Microscope image correction method and microscope image correction program

Abstract

To provide a microscope image correction method that can correct an original image obtained with a microscope and calculate a corrected image with a higher signal-to-noise ratio.SOLUTION: A microscope image correction method includes the steps of: calculating a power root image of a sample from an original image of the sample; calculating a spatial frequency spectrum of the power root image of the sample from the power root image of the sample; calculating a corrected spatial frequency spectrum of the power root image of the sample by applying a high-frequency rejection filter to the spatial frequency spectrum of the power root image of the sample; calculating a corrected power root image of the sample from the corrected spatial frequency spectrum of the power root image of the sample; and calculating a corrected image of the sample from the corrected power root image of the sample.SELECTED DRAWING: Figure 4

Description

[Technical Field]

[0001] This disclosure relates to a microscope image correction method and a microscope image correction program. [Background technology]

[0002] Japanese Patent Publication No. 2018-533769 (Patent Document 1) discloses two-photon excitation microscopy (TPM). [Prior art documents] [Patent Documents]

[0003] [Patent Document 1] Special Publication No. 2018-533769 [Overview of the project] [Problems that the invention aims to solve]

[0004] The purpose of this disclosure is to provide a microscope image correction method and a microscope image correction program that can correct original images obtained with a microscope and calculate corrected images having a higher signal-to-noise ratio. [Means for solving the problem]

[0005] The first aspect of the present disclosure is a microscopic image correction method comprising the steps of: calculating a root power image of a sample from an original image of a sample; calculating a spatial frequency spectrum of the root power image of a sample from the root power image of a sample; applying a high-frequency rejection filter to the spatial frequency spectrum of the root power image of a sample to calculate a corrected spatial frequency spectrum of the root power image of a sample; calculating a corrected root power image of a sample from the corrected spatial frequency spectrum of the root power image of a sample; and calculating a corrected image of a sample from the corrected root power image of a sample.

[0006] The second aspect of the microscope image correction method of this disclosure is a step of calculating a root-of-power image of a sample from an original image of the sample; a step of calculating a first spatial frequency spectrum of the root-of-power image of the sample from the root-of-power image of the sample; a step of applying a high-frequency rejection filter to the first spatial frequency spectrum of the root-of-power image of the sample to calculate a first corrected spatial frequency spectrum of the root-of-power image of the sample; and a step of deconvolving the first corrected spatial frequency spectrum of the root-of-power image of the sample with the optical transfer function of the optical system calculated by Fourier transforming the root of the point image distribution function of the optical system used to obtain the original image of the sample. The method comprises the steps of: calculating a second corrected spatial frequency spectrum of the root-of-power image of the sample; calculating a corrected root-of-power image of the sample from the second corrected spatial frequency spectrum of the root-of-power image of the sample; calculating a provisional corrected image of the sample by raising the corrected root-of-power image of the sample to a power; calculating a second spatial frequency spectrum of the provisional corrected image of the sample from the provisional corrected image of the sample; and calculating a third corrected spatial frequency spectrum of the original image of the sample by deconvolving the second spatial frequency spectrum of the provisional corrected image of the sample with the corrected optical transfer function of the optical system. The corrected optical transfer function of the optical system is calculated by performing a Fourier transform on the first corrected point image distribution function. The first corrected point image distribution function is calculated by raising the second corrected point image distribution function to a power. The second corrected point image distribution function is calculated by deconvolving the product of the Fourier transform of the root-of-power of the point image distribution function of the optical system obtained from the original image of the sample and a high-frequency rejection filter with the optical transfer function of the optical system, and then performing an inverse Fourier transform. The second aspect of the present disclosure of the microscope image correction method comprises the step of calculating a corrected image of a sample from a third corrected spatial frequency spectrum of the original image of the sample.

[0007] A third aspect of the present disclosure of a microscope image correction method comprises the steps of: calculating a root-of-power image of a sample from an original image of a sample; calculating a first spatial frequency spectrum of the root-of-power image of a sample from the root-of-power image of the sample; calculating a first corrected spatial frequency spectrum of the root-of-power image of a sample by deconvolving the first spatial frequency spectrum of the root-of-power image of the sample with the optical transfer function of the optical system, which is calculated by Fourier transforming the root-of-power of the point image distribution function of the optical system used to obtain the original image of the sample; calculating a corrected root-of-power image of a sample from the first corrected spatial frequency spectrum of the root-of-power image of the sample; calculating a provisional corrected image of a sample by raising the corrected root-of-power image of the sample to a power; calculating a second spatial frequency spectrum of the provisional corrected image of a sample from the provisional corrected image of the sample; and calculating a second corrected spatial frequency spectrum of the original image of a sample by deconvolving the second spatial frequency spectrum of the provisional corrected image of the sample with the corrected optical transfer function of the optical system. The corrected optical transfer function is calculated by Fourier transforming the first corrected point image distribution function. The first corrected point image distribution function is calculated by raising the second corrected point image distribution function to a power. The second corrected point image distribution function is calculated by deconvolving the product of the Fourier transform of the power root of the point image distribution function of the optical system used to obtain the original image of the sample and a high-frequency rejection filter, using the optical transfer function of the optical system, and then performing an inverse Fourier transform. The third aspect of the microscopic image correction method of this disclosure comprises the step of calculating a corrected image of the sample from the second corrected spatial frequency spectrum of the original image of the sample.

[0008] The microscope image correction program of this disclosure causes a processor to execute each step of any of the first to third aspects of the microscope image correction method of this disclosure. [Effects of the Invention]

[0009] According to the microscope image correction method and microscope image correction program of this disclosure, it is possible to correct the original image obtained by a microscope and calculate a corrected image having a higher signal-to-noise ratio. [Brief explanation of the drawing]

