the change of the image by the image preparation method of this embodiment, and the process of a fundus image processing apparatus. It is a figure which shows the image produced by the image production method which concerns on this embodiment, and a fundus image processing apparatus about a healthy person. It is a figure which shows the image produced by the image production method which concerns on this embodiment, and a fundus image processing apparatus about a patient with diabetic macular edema.

Hereinafter, preferred embodiments of the present disclosure will be described in detail with reference to the drawings.
<1. Basic concept of the present disclosure>
First, the basic concept of the present disclosure will be described. FIG. 1 is a diagram illustrating the distribution of tissues in the fundus in each of the color fundus image and the color space. FIG. 1A is a diagram showing the position of each tissue in a color fundus image. FIG. 1B is a diagram showing the two-dimensional color distribution of each tissue in the color fundus image in association with the R component and the G component. In FIG. 1, the disc region of the optic nerve head, the Cup region of the optic nerve head, the PPA (β-Peripillary Atrophy) region around the optic nerve head, the arterial blood vessel (Artery), A rough distribution of venous blood vessels (Vein), macular (Macular), and regions other than these six regions (Bacground) is shown. As shown in FIG. 1B, the difference between the Disc region and the Cup region from other regions is clarified by the R component, and the difference between the Disc region and the Cup region is also clarified by the G component. However, the macular region is difficult to distinguish from arterial blood vessels and venous blood vessels in the R component, and difficult to distinguish from arterial blood vessels and PPA in the G component.

  FIG. 2 is a diagram illustrating a fundus image and a distribution of tissues in the eyeball in different color spaces from those in FIG. FIG. 2 (a) is a diagram showing the position of each tissue in the fundus image, and is the same as FIG. 1 (a). FIG. 2B is a diagram showing the two-dimensional color distribution of each tissue in the color fundus image in association with the RB color difference component and the GB color difference component. The RB color difference component is a component obtained by subtracting the B component from the R component, and the GB color difference component is a component obtained by subtracting the B component from the G component. Also in FIG. 2, as in FIG. 1, the disc region of the optic disc, the Cup region of the optic disc, the PPA region around the optic disc, arterial blood vessels (Artery), venous blood vessels (Vein), macular (Macular), showing a rough distribution of regions (Bacground) other than these six regions. As shown in FIG. 2B, the macular region is difficult to distinguish from arterial blood vessels and venous blood vessels in both the RB component and the GB component.

  FIG. 3 is a diagram illustrating a relationship between a color fundus image, a tomographic structure, and a spatial frequency characteristic. 3A is a color fundus image of a healthy person, FIG. 3B is a macular cross section of the healthy person, and FIG. 3C is a frequency dimension converted to the color fundus image of FIG. 3A. The frequency component data is a spatial frequency characteristic plotted on a one-dimensional spectrum. 3 (d) is a color fundus image in the case of relatively advanced age-related macular degeneration, FIG. 3 (e) is a macular cross section in the case of age-related macular degeneration, and FIG. 3 (f) is FIG. It is a spatial frequency characteristic of the color fundus image of (d). 3C and 3F, the horizontal axis corresponds to the spatial frequency and is expressed in the same unit as the pixel, “0” is direct current, and “1” is the diameter of the circular fundus image (the diameter of the eyeball). Is equivalent to a spatial frequency having a period of “1” and “2” corresponds to a spatial frequency having a period of ½ of the same diameter.

  As can be seen by comparing FIG. 3B and FIG. 3E, when a macular lesion such as age-related macular degeneration progresses, fluid exudes from the new blood vessels and pushes up the macula. Thereby, the macular region appears enlarged in the fundus image. As can be seen by comparing FIG. 3C and FIG. 3F, in the case of age-related macular degeneration, the spatial frequency component of R becomes larger than the G component and the B component when the frequency value is 15 or more. However, in the case of relatively early age-related macular degeneration, the spatial frequency characteristics expressed by the fundus image corresponding to FIG. 3D and the one-dimensional spectrum corresponding to FIG. Is often difficult to recognize.

  As described above, as shown in FIG. 1B and FIG. 2B, the macular region is obtained only from color information such as R component, G component, RB color difference component, and GB color difference component, and the like. It is difficult to distinguish it from an area. However, as shown in FIGS. 3B and 3E, the three-dimensional tomographic image observed by OCT has a remarkable feature in the depth direction, and the macular region of a healthy person is centered on the fovea. Since there are concentric depressions, it is assumed that the luminance distribution reflected by the illumination light of the fundus camera has a spatial frequency distribution different from that of other retinal regions. Further, as shown in FIG. 3 (e), it is inferred that if the swell occurs due to macular edema, or if macular tissue changes or lacks due to macular degeneration, a significant change occurs in the spatial frequency distribution. However, in the one-dimensional spectrum shown in FIG. 3C and FIG. 3F, a low-frequency component having a frequency value of 5 or less (reflecting the three-dimensional spherical shape of the fundus) is prominent. Therefore, the frequency component of the macular region included in the frequency range of 5 to 15 or less is masked, and the feature of the macular region is difficult to identify. Therefore, in the present disclosure, the fundus image in which the low frequency component is cut and the macular region is emphasized is reconstructed using such a spatial frequency characteristic of the macular region.

