Measurement of image quality of blood cell images

Abstract

By subjecting one or more images taken by a visual analysis system to a multi-layer analysis, such a visual analysis system may be automatically focused (or the focus of such a system may be automatically corrected). In such an analysis, cell boundaries may be identified in an input image based on the brightness values ​​of the pixels of the input image. Based on the identified cell boundaries, a predicted nominal focus value may be determined, which may give a focus distance (e.g., the distance between the focal plane of the camera used when the image was taken and the actual focal plane corresponding to the in-focus image). This focus distance may then be used to (re)focus the camera or for other purposes (e.g., generating an alert).

Description

[Technical Field]

[0001] Cross-references to related applications This application claims priority to and is a formal application of Provisional Patent Application No. 63 / 305,890, entitled “Measure image quality of flow blood cell images,” filed with the United States Patent and Trademark Office on 2 February 2022. That application is incorporated herein by reference in its entirety. [Background technology]

[0002] Hematological analysis is one of the most widely performed medical tests to provide an overview of a patient's health. A blood sample can be taken from the patient's body and stored in a test tube containing an anticoagulant to prevent clotting. A whole blood sample typically contains three major classes of blood cells: red blood cells (erythrocytes), white blood cells (leukocytes), and platelets (thrombocytes). Each class can be further divided into subclasses of its components. For example, the five major types or subclasses of white blood cells (WBCs) have different shapes and functions. White blood cells may include neutrophils, lymphocytes, monocytes, eosinophils, and basophils. There are also subclasses of red blood cell types. The appearance of particles in a sample may vary depending on pathological conditions, cell maturity, and other causes. Subclasses of red blood cells may include reticulocytes and nucleated red blood cells.

[0003] This analysis may involve imaging samples containing blood cells, and the higher the quality of these images, the more suitable they are for the analysis. However, capturing high-quality images presents many challenges. For example, ensuring that images are in focus can be difficult because changes in temperature or other factors related to the operation of the analyzer may require refocusing of previously focused optics. In addition, some types of focusing are not effective for all types of blood cells that may be present in a sample (for example, a focusing method based on feature extraction may be suitable for red blood cells but not for white blood cells). Therefore, there is a need to improve techniques related to detecting out-of-focus images and / or automatic focusing of the analyzer's optics, including providing a fast and reliable method for evaluating the quality of focus and / or automatically refocusing as needed. [Overview of the Initiative]

[0004] Embodiments of the present disclosure may be used to determine the focal length of a camera based on an image representing one or more blood cells.

[0005] One embodiment may provide a system having a processor and an imaging device. Such a system may be configured to acquire multiple cells using the imaging device, with each of the multiple images containing at least one blood cell. Once acquired, the system identifies cell boundaries in at least one of the images. Based on the cell boundaries, the system can generate multiple rings, each of which is offset from the cell boundaries. Finally, the system can determine a predicted nominal focal value based on the brightness values ​​of the pixels located within the multiple rings.

[0006] In further embodiments, there may be a method by which multiple images are acquired from an imaging device. Each of the multiple images contains an image of at least one blood cell. A cell boundary is then identified in at least one of the images. Multiple rings are then generated based on the cell boundary, with each of the multiple rings offset from the cell boundary. A nominal focal value predicted for at least one image can then be determined based on the brightness values ​​of the pixels located within the multiple rings. Other embodiments are also disclosed.

[0007] This specification lists the present invention in detail and concludes with claims that expressly assert the invention, but the invention is considered to be better understood from the following description of specific examples in conjunction with the accompanying drawings. In the accompanying drawings, similar reference numerals indicate the same element. [Brief explanation of the drawing]

[0008] [Figure 1] This is a partially cross-sectional, schematic diagram, not to actual scale, illustrating the operation of an exemplary flow cell, autofocus system, and high optical resolution imaging device for analyzing sample images using digital image processing. [Figure 2] This is a diagram of a visual inspection system using slides, according to one embodiment. [Figure 3] This figure shows a process that may be used to refocus an imaging device according to one embodiment. [Figure 4] Figure 4A is an exemplary image of a blood cell with the focus shifted towards the "positive" side, according to one embodiment. Figure 4B is an exemplary image of a blood cell with the focus shifted towards the "negative" side, according to one embodiment. [Figure 5] This flowchart shows a method that may be used in an architecture related to one embodiment. [Figure 6] Figure 6A shows the identification of approximate cell boundaries according to one embodiment. Figure 6B shows the placement of rings on a blood cell image according to one embodiment. [Figure 7] It is a graphical representation of the average brightness values of rings on a blood cell image according to one embodiment. [Figure 8] It is a graphical representation of the average brightness values of rings on blood cell images with various focal points according to one embodiment. [Figure 9] It is a detailed view of a graphical representation of the average brightness values of rings on a blood cell image according to one embodiment. Mode for Carrying Out the Invention

[0009] The drawings are not intended to be limiting in any way, and it is contemplated that various embodiments of the present invention may be implemented in various other manners, including those not necessarily depicted in the drawings. The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate several aspects of the present invention and, together with the description, serve to explain the principles of the invention. However, it is to be understood that the present invention is not limited to the configurations shown.

[0010] The present disclosure relates to apparatuses, systems, and methods for analyzing blood samples containing blood cells. In one embodiment, the disclosed technology may be used in the context of an automated imaging system comprising an analyzer, which may be, for example, a visual analyzer. In some embodiments, the visual analyzer may further comprise a processor that facilitates automated conversion and / or analysis of images.

[0011] According to some aspects of the present disclosure, a system comprising a visual analyzer may be provided for obtaining an image of a sample containing particles (e.g., blood cells) suspended in a liquid. Such a system may be useful for characterizing particles in biological fluids, such as detecting and quantifying red blood cells, reticulocytes, nucleated red blood cells, platelets, and leukocytes, including, for example, leukocyte differential counting, categorization and subcategorization, and analysis. Other similar applications are also contemplated, such as characterizing blood cells derived from other fluids.

[0012] Identification of blood cells in blood samples is an exemplary application to which this subject is particularly suited, but other types of body fluid samples may be used. For example, embodiments of the disclosed techniques may be used for the analysis of non-blood body fluid samples containing blood cells (e.g., leukocytes and / or erythrocytes), such as serum, bone marrow, lavage fluid, exudate, cerebrospinal fluid, pleural fluid, peritoneal fluid, and amniotic fluid. It is also possible that the sample is a solid tissue sample, such as a biopsy sample processed to produce a cell suspension. The sample may also be a suspension obtained from processing a fecal sample. The sample may also be a laboratory sample or production line sample containing particles, such as a cell culture sample. The term sample may be used to refer to a sample obtained from a patient or laboratory, or a fraction, part, or aliquot thereof. In some processes, the sample may be diluted, divided into several parts, or stained.

[0013] In some embodiments, the sample is automatically presented, imaged, and analyzed. In the case of blood samples, the sample may be significantly diluted with a suitable diluent or physiological saline to reduce the degree to which some cells are obscured by other cells in the undiluted or low-diluted sample. Cells may be treated with agents that enhance the contrast of certain cellular features, such as permeabilizing agents that make the cell membrane permeable, or tissue stains that adhere to and highlight features such as granules and nuclei. In some cases, it may be desirable to stain aliquots of the sample for counting and characterizing particles including reticulocytes, nucleated red blood cells, and platelets, as well as for differentiation, characterization, and analysis of leukocytes. In other cases, samples containing red blood cells may be diluted before introduction into a flow cell and / or imaging in or by other means within the flow cell.

