Device and method for separating images

Abstract

A first aspect of the present disclosure relates to a device for unmixing an image of a sample having fluorescent dyes.SOLUTION: The device is configured to obtain a mixed image of the sample, determine an unmixed image based on the mixed image, determine a noise map image based on the mixed image, determine a signal-to-noise ratio image based on the unmixed image and the noise map image, determine a denoised signal-to-noise ratio image based on the signal-to-noise ratio image, and determine a noise-reduced unmixed image of the sample based on the denoised signal-to-noise ratio image and the noise map image.SELECTED DRAWING: Figure 1

Description

[Technical Field]

[0001] The present disclosure relates to a device for isolating an image of a sample having a fluorescent dye and a method therefor. [Background technology]

[0002] Fluorescence imaging allows for the labeling of various targets within a sample using multiple fluorescent dye markers. However, spectral overlap between dyes (i.e., fluorophores) is common, which can complicate image clarity. One approach to address this is linear unmixing. Linear unmixing optimizes the signal-to-noise ratio and allows for clearer dye differentiation. However, for image regions with stronger spectral overlap and higher noise levels, linear unmixing can increase the noise level in pixels due to colocalization of corresponding dyes. For images with high noise levels, unmixing can be applied for noise reduction. However, when the number of dye components in an image pixel is large, these approaches are not useful for correctly distinguishing all components.

[0003] Fluorescence images can be affected by Poisson noise, where the noise level is approximately proportional to the square root of the intensity. This "noise law" can be lost after separation, such that the noise can no longer be estimated from the intensity at a pixel. The loss / degradation of the noise law limits the use of linear separation methods in combination with other processing steps, such as denoising. Denoising with Gaussian filters is not affected by the linear behavior of this filter. The denoising quality of Gaussian filters is limited.

[0004] If denoising is performed prior to separation, the denoiser can operate based on the Poisson noise law. However, even small deviations in filter behavior in different channels can be amplified by linear separation, resulting in visible artifacts. If separation is performed prior to denoising, the input image to the denoiser no longer follows the noise law. Therefore, after denoising, there is no known relationship between intensity and noise, and the denoiser cannot estimate the noise. The noise behavior after linear separation can also mislead other processing techniques, such as deconvolution and AI solutions with pre-trained models. Various improvements are needed. Summary of the Invention [Problem to be solved by the invention]

[0005] An object of the present disclosure is to improve image separation of samples with fluorescent dyes. [Means for solving the problem]

[0006] This object is solved by the disclosed embodiments as particularly defined by the subject matter of the independent claims. The dependent claims provide information on further embodiments. Various aspects, and various embodiments of these aspects, that provide additional features and advantages are also disclosed in this summary and in the following description.

[0007] A first aspect of the present disclosure is a device for separating an image of a sample having a fluorescent dye, the device comprising: - Obtain a mixed image of the sample, - Identifying separated images based on the mixed images; -Identifying a noise map image based on the mixed image; - determining a signal-to-noise ratio image based on the separation image and the noise map image; - Based on the signal-to-noise ratio image, identify the denoised signal-to-noise ratio image, - a device configured to determine a noise-reduced and separated image of a sample based on a denoised signal-to-noise ratio image and a noise map image.

[0008] In fluorescence imaging, a mixed image can be described as an image in which the detected fluorescence signal is a combination of emissions from multiple fluorophores (or dyes) with overlapping emission spectra. This overlap causes the fluorescence signals from different fluorophores to "blend," making it difficult to distinguish or quantify the contribution of each fluorophore in the image. Additionally or alternatively, spectral portions of a mixed image may arise from the autofluorescence properties of the sample.

[0009] Separation is the process of (at least partially) distinguishing overlapping signals and attributing them to individual fluorophores. This process can be achieved using computational techniques based on the known spectral properties of the fluorophores. Separation can also be referred to as "spectral separation" and is a standard technique, for example, in multicolor fluorescence microscopy, where multiple fluorophores are used simultaneously to stain the sample to be microscopically imaged. In particular, if the number of dyes and the number of detection channels are not the same, separation can be based on the pseudoinverse of the mixing matrix. In particular, there can be more collection channels than dyes. However, the opposite can also be true: there can be more dyes than collection channels.