<2. Configuration of fundus image processing apparatus>
FIG. 4 is a hardware configuration diagram of a fundus image processing apparatus according to an embodiment of the present disclosure. The fundus image processing apparatus 100 according to the present embodiment can be realized by a general-purpose computer. As shown in FIG. 41, a CPU (Central Processing Unit) 1 and a RAM (Random Access Memory) that is a main memory of the computer. 2, a large-capacity storage device 3 such as a hard disk or flash memory for storing programs and data executed by the CPU 1, an instruction input I / F (interface) 4 such as a keyboard and a mouse, and a data storage medium A data input / output I / F (interface) 5 for data communication with an external apparatus and a display unit 6 which is a display device such as a liquid crystal display are provided, and are connected to each other via a bus. The fundus image processing apparatus 100 according to the present embodiment also executes an image creation method according to an embodiment of the present disclosure.

  FIG. 5 is a functional block diagram showing a configuration of the fundus image processing apparatus according to the present embodiment. In FIG. 5, 10 is an image size enlarging means, 20 is a frequency dimension converting means, 30 is a frequency component modifying means, 40 is a spatial dimension converting means, 50 is a specific color image extracting means, 60 is a quantitative evaluation means, and 70 is a fundus image. Storage means 80 is a processing data storage means.

  When the number of pixels of the color fundus image is Xs × Ys, the image size enlarging means 10 defines an integer Ns of 2 n (n is a positive integer) that satisfies Ns / 2 <Ys ≦ Xs <Ns, This is means for adding a dummy pixel to the color fundus image and converting it to an enlarged color fundus image of Ns × Ns pixels. The frequency dimension conversion means 20 is a means for performing frequency dimension conversion on the image of each color component of the color fundus image and converting it into frequency component data composed of complex components for each color component. The frequency component modifying unit 30 sets a predetermined frequency range for each color component in a low frequency range lower than a predetermined frequency with respect to the complex number component of each color component of the frequency component data, and an R component modified frequency range. Is set to be narrower than the modified frequency range of the G component and the B component and attenuated, thereby modifying the frequency component data for each color component.

  The spatial dimension conversion means 40 is a means for performing a frequency dimension inverse conversion on the frequency component data for each modified color component to create a corrected color fundus image composed of three color components of RGB. The specific color image extracting unit 50 is a unit that extracts the R component image of the corrected color fundus image as a specific color image in which the macular region is particularly emphasized. The quantitative evaluation means 60 searches for the maximum value position which is the coordinates of the pixel having the maximum value with respect to the specific color image, sets an analysis range of a predetermined size around the searched maximum value position, and sets the maximum Set a threshold value by multiplying the value by a real value less than 1, count the number of pixels with a value greater than or equal to the threshold value in the analysis range, and calculate the ratio between the count value and the total number of pixels in the analysis range. It is a means for calculating as a progress parameter.

  The image size enlarging means 10, the frequency dimension converting means 20, the frequency component modifying means 30, the spatial dimension converting means 40, the specific color image extracting means 50, and the quantitative evaluation means 60 are programs that the CPU 1 stores in the storage device 3. It is realized by executing. The fundus image storage unit 70 is a storage unit that stores a color fundus image, which is a target for extracting a macular region, captured in full color using a visible light / light source type fundus camera, and is realized by the storage device 3. When the color fundus image is read by the fundus image processing apparatus and processing is performed as it is, the RAM 2 serves as the fundus image storage unit 70.

  The color fundus image is image data recorded with three color components of R (red), G (green), and B (blue), and is obtained by photographing the fundus of the subject. In the present embodiment, a full-color fundus image recorded with 8-bit 256 gradations for each RGB color is used. The processing data storage means 80 is a storage means for storing data such as an image created by the space dimension conversion means 40, the specific color image extraction means 50, etc., and quantitative evaluation calculated by the quantitative evaluation means 60, etc. 3 is realized.

  Each component means shown in FIG. 5 is actually realized by installing a dedicated program in hardware such as a computer and its peripheral devices as shown in FIG. That is, the computer executes the contents of each means according to a dedicated program. In this specification, the computer means a device having an arithmetic processing unit such as a CPU and capable of data processing, and is not only a general-purpose computer such as a personal computer but also a portable terminal such as a tablet terminal or a smartphone. Including.

  In the storage device 3 shown in FIG. 4, a dedicated program for causing a computer to execute an image creation method while operating the CPU 1 and causing the computer to function as a fundus image processing device is installed. When the dedicated program is executed by the computer, the CPU 1 causes the image size enlarging means 10, the frequency dimension converting means 20, the frequency component modifying means 30, the space dimension converting means 40, the specific color image extracting means 50, and the quantitative evaluation means 60. As a function. That is, the image creation method is realized by a computer and functions as a fundus image processing apparatus. Such a program can be distributed via a computer network, or can be distributed by being recorded on various recording media such as CD, DVD, Blu-ray. The storage device 3 not only functions as the fundus image storage unit 70 and the processing data storage unit 80 but also stores various data necessary for the image creation method and the processing as the fundus image processing device.

<3. Image creation method>
<3.1. Pretreatment>
First, a color fundus image to be processed is prepared. As a color fundus image, if there is an image file taken in full color by a digital fundus camera, it can be used as it is. In addition, if it is an old one recorded on a photographic medium by an analog fundus camera, the stored digital color negative / positive film, photographic paper, instant photo, etc. are read in full color with a scanner. Get color fundus image file. Generally, a color fundus image is obtained by photographing in full color using a visible light / light source type fundus camera. The acquired color fundus image is stored in the fundus image storage means 70 of the fundus image processing apparatus. In the present embodiment, a color image of R, G, B component 8-bit 256 gradation is prepared as a color fundus image.

