Illumination light field correction method and system for quantitative fluorescence resonance energy transfer microscopic imaging

Abstract

The invention relates to an illumination light field correction method and system for quantitative fluorescence resonance energy transfer microscopic imaging. The method comprises the following steps: S1, establishing an image non-uniformity correction model; s2, measuring a dark field signal distribution function; s3, measuring a background signal distribution function and a background correction coefficient; s4, measuring a foreground signal distribution function and a foreground correction coefficient; s5, repeating the steps S2-S4 to respectively obtain dark field, background and foreground signal distribution functions and background and foreground correction coefficients of the FRET three channels; and S6, based on an image non-uniformity correction model, correcting the non-uniformity of the FRET three-channel fluorescence image of the living cell by using the measured signal distribution function and the correction coefficient. According to the method, reference samples do not need to be prepared repeatedly, illumination non-uniformity in a microscopic imaging system can be effectively corrected only through one-time measurement, the proportion of FRET three-channel image gray levels can be kept, and therefore the accuracy of quantitative FRET measurement is guaranteed while the available data amount in a single-frame large-view image is greatly increased.

Description

Technical Field

[0001] The present invention relates to the technical field of image correction of optical microscopic imaging technology, and in particular to an illumination light field correction method and system for quantitative fluorescence resonance energy transfer (FRET) microscopic imaging. Background Art

[0002] Fluorescence microscopy is an imaging technique widely used in biomedical research, which can realize the visualization of the internal structure and function of cells. In the field of fluorescence microscopy, quantitative FRET microscopy is currently the only imaging technology that can monitor weak, reversible and dynamic molecular events in living cells in real time.

[0003] Accurate measurement of fluorescence intensity is essential for obtaining reliable quantitative FRET analysis results. However, due to the uneven illumination of the optical system itself, as well as the interference of background fluorescence and camera noise, the phenomenon of uneven brightness and darkness occurs during imaging, which seriously affects the accuracy of quantitative FRET analysis.

[0004] In response to this problem, the prior art has proposed a variety of illumination correction methods. The most common one is the forward-looking method using a reference image (Model MA. Intensity calibration and shading correction for fluorescence microscopes. Curr Protoc Cytom. 2006 Aug; Chapter 10: Unit 10. 14.). The forward-looking method requires taking a pure white or uniform grayscale reference image in advance, and estimating the illumination distribution by measuring the intensity distribution of the reference image for subsequent correction. This method is relatively simple to operate, but it requires the preparation of a reference image in advance. The reference image is generally collected by selecting a fluorescent solution as a sample, but the thickness of the fluorescent solution is not consistent with the thickness of the actual sample to be calibrated, and the acquisition conditions for obtaining the reference image are not completely consistent with the conditions for imaging the actual sample to be calibrated. All of the above limit the use of this method.

[0005] The other is a retrospective correction method based on multiple images. Although this method does not require a reference standard, it has certain requirements on the number of input training images. When the number is low, it is difficult to ensure the correction effect. Moreover, for multi-channel fluorescence imaging, this correction method starts from the image itself, and it is difficult to ensure the proportional relationship between the grayscale of multi-channel images. Therefore, it is urgent to develop a new illumination correction technology to improve the accuracy and reliability of quantitative analysis in fluorescence microscopy. Summary of the invention

[0006] In order to solve the problems existing in the prior art, the present invention provides an illumination light field correction method and system for quantitative fluorescence resonance energy transfer microscopy imaging, establishes a correction model including three factors causing image unevenness, namely, uneven illumination light source, uneven background fluorescence and inconsistent proportional relationship between multi-channel image grayscale, and proposes corresponding model parameter determination and correction methods.

[0007] The method of the present invention is implemented by the following technical scheme: A method for calibrating the illumination light field for quantitative fluorescence resonance energy transfer microscopy imaging, comprising the following steps:

[0008] S1. Establish an image non-uniformity correction model, divide the image into foreground signal, background signal and dark field signal, measure the distribution of different signals, correct their non-uniformity, and thus obtain an image with uniform illumination;

[0009] S2, measuring the dark field signal distribution function, turning off the illumination source of the microscopic imaging system and shielding the interference of the external ambient light, continuously collecting multiple frames of images at the minimum exposure time of the camera, and averaging the collected images to obtain the dark field signal distribution function;

[0010] S3, determining the background signal distribution function and the background correction coefficient, taking the culture medium without cells as the sample, continuously collecting multiple frames of images under the same imaging conditions as the cells, averaging the collected images, and then subtracting the dark field signal distribution function obtained in step S2 to obtain the background signal distribution function, and calculating the background correction coefficient pixel by pixel based on the obtained background signal distribution function;

[0011] S4, determining a foreground signal distribution function and a foreground correction coefficient, taking a fluorescent calibration slide as a sample, continuously acquiring multiple frames of images without overexposure, averaging the acquired images, and then subtracting a dark field signal distribution function and a background signal distribution function to obtain a foreground signal distribution function, and calculating a foreground correction coefficient pixel by pixel based on the acquired foreground signal distribution function;

[0012] S5, repeating steps S2-S4 to obtain the dark field, background and foreground signal distribution functions of the three FRET channels and the background and foreground correction coefficients respectively;

[0013] S6. Based on the image non-uniformity correction model, the non-uniformity of the living cell FRET three-channel fluorescence image is corrected using the measured signal distribution function and correction coefficient; based on the image non-uniformity correction model established in step S1, according to the FRET three-channel background signal distribution function and background correction coefficient obtained in step S5, the average grayscale of the background uniform area is used as the ideal grayscale, and the living cell FRET three-channel fluorescence image is subjected to pixel-by-pixel background correction; according to the foreground signal distribution function and foreground correction coefficient obtained in step S5, the average grayscale of the foreground uniform area is used as the ideal grayscale, and the living cell FRET three-channel fluorescence image is subjected to pixel-by-pixel foreground correction, and finally the corrected uniform FRET three-channel fluorescence image is output.