[0010] [Figure 1] This is a schematic diagram showing the microscope systems of Embodiments 1 to 3. [Figure 2] This is a block diagram illustrating the functional configuration of the computer included in the microscope systems of Embodiments 1 to 3. [Figure 3] This is a block diagram illustrating the functional configuration of the image processing unit in Embodiment 1. [Figure 4] This diagram shows a flowchart of the correction method for the original image obtained with the microscope of Embodiment 1. [Figure 5] This figure shows flowcharts of the denoising process in Embodiment 1 and Embodiment 2. [Figure 6] This diagram shows a flowchart of the steps for calculating the corrected image in Embodiment 1. [Figure 7] This figure shows the relationship between the original images and their root power images from multiphoton excitation microscopes such as two-photon or three-photon excitation microscopes, confocal microscopes, or super-resolution microscopes, the one-photon excitation images from a one-photon excitation scanning microscope with a single-element photodetector and no pinhole, and their Fourier transforms (spatial frequency spectra). [Figure 8] This is a block diagram illustrating the functional configuration of the image processing unit in Embodiment 2. [Figure 9] This diagram shows a flowchart of the correction method for the original image obtained with the microscope of Embodiment 2. [Figure 10] This figure shows a flowchart of the second deconvolution process step in Embodiment 2. [Figure 11] This is a block diagram illustrating the functional configuration of the image processing unit in Embodiment 3. [Figure 12] This diagram shows a flowchart of the correction method for the original image obtained with the microscope of Embodiment 3. [Figure 13] This figure shows a flowchart of the first deconvolution process step in Embodiment 3. [Figure 14]This figure shows a flowchart of the second deconvolution process step in Embodiment 3. [Figure 15] This is a diagram of the first original image. [Figure 16] This figure shows the corrected image for Example 1-1. [Figure 17] This figure shows the corrected image for Example 2-1. [Figure 18] This figure shows the corrected image for Example 3-1. [Figure 19] This figure shows a graph representing the improvement rate of the signal-to-noise ratio of the images from Examples 1-1 to 3-2. [Modes for carrying out the invention]

[0011] The embodiments are described below. Note that identical components will be given the same reference numerals, and their descriptions will not be repeated.

[0012] (Embodiment 1)

[0013] The first embodiment of the microscope system 1 will be described with reference to Figures 1 to 3. As shown in Figure 1, the microscope system 1 comprises a microscope 2, a computer 3, and a display device 4. Microscope 2 is a microscope that can obtain an original image of the sample 12. In this embodiment, microscope 2 is a two-photon excitation microscope.

[0014] <Microscope Configuration>

[0015] As shown in Figure 1, the microscope 2 includes an illumination optical system 5, a sample stage 13, and an observation optical system 15.

[0016] The illumination optical system 5 includes, for example, a light source 6, a scanning mirror 8, lenses 9a, 9b, 9c, 9d, a dichroic mirror 10, and an objective lens 11.

[0017] Light source 6 outputs light 7. Light source 6 is, for example, a femtosecond ultrashort pulse laser such as a titanium-sapphire laser. Light 7 is, for example, a femtosecond pulsed laser beam. Light 7 passes through lens 9a and enters lens 9b. Light 7 is spread by lenses 9a and 9b. Light 7 is collimated by lens 9b and enters scanning mirror 8. Scanning mirror 8 reflects and scans light 7. Scanning mirror 8 is, for example, a galvanometer mirror. Light 7 passes through lens 9c and enters lens 9d. Light 7 is collimated by lens 9d and enters dichroic mirror 10. Dichroic mirror 10 reflects light 7 toward sample 12. Light 7 is focused by objective lens 11 and illuminates a portion of sample 12. The focal point of light 7 is scanned by the scanning mirror 8 within a plane perpendicular to the direction of incidence of light 7 on the sample 12.

[0018] Sample 12 is supported by a sample stage 13. Sample 12 is not particularly limited, but may be, for example, a living cell, molecule, protein, or DNA. Sample 12 is labeled with a fluorescent substance, such as a fluorescent dye molecule. When light 7 is shone on sample 12, light 7 is absorbed by the fluorescent substance that labels sample 12. The fluorescent substance is excited by absorbing light 7 with two photons and emits light 14. Light 14 is incoherent light, such as fluorescence light. Thus, light 14 is emitted from sample 12.

[0019] The observation optical system 15 includes, for example, an objective lens 11, a dichroic mirror 10, and a photodetector 16. Light 14 emitted from the sample 12 is collimated by the objective lens 11 and incident on the dichroic mirror 10. The light 14 passes through the dichroic mirror 10 and is incident on the photodetector 16. The photodetector 16 detects the intensity of the light 14 emitted from each part of the sample 12. The photodetector 16 is, for example, a photomultiplier tube (PMT).

[0020] Referring to Figure 1, computer 3 is communicatively connected to microscope 2. Computer 3 controls microscope 2, performs image reconstruction processing to construct the original image of sample 12, and performs image processing to correct the original image of sample 12. Computer 3 is, for example, a personal computer, microcomputer, cloud server, or smart device (such as a smartphone or tablet). Display device 4 is communicatively connected to computer 3. Display device 4 is, for example, a liquid crystal display device or an organic EL display device. Display device 4 displays, for example, the operating status of microscope 2, the original image of sample 12, and the corrected image of sample 12.

[0021] <Configuration of Computer 3>

[0022] Referring to Figure 1, the computer 3 includes a processor 20 and a storage device 22. The storage device 22 is, for example, RAM (Random Access Memory), ROM (Read Only Memory), or a hard disk. The storage device 22 stores computer programs executed by the processor 20, original images of the sample 12, and corrected images of the sample 12. The computer program includes, for example, an image acquisition program for acquiring original images of the sample 12 and an image correction program for correcting the original images of the sample 12 and calculating corrected images of the sample 12. The image acquisition program and the image correction program may be stored in a non-temporary computer-readable storage medium.

[0023] The processor 20 is composed of, for example, a CPU (Central Processing Unit), a GPU (Graphics Processing Unit), or an FPGA (field-programmable gate array). The processor 20 executes the computer program stored in the memory device 22. By executing the computer program, the computer 3 realizes the functions shown in Figures 2 and 3. The functional configuration of the computer 3 will be explained with reference to Figures 2 and 3.

[0024] As shown in Figure 2, the computer 3 includes an original image acquisition unit 25 and an image processing unit 30. The original image acquisition unit 25 is realized by the processor 20 executing an image acquisition program. The image processing unit 30 corrects the original image of the sample 12 obtained by the microscope 2 and calculates a corrected image of the sample 12. The image processing unit 30 is realized by the processor 20 executing an image correction program.

[0025] As shown in Figure 3, the image processing unit 30 includes a root image calculation unit 31, a denoising unit 32, and a corrected image calculation unit 35. The root image calculation unit 31 performs the step of calculating a root image from the original image of the sample 12 (step S1, see Figure 4). The denoising unit 32 performs the denoising step (step S2, see Figures 4 and 5). The corrected image calculation unit 35 performs the step of calculating a corrected image of the sample 12 (step S3, see Figures 4 and 6).

[0026] <Method for obtaining the original image of sample 12>

[0027] The method for acquiring the original image of sample 12 is performed by the original image acquisition unit 25. Specifically, the original image acquisition unit 25 controls the light source 6 to emit light 7 from the light source 6. The light 7 passes through lens 9a and enters lens 9b. The light 7 is spread by lenses 9a and 9b. The light 7 is collimated by lens 9b and enters the scanning mirror 8. The light 7 is reflected by the scanning mirror 8. The light 7 passes through lenses 9c and 9d and is reflected by the dichroic mirror 10. The light 7 is focused by the objective lens 11 and illuminates a part of sample 12. The original image acquisition unit 25 controls the scanning mirror 8 to scan the focal point of the light 7 in sample 12 within a plane perpendicular to the direction of incidence of the light 7 in sample 12. The photodetector 16 detects the intensity of the light 14 emitted from each part of sample 12. The original image acquisition unit 25 constructs a two-dimensional original image of the sample 12 from the intensity of the light 14 emitted from each part of the sample 12 and detected by the photodetector 16. For example, if the microscope 2 is a two-photon excitation microscope, the original image of the sample 12 is the two-photon excitation image I2. If the microscope 2 is a three-photon excitation microscope, the original image of the sample 12 is the three-photon excitation image I3.

