Correlation image analysis for 3D biopsy

Abstract

The present invention relates to image analysis of pathological images. To improve the reliability of image analysis of pathological images, a method for providing support in identifying at least one feature of a tissue sample in a microscope image is provided. The method comprises the following steps: providing a first image of a first microscope modality representing an extent of the tissue sample, providing a second image of a second microscope modality representing the extent of the tissue sample, obtaining first information about the at least one feature by applying a first high-intensity filter to the first image to generate a first high-intensity image or by applying a first low-intensity filter to the first image to generate a first low-intensity image, obtaining second information about the at least one feature by applying a second high-intensity filter to the second image to generate a second high-intensity image or by applying a second low-intensity filter to the second image to generate a second low-intensity image, calculating a correlation between an image pair including one of the first high-intensity image and the first low-intensity image and one of the second low-intensity image and the second high-intensity image to associate the first information about the at least one feature with the second information, and outputting the calculated correlation to provide support in identifying the at least one feature of the tissue sample.

Description

Technical Field

[0001] The present invention relates to image analysis of pathological images. In particular, the present invention relates to a method, a data processing device and a system, as well as a computer program element and a computer-readable medium for providing support in identifying at least one feature of a tissue sample in a microscope image. Background Art

[0002] Pathological examination of tissue samples, such as human tissue samples or biopsies, may involve the identification of certain features of interest within the tissue sample. For example, in three-dimensional biopsy analysis, it may be necessary to detect gradual increases in the density of nuclear structures or changes in the internal morphology of the sample. Various microscopy modalities have been developed to provide volumetric or two-dimensional microscopic imaging of, for example, light absorption, reflection, or scattering contrast in biological tissue.

[0003] US 9224301 B2 describes image analysis of darkfield images. Particles to be distinguished are identified in fluorescence analysis and can be removed by analyzing at least one darkfield image. However, due to image artifacts or simple signal intensity variations across the tissue sample, for example, the features of interest may not be correctly identified. Summary of the Invention

[0004] There may be a need to improve the reliability of image analysis of pathology images.

[0005] The objects of the invention are solved by the subject-matter of the independent claims, wherein further embodiments are contained in the dependent claims. It should be noted that the aspects of the invention described below also apply to the method, the data processing apparatus, the system, the computer program element and the computer-readable medium.

[0006] A first aspect of the present invention relates to a method for providing support in identifying at least one feature of a tissue sample in a microscopic image. The method includes the steps of: providing a first image of a first microscopic modality representing an area of ​​the tissue sample; providing a second image of a second microscopic modality representing the area of ​​the tissue sample; obtaining first information of the at least one feature by applying a first high-intensity filter to the first image to generate a first high-intensity image or applying a first low-intensity filter to the first image to generate a first low-intensity image; obtaining second information of the at least one feature by applying a second high-intensity filter to the second image to generate a second high-intensity image or applying a second low-intensity filter to the second image to generate a second low-intensity image; calculating a correlation between an image pair including one of the first high-intensity image and the first low-intensity image and one of the second high-intensity image and the second low-intensity image to associate the first information and the second information of the at least one feature; and outputting the calculated correlation to provide support in identifying the at least one feature of the tissue sample.

[0007] In other words, in order to enhance the detection of predefined features (e.g., nuclei) of a tissue sample, it is proposed to associate the detection information of the predefined features in an image of a first microscopy modality with the detection information of the same features in a second microscopy modality; that is, the feature information from the second modality is used as supplementary feature information to the first modality.

[0008] The tissue sample can be obtained from a sample selected from, for example, a liver sample, a kidney sample, a muscle sample, a brain sample, a lung sample, a skin sample, a thymus sample, a spleen sample, a gastrointestinal tract sample, a pancreas sample, a prostate sample, a breast sample, or a thyroid sample. The tissue sample can be derived from, for example, a human sample, or a sample from an animal, such as a mouse sample, a rat sample, a monkey sample, or a dog sample.

