Peripheral blood cell morphology auxiliary extraction method under microscopic imaging

Abstract

The invention relates to the technical field of image recognition, in particular to a peripheral blood cell morphology auxiliary extraction method under microscopic imaging, and the method comprises the steps: obtaining a lymphocyte image, in which to-be-recognized lymphocytes comprise common lymphocytes, large-particle lymphocytes and heterotypic lymphocytes; on the basis of the distance between the cell nucleus and the cytoplasm of the to-be-recognized lymphocyte, screening to obtain common lymphocytes; according to the dense condition of the peripheral lymphocytes of the remaining to-be-recognized lymphocytes and the extrusion condition of the outer edges of the cytoplasm of the remaining to-be-recognized lymphocytes, large-particle lymphocytes and heterotypic lymphocytes are obtained through screening, and the accuracy of lymphocyte classification can be improved.

Description

Technical Field

[0001] The present invention relates to the technical field of image recognition, and particularly to a method for assisting in extracting the morphology of peripheral blood cells under microscopic imaging. Background Art

[0002] Detecting human peripheral blood cells through microscopic imaging is a common blood routine detection method. As a common and important type of cell, lymphocytes are crucial for their morphological detection. Lymphocytes are divided into ordinary lymphocytes, large granular lymphocytes, and atypical lymphocytes. Whether it is for teaching work in the medical field or other uses, the differentiation of lymphocytes has very wide applications. However, when differentiating lymphocytes, due to the small differences between different types of lymphocytes, especially between large granular lymphocytes and atypical lymphocytes, the accuracy of lymphocyte classification is affected. Summary of the Invention

[0003] In order to solve the technical problem of the low accuracy of lymphocyte classification in the existing lymphocyte differentiation method, the purpose of the present invention is to provide a method for assisting in extracting the morphology of peripheral blood cells under microscopic imaging, and the specific technical solution adopted is as follows:

[0004] In the first aspect of the present invention, a method for assisting in extracting the morphology of peripheral blood cells under microscopic imaging is provided, including:

[0005] Obtain lymphocyte images, where the lymphocytes to be recognized in the lymphocyte images include ordinary lymphocytes, large granular lymphocytes, and atypical lymphocytes;

[0006] Based on the distance between the nucleus and cytoplasm of the lymphocytes to be recognized, screen out ordinary lymphocytes;

[0007] According to the density of lymphocytes around the remaining lymphocytes to be recognized and the extrusion condition of the outer edge of the cytoplasm of the remaining lymphocytes to be recognized, screen out large granular lymphocytes;

[0008] Based on the screened large granular lymphocytes, obtain atypical lymphocytes.

[0009] In an exemplary embodiment, the screening out ordinary lymphocytes based on the distance between the nucleus and cytoplasm of the lymphocytes to be recognized includes:

[0010] Obtain the distances from the outer edge of the nucleus to the outer edge of the cytoplasm in multiple directions starting from the center of the nucleus of the lymphocytes to be recognized, and obtain the average distance and the variance of the distances;

[0011] Based on the area of the lymphocyte to be identified, as well as the mean distance and the distance variance, obtain the first possibility that the lymphocyte to be identified belongs to a normal lymphocyte; the first possibility is inversely proportional to the area, inversely proportional to the distance variance, and inversely proportional to the mean distance.

[0012] Screen out normal lymphocytes according to the first possibility.

[0013] In an exemplary embodiment, the screening out of large granular lymphocytes according to the density of lymphocytes around the remaining lymphocytes to be identified and the squeezing condition of the outer edge of the cytoplasm of the remaining lymphocytes to be identified includes:

[0014] According to the density of lymphocytes around the lymphocytes to be identified remaining for the first time and the squeezing condition of the outer edge of the cytoplasm of the lymphocytes to be identified remaining for the first time, screen out large granular lymphocytes for the first time;

[0015] Based on the similarity of the nuclear shapes of the lymphocytes to be identified remaining for the second time and the large granular lymphocytes screened out for the first time, as well as the cell distance, screen out large granular lymphocytes for the second time.