[0010] A noise map image (or simply noise map) is a representation of the noise level or noise distribution across an image. It highlights where noise is present and can provide insight into the intensity and characteristics of noise across different regions of an image. To create a noise map, the noise level at each pixel (or in a local patch of an image) can be estimated. This estimation can be done based on statistical methods (e.g., calculating local variance) or by comparing the noisy image with a "clean" reference image, if available. A noise map can represent noise as a pixel-by-pixel value (each pixel indicates the noise level at that point) or in small patches (representing the average noise level across a small region). Larger values ​​in the noise map may correspond to higher levels of noise, often visualized as brighter intensities. In general, a noise map image provides a structured means for seeing how noise affects different regions of an image. A noise map image may have the same structure as the data on which it is based, i.e., the original image. A (raw) image may contain pixels in multiple dimensions, e.g., x, y, z, and time. The pixel data types may be different: the pixels of the noise map image may have floating-point values, whereas the input image may have 8-bit or 16-bit integer values. The noise map image may not be intended to be shown to the user.

[0011] A signal-to-noise ratio image (or signal-to-noise ratio (SNR) map) is a visual representation that shows the ratio of signal strength to noise level across an image. SNR maps can highlight the quality of an image in terms of signal clarity relative to background noise. A higher SNR indicates clearer and more reliable signal content, while a lower SNR indicates areas where noise may obscure important details.

[0012] Noise reduction can be performed based on the signal-to-noise ratio image. Such noise reduction can be based on the assumption that the noise in the signal-to-noise ratio map is distributed according to a Gaussian normal distribution. Or, in other words, a noise reduction that works best when the noise is normally distributed can be selected. After denoising the signal-to-noise ratio image, it is necessary to transform it back to the actual image. This is done by using a noise map image. The noise map image determined based on the blended image can be used as the noise map. Alternatively, a different noise map image can be used.

[0013] According to the embodiment of the first aspect, noise can be effectively reduced in multiple fluorescence images.

[0014] One embodiment of the first aspect is a device for separating an image of a sample having a fluorescent dye, comprising: - Noise reduction - configured to identify the noise map image The noise map image for identifying the noise reduced image relates to a device that is based on the noise reduced image.

[0015] The noise reduced image may be the original image. Additionally or alternatively, the noise reduced image may be a noise map image, which may result in a smoother noise map image.

[0016] The noise reduction for reducing noise and resulting in a noise-reduced image may be configured to reduce Poisson noise. Poisson noise, also known as shot noise, may be a type of image noise caused by random arrival of photons at an imaging sensor. Poisson noise may be dominant in low-light conditions and / or photon-limited scenarios. In a Poisson distribution process, the noise variance may be equal to the mean, resulting in signal-dependent non-uniform noise. Brighter areas in the image exhibit higher noise levels due to the detection of more photons. Poisson noise may manifest as grainy variations in image intensity. Techniques such as wavelet thresholding and / or Poisson-specific denoising methods can be used to reduce its effects.

[0017] One embodiment of the first aspect is a device for separating an image of a sample having a fluorescent dye, comprising: - A device configured to determine a noise-reduced noise map image based on a blended image.

[0018] Separation may alter the noise characteristics of the image, e.g., noise distribution. Basing the noise map image on the separated image may result in a more realistic noise propagation, i.e., a more realistic noise map image.

[0019] One embodiment of the first aspect relates to a device for separating an image of a sample having a fluorescent dye, wherein the denoised signal-to-noise image is further based on the denoised blended image.

[0020] The noise reduction / blending image may be a single value that represents the noise characteristics. This may be, for example, the average noise level or the standard deviation.