[0028] The original image acquisition unit 25 may further control the objective lens 11 to move the focal point of the light 7 in the sample 12 in the direction of incidence of the light 7 in the sample 12. The original image acquisition unit 25 may stack the two-dimensional original images of the sample 12 to construct a three-dimensional original image of the sample 12 as the original image of the sample 12.

[0029] <Method for correcting the original image of sample 12 obtained with microscope 2>

[0030] Referring to Figures 4 to 7, the method for calculating a corrected image of sample 12 from the original image of sample 12 obtained by microscope 2 will be explained. The method for calculating the corrected image of sample 12 is performed by the image processing unit 30. Below, we will mainly explain the case where microscope 2 is a two-photon excitation microscope and the original image of sample 12 is the two-photon excitation image I2 of sample 12.

[0031] To obtain the original image of sample 12 (for example, the two-photon excitation image I2), sample 12 is excited by one photon using light having the same wavelength as the light 7 irradiated onto sample 12. The image obtained by this excitation is defined as the one-photon excitation image I1 (see Figure 7). An image obtained by the power root of the original image and having equivalent resolution to the one-photon excitation image I1 is defined as the power root image I1' (see Figure 7). Original image of the sample I m When is an m-photon excitation image, the root means the m-th root, the power means the m-th power, and the root image I1' is the original image I m This refers to the image obtained by raising (m) to the power of (1 / m) (where m is a natural number).

[0032] As shown in Figure 4, the root image I1' is calculated from the original image of sample 12 (step S1). The root image I1' is calculated by taking the root of the original image of sample 12. For example, if the original image of sample 12 is the two-photon excitation image I2 of sample 12, the root image I1' is calculated by equation (1). That is, the light intensity of each pixel in the root image I1' is calculated by calculating the square root of the light intensity of each pixel in the two-photon excitation image I2.

number

[0033] For example, if the original image of sample 12 is the three-photon excitation image I3 of sample 12, the root image I1' is calculated by equation (2). That is, the light intensity of each pixel in the root image I1' is calculated by calculating the cube root of the light intensity of each pixel in the three-photon excitation image I3.

number

[0034] As shown in Figures 4 and 5, a denoising step is performed (step S2).

[0035] Specifically, as shown in Equation (3), the square root image I1' is Fourier-transformed to calculate the spatial frequency spectrum F1' of the square root image I1' (step S2-1). In this specification, FT represents the Fourier transform.

Number

[0036] As shown in Equation (4), by multiplying the spatial frequency spectrum F1' of the square root image I1' by a high-frequency removal filter G(k x , k<\(0000004\)>, k<\(0000005\)>)(hereinafter sometimes referred to as the "high-frequency removal filter G"), the corrected spatial frequency spectrum F<\(0000006\)>' of the square root image I1' is calculated (step S2-2). Therefore, in the corrected spatial frequency spectrum F<\(0000007\)>' of the square root image I1', high-frequency noise is removed.

Number

[0037] The high-frequency removal filter G is given, for example, by Equation (5) on the two axes of the horizontal wave number k<\(0000008\)> and the vertical wave number k<\(0000009\)>. The vertical direction means the incident direction of the light 7 in the sample 12. The horizontal direction means the direction perpendicular to the incident direction of the light 7 in the sample 12.

Number

[0038] k<\(0000010\)> is given by Equation (6). The x direction means one of the directions perpendicular to the incident direction of the light 7 in the sample 12. The y direction means the direction perpendicular to the x direction among the directions perpendicular to the incident direction of the light 7 in the sample 12. k<\(0000011\)> means the wave number in the x direction, and k<\(0000012\)> means the wave number in the y direction.

Number

[0039] K is given by equation (7).

number

[0040] In equation (5), angle α is the maximum angle of incidence of light that passes through the objective lens 11 and enters the sample 12 (see Figure 1). Angle α is determined by the numerical aperture NA of the objective lens 11 and the refractive index n of the medium from the objective lens 11 to the sample 12. Specifically, angle α is given by equation (8).

number

[0041] As shown in Figures 4 and 6, a corrected image of sample 12 is calculated (step S3).

[0042] Specifically, as shown in equation (9), the corrected spatial frequency spectrum F of the root image I1' 1a ' is inverse Fourier transformed to obtain the root-corrected image I 1a Calculate ' (Step S3-1). In this specification, FT -1 This represents the inverse Fourier transform.

number

[0043] Root Corrected Image I 1a Corrected image I of sample 12 2f Calculate (Step S3-2). Specifically, the root-corrected image I 1a By raising ' to a power, the corrected image I of sample 12 is obtained. 2f The following is calculated. For example, if the original image of sample 12 is the two-photon excitation image I2 of sample 12, then the root-corrected image I is calculated as shown in equation (10). 1a By squaring ', the corrected image I of sample 12 is obtained. 2fCalculate the root-corrected image I. 1a By calculating the square of the light intensity of each pixel of ', the corrected image I of sample 12 is obtained. 2f The light intensity of each pixel is calculated. The high-frequency rejection filter G corrects the image I of sample 12. 2f Then, high-frequency noise is removed. Therefore, corrected image I 2f The signal-to-noise ratio (contrast ratio) will improve.

number

[0044] If the original image of sample 12 is the three-photon excitation image I3 of sample 12, then the root-corrected image I is as shown in equation (11). 1a By cubing ', the corrected image I of sample 12 is obtained. 3f Calculate the root-corrected image I. 1a By calculating the cube of the light intensity of each pixel of ', the corrected image I of sample 12 is obtained. 3f The light intensity of each pixel is calculated. The high-frequency rejection filter G corrects the image I of sample 12. 3f Then, high-frequency noise is removed. Therefore, corrected image I 3f The signal-to-noise ratio (contrast ratio) will improve.

number

[0045] The effects of this embodiment will now be explained.

