Methods and systems for whole slide imaging

Abstract

A method (and system) for obtaining an adaptive calibration transform for whole slide imaging is provided, where the adaptive calibration transform is used to output the color of the blank area of the slide in the whole slide image (WSI) as the desired color. The method includes first receiving a WSI of a color target in a device - dependent color space, where the WSI of the color target is from a scanner represented in the device - dependent color space. The color target includes at least one white color patch and three (near) primary color patches. A calibration transform is obtained for transforming the device - dependent color space to a device - independent color space. A chromaticity adaptation transform is derived using the color of the white color patch in the device - dependent color space, the color of the blank (glass) area in the device - independent color space, and the desired blank (glass) color in the target color space. The adaptive calibration transform is obtained by applying the chromaticity adaptation transform to the calibration transform.

Description

Technical Field

[0001] The present invention relates to methods and systems for improving calibration of whole slide imaging. Specifically, the present invention relates to adapting a calibration transform to output a blank region of a slide in a whole slide image as a desired color. Background Art

[0002] Pathology slides containing stained samples of body tissue can be examined under a microscope to analyze the samples and diagnose abnormalities. The variations and number of characteristics of different organs, anatomical sites, and lesions mean that classifying abnormalities from pathology slides is a time-consuming task for doctors and requires a high level of experience to accurately classify abnormalities.

[0003] Digital pathology allows pathology slides to be converted into digital slides that can be viewed and analyzed on a computer. Digital pathology scanners acquire whole slide images (WSIs) of pathology slides, aiming to provide pathologists with an accurate and precise view of the slide content. Since the presence or absence of color in the WSI directly affects the performance of pathologists and the reliability of clinical diagnosis, the colors reproduced by the scanner should be accurate and consistent. In addition, the use of WSIs in computational pathology increases the need for more stringent color fidelity in the images.

[0004] Different experiments have shown that using color calibration in digital pathology can improve color accuracy and consistency. However, the visually unpleasant effects introduced by reference and sample mismatches, both in terms of intensity and white color cast, reduce the practicality of calibration. Attempts have been made to minimize the impact by adding empirical static gains to the system, but these may not be applicable to different systems and may increase the differences between different systems.

[0005] Color calibration in digital pathology scanners aims to provide optimal color reproduction (closest to the true sample color) for output devices such as display monitors, while reducing variations between scanners. Ideally, the color target (e.g., color model) used as a reference in color calibration should match the pathology slide characteristics. However, due to the lack of standardization, the targets used in digital pathology scanners are general targets, such as film-based IT8 targets. The physical differences between the calibration target and the pathology slide may introduce a colored white projection on the tissue image.

[0006] Therefore, there is a need for methods to provide a consistent and desired background color in whole slide imaging. Summary of the Invention

[0007] The present invention is defined by the claims.

[0008] According to an example of one aspect of the present invention, a method for obtaining an adaptive calibration transformation for whole slide imaging is provided, where the adaptive calibration transformation is used to output the color of the blank area of the slide in the whole slide image WSI obtained by the scanner as a desired color in a device-independent color space. The method includes:

[0009] Receiving a WSI of a color target in a device-dependent color space, where the WSI of the color target is a WSI of the color target from the scanner represented in the device-dependent color space, and where the color target includes at least one white color patch and three color patches;

[0010] Obtaining a calibration transformation for changing the device-dependent color space to a device-independent color space;

[0011] Determining a chromaticity adaptive transformation based on the color of the white color patch in the device-dependent color space, the color of the blank area in the device-dependent color space, and the desired color in the target color space (i.e., the color of the blank area of the slide); and

[0012] Obtaining an adaptive calibration transformation based on applying the chromaticity adaptive transformation to the calibration transformation.

[0013] The whole slide image shows the background color of the blank area (e.g., the location without tissue), but due to differences in RGB sensors, light sources, and operating conditions, the background color may vary among different scanners. The whole slide image is typically processed to a standard color space, mainly to enable the display to accurately show the image. The processing usually involves transforming the WSI using a mapping matrix or a lookup table obtained as a result of the scanner color calibration, which transforms the scanner-related WSI to a device-independent color space, and the device-independent color space can be further transformed to an output device color space, such as the display sRGB. However, such transformations commonly used in digital pathology do not reproduce the desired and consistent background color, especially the blank area of the slide.