[0009] The first and second microscopy modalities can include various optical microscopy modalities that allow the study of biological structures. Microscopy modalities can include fluorescence microscopy imaging (such as epifluorescence microscopy), total internal reflection fluorescence (TIRF) microscopy, confocal microscopy, or multiphoton excitation microscopy. Microscopy modalities can also include absorption-based microscopy modalities (such as brightfield microscopy), stimulated emission microscopy, photoacoustic microscopy, or optical projection tomography (OPT). In addition, microscopy modalities can include scattering-based microscopy (such as darkfield microscopy) and optical coherence tomography (OCT). As described in detail below with respect to Figure 1 and Figures 2A to 2C As will be explained in the exemplary embodiments, the first and second microscopy modalities may be dark field and fluorescence microscopy imaging.

[0010] A high-intensity or low-intensity filter is applied to the first and second images to extract information about features. For example, the presence of certain cells can be detected using fluorescently labeled antibodies and specific staining of their cell membranes using a fluorescence microscope. Information about cell nuclei can also be extracted using nuclear staining and fluorescence microscope images. A high-intensity filter can be applied to a fluorescence microscope image to identify pixels representing areas with high sample density and little or no scattering, indicating areas with multiple cell nuclei. For example, to extract information about tubules in a fluorescence microscope image, a low-intensity filter can be applied to the fluorescence microscope image to identify pixels representing areas with low sample density and high scattering, indicating areas surrounding tubules in a tissue sample. As another example, to extract information about cell nuclei in a darkfield microscope image, a low-intensity filter can be applied to the darkfield image to identify pixels representing areas with high sample density and little or no scattering, indicating areas with many cell nuclei. To extract information about tubules in a darkfield microscope image, a high-intensity filter can be applied to identify pixels representing areas with low sample density and high scattering, indicating areas surrounding tubules in a tissue sample. The high-intensity or low-intensity filter can be a threshold filter. Thresholding identifies pixels that have intensity values ​​within a specific range. In one example, thresholding using a high intensity filter can identify pixels above a specific threshold. In another example, thresholding using a low intensity filter can identify pixels below a specific threshold. Various thresholding techniques can be used, including but not limited to: global thresholding, local thresholding, thresholding based on histogram shape, thresholding based on clustering, and thresholding based on object properties. Global thresholding applies the same threshold to every pixel in the entire image. Local thresholding can be applied to situations where the background itself in the image varies in brightness. Color images, such as fluorescence images, can also be thresholded. One approach is to assign a separate threshold to each RGB component of the image and then combine them using an AND operation.

[0011] It should also be noted that the term "generating a high-intensity or low-intensity image" refers to obtaining high-intensity or low-intensity image data. Displaying the high-intensity or low-intensity image is not required.

[0012] Correlation can use a correlation coefficient as a measure of similarity between two images in an image pair for each position. In order to associate first information and second information of at least one feature, the image pair may include one of a first high-intensity image and a first low-intensity image and one of a second high-intensity image and a second low-intensity image. In other words, the image pair may include at least one of the following combinations: i) a first high-intensity image and a second high-intensity image, ii) a first high-intensity image and a second low-intensity image, iii) a first low-intensity image and a second high-intensity image, and iv) a first low-intensity image and a second low-intensity image. The combination of images in the image pair depends on the first microscopy modality and the second microscopy modality and the features to be detected. For example, in order to extract information of a cell nucleus, the image pair may include a high-intensity fluorescence microscopy image and a low-intensity dark-field microscopy image, as in Figure 1 and Figures 2A to 2C As explained in

[15] , this is because contrast enhancement works differently in the two modalities. On the other hand, to extract information about the cell nucleus, the image pair can include a high-intensity fluorescence microscopy image and a high-intensity brightfield image, because contrast enhancement works similarly in both modalities. The result will be greatest for locations where there is a correspondence between the two images. These locations represent the pixels of the features to be identified in the first and second images. As will be detailed below, Figure 1 As explained in the exemplary embodiment of FIG, the correlation can be based on Boolean operations or other more complex methods. Therefore, the identification of features can be performed based on the correlation coefficient, for example manually or automatically using an algorithm.