[0046] The microscope image correction method of this embodiment includes the steps of: calculating the root power image I1' of sample 12 from the original image of sample 12 (for example, the two-photon excitation image I2) (step S1); calculating the spatial frequency spectrum F1' of the root power image I1' of sample 12 from the root power image I1' of sample 12 (step S2-1); and applying a high-frequency rejection filter G to the spatial frequency spectrum F1' of the root power image I1' of sample 12 to obtain the corrected spatial frequency spectrum F 1aStep S2-2: Calculate the ' and correct the spatial frequency spectrum F of the root image I1' of sample 12. 1a 'Root-corrected image I of sample 12 1a The steps to calculate ' (step S3-1) and the root-corrected image I of sample 12 1a Corrected image I of sample 12 2f The process includes a step (step S3-2) for calculating [the specified value].

[0047] The microscope image correction program of this embodiment causes the processor 20 to execute each step of the microscope image correction method of this embodiment.

[0048] As shown in Figure 7, when sample 12 is single-photon excited using the illumination optical system 5, single-photon excitation occurs throughout the entire portion of sample 12 through which light 7 passes (see the real-space point image I1 in Figure 7). Therefore, light is emitted from the entire sample 12. In a single-photon excitation scanning microscope, where the photodetector is a single element and does not have a pinhole, the light emitted from the entire sample 12 is summed up and detected. In contrast, the light 14 detected by multi-photon excitation microscopes such as two-photon or three-photon excitation microscopes, confocal microscopes, or super-resolution microscopes (Image Scanning Microscopy (ISM)) is limited to only the portion of sample 12 near the focal point of the objective lens 11 (see original image I in Figure 7). m (See point image in real space). The reason for this is that the point image distribution function of these microscopes is related to the point image distribution function of a simple one-photon excitation scanning microscope without a pinhole by raising it to the power of m. Therefore, the amount of light detected by the photodetector 16 increases further in areas where the illumination light intensity is strong, while the amount of light detected by the photodetector 16 decreases further in areas where the illumination light intensity is weak.

[0049] As a result, original image I m The resolution of is greater than the resolution of image I1. In other words, in the spatial frequency domain, the resolution of the original image I m The cutoff frequency is greater than the cutoff frequency of image I1 (original image I in Figure 7). m(See spatial frequencies of the original image I1). The cutoff frequencies of these images are given by the illumination optical system 5. m The resolution of the root image I1' obtained by taking the root of is equivalent to the resolution of image I1 (see the real-space point image column in Figure 7). Therefore, similar to image I1, the cutoff frequency of the root image I1' is the same as that of the original image I m This becomes lower than the cutoff frequency and is equivalent to the cutoff frequency in image I1 (see the spatial frequency column in Figure 7).

[0050] The signal component from sample 12 in the root image I1' is the same as the original image I of sample 12. m The signal components from sample 12 are distributed in a smaller spatial frequency range than those from sample 12. In contrast, the noise components in the root image I1' are distributed in a smaller spatial frequency range than those in the original image I of sample 12. m Similar to the noise components in the original image, these components are randomly distributed across the entire spatial frequency domain. Therefore, by performing a denoising step (S2) in which a high-frequency rejection filter G is applied to the spatial frequency spectrum F1' of the root image I1', noise distributed over a wider spatial frequency domain can be removed. In this way, a corrected image with a higher signal-to-noise ratio (contrast ratio) can be calculated.

[0051] (Embodiment 2)

[0052] The microscope system 1 of Embodiment 2 will be described with reference to Figures 1, 2, and 8. The microscope system 1 of Embodiment 2 is configured similarly to the microscope system 1 of Embodiment 1, but the functional configuration of the computer 3 is different.

[0053] As shown in Figure 8, the image processing unit 30 of this embodiment is configured similarly to the image processing unit 30 of Embodiment 1, but differs from the image processing unit 30 of Embodiment 1 in that it further includes a first deconvolution processing unit 33 and a second deconvolution processing unit 34.

[0054] The root image calculation unit 31 performs the step of calculating a root image I1' from the original image of sample 12 (for example, the two-photon excitation image I2) (step S1, see Figure 9). The denoising unit 32 performs the denoising step (step S2, see Figure 9). The first deconvolution unit 33 performs the first deconvolution step (step S3, see Figure 9). The second deconvolution unit 34 performs the second deconvolution step (step S4, see Figures 9 and 10). The corrected image calculation unit 35 performs the step of calculating a corrected image of sample 12 (step S5, see Figure 9).

[0055] Referring to Figures 5, 9, and 10, the method for correcting the original image of sample 12 in Embodiment 2 will be described. The method for calculating the corrected image of sample 12 is performed by the image processing unit 30. Below, we will mainly describe the case where the microscope 2 is a two-photon excitation microscope and the original image of sample 12 is the two-photon excitation image I2 of sample 12. The method for correcting the original image of sample 12 in Embodiment 2 is similar to the method for correcting the original image of sample 12 in Embodiment 1, but differs from the method for correcting the original image of sample 12 in Embodiment 1 mainly in that it further comprises a first deconvolution processing step (step S3) and a second deconvolution processing step (step S4).

[0056] The light 7 irradiated onto sample 12 is expressed as an electric field Eexp(iθ) having amplitude E and phase θ. The intensity of light 7 is |E 2As light 7 is absorbed by sample 12, sample 12 is excited by light 7. In the absorption of light 7 intensity by sample 12, information about the phase θ of the electric field of light 7 is lost, and the amplitude E of the electric field of light 7 is squared. Therefore, in the spatial frequency spectrum of the original image of sample 12, the amplitude component of the spatial frequency component close to the cutoff frequency of the illumination optical system 5 decreases, and the S / N ratio (contrast ratio) of the original image of sample 12 decreases. If the original image of sample 12 is the one-photon excitation image I1 of sample 12, the decrease in the amplitude component of the spatial frequency component close to the cutoff frequency of the illumination optical system 5 occurs once, but if the original image of sample 12 is the two-photon excitation image I2 of sample 12, the decrease in the amplitude component of the spatial frequency component close to the cutoff frequency of the illumination optical system 5 occurs twice. Therefore, in the two-photon excitation image I2 of sample 12, a first deconvolution processing step (step S3) and a second deconvolution processing step (step S4) are performed to recover the amplitude component of the spatial frequency component close to the cutoff frequency of the illumination optical system 5.

[0057] Specifically, referring to Figures 5 and 9, steps S1 and S2 in this embodiment are the same as steps S1 and S2 in Embodiment 1.

[0058] As shown in Figure 9, the first deconvolution process step (step S3) is performed after the denoising process (step S2).