[0014] Detailed specifications of sample preparation apparatus and methods for sample dilution, permeabilization, and tissue staining may generally be achieved using precision pumps and valves operated by one or more programmable controllers. Examples of these may be found in patents such as U.S. Patent No. 7,319,907. Similarly, techniques for determining specific cell categories and / or subcategories by cell attributes such as relative size and color may be found in U.S. Patent No. 5,436,978 in relation to leukocytes. The disclosures of these patents are incorporated herein by reference in their entirety.

[0015] Next, looking at the drawings, Figure 1 schematically shows an exemplary flow cell 22 for transporting the sample fluid through the field of view zone 23 of a high optical resolution imaging device 24 in a configuration for imaging microscopic particles in a sample flow 32 using digital image processing. The flow cell 22 is coupled to a source 25 of the sample fluid, which may be treated such as contact with a particle contrast agent composition and heating. The flow cell 22 is also coupled to one or more sources 27 of particle and / or intracellular organelle alignment liquid (PIOAL), such as a clear glycerin solution having a viscosity higher than that of the sample fluid. Examples of PIOAL are disclosed in U.S. Patent Nos. 9,316,635 and 10,451,612, the entirety of which is incorporated herein by reference.

[0016] The sample fluid is injected into the flow cell 22 through a flat opening at the distal end 28 of the sample supply tube 29 once the PIOAL flow is substantially established, resulting in a stable, symmetrical laminar flow of PIOAL above and below (or on both sides of) the ribbon-shaped sample flow. The sample flow and PIOAL flow may be supplied by a precision measuring pump that moves the PIOAL along the injected sample fluid along a flow path that narrows considerably. The PIOAL surrounds and compresses the sample fluid within the zone 21 where the flow path narrows. Thus, the reduction in flow path thickness in zone 21 can contribute to the geometric focus of the sample flow 32. The sample flow 32, surrounded and transported downstream of the zone 21 that narrows with the PIOAL, passes in front of the high optical resolution imaging device 24 or otherwise passes through the field of view zone 23 of the device 24, where an image is acquired using, for example, a CCD 48. The processor 18 can receive pixel data as input from the CCD 48. The sample fluid ribbon flows along with PIOAL to discharge 33.

[0017] As shown here, the narrowing zone 21 can have a proximal flow channel portion 21a having a proximal thickness PT and a distal flow channel portion 21b having a distal thickness DT such that the distal thickness DT is less than the proximal thickness PT. Thus, the sample fluid can be injected through the distal end 28 of the sample tube 29 at a position distal to the proximal portion 21a and proximal to the distal portion 21b. Thus, the sample fluid can enter the PIOAL enclosure when the PIOAL flow is compressed by zone 21. Here, the sample fluid injection tube has a distal outlet port through which the sample fluid is injected into the flowing coating fluid, and the distal outlet port is defined by the reduction in the flow channel size of the flow cell.

[0018] A digital high-optical-resolution imaging device 24, equipped with an objective lens 46, is guided along an optical axis intersecting the ribbon-shaped sample flow 32. The relative distance between the objective lens 46 and the flow cell 33 can be varied by operating a motor drive 54 to resolve and collect a focused digital image on the photosensor array. Additional information regarding the construction and operation of an exemplary flow cell, as shown in Figure 1, is provided in U.S. Patent No. 9,322,752, filed March 17, 2014, entitled "Flowcell Systems and Methods for Particle Analysis in Blood Samples." The disclosures of that patent are incorporated herein by reference in their entirety.

[0019] The embodiments of the disclosed technology may be applied in contexts other than the flow cell system shown in Figure 1. For example, Figure 2 shows a slide-based visual inspection system 200 in which the embodiments of the disclosed technology may be used. In the system shown in Figure 2, a slide 202 containing a sample, such as a blood sample, is placed in a slide holder 204. The slide holder 204 may be adapted to hold several slides or just one slide, as shown in Figure 2. An imaging device 206, comprising an optical system 208 and an image sensor 210, is adapted to capture image data representing the sample on slide 202. Furthermore, a light-emitting device (not shown) may be used to control the light environment and thus obtain image data that is easier to analyze.

[0020] Image data captured by the image acquisition device 206 can be transferred to the image processing device 212. The image processing device 212 may be an external device, such as a personal computer, connected to the image acquisition device 206. Alternatively, the image processing device 212 may be integrated into the image acquisition device 206. The image processing device 212 may include a memory 216 and an associated processor 214 configured to determine the changes necessary to determine the difference between the actual focus and the correct focus of the image acquisition device 206. When determining the difference, a command may be sent to the steering motor system 218. Based on the command from the image processing device 212, the steering motor system 218 can change the distance z between the slide 202 and the optical system 208.

[0021] In a system as shown in Figure 1 or Figure 2, components such as a motor drive 54 or steering motor system 218 may be used in a process as shown in Figure 3 to continuously refocus the imaging device (e.g., a digital high optical resolution imaging device 24 or an image acquisition device 206) to maximize the quality of images generated for later analysis. In a process as shown in Figure 3, an image can be taken 301 and the image can be analyzed 302 to identify the deviation between the expected focal plane (i.e., the plane to which the imaging device was focused when the image was taken) and the correct focal plane (i.e., the plane to which the imaging device needed to be focused in order to take a focused image). If the deviation is zero (i.e., the plane to which the imaging device was focused when the image was taken was the correct focal plane), the process can continue taking images until all images required for analysis have been taken. Alternatively, if the deviation is not zero (e.g., there is a 2 μm difference between the expected focal plane and the correct focal plane), the focal plane of the imaging device can be adjusted to account for this deviation 303. For example, if the correct focal plane is 2 μm below the expected focal plane, the imaging device may be moved 2 μm closer to the sample using components such as the motor drive 54 or the steering motor system 218, or the sample may be moved 2 μm closer to the imaging device (for example, by moving the flow cell holding the sample) so that the expected and correct focal planes are the same in the next image. This process is then repeated until image acquisition is complete 304, thereby ensuring that deviations caused by factors such as temperature changes are detected and addressed.

[0022] As those skilled in the art will understand, a typical human white blood cell (WBC) can be considered a sphere when recorded by the optical elements of a flow imaging device, and is captured as a two-dimensional (2D) circular cell image with internal structure. Alternatively, a typical human red blood cell (RBC) is a biconcave disk, with a thickness of 0.8–1 μm at the center and 2–2.5 μm at the edge, and when viewed at the correct focus, the brightness contrast between the center and the edge decreases. Therefore, creating a system that can capture and analyze the different physical characteristics of both RBCs and WBCs is a challenging task.

[0023] One of the main obstacles in evaluating WBC images is that WBCs have various nuclei and granules at their cell boundaries, which can make the application of image pattern recognition difficult. Briefly referring to Figures 4A-4C, when WBCs are captured in focus, the image includes relatively sharp cell boundaries and details inside the cell, as shown in Figure 4B. However, if the system captures out-of-focus images (e.g., Figures 4A and 4C), the boundaries may be blurred, and the internal structure of the cell may become indistinguishable. Furthermore, because the system does not have a standard WBC template, the unknown properties of the internal structure complicate automated image analysis (e.g., determining whether an image is in focus or out of focus).

[0024] To overcome these problems, the systems and / or methods disclosed herein may rely on the interaction between the illumination device and the WBC. For example, in some embodiments, the illumination device may generate various “halo” patterns around cell boundaries as the quality of focus fluctuates (see Figures 4A–4C). These “halo” patterns are independent of the internal structure of the WBC and can therefore be used during image processing (i.e., as discussed herein) to evaluate images of any / all WBC subtypes, e.g., neutrophils, monocytes, lymphocytes, etc. More specifically, in some embodiments, the “halo” patterns may be characterized by a ring-based feature extraction process and analyzed using this feature extraction process. This feature extraction process may involve analyzing the image in a specific color space (e.g., a color space with lightness values).