[0016] In an exemplary embodiment, the screening out of large granular lymphocytes for the first time according to the density of lymphocytes around the lymphocytes to be identified remaining for the first time and the squeezing condition of the outer edge of the cytoplasm of the lymphocytes to be identified remaining for the first time includes:

[0017] Obtain the first quantity of the second lymphocytes to be identified within a preset range around the first lymphocyte to be identified; the first lymphocyte to be identified is any one of the lymphocytes to be identified remaining for the first time, and the second lymphocytes to be identified are other lymphocytes to be identified remaining for the first time except the first lymphocyte to be identified;

[0018] Obtain the second quantity of the third lymphocytes to be identified in contact with the first lymphocyte to be identified; the third lymphocytes to be identified are other lymphocytes to be identified remaining for the first time except the first lymphocyte to be identified;

[0019] According to the first quantity, the second quantity, and the squeezing condition of the outer edge of the cytoplasm, obtain the second possibility that the first lymphocyte to be identified belongs to a large granular lymphocyte;

[0020] Judge whether the first lymphocyte to be identified belongs to a large granular lymphocyte according to the second possibility.

[0021] In an exemplary embodiment, the process of obtaining the second quantity includes:

[0022] Segment the outer edge of the cytoplasm of the first lymphocyte to be recognized according to the gray value to obtain a plurality of line segments;

[0023] Obtain the second quantity, where the second quantity is the number of the first line segments, and the first line segments are the line segments with gray values less than a preset threshold.

[0024] In an exemplary embodiment, the process of obtaining the extrusion condition of the outer edge of the cytoplasm of the remaining lymphocytes to be recognized for the first time includes:

[0025] Obtain the lengths of the line segments in each of the remaining lymphocytes to be recognized for the first time with the second quantity greater than 1, and obtain the average length;

[0026] Obtain the standard line segment length according to the average length;

[0027] Obtain the length difference between the lengths of the line segments in the first lymphocyte to be recognized and the standard line segment length, and obtain the average length difference corresponding to the first lymphocyte to be recognized.

[0028] In an exemplary embodiment, the calculation formula of the second possibility is as follows:

[0029]

[0030] where D i is the second possibility that the i-th remaining lymphocyte to be recognized for the first time belongs to a large granular lymphocyte, N i is the second quantity corresponding to the i-th remaining lymphocyte to be recognized for the first time, M i is the first quantity corresponding to the i-th remaining lymphocyte to be recognized for the first time, L i,n is the length of the n-th first line segment in the i-th remaining lymphocyte to be recognized for the first time, L0 is the standard line segment length, is the negative correlation normalization of , and norm is the normalization function.

[0031] In an exemplary embodiment, the second screening for large granular lymphocytes based on the similarity of the nuclear shapes of the remaining lymphocytes to be recognized for the second time and the large granular lymphocytes obtained by the first screening, and the cell distance includes;

[0032] Obtain the difference in the roundness of the nuclei between the fourth lymphocyte to be recognized and the reference large granular lymphocyte; the fourth lymphocyte to be recognized is any one of the remaining lymphocytes to be recognized for the second time, and the reference large granular lymphocyte is any one of the large granular lymphocytes obtained by the first screening;

[0033] Obtain the cell distance between the fourth lymphocyte to be identified and the reference large granular lymphocyte;

[0034] According to the difference in shape roundness and the cell distance, obtain the reference weight of the reference large granular lymphocyte for the fourth lymphocyte to be identified, where the reference weight is inversely proportional to the difference in shape roundness and inversely proportional to the cell distance;

[0035] Weight the second possibility that the reference large granular lymphocyte belongs to the large granular lymphocyte according to the reference weight and calculate the average value to obtain the third possibility that the fourth lymphocyte to be identified belongs to the large granular lymphocyte;

[0036] Judge whether the fourth lymphocyte to be identified belongs to the large granular lymphocyte according to the third possibility.

[0037] In an exemplary embodiment, the process of obtaining the difference in shape roundness includes:

[0038] Obtain the circumscribed circle of the nucleus of the fourth lymphocyte to be identified, and obtain the ratio of the area of the nucleus of the fourth lymphocyte to be identified to the area of its corresponding circumscribed circle as the shape roundness of the fourth lymphocyte to be identified;

[0039] Obtain the circumscribed circle of the nucleus of the reference large granular lymphocyte, and obtain the ratio of the area of the nucleus of the reference large granular lymphocyte to the area of its corresponding circumscribed circle as the shape roundness of the reference large granular lymphocyte;

[0040] Obtain the difference between the shape roundness of the fourth lymphocyte to be identified and the shape roundness of the reference large granular lymphocyte.