[0059] Specifically, as shown in equation (12), the corrected spatial frequency spectrum F of the root image I1' 1a ' is deconvolved with the optical transfer function OTF1 of the optical system (illumination optical system 5) to obtain the corrected spatial frequency spectrum F of the root image I1'. 1b The following is calculated: '. As a result, noise with frequencies below the cutoff frequency of the illumination optical system 5 is reduced. The first deconvolution process includes, for example, a deconvolution process using OTF1 as a Wiener filter. In equation (12), w1 represents the Wiener constant.

number

[0060] The optical transfer function OTF1 is calculated by performing a Fourier transform on the root of the point image distribution function PSF2 of the optical system (illumination optical system 5) that obtains the original image of sample 12 (e.g., the two-photon excitation image I2). For example, if the original image of sample 12 is the two-photon excitation image I2 of sample 12, OTF1 is given by equation (13).

number

[0061] The objective of this embodiment is to improve the signal-to-noise ratio (SNR) of the original image of the sample 12 obtained by the microscope 2. Therefore, the first deconvolution process may be performed using the optical transfer function OTF1, or it may be performed using the modulation transfer function MTF1, which is the absolute value of the optical transfer function OTF1.

[0062] As shown in Figure 10, the second deconvolution process step (step S4) is performed after the first deconvolution process step (step S3).

[0063] Specifically, as shown in equation (14), the corrected spatial frequency spectrum F of the root image I1' 1b ' is inverse Fourier transformed to obtain the root-corrected image I 1b Calculate ' (Step S4-1).

number

[0064] Root Corrected Image I 1b By raising ' to a power, the provisional corrected image I of sample 12 is obtained. 2b Calculate (Step S4-2). For example, if the original image of sample 12 is the two-photon excitation image I2 of sample 12, then the root-corrected image I is calculated as shown in equation (15). 1b By squaring ', the provisional corrected image I of sample 12 is obtained. 2b This calculates the root-corrected image I. 1bBy calculating the square of the light intensity of each pixel of ', the provisional corrected image I of sample 12 is obtained. 2b The light intensity of each pixel is calculated. Provisional corrected image I of sample 12. 2b In this process, the denoising process (step S2) removes high-frequency noise with frequencies above the cutoff frequency of the illumination optical system 5, and the first deconvolution process step (step S3) reduces noise with frequencies below the cutoff frequency of the illumination optical system 5.

number

[0065] As shown in equation (16), the provisional corrected image I of sample 12 2b Performing a Fourier transform on this, we obtain the provisional corrected image I of sample 12. 2b The spatial frequency spectrum F 2b Calculate (Step S4-3).

number

[0066] As shown in equation (17), the provisional corrected image I of sample 12 2b The spatial frequency spectrum F 2b The corrected optical transfer function OTF of the optical system (illumination optical system 5) 2b By deconvolution, the corrected spatial frequency spectrum F of the original image of sample 12 is obtained. 2c The second deconvolution process is performed, for example, OTF. 2b This includes a deconvolution process using as a Wiener filter. In equation (17), w2 represents the Wiener constant.

number

[0067] The objective of this embodiment is to improve the signal-to-noise ratio (SNR) of the original image of sample 12 obtained with microscope 2. Therefore, the second deconvolution process uses the corrected optical transfer function OTF.2b This may be done using the corrected optical transfer function OTF. 2b The MTF is the absolute value of the corrected modulation transfer function. 2b It may be done in this way.

[0068] Corrected Optical Transfer Function (OTF) 2b This is the corrected point image distribution function (PSF). 2b It is calculated by performing a Fourier transform. Specifically, the corrected optical transfer function OTF 2b This is given by equation (18).

number

[0069] Corrected point image distribution function (PSF) 2b This is the corrected point image distribution function (PSF). 1b It is calculated by raising to a power. For example, if the original image of sample 12 is the two-photon excitation image I2 of sample 12, then PSF 2b This is given by equation (19).

number

[0070] Corrected point image distribution function (PSF) 1b This is calculated by deconvolving the product of the Fourier transform of the root of the point image distribution function PSF2 of the optical system (illumination optical system 5) that obtains the original image of sample 12 and the high-frequency rejection filter G using the optical transfer function OTF1, and then performing an inverse Fourier transform. For example, if the original image of sample 12 is the two-photon excitation image I2 of sample 12, then PSF 1b This is given by equation (20).

number

[0071] The objective of this embodiment is to improve the signal-to-noise ratio (SNR) of the original image of sample 12 obtained with microscope 2. Therefore, the corrected point image distribution function PSF is used. 1bTo calculate this, the product of the Fourier transform of the root of the point image distribution function PSF2 and the high-frequency rejection filter G may be deconvolved with respect to the optical transfer function OTF1, or with respect to the modulation transfer function MTF1, which is the absolute value of the optical transfer function OTF1.

[0072] Referring to Figure 9, the corrected spatial frequency spectrum F of the original image of sample 12 is as shown in equation (21). 2c The corrected image I of sample 12 is obtained by performing an inverse Fourier transform on it. 2f Calculate (Step S5).

number

[0073] In addition to the effects of Embodiment 1, this embodiment provides the following effects.

[0074] The microscope image correction method of this embodiment includes the steps of: calculating a root-of-power image I1' of sample 12 from the original image of sample 12 (for example, two-photon excitation image I2) (step S1); calculating a first spatial frequency spectrum (spatial frequency spectrum F1') of the root-of-power image I1' of sample 12 from the root-of-power image I1' of sample 12 (step S2-1); and applying a high-frequency rejection filter G to the first spatial frequency spectrum of the root-of-power image I1' of sample 12 to obtain a first corrected spatial frequency spectrum (corrected spatial frequency spectrum F1') of the root-of-power image I1' of sample 12. 1a Step S2-2 involves calculating the first corrected spatial frequency spectrum of the root image I1' of sample 12 by deconvolving it with the optical transfer function (optical transfer function OTF1) of the optical system, which is calculated by Fourier transforming the root of the point image distribution function of the optical system (illumination optical system 5) that obtains the original image of sample 12, thereby obtaining the second corrected spatial frequency spectrum (corrected spatial frequency spectrum F) of the root image I1' of sample 12. 1b Step (Step S3) to calculate ') and the second corrected spatial frequency spectrum of the root image I1' of sample 12 to obtain the root corrected image I 1bThe step of calculating ’ (step S4-1), and the square root corrected image I of the sample 12 1b By raising ’ to the power, the provisional corrected image I of the sample 12 2b The step of calculating (step S4-2), and the provisional corrected image I of the sample 12 2b From the provisional corrected image I of the sample 12 to the provisional corrected image I of the sample 12 2b The second spatial frequency spectrum (spatial frequency spectrum F 2b ) The step of calculating (step S4-3), and the provisional corrected image I of the sample 12 2b The second spatial frequency spectrum of is deconvolved with the correction optical transfer function of the optical system (correction optical transfer function OTF 2b ) to calculate the third corrected spatial frequency spectrum of the original image of the sample 12 (corrected spatial frequency spectrum F 2c ) The step of calculating (step S4-4). The correction optical transfer function is calculated by performing a Fourier transform on the first corrected point spread function (corrected point spread function PSF 2b ). The first corrected point spread function is calculated by raising the second corrected point spread function (corrected point spread function PSF 1b ) to the power. The second corrected point spread function is obtained by deconvolving the product of the Fourier transform of the square root of the point spread function of the optical system for obtaining the original image of the sample 12 and the high-frequency removal filter G with the optical transfer function of the optical system (optical transfer function OTF1), and then performing an inverse Fourier transform. The microscope image correction method of the present embodiment includes the step of calculating the corrected image I of the sample 12 from the third corrected spatial frequency spectrum of the original image of the sample 12 (step S5). 2f

[0075] The microscope image correction program of the present embodiment causes the processor 20 to execute each step of the microscope image correction method of the present embodiment.