[0013] Because the first image from the first microscopy modality and the second image from the second microscopy modality may not have the same image artifacts, or the intensity of the same signal may vary across the entire tissue sample, the correlation can eliminate or reduce their negative effects. Consequently, the recognition process can be more reliable. This can be beneficial for machine-based recognition processes using image segmentation and feature extraction methods. Therefore, these method steps can be performed during the image adjustment phase, which is the next stage of image segmentation and feature extraction. In other words, these method steps can manipulate the microscopy image to eliminate or reduce the negative effects of image artifacts and / or signal variations across the entire tissue sample, thereby making it meet the requirements of the next stage of further image processing. As a result, features of interest can be correctly identified with a higher probability.

[0014] According to an embodiment of the present invention, the first microscopy modality and the second microscopy modality are different modalities selected from at least one of the following: fluorescence microscopy imaging, dark-field microscopy imaging, and bright-field microscopy imaging.

[0015] According to an embodiment of the present invention, the first microscopy modality is fluorescence microscopy imaging and the second microscopy modality is dark-field microscopy imaging.

[0016] Fluorescence microscopy allows visualization of fluorescently labeled structures, such as cells stained with fluorescently labeled antibodies on their cell membranes or by staining the nucleus with nuclear stains or so-called intercalating dyes. Darkfield microscopy allows identification of boundaries and internal cavities. Combining the two imaging modalities can thus provide additional information about the tissue. For example, it can allow for a better assessment of the location of tubules within the imaged tissue.

[0017] According to an embodiment of the present invention, the image pair comprises a first high-intensity image and a second low-intensity image or a first low-intensity image and a second high-intensity image.

[0018] This can be beneficial in situations where the contrast enhancement of the two microscopy modalities works differently. In one example, an absorption-based imaging modality (such as brightfield) is combined with a scattering-based imaging modality (such as darkfield or OCT). In another example, a fluorescence imaging modality is combined with a scattering-based imaging modality. For example, a high-intensity darkfield image and a low-intensity fluorescence image represent areas where the sample density is low and there is a lot of scattering, such as around a catheter or tubule in the sample. On the other hand, a low-intensity darkfield image and a high-intensity fluorescence image represent areas where the sample density is high and there is little or no scattering, indicating an area with many cell nuclei, indicating abnormal tissue activity.

[0019] According to an embodiment of the present disclosure, the correlation is calculated based on Boolean operations.

[0020] For example, a pixel-by-pixel Boolean multiplication can be performed to obtain a cross-correlation matrix. Other more complex methods can also be used, which will be discussed in detail below. Figure 1 The exemplary embodiments are explained in detail.

[0021] A second aspect of the present invention relates to a system for supporting identification of at least one feature of a tissue sample in a microscope image. The system includes a data processing device and a display as described above and below. The display is configured to display at least one of a first image and a second image, and a calculated correlation between first extracted information and second extracted information of at least one feature output from the data processing device.

[0022] A third aspect of the present invention relates to a data processing device for supporting the identification of at least one feature of a tissue sample in a microscope image. The data processing device includes an input unit, an information extraction unit, a correlation unit, and an output unit. The input unit is configured to receive a first image of a first microscope modality representing an area of ​​the tissue sample and a second image of a second microscope modality representing the area of ​​the tissue sample. The information extraction unit is configured to obtain first information about the at least one feature by applying a first high-intensity filter to the first image to generate a first high-intensity image or by applying a first low-intensity filter to the first image to generate a first low-intensity image, and to obtain second information about the at least one feature by applying a second high-intensity filter to the second image to generate a second high-intensity image or by applying a second low-intensity filter to the second image to generate a second low-intensity image. The correlation unit is configured to calculate a correlation between an image pair comprising one of the first high-intensity image and the first low-intensity image and one of the second high-intensity image and the second low-intensity image to correlate the first information and the second information about the at least one feature. The output unit is configured to output the calculated correlation to provide support for identifying the at least one feature of the tissue sample.