[0021] One embodiment of the first aspect is a device for separating an image of a sample having a fluorescent dye, wherein a noise map image for identifying a noise reduced image is determined based on one or more of the following: -Separate images and a noise map image based on the blended image; - Signal to noise ratio image and - Noise reduction - Signal to noise ratio images and devices identified in parallel.

[0022] Parallelizing one or more of these processing steps may result in a more efficient implementation of a device according to the first aspect.

[0023] One embodiment of the first aspect is a device for separating an image of a sample having a fluorescent dye, wherein the denoised and blended image comprises one or more of the following: -Separate images and a noise map image based on the blended image; -Related to devices identified in parallel with signal-to-noise images.

[0024] Parallelizing one or more of these processing steps may result in a more efficient implementation of a device according to the first aspect.

[0025] One embodiment of the first aspect relates to a device for separating an image of a sample having a fluorescent dye, wherein the noise map image is based on a weighted inverse of the blended image.

[0026] The weighted inverse of the blended image may be, for example, the quadratic inverse of the blending matrix.

[0027] One embodiment of the first aspect is a device for separating an image of a sample having a fluorescent dye, comprising: - noise correction image, i.e. --A separate image, --Noise reduction and separation image, --Related to a device configured to identify based on a noise map image.

[0028] By using all three types of information, the Poisson characteristics of the image noise in the original image can be preserved and further processing can be facilitated.

[0029] One embodiment of the first aspect is a device for separating an image of a sample having a fluorescent dye, comprising: - determining a noise estimate based on the blended image; - A device configured to determine a noise-compensated image based on a noise estimate.

[0030] This may also help to preserve the Poisson characteristics of the noise prevalent in the original source image.

[0031] One embodiment of the first aspect relates to a device for separating an image of a sample having a fluorescent dye, wherein the separated image is weighted by a blending factor not less than 20% and the noise-reduced separated image is weighted by a blending factor not greater than 80% to identify a noise-corrected image based on the separated image and the noise-reduced separated image.

[0032] The blending coefficients are limited to reduce noise levels below 20%, which has proven sufficient for all denoisers and deconvolution routines tested. Some denoisers may be confused if the data contains fractional photon counts. If the input images are recorded in photon counting mode, this problem cannot occur if the output data depth is the same as the input and no scaling is performed on the separation matrix. A final discretization step can be applied to other images.

[0033] One embodiment of the first aspect relates to a device for separating an image of a sample having a fluorescent dye, wherein a denoising and signal-to-noise ratio image is determined based on a wavelet filter.

[0034] The denoiser may be a Dual Tree-Complex Wavelet Filter (DTCWF) with a LeGall 5,3-tap filter for the first level and a Kingsbury quarter-sample-shifted quadrature (Q-shifted) 32,32-tap filter for the higher levels. Wavelet shrinkage follows a slightly modified Sendur and Selesnick procedure (L. Sendur & I. Selesnick, "Bivariate shrinkage with local variance estimation," IEEE Signal Process. Lett., 9, 438-441 (2003)).

[0035] A second aspect of the present disclosure is a method for separating an image of a sample having a fluorescent dye, the method comprising: - acquiring a mixed image of the sample; - identifying a separated image based on the blended image; - determining a noise map image based on the blended image; - determining a signal-to-noise ratio image based on the separation image and the noise map image; - identifying a denoised signal-to-noise ratio image based on the signal-to-noise ratio image; - determining a noise-reduced and separated image of the sample based on the denoised signal-to-noise ratio image and the noise map image.

[0036] Further steps and / or parameters may be based on the functionality described in relation to the first aspect and its embodiments.

[0037] A third aspect of the present disclosure relates to a computer program for performing the method according to the previous aspect when the computer program is run on a processor.

[0038] The computer program may be configured to be executed on a device according to the first aspect of the present disclosure.

[0039] Another aspect of the present disclosure relates to a computing device having a processor configured to perform a method according to any one of the preceding aspects / embodiments.