[0025] For illustrative purposes only, most of the examples in this disclosure and herein focus on the HSV color space and / or HSL color space. However, these are non-limiting examples, and other color spaces such as HSV, HSL, Munsell color system, LCh, NCS, and CIELCH are also included. uv , CIELCH ab It should be understood that other color spaces may be used, such as CIECAM02, or current or future color spaces for evaluating lightness levels / values. It should also be understood that while the systems / methods discussed herein primarily rely on lightness levels, some embodiments may utilize color spaces that do not have lightness factors. For example, in some embodiments, the system may utilize / combine the characteristics of the RGB color space to provide similar / equivalent information to lightness levels.

[0026] As discussed herein, the system takes images over time, for example, using the image acquisition device 24 in Figure 1. When an image is taken out of focus, the image is either "positive" or "negative" in focus, depending on whether the image acquisition device is too close or too far away. For example, Figure 4A represents an image that is "positively" focused, meaning the image acquisition device is too close to the cell compared to the ideal position, while Figure 4C represents an image that is "negatively" focused, meaning the image acquisition device is too far from the cell compared to the ideal position.

[0027] Referring next to Figure 5, an exemplary process for measuring the image quality of a flow hematopoietic image is shown. As described above, before image analysis is performed, the image may be placed in a color space that facilitates the extraction of lightness values. Thus, in some embodiments, the imaging device may be a specialized imaging device capable of capturing images in an appropriate color space 501. As an addition or alternative, the system may convert the image captured by the imaging device to an appropriate color space before analysis 501. Regardless of the method used, once the image is in an appropriate color space, the system may separate the foreground of the image from the background of the image by evaluating lightness values ​​from the image (e.g., V for HSV or L for HSL) 502. As discussed herein, even when the focus is off from the WBC or RBC due to the “halo” effect (shown in Figures 4A, 4B, and 4C), it is possible to determine where the cells (i.e., foreground) end and where the biological fluid (i.e., background) begins.

[0028] More specifically, in order to separate the foreground from the background of an image 502, the system may evaluate each pixel of the captured image against a threshold brightness value (e.g., V=204). Briefly referring to Figure 6A, in some embodiments, if a pixel is above the threshold, it is defined as the foreground 602, and if it is below the threshold, it is defined as the background 601. In some embodiments, the threshold may be product-specific, or in other words, two systems may have different thresholds based on their respective design factors. For example, two products may have different lighting devices or different image capturing devices, and therefore, images captured by those systems may require different thresholds.

[0029] Therefore, in some embodiments, the brightness threshold may be determined / identified during product design or manufacturing through training with machine learning / human insights. In alternative embodiments, the system may automatically (i.e., without human insights) determine the brightness threshold using one or more known image analysis techniques. For example, this could involve creating a histogram of brightness values ​​to determine the most accurate brightness value or brightness range associated with the most significant brightness transitions.

[0030] In further embodiments, the system may begin evaluating pixels from the center point of the image and then expand outwards. Thus, since the image being analyzed should contain a single cell 501 (for example, because it was taken as a single-cell image or extracted from a multi-cell image), starting the evaluation from the center can reduce the number of pixels that need to be analyzed. In other words, once the system has determined the boundary of the entire cell 603, the system can proceed without having to analyze the rest of the image outside. It should be understood that various other edge detection or foreground segmentation methods may be used to separate the two 502. In some embodiments, the interface or intersection of the background 601 and foreground 602 of the captured image is defined as the cell boundary 603. Thus, as discussed herein, the system may create and / or superimpose virtual cell boundaries within the image based on the determined foreground and background 503.

[0031] An approximate cell boundary 603 is determined 503, and multiple rings may be generated, which may be offset from the cell boundary 504. Thus, in some embodiments, as also shown in Figure 6B, the system may determine the cell boundary 603, and multiple rings (e.g., 610, 611, 612, 613, 614, and 615) may be generated 504. The rings 610, 611, 612, 613, 614, and 615 are for illustrative purposes only, and it should be understood that these should not be considered as the only rings identified / created by the system. Thus, although only a total of six rings are shown, in some alternative embodiments there may be many more rings (e.g., 7, 8, 9, 10, 50, 100, 1,000, 10,000, 100,000, etc.). In other embodiments, fewer rings may be used (e.g., 2 to 5 rings).

[0032] As shown in Figure 6B, the system may utilize smaller rings (e.g., 613, 614, and 615, respectively) that are one, two, or three levels below the cell boundary, as well as larger rings (e.g., 612, 611, and 610, respectively) that are one, two, or three levels below the cell boundary. In some embodiments, the position and / or size of the rings are based on the approximate cell boundary 603. For example, one method for creating / identifying one or more smaller rings is to apply morphological image erosion from the cell boundary. Alternatively, creating / identifying one or more larger rings may involve applying morphological image dilation from the cell boundary. In yet another embodiment, the system may fit the cell boundary to an approximate ellipse. The approximate ellipse is then enlarged and / or reduced to generate the rings. In further embodiments, the enlargement and reduction of the approximate ellipse may be based on an offset distance or coefficient. For example, an approximate ellipse may be scaled based on units of measurement (e.g., 1 μm, 0.1 μm, etc.) or pixel values ​​(e.g., 1 pixel, 5 pixels, 10 pixels, etc.).

[0033] Returning to Figure 5, multiple rings are generated around the cell boundary 504, and the system then determines the average brightness value for each group of pixels between all adjacent rings 505. Specifically, the system determines the average brightness value for all pixels captured between two rings 505. For example, averaging all brightness values ​​of pixels between ring 610 and ring 611 determines a single average brightness value for the area between 610 / 611 505 (i.e., bin index 7 in Figure 7). Similar calculations are performed for the areas between ring 611 and ring 612, between ring 612 and ring boundary 603, between ring boundary 603 and ring 513, between ring 613 and 614, and between ring 614 and 615.

[0034] 505 All average lightness values ​​are calculated for each pair of adjacent rings, and the system identifies various image characteristics based on the average lightness values. 506 In some embodiments, as shown in Figure 7, the system may plot the average lightness values ​​against the bin index (e.g., in a Cartesian coordinate system). While various plots / graphs are discussed and shown in this disclosure, it should be understood that they are for illustrative purposes only. Thus, as will be understood by those skilled in the art, in practice, such plots / graphs do not have to be generated, and instead, a function or other numerical representation relating the index to lightness may be created and analyzed using the disclosed technique. In other words, the disclosed system and / or method may or may not provide any kind of externally visible plot, such as those shown in Figures 7-8.

[0035] As a non-limiting example, Figure 7 shows an embodiment in which a bin index value of 1 is assigned to the area inside the smallest circle (i.e., ring 615), and a lightness value of approximately 0.5 (i.e., V value) is determined. The determined average lightness value is then plotted and associated with bin index 1 of 701. In further embodiments, as shown, the remaining areas (e.g., rings 611 and 612, ring 612 and ring boundary 603, ring boundary 603 and ring 613, rings 613 and 614, and the area between rings 614 and 615) are added to the plot with their respective values. A lightness curve is created by plotting the bin index on the X axis and the corresponding lightness value on the Y axis 710. In an alternative embodiment, the system may plot the bin index on the Y axis and the corresponding lightness value on the X axis. Furthermore, in some embodiments, the values ​​on the X axis may increase from the outside rather than from the inside.