[0014] According to the present invention, to adapt the background color, the chromaticity adaptive transformation is determined based on the color of the (glass) blank area of the slide and the desired color. The (final) chromaticity adaptive transformation transforms the white point of the scanner to an optimal white point, where the optimal white point is derived from adjusting the transformed color of the blank area of the slide (i.e., the glass area).

[0015] The chromaticity adaptive transformation is applied to the calibration transformation to obtain an adaptive calibration transformation. The adaptive calibration transformation can then be applied to any WSI, and the output image will have a background color (blank slide area), and the background color is represented as the desired color in the device-independent color space.

[0016] For the avoidance of doubt, unless otherwise stated, "device - related color space" refers to the color space used by a scanner. "Device - independent color space" is intended to refer to a color space that is independent of the scanner and independent of any output device. This device - independent color space can be a color space that is not normally used as an output color space (such as XYZ).

[0017] In this document, whenever the color of a feature is mentioned (such as the color of a white patch or the color of a glass area), this refers to the color in the specific color space that is currently being discussed.

[0018] A color target, for example, includes three color patches that are primary colors or close to primary colors. Obtaining a calibration transformation can then be based on mapping a set of original RGB values obtained by the scanner from the color target to a set of reference chromaticity values and obtaining a transformation function to calibrate the scanner.

[0019] The target color space can be the same color space as the device - independent color space or a different, output - device - related color space (such as the display sRGB). However, the device - independent color space is typically used for intermediate color transformations and is not visually intuitive. Therefore, the desired color is preferably defined in the output - device - related color space. For example, the neutral white in the displayed color white is defined as R = G = B.

[0020] Determining the chromaticity adaptation transformation can then include:

[0021] Applying the adaptive calibration transformation to the color of the blank area in the device - related color space;

[0022] Further transforming the color of the blank area to the target color space;

[0023] In the same target color space, determining the color difference between the color of the blank area and the desired color; and

[0024] Iteratively modifying the white point in the device - independent color space to obtain an adaptive white point in the device - independent color space such that the adaptive calibration transformation produces the minimum color difference.

[0025] Although the desired color is defined in the target color space, the chromaticity adaptation method utilizes the input and output white points in the device - independent color space. This is because the techniques used to calculate the chromaticity adaptation transformation (such as chromaticity adaptation matrices, for example, Bradford, Von Kries, and XYZ scaling) are all designed in the device - independent color space. Therefore, the adaptation of the chromaticity adaptation matrix occurs in the device - independent color space.

[0026] Modifying the scanner white point, for example, includes determining a chromaticity adaptation matrix in the case where a white point in a device-independent color space given as a source white point and an adaptive scanner white point as a target white point are included.

[0027] As described above, the final (optimal) scanner white point may not be intuitively definable in a device-independent color space. Therefore, the optimization process is not predetermined but results in an optimal scanner white point based on processing the actual and desired colors of the blank (glass) area of the slide in the target color space. This target color space can be visualized and made intuitive, for example, on a monitor or a printer.

[0028] The device-independent color space can be one of the following:

[0029] The XYZ color space; or

[0030] The Lab color space.

[0031] The target color space includes, for example, a linear color space and can include one of the following:

[0032] The sRGB color space;

[0033] The SWOP CMYK color space;

[0034] The Adobe RGB color space; and

[0035] The ProPhoto RGB color space.

[0036] Determining the chromaticity adaptation transformation can be based on one or more of the following:

[0037] The XYZ scaling method;

[0038] The Bradford method; or

[0039] The Von Kries method.

[0040] The desired color can be neutral white, and the target color space can be the sRGB color space, where neutral white is defined as R = G = B in the sRGB color space.

[0041] The present invention also provides a computer program, which includes code means for implementing any one of the above methods when the program runs on a processing system.