<3.2. Process Overview>
Next, processing operations of the fundus image processing apparatus illustrated in FIGS. 4 and 5 will be described together with an image creation method according to an embodiment of the present disclosure. FIG. 6 is a flowchart showing an image creation method as well as an outline of the process of the fundus image processing apparatus of the present embodiment. As described above, the color fundus image to be processed is 8-bit 256-gradation image data for each RGB color. Therefore, a color fundus image having the number of pixels Xs in the x direction and the number of pixels Ys in the y direction is represented by Image (x, y) when variables c = 0 (Red), 1 (Green), and 2 (Blue) indicating color components are used. , C) = 0 to 255 (x = 0,..., Xs-1; y = 0,..., Ys-1; c = 0, 1, 2).

First, the image size enlarging means 10 enlarges the image size by adding dummy pixels to the color fundus image Image (x, y, c) (step S100). In step S100, which is an image size enlargement step, specifically, dummy pixels are added to the color fundus image Image (x, y, c), and the color fundus image Src (x, y, c) of Ns × Ns pixels. c). As described above, when the original color fundus image Image (x, y, c) is Xs × Ys pixels, Ns is an integer of 2 ⁿ satisfying Ns / 2 <Xs, Ys ≦ Ns ( ⁿ of 2 (Integer power). That is, Ns is a number equal to or greater than Xs, Ys, and is an integer of 2 ⁿ that is less than twice Xs, Ys. Here, n is a positive integer. Normally, the relationship between Xs and Ys is Ys ≦ Xs in this embodiment, but is not limited to this condition.

  The color fundus image Src (x, y, c) is also image data of 8 bits and 256 gradations for each color of RGB. Therefore, the color fundus image Src (x, y, c) having the number of pixels Ns in the x direction and the number of pixels Ns in the y direction has variables c = 0 (Red), 1 (Green), and 2 (Blue) indicating the color components. Then, it is defined as Image (x, y, c) = 0 to 255 (x = 0,..., Ns−1; y = 0,..., Ns−1; c = 0, 1, 2). Is done. The dummy pixel is added in order to perform high-speed processing on the CPU by applying a known fast Fourier transform algorithm (FFT) when performing frequency dimension conversion. In this embodiment, the pixel value of the dummy pixel is set to R = G = B = 0 representing black so that a bias based on the dummy pixel is not applied.

  FIG. 7 is a diagram illustrating a state of an image change caused by the image creation method and the fundus image processing apparatus according to the present embodiment. FIG. 7A is a diagram illustrating a relationship between image regions in the enlarged color fundus image Src (x, y, c). FIG. 7 shows a case where Xs = 700, Ys = 605, and Ns = 1024. In FIG. 7A, since the number of pixels Xs in the x direction of the original image area corresponding to the original color fundus image Image (x, y, c) is larger than the number of pixels Ys in the Y direction, It has a shape. In addition, since the enlarged image area that is the enlarged color fundus image Src (x, y, c) after adding the dummy pixels and the image size is enlarged is equal to the number of pixels Ns in the x direction and the number of pixels Ns in the Y direction, It has become a shape. Therefore, in the rectangular original image area shown in FIG. 7A, the fundus image is expressed as in the original Image (x, y, c) fundus image. In a portion other than the image area, the color is black (R = G = B = 0).

  Next, the frequency dimension conversion means 20 performs frequency dimension conversion on the enlarged color fundus image (step S200). In step S200, which is a frequency dimension conversion step, the frequency dimension conversion means 20 first normalizes each color value of each pixel. Specifically, a color value that takes an integer value of 0 to 255 is normalized to a real value of 0 to 1 by executing processing according to the following [Equation 1].

[Formula 1]
Src ′ (x, y, c) = Src (x, y, c) / 255

  As shown in the above [Equation 1], by dividing the value of each pixel of the enlarged color fundus image Src (x, y, c) by 255, an enlarged color fundus image Src ′ () in which the value of each pixel is normalized. x, y, c) is obtained.

  Subsequently, the frequency dimension conversion means 20 performs frequency dimension conversion on the normalized enlarged color fundus image Src ′ (x, y, c). As a specific method of frequency dimension conversion, various methods can be used as long as the method can convert an image into frequency-dimensional frequency component data. In this embodiment, two-dimensional Fourier transform is used. ing. Specifically, frequency component data composed of complex components is obtained by executing processing according to the following [Equation 2].

[Formula 2]
Fr (u, v, c) = Σ y = 0, Ns-1 Σ x = 0, Ns-1 Src' (x, y, c) · cos (2π (ux + vy) / Ns)
Fi (u, v, c) = Σ y = 0, Ns-1 Σ x = 0, Ns-1 Src' (x, y, c) · sin (2π (ux + vy) / Ns)

  [Expression 2] is an expression for realizing a known two-dimensional discrete Fourier transform. Since executing a double convolution operation based on [Equation 2] entails an enormous calculation load, the vertical and horizontal size Ns of the image is enlarged to a power of 2 as described above, and a known two-dimensional fast Fourier transform algorithm Can be applied (the specific description of the algorithm is omitted). As a result, when Ns = 1024, the normal CPU can execute the Fourier transform in almost real time in less than 1 second. In [Formula 2], Fr (u, v, c) is a real number component of the two-dimensional frequency component data, and Fi (u, v, c) is an imaginary number component of the two-dimensional frequency component data. U and v are horizontal and vertical spatial frequencies (u = 0,..., Ns-1; v = 0,..., Ns-1;) corresponding to x and y in the spatial dimension. The minimum unit is a spatial frequency with a period of Ns / 2, and the maximum frequency is Ns / 2 (both u and v are turned back to Ns / 2 or more, the component of u = v = Ns−1 is u = v = 0. DC component is the same as this component). c is a color component corresponding to c in the enlarged color fundus image Src (x, y, c), and c = 0, 1, and 2 correspond to (Red) (Green) (Blue), respectively. For the enlarged color fundus image Src ′ (x, y, c), processing according to [Equation 2] is executed for each color component, and frequency component data corresponding to each color component is obtained.