[0014] The system of the present invention is implemented by the following technical scheme: an illumination light field correction system for quantitative fluorescence resonance energy transfer microscopy imaging, comprising:

[0015] Image non-uniformity correction model building module, used to divide the image into foreground signal, background signal and dark field signal, and to obtain an image with uniform illumination by measuring the distribution of different signals and correcting their non-uniformity;

[0016] The dark field signal distribution function determination module turns off the illumination source of the microscopic imaging system and shields the interference of the external ambient light, continuously collects multiple frames of images at the minimum exposure time of the camera, and averages the collected images to obtain the dark field signal distribution function;

[0017] A background signal distribution function and background correction coefficient determination module, which uses a cell-free culture medium as a sample, continuously collects multiple frames of images under the same imaging conditions as the cells, averages the collected images, and then subtracts the dark field signal distribution function obtained by the dark field signal distribution function determination module to obtain a background signal distribution function, and calculates the background correction coefficient pixel by pixel based on the obtained background signal distribution function;

[0018] The foreground signal distribution function and foreground correction coefficient determination module uses a fluorescent calibration slide as a sample, continuously collects multiple frames of images without overexposure, averages the collected images, and then subtracts the dark field signal distribution function and the background signal distribution function to obtain the foreground signal distribution function, and calculates the background correction coefficient pixel by pixel based on the obtained foreground signal distribution function;

[0019] The FRET three-channel measurement module repeats the dark field signal distribution function measurement module, the background signal distribution function and background correction coefficient measurement module, and the foreground signal distribution function and foreground correction coefficient measurement module to obtain the dark field, background and foreground signal distribution functions and background and foreground correction coefficients of the FRET three-channel respectively;

[0020] The non-uniformity correction module of the living cell FRET three-channel fluorescence image is used to perform pixel-by-pixel background correction on the living cell FRET three-channel fluorescence image according to the FRET three-channel background signal distribution function and background correction coefficient obtained by the FRET three-channel measurement module through the established image non-uniformity correction model, taking the average grayscale of the background uniform area as the ideal grayscale, and performing pixel-by-pixel foreground correction on the living cell FRET three-channel fluorescence image according to the FRET three-channel foreground signal distribution function and foreground correction coefficient obtained by the FRET three-channel measurement module, taking the average grayscale of the foreground uniform area as the ideal grayscale, and finally outputting the corrected uniform FRET three-channel fluorescence image.

[0021] Compared with the prior art, the present invention has the following advantages and beneficial effects:

[0022] 1. The present invention can effectively correct the illumination non-uniformity in the microscopic imaging system and the inconsistency in the ratio between the grayscale of multi-channel images, thereby significantly improving the accuracy and reliability of quantitative FRET measurement.

[0023] 2. The present invention uses a fluorescent calibration slide to perform illumination calibration, and only requires one measurement of the system's calibration parameters, which simplifies the calibration process and eliminates the cumbersome operations of the traditional forward-looking calibration method.

[0024] 3. The present invention uses a fluorescent calibration slide for illumination calibration. Within a certain focal range, the foreground signal distribution function at different focal planes is the same, so there is no need to adjust the fluorescent calibration slide to the focal plane corresponding to the cell sample to measure the calibration parameters.

[0025] 4. The present invention uses a fluorescent calibration slide for illumination correction, and its foreground signal distribution function under different imaging conditions can be linearly converted, so even if the imaging conditions and cells are different, the foreground correction coefficient can be accurately determined.

[0026] 5. The present invention uses a fluorescent calibration slide for illumination correction. The fluorescent medium thereof is more uniform than the fluorescent medium of the fluorescent solution in the traditional prospective correction method, thereby eliminating the influence of the uneven concentration of the fluorescent sample itself on the determination of the foreground correction coefficient.

[0027] 6. The present invention uses the same substrate solution as the cell sample, and can approximately correct the background fluorescence inhomogeneity caused by the substrate solution in the cell sample, so that it can be applied to samples in various different substrate solutions and can effectively correct the imaging inhomogeneity caused by the background.

[0028] 7. The present invention performs maximum value normalization correction on the fluorescence signal of the measured area based on the established illumination correction model; this normalization processing based on the physical model can effectively correct the measurement error caused by uneven illumination; taking the MCF7 cell sample as an example, when not corrected, the error of fluorescence intensity in different areas of the sample reaches more than 50%, while after correction by the method of the present invention, the error is reduced to less than 10%, thereby greatly improving the amount of available data in a single-frame large-field-of-view image while ensuring the accuracy of quantitative FRET measurement.