[0041] In an exemplary embodiment, the calculation formula of the third possibility is as follows:

[0042]

[0043] Where Y j is the third possibility that the j-th second remaining lymphocyte to be identified belongs to the large granular lymphocyte, D k is the second possibility that the k-th reference large granular lymphocyte belongs to the large granular lymphocyte, K is the number of reference large granular lymphocytes, d j,k is the cell distance between the j-th second remaining lymphocyte to be identified and the k-th reference large granular lymphocyte, M j is the shape roundness of the j-th second remaining lymphocyte to be identified, M k is the shape roundness of the k-th reference large granular lymphocyte, f(d j,k *|Mj -M k |) represents the negative correlation normalization of d j,k *|M j -M k |negative correlation normalization of

[0044] The present invention has the following beneficial effects: In the method for assisting in extracting the morphology of peripheral blood cells under microscopic imaging provided by the present invention, since the shape differences between normal lymphocytes, large granular lymphocytes, and atypical lymphocytes are relatively large, first, according to the distance between the nucleus and cytoplasm of lymphocytes, normal lymphocytes are screened out first, and then the screening focuses on large granular lymphocytes and atypical lymphocytes, reducing the interference caused by normal lymphocytes during the screening process of large granular lymphocytes and atypical lymphocytes. Since the density of lymphocytes around large granular lymphocytes and atypical lymphocytes and the degree of extrusion of lymphocytes by other cells are different, therefore, according to the density of lymphocytes around the lymphocytes to be identified and the degree of extrusion of the outer edge of the cytoplasm of the remaining lymphocytes to be identified, large granular lymphocytes are screened out, and then atypical lymphocytes are obtained. In this way, the accuracy of lymphocyte classification can be improved. BRIEF DESCRIPTION OF THE DRAWINGS

[0045] Figure 1 is a flowchart of a method for assisting in extracting the morphology of peripheral blood cells under microscopic imaging provided by an embodiment of the present invention;

[0046] Figure 2 is a schematic diagram of a lymphocyte image provided by an embodiment of the present invention;

[0047] Figure 3 is a schematic diagram of various lymphocytes provided by an embodiment of the present invention;

[0048] Figure 4 is a flowchart for identifying normal lymphocytes provided by an embodiment of the present invention;

[0049] Figure 5 is a schematic diagram of the aggregation of large granular lymphocytes provided by an embodiment of the present invention;

[0050] Figure 6 is a flowchart for overall screening of large granular lymphocytes provided by an embodiment of the present invention;

[0051] Figure 7 is a flowchart for the first screening of large granular lymphocytes provided by an embodiment of the present invention;

[0052] Figure 8 is a schematic diagram of the gray color of the outer edge of the cytoplasm provided by an embodiment of the present invention;

[0053] Figure 9 It is a flowchart for obtaining the second quantity provided by an embodiment of the present invention;

[0054] Figure 10 It is a schematic structural diagram of abnormal lymphocytes provided by an embodiment of the present invention;

[0055] Figure 11 It is a flowchart for obtaining the extrusion condition of the outer edge of the cytoplasm provided by an embodiment of the present invention;

[0056] Figure 12 It is a schematic structural diagram of free large granular lymphocytes provided by an embodiment of the present invention;

[0057] Figure 13 It is a flowchart for the second screening of large granular lymphocytes provided by an embodiment of the present invention;

[0058] Figure 14 It is a flowchart for obtaining the difference in roundness of shape provided by an embodiment of the present invention. Detailed implementation manners

[0059] In order to further elaborate on the technical means and effects adopted by the present invention to achieve the predetermined invention purpose, the following combines the drawings and preferred embodiments to detail the specific implementation manners, structures, features and their effects of the present invention. In the following description, different "one embodiment" or "another embodiment" do not necessarily refer to the same embodiment. In addition, the specific features, structures or characteristics in one or more embodiments can be combined in any suitable form.

[0060] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by those skilled in the technical field to which the present invention belongs. All data information collected in this application has been obtained through full consent authorization, and the collection, use and processing of relevant information need to comply with the relevant laws, regulations and standards of relevant countries and regions.

[0061] This embodiment provides a method for assisting in the extraction of peripheral blood cell morphology under microscopic imaging, as Figure 1 shown, including:

[0062] Step 1: Obtain lymphocyte images, where the lymphocytes to be recognized in the lymphocyte images include normal lymphocytes, large granular lymphocytes and abnormal lymphocytes.