[0040] Another aspect of the present disclosure relates to a computer program product having instructions that, when executed by a computer system, cause the computer system to perform the method according to any one of the preceding aspects / embodiments.

[0041] Another aspect of the present disclosure relates to a computer-readable medium having instructions that, when executed by a computer system, cause the computer system to perform the method according to any one of the preceding aspects / embodiments.

[0042] Further advantages and features result from the following embodiments, some of which refer to the drawings. The drawings do not necessarily show the embodiments to scale. Dimensions of various features may be enlarged or reduced, particularly for clarity of illustration. To this end, the drawings are at least partially schematic. [Brief explanation of the drawings]

[0043] [Figure 1] FIG. 1 is a block diagram for image separation and denoising according to one embodiment of the present disclosure. [Figure 2] FIG. 10 is a block diagram for image separation and denoising according to another embodiment of the present disclosure. [Figure 3] FIG. 10 illustrates the separation / denoising results of one embodiment according to the present disclosure. [Figure 4] 10 is a table showing separation / denoising results of an embodiment according to the present disclosure. [Figure 5] FIG. 1 illustrates a system for use in connection with devices and / or methods according to embodiments of the present disclosure. DETAILED DESCRIPTION OF THE INVENTION

[0044] While some aspects have been described in the context of apparatus (or systems) in this disclosure, it is apparent that the description of these aspects also represents a description of a corresponding method, where blocks or devices correspond to steps or features of steps.

[0045] Similarly, aspects described in the context of a step also represent a description of a corresponding device, or a corresponding block, item or feature of a system, which may in particular be distributed across different locations and configured to exchange information between the different locations with respective communication means.

[0046] In general, the disclosure of a described method also applies to a corresponding device (or apparatus) for performing the method or a corresponding system including one or more devices, and vice versa. For example, if a particular step is described, the corresponding device may include features for performing the described step, even if the features are not explicitly described or shown in the figures. On the other hand, for example, if a particular device is described based on functional units, the corresponding method may include one or more steps for performing the described function, even if these steps are not explicitly described or shown in the figures. Similarly, a system may be provided with features of the corresponding device or features for performing the particular step. Features of the various exemplary aspects and embodiments described above or below may be combined unless expressly stated otherwise.

[0047] As used herein, the term "and / or" includes any and all combinations of one or more of the associated listed items and may be abbreviated as " / ". Words such as "exemplarily," "for example," or "particularly" denote a conditional or optional feature that may be combined with any other feature (mandatory, conditional, or optional) of an aspect or embodiment of the disclosure, unless expressly stated otherwise.

[0048] In the following description, reference is made to the accompanying drawings that form a part of this disclosure and in which are shown by way of illustration specific aspects in which the disclosure can be understood, and in which like reference numerals refer to like features, or at least functionally or structurally similar features.

[0049] FIG. 1 shows a flowchart 100 for processing noise reduced and separated images.

[0050] The first step is to load an initial image 101. The initial image 101 may be a raw image acquired by a microscope, laparoscope, or any other camera that captures images from a stained sample. The initial image 101 has signals from multiple stains of the stained sample or signals from autofluorescent materials of the sample. The initial image 101 is separated by using a pseudo-inverse matrix (U) 102, particularly for multiple dyes and multiple detection channels, to obtain a separated image 103.

[0051] Independently of the separation 103, the initial image 101 is also processed by a noise model 104 to receive a first noise map image 105 of the initial image 101. The first noise map image (E) 105 may also be based on a pseudo-inverse matrix U, whereby a noise map value is determined for a dye d at a pixel p and for a recording channel c, for example as follows:

number

[0052] The signal to noise ratio image 107 is determined by a noise processing function 106, which allows for, e.g.

number

[0053] The signal-to-noise ratio image 107 is then denoised by a denoising function 108 that assumes Gaussian noise. The result is a denoised SNR map 109.