[0029] 8. The illumination correction method provided by the present invention has the advantages of simple operation, wide applicability, high correction accuracy, etc. It can effectively solve the deficiencies in the prior art and provide a reliable solution for fluorescence quantitative analysis in the biomedical field, especially quantitative FRET imaging. BRIEF DESCRIPTION OF THE DRAWINGS

[0030] Figure 1 is a flow chart of the method of the present invention;

[0031] Figure 2 is a schematic diagram of the optical path of the quantitative FRET microscopy imaging system;

[0032] Figure 3 is a flow chart for determining background signal distribution function and background correction coefficient;

[0033] Figure 4 (a) is the imaging spectrum of the fluorescence calibration slide obtained in the AA detection channel;

[0034] Figure 4 (b) is the imaging spectrum of the fluorescence calibration slide obtained in the DA detection channel;

[0035] Figure 4 (c) is the imaging spectrum of the fluorescence calibration slide obtained in the DD detection channel;

[0036] Figure 5 is a linear relationship graph of foreground correction coefficient and optical power;

[0037] Figure 6 is a linear plot of the foreground correction factor versus exposure time;

[0038] Figure 7 It is the Quantile-Quantile diagram of foreground correction coefficient and focal position;

[0039] Figure 8 is a flow chart for determining a foreground signal distribution function and a foreground correction coefficient;

[0040] Fig. 9It is a flow chart for determining the dark field, background and foreground signal distribution functions and background and foreground correction coefficients of the three channels of FERT;

[0041] Fig.10 This is a comparison of cell fluorescence images before and after illumination correction;

[0042] Fig.11 (a) is a box plot of the quantitative FRET analysis before and after illumination correction of the system correction factor G;

[0043] Fig.11 (b) is a box plot of the k factor before and after illumination correction for quantitative FRET analysis;

[0044] Fig.11 (c) is the efficiency E D Box plots of quantitative FRET analysis before and after illumination correction;

[0045] Fig.11 (d) is the acceptor-donor concentration ratio R C Box plots of quantitative FRET analysis before and after illumination correction;

[0046] Fig.12 (a) is the error bar graph of the quantitative FRET analysis before and after illumination correction of the system correction factor G;

[0047] Fig.12 (b) is the error bar graph of the k factor before and after illumination correction for quantitative FRET analysis;

[0048] Fig.12 (c) is the efficiency E D Error bars of quantitative FRET analysis before and after illumination correction;

[0049] Fig.12 (d) is the acceptor-donor concentration ratio R C Error bar plot of quantitative FRET analysis before and after illumination correction. DETAILED DESCRIPTION

[0050] The present invention is further described in detail below in conjunction with embodiments and drawings, but the embodiments of the present invention are not limited thereto.

[0051] Example

[0052] like Figure 1 As shown, this embodiment provides an illumination light field correction method for quantitative fluorescence resonance energy transfer microscopy imaging, comprising the following steps:

[0053] S1. Establish an image non-uniformity correction model, divide the image into foreground signal, background signal and dark field signal, measure the distribution of different signals, correct their non-uniformity, and thus obtain an image with uniform illumination;

[0054] S2, measuring the dark field signal distribution function, turning off the illumination source of the microscopic imaging system and shielding the interference of the external ambient light, continuously collecting multiple frames of images at the minimum exposure time of the camera, and averaging the collected images to obtain the dark field signal distribution function;

[0055] S3, determining the background signal distribution function and the background correction coefficient, taking the culture medium without cells as the sample, continuously collecting multiple frames of images under the same imaging conditions as the cells, averaging the collected images, and then subtracting the dark field signal distribution function obtained in step S2 to obtain the background signal distribution function, and calculating the background correction coefficient pixel by pixel based on the obtained background signal distribution function;

[0056] S4, determining a foreground signal distribution function and a foreground correction coefficient, taking a fluorescent calibration slide as a sample, continuously acquiring multiple frames of images without overexposure, averaging the acquired images, and then subtracting a dark field signal distribution function and a background signal distribution function to obtain a foreground signal distribution function, and calculating a foreground correction coefficient pixel by pixel based on the acquired foreground signal distribution function;

[0057] S5, repeating steps S2-S4 to obtain the dark field, background and foreground signal distribution functions of the three FRET channels and the background and foreground correction coefficients respectively;

[0058] S6. Based on the image non-uniformity correction model, the non-uniformity of the living cell FRET three-channel fluorescence image is corrected using the measured signal distribution function and correction coefficient; based on the image non-uniformity correction model established in step S1, according to the FRET three-channel background signal distribution function and background correction coefficient obtained in step S5, the average grayscale of the background uniform area is used as the ideal grayscale, and the living cell FRET three-channel fluorescence image is subjected to pixel-by-pixel background correction; according to the foreground signal distribution function and foreground correction coefficient obtained in step S5, the average grayscale of the foreground uniform area is used as the ideal grayscale, and the living cell FRET three-channel fluorescence image is subjected to pixel-by-pixel foreground correction, and finally the corrected uniform FRET three-channel fluorescence image is output.

[0059] like Figure 2As shown in the figure, it is a schematic diagram of the optical path of the quantitative FRET microscopy imaging system. In the figure, the LED light source is refracted by the collimator, reflector and dichroic mirror, and irradiates the sample through a part of the objective lens. The emitted light of the sample is amplified by the objective lens and refracted by the dichroic mirror and reflector, and then forms an intermediate image through the tube lens. The intermediate image is divided into two channels by a relay lens, a dichroic mirror and a filter, and finally imaged by a camera sensor. The factors that cause image non-uniformity include three aspects: first, the light source irradiated to the surface of the sample tends to attenuate from the middle to the surroundings; second, the non-uniformity of the spontaneous fluorescence of the sample background; and third, the difference in the proportional change of the image grayscale between each channel. Taking the above three factors into full consideration, step S1 establishes the image non-uniformity correction model as follows:

[0060]

[0061] F cell (x,y)=I cell (x,y)-B(x,y)-D(x,y)

[0062] in, is the corrected cell gray value, FCF(x,y) is the foreground correction coefficient, F cell (x, y) is the foreground signal distribution function of the cell fluorescence image, I cell (x, y) is the measured gray value of the cell fluorescence image, BCF(x, y) is the background correction coefficient, B(x, y) is the background signal distribution function of the cell fluorescence image, and D(x, y) is the dark field signal distribution function.