[0042] The present invention also provides a system for obtaining an adaptive calibration transformation for whole-slide imaging, where the adaptive calibration transformation is used to output the color of the blank area of the slide in the whole-slide image WSI obtained by the scanner as a desired color in a device-independent color space. The system includes:

[0043] A scanner for obtaining a WSI in a device - related color space;

[0044] A processor configured to:

[0045] Receive a WSI of a color target in a device - related color space, where the WSI of the color target is the WSI of the color target from the scanner represented in the device - related color space, and where the color target includes at least one white color patch;

[0046] Obtain a calibration transformation for transforming the device - related color space to a device - independent color space;

[0047] Determine a chromaticity adaptation transformation based on the color of the white patch in the device - related color space, the color of the blank area in the device - related color space, and the desired color in the target color space; and

[0048] Obtain an adaptive calibration transformation based on applying the chromaticity adaptation transformation to the calibration transformation.

[0049] The device - independent color space includes, for example:

[0050] The XYZ color space; or

[0051] The Lab color space.

[0052] The color target is composed of, for example, at least one white color patch and at least three color patches, and includes:

[0053] The IT 8.7 / 1 color target;

[0054] The Macbeth color checker;

[0055] The Sierra color target;

[0056] The Chromacal color target; or

[0057] The MGH color target.

[0058] The processor is, for example, adapted to determine the chromaticity adaptation transformation by:

[0059] Applying the adaptive calibration transformation to the color of the blank area in the device - related color space;

[0060] Applying a further transformation to the color of the blank area, further transforming to the target color space;

[0061] Determining the color difference between the blank area in the target color space and the desired color; and

[0062] Iteratively modify the white point in the device-independent color space to obtain an adaptive scanner white point in the device-independent color space such that the adaptive calibration transform produces a minimum color difference.

[0063] Modifying the scanner white point can include determining a chromatic adaptation matrix given a white point in the device-independent color space as the source white point and an optimal scanner white point as the destination white point.

[0064] The target color space includes, for example, a linear color space and includes one of the following:

[0065] The sRGB color space;

[0066] The SWOP CMYK color space;

[0067] The Adobe RGB color space; and

[0068] The ProPhoto RGB color space.

[0069] These and other aspects of the invention will become apparent from the embodiments described below and will be elucidated with reference to the embodiments described hereinafter. BRIEF DESCRIPTION OF THE DRAWINGS

[0070] For a better understanding of the present invention and to more clearly show how it may be implemented, reference will now be made, by way of example only, to the accompanying drawings, in which:

[0071] Figure 1 An example of a color target in a glass slide for whole slide imaging is shown;

[0072] Figure 2 An enlarged illustration of the difference between the original scanner acquisition and the scanner output after calibration is shown;

[0073] Figure 3 An enlarged illustration of the blank glass slide area and after calibration is shown;

[0074] Figure 4a A schematic diagram of a method for obtaining an adaptive calibration transform is shown;

[0075] Figure 4b The process of calculating the chromatic adaptation matrix is shown in more detail;

[0076] Figure 5 An enlarged illustration of the blank glass slide area 104 before and after calibration using the adaptive calibration transform is shown;

[0077] FIG. 6 shows the distribution of color target color patches in the chromaticity diagram; and

[0078] Figure 7Shows regions of the whole slide image before and after calibration. Detailed Description

[0079] The present invention will be described with reference to the accompanying drawings.

[0080] It should be understood that the detailed description and specific examples, while indicating exemplary embodiments of the devices, systems, and methods, are intended for illustrative purposes only and are not intended to limit the scope of the present invention. These and other features, aspects, and advantages of the devices, systems, and methods of the present invention will be better understood from the following description, the appended claims, and the drawings. It should be understood that the drawings are merely schematic and are not drawn to scale. It should also be understood that the same reference numerals are used throughout the drawings to represent the same or similar parts.

[0081] The present invention provides a method (and system) for obtaining an adaptive calibration transform for whole slide imaging, where the adaptive calibration transform is used to output the color of the blank region of the whole slide image (WSI) as a desired color. The method includes: first receiving a WSI of a color target in a device - related color space, where the WSI of the color target is from a scanner represented in the device - related color space. The color target contains at least one white color patch and three (near) primary color patches. A calibration transform is obtained for transforming the device - related color space to a device - independent color space. A chromaticity adaptive transform is derived using the color of the white color patch in the device - independent color space, the color of the blank (glass) region in the target color space, and the desired blank color in the target color space. The adaptive calibration transform is obtained by applying the chromaticity adaptive transform to the calibration transform.