  FIG. 7B is a diagram showing a spectrum distribution in which frequency component data after frequency dimension conversion is plotted in two dimensions. As shown in FIG. 7B, the range of the frequency component data is apparently Ns × Ns, which is equal to the number of pixels of the enlarged color fundus image Src (x, y, c) after the image size enlargement. , Ns × Ns / 2 effective pixels, and pixels having the same pixel value or a pixel value obtained by inverting the sign are present in a point-symmetric manner (the original image is Ns × Ns pixels, and the pixel after Fourier transform is The combined real and imaginary numbers are Ns × Ns × 2 pixels, which apparently doubles, but the Ns × Ns pixels are copied point-symmetrically). That is, from the characteristics of the two-dimensional Fourier transform, a frequency distribution in which the center of the range of Ns × Ns is the maximum frequency Ns / 2-1 and the four corner frequencies are 0 (direct current) and the four regions are point-symmetric from the center. Will be shown.

  Next, the frequency component modification unit 30 modifies the frequency component with respect to the frequency component data (step S300). In step S300, which is a frequency component modification step, the frequency component modification means 30 first executes a process according to the following [Equation 3] to thereby obtain a two-dimensional Euclidean distance r (u, v from the origin of the DC component. ) Is calculated.

[Formula 3]
When u <Ns / 2, du = u, when u ≧ Ns / 2, du = Ns-1-u
When v <Ns / 2, dv = v, when v ≧ Ns / 2, dv = Ns-1-v
r (u, v) = (du ² + dv ² ) ^1/2

  As shown in FIG. 7B, du and dv calculated by the processing according to [Formula 3] have values of u and v of Ns / 2 or more (the right half and the lower half in FIG. 7B). Then, the values of du and dv change in the opposite direction to the values of u and v. The two-dimensional Euclidean distance r (u, v) calculated by the processing according to [Equation 3] starts from the four corner vertices of the Ns × Ns square shown in FIG. It is expressed as a distance from, and shows the geometric mean of spatial frequencies in the horizontal and vertical directions.

  Next, the frequency component modifying means 30 modifies the frequency components in the modified frequency range with the predetermined frequency range as the modified frequency range to be modified. Specifically, the frequency component in the modified frequency range is attenuated. The degree of attenuation can be set as appropriate, but in this embodiment, the attenuation is removed so that the numerical value becomes zero. The modified frequency range is set to a low frequency range lower than a predetermined frequency and a high frequency range higher than the predetermined frequency. In this case, when Ns = 1024, the predetermined frequency can be 10 (the unit of spatial frequency is expressed in cycles per m in the international unit system, but the unit of frequency is not shown in this specification). The low frequency region corresponds to a two-dimensional Euclidean distance r (u, v) of 10 or less. In the case of Ns = 1024 as the high frequency region, the threshold value is set to 15, and the two-dimensional Euclidean distance r (u, v) corresponds to 15 or more. The high frequency range is not a high frequency in the entire frequency range, but is referred to as a high frequency range in the sense that it is higher than a predetermined frequency. The threshold value expressing the low frequency region and the high frequency region also changes in proportion to the value of Ns.

  The modified frequency range is set for all color components of R, G, and B in the low frequency range. However, the modified frequency range is set to only the R component for the high frequency range. Therefore, the R component is modified at two locations, the modified frequency range in the low frequency region and the modified frequency range in the high frequency region, but the G component and the B component are at one location in the modified frequency range in the low frequency region. Only the modification is done. The modified frequency range can be set as appropriate. When expressed in correspondence with the two-dimensional Euclidean distance r (u, v), in the low frequency range, the modified frequency range is from the lower limit r1 (c) to the upper limit r2 (c). (C) indicates that a lower limit and an upper limit are set for each of R, G, and B colors corresponding to c = 0, 1, and 2, respectively. In the high frequency range, the modified frequency range of the R component is from the lower limit rh1 to the upper limit rh2.

  FIG. 7C shows a modified frequency range of the frequency component data of the R component. FIG. 7C has the same axial direction and size as FIG. 7B, and du and dv in FIG. In FIG. 7C, although r1 (0) is not specified, r1 (0) = 1, and in the low frequency region, the two-dimensional Euclidean distance r (u, v) from the four corners is r1. The range is from (0) to r2 (0), and is a shaded range of fan-shaped regions centered on the four corners. Note that the DC components at the four corners where the two-dimensional Euclidean distance r (u, v) = 0 are not modified. The high frequency range is a range where the two-dimensional Euclidean distance r (u, v) from the four corners is rh1 or more and rh2 or less, and rh2 is set to Ns / 2-1 which is the maximum frequency. Specifically, all the frequency components of rh1 or higher are to be modified, and the shaded range reaches the center of the frequency component data. Therefore, the frequency component modification means 30 modifies the frequency component in the modified frequency range by executing the processing according to the following [Equation 4].

[Formula 4]
For c = 0, 1, 2
In r1 (c) ≦ r (u, v) ≦ r2 (c) and rh1 ≦ r (u, v) ≦ rh2,
Fr (u, v, c) = 0
Fi (u, v, c) = 0

  In the example of [Equation 4], both the real component Fr (u, v, c) and the imaginary component Fi (u, v, c) are removed as “0”, but even if they are not removed, they are greatly attenuated. That's fine. For the R component, when processing according to [Formula 4] is executed to remove the frequency component in the modified frequency range, only the component sandwiched between the low frequency region and the high frequency region remains. Therefore, in the example of FIG. 7C, only the frequency components that are not shaded and whose two-dimensional Euclidean distance r (u, v) is in the range of r2 (0) +1 to rh1-1 remain. Become.