[0063] Step 2: Screen out normal lymphocytes based on the distance between the nucleus and cytoplasm of the lymphocytes to be recognized.

[0064] Step 3: Filter out large granular lymphocytes based on the density of lymphocytes around the remaining lymphocytes to be recognized and the degree of extrusion of the outer edge of the cytoplasm of the remaining lymphocytes to be recognized.

[0065] Step 4: Obtain atypical lymphocytes based on the large granular lymphocytes filtered out.

[0066] The following combines the accompanying drawings to specifically illustrate the specific implementation process of each step.

[0067] Step 1: Obtain a lymphocyte image, where the lymphocytes to be recognized in the lymphocyte image include normal lymphocytes, large granular lymphocytes, and atypical lymphocytes.

[0068] In an exemplary embodiment, the blood of a patient is analyzed by an automated microscopic blood analyzer, or the blood on a slide of the patient is detected by an optical microscope to obtain an initial blood image, and a preliminary analysis is performed on it to obtain the positions and quantities of various peripheral blood cells, and the positions and sizes of each lymphocyte among them are obtained, that is, a lymphocyte microscopic distribution map, that is, a lymphocyte image. As Figure 2 shown, the darker large cell area in the figure is the lymphocyte area.

[0069] It should be understood that in the initial blood image, there will be information on various peripheral blood cells. However, due to the large differences among various major types of cells in peripheral blood cells, the automated microscopic blood analyzer can classify peripheral blood cells more accurately, so as to obtain a lymphocyte image, which only contains lymphocytes.

[0070] Therefore, in this embodiment, the obtained lymphocyte image is an image that has been initially recognized for lymphocytes, that is, other types of cells have been filtered out, and the lymphocyte image only contains lymphocytes, which is convenient for subsequent classification of lymphocytes. The lymphocytes to be recognized in the lymphocyte image include normal lymphocytes, large granular lymphocytes, and atypical lymphocytes.

[0071] Step 2: Filter out normal lymphocytes based on the distance between the nucleus and cytoplasm of the lymphocytes to be recognized.

[0072] For the obtained lymphocyte images, since there are three types of lymphocytes in lymphocytes: normal lymphocytes, large granular lymphocytes, and atypical lymphocytes, which mainly exist in the form of unactivated normal lymphocytes under normal circumstances, providing long-term immune surveillance and specific immune response functions for the human body. After a virus invasion, normal lymphocytes will be activated and differentiated into large granular lymphocytes or atypical lymphocytes, thus enhancing the immune response. Especially in the early stage of the immune response and the response to specific pathogens in the later stage, they play an important role. However, in lymphocyte images, there are significant morphological differences between normal lymphocytes and activated lymphocytes (such as large granular lymphocytes and atypical lymphocytes). Moreover, due to the different cell functions, the morphological differences between these activated immune cells and normal lymphocytes are particularly obvious. As Figure 3 shown, it is a schematic diagram of various lymphocytes. Normal lymphocytes are usually smaller, with a thinner and evenly distributed cytoplasm, while large granular lymphocytes and atypical lymphocytes are usually larger, with a thicker and unevenly distributed cytoplasm. Therefore, when screening lymphocytes in lymphocyte images, based on these morphological differences, normal lymphocytes can be first identified and distinguished, so as to effectively extract large granular lymphocytes and atypical lymphocytes.

[0073] Since both cytoplasm and nucleus exist in lymphocytes, the difference is that the cytoplasm distribution in all directions of the nucleus of normal lymphocytes is thinner, while the cytoplasm in all directions of the nucleus of large granular lymphocytes and atypical lymphocytes is thicker. Therefore, to distinguish normal lymphocytes, large granular lymphocytes, and atypical lymphocytes, for each lymphocyte, the lymphocyte needs to be divided into two regions, namely the nucleus and the cytoplasm.

[0074] In an exemplary embodiment, the gray values of the nucleus and the cytoplasm are different. The gray value of the nucleus region is smaller, and the gray value of the cytoplasm region is larger. Therefore, first, through edge segmentation, the lymphocyte is segmented into two regions, and then a gray value threshold is preset. By comparing the average gray values of the two regions with this gray value threshold, the region with a gray value smaller than the gray value threshold, that is, the region with a smaller gray value, is identified as the nucleus region, and the region with a gray value greater than or equal to the gray value threshold, that is, the region with a larger gray value, is identified as the cytoplasm region.