[0036] Referring to Figure 8, three cell images (i.e., 810, 820, and 830) are shown alongside their respective graph plots (i.e., 811, 821, and 831). As shown, image 810 represents a positively out-of-focus image (i.e., the imaging device is too close to the cell compared to the ideal position, "positively focused"), image 820 represents a focused image, and image 830 represents a negatively out-of-focus image (i.e., the imaging device is too far from the cell compared to the ideal position, "negatively focused"). Thus, in some embodiments, as shown, different brightness curves are created for each focus level when the brightness values ​​are plotted as a function of the bin index.

[0037] Therefore, in some embodiments, the system can analyze the function of each curve (e.g., 811, 821, and 831) to determine whether the image is out of focus, and if so, to what extent and in what direction (i.e., positive or negative). Thus, the system can identify various image characteristics (e.g., inflection points, left marks, and right marks) based on the function / curve of the average brightness value compared to the number of bin indices.506 As a non-limiting example, Figure 9 shows a magnified representation of plot 821 (i.e., a graph plot of the in-focus image 820).

[0038] In some embodiments, as shown in Figure 9, the inflection point 901 may be identified / determined as the point where the V value (i.e., the brightness value) changes most rapidly, and this point may be determined by taking the first derivative of the V curve. In further embodiments, the left mark 902 may be identified / determined as the trough to the left of the inflection point. In alternative embodiments, or if no trough is found to the left of the inflection point, the left mark may be calculated as the peak of the second derivative of the V curve to the left of the inflection point 901. In another embodiment, the right mark 903 may be identified / determined as the peak to the right of the inflection point. In alternative embodiments, or if no peak is found to the right of the inflection point, the right mark may be calculated by taking the second derivative of the V curve to the right of the inflection point 901.

[0039] In some embodiments, once the inflection point 901, left mark 902, and right mark 903 are determined, the system may then calculate several numerical features. More specifically, the system may calculate the following: a. "Inflection_right_mark_distance". This is equal to the distance between inflection point 901 and right mark 902 on the V curve. b. "Left_right_mark_distance". This is equal to the distance between the left and right marks on the V-curve. c. "V_right_mark_value". Equals to the V value of the right mark. d. "V_left_mark_value". Equals to the V value of the left mark.

[0040] Once the characteristics are identified 506 and / or the numerical features are calculated, in some embodiments the system may calculate a predicted nominal focal value 507. As a non-limiting example, if the determined nominal focal value is equal to 0.0 μm, the imaging device can be assumed to be in perfect focus (i.e., the system is in focus). Or, when the quality of focus deteriorates and the nominal focal value moves from 0.0 μm in either the positive or negative direction. The system may use the four numerical features described above (i.e., inflection_right_mark_distance, left_right_mark_distance, V_right_mark_value, and V_left_mark_value) to correlate with the nominal focal value and measure the quality of focus. In other words, the system may use the numerical features described above to predict the nominal focal value.

[0041] In some embodiments, the system may mathematically define a function that takes numerical features as input and outputs a predicted nominal focus in the following form: predicted nominal focus = f(inflection_right_mark_distance, left_right_mark_distance, V_right_mark_value, V_left_mark_value). The specific process that this type of function may use to provide the output may vary from case to case. For example, it may be defined using a neural network, a support vector machine, polynomial regression, linear regression, etc. As a non-restrictive example, the system may use a quadratic polynomial regression defined as follows: Expected nominal focus = -6.8694e-01 * inflection_right_mark_distance + 7.1640e-02 * left_right_mark_ distance + -9.8570e+02 * V_right_mark_value + 1.6144e+01 * V_left_mark_value + 2.9647e-02 * inflection_right_mark_distance * inflection_right_mark_distance + -7.2789e-03 * left_right_mark_distance * left_right_mark_distance + 5.6250e+02 * V_right_mark_value * V_right_mark_value + -1.0997e+01 * V_left_mark_value * V_left_mark_value + 4.2827e+02 Other similar formulas may also be used. For example, instead of using the values ​​shown in the example formula above, the quadratic polynomial may be expressed in the following form: F Predicted = LC IRMD * inflection_right_mark_distance + LC LRMD * left_right_mark_distance + LC VRMV* V_right_mark_value + LC VLMV * V_left_mark_value + SC IRMD * inflection_right_mark_distance 2 + SC LRMD * left_right_mark_distance 2 + SC VRMV * V_right_mark_value 2 + SC VLMV * V_left_mark_value 2 In the above exemplary formula, F Predicted is the predicted nominal focus. LC IRMD is -6.8*10 -1 to -6.9*10 -1 real number. LC LRMD is 7.1*10 -2 to 7.2*10 -2 real number. LC VRMV is -9.8*10 2 to -9.9*10 2 real number. LC VLMV is 1.6*10 1 to 1.7*10 1 real number. SC IRMD is 2.9*10 -2 to 3.0*10 -2 real number. SC LRMD is 7.2*10 -3 to 7.3*10 -3 real number. SC VRMV is 5.6*10 2 to 5.7*10 2 is a real number, SC VLMV is 4.2*10 2 to 4.3*10 2 is a real number. Other formulas and classes of formulas may also be used, therefore, it should be understood that the above formulas, whether expressed as ranges or discrete values, are merely illustrative and should not be treated as limiting.

[0042] Information that can be used as a basis for deriving the formulas shown above, whether from polynomial regression, neural networks, or other sources, can be obtained in various ways. For example, ground truth images for training a system to generate predicted nominal focal values ​​can be obtained through human annotation of images generated during the normal operation of the analyzer (e.g., a human examines the images and, based on their experience, labels any differences between the actual focal plane and the optimal focal plane, thereby training the system to distinguish between in-focus and out-of-focus cell images), but they can also be obtained in other ways. For example, by using an analyzer, it is possible to intentionally change the relationship between the image and the sample being imaged after an in-focus image has been taken, thereby capturing both an in-focus image and an image that is out of focus by a known amount.

[0043] Table 1 shows examples of how intentional offsets of the ideal focal length can be used as part of a training procedure. In various examples, the camera or camera lens is set to a first ideal focal position to photograph focused blood cells. The camera or camera lens is then offset in either direction to establish a training set of out-of-focus data. For example, the camera or camera lens may start from position X, which correlates to the ideal focal quality position (e.g., zero offset). It may then be offset in both directions by fixed intervals (e.g., intervals of 0.1 microns, 0.2 microns, 0.3 microns, 0.4 microns, or 0.5 microns), for example, -1 to +1 micron in each direction, -2 to +2 microns in each direction, -3 to +3 microns in each direction, -4 to +4 microns in each direction, or -5 to +5 microns in each direction. In the context of Table 1, X represents the starting position, and n represents the offset increment (e.g., 0.3 microns) that defines the fixed interval that the camera offsets in each sample run. Other methods are also possible, such as moving in variable increments, moving in different increments for different directions (e.g., moving away from the flow cell in 0.3 micron increments and closer to the flow cell in 0.2 micron increments), or obtaining images from a different number of positions than those shown in Table 1 (e.g., moving 6n closer to the flow cell and 4n further away from the flow cell). Different types of training data preparation are also possible, such as providing a set of images to a human reviewer and asking that reviewer to specify the offset distance for each image. Therefore, the explanation of how an intentional offset of the ideal focal length can be used as part of the training procedure should be understood as illustrative only and should not be treated as suggesting any limitations on the protections provided by this document or related documents.

[0044] Training data can be used to establish a reliable source of truth (e.g., images correlated to an ideal focal position, or images correlated to a certain degree of offset), as described above. A relevant formula or algorithm used to establish the expected nominal focus is trained, in which the parameters / function elements constituting the formula or algorithm for the predicted nominal focus are refined throughout the training process to match the reliable source of truth. In this way, a final, definitive formula or algorithm for the predicted nominal focus is established.