[0054] Independently of identifying the separated image 103 and the first noise map image 105, the original image 101 can be additionally used to identify the second noise map image 113. To identify the second noise map image 113, the initial image 101 is denoised by a denoising function 110. This may assume Poisson statistics for the noise in the initial image, or may be a denoising function that works best for (at least) Poisson-distributed noise. The denoised-mixed image 111 is used to estimate the noise propagation 113 based on a second noise model 112. Basing the noise map 113 on the denoised but mixed initial image results in a smoother noise map image 113.

[0055] The denoising function 108 operates based on the estimated noise level 116 from the denoising unit 110 of the blended image 101. The square root of P can be used as the noise level for the denoising unit 108.

[0056] The final denoised image 115 is determined by the inverse transformation I of the denoised SNR map 109 and based on the second noise map image (M) 113, e.g., O d,p =I d,p M d,p is.

[0057] The denoiser 110 does not introduce the first-mentioned artifacts into the noise map 113 because the noise propagation is additive rather than subtractive. The noise estimate 116 is preferred in the denoiser 108 because it is more stable than the estimate 107.

[0058] In particular, if the noise estimate 116 is not a single value but a vector of, for example, about 256 floating-point values ​​containing information about the noise level and spatial structure, a mechanism similar to a neuron network denoiser can be implemented.

[0059] Using the smooth second noise map 113 rather than the first noise map image 105 for the denoising SNR map inverse transform 114 ensures that no additional noise is introduced in this step.

[0060] FIG. 2 shows a flowchart 200 for processing noise reduced and separated images.

[0061] An initial image 201 with multiple fluorescent signals is separated by applying a pseudo-inverse matrix 202, resulting in a separated image 203.

[0062] Separately, a noise model 204 is input with the mixed image 201 to obtain a noise map image 205 .

[0063] The separated image 203 and noise map image 205 are then converted into an SNR map in block 206. The SNR map is denoised by a denoiser 207 and converted back into a denoised separated image 211 by function 210. The denoiser 207 uses a noise estimate 209 from the initial image 201, calculated in a noise estimator 208.

[0064] The noisy separated image (C) 203, the denoised image (K) 211, the noise map image (E) 205 and the estimated noise level (I) 209 are combined by a mixer 212 to produce the desired image 213. This can be done in particular by the following process:

number

[0065] Here, the blending factor f may be limited to reduce noise levels below 20%, which has proven sufficient for all tested denoisers and deconvolution routines. Some denoisers may be confused if the data contains fractional photon counts. If the input image is recorded in photon counting mode, this problem cannot occur if the output data depth is the same as the input and no scaling is performed on the decomposition matrix. A final discretization step can be applied to other images. In the results shown in Figure 3, a discretization step that rounds to integer photon counts is used.

[0066] In a preferred embodiment, the denoiser by L. Sendur and I. Selesnick, "Bivariate shrinkage with local variance estimation," IEEE Signal Process, Lett., 9, 438-441 (2003), can be used in the denoising block 207. This denoiser provides sufficient quality and high processing speed.

[0067] 3 shows the denoised separation result 300 according to the method of FIG. 2 of the present disclosure in comparison with a prior art implementation. For this purpose, a denoiser 207 on the SNR map was implemented according to the denoiser by L. Sendur and I. Selesnick, "Bivariate shrinkage with local variance estimation," IEEE Signal Process, Lett., 9, 438-441 (2003).

[0068] A separated image 301 from a conventional linear separation method is shown on the left. A separated image 302 from separation according to the method of the second embodiment is shown on the right. As can be seen, areas with high crosstalk in the original image show higher noise in image 301 compared to image 302.

[0069] The results in FIGS. 3 and 4 are obtained by the method and the L. Sendur and I. Selesnick filter in the denoising block 207.

[0070] FIG. 4 shows a table with a comparison of separated images with ground truth confocal images in dB PSNR (Peak Signal-to-Noise Ratio). A typical mixing matrix is ​​applied to the ground truth image. This image is then separated by different techniques and the results are compared with the ground truth image. The table shows the comparison results for different levels of noise 401. Column 402 contains the results for the prior art linear separation method. Column 403 contains the results obtained by separation according to the second embodiment. Note that the images for column 403 still contain noise, since the goal of the method is to preserve the noise law, not to remove it.