[0075] Therefore, based on the distance between the nucleus and the cytoplasm of the lymphocyte to be identified, normal lymphocytes are screened. In an exemplary embodiment, as Figure 4 shown, the following gives a specific recognition process of normal lymphocytes:

[0076] Step 2-1: Obtain the distances from the outer edge of the nucleus to the outer edge of the cytoplasm in multiple directions starting from the center of the nucleus of the lymphocyte to be identified, and obtain the distance mean and the distance variance.

[0077] Construct a two-dimensional coordinate system according to the length direction and width direction of the lymphocyte image, and map the lymphocyte image into this two-dimensional coordinate system, so that each pixel point in the lymphocyte image corresponds to a two-dimensional coordinate point.

[0078] For any lymphocyte to be recognized, obtain the central position of the nucleus of the lymphocyte to be recognized. For example, average the abscissas of each position of the nucleus of the lymphocyte to be recognized, and average the ordinates of each position of the nucleus to obtain the average abscissa value and the average ordinate value as the coordinates of its central position.

[0079] Starting from the center of the nucleus of the lymphocyte to be recognized, emit rays in multiple directions around. The multiple directions around are multiple directions within 360° around. The number of directions and the specific angles are set according to the actual situation. For example: 30° is one direction, and there are a total of 12 directions.

[0080] In each direction, the ray in this direction intersects with the outer edge of the nucleus of the lymphocyte to be recognized and the outer edge of the cytoplasm. Then, obtain the distance between these two intersection points, that is, obtain the distance from the outer edge of the nucleus to the outer edge of the cytoplasm in this direction. One distance is obtained for each direction, so that multiple distances can be obtained. Then calculate the mean and variance of these distances to obtain the distance mean and distance variance.

[0081] Step 2-2: Obtain the first possibility that the lymphocyte to be recognized belongs to a normal lymphocyte according to the area of the lymphocyte to be recognized, and the distance mean and distance variance.

[0082] Since the size of normal lymphocytes is smaller than that of large granular lymphocytes and abnormal lymphocytes, the distance from the outer edge of the nucleus to the outer edge of the cytoplasm is smaller, and the distribution of its cytoplasm is relatively uniform. Therefore, according to the area of the lymphocyte to be recognized, and the distance mean and distance variance, obtain the first possibility that the lymphocyte to be recognized belongs to a normal lymphocyte. The smaller the area of the lymphocyte to be recognized, the greater the first possibility; the smaller the distance mean, the greater the first possibility; the smaller the distance variance, the greater the first possibility. Among them, the area of the lymphocyte to be recognized is the number of pixel points contained in the area of the lymphocyte to be recognized.

[0083] In an exemplary embodiment, a specific calculation method of the first possibility is given as follows:

[0084] P x = 1 - norm(σ x *S x *L x );

[0085] Where Px is the first possibility that the x-th lymphocyte to be recognized belongs to a normal lymphocyte, S x is the area of the x-th lymphocyte to be recognized, σ x is the distance variance of the x-th lymphocyte to be recognized, L x is the average distance of the x-th lymphocyte to be recognized.

[0086] norm represents the normalization function. The normalization method here can be: obtain the maximum and minimum values in the product of the area, distance variance, and average distance of the lymphocytes to be recognized corresponding to each lymphocyte to be recognized, and then normalize the product of the area, distance variance, and average distance of the lymphocytes to be recognized corresponding to each lymphocyte to be recognized to 0 - 1 using the maximum and minimum value normalization method.

[0087] Step 2 - 3: Screen out normal lymphocytes based on the first possibility.

[0088] The greater the first possibility, the more likely the lymphocyte to be recognized belongs to a normal lymphocyte. In an exemplary embodiment, a normal lymphocyte possibility threshold is preset. The value range of this normal lymphocyte possibility threshold is 0 - 1, which is set according to the actual situation. The larger the set normal lymphocyte possibility threshold, the higher the strictness of screening for normal lymphocytes. In this embodiment, 0.7 is taken as an example.