[0082] The colors used in this document refer to the coordinates that make up a specified color in a color space. For example, the RGB color model can be used to refer to three - dimensional colors, and the color space can be one of sRGB, Adobe RGB, ProPhoto RGB, scRGB, CIE RGB, etc. The RGBA or CMYK color models can also be used.

[0083] Figure 1An example of a color target 102 in a glass slide 103 for whole slide imaging is shown. Due to the lack of color standardization in digital pathology, scanner manufacturers and / or users generally follow a common color workflow recommended by the International Color Consortium (ICC). The workflow helps transform the original scanner colors acquired in a device - related color space (e.g., RGB) to a device - independent color space (e.g., XYZ or LAB color space), so that an output device (e.g., an LCD monitor or a printer) can represent the colors acquired by the scanner without losing information. By mapping the original RGB values of different color patches in the color target 102 acquired by the scanner to the reference chromaticity values of the patches, a transformation function can be derived during color calibration. Some examples of color models 102 used in microscopy and digital pathology are IT8.7 / 1 color targets, Macbeth color checkers, Chromacal color targets, Sierra color targets, or MGH color targets.

[0084] Color targets (or models) 102 are imported from film - based analog cameras, flatbed scanners, or the filter field, and thus the samples do not fully represent general microscopy, especially digital pathology. Film - based IT8 and Macbeth color checkers are the most commonly used color targets 102 in digital pathology, with a total thickness of approximately 0.1 mm, while the thickness of a pathology glass slide 103 is approximately 0.002 mm - 0.006 mm. Therefore, the color target 102 transmits less light than the pathology glass slide. Similarly, Chromacal and MGH color targets 102 are based on chips and stained polymer films mainly used in filters, respectively. Figure 1 The color target 102 shown in is an example of a film color target 102 on a pathology glass slide 103. As Figure 1 shown, the film color target 102 can be wider than the pathology glass slide 103.

[0085] The mismatch between the substrate and color between the color of the pathology glass slide and the calibration target 103 results in an unnatural effect in the calibration matrix, making it look color - biased when calibrating the original image.

[0086] Figure 2 An enlarged view showing the difference between the original scanner acquisition and the scanner output after calibration is shown.

[0087] Figure 2 a shows the reference chromaticity values.

[0088] Figure 2 b shows the original scanner acquisition of the color target 102.

[0089] Figure 2 c shows the scanner output of the color target 102 after calibration.

[0090] Figure 2 For example, the sRGB color space representation of the IT8 white color patch (labeled GS0) is schematically shown based on its reference chromaticity values and the scanner output after the original scanner acquisition and calibration. Figure 2 The color of b (before calibration) and Figure 2 the color of c (after calibration) show the stained white. When the same calibration is applied to the blank slide area 104, it also appears to have another strong projection, resulting in an unnatural background for the tissue image.

[0091] Figure 3 Magnified views of the blank slide area 104 before and after calibration are shown.

[0092] Figure 3 a shows the original scanner acquisition of the blank slide area 104.

[0093] Figure 3 b shows the scanner output of the blank slide area 104 after calibration.

[0094] Figure 3 For example, the sRGB representation of the blank slide area 104 after the original scanner acquisition and calibration is schematically shown. For example, Figure 3 a may represent the average RGB values [0.74, 0.74, 0.72], while Figure 3 b represents the average RGB values [0.68, 0.75, 0.72]. This schematically shows that the calibration transformation obtained from the color target 102 and the reference values does not represent the blank slide area 104 in the desired (neutral white) color.

[0095] Therefore, the calibration transformation must be adapted so that when it is applied to the full slide image, the blank glass area 104 specifically becomes the desired color with the desired intensity, such as neutral white (R = G = B), while maintaining the object colors (e.g., tissue).

[0096] Figure 4a A schematic diagram of a method for obtaining an adaptive calibration transformation 416 is shown. The following example assumes that the transformation is a matrix, but a matrix is not required, and those skilled in the art will know that there are other methods for performing the transformation (e.g., look-up tables, TRC matrix transformations).

[0097] The first step is to perform a conventional color calibration.

[0098] Given the original RGB color patch color 106 of the color target 102 captured by the scanner and the reference chromaticity value 404 of the color patch (in the XYZ or Lab color space), a calibration matrix 406, e.g., of size 3x3, is calculated by minimizing the color difference between the transformed RGB color 106 and the reference chromaticity value 404. Matrix-based calibration is typically employed because it has superior perceptual quality in pathology slide images. However, the present invention is applicable to any calibration transformation.