  Specific values of the modified frequency range in the low frequency range and the high frequency range can be set as appropriate. In this embodiment, since Ns = 1024, the lower limit of the modified frequency range in the low frequency range is set to r1 (0) = r1 (1) = r1 (2) = 1 for each color component, and in the low frequency range. Regarding the upper limit of the modified frequency range, r2 (0) = 2 for R, r2 (1) = 5 for G, and r2 (2) = 3 for B. Therefore, in the low frequency range, the G component is removed in the widest range and the R component is removed in the narrowest range. The reason why the G component is removed in the widest range is that the G component contains a large amount of various fundus tissues such as blood vessels, and therefore tissues other than the macula, such as the optic disc included in the low frequency range, are removed. However, when observing in full color, it is easier to understand the positional relationship of the macular region when the blood vessel is displayed, so a high frequency region containing a lot of blood vessel components remains. The reason why the R component is removed in the narrowest range is to quantitatively analyze the macular region. In this embodiment, the lower limit of the modified frequency range in the high frequency region of the R component is rh1 = 15, and the upper limit of the modified frequency range in the high frequency region of the R component is rh2 = 511. For the R component, the modified frequency range in the low frequency region is made the narrowest compared to other color components, and the modified frequency range in the high frequency region is also modified. As a result, it is possible to remove a high frequency region containing many blood vessel components that hinder quantitative analysis, leave a low frequency region containing many macular regions, and quantitatively analyze the macular region using the R component. it can. In proportion to the value of Ns, the upper and lower limits of the modified frequency range corresponding to the two-dimensional Euclidean distance r (u, v) also change.

  As described above, when the process of removing the frequency components in the low frequency region and the high frequency region is performed, the luminance of the corrected color fundus image generated by performing spatial dimension conversion described later is significantly reduced. In particular, since the frequency components in the low frequency range are remarkably large in all the color components, if these are removed, the luminance of the corrected color fundus image is significantly reduced in all the color components. Therefore, the frequency component modifying means 30 further sets a magnification s (c) larger than 1 for each color component, and executes processing according to the following [Equation 5] for the frequency component data after the frequency component modification. By doing so, the frequency component is increased.

[Formula 5]
Fr ′ (u, v, c) = Fr (u, v, c) · s (c)
Fi ′ (u, v, c) = Fi (u, v, c) · s (c)

  The processing according to [Formula 5] is performed for all frequencies corresponding to all u and v. As the magnification s (c), different values can be set for the respective colors. The magnification s (c) needs to be larger than 1 although it is influenced by the photographing conditions. In the present embodiment, s (0) = s (1) = s (2) = 4.0. In this way, by increasing the values of all the frequency components of all the color components, the brightness of the corrected color fundus image generated by performing the spatial dimension conversion described later is brightened, thereby preventing a decrease in visibility. be able to.

  Next, the space dimension conversion means 40 performs space dimension conversion on the modified frequency component data (step S400). In step S400 which is a space dimension conversion step, space dimension conversion which is frequency dimension inverse conversion is performed. The spatial dimension transformation is a transformation that returns the frequency component data obtained by transforming the spatial dimension image to the frequency dimension to the original spatial dimension, and is therefore a frequency dimension inverse transformation that is an inverse transformation of the frequency dimension transformation. Therefore, as a specific method of spatial dimension conversion, any method may be used originally, but it is necessary to match the frequency dimension conversion method. In this embodiment, since the two-dimensional Fourier transform is used as the frequency dimension conversion, the two-dimensional inverse Fourier transform is used as the spatial dimension conversion.

  The spatial dimension conversion means 40 obtains a spatial dimension inversely converted image by executing processing according to the following [Equation 6].

[Formula 6]
Des (x, y, c) = [Σ _{y = 0, Ns-1} Σx _{= 0, Ns-1} Fi (u, v, c) · cos (2π (ux + vy) / Ns)] / (Ns · Ns )-[Σ _{y = 0, Ns-1} Σx _{= 0, Ns-1} Fi (u, v, c) · sin (2π (ux + vy) / Ns)] / (Ns · Ns)

  [Expression 6] is an expression that realizes a known inverse two-dimensional discrete Fourier transform. Since executing a double convolution operation based on [Equation 6] entails a huge calculation load, a well-known two-dimensional fast Fourier inverse transform algorithm is applied as in the case of frequency dimensional transformation (specific description of the algorithm). Is omitted). As a result, when Ns = 1024, the inverse Fourier transform can be executed in almost real time in less than 1 second with a normal CPU. In [Formula 6], Des (x, y, c) is a real number image having the same size (the same number of pixels) as the fundus image Src (x, y, c) after size enlargement. Take the following real values:

  The spatial dimension conversion means 40 further performs processing for returning the real image Des (x, y, c) to an image. Specifically, the value of each pixel is converted from a real value from 0 to 1 to an integer value from 0 to 255. This is performed by multiplying each pixel of the real image Des (x, y, c) by 255 and then converting it to an integer. This process is the reverse process corresponding to the normalization in [Formula 1].

  The spatial dimension conversion means 40 further performs negative inversion of each pixel. By performing negative reversal, the luminance of the macular region becomes higher than that of other fundus tissues, and visibility can be improved. In negative inversion, a larger value is replaced with a smaller value, and a smaller value is replaced with a larger value, thereby inverting the magnitude relationship. In this embodiment, negative inversion is performed by subtracting from the maximum value (255 in the case of 8 bits). The conversion to the integer value and the negative inversion can be performed as individual processes, respectively, but can also be performed by collectively executing the process according to the following [Equation 7].