[0063] In this embodiment, the specific process of determining the dark field signal distribution function in step S2 is as follows:

[0064]

[0065] Where D(x,y) is the dark field signal distribution function, I D(i) (x, y) is the i-th frame of the image acquired continuously by setting the camera's minimum exposure time with the illumination source of the microscopic imaging system turned off and other ambient light shielded, and n is the total number of image frames acquired continuously.

[0066] The imaging process of a digital camera can be viewed as a mapping of the light reflected (or illuminated) by an object to the grayscale value of the image; this process includes the light entering the lens, the image sensor converting the incident photons during the exposure time into electrons to form an analog signal, the output amplifier forming a voltage signal, and then the analog-to-digital converter converting it into a digital signal, and finally mapping it to the grayscale value of the image. However, each pixel has an inherent dark current bias value, which will randomly fluctuate between different pixels, forming a dark field signal distribution. Therefore, measuring the dark field signal distribution function in a dark environment is to calibrate the initial grayscale bias of each pixel and a part of the noise signal.

[0067] In this embodiment, Figure 3 As shown, the specific process of step S3 for determining the background signal distribution function and the background correction coefficient is as follows:

[0068] S301, determining the background signal distribution function B(x, y); taking the culture medium without cells as the sample, continuously collecting multiple frames of images under the same imaging conditions as the cells, averaging the collected images, and then subtracting the dark field signal distribution function obtained in step S2 to obtain the background signal distribution function:

[0069]

[0070] Among them, B(x,y) is the background signal distribution function, I B(i) (x, y) is the i-th frame of image acquired continuously under the same imaging conditions as the cells by turning on the illumination source of the microscopic imaging system, D(x, y) is the dark field signal distribution function obtained in step S2, and n is the total number of image frames acquired continuously;

[0071] S302, obtain the relative background center point (x B ,y B ); The collected background signal distribution function is processed by median filtering. The neighborhood window size selected by median filtering is 3pixel*3pixel, and a pixel window of 100pixel*100pixel is selected. The pixel window is traversed in turn to calculate its average gray value, and the center point of the pixel window with the largest average gray value is selected as the relative background center point, which is recorded as (x B ,y B );

[0072] S303, determining the background correction factor BCF (Background Correction Factor); calculating the background correction factor BCF (x, y) pixel by pixel for the collected image based on the acquired background signal distribution function, and the calculation formula is:

[0073]

[0074] Among them, BCF(x,y) is the background correction coefficient, B(x,y) is the background signal distribution function, and B(x B ,y B ) is the intensity relative to the background center point.

[0075] In this embodiment, when determining the foreground signal distribution function and the foreground correction coefficient in step S4, it is necessary to turn on the illumination light source of the microscope imaging system and collect an image of the fluorescence calibration slide. The fluorescence calibration slide usually has a variety of different colors, which correspond to different imaging channels. In order to ensure accurate calibration of each imaging channel, the fluorescence calibration slide used should be consistent with the spectral composition of the sample to be calibrated. In other words, different imaging channels should be calibrated using fluorescence calibration slides of different colors, such as Figure 4 (a) Figure 4 (b) Figure 4 (c). Among them, Figure 4 (a) is the imaging spectrum obtained by the AA detection channel, which shows that the spectral components of the fluorescence calibration slides of ORANGE and RED colors are consistent with the AA channel of actual cell imaging. Therefore, the AA channel can be corrected by selecting the ORANGE and RED fluorescence calibration slides; Figure 4 (b) is the imaging spectrum obtained by the DA detection channel, which shows that the spectral components of the ORANGE and YELLOW color fluorescence calibration slides are more consistent with the DA channel of actual cell imaging. Therefore, the DA channel can be corrected by selecting the ORANGE and YELLOW fluorescence calibration slides; Figure 4 (c) is the imaging spectrum obtained by the DD detection channel, which shows that the spectral components of the fluorescence calibration slides of ORANGE, YELLOW and GREEN colors are more consistent with the DD channel of actual cell imaging. Therefore, the DD channel can be corrected by selecting the YELLOW and GREEN fluorescence calibration slides.

[0076] In fluorescence microscopy, the fluorescence intensity is related to the excitation light power I ex , exposure time t, fluorescence quantum yield Q, extinction coefficient ε, etc. are linearly related as shown below:

[0077] I(x,y)=K(x,y)·I ex εQL1L2ct+D(x,y)

[0078] Among them, I(x,y) is the gray value of the image, I ex is the excitation light power, ε is the extinction coefficient, L1 and L2 are the transmittance of laser and emission light paths respectively, t is the exposure time, Q is the fluorescence quantum yield, c is the solubility of the fluorescent solution, K(x,y) is a constant related to the measurement system and characterizes the inhomogeneity of the fluorescence image.

[0079] After experimental verification, the foreground correction coefficient is mainly related to the field of view corresponding to different positions, and has a certain linear conversion relationship with the light power and exposure time, such as Figure 5 and Figure 6 As shown. Among them, Figure 5The linear change of a single point as the optical power increases and the linear conversion relationship between different points are demonstrated; Figure 6 The linear change of a single point with increasing exposure time and the linear conversion relationship between different points are demonstrated.