[0071] Column 404 shows a comparison result of first performing denoising with Dabov's BM3D denoising filter and then performing prior art separation (K. Dabov, A. Foi, V. Katkovnik, and K. Egiazarian; "Image denoising by sparse 3d transform domain collaborative filtering"; IEEE Transactions on image processing; 16(82):3736-3745 (2007)). Column 405 shows a comparison result of first performing prior art separation and then performing BM3D denoising. Column 406 shows a comparison result of performing separation according to the second embodiment and then performing BM3D denoising. For typical photon counts in confocal images of 10 to 50 photon counts, the method of the second embodiment improves denoising quality by about 3 dB. Column 406 shows the results of a method according to one embodiment of the present disclosure, where denoising and separation in a first step is followed by a BM3D denoiser in a second step. During the first step, a noisy image with a Poisson noise law is generated. During the second step, the image is denoised.

[0072] Another embodiment relates to a method for separating a fluorescence microscope image by the following steps: performing a linear separation of an original image 101, resulting in a separated image 103; calculating a noise map image 105 of the separated image 103 from the original image; calculating a signal-to-noise ratio image 107 from the separated image 103 and the noise map image 105; calculating a denoised signal-to-noise ratio image 109 from the signal-to-noise ratio image 107, and inverting the denoised signal-to-noise ratio image 109 using the noise map image.

[0073] The noise map image for the inverse transform may be the denoised noise map image 113. The noise estimate used by the denoiser 108 may be estimated from the original image 101 rather than from the signal-to-noise ratio image 107. The noisy separated image 107, the denoised separated image 211 and the noise map image may be mixed to produce the result image 213. Thus, the mixer 212 may use the noise estimate from the original image (rather than from the separated image).

[0074] Some embodiments relate to a microscope that includes a system such as that described in connection with one or more of Figures 1-4. Alternatively, the microscope may be part of or connected to a system such as that described in connection with one or more of Figures 1-4.

[0075] FIG. 5 shows a schematic diagram of a system 500 configured to perform the methods described herein. The system 500 includes a microscope 510 and a computer system 520. The microscope 510 is configured to capture images and is connected to the computer system 520. The computer system 520 is configured to perform at least a portion of the methods described herein. The computer system 520 may be configured to execute machine learning algorithms. The computer system 520 and the microscope 510 may be separate entities or may be integrated into a common housing. The computer system 520 may be part of a central processing system of the microscope 510 and / or may be part of a subordinate component of the microscope 510, such as a sensor, actor, camera, or lighting unit of the microscope 510.

[0076] The computer system 520 may be a local computing device (e.g., a personal computer, laptop, tablet computer, or mobile phone) with one or more processors and one or more storage devices, or may be a distributed computing system (e.g., a cloud computing system with one or more processors and one or more storage devices distributed across various locations, such as local clients and / or one or more remote server farms and / or data centers). The computer system 520 may include any circuit or combination of circuits. In one embodiment, the computer system 520 may include one or more processors, which may be of any type. As used herein, a processor may contemplate any type of computing circuit, such as, but not limited to, a microprocessor of a microscope or microscope component (e.g., a camera), a microcontroller, a complex instruction set computing (CISC) microprocessor, a reduced instruction set computing (RISC) microprocessor, a very long instruction word (VLIW) microprocessor, a graphics processor, a digital signal processor (DSP), a multi-core processor, a field programmable gate array (FPGA), or any other type of processor or processing circuit. Other types of circuitry that may be included in computer system 520 may be custom circuitry, application specific integrated circuits (ASICs), etc., such as one or more circuits (e.g., communications circuits) used in wireless devices such as cell phones, tablet computers, laptop computers, two-way radios, and similar electronic systems. Computer system 520 may also include one or more storage devices, which may include one or more memory elements suitable for a particular application, such as main memory in the form of random access memory (RAM), one or more hard drives and / or one or more drives that handle removable media, such as compact discs (CDs), flash memory cards, digital video discs (DVDs), etc.Computer system 520 may also include a display device, one or more speakers and a controller which may include a keyboard and / or a mouse, trackball, touch screen, voice recognition device, or any other device that allows a user of the system to input information to and receive information from computer system 520.