[0099] The present invention then applies chromaticity adaptation to the calibration transformation, e.g., the calibration matrix 406.

[0100] The determination of the chromaticity adaptation matrix (or more generally other functions) is shown as step 414.

[0101] It receives the native scanner white (i.e., the color of the target white color patch 106 in the device - related color space) as input. This is referred to as "device - related white" in Figure 4a . By applying the final adaptive calibration matrix to the color target white color patch 106, the scanner white point in the device - independent color space can be derived from this device - related white.

[0102] The term "white point" is generally used to refer to the white representation in the device - independent color space, while the term "scanner white" refers to the color of the white color patch represented in the scanner color space.

[0103] Another input is the color of the blank slide 104 in the device - related color space (referred to as "device - related glass" in Figure 4a ).

[0104] Another input is the desired color 408 of the glass region in the target color space.

[0105] Another input is the calibration matrix 406.

[0106] The determination of the chromaticity adaptation matrix is, for example, an iterative minimization process that will be explained in detail below. The resulting chromaticity adaptation matrix is applied to the calibration matrix to derive the adaptive calibration matrix 416.

[0107] This adaptive calibration matrix can then be used to transform from the device - related color space to the device - independent color space, where the blank glass region is transformed to a color in the device - independent color space that corresponds to the desired color 408 in the target color space.

[0108] In this example, the device - related color space is RGB, the device - independent color space is XYZ, and the target color space is sRGB. The desired color 408 in the target sRGB color space can be set, for example, such that R = G = B and a comfortable viewing luminance level is obtained at 85% and intensity clipping is avoided in case the tissue content is brighter than the glass area.

[0109] Figure 4b Steps for determining the chromaticity adaptation matrix and the resulting adaptive calibration matrix are shown in more detail. It shows in more detail Figure 4a steps 414 and 416.

[0110] The process in this example is an iterative process and is implemented as a minimization problem. It generates the chromaticity adaptation matrix as a series of iterative adjustments aimed at minimizing the color difference in the target color space between the desired color 408 of the blank glass area and the actual blank glass color (in the target color space) obtained through the adaptive calibration matrix.

[0111] The target color space can be a device - independent color space. In such a case, the desired color is set in the XYZ color space. For example, the desired color in the color - independent color space (XYZ) can be selected or even derived using the inverse transform from the monitor color space.

[0112] However, Figure 4a and Figure 4b shows a preferred implementation where the target color space is the color space used by the output device to enable a more intuitive setting of the desired color.

[0113] For example, the XYZ device - independent color space is used for color transformation and is thus not visually intuitive. The desired color is therefore preferably defined in the target color space (such as sRGB, where neutral white is defined as R = G = B).

[0114] However, the chromaticity adaptation method is performed in the device - independent color space rather than in the target color space. It utilizes the transformation between the input and output white points in the device - independent color space. Thus, it can be considered a modification of the white point (i.e., in the device - independent color space).

[0115] In step 500, the native scanner white ("device - related white") is converted to the scanner white point in the device - independent color space (shown as "scanner white point XYZ") by using the calibration matrix 406.

[0116] The calibration matrix is only applied once to the scanned color target. As described below, the process finds a transformation (e.g., a matrix) that adapts to the scanner white point. The corresponding transformation is then applied to the calibration matrix. At the end of the process, this adaptive scanner white point can be considered the optimal scanner white point in the device-independent color space, and it corresponds to the mapping from the device-dependent color space to the device-independent color space, which results in a blank glass area with the desired color.

[0117] The scanner white point is adapted for the new scanner white point as a series of iterations in step 501. In each iteration, the next iteration of the resulting chromaticity adaptation matrix is calculated in step 502. It is applied to the calibration matrix 406 to derive the next iteration of the adaptive calibration matrix 503.

[0118] The next iteration of the adaptive calibration matrix is applied to the scanned blank glass area 104 in the device-dependent color space (RGB in this example), shown as "device-dependent glass" in Figure 4b to obtain the glass color in the device-independent color space, shown as "glass XYZ" in Figure 4b .