[0080] At the same time, for linear imaging systems, changing the focal position within a certain range will not significantly affect the determination of the foreground correction coefficient, such as Figure 7 shown. Figure 7 The Quantile-Quantile plot between the fluorescence calibration slide image with a fixed focal plane and the fluorescence calibration slide image with different focal planes shows that the quantile points between different focal planes are basically maintained on the same fitting straight line. This shows that within a certain range of focal plane changes, the measurement results of the foreground correction coefficient remain stable. In summary, changing the light intensity, exposure time, and focal position within a certain range will not have a significant impact on the measurement of the foreground correction coefficient.

[0081] Specifically, Figure 8 As shown, the specific process of step S4 is as follows:

[0082] S401, determining the foreground signal distribution function F(x, y); taking the fluorescent calibration slide as a sample, continuously collecting multiple frames of images without overexposure, averaging the collected images, and then subtracting the dark field signal distribution function and the background signal distribution function to obtain the foreground signal distribution function. The specific process is as follows:

[0083]

[0084] Among them, F(x,y) is the foreground signal distribution function, I F(i) (x, y) is the i-th frame image acquired continuously without overexposure by turning on the illumination source of the microscopic imaging system and taking the fluorescent calibration slide as the sample, B(x, y) is the background signal distribution function obtained in step S3, D(x, y) is the dark field signal distribution function obtained in step S2, and n is the total number of image frames acquired continuously;

[0085] S402, obtain the relative foreground center point (x F ,y F ); The collected foreground distribution function is processed by median filtering. The neighborhood window size selected by the median filter is 3pixel*3pixel, and a pixel window of 100pixel*100pixel is selected. The pixel window is traversed in turn to calculate its average gray value, and the center point of the pixel window with the largest average gray value is selected as the relative foreground center point, which is recorded as (x F ,y F );

[0086] S403, determining the foreground correction factor FCF (Foreground Correction Factor); calculating the foreground correction factor FCF (x, y) pixel by pixel for the FRET three-channel image based on the acquired foreground signal distribution function, and the calculation formula is:

[0087]

[0088] Among them, FCF(x,y) is the foreground correction coefficient, F(x,y) is the foreground signal distribution function, and F(x F ,y F ) is the intensity relative to the foreground center point.

[0089] In this embodiment, Fig. 9 As shown, the specific process of step S5 is as follows:

[0090] S501, repeat step S2 to measure the dark field signal distribution function of the three FRET channels:

[0091]

[0092]

[0093]

[0094] Among them, D AA (x,y),D DA (x,y),D DD (x, y) are the dark field signal distribution functions corresponding to the three FRET imaging channels respectively. The i-th frame of image is collected continuously by turning off the illumination source of the microscopic imaging system and shielding other ambient light for the three FRET imaging channels respectively, and n is the total number of image frames collected continuously by the three FRET imaging channels respectively;

[0095] S502, repeat step S3 to determine the background signal distribution function and background correction coefficient of the three FRET channels;

[0096]

[0097]

[0098]

[0099]

[0100]

[0101]

[0102] Among them, B AA (x,y),B DA (x,y),B DD (x, y) are the background signal distribution functions corresponding to the three FRET imaging channels, BCF AA (x,y), BCF DA (x,y), BCF DD (x, y) are the background correction coefficients corresponding to the three FRET channels respectively. The i-th frame image is acquired by turning on the illumination source of the microscopic imaging system for the three FRET imaging channels and continuously acquiring the image under the same imaging conditions as the cells. AA (x,y),D DA (x,y),D DD (x, y) is the dark field signal distribution function of the three channels of FRET obtained in step S501, n is the total number of image frames continuously collected by the three FRET imaging channels, They are the intensities of the relative background center points corresponding to the three FRET imaging channels;

[0103] S503, repeat step S4 to determine the foreground signal distribution function and foreground correction coefficient of the three FRET channels:

[0104]

[0105]

[0106]

[0107]

[0108]

[0109]

[0110] Among them, F AA (x,y),F DA (x,y),F DD (x, y) are the foreground signal distribution functions corresponding to the three FRET imaging channels, FCF AA (x,y), FCF DA (x,y), FCF DD (x, y) are the foreground correction coefficients corresponding to the three FRET channels respectively. The illumination source of the microscope imaging system is turned on for the three FRET imaging channels respectively, and the fluorescence calibration slide is used as a sample. The i-th frame image is continuously collected without overexposure. AA (x,y),BDA (x,y),B DD (x, y) is the FRET three-channel background signal distribution function D obtained in step S502 AA (x,y),D DA (x, y)D DD (x, is the FRET three-channel dark field signal distribution function obtained in step S501, n is the total number of image frames continuously collected by the three FRET imaging channels, They are the relative intensities of the foreground center point corresponding to the three FRET imaging channels.

[0111] In this embodiment, the specific process of step S6 is as follows:

[0112]

[0113]

[0114]

[0115]

[0116]

[0117]

[0118] in, The grayscale values ​​of the cell fluorescence images of the three channels after FRET correction, FCF AA (x,y), FCF DA (x,y), FCF DD (x, y) are the foreground correction coefficients corresponding to the three FRET channels respectively, are the foreground signal distribution functions of the cell fluorescence image in the three FRET channels, The measured gray values ​​of the cell fluorescence images of the three channels of FRET, BCF AA (x,y), BCF DA (x,y), BCF DD (x, y) are the background correction coefficients corresponding to the three FRET channels, B AA (x,y),B DA (x,y),B DD (x, y) are the background signal distribution functions corresponding to the three FRET imaging channels, respectively. AA (x,y),D DA (x,y),D DD (x, y) are the dark field signal distribution functions corresponding to the three FRET imaging channels respectively.