[0077] Some or all of the steps may be performed by (or using) a hardware apparatus, such as, for example, a processor, microprocessor, programmable computer, or electronic circuitry. In some embodiments, any one or more of the critical steps may be performed by such an apparatus.

[0078] Depending on certain implementation requirements, embodiments of the present invention may be implemented in hardware or software. This implementation may be performed by a non-transitory storage medium, such as a digital storage medium, for example, a floppy disk, a DVD, a Blu-ray, a CD, a ROM, a PROM, an EPROM, an EEPROM, or a FLASH memory, on which electronically readable control signals are stored, which cooperate (or can cooperate) with a programmable computer system to implement the respective methods. Therefore, the digital storage medium may be computer-readable.

[0079] Some embodiments of the present invention include a data carrier having electronically readable control signals that can cooperate with a programmable computer system to perform any of the methods described herein.

[0080] Generally, embodiments of the present invention may be implemented as a computer program product comprising program code that is operative to perform any of the methods when the computer program product is run on a computer, and that may be stored, for example, on a machine-readable carrier.

[0081] Further embodiments comprise the computer program for performing any of the methods described herein, stored on a machine readable carrier.

[0082] In other words, an embodiment of the present invention is, therefore, a computer program having a program code for performing any of the methods described herein when the computer program runs on a computer.

[0083] Therefore, another embodiment of the invention is a recording medium (or data carrier or computer readable medium) containing a computer program stored thereon for performing any of the methods described herein when executed by a processor. The data carrier, digital recording medium or recording medium is typically tangible and / or non-transitory. Another embodiment of the invention is an apparatus as described herein, comprising a processor and a recording medium.

[0084] A further embodiment of the present invention is, therefore, a data stream or a sequence of signals representing the computer program for performing any of the methods described herein, the data stream or sequence of signals being for example adapted to be transmitted via a data communication connection, for example the Internet.

[0085] Another embodiment comprises a processing means, for example a computer, or a programmable logic device configured to or adapted to perform any of the methods described herein.

[0086] Another embodiment comprises a computer having installed thereon the computer program for performing any of the methods described herein.

[0087] Another embodiment of the present invention includes an apparatus or system configured to transfer (e.g., electronically or optically) a computer program for implementing any of the methods described herein to a receiver. The receiver may be, for example, a computer, a mobile device, a storage device, etc. The apparatus or system may include, for example, a file server for transferring the computer program to the receiver.

[0088] In some embodiments, a programmable logic device (e.g., a field programmable gate array) may be used to perform some or all of the functionality of the methods described herein. In some embodiments, a field programmable gate array may cooperate with a microprocessor to perform any of the methods described herein. In general, the methods may be advantageously performed by any hardware device.

[0089] As used herein, the term "and / or" includes any and all combinations of one or more of the associated listed items and may be abbreviated as " / ".

[0090] While some aspects have been described in the context of an apparatus, it will be apparent that these aspects also represent a description of a corresponding method, where a block or apparatus corresponds to a step or feature of a step, and similarly, aspects described in the context of a step also represent a description of a corresponding block or item or feature of a corresponding apparatus. [Explanation of symbols]