[0119] The transformation 503 from the device-independent color space (XYZ) is then performed to derive the glass color in the target color space (sRGB in this example). The color difference from the desired color 408 in the same target color space is then performed in step 505 to give an error value. This color error is calculated, for example, for the color difference between RGB (or XYZ) coordinates, using known DeltaE or Eucledian methods.

[0120] This error value is used to adjust the calculation of the chromaticity adaptation matrix in the next iteration. Thus, the error value is provided for the next iteration. For this repeated iterative process, the method returns to step 501.

[0121] The iterative processing of steps 501 and 502 involves the white point in the device-independent color space given as the source white point and the adaptive scanner white point as the destination white point to determine the chromaticity adaptation matrix. When the iteration is complete, the adaptive white point set in step 501 will correspond to the optimal scanner white point, which is the scanner white point that produces the desired glass color in the target color space.

[0122] The setting of the destination white point in step 501 can be achieved, for example, using the Nelder-Mead (downhill simplex method), which is a numerical method for finding the minimum or maximum of an objective function. The method can be repeated until it meets predefined (multiple) constraints, such as: the maximum number of iterations, the maximum number of function evaluations, and the tolerance of the difference between two consecutive function values.

[0123] Finally, a chromaticity adaptation matrix is obtained such that an optimal scanner white point in a device-independent color space is reached, and the optimal scanner white point corresponds to the desired color of the glass area in the target color space.

[0124] The final iteration of the adaptive calibration matrix 416 is output.

[0125] The error signal can be considered to correspond to the error from the optimal scanner white point, because the iteration ends when the adaptive scanner white point is optimal. Therefore, the adaptive calibration matrix is calculated based on the chromaticity adaptation matrix that transforms the native scanner white point of the scanner into the optimal scanner white point. Thus, this process optimizes the way the native scanner white point is processed.

[0126] From Figure 4b It can be seen that the adaptive calibration matrix 416 is translated into a device-independent color space. By applying an appropriate transformation from the device-independent color space to the desired target color space, the same adaptive calibration matrix can be used for different output devices, i.e., different target color spaces.

[0127] The obtained adaptive calibration matrix can thus be applied to a WSI image from a specific scanner with a specific slide format to achieve the desired output color of the glass area in a device-independent color space. The user can then apply a further transformation to map to their desired target color space. Thus, the obtained adaptive calibration matrix can be combined with a further transformation to achieve a direct translation from a device-dependent (scanner) color space to a desired target (e.g., output device) color space.

[0128] Before the WSI is displayed on a monitor, a Gamma transformation can be applied to the WSI in the output device (e.g., monitor) color space. A transformation from a linear target space to a desired output color space such as sRGB can also be performed.

[0129] Therefore, the process of deriving the adaptive calibration matrix is a calibration process that prepares a specific scanner type and slide type to be able to generate an image with known and desired output colors for the blank glass area.

[0130] Figure 5 An enlarged view of the blank slide area 104 before and after calibration using the adaptive calibration transformation 416 is shown.

[0131] Figure 5 a shows the original scanner acquisition of the blank glass area 104.

[0132] Figure 5b shows the scanner output of the blank glass area 104 after calibration using the adaptive calibration matrix 416.

[0133] Figure 5 Schematically shows the sRGB representation of the blank glass area 104 as the original scanner image and after applying the adaptive calibration matrix 416. Figure 5 a schematically represents the RGB coordinates [0.74, 0.74, 0.72], and Figure 5 b represents the average RGB coordinates [0.85, 0.85, 0.85], which correspond to the desired color 408 used in the optimization. This only represents one example of possible color variations.

[0134] Figure 6 shows the distribution of the color target patches in the chromaticity diagram.

[0135] Figure 6a shows the original color patches after applying the calibration matrix 406.

[0136] Figure 6b shows the original color patches after applying the adaptive calibration matrix 416.

[0137] The color target patches are given by black dots with label names. The triangle shown by the white dashed line represents the sRGB color space. Figure 6 shows the distribution of the IT8 color target patches collected by the scanner in the chromaticity diagram after only applying the calibration matrix 406 in Figure 6a and after applying the adaptive calibration matrix 416 in Figure 6b. It shows a slight change in the overall color distribution. For example, in Figure 6a, the color patch L18 602a seems to be located at the upper corner of the sRGB triangle, while in Figure 6b, the color patch 602b moves slightly to the lower right. However, the color patch coverage rate in the sRGB color space is comparable in both transformations. It shows that the chromaticity adaptation of the calibration matrix 406 does not limit the reproduction of colors in the sRGB color space.