[0119] After the illumination correction, the images collected by the quantitative FRET microscopy system have good brightness, contrast and uniformity. Taking MCF7 cells as an example, a single cell is selected and moved along the diagonal of the image sensor to measure the change in the grayscale level of the cell. Before correction, there is a significant difference in the grayscale distribution of the cell along the diagonal, but after correction, the grayscale distribution level of the cell is significantly improved, such as Fig.10 In addition, under the above imaging conditions, the photobleaching level was maintained below 5% within five minutes, which basically eliminated the influence of photobleaching.

[0120] These results indicate that illumination correction technology can effectively correct non-uniform factors in microscopic imaging systems and provide high-quality, uniform image data for subsequent quantitative FRET analysis, which is crucial to improving the accuracy and reliability of FRET quantitative analysis.

[0121] FRET technology is a method of indirectly detecting molecular interactions by measuring changes in fluorescence intensity. In FRET, the calculation of the fluorescent sample image first requires background subtraction. The background is usually selected by circling the area near the ROI in the fluorescent sample image as the background value for subtraction, but this operation is cumbersome and has high requirements for the circles; then the calculation is performed based on the image after background subtraction, so the background non-uniformity seriously affects the calculation accuracy of quantitative FRET. Due to the uneven illumination, the fluorescence intensity of the ROI that is not in the center of the field of view is reduced, which greatly affects the amount of available data and SBR (Signal Background Ratio) in a single-frame large field of view image. Therefore, the present invention corrects the background and foreground non-uniformity, thereby simplifying the background subtraction process, while increasing the amount of available data and SBR while ensuring the accuracy of quantitative FRET measurement. Taking the MCF7 cell standard plasmid C4Y as an example, Fig.11 (a) Fig.11 (b) Fig.11 (c) Fig.11 (d) Fig.12 (a) Fig.12 (b) Fig.12 (c) Fig.12 As shown in (d), after the illumination correction process, the standard deviation of the system correction factor G is reduced from the original 0.827 to 0.278, and the standard deviation of the k factor is reduced from the original 0.129 to 0.112. D The standard deviation of the original 0.022 was reduced to 0.007, and the acceptor-donor concentration ratio R C The standard deviation of the original FRET was reduced from 0.19 to 0.165. The quantitative FRET results were effectively improved, further demonstrating the important role of illumination correction technology in improving the accuracy of FRET quantification.

[0122] Based on the same inventive concept, the present invention also proposes an illumination light field correction system for quantitative fluorescence resonance energy transfer microscopy imaging, comprising:

[0123] Image non-uniformity correction model building module, used to divide the image into foreground signal, background signal and dark field signal, and to obtain an image with uniform illumination by measuring the distribution of different signals and correcting their non-uniformity;

[0124] The dark field signal distribution function determination module turns off the illumination source of the microscopic imaging system and shields the interference of the external ambient light, continuously collects multiple frames of images at the minimum exposure time of the camera, and averages the collected images to obtain the dark field signal distribution function;

[0125] A background signal distribution function and background correction coefficient determination module, which uses a cell-free culture medium as a sample, continuously collects multiple frames of images under the same imaging conditions as the cells, averages the collected images, and then subtracts the dark field signal distribution function obtained by the dark field signal distribution function determination module to obtain a background signal distribution function, and calculates the background correction coefficient pixel by pixel based on the obtained background signal distribution function;

[0126] The foreground signal distribution function and foreground correction coefficient determination module uses a fluorescent calibration slide as a sample, continuously collects multiple frames of images without overexposure, averages the collected images, and then subtracts the dark field signal distribution function and the background signal distribution function to obtain the foreground signal distribution function, and calculates the background correction coefficient pixel by pixel based on the obtained foreground signal distribution function;

[0127] The FRET three-channel measurement module repeats the dark field signal distribution function measurement module, the background signal distribution function and background correction coefficient measurement module, and the foreground signal distribution function and foreground correction coefficient measurement module to obtain the dark field, background and foreground signal distribution functions and background and foreground correction coefficients of the FRET three-channel respectively;

[0128] The non-uniformity correction module of the living cell FRET three-channel fluorescence image is used to perform pixel-by-pixel background correction on the living cell FRET three-channel fluorescence image according to the FRET three-channel background signal distribution function and background correction coefficient obtained by the FRET three-channel measurement module through the established image non-uniformity correction model, taking the average grayscale of the background uniform area as the ideal grayscale, and performing pixel-by-pixel foreground correction on the living cell FRET three-channel fluorescence image according to the FRET three-channel foreground signal distribution function and foreground correction coefficient obtained by the FRET three-channel measurement module, taking the average grayscale of the foreground uniform area as the ideal grayscale, and finally outputting the corrected uniform FRET three-channel fluorescence image.

[0129] The above embodiments are preferred implementation modes of the present invention, but the implementation modes of the present invention are not limited to the above embodiments. Any other changes, modifications, substitutions, combinations, and simplifications that do not deviate from the spirit and principles of the present invention should be equivalent replacement methods and are included in the protection scope of the present invention.