[0023] For the data processing apparatus, the same explanations as for the above-described method apply. Thus, the data processing apparatus can provide image adjustments to remove image artifacts and signal variations across the entire tissue sample in the microscope image, thereby making the microscope image suitable and reliable for further processing in machine-based image processing methods, such as image segmentation and feature extraction.

[0024] The term "unit" as used herein may include or refer to a portion of an application specific integrated circuit (ASIC), an electronic circuit, a processor (shared, dedicated, or grouped) and / or memory (shared, dedicated, or grouped) that executes one or more software or firmware programs, a combinational logic circuit, and / or other suitable components that provide the described functionality.

[0025] According to an embodiment of the present invention, the data processing apparatus includes a feature recognition unit configured to recognize at least one feature based on at least one of the first image and the second image and the calculated correlation.

[0026] In other words, the data processing device can perform feature recognition after the adjustment. Since the microscope image after the adjustment is more reliable, the chance of correctly identifying the feature of interest in the tissue sample can be increased.

[0027] According to an embodiment of the present invention, the first microscopy modality and the second microscopy modality are different modalities selected from at least one of the following: fluorescence microscopy imaging, dark-field microscopy imaging, and bright-field microscopy imaging.

[0028] According to an embodiment of the present invention, the first microscopy modality is fluorescence microscopy imaging and the second microscopy modality is dark-field microscopy imaging.

[0029] Thus, the boundaries and internal cavities can be visualized using darkfield microscopy, and fluorescently labeled structures (e.g., cancer cells) can be visualized using fluorescently labeled microscopy. In this way, both structures can be combined in one three-dimensional representation.

[0030] According to an embodiment of the present invention, the image pair comprises a first high-intensity image and a second low-intensity image or a first low-intensity image and a second high-intensity image.

[0031] According to an embodiment of the present invention, the correlation unit is configured to calculate the correlation based on Boolean operations.

[0032] A fourth aspect of the invention relates to a computer program element for instructing the means described above and below, which, when executed by a processing unit, is adapted to perform the steps of the method described above and below.

[0033] A fifth aspect of the invention relates to a computer-readable medium storing a program element.

[0034] These and other aspects of the invention will be apparent from and elucidated with reference to the embodiments described hereinafter. BRIEF DESCRIPTION OF THE DRAWINGS

[0035] Exemplary embodiments of the present invention will be described below with reference to the accompanying drawings:

[0036] Figure 1 A flow chart illustrating a method for providing support in identifying at least one feature of a tissue sample in a microscope image according to some embodiments of the present disclosure is shown;

[0037] Figure 2A shows a fluorescence microscopy image of a rat liver region according to some embodiments of the present disclosure;

[0038] Figure 2B shows a dark field microscopy image of the same rat liver region according to some embodiments of the present disclosure;

[0039] Figure 2C shows an overlay of a filtered fluorescence microscopy image and a filtered dark-field microscopy image of the same rat liver region according to some embodiments of the present disclosure;

[0040] Figure 3A shows a fluorescence microscopy image of a human prostate region according to some embodiments of the present disclosure;

[0041] Figure 3Bshows a dark field microscopy image of the same human prostate region according to some embodiments of the present disclosure;

[0042] Figure 3C shows an overlay of a filtered fluorescence microscopy image and a filtered dark-field microscopy image of the same human prostate region according to some embodiments of the present disclosure;

[0043] Figure 4 shows a data processing apparatus for providing support in identifying at least one feature of a tissue sample in a microscope image according to some embodiments of the present disclosure;

[0044] Figure 5 A system for providing support in identifying at least one feature of a tissue sample in a microscope image is shown, according to some embodiments of the present disclosure. DETAILED DESCRIPTION