[0091] 100 Image Processing Flowchart 101 Images with multiple fluorescent signals 102 Pseudo-inverse matrix 103 Separated Images 104 Noise Model 105 First noise map image 106 Noise Processing 107 SNR maps 108 Noise Remover 109 Noise Reduction / SNR Map 110 Noise Remover 111 Noise Removal and Mixed Images 112 Noise Model 113 Noise Removal / Noise Map 114 Reverse conversion to image 115 Noise Removal and Mixed Images 116 Noise Estimate 200 Image Processing Flowchart 201 Images with multiple fluorescent signals 202 Pseudo-inverse matrix 203 Separated Images 204 Noise Model 205 noise map images 206 Noise Processing 207 Noise Remover 208 Noise Model 209 Noise Removal and Noise Map Images 210 Reverse conversion to image 211 Noise removal and separation images 212 Mixer 213 Noise removal and noise correction separated image 300 Implementation Results 301 Separated Images 302 Separated Images 401 Noise Level 402 PSNR for common separations 403 PSNR for noise removal and separation 404 PSNR for BM3D denoising filter 405 PSNR for Mixture Denoising 406 PSNR for noise removal and separation 500 Systems 510 Microscope 520 Computer Systems

Claims

1. 1. A device for separating an image of a sample having a fluorescent dye, said device comprising: - obtaining a mixed image (101) of the sample, - identifying a separated image (103) based on said mixed image, - determining a noise map image (105) based on said blended image; - determining a signal-to-noise ratio image (107) based on said separation image (103) and said noise map image (105); - determining a denoised signal-to-noise image (109) based on said signal-to-noise image (107); - configured to determine a noise-reduced and separated image (115) of said sample based on said denoised and signal-to-noise ratio image (109) and noise map images (105, 113), device.

2. The device comprises: - configured to determine a noise reduced and noise mapped image (113), The noise map image for identifying the noise-reduced and separated image (115) is based on the noise-reduced image (111), The device of claim 1.

3. The device comprises: - configured to determine said noise-reduced and noise-mapped image (113) based on said blended image (101); The device of claim 2.

4. The device comprises: - configured to identify a noise-reduced blended image (116) based on said blended image (101), The denoised signal-to-noise ratio image (109) is further based on the noise reduced blended image (116).

4. A device according to any one of claims 1 to 3.

5. The noise map image (113) for identifying the noise reduced and separated image (115) is said separated image (103), - said noise map image (105) based on said blended image (101); said signal-to-noise ratio image (107), - said denoised signal-to-noise image (109); Identified in conjunction with one or more of 5. A device according to any one of claims 1 to 4.

6. The denoised and blended image (116) said separated image (103), - said noise map image (105) based on said blended image (101); said signal-to-noise ratio image (107), Identified in conjunction with one or more of 6. A device according to any one of claims 1 to 5.

7. the noise map image (105, 113) is based on a weighted inverse of the blended image; 7. A device according to any one of claims 1 to 6.

8. The device comprises: said separated image (203); --the noise-reduced and separated image (211); --the noise map image (205); and determining a noise-compensated image (213) based on the noise-compensated image. A device according to any one of claims 1 to 7.

9. The device comprises: - determining a noise estimate (209) based on said blended image; - determining the noise-compensated image (213) based on the noise estimate, The device of claim 8.

10. To identify the noise-corrected image (213) based on the separated image (203) and the noise-reduced and separated image (211), the separated image (203) is weighted by a blending factor not less than 20% and the noise-reduced and separated image (211) is weighted by a blending factor not greater than 80%. The device of claim 9.

11. The denoised signal-to-noise ratio image (109) is determined based on a wavelet filter. A device according to any one of claims 1 to 10.

12. 1. A method for separating an image of a sample having a fluorescent dye, the method comprising: - acquiring a mixed image (101) of the sample; - identifying a separated image (103) based on said mixed image; - determining a noise map image (105) based on said blended image; - determining a signal-to-noise ratio image (107) based on said separation image (103) and said noise map image (105); - determining a denoised signal-to-noise image (109) based on said signal-to-noise image (107); - determining a noise-reduced and separated image (115) of said sample based on said denoised and signal-to-noise ratio image (109) and noise map images (105, 113); A method having the following.

13. 13. A computer program comprising a program code for performing the method of claim 12 when the computer program runs on a processor.

Images