Claims

1. A method for calibrating illumination light field for quantitative fluorescence resonance energy transfer microscopy, characterized in that: The following steps are involved: S1. Establish an image non-uniformity correction model, divide the image into foreground signal, background signal and dark field signal, measure the distribution of different signals, correct their non-uniformity, and thus obtain an image with uniform illumination; S2, measuring the dark field signal distribution function, turning off the illumination source of the microscopic imaging system and shielding the interference of the external ambient light, continuously collecting multiple frames of images at the minimum exposure time of the camera, and averaging the collected images to obtain the dark field signal distribution function; S3, determining the background signal distribution function and the background correction coefficient, taking the culture medium without cells as the sample, continuously collecting multiple frames of images under the same imaging conditions as the cells, averaging the collected images, and then subtracting the dark field signal distribution function obtained in step S2 to obtain the background signal distribution function, and calculating the background correction coefficient pixel by pixel based on the obtained background signal distribution function; S4, determining a foreground signal distribution function and a foreground correction coefficient, taking a fluorescent calibration slide as a sample, continuously acquiring multiple frames of images without overexposure, averaging the acquired images, and then subtracting a dark field signal distribution function and a background signal distribution function to obtain a foreground signal distribution function, and calculating a foreground correction coefficient pixel by pixel based on the acquired foreground signal distribution function; S5, repeating steps S2-S4 to obtain the dark field, background and foreground signal distribution functions of the three FRET channels and the background and foreground correction coefficients respectively; S6. Based on the image non-uniformity correction model, the non-uniformity of the living cell FRET three-channel fluorescence image is corrected using the measured signal distribution function and correction coefficient; based on the image non-uniformity correction model established in step S1, according to the FRET three-channel background signal distribution function and background correction coefficient obtained in step S5, the average grayscale of the background uniform area is used as the ideal grayscale, and the living cell FRET three-channel fluorescence image is subjected to pixel-by-pixel background correction; according to the foreground signal distribution function and foreground correction coefficient obtained in step S5, the average grayscale of the foreground uniform area is used as the ideal grayscale, and the living cell FRET three-channel fluorescence image is subjected to pixel-by-pixel foreground correction, and finally the corrected uniform FRET three-channel fluorescence image is output.

2. The illumination light field correction method for quantitative fluorescence resonance energy transfer microscopy according to claim 1, characterized in that: Step S1 establishes an image non-uniformity correction model as follows: F cell (x,y)=I cell (x,y)-B(x,y)-D(x,y) in, is the corrected cell gray value, FCF(x,y) is the foreground correction coefficient, F cell (x, y) is the foreground signal distribution function of the cell fluorescence image, I cell (x, y) is the measured gray value of the cell fluorescence image, BCF(x, y) is the background correction coefficient, B(x, y) is the background signal distribution function of the cell fluorescence image, and D(x, y) is the dark field signal distribution function.

3. The illumination light field correction method for quantitative fluorescence resonance energy transfer microscopy according to claim 1, characterized in that: The specific process of determining the dark field signal distribution function in step S2 is as follows: Where D(x,y) is the dark field signal distribution function, I D(i) (x, y) is the i-th frame of the image acquired continuously by setting the camera's minimum exposure time with the illumination source of the microscopic imaging system turned off and other ambient light shielded, and n is the total number of image frames acquired continuously.

4. The illumination light field correction method for quantitative fluorescence resonance energy transfer microscopy according to claim 1, characterized in that: The specific process of step S3 is as follows: S301, determining the background signal distribution function B(x, y); taking the culture medium without cells as the sample, continuously collecting multiple frames of images under the same imaging conditions as the cells, averaging the collected images, and then subtracting the dark field signal distribution function obtained in step S2 to obtain the background signal distribution function: Among them, B(x,y) is the background signal distribution function, I B(i) (x, y) is the i-th frame of image acquired continuously under the same imaging conditions as the cells by turning on the illumination source of the microscopic imaging system, D(x, y) is the dark field signal distribution function obtained in step S2, and n is the total number of image frames acquired continuously; S302, obtain the relative background center point (x B ,y B ); The collected background signal distribution function is processed by median filtering. The neighborhood window size selected by the median filter is m pixel*m pixel, and a pixel window of p pixel*p pixel size is selected. The pixel window is traversed in turn to calculate its average gray value, and the center point of the pixel window with the largest average gray value is selected as the relative background center point, which is recorded as (x B ,y B ); S303, determining the background correction coefficient BCF(x,y); calculating the background correction coefficient BCF(x,y) pixel by pixel for the collected image based on the acquired background signal distribution function, and the calculation formula is: Among them, BCF(x,y) is the background correction coefficient, B(x,y) is the background signal distribution function, and B(x B ,y B ) is the intensity relative to the background center point.

5. The illumination light field correction method for quantitative fluorescence resonance energy transfer microscopy imaging according to claim 1, characterized in that: The specific process of step S4 is as follows: S401, determining the foreground signal distribution function F(x, y); taking the fluorescent calibration slide as a sample, continuously collecting multiple frames of images without overexposure, averaging the collected images, and then subtracting the dark field signal distribution function and the background signal distribution function to obtain the foreground signal distribution function. The specific process is as follows: Among them, F(x,y) is the foreground signal distribution function, I F(i) (x, y) is the i-th frame image acquired continuously without overexposure by turning on the illumination source of the microscopic imaging system and taking the fluorescent calibration slide as the sample, B(x, y) is the background signal distribution function obtained in step S3, D(x, y) is the dark field signal distribution function obtained in step S2, and n is the total number of image frames acquired continuously; S402, obtain the relative foreground center point (x F ,y F ); The collected foreground distribution function is processed by median filtering. The size of the neighborhood window selected by the median filter is n pixel*n pixel, and a pixel window of size q pixel*q pixel is selected. The pixel window is traversed in turn to calculate its average gray value, and the center point of the pixel window with the largest average gray value is selected as the relative foreground center point, which is recorded as (x F ,y F ); S403, determining the foreground correction coefficient FCF(x,y); calculating the foreground correction coefficient FCF(x,y) pixel by pixel for the FRET three-channel image based on the acquired foreground signal distribution function, and the calculation formula is: Among them, FCF(x,y) is the foreground correction coefficient, F(x,y) is the foreground signal distribution function, and F(x F ,y F ) is the intensity relative to the foreground center point.