[0045] Figure 1 At least one feature 14, 16 for identifying a tissue sample in a microscope image according to some embodiments of the present disclosure is shown (see Figures 2A-2C and Figures 3A-3C ) is a flowchart of a method 100 provided in support. Figures 2A to 2C A set of rat liver images are shown to illustrate the method steps according to an exemplary embodiment of the present disclosure. Specifically, fluorescence microscope images of rat liver regions collected under 620 nm laser excitation are shown in FIG. Figure 2A In this example, SiR DNA nuclear dye (Spirochrome, Switzerland) was used. This is a far-red, fluorescent, cell-permeable and highly specific probe for DNA (no specificity for tumor versus normal cells). Dark-field microscopy images of the same rat liver region acquired under 840 nm laser irradiation are shown in Figure 2. Figure 2B The overlay of the filtered fluorescence microscopy image and the filtered dark field microscopy image is shown in Figure 2C middle.

[0046] In step 110, a first image 10 of a first microscopy modality representing a range of a tissue sample is provided. For example, the first image 10 may be Figure 2A Fluorescence microscopy images are shown, and the first microscopy modality may be fluorescence microscopy imaging.

[0047] In step 120, a second image 12 of a second microscopic modality representing said area of ​​the tissue sample is provided. For example, the second image 12 may be as follows: Figure 2B As shown in the dark field microscopy image, the second microscopy modality may be dark field microscopy imaging.

[0048] In step 130, first information of at least one feature 14, 16 is obtained by applying a first high-intensity filter to first image 10 to generate a first high-intensity image (not shown). Alternatively, first information of at least one feature 14, 16 is obtained by applying a first low-intensity filter to first image 10 to generate a first low-intensity image (not shown). Note that the first high-intensity or low-intensity image refers to the first high-intensity or low-intensity image data. Therefore, the first high-intensity or low-intensity image does not need to be displayed.

[0049] The choice of the first high intensity filter or the first low intensity filter depends on the features to be identified. For example, Figure 2A The high-intensity region 14 in the fluorescence microscope image (i.e., the first image 10) in FIG. 1 represents an area with high sample density and no or little scattering, indicating a range with many cell nuclei. Therefore, a high-intensity filter can be applied to Figure 2A On the other hand, a low intensity filter can be applied to the fluorescence microscopy image to obtain information about the cell nucleus. Figure 2A Fluorescence microscopy images of the sample were used to identify areas with low sample density and high scatter, indicating areas around ducts or tubules.

[0050] The high intensity filter can be defined as

[0051]

[0052] Where T is the threshold, x, y are the coordinates of the first or second image f(x,y), and g(x,y) is the threshold image after applying the high intensity filter.

[0053] The low intensity filter can be defined as

[0054]

[0055] The threshold T can be a manual threshold defined by the user. Alternatively, the threshold can be determined from the histogram by an automated method. For global thresholding, the threshold T is the same for every pixel in the entire image. For local thresholding, the threshold T varies across the entire image and can therefore be defined as T(x,y).

[0056] In step 140, second information about at least one feature 14, 16 is obtained by applying a second high-intensity filter to the second image 12 to generate a second high-intensity image (not shown). Alternatively, second information about at least one feature 14, 16 is obtained by applying a second low-intensity filter to the second image to generate a second low-intensity image (not shown). Note that the second high- or low-intensity image refers to second high- or low-intensity image data. The second high- or low-intensity image need not be displayed.

[0057] The choice of the second high intensity filter or the second low intensity filter also depends on the features to be identified. For example, Figure 2BThe high intensity region 16 in the dark field microscope image (i.e., the second image 12) in FIG. 1 represents an area with low sample density and a large amount of scattering, indicating the area around the catheter or thin tube. Therefore, a high intensity filter can be applied to Figure 2B On the other hand, a low intensity filter can be applied to dark field microscopy images to obtain information about ducts or thin tubes. Figure 2B Darkfield microscopy images were taken to identify areas of high sample density with little or no scattering, indicating areas with a high number of cell nuclei.

[0058] In step 150 , correlation of an image pair is calculated to correlate first and second information of at least one feature, wherein the image pair includes one of a first high-intensity image and a first low-intensity image and one of a second high-intensity image and a second low-intensity image.