[Formula 7]
Fmg (x, y, c) = [1-Des (x, y, c)]. 255

  In [Formula 7], Fimg (x, y, c) is a corrected color fundus image, and each pixel is integerized and takes an integer value of 0 or more and 255 or less. Finally, an area corresponding to the original color fundus image may be obtained. Therefore, the processing of [Equation 7] is performed on the effective image area in the range of 0 ≦ x ≦ Xs−1 and 0 ≦ y ≦ Ys−1, thereby correcting the color having the number of pixels of Xs × Ys. A fundus image is created. This is because, after creating a corrected enlarged color fundus image of Ns × Ns pixels of RGB three color components, pixels corresponding to dummy pixels are removed from the corrected enlarged color fundus image, and a corrected color fundus image of Xs × Ys pixels is obtained. It is equivalent to the process of converting to

  As described above, since the obtained corrected color fundus image removes the low frequency region in a different frequency range for each color, the contrast of the macular region is improved. Therefore, by displaying and outputting the corrected color fundus image on the display unit 6, the visibility of the macular region is improved, and it becomes easy to make an accurate determination even at the time of diagnosis by a doctor.

  Next, the specific color image extraction unit 50 extracts a specific color image from the full-color corrected color fundus image (step S500). In step S500, which is a specific color image extraction step, specifically, only the R component image is extracted from the three color components of R, G, and B. Since the specific color image is an R component of the corrected color fundus image Fmg (x, y, c), c = 0, and is represented by Fimg (x, y, 0). Since this R component image removes blood vessel components and the like other than the macular region, displaying it on the display unit 6 along with the corrected color fundus image not only improves the visibility of the macular region, but also visually Thus, it becomes possible to quantitatively analyze the degree of progression of macular lesions, which is difficult.

  Next, the quantitative evaluation means 60 performs a quantitative evaluation using the specific color image (step S600). In step S600, which is a quantitative evaluation step, specifically, an evaluation value serving as a progress parameter is calculated using the specific color image, and quantitative evaluation is performed using the progress parameter. The progress parameter can be paraphrased as a disease state parameter indicating a disease state. In step S600, the quantitative evaluation means 60 first sets a predetermined area of the image area Xs × Ys as a target area effective for applying the quantitative analysis. The reason for setting the target region is to specify the center of the macular region existing in the vicinity of the center in the image region. The target region may be set anywhere in the image region as long as it does not include a bright region that is not an analysis target other than the macular region, but is preferably as close to the center as possible. In the present embodiment, both the x direction and the y direction are 1/2, and the total number of pixels is 1/4, that is, Xs / 4 ≦ x ≦ Xs · 3-4-1, Ys / 4 ≦ y ≦ Ys. -It becomes an area specified by 3 / 4-1.

  Subsequently, the quantitative evaluation unit 60 detects the coordinates (xp, yp) of the pixel having the maximum value Maxr and the maximum value Maxr in the target region. The relationship between the maximum value Maxr and the pixels of the specific color image is expressed by the following [Equation 8].

[Formula 8]
Maxr = Fimg (xp, yp, 0) = MAX _{Xs / 4 ≦ x ≦ Xs · 3 / 4-1, Ys / 4 ≦ y ≦ Ys · 3 / 4-1} Fimg (x, y, 0)

In [Equation 8], “MAX _{Xs / 4 ≦ x ≦ Xs · 3 / 4-1, Ys / 4 ≦ y ≦ Ys · 3 / 4-1} Fimg (x, y, 0)” is a pixel in the target region The maximum value (MAX) is shown. Next, the quantitative evaluation means 60 sets an analysis range with the pixel having the maximum value Maxr as the analysis center. The reason for setting the analysis range is to exclude regions other than the macular region such as the non-fundus region and the imaging frame included in the periphery of the outline of the specific color image from the analysis target. The size (size) of the analysis range may be any as long as the macular region to be analyzed is completely included, but is preferably equal to the target region. In the present embodiment, the number of pixels dx in the x direction in the analysis range and the number of pixels dy in the y direction are ½ of the number of pixels Xs in the x direction and the number of pixels Ys in the y direction, respectively. In this case, four corners (x1, y1), (x2, y1), (x2, y2), and (x1, y2) of the rectangular analysis range are x1 = xp−dx / 2 and x2 = xp + dx / 2−. 1, y1 = yp−dy / 2, y2 = yp + dy / 2-1.

  Then, the quantitative evaluation means 60 sets a value Maxr · γ obtained by multiplying the maximum value Maxr of the target region by a predetermined ratio γ as a threshold value. γ is preferably 0.7 or more and 0.9 or less. In this embodiment, γ = 0.8. The quantitative evaluation means 60 calculates Rasior by executing processing according to the following [Equation 9].

[Formula 9]
Rasior = Σ _{Fimg (x, y, 0)> Maxr · γ; x1 ≦ x ≦ x2, y1 ≦ y ≦ y2 1.100} / (dx · dy)

  As shown in [Equation 9], Rasior is a ratio of the number of pixels having a value larger than the threshold value Maxr · γ in the analysis range to the number of pixels dx · dy in the analysis range, expressed as a percentage. . A pixel having a value larger than the threshold value Maxr · γ is expressed brightly and is likely to be a macular region. Therefore, Rasior indicates a ratio of a region that is highly likely to be a macular region to an analysis range. Become. Therefore, it can be considered that the macular region is expanded as the value of Rasior is larger.