6. The illumination light field correction method for quantitative fluorescence resonance energy transfer microscopy according to claim 1, characterized in that: The specific process of step S5 is as follows: S501, repeat step S2 to measure the dark field signal distribution function of the three FRET channels: Among them, D AA (x,y),D DA (x,y),D DD (x, y) are the dark field signal distribution functions corresponding to the three FRET imaging channels respectively. The i-th frame of image is collected continuously by turning off the illumination source of the microscopic imaging system and shielding other ambient light for the three FRET imaging channels respectively, and n is the total number of image frames collected continuously by the three FRET imaging channels respectively; S502, repeat step S3 to determine the background signal distribution function and background correction coefficient of the three FRET channels: Among them, B AA (x,y),B DA (x,y),B DD (x, y) are the background signal distribution functions corresponding to the three FRET imaging channels, BCF AA (x,y), BCF DA (x,y), BCF DD (x, y) are the background correction coefficients corresponding to the three FRET channels respectively. The i-th frame image is acquired by turning on the illumination source of the microscopic imaging system for the three FRET imaging channels and continuously acquiring the image under the same imaging conditions as the cells. AA (x,y),D DA (x,y),D DD (x, y) is the dark field signal distribution function of the three channels of FRET obtained in step S501, n is the total number of image frames continuously collected by the three FRET imaging channels, They are the intensities of the relative background center points corresponding to the three FRET imaging channels; S503, repeat step S4 to determine the foreground signal distribution function and foreground correction coefficient of the three FRET channels: Among them, F AA (x,y),F DA (x,y),F DD (x, y) are the foreground signal distribution functions corresponding to the three FRET imaging channels, FCF AA (x,y), FCF DA (x,y), FCF DD (x, y) are the foreground correction coefficients corresponding to the three FRET channels respectively. The illumination source of the microscope imaging system is turned on for the three FRET imaging channels respectively, and the fluorescence calibration slide is used as a sample. The i-th frame image is continuously collected without overexposure. AA (x,y),B DA (x,y),B DD (x, y) is the FRET three-channel background signal distribution function D obtained in step S502 AA (x,y),D DA (x, y)D DD (x, is the FRET three-channel dark field signal distribution function obtained in step S501, n is the total number of image frames continuously collected by the three FRET imaging channels, They are the relative intensities of the foreground center point corresponding to the three FRET imaging channels.

7. The illumination light field correction method for quantitative fluorescence resonance energy transfer microscopy according to claim 1, characterized in that: The specific process of step S6 is as follows: in, The grayscale values ​​of the cell fluorescence images of the three channels after FRET correction, FCF AA (x,y), FCF DA (x,y), FCF DD (x, y) are the foreground correction coefficients corresponding to the three FRET channels respectively. are the foreground signal distribution functions of the cell fluorescence image in the three FRET channels, The measured gray values ​​of the cell fluorescence images of the three channels of FRET, BCF AA (x,y), BCF DA (x,y), BCF DD (x, y) are the background correction coefficients corresponding to the three FRET channels, B AA (x,y),B DA (x,y),B DD (x, y) are the background signal distribution functions corresponding to the three FRET imaging channels, respectively. AA (x,y),D DA (x,y),D DD (x, y) are the dark field signal distribution functions corresponding to the three FRET imaging channels respectively.

8. An illumination light field correction system for quantitative fluorescence resonance energy transfer microscopy, characterized in that: include: Image non-uniformity correction model building module, used to divide the image into foreground signal, background signal and dark field signal, and to obtain an image with uniform illumination by measuring the distribution of different signals and correcting their non-uniformity; The dark field signal distribution function determination module turns off the illumination source of the microscopic imaging system and shields the interference of the external ambient light, continuously collects multiple frames of images at the minimum exposure time of the camera, and averages the collected images to obtain the dark field signal distribution function; A background signal distribution function and background correction coefficient determination module, which uses a cell-free culture medium as a sample, continuously collects multiple frames of images under the same imaging conditions as the cells, averages the collected images, and then subtracts the dark field signal distribution function obtained by the dark field signal distribution function determination module to obtain a background signal distribution function, and calculates the background correction coefficient pixel by pixel based on the obtained background signal distribution function; The foreground signal distribution function and foreground correction coefficient determination module uses a fluorescent calibration slide as a sample, continuously collects multiple frames of images without overexposure, averages the collected images, and then subtracts the dark field signal distribution function and the background signal distribution function to obtain the foreground signal distribution function, and calculates the background correction coefficient pixel by pixel based on the obtained foreground signal distribution function; The FRET three-channel measurement module repeats the dark field signal distribution function measurement module, the background signal distribution function and background correction coefficient measurement module, and the foreground signal distribution function and foreground correction coefficient measurement module to obtain the dark field, background and foreground signal distribution functions and background and foreground correction coefficients of the FRET three-channel respectively; The non-uniformity correction module of the living cell FRET three-channel fluorescence image is used to perform pixel-by-pixel background correction on the living cell FRET three-channel fluorescence image according to the FRET three-channel background signal distribution function and background correction coefficient obtained by the FRET three-channel measurement module through the established image non-uniformity correction model, taking the average grayscale of the background uniform area as the ideal grayscale, and performing pixel-by-pixel foreground correction on the living cell FRET three-channel fluorescence image according to the FRET three-channel foreground signal distribution function and foreground correction coefficient obtained by the FRET three-channel measurement module, taking the average grayscale of the foreground uniform area as the ideal grayscale, and finally outputting the corrected uniform FRET three-channel fluorescence image.

Images