[0045] In some examples, this training step is performed for each distinct group of blood cells, such as red blood cells in one sample and white blood cells in another, so that the system is trained to identify the quality of focus from smaller cells (e.g., red blood cells) and larger cells (e.g., white blood cells). The various types of cells used to train the system may include red blood cells, platelets, and various groups of white blood cells (neutrophils, lymphocytes, monocytes, eosinophils, and basophils). In other examples, the system is trained on only specific types of cells (e.g., only red blood cells, only white blood cells, or only specific types of white blood cells such as only neutrophils).

[0046] [Table 1]

[0047] Therefore, in some embodiments, the predicted nominal focus may be calculated for any single WBC image in a blood sample, and thus the system can measure the focus / image quality of any single WBC image. In further embodiments, the predicted nominal focus may then be used in downstream components. As a non-limiting example, the downstream component may be an image classification algorithm that can label WBC images based on the biological properties of the WBC images. Furthermore, if the system determines that the focus / image quality is poor in a single image or across multiple images, the system may flag the classification result and / or invalidate the classification result. In fact, we intend that focus evaluation and / or autofocus as described herein may be used to improve image quality in any type of device that captures flow images of blood cells. It should also be understood that the disclosed systems and / or methods may utilize images captured as cells are presented through an imaging region (for example, using a flow-cell system with an imaging device configured to capture images of cells as they pass through an analysis region of the flow cell), and the cells may be one at a time or may include multiple cells presented and captured within a single frame, and then the cellular images may be separated at the cellular level using a computational methodology. Accordingly, the above figures and related discussions should not be used to assert the interests of this document or to suggest limitations on the scope of protection provided by other documents otherwise relating to this document.

[0048] In another embodiment, the predicted nominal focus may be available at the sample level. Furthermore, statistics such as the mean, median, percentile, and / or other mean of the predicted nominal focus of all or some WBC images may indicate whether a systematic shift in focus occurred during the imaging process. In addition, the variability or dispersion of the predicted nominal focus of all or some WBC images may provide information related to the stability of the flow during image acquisition. Thus, as described above, various metrics associated with one or more images may be stored to enhance the usefulness of the images downstream.

[0049] In another embodiment, the system may automatically perform proactive steps to correct the predicted nominal focus. For example, the system may perform an autofocus process, thereby adjusting the image acquisition device or other components to improve the nominal focus. For example, the system may use mechanical parts to adjust the position of the image acquisition device based on the predicted nominal focus. Thus, the predicted nominal focus value can serve as an indicator of focus quality; that is, the closer the absolute value of the nominal focus is to 0.0 μm, the better the focus quality. In some embodiments, the sign of the predicted nominal focus value may indicate the direction of the system that is out of focus.

[0050] The architectures illustrated and discussed in the context of Figures 1 to 9 may be implemented in various ways. For example, such an architecture could be implemented as a convolutional neural network (CNN) that utilizes a regression method to better measure focus quality. In this type of implementation, the specific values ​​used to analyze the image and generate the output can be optimized by training the neural network to minimize the regression error between the known focus position and the calculated focus position using blood cell images with known focus positions.

[0051] Modifications of the methods that may utilize focusing techniques as described herein are also possible. For example, an autofocusing process as described herein may be implemented to run a series of samples to determine how the camera should be focused, adjusting the focus on run units rather than on image units. Similarly, instead of automatically refocusing the camera, a warning may be generated using a certain focus position (e.g., if the difference between the expected and correct focal planes exceeds a threshold, or if there is a trend indicating that the focus is fluctuating), and the user may then decide whether to refocus the analyzer or continue the current imaging operation. Autofocusing as described herein may also / alternatively be incorporated into a periodic (e.g., routine) quality control process. Data collected during autofocusing may then be used to improve the operation of the system. For example, if it is found that adjustments made during autofocusing are consistently unidirectional, this may be used as a diagnostic indicator that there is a system defect in the mechanical or optical components of the analyzer that, when corrected, could reduce the need for automatic refocusing. As another example of how autofocus as described herein may be applied, consider that in some cases, even if the focus is acceptable, different focus positions within the acceptable range may result in slightly clearer perception of different features in the image. In such cases, the focus information may be used to characterize the image captured by the system (for example, as being closer to or further from the sample, while still within the acceptable range) so that downstream processing can be optimized as needed depending on which features are detected (for example, by applying a sharpening kernel if it may be more difficult to identify a particular feature based on the characterization). Therefore, the image-by-image autofocus described above should be understood as merely illustrative and should not be treated as suggesting any limitations on the protections provided by this document or related documents.

[0052] Modifications are possible in how the focusing methods described herein may be implemented. For example, in some cases, the method shown in Figure 5 may include additional steps to facilitate image processing and use. Examples of such additional steps include pre-processing or reformatting the image before analysis. In an example where processing is performed using a CNN trained to accept a specific type of image (e.g., a 128×128×3 RGB image) as input, this may include resizing the image to match the training input and converting the image to an appropriate color space before analysis. Thus, specific implementations may vary from case to case depending on the potential context and application in which the disclosed technology may be used, and the examples, figures, and descriptions described herein should be understood as non-limiting.

[0053] As further examples of possible implementations and uses of the disclosed technology, the following examples are provided as non-exhaustive ways in which the teachings herein may be combined or applied. It should be understood that the following examples are not intended to limit the scope of claims that may be presented at any time in this application or in subsequent applications. No waiver is intended. The following examples are provided solely for illustrative purposes. Various teachings herein are intended to be constructed and applied in numerous other ways. Also, some modifications are intended to omit certain features mentioned in the following examples. Accordingly, none of the embodiments or features mentioned below should be considered essential unless specifically and expressly indicated thereafter by the inventors or their successors. If any claims are presented in this application or a subsequent application relating to this application that include additional features beyond those mentioned below, those additional features should not be considered added for patentability reasons.

[0054] Example 1 A system comprising: a processor; an image acquisition device; and a non-temporary computer-readable medium storing instructions causing the processor to perform a set of operations including: acquiring a plurality of images from the image acquisition device, each of which contains blood cells; identifying cell boundaries in at least one image; generating a plurality of rings based on the cell boundaries, each of which is offset from the cell boundaries; and determining a predicted nominal focal value for at least one image based on the brightness values ​​of pixels located within the plurality of rings.

[0055] Example 2 The system in Example 1 determines the predicted nominal focal value for at least one image, further based on the brightness values ​​of pixels located between each pair of adjacent rings in a group of rings.

[0056] Example 3 The system of Example 1 further includes acquiring multiple images and converting each of the multiple images into a color space with brightness values.

[0057] Example 4 The system of Example 1 further includes identifying cell boundaries in at least one image, and separating at least one image into foreground and background based on predetermined brightness values.

[0058] Example 5 The system of Example 1 further comprises generating multiple rings by using morphological expansion of cell boundaries to generate at least one larger ring, where an additional larger ring is generated by using morphological expansion of a previously generated larger ring, and generating at least one smaller ring by using morphological erosion of cell boundaries, where an additional smaller ring is generated by using morphological erosion of a previously generated smaller ring.

[0059] [Example 6] The system of Example 1 further comprises at least one of the following: generating multiple rings by identifying the elliptical shape that best fits based on cell boundaries and generating at least one larger ring based on the offset distance, wherein the additional larger ring is generated from the previously generated larger ring based on the offset distance; and identifying the elliptical shape that best fits based on cell boundaries and generating at least one smaller ring based on the offset distance, wherein the additional smaller ring is generated from the previously generated smaller ring based on the offset distance.