[0059] In an example where at least one feature to be identified includes a cell nucleus, a correlation between a high-intensity fluorescence microscopy image and a low-intensity dark-field microscopy image may be calculated because both images represent areas where the sample density is high and there is no or little scattering, indicating an area with many cell nuclei.

[0060] In another example, where the at least one feature to be identified includes ducts and cell nuclei, a correlation between a low-intensity fluorescence microscopy image and a high-intensity dark-field microscopy image can be calculated because both images represent areas of low sample density with a large amount of scatter, indicating the presence of ducts and tubules.

[0061] It is also possible to combine the two examples above to create an overlay that enables Figure 2C In the overlay image 18 shown, two features are identified: cell nuclei and ducts or tubules. In this example, correlation is calculated based on Boolean operations. A binary image is used, where each pixel can have only one of two values, indicating whether it is part of at least one feature to be identified. Although more advanced correlation methods can be used, correlation is calculated by multiplexing the binary images of the image pair. In the overlay image 18, it is easy to see areas 14 with high sample density and little or no scattering, indicating areas with many cell nuclei, and areas 16 with low sample density and extensive scattering, indicating areas with ducts or tubules.

[0062] In addition to Boolean operations, more complex methods can be used to calculate correlations, such as

[0063]

[0064] Where r is the correlation coefficient, m and n are the first or second image A mn and B mn The coordinates of It's A mnThe average value of It's B mn The average value of .

[0065] In step 160 , the calculated correlation is output to provide support in identifying at least one feature of the tissue sample.

[0066] Figures 3A to 3C A set of human prostate images is shown to illustrate the method steps according to an exemplary embodiment of the present disclosure. A fluorescence microscope image 10 of a human prostate region collected under 620 nm laser excitation is shown in FIG. Figure 3A In the experiment, SiR DNA nuclear dye (Spirochrome, Switzerland) was used, which is a far-red, fluorescent, cell-permeable and highly specific DNA probe (not specific for tumor and normal cells). Specifically, dark field microscopy images 12 of the same rat liver region collected under 530 nm laser irradiation are shown in FIG. Figure 3B The superimposed image 18 of the filtered fluorescence microscopy image and the filtered dark field microscopy image is shown in FIG. Figure 3C In the Figures 2A to 2C Similar analysis was performed in .

[0067] Correlating feature information obtained from two different microscopy modalities can resolve issues with image artifacts or simple signal intensity variations across a sample, making the feature identification process reliable. This could be beneficial for machine-based methods used to provide reliable and rapid tissue analysis.

[0068] Optionally, in step 170 , at least one feature is identified based on at least one of the first and second images and the calculated correlation.

[0069] In an example, this identification step may be performed manually by a user.

[0070] In another example, the identification step can be performed automatically based on image segmentation and feature extraction methods. For example, the calculated correlation (e.g., the resulting cross-correlation matrix) can be used as a pixel-by-pixel weighting function, which can be multiplied by the first image or the second image to identify a single feature (e.g., a duct and tubule or a cell nucleus), or multiplied by the sum of both images to identify two features. Image segmentation and feature extraction can be performed on at least one of the first weighted image and the second weighted image.

[0071] Figure 4A data processing apparatus 200 providing support in identifying at least one feature of a tissue sample in a microscope image according to some embodiments of the present disclosure is shown. The data processing apparatus 200 includes an input unit 210, an information extraction unit 220, a correlation unit 230, and an output unit 240, which may include or be part of an ASIC, an electronic circuit, a processor (shared, dedicated, or grouped) and / or memory (shared, dedicated, or grouped) executing one or more software or firmware programs, combinational logic circuitry, and / or other suitable components that provide the described functionality.

[0072] The input unit 210 is configured to receive a first image of a first microscopy modality representing an extent of a tissue sample and a second image of a second microscopy modality representing the extent of the tissue sample.