[0138] Figure 7 Shows the areas of the full glass slide image before and after calibration.

[0139] Figure 7 a shows the original scanner acquisition of the WSI.

[0140] Figure 7 b shows the scanner output after applying the calibration matrix 406 to the WSI.

[0141] Figure 7 c shows the scanner output after applying the adaptive calibration matrix 416 to the WSI.

[0142] Compared with Figure 7 b, the blank glass slide area 104 in Figure 7In c, it appears as neutral white, looks more natural, and provides higher visual contrast. Figure 7 Shows a cropped area of the WSI, where in Figure 7 the uncalibrated original image of a and only in Figure 7 the image calibrated using the calibration matrix 406 in b, different shades of the blank glass slide area 104 can be seen. Using Figure 7 the same original image calibrated with the adaptive calibration matrix 416 in c shows a neutral white background specified by the desired color 408.

[0143] The adaptive calibration matrix enables, for example, a scanner with multiple cameras (e.g., high resolution and low resolution) to generate the same output color (especially the background glass color) in the target color space. The low-resolution camera can be used, for example, for tissue detection, and the high-resolution camera can be used for tissue scanning.

[0144] The system for obtaining an adaptive calibration transformation according to the present invention combines a scanner for obtaining a WSI in a device-related color space and a processor for performing the above-described processing steps.

[0145] A person skilled in the art will be able to easily develop a processor for performing any of the methods described herein. Thus, Figure 4a and Figure 4b each step of the flowchart can represent different actions performed by the processor and can be executed by the corresponding modules that process the processor.

[0146] The processor can be implemented in various ways using software and / or hardware to perform the various required functions. The processor typically employs one or more microprocessors, and one or more microprocessors can be programmed using software (e.g., microcode) to perform the required functions. The processor can be implemented as a combination of dedicated hardware for performing certain functions and one or more programmed microprocessors and associated circuitry for performing other functions.

[0147] Examples of circuits that can be employed in various embodiments of the present disclosure include, but are not limited to, conventional microprocessors, application-specific integrated circuits (ASICs), and field-programmable gate arrays (FPGAs).

[0148] In various implementations, the processor can be associated with one or more storage media, such as volatile and non-volatile computer memories, such as RAM, PROM, EPROM, and EEPROM. The storage media can be encoded with one or more programs that, when executed on one or more processors and / or controllers, perform the required functions. The various storage media can be fixed within the processor or controller or can be removable, such that the one or more programs stored thereon can be loaded into the processor.

[0149] By studying the drawings, the disclosure, and the appended claims, those skilled in the art can understand and realize variations of the disclosed embodiments when practicing the claimed invention. 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.

[0150] A single processor or other unit may implement the functions of several items recited in the claims. The fact that certain measures are recited in mutually different dependent claims does not indicate that a combination of these measures cannot be used advantageously.

[0151] A computer program may be stored / distributed on a suitable medium, such as an optical storage medium or a solid-state medium 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 telecommunication systems.

[0152] If the term "adapted to" is used in the claims or the description, note that the term "adapted to" is intended to be equivalent to the term "configured to".

[0153] Any reference signs in the claims should not be construed as limiting the scope.