[0060] Example 7 The system of Example 1 further includes determining a nominal focal value predicted for at least one image based on the brightness values ​​of pixels located within a plurality of rings, by performing operations including: generating a V-curve of average brightness values ​​associated with each area between two adjacent rings for each known distance of the plurality of rings from the cell boundary; identifying an inflection point of the V-curve, where the inflection point is equal to the peak of the first derivative of the V-curve; identifying a left mark of the V-curve, where the left mark is equal to the peak of the second derivative of the V-curve to the left of the inflection point; and identifying a right mark of the V-curve, where the right mark is equal to the trough of the second derivative of the V-curve to the right of the inflection point.

[0061] Example 8 The system in Example 7 involves determining the predicted nominal focal value for at least one image, further comprising calculating the distance between the inflection point and the right mark, the distance between the left mark and the right mark, the V value for the right mark, and the V value for the left mark, such that the predicted nominal focal value is a function of the distance between the inflection point and the right mark, the distance between the left mark and the right mark, the V value for the right mark, and the V value for the left mark.

[0062] Example 9 The system of Example 1, further comprising a set of actions that disable at least one image based on the predicted nominal focal value.

[0063] Example 10 The system of Example 1, further comprising: obtaining a set of operations to obtain a set of predicted nominal focal points, each nominal focal point corresponding to a different image among a set of images; determining the median of the set of predicted nominal focal points; and translating the image acquisition device based on the median of the set of predicted nominal focal points.

[0064] Example 11 The system of Example 1, further comprising: obtaining a set of operations to obtain a set of predicted nominal focal points, each nominal focal point corresponding to a different image among a set of images; determining the median of the set of predicted nominal focal points; and evaluating the flow stability of a blood sample containing blood cells based on the median of the set of predicted nominal focal points.

[0065] Example 12 A method comprising: acquiring multiple images from an imaging device, each of which contains blood cells; identifying cell boundaries in at least one of the images; generating multiple rings based on the cell boundaries, each of which is offset from the cell boundaries; and determining a predicted nominal focal value for at least one of the images based on the brightness values ​​of pixels located within the multiple rings.

[0066] Example 13 The method in Example 12 determines the predicted nominal focal value for at least one image, further based on the brightness values ​​of pixels located between each pair of adjacent rings in a group of rings.

[0067] Example 14 The method of Example 12 further includes identifying cell boundaries in at least one image, and separating at least one image into foreground and background based on predetermined brightness values.

[0068] Example 15 The method of Example 12 further comprises generating multiple rings by generating at least one larger ring using morphological expansion of cell boundaries, where an additional larger ring is generated by generating at least one larger ring using morphological expansion of a previously generated larger ring, and generating at least one smaller ring using morphological erosion of cell boundaries, where an additional smaller ring is generated by generating at least one smaller ring using morphological erosion of a previously generated smaller ring.

[0069] Example 16 The method of Example 12 further comprises at least one of the following: generating multiple rings by identifying the elliptical shape that best fits based on cell boundaries and generating at least one larger ring based on the offset distance, wherein the additional larger ring is generated from the previously generated larger ring based on the offset distance; and identifying the elliptical shape that best fits based on cell boundaries and generating at least one smaller ring based on the offset distance, wherein the additional smaller ring is generated from the previously generated smaller ring based on the offset distance.

[0070] Example 17 The method of Example 12 further includes determining a nominal focal value predicted for at least one image based on the brightness values ​​of pixels located within a plurality of rings, generating a V-curve of average brightness values ​​associated with each area between two adjacent rings for each known distance of the plurality of rings from the cell boundary, identifying an inflection point of the V-curve where the inflection point is equal to the peak of the first derivative of the V-curve, identifying a left mark of the V-curve where the left mark is equal to the peak of the second derivative of the V-curve to the left of the inflection point, and identifying a right mark of the V-curve where the right mark is equal to the trough of the second derivative of the V-curve to the right of the inflection point.

[0071] Example 18 The method of Example 17 involves determining the predicted nominal focal value for at least one image, further comprising calculating the distance between the inflection point and the right mark, the distance between the left mark and the right mark, the V value for the right mark, and the V value for the left mark, such that the predicted nominal focal value is a function of the distance between the inflection point and the right mark, the distance between the left mark and the right mark, the V value for the right mark, and the V value for the left mark.

[0072] Example 19 The method of Example 12 further includes disabling at least one image based on the predicted nominal focal value.

[0073] Example 20 The method of Example 12 further comprises obtaining a plurality of predicted nominal focal values, each nominal focal value corresponding to a different image among a plurality of images; determining the median of the plurality of predicted nominal focal values; and translating the image acquisition device based on the median of the plurality of predicted nominal focal values.

[0074] Example 21 The method of Example 12 further comprises obtaining a plurality of predicted nominal focal values, each nominal focal value corresponding to a different image among a plurality of images; determining the median of the plurality of predicted nominal focal values; and evaluating the flow stability of a blood sample containing blood cells based on the median of the plurality of predicted nominal focal values.

[0075] Example 22 A machine comprising a camera and means for determining the focal length of the camera based on images representing one or more blood cells.

[0076] Each of the calculations or operations described herein may be performed using a computer or other processor having hardware, software, and / or firmware. Various method steps may be performed by modules, which may include any of various digital and / or analog data processing hardware and / or software configured to perform the method steps described herein. Modules may optionally include data processing hardware adapted to perform one or more of those steps by having appropriate machine programming code related to those steps, and modules for two or more steps (or parts of two or more steps) may be integrated on a single processor board or separated on different processor boards in any of various integrated and / or distributed processing architectures. These methods and systems often use tangible media that embody machine-readable code having instructions for performing the method steps described above. Preferred tangible media may include memory (including volatile memory and / or non-volatile memory), storage media (such as magnetic recording on floppy disks, hard disks, tapes, optical memory such as CDs, CD-R / Ws, CD-ROMs, DVDs, or any other digital or analog storage media).

[0077] All patents, patent publications, patent applications, journal articles, books, technical references, etc., discussed herein are incorporated herein by reference in their entirety for any purpose.

[0078] Different configurations of the components shown in the drawings or described above, as well as components and steps not shown or described, are possible. Similarly, some features and subcombinations may be useful and may be used without reference to other features and subcombinations. Embodiments of the present invention have been described for illustrative purposes only and not for limiting purposes, but alternative embodiments will become apparent to the reader of this patent. In some cases, method steps or operations may be carried out or performed in a different order, or operations may be added, deleted, or modified. In particular aspects of the present invention, it will be understood that a single component may be replaced by multiple components, or multiple components may be replaced by a single component, in order to provide an element or structure or to perform one or more given functions. Such substitutions are considered to be within the scope of the invention unless they do not function to carry out a particular embodiment of the present invention. Accordingly, the claims should not be treated as being limited to the examples, drawings, embodiments, and descriptions provided above, and the terms should be understood to have the scope provided when their broadest reasonable interpretation is given by a general dictionary, and should be understood to have that meaning when used in the claims, unless a term or phrase is indicated to have a specific meaning under the heading "Explicit Definitions."

[0079] Explicit definition In the above examples and claims, it should be understood that the phrase "based on" something else means that it is at least partially determined by what is indicated as being based on. To indicate that something must be entirely determined on something else, it is written as "based solely on" what must be entirely determined by.

[0080] In the above examples and claims, the phrase “means for determining the focal length of a camera based on images representing one or more blood cells” should be understood as a means-plus-function limitation under Section 112(f) of the U.S. Patent Act, where the function is “determining the focal length of a camera based on images representing one or more blood cells,” and the corresponding structure is a computer configured to use algorithms such as those shown in Figures 5 to 9 and described in the accompanying descriptions.