[0076] ​When the light 7 from the illumination optical system 5 is absorbed by the sample 12, the amplitude E of the electric field of the light 7 is squared, which reduces the amplitude component of the spatial frequency component close to the cutoff frequency of the illumination optical system 5 in the spatial frequency spectrum of the original image of the sample 12. As a result, the signal-to-noise ratio (contrast ratio) of the original image of the sample 12 decreases. In this embodiment, since the first deconvolution processing step (step S3) and the second deconvolution processing step (step S4) are performed, the amplitude component of the spatial frequency component close to the cutoff frequency of the illumination optical system 5 can be restored. Corrected image I of sample 12 2f The signal-to-noise ratio (contrast ratio) improves.

[0077] (Embodiment 3)

[0078] The microscope system 1 of Embodiment 3 will be described with reference to Figures 1, 2, and 11. The microscope system 1 of Embodiment 3 is configured similarly to the microscope system 1 of Embodiment 2, but the functional configuration of the computer 3 is different.

[0079] As shown in Figure 11, the image processing unit 30 of this embodiment is configured similarly to the image processing unit 30 of embodiment 2, but differs from the image processing unit 30 of embodiment 2 in that the denoising processing unit 32 is omitted.

[0080] The root image calculation unit 31 performs the step of calculating a root image I1' from the original image of the sample 12 (for example, the two-photon excitation image I2) (step S1, see Figure 12). The first deconvolution processing unit 33 performs the first deconvolution processing step (step S2, see Figures 12 and 13). The second deconvolution processing unit 34 performs the second deconvolution processing step (step S3, see Figures 12 and 14). The corrected image calculation unit 35 performs the step of calculating a corrected image of the sample 12 (step S4, see Figure 12).

[0081] Referring to Figures 12 to 14, the method for correcting the original image of sample 12 in Embodiment 3 will be described. The method for calculating the corrected image of sample 12 is performed by the image processing unit 30. Below, we will mainly describe the case where the microscope 2 is a two-photon excitation microscope and the original image of sample 12 is the two-photon excitation image I2 of sample 12. The method for correcting the original image of sample 12 in Embodiment 3 is similar to the method for correcting the original image of sample 12 in Embodiment 2, but differs from the method for correcting the original image of sample 12 in Embodiment 2 mainly in that it does not include the denoising step (S3) of Embodiment 2.

[0082] Specifically, referring to Figure 12, step S1 in this embodiment is the same as step S1 in embodiment 2.

[0083] As shown in Figure 12, after the step of calculating the root image I1' (step S1), the first deconvolution process step (step S2) is performed.

[0084] Specifically, referring to Figure 13, the root-of-power image I1' is Fourier transformed to calculate the spatial frequency spectrum F1 of the root-of-power image I1', as shown in equation (3) of Embodiment 1 (step S2-1). Step S2-1 in this embodiment is the same as step S2-1 in Embodiment 1.

[0085] As shown in equation (22), the spatial frequency spectrum F1 of the root image I1' is deconvolved with the optical transfer function OTF1 of the optical system (illumination optical system 5) to obtain the corrected spatial frequency spectrum F of the root image I1'. 1d ' is calculated (step S2-2). As already described in Embodiment 2, the optical transfer function OTF1 is calculated by performing a Fourier transform on the root of the point image distribution function PSF2 of the optical system (illumination optical system 5) that obtains the original image of the sample 12 (e.g., two-photon excitation image I2). The first deconvolution process includes, for example, a deconvolution process using OTF1 as a Wiener filter. w1 in equation (22) represents the Wiener constant.

number

[0086] The objective of this embodiment is to improve the signal-to-noise ratio (contrast ratio) of the original image of the sample 12 obtained by the microscope 2. Therefore, the first deconvolution process may be performed using the optical transfer function OTF1, or may be performed using the modulation transfer function MTF1 which is the absolute value of the optical transfer function OTF1.

[0087] As shown in FIG. 12, after the first deconvolution process step (step S2), a second deconvolution process step (step S3) is performed.

[0088] Specifically, referring to FIG. 14, as shown in Equation (23), the inverse Fourier transform of the corrected spatial frequency spectrum F 1d ’ of the square root image I1’ is performed to calculate the square root corrected image I 1d ’ (step S3-1).

Equation

[0089] As shown in Equation (24), by raising the square root corrected image I 1d ’ to a power, a provisional corrected image I 2d of the sample 12 is calculated (step S3-2). For example, when the original image of the sample 12 is the two-photon excitation image I2 of the sample 12, the square root corrected image I 1d ’ is squared to calculate the provisional corrected image I 2d of the sample 12. Specifically, by calculating the square of the light intensity of each pixel of the square root corrected image I 1d ’, the light intensity of each pixel of the provisional corrected image I 2d of the sample 12 is calculated. In the provisional corrected image I 2d of the sample 12, noise having a frequency lower than the cut-off frequency of the illumination optical system 5 is reduced by the first deconvolution process step (step S2).

Equation

[0090] As shown in equation (25), the provisional corrected image I of sample 12 2d Performing a Fourier transform on this, we obtain the provisional corrected image I of sample 12. 2d The spatial frequency spectrum F 2d Calculate (Step S3-3).

number

[0091] As shown in equation (26), the provisional corrected image I of sample 12 2d The spatial frequency spectrum F 2d Correction of the optical system's optical transfer function (OTF) 2b By deconvolution, the corrected spatial frequency spectrum F of the original image of sample 12 is obtained. 2e The second deconvolution process is performed, for example, OTF. 2b This includes a deconvolution process using as a Wiener filter. In equation (26), w2 represents the Wiener constant.

number

[0092] The objective of this embodiment is to improve the signal-to-noise ratio (SNR) of the original image of sample 12 obtained with microscope 2. Therefore, the second deconvolution process uses the corrected optical transfer function OTF. 2b This may be done using the corrected optical transfer function OTF. 2b The MTF is the absolute value of the corrected modulation transfer function. 2b It may be done in this way.