[0073] The information extraction unit 220 is configured to obtain first information about at least one feature by applying a first high-intensity filter to the first image to generate a first high-intensity image, or by applying a first low-intensity filter to the first image to generate a first low-intensity image, and to obtain second information about the at least one feature by applying a second high-intensity filter to the second image to generate a second high-intensity image, or by applying a second low-intensity filter to the second image to generate a second low-intensity image. For example, the first microscopy modality and the second microscopy modality are different modalities selected from at least one of the following: fluorescence microscopy, dark-field microscopy, and bright-field microscopy. In an example, the first microscopy modality is fluorescence microscopy, and the second microscopy modality is dark-field microscopy.

[0074] Correlation unit 230 is configured to calculate a correlation for an image pair including one of a first high-intensity image and a first low-intensity image, and one of a second high-intensity image and a second low-intensity image, to associate first information and second information about at least one feature. For example, the image pair includes a first high-intensity image and a second low-intensity image, or a first low-intensity image and a second high-intensity image. For example, correlation unit 206 is configured to calculate the correlation based on a Boolean operation.

[0075] The output unit 240 is configured to output the calculated correlation to provide support in identifying at least one feature of the tissue sample.

[0076] Optionally, the data processing apparatus further includes a feature recognition unit 250. The feature recognition unit 250 is configured to recognize at least one feature based on at least one of the first image and the second image and the calculated correlation.

[0077] Figure 5A system 300 for providing support in identifying at least one feature of a tissue sample in a microscope image according to some embodiments of the present disclosure is shown. The system 300 includes a data processing device 200 as described above and below and a display 310. The display 310 is configured to display at least one of the first and second images and a calculated correlation between first extracted information and second extracted information of at least one feature output from the data processing device.

[0078] In a further exemplary embodiment of the present invention, a computer program or a computer program element is provided, which is characterized in that it is adapted to execute the method steps of the method according to one of the preceding embodiments on a suitable system.

[0079] The computer program element can thus be stored on a computer unit, which can also be part of an embodiment of the present invention. The computer unit can be adapted to execute or cause the execution of the steps of the above-described method. Furthermore, it can be adapted to operate components of the above-described apparatus. The computer unit can be adapted to operate automatically and / or execute user commands. The computer program can be loaded into a working memory of a data processor. The data processor can thus be equipped to execute the method of the present invention.

[0080] This exemplary embodiment of the invention covers both a computer program that right from the beginning uses the invention and a computer program that by means of an up-date turns an existing program into a program that uses the invention.

[0081] Furthermore, the computer program element may be able to provide all necessary steps to implement the procedures of the exemplary embodiment of the method as described above.

[0082] According to a further exemplary embodiment of the present invention, a computer-readable medium, such as a CD-ROM, is provided, wherein the computer-readable medium has a computer program element stored thereon, said computer program element being described in the preceding section.

[0083] The computer program may be stored and / or distributed on suitable media such as optical storage media or solid-state media provided together with or as part of other hardware, but may also be distributed in other forms such as via the Internet or other wired or wireless telecommunications systems.

[0084] However, the computer program may also be presented over a network like the World Wide Web and may be downloaded from such a network into the working memory of a data processor. According to a further exemplary embodiment of the present invention, a medium for making a computer program element available for downloading is provided, said computer program element being arranged to perform a method according to one of the aforementioned embodiments of the invention.

[0085] It should be noted that embodiments of the present invention are described with reference to different subject matter. Specifically, some embodiments are described with reference to method-type claims, while other embodiments are described with reference to apparatus-type claims. However, those skilled in the art will appreciate from the above and following descriptions that, unless otherwise indicated, any combination of features relating to different subject matter, in addition to any combination of features belonging to the same category of subject matter, is also considered to be disclosed with this application. However, all features can be combined to provide synergistic effects that are greater than the sum of their individual features.

[0086] Although the present invention has been illustrated and described in detail in the drawings and the foregoing description, such illustration and description are to be considered illustrative or exemplary rather than restrictive. The present invention is not limited to the disclosed embodiments. Other variations to the disclosed embodiments can be understood and effected by those skilled in the art in practicing the claimed invention from a study of the drawings, the disclosure, and the appended claims.