[0081] In the above examples and claims, the term “set” should be understood as one or more things that are grouped together. The present invention also includes the following embodiments. [Aspect 1] Processor and Image capture device and A non-temporary computer-readable medium, wherein the processor, The process involves acquiring multiple images from the aforementioned image acquisition device, wherein each of the multiple images contains blood cells. Identifying cell boundaries in at least one image, The process involves generating a plurality of rings based on the cell boundary, wherein each of the plurality of rings is offset from the cell boundary. Based on the brightness values ​​of the pixels arranged within the plurality of rings, the nominal focal value predicted for at least one image is determined, A non-temporary computer-readable medium that stores instructions for performing a set of actions including, A system equipped with these features. [Aspect 2] The system according to embodiment 1, further comprising acquiring multiple images and converting each of the multiple images into a color space having brightness values. [Aspect 3] The system according to embodiment 1, wherein identifying the cell boundary in the at least one image further comprises separating the at least one image into foreground and background based on predetermined brightness values. [Aspect 4] The generation of the aforementioned multiple rings Using the morphological expansion of the cell boundary, generate at least one larger ring, wherein the additional larger ring is generated using the morphological expansion of the previously generated larger ring, The method involves generating at least one smaller ring using morphological erosion of the cell boundary, wherein additional smaller rings are generated using morphological erosion of the previously generated smaller rings, and the generation of at least one smaller ring is also achieved by this method. The system according to embodiment 1, further comprising at least one of the following. [Aspect 5] The generation of the aforementioned multiple rings Identifying the best-fitting elliptical shape based on the cell boundary, and generating at least one larger ring based on the offset distance, wherein the additional larger ring is generated from the previously generated larger ring based on the offset distance, Identifying the best-fitting elliptical shape based on the cell boundary, and generating at least one smaller ring based on the offset distance, wherein additional smaller rings are generated from the previously generated smaller rings based on the offset distance, and The system according to embodiment 1, further comprising at least one of the following. [Aspect 6] Determining the predicted nominal focal value for at least one image based on the brightness values ​​of the pixels arranged within the plurality of rings, To generate a V-curve of average brightness values ​​associated with each area between two adjacent rings for each of the plurality of rings at a known distance from the cell boundary, Identifying an inflection point of a V-curve, wherein the inflection point is equal to the peak of the first derivative of the V-curve. Identifying the left mark of the V-curve, wherein the left mark is equal to the peak of the second derivative of the V-curve to the left of the inflection point. Identifying the right-hand mark of the V-curve, wherein the right-hand mark is equal to the trough of the second derivative of the V-curve to the right of the inflection point. The system according to embodiment 1, further comprising identifying a plurality of characteristics based on the brightness value by performing an operation including the following. [Aspect 7] The system according to embodiment 6, wherein determining the predicted nominal focal value for at least one image further includes calculating the distance between the inflection point and the right mark, the distance between the left mark and the right mark, the V value for the right mark, and the V value for the left mark, the predicted nominal focal value is a function of the distance between the inflection point and the right mark, the distance between the left mark and the right mark, the V value for the right mark, and the V value for the left mark. [Aspect 8] The system according to embodiment 1, wherein the set of operations further includes disabling the at least one image based on the predicted nominal focal value. [Aspect 9] The aforementioned set of operations is The method involves obtaining multiple predicted nominal focal values, where each nominal focal value corresponds to a different image among the multiple images. The median of the aforementioned multiple predicted nominal focal values ​​is determined, The image acquisition device is translated based on the median of the plurality of predicted nominal focal points. The system according to embodiment 1, further comprising: [Aspect 10] The aforementioned set of operations is The method involves obtaining multiple predicted nominal focal values, where each nominal focal value corresponds to a different image among the multiple images. The median of the aforementioned multiple predicted nominal focal values ​​is determined, Based on the median of the multiple predicted nominal focal values, the stability of the flow of the blood sample containing the blood cells is evaluated. The system according to embodiment 1, further comprising: [Aspect 11] The process of acquiring multiple images from an image acquisition device, wherein each of the multiple images contains blood cells, Identifying cell boundaries in at least one image, The process involves generating a plurality of rings based on the cell boundary, wherein each of the plurality of rings is offset from the cell boundary. Based on the brightness values ​​of the pixels arranged within the plurality of rings, the nominal focal value predicted for at least one image is determined, A method that includes this. [Aspect 12] The method according to embodiment 11, wherein identifying the cell boundaries in the at least one image further comprises separating the at least one image into foreground and background based on predetermined brightness values. [Aspect 13] The generation of the aforementioned multiple rings Using the morphological expansion of the cell boundary, generate at least one larger ring, wherein the additional larger ring is generated using the morphological expansion of the previously generated larger ring, The method involves generating at least one smaller ring using morphological erosion of the cell boundary, wherein additional smaller rings are generated using morphological erosion of the previously generated smaller rings, and the generation of at least one smaller ring is also achieved by this method. The method according to embodiment 11, further comprising at least one of the above. [Aspect 14] The generation of the aforementioned multiple rings Identifying the best-fitting elliptical shape based on the cell boundary, and generating at least one larger ring based on the offset distance, wherein the additional larger ring is generated from the previously generated larger ring based on the offset distance, Identifying the best-fitting elliptical shape based on the cell boundary, and generating at least one smaller ring based on the offset distance, wherein additional smaller rings are generated from the previously generated smaller rings based on the offset distance, and The method according to embodiment 11, further comprising at least one of the above. [Aspect 15] The predicted nominal focal value for at least one image is determined based on the brightness values ​​of the pixels located within the plurality of rings. To generate a V-curve of average brightness values ​​associated with each area between two adjacent rings for each of the plurality of rings at a known distance from the cell boundary, Identifying an inflection point of a V-curve, wherein the inflection point is equal to the peak of the first derivative of the V-curve. Identifying the left mark of the V-curve, wherein the left mark is equal to the peak of the second derivative of the V-curve to the left of the inflection point. Identifying the right-hand mark of the V-curve, wherein the right-hand mark is equal to the trough of the second derivative of the V-curve to the right of the inflection point. The method according to embodiment 11, further comprising identifying a plurality of characteristics based on the brightness value by performing an operation including the following. [Aspect 16] The method according to aspect 16, wherein determining the predicted nominal focal value for at least one image further includes calculating the distance between the inflection point and the right mark, the distance between the left mark and the right mark, the V value for the right mark, and the V value for the left mark, the predicted nominal focal value is a function of the distance between the inflection point and the right mark, the distance between the left mark and the right mark, the V value for the right mark, and the V value for the left mark. [Aspect 17] The method according to embodiment 11, further comprising disabling the at least one image based on the predicted nominal focal value. [Aspect 18] The method involves obtaining multiple predicted nominal focal values, where each nominal focal value corresponds to a different image among the multiple images. The median of the aforementioned multiple predicted nominal focal values ​​is determined, The image acquisition device is translated based on the median of the plurality of predicted nominal focal points. The method according to embodiment 11, further comprising: [Aspect 19] The method involves obtaining multiple predicted nominal focal values, where each nominal focal value corresponds to a different image among the multiple images. The median of the aforementioned multiple predicted nominal focal values ​​is determined, Based on the median of the multiple predicted nominal focal values, the stability of the flow of the blood sample containing the blood cells is evaluated. The method according to embodiment 11, further comprising: [Aspect 20] Camera and, A machine comprising means for determining the focal length of the camera based on images representing one or more blood cells. [Explanation of Symbols]