[0093] Corrected Optical Transfer Function (OTF) 2b This is the corrected point image distribution function (PSF). 2b It is calculated by performing a Fourier transform. Specifically, the corrected optical transfer function OTF 2b This is given by equation (27). PSF in equation (27) 2bThis represents the corrected point image distribution function obtained by performing a first deconvolution process on the point image distribution function of the illumination optical system 5 that excites sample 12 with two photons.

number

[0094] Corrected point image distribution function (PSF) 2b This is the corrected point image distribution function (PSF). 1b It is calculated by raising to a power. For example, if the original image of sample 12 is the two-photon excitation image I2 of sample 12, then PSF 2b This is given by equation (28).

number

[0095] Corrected point image distribution function (PSF) 1b This is calculated by deconvolving the product of the Fourier transform of the root of the point image distribution function PSF2 of the optical system used to obtain the original image of sample 12 and the high-frequency rejection filter G using the optical transfer function (optical transfer function OTF1), and then performing an inverse Fourier transform. For example, if the original image of sample 12 is the two-photon excitation image I2 of sample 12, then PSF 1b This is given by equation (29).

number

[0096] The objective of this embodiment is to improve the signal-to-noise ratio (SNR) of the original image of sample 12 obtained with microscope 2. Therefore, the corrected point image distribution function PSF is used. 1b To calculate this, the product of the Fourier transform of the root of the point image distribution function PSF2 and the high-frequency rejection filter G may be deconvolved with respect to the optical transfer function OTF1, or with respect to the modulation transfer function MTF1, which is the absolute value of the optical transfer function OTF1.

[0097] Referring to Figure 12, the corrected spatial frequency spectrum F of the original image of sample 12 is given by equation (30).2e The corrected image I of sample 12 is obtained by performing an inverse Fourier transform on it. 2f Calculate (Step S4).

number

[0098] The effects of this embodiment are the same as those of Embodiment 2, as follows:

[0099] The microscope image correction method of this embodiment includes the steps of: calculating a root-of-power image I1' of sample 12 from the original image of sample 12 (for example, two-photon excitation image I2) (step S1); calculating a first spatial frequency spectrum (spatial frequency spectrum F1') of the root-of-power image I1' of sample 12 from the root-of-power image I1' of sample 12 (step S2-1); and deconvolving the first spatial frequency spectrum of the root-of-power image I1' of sample 12 with the optical transfer function (optical transfer function OTF1) of the optical system, which is calculated by Fourier transforming the root of the point image distribution function of the optical system (illumination optical system 5) that obtains the original image of sample 12, thereby obtaining a first corrected spatial frequency spectrum (corrected spatial frequency spectrum F1') of the root-of-power image I1' of sample 12. 1d Step (Step S2-2) to calculate ') and the first corrected spatial frequency spectrum of the root image I1' of sample 12 to obtain the root-corrected image I 1d The steps to calculate ' (step S3-1) and the root-corrected image I of sample 12 1d By raising ' to a power, the provisional corrected image I of sample 12 is obtained. 2d Step S3-2: Calculate the value of the sample 12 and the provisional corrected image I 2d Preliminary corrected image I of sample 12 2d The second spatial frequency spectrum (spatial frequency spectrum F 2d Step S3-3 involves calculating the value of sample 12 and the provisional corrected image I 2d The second spatial frequency spectrum is obtained using the corrected optical transfer function (OTF) of the optical system. 2b By deconvolving with ), the second corrected spatial frequency spectrum (corrected spatial frequency spectrum F) of the original image of sample 12 is obtained.2e The process includes the step of calculating the first corrected point image distribution function (PSF). The corrected optical transfer function is the first corrected point image distribution function (PSF). 2b The first corrected point image distribution function is calculated by performing a Fourier transform on the second corrected point image distribution function (corrected point image distribution function PSF). 1b The second corrected point image distribution function is calculated by raising the second corrected point image distribution function to a power. The second corrected point image distribution function is calculated by deconvolving the product of the Fourier transform of the root of the point image distribution function of the optical system that obtains the original image of sample 12 and the high-frequency rejection filter G with the optical transfer function of the optical system (optical transfer function OTF1), and then performing an inverse Fourier transform. The microscope image correction method of this embodiment calculates the corrected image I of sample 12 from the second corrected spatial frequency spectrum of the original image of sample 12. 2f The process includes a step (step S4) of calculating [something].

[0100] The microscope image correction program of this embodiment causes the processor 20 to execute each step of the microscope image correction method of this embodiment.

[0101] When the light 7 from the illumination optical system 5 is absorbed by the sample 12, the amplitude E of the electric field of the light 7 is squared, which reduces the amplitude component of the spatial frequency component close to the cutoff frequency of the illumination optical system 5 in the spatial frequency spectrum of the original image of the sample 12. As a result, the signal-to-noise ratio (contrast ratio) of the original image of the sample 12 decreases. In this embodiment, since the first deconvolution processing step (step S2) and the second deconvolution processing step (step S3) are performed, the amplitude component of the spatial frequency component close to the cutoff frequency of the illumination optical system 5 can be restored. Corrected image I of sample 12 2f The signal-to-noise ratio (contrast ratio) improves.

[0102] (Examples)

[0103] Referring to Figures 15 to 19, the operation and effects of Embodiments 1 to 3 will be explained while showing Examples 1-1 to 3-2.

[0104] As source images for sample 12, a first source image with a signal-to-noise ratio (SNR) of 10 (see Figure 15) and a second source image with an SNR of 200 were prepared. The first and second source images were created by adding computer-generated Gaussian-distributed noise to computer-generated point images.

[0105] Example 1-1 (see Figure 16) is one embodiment of Embodiment 1, and is a corrected image calculated by denoising the first original image as in Embodiment 1 (see Figure 4). Example 1-2 is another embodiment of Embodiment 1, and is a corrected image calculated by denoising the second original image as in Embodiment 1 (see Figure 4).

[0106] Example 2-1 (see Figure 17) is one embodiment of Embodiment 2, and is a corrected image calculated by denoising, first deconvolution processing and second deconvolution processing (see Figure 9) of the first original image as in Embodiment 2. Example 2-2 is another embodiment of Embodiment 2, and is a corrected image calculated by denoising, first deconvolution processing and second deconvolution processing (see Figure 9) of the second original image as in Embodiment 2.

[0107] Example 3-1 (see Figure 18) is one embodiment of Embodiment 3, and is a corrected image calculated by performing a first deconvolution process and a second deconvolution process (see Figure 12) on the first original image as in Embodiment 3. Example 3-2 is another embodiment of Embodiment 3, and is a corrected image calculated by performing a first deconvolution process and a second deconvolution process (see Figure 12) on the second original image as in Embodiment 3.