[0087] In the claims, the word "comprising" does not exclude other elements or steps, and the indefinite article "a" or "an" does not exclude a plurality. A single processor or other unit may perform the functions of several items recited in a claim. The fact that certain measures are recited in mutually different dependent claims does not indicate that a combination of these measures cannot be used to advantage. Any reference signs in the claims should not be construed as limiting the scope.

Claims

1. A method (100) for identifying at least one feature of a tissue sample in a microscope image, comprising the following steps: providing (110) a first image (110) of a fluorescence microscopy imaging modality representing an extent of the tissue sample; providing (120) a second image (120) in a dark field microscopy imaging modality representing the extent of the tissue sample; and generating (130) a first high-intensity image by applying a first high-intensity filter to identify pixels in the first image that are above a threshold value, thereby obtaining first information of the at least one feature (14, 16); generating (140) a second low-intensity image by applying a second low-intensity filter to identify pixels in the second image that are below a threshold value, thereby obtaining second information of the at least one feature; calculating (150) a correlation between the first high-intensity image and the second low-intensity image to associate the first information and the second information of the at least one feature; outputting (160) the calculated correlation for identifying the at least one characteristic of the tissue sample; or generating (130) a first low-intensity image by applying a first low-intensity filter to identify pixels in the first image that are below a threshold value, thereby obtaining first information of the at least one feature (14, 16); generating (140) a second high-intensity image by applying a second high-intensity filter to identify pixels in the second image that are above a threshold value, thereby obtaining second information of the at least one feature; calculating (150) a correlation between the first low-intensity image and the second high-intensity image to associate the first information and the second information of the at least one feature; The calculated correlation is output (160) for use in identifying the at least one characteristic of the tissue sample.

2. The method according to claim 1, wherein The correlation is calculated based on Boolean operations.

3. A data processing device (200) for providing a method for identifying at least one feature of a tissue sample in a microscope image, comprising: Input unit (210); an information extraction unit (220); Correlation unit (230); as well as Output unit (240); wherein the input unit is configured to receive a first image of a fluorescence microscopy imaging modality representing a range of the tissue sample and a second image of a dark field microscopy imaging modality representing the range of the tissue sample; And among them, The information extraction unit is configured to generate a first high-intensity image by applying a first high-intensity filter to identify pixels above a threshold in the first image, thereby obtaining first information about the at least one feature, and generate a second low-intensity image by applying a second low-intensity filter to identify pixels below a threshold in the second image, thereby obtaining second information about the at least one feature; The correlation unit is configured to calculate a correlation between the first high-intensity image and the second low-intensity image to associate the first information and the second information of the at least one feature; The output unit is configured to output the calculated correlation for identifying the at least one feature of the tissue sample; Or among them, The information extraction unit is configured to generate a first low-intensity image by applying a first low-intensity filter to identify pixels below a threshold in the first image, thereby obtaining first information about the at least one feature, and generate a second high-intensity image by applying a second high-intensity filter to identify pixels above a threshold in the second image, thereby obtaining second information about the at least one feature; The correlation unit is configured to calculate a correlation between the first low-intensity image and the second high-intensity image to associate the first information and the second information of the at least one feature, The output unit is configured to output the calculated correlation for identifying the at least one feature of the tissue sample.

4. The data processing apparatus according to claim 3, wherein: The correlation unit is configured to calculate the correlation based on a Boolean operation.

5. A system (300) for providing support in identifying at least one feature of a tissue sample in a microscope image, comprising: The data processing device according to any one of claims 3 to 4; as well as Display (310); The display is configured to display a correlation between at least one of the first image and the second image and the first extracted information and the second extracted information of the at least one feature calculated and output from the data processing device.

6. A computer program product comprising a computer program for instructing, when executed by a processing unit, an apparatus according to any one of claims 3 to 4 to be adapted to perform the steps of the method according to any one of claims 1 to 2.

7. A computer-readable medium storing the computer program according to claim 6.

Images