[0082] 18 processors 21 Zones 21a Proximal flow channel 21b Distal channel portion 22 Flow Cells 23 Field of View Zones 24 Imaging devices 25. Source of sample fluid 28 Distal end 29 Sample supply tube 32 Sample flow 33 Emission 46 Objective lens 48 CCD 54 Motor Drive 200 Visual Inspection Systems Slide 202 204 Slide Holder 206 Image Capture Devices 208 Optical system 210 Image Sensor 212 Image Processing Devices 214 processors 216 memory 218 Steering Motor System 601 Background 602 Foreground 603 Cell Boundary 610, 611, 612, 613, 614, 615 Rings Cell images 810, 820, 830 811, 821, 831 Graph plots, curves 901 Inflection point 902 Left Mark 903 Right Mark

Claims

1. Processor and Image capture device and A non-temporary computer-readable medium, wherein the processor, The process involves acquiring multiple images from the aforementioned image acquisition device, wherein each of the multiple images contains blood cells. Identifying cell boundaries in at least one image, The process involves generating a plurality of rings based on the cell boundary, wherein each of the plurality of rings is offset from the cell boundary. By identifying multiple image characteristics for each of the brightness values ​​of pixels arranged within the plurality of rings, the nominal focal value predicted for at least one image based on the brightness value is determined. A non-temporary computer-readable medium that stores instructions for performing a set of actions including, Equipped with, The plurality of image characteristics include the first and second derivatives of the brightness value. system.

2. The system according to claim 1, wherein acquiring multiple images further includes converting each of the multiple images into a color space having brightness values, and each of the multiple images contains a white blood cell.

3. The system according to claim 1, wherein identifying the cell boundary in the at least one image further comprises separating the at least one image into foreground and background based on predetermined brightness values.

4. The generation of the aforementioned multiple rings Using the morphological expansion of the cell boundary, generate at least one larger ring, wherein the additional larger ring is generated using the morphological expansion of the previously generated larger ring, The process involves generating at least one smaller ring using morphological erosion of the cell boundary, wherein additional smaller rings are generated using morphological erosion of the previously generated smaller rings, and the process involves generating at least one smaller ring. The system according to claim 1, further comprising at least one of the following.

5. The generation of the aforementioned multiple rings Identifying the best-fitting elliptical shape based on the cell boundary, and generating at least one larger ring based on the offset distance, wherein the additional larger ring is generated from the previously generated larger ring based on the offset distance, Identifying the best-fitting elliptical shape based on the cell boundary, and generating at least one smaller ring based on the offset distance, wherein additional smaller rings are generated from the previously generated smaller rings based on the offset distance, The system according to claim 1, further comprising at least one of the following.

6. Identifying the multiple image characteristics for each of the aforementioned brightness values ​​is To generate a V-curve of average brightness values ​​associated with each area between two adjacent rings for each of the plurality of rings at a known distance from the cell boundary, Identifying the inflection point of the V-curve, wherein the inflection point is equal to the peak of the first derivative of the V-curve. Identifying the left mark of the V-curve, wherein the left mark is equal to the peak of the second derivative of the V-curve to the left of the inflection point. Identifying the right-hand mark of the V-curve, wherein the right-hand mark is equal to the trough of the second derivative of the V-curve to the right of the inflection point. The system according to claim 1, including the following:

7. The system according to claim 6, wherein determining the predicted nominal focal value for at least one image further includes calculating the distance between the inflection point and the right mark, the distance between the left mark and the right mark, the V value for the right mark, and the V value for the left mark, the predicted nominal focal value is a function of the distance between the inflection point and the right mark, the distance between the left mark and the right mark, the V value for the right mark, and the V value for the left mark.

8. The system according to claim 1, wherein the set of operations further includes disabling the at least one image based on the predicted nominal focal value.

9. The aforementioned set of operations is The method involves obtaining multiple predicted nominal focal values, where each nominal focal value corresponds to a different image among the multiple images. The median of the aforementioned multiple predicted nominal focal values ​​is determined, The image acquisition device is translated based on the median of the plurality of predicted nominal focal points. The system according to claim 1, further comprising:

10. The aforementioned set of operations is The method involves obtaining multiple predicted nominal focal values, where each nominal focal value corresponds to a different image among the multiple images. The median of the aforementioned multiple predicted nominal focal values ​​is determined, Based on the median of the multiple predicted nominal focal values, the stability of the flow of the blood sample containing the blood cells is evaluated. The system according to claim 1, further comprising:

11. The process of acquiring multiple images from an image acquisition device, wherein each of the multiple images contains blood cells, Identifying cell boundaries in at least one image, The process involves generating a plurality of rings based on the cell boundary, wherein each of the plurality of rings is offset from the cell boundary. By identifying multiple image characteristics for each of the brightness values ​​of pixels arranged within the plurality of rings, the nominal focal value predicted for at least one image based on the brightness value is determined. Includes, The plurality of image characteristics include the first and second derivatives of the brightness value. method.

12. The method according to claim 11, wherein identifying the cell boundary in the at least one image further comprises separating the at least one image into foreground and background based on predetermined brightness values.

13. The generation of the aforementioned multiple rings Using the morphological expansion of the cell boundary, generate at least one larger ring, wherein the additional larger ring is generated using the morphological expansion of the previously generated larger ring, The process involves generating at least one smaller ring using morphological erosion of the cell boundary, wherein additional smaller rings are generated using morphological erosion of the previously generated smaller rings, and the process involves generating at least one smaller ring. The method according to claim 11, further comprising at least one of the above.

14. The generation of the aforementioned multiple rings Identifying the best-fitting elliptical shape based on the cell boundary, and generating at least one larger ring based on the offset distance, wherein the additional larger ring is generated from the previously generated larger ring based on the offset distance, Identifying the best-fitting elliptical shape based on the cell boundary, and generating at least one smaller ring based on the offset distance, wherein additional smaller rings are generated from the previously generated smaller rings based on the offset distance, The method according to claim 11, further comprising at least one of the above.

15. Identifying the multiple image characteristics for each of the aforementioned brightness values ​​is To generate a V-curve of average brightness values ​​associated with each area between two adjacent rings for each of the plurality of rings at a known distance from the cell boundary, Identifying the inflection point of the V-curve, wherein the inflection point is equal to the peak of the first derivative of the V-curve. Identifying the left mark of the V-curve, wherein the left mark is equal to the peak of the second derivative of the V-curve to the left of the inflection point. Identifying the right-hand mark of the V-curve, wherein the right-hand mark is equal to the trough of the second derivative of the V-curve to the right of the inflection point. The method according to claim 11, including the method described in claim 11.

16. The method of claim 15, wherein determining the predicted nominal focal value for the at least one image further includes calculating the distance between the inflection point and the right mark, the distance between the left mark and the right mark, the V value for the right mark, and the V value for the left mark, the predicted nominal focal value is a function of the distance between the inflection point and the right mark, the distance between the left mark and the right mark, the V value for the right mark, and the V value for the left mark.

17. The method according to claim 11, further comprising disabling the at least one image based on the predicted nominal focal value.

18. The method involves obtaining multiple predicted nominal focal values, where each nominal focal value corresponds to a different image among the multiple images. The median of the aforementioned multiple predicted nominal focal values ​​is determined, The image acquisition device is translated based on the median of the plurality of predicted nominal focal points. The method according to claim 11, further comprising:

19. The method involves obtaining multiple predicted nominal focal values, where each nominal focal value corresponds to a different image among the multiple images. The median of the aforementioned multiple predicted nominal focal values ​​is determined, Based on the median of the multiple predicted nominal focal values, the stability of the flow of the blood sample containing the blood cells is evaluated. The method according to claim 11, further comprising:

20. The method according to claim 11, wherein the blood cells are white blood cells.

Images