[0108] Figure 19 shows the improvement rates of the signal-to-noise ratio (SNR) of the images from Examples 1-1 to 3-2. The improvement rate of the SNR of the image in Example 1-1 is calculated by dividing the SNR of the corrected image in Example 1-1 by the SNR of the first original image (i.e., 10). The improvement rate of the SNR of the image in Example 1-2 is calculated by dividing the SNR of the corrected image in Example 1-2 by the SNR of the second original image (i.e., 200). The improvement rate of the SNR of the image in Example 2-1 is calculated by dividing the SNR of the corrected image in Example 2-1 by the SNR of the first original image. The improvement rate of the SNR of the image in Example 2-2 is calculated by dividing the SNR of the corrected image in Example 2-2 by the SNR of the second original image. The improvement rate of the SNR of the image in Example 3-1 is calculated by dividing the SNR of the corrected image in Example 3-1 by the SNR of the first original image. The improvement rate of the SNR of the image in Example 3-2 is calculated by dividing the SNR of the corrected image in Example 3-2 by the SNR of the second original image.

[0109] As shown in the signal-to-noise ratio improvement rates of Examples 1-1 and 1-2, the signal-to-noise ratio of the corrected image is improved by applying the denoising process of Embodiment 1 (see Figure 4) to the original image. As shown in the signal-to-noise ratio improvement rates of Examples 3-1 and 3-2, the signal-to-noise ratio of the corrected image is significantly improved by applying the first deconvolution process and the second deconvolution process of Embodiment 3 (see Figure 12) to the original image. As shown in the signal-to-noise ratio improvement rates of Examples 2-1 and 2-2, the signal-to-noise ratio of the corrected image is further improved by applying the denoising process, the first deconvolution process, and the second deconvolution process of Embodiment 2 (see Figure 9) to the original image.

[0110] Embodiments 1 to 3 and their variations disclosed herein should be considered in all respects to be illustrative and not restrictive. For example, microscope 2 may be a multiphoton excitation microscope other than a two-photon excitation microscope, a confocal microscope, or a super-resolution microscope. The original image of sample 12 may be obtained by a multiphoton excitation microscope other than a two-photon excitation microscope, a confocal microscope, or a super-resolution microscope. The scope of this disclosure is indicated by the claims rather than the above description and is intended to include all modifications in the meaning and scope equivalent to the claims. [Explanation of symbols]

[0111] 1 Microscope system, 2 Microscope, 3 Computer, 4 Display device, 5 Illumination optics, 6 Light source, 7 Light, 8 Scanning mirror, 9a, 9b, 9c, 9d Lenses, 10 Dichroic mirror, 11 Objective lens, 12 Sample, 13 Sample stage, 14 Light, 15 Observation optics, 16 Photodetector, 20 Processor, 22 Storage device, 25 Original image acquisition unit, 30 Image processing unit, 31 Root power image calculation unit, 32 Denoising unit, 33 First deconvolution processing unit, 34 Second deconvolution processing unit, 35 Corrected image calculation unit.

Claims

1. A step of calculating the root image of the sample from the original image of the sample, The steps include: calculating the spatial frequency spectrum of the root image of the sample from the root image of the sample; The steps include: applying a high-frequency rejection filter to the spatial frequency spectrum of the root image of the sample to calculate the corrected spatial frequency spectrum of the root image of the sample; The steps include: calculating the corrected root image of the sample from the corrected spatial frequency spectrum of the root image of the sample; A microscope image correction method comprising the step of calculating a corrected image of a sample from the corrected root image of the sample.

2. A step of calculating the root image of the sample from the original image of the sample, The steps include: calculating the first spatial frequency spectrum of the root image of the sample from the root image of the sample; The steps include: applying a high-frequency rejection filter to the first spatial frequency spectrum of the root image of the sample to calculate the first corrected spatial frequency spectrum of the root image of the sample; The steps include: calculating a second corrected spatial frequency spectrum of the root image of the sample by deconvolving the first corrected spatial frequency spectrum of the root image of the sample with the optical transfer function of the optical system, which is calculated by performing a Fourier transform on the root of the point image distribution function of the optical system used to obtain the original image of the sample; The steps include: calculating the corrected root image of the sample from the second corrected spatial frequency spectrum of the root image of the sample; A step of calculating a provisional corrected image of the sample by raising the corrected root image of the sample to a power, The steps include: calculating the second spatial frequency spectrum of the provisionally corrected image of the sample from the provisionally corrected image of the sample; The process includes the step of calculating a third corrected spatial frequency spectrum of the original image of the sample by deconvolving the second spatial frequency spectrum of the provisionally corrected image of the sample with the corrected optical transfer function of the optical system, wherein the corrected optical transfer function is calculated by performing a Fourier transform on a first corrected point image distribution function, the first corrected point image distribution function is calculated by raising the second corrected point image distribution function to a power, the second corrected point image distribution function is calculated by deconvolving the product of the Fourier transform of the power root of the point image distribution function of the optical system used to obtain the original image of the sample and the high-frequency rejection filter with the optical transfer function, and then performing an inverse Fourier transform. A microscope image correction method comprising the step of calculating a corrected image of a sample from the third corrected spatial frequency spectrum of the original image of the sample.

3. A step of calculating the root image of the sample from the original image of the sample, The steps include: calculating the first spatial frequency spectrum of the root image of the sample from the root image of the sample; The steps include: calculating the first corrected spatial frequency spectrum of the root image of the sample by deconvolving the first spatial frequency spectrum of the root image of the sample with the optical transfer function of the optical system, which is calculated by performing a Fourier transform on the root of the point image distribution function of the optical system used to obtain the original image of the sample; The steps include: calculating the corrected root image of the sample from the first corrected spatial frequency spectrum of the root image of the sample; A step of calculating a provisional corrected image of the sample by raising the corrected root image of the sample to a power, The steps include: calculating the second spatial frequency spectrum of the provisionally corrected image of the sample from the provisionally corrected image of the sample; The process comprises the steps of calculating the second corrected spatial frequency spectrum of the original image of the sample by deconvolving the second spatial frequency spectrum of the provisionally corrected image of the sample with the corrected optical transfer function of the optical system, wherein the corrected optical transfer function is calculated by performing a Fourier transform on a first corrected point image distribution function, the first corrected point image distribution function is calculated by raising the second corrected point image distribution function to a power, and the second corrected point image distribution function is calculated by deconvolving the product of the Fourier transform of the power root of the point image distribution function of the optical system used to obtain the original image of the sample and a high-frequency rejection filter with the optical transfer function, and then performing an inverse Fourier transform. A microscope image correction method comprising the step of calculating a corrected image of a sample from the second corrected spatial frequency spectrum of the original image of the sample.

4. A microscope image correction program that causes a processor to perform each step of the microscope image correction method described in any one of claims 1 to 3.

Images