  Examples of macular lesions include macular degeneration (age-related macular degeneration), macular edema (diabetic retinopathy), macular hole / macular anterior membrane (posterior vitreous detachment due to aging, vitreous pocket). In macular degeneration and macular edema, the above-mentioned Rasior is increased as compared with healthy eyes, and particularly in macular edema, it is significantly increased. Therefore, the progression of macular lesions can be quantitatively evaluated by looking at the value of Rasior. A larger Rasior indicates that the macular lesion progresses and the symptoms worsen. By setting a predetermined threshold value as a progress threshold value and comparing Rasior with the progress threshold value, it is possible to determine whether the progress is progressing or normal. For example, if the progress threshold is set to 5% or the like, if Rasio is greater than the progress threshold 5%, it is determined as a macular lesion, and if Rasio is 5% or less, It can be determined as normal. Furthermore, by setting a plurality of progress threshold values, comparing the Rasior with each progress threshold value and specifying which range it belongs to, it is possible to determine the progress level step by step.

  FIG. 8 is a diagram illustrating an image created by the image creation method and the fundus image processing apparatus according to the present embodiment for a healthy person. Among these, FIG. 8A shows a color fundus image obtained by imaging. FIG. 8B shows a corrected color fundus image after the spatial dimension conversion by the spatial dimension conversion means 40 in step S400. FIG. 8C shows the specific color image (R component) after the specific color image extraction by the specific color image extraction means 50 in step S500.

  In the color fundus image obtained by imaging in FIG. 8A, it is difficult to identify the macular region by the naked eye. On the other hand, in the corrected color fundus image of FIG. 8B, the central macular region can be identified by the naked eye. This is because the low frequency region is removed from the R, G, and B color components. In particular, in the example of FIG. 8B, since the low frequency component is also removed from the G component, the macular region can be easily identified. As shown in FIG. 8C, in the specific color image created by extracting only the R component as a grayscale image, the center of the macular region is further increased compared to the corrected color fundus image of FIG. The fossa is easy to identify. A rectangle shown in FIG. 8C indicates the analysis range. As a result of performing the quantitative evaluation by the quantitative evaluation means 60 in step S600 on the analysis range shown in FIG. 8C, the maximum value (Maxr) 231 and the average value 45 are calculated in the analysis range of 352 pixels × 352 pixels. The number of pixels equal to or greater than the threshold value (Maxr · γ: maximum value 80%) was 1188, and the ratio (Radio) was 0.95%. When 5% is set as the progress threshold value, it is determined to be normal because Rasior is 5% or less.

  FIG. 9 is a diagram illustrating an image created by the image creation method and fundus image processing apparatus according to the present embodiment for a patient with diabetic macular edema (diabetic retinopathy). Among these, FIG. 9A shows a color fundus image obtained by imaging. FIG. 9B shows a corrected color fundus image after the spatial dimension conversion by the spatial dimension conversion means 40 in step S400. FIG. 9C shows the specific color image (R component) after the specific color image extraction by the specific color image extraction means 50 in step S500.

  In the color fundus image obtained by imaging in FIG. 9A, it is difficult to identify the macular region by the naked eye. On the other hand, in the corrected color fundus image of FIG. 9B, the central macular region can be identified by the naked eye. This is because, in each of the R, G, and B color components, the low frequency range of the R component is removed in a narrower range than that of G and B. In particular, in the example of FIG. 9B, since the components in the low frequency range are removed from all the color components, the macular region can be easily identified. As shown in FIG. 9C, in the specific color image created by extracting only the R component as a grayscale image, the center of the macular region is further improved compared to the corrected color fundus image of FIG. 9B. The fossa is easy to identify. A rectangle shown in FIG. 9C indicates the analysis range. As a result of performing quantitative evaluation by the quantitative evaluation means 60 in step S600 on the analysis range shown in FIG. 9C, the maximum value (Maxr) 255, the average value 85, and so on in the analysis range of 352 pixels × 352 pixels. The number of pixels equal to or greater than the threshold (Maxr · γ: maximum value 80%) was 12660, and the ratio (Rasio) was 10.21%. When 5% is set as the progress threshold value, since Rasior is larger than the progress threshold value 5%, it is determined as a macular lesion.

  The preferred embodiments of the present disclosure have been described above. However, the present disclosure is not limited to the above embodiments, and various modifications can be made. For example, in the above-described embodiment, a full-color image of 8 bits for each RGB is used as a color fundus image obtained by photographing. However, the present invention is not limited to this, and fundus images of various modes can be used. For example, in recent years, the gradation of commercial digital cameras has been expanded to 10 bits or more, and it is also possible to acquire full-color images such as 1024 gradations for each color, and color images having a larger number of gradations are used. May be. Further, a color image or a monochrome image having a smaller number of gradations may be used. For example, in the case of a monochrome image captured with an infrared light source, the same processing as the R component of the specific color image may be applied to the monochrome image.