Claims

1. A method for obtaining an adaptive calibration transform (416) for whole slide imaging, wherein the adaptive calibration transform (416) is used to output the color of the blank area of the slide (104) in a whole slide image WSI obtained by a scanner as a desired color (408) in a device-independent color space, the method comprising: Receive a WSI of a color target in a device - related color space, where the WSI of the color target is a WSI of a color target (102) from the scanner represented in the device - related color space, and where the color target (102) includes at least one white color patch (106) and three color patches; Obtain a calibration transform (406) for transforming the device - related color space to the device - independent color space; Determine a chromaticity adaptation transform (414) based on the color of the white color patch in the device - related color space, the color of the blank area in the device - related color space, the calibration transform (406), and the desired color in the target color space; and Obtain an adaptive calibration transform (416) based on applying the chromaticity adaptation transform (414) to the calibration transform (406); Where obtaining the calibration transform (406) is achieved based on calibrating the scanner, which is performed by mapping a set of original RGB values obtained by the scanner from the color target (102) to a set of reference chromaticity values (404) and obtaining the calibration transform (406); Where determining the final adaptive calibration transform includes performing the following steps iteratively: (501) Modify the white point in the device - independent color space to obtain an adaptive white point in the device - independent color space; (502) Calculate the next iteration of the chromaticity adaptation transform; (503) Apply the adaptive calibration transform to the color of the blank area in the device - related color space; (504) Apply a further transform to the color of the blank area to derive the color of the blank area in the target color space; and (505) Determine the color difference between the blank area in the target color space and the desired color in the target color space.

2. The method according to claim 1, wherein the device-independent color space comprises: XYZ color space; or Lab color space.

3. The method according to claim 1 or 2, wherein the color target (102) consists of at least one white color patch and at least three color patches, such as: IT 8.7 / 1 color target; Macbeth ColorChecker; Sierra color target; Chromacal color target; or MGH color target.

4. The method according to claim 1, wherein modifying the scanner white point comprises: Determine a chromaticity adaptation matrix given the white point in the device - independent color space as the source white point and the adaptive scanner white point as the destination white point.

5. The method according to claim 1 or 2, wherein the target color space comprises a linear color space and comprises: sRGB color space; SWOP CMYK color space; Adobe RGB color space; Or ProPhoto RGB color space.

6. The method according to claim 1 or 2, wherein determining the chromaticity adaptive transform (414) is based on one or more of the following: XYZ scaling method; Bradford method; and von Kries method 7. The method according to claim 1 or 2, wherein the desired color (408) is neutral white, and the target color space is the linear sRGB color space, and wherein neutral white is defined as R = G = B.

8. A computer program product comprising code means for implementing the method according to any one of claims 1 to 7 when the program is run on a processing system.

9. A system for obtaining an adaptive calibration transform (416) for whole slide imaging, wherein the adaptive calibration transform (416) is for outputting the color of the blank area of the slide (104) in a whole slide image WSI obtained by a scanner as a desired color (408) in a device-independent color space, the system comprising: A scanner for obtaining a WSI in a device - related color space; A processor configured to: Receive a WSI of a color target in a device - related color space, where the WSI of the color target is a WSI of a color target (102) from the scanner represented in the device - related color space, and where the color target (102) includes at least one white color patch (106); Based on calibrating the scanner, a calibration transform (406) for transforming the device - related color space into a device - independent color space is obtained. The calibration of the scanner is performed by: mapping a set of original RGB values obtained by the scanner from the color target (102) to a set of reference chromaticity values (404) to obtain the calibration transform (406); Determine a chromaticity adaptation transform (414) based on the color of the white patch in the device - related color space, the color of the blank area in the device - related color space, the calibration transform (406), and the desired color in the target color space; And Based on applying the chromaticity adaptation transform (414) to the calibration transform (406), an adaptive calibration transform (416) is obtained; Wherein the processor is adapted to determine a final adaptive calibration transform by iteratively performing the following steps: Modify the white point in the device - independent color space to obtain an adaptive white point in the device - independent color space; Calculate the next iteration of the chromaticity adaptation transform; Apply the adaptive calibration transform to the color of the blank area in the device - related color space; Apply a further transform to the color of the blank area to derive the color of the blank area in the target color space; and Determine the color difference between the blank area in the target color space and the desired color.

10. The system according to claim 9, wherein the device-independent color space comprises: XYZ color space; Or Lab color space, And wherein the color target (102) consists of at least one white patch and at least three color patches and includes: IT 8.7 / 1 color target; Macbeth color checker; Sierra color target; Chromacal color target; or MGH color target.

11. The system according to claim 10, wherein modifying the scanner white point comprises: Determine the chromaticity adaptation matrix given the white point in the device - independent color space as the source white point and the adaptive scanner white point as the destination white point.

12. The system according to any one of claims 9 to 11, wherein the target color space includes a linear color space and includes one of the following: sRGB color space; SWOP CMYK color space; Adobe RGB color space; or ProPhoto RGB color space.

Images