1 ... CPU (Central Processing Unit)
2 ... RAM (Random Access Memory)
3 ... Storage device 4 ... Instruction input I / F
5. Data input / output I / F
DESCRIPTION OF SYMBOLS 6 ... Display part 10 ... Image size expansion means 20 ... Frequency dimension conversion means 30 ... Frequency component modification means 40 ... Spatial dimension conversion means 50 ... Specific color image extraction means 60 ... Quantitative evaluation means 70: fundus image storage means 80 ... processing data storage means 100 ... fundus image processing apparatus

Claims (12)

An image creation method for creating a corrected color fundus image in which a macular region is emphasized based on a color fundus image composed of three color components of RGB,
A frequency dimension conversion step of performing frequency dimension conversion on an image for each color component of the color fundus image and converting the frequency component data to complex frequency components for each color component;
In a low frequency range lower than a predetermined frequency, the predetermined frequency range is set as a modified frequency range for each color component, the modified frequency range of the R component is set narrower than the modified frequency range of the G component or the B component, A frequency component modification step for attenuating the value of the modified frequency range in the frequency component data of
A spatial dimension conversion step of performing frequency dimension inverse transformation on the frequency component data for each modified color component to create a corrected color fundus image composed of three color components of RGB;
An image creating method characterized by comprising:   In the frequency component modification step, a predetermined frequency range is further set as a modified frequency range in a high frequency region higher than the predetermined frequency with respect to the frequency component data of the R component, and the frequency component data of the R component The image creation method according to claim 1, wherein a value of the modified frequency range is attenuated.   The modified frequency range is defined by a sector region in which the two-dimensional Euclidean distance from the position of the DC component is a predetermined range, and the position of the DC component is the four corners of the frequency component data expressed by a square. The image creation method according to claim 1, wherein the image creation method is an image creation method.   4. The frequency component modification step further performs a process of increasing the frequency component data value for each color component at a predetermined magnification, according to claim 1. The image creation method described.   The spatial dimension conversion step further performs a process of subtracting a negative value from the maximum possible values for the values of the three color components of RGB when creating the corrected color fundus image, and performing negative inversion. The image creation method according to any one of claims 1 to 4. After the spatial dimension conversion step,
6. The specific color image extracting step of extracting the R component image of the corrected color fundus image as a specific color image in which a macular region is particularly emphasized. The image creation method according to item. For the specific color image,
Search for the maximum position that is the coordinates of the pixel with the maximum value,
Set the analysis range of a predetermined size around the searched maximum value position,
Set a threshold value obtained by multiplying the maximum value by a real value less than 1,
Count the number of pixels having a value equal to or greater than the threshold value in the analysis range,
A quantitative evaluation step of calculating a ratio of the counted number of pixels and the total number of pixels in the analysis range as a progress parameter;
The image creating method according to claim 6, further comprising: When the number of pixels of the color fundus image is Xs × Ys, an integer Ns of 2 n (n is a positive integer) that satisfies Ns / 2 <Ys ≦ Xs <Ns is defined,
An image size enlargement step of adding a dummy pixel to the color fundus image and converting it to an enlarged color fundus image of Ns × Ns pixels;
In the frequency dimension conversion step, a two-dimensional fast Fourier transform is performed on a color component image composed of Ns × Ns pixels for each color component of the enlarged color fundus image, and Ns × Ns pixels for each color component. It converts to frequency component data composed of complex number components,
The frequency component modification step modifies the frequency component data composed of complex components of the Ns × Ns pixels,
In the spatial dimension conversion step, two-dimensional fast Fourier inverse transform is performed for each color component of the modified frequency component data of the Ns × Ns pixels, and a corrected enlarged color fundus image of Ns × Ns pixels of three RGB components The pixel corresponding to the dummy pixel is removed from the corrected enlarged color fundus image, and converted to the corrected color fundus image of Xs × Ys pixels. The image creation method as described in any one of Claims.   The frequency component modification step is configured such that, when Ns = 1024, the two-dimensional Euclidean distance is set within a range of 1 to 10 as a modified frequency range in the low frequency range. The image creation method described. A fundus image processing apparatus for creating a corrected color fundus image in which a macular region is emphasized based on a color fundus image composed of RGB three-color components,
Frequency dimensional conversion means for performing frequency dimensional conversion on each color component image of the color fundus image and converting the frequency component data to complex frequency components for each color component;
For the complex number component of each color component of the frequency component data,
In a low frequency range lower than a predetermined frequency, a predetermined frequency range is set as a modified frequency range for each color component, and the modified frequency range of the R component is set to be narrower than the modified frequency ranges of the G component and the B component and attenuated. The frequency component modification means for modifying the frequency component data for each color component,
A spatial dimension conversion means for performing frequency dimension inverse transformation on the frequency component data for each modified color component and creating a corrected color fundus image composed of three color components of RGB;
A fundus image processing apparatus comprising: Computer
A frequency dimension conversion step of performing frequency dimension conversion on the image of each color component of the color fundus image and converting the frequency component data to complex frequency components for each color component;
With respect to the complex number component of each color component of the frequency component data, in a low frequency range lower than a predetermined frequency, the predetermined frequency range for each color component is set as the modified frequency range, and the modified frequency range of the R component is set as the G component or B component. A frequency component modification step for modifying the frequency component data for each color component by setting the component to be narrower than the modified frequency range and attenuating it.
A spatial dimension conversion step of performing a frequency dimension inverse transform on the frequency component data for each modified color component to create a corrected color fundus image composed of three color components of RGB;
Program to function as. Computer
A frequency dimension conversion step of performing frequency dimension conversion on the image of each color component of the color fundus image and converting the frequency component data to complex frequency components for each color component;
With respect to the complex number component of each color component of the frequency component data, in a low frequency range lower than a predetermined frequency, the predetermined frequency range for each color component is set as the modified frequency range, and the modified frequency range of the R component is set as the G component or B component. A frequency component modification step for modifying the frequency component data for each color component by setting the component to be narrower than the modified frequency range and attenuating it.
A spatial dimension conversion step of performing a frequency dimension inverse transform on the frequency component data for each modified color component to create a corrected color fundus image composed of three color components of RGB;
A recording medium on which a program for functioning as a recording medium is recorded.