[0089] Determine the lymphocytes to be recognized corresponding to the first possibility that is greater than or equal to the normal lymphocyte possibility threshold as normal lymphocytes. Set the lymphocytes to be recognized in the lymphocyte image other than normal lymphocytes as the remaining lymphocytes to be recognized for the first time. Then, the remaining lymphocytes to be recognized in the lymphocyte image are large granular lymphocytes and atypical lymphocytes. The similarity between large granular lymphocytes and atypical lymphocytes is relatively high. Therefore, after excluding normal lymphocytes in the lymphocytes, further detailed differentiation of large granular lymphocytes and atypical lymphocytes can be carried out.

[0090] Step 3: Screen out large granular lymphocytes according to the density of lymphocytes around the remaining lymphocytes to be recognized and the degree of extrusion of the outer edge of the cytoplasm of the remaining lymphocytes to be recognized.

[0091] This step is used to screen out large granular lymphocytes among large granular lymphocytes and atypical lymphocytes. After obtaining large granular lymphocytes, the remaining lymphocytes are atypical lymphocytes.

[0092] Due to the different functions and action times of large granular lymphocytes and atypical lymphocytes in the immune system, there are also different differences between large granular lymphocytes and atypical lymphocytes in lymphocyte images. When a virus invades the human body, the body's immune system will be quickly activated. Large granular lymphocytes will quickly concentrate in the infected area to quickly identify and kill infected cells or pathogens. Therefore, in lymphocyte images, large granular lymphocytes are mostly in an aggregated state, and the possibility of contact with other cells is greater. Compared with atypical lymphocytes, the cytoplasm in the cytoplasm of large granular lymphocytes is more transparent.

[0093] As Figure 5 shown, the circled area is the aggregation area, and there will be an aggregated change among the large granular lymphocytes in the circled area. Due to the influence of aggregation, the cell membrane of large granular lymphocytes will be squeezed by other surrounding cells and show a certain deformation. Therefore, based on this feature, large granular lymphocytes can be screened for the first time, and the large granular lymphocytes in the aggregated cells can be screened out, and then the free large granular lymphocytes can be screened out.

[0094] In an exemplary embodiment, as Figure 6 shown, the screening process of large granular lymphocytes generally includes the following two steps:

[0095] Step 3-1: According to the density of lymphocytes around the remaining lymphocytes to be recognized for the first time and the extrusion situation of the outer edge of the cytoplasm of the remaining lymphocytes to be recognized for the first time, large granular lymphocytes are screened for the first time.

[0096] In the area where large granular lymphocytes aggregate, their cell density is higher and they are more easily squeezed by other cells. Since the color of the cytoplasm inside large granular lymphocytes is lighter, when they are squeezed by other cells, the color of the edge of their mutual contact area will become darker. Therefore, large granular lymphocytes can be screened according to the aggregation area of large granular lymphocytes.

[0097] In an exemplary embodiment, as Figure 7 shown, the following gives the specific implementation process of the first screening:

[0098] Step 3-1-1: Obtain the first quantity of the second lymphocytes to be recognized within a preset range around the first lymphocyte to be recognized.

[0099] For the sake of convenience of description, it is assumed that the first lymphocyte to be recognized is any one of the remaining lymphocytes to be recognized for the first time. The number of the remaining lymphocytes to be recognized for the first time within a preset range around the first lymphocyte to be recognized is obtained. Specifically, the center of the nucleus of the first lymphocyte to be recognized is used as the center. The size of the preset range around is set according to the actual situation. In an exemplary embodiment, 4 times the radius of the circumscribed circle of the first lymphocyte to be recognized is used as the radius of the preset range around. A circle is drawn with the center position of the nucleus of the first lymphocyte to be recognized as the center of the circle, and the obtained circular area is the preset range around the first lymphocyte to be recognized.

[0100] It is assumed that the remaining lymphocytes to be recognized for the first time within the preset range around the first lymphocyte to be recognized are the second lymphocytes to be recognized. Then, both the second lymphocyte to be recognized and the first lymphocyte to be recognized are the remaining lymphocytes to be recognized for the first time, and the second lymphocyte to be recognized and the first lymphocyte to be recognized are not the same lymphocyte to be recognized.

[0101] The number of the remaining lymphocytes to be recognized for the first time within the preset range around the first lymphocyte to be recognized is obtained, that is, the number of the second lymphocytes to be recognized within the preset range around the first lymphocyte to be recognized is obtained, and is defined as the first quantity.

[0102] Step 3-1-2: Obtain the second quantity of the third lymphocytes to be recognized that are in contact with the first lymphocyte to be recognized.

[0103] As can be seen from the above, there are other cells in contact with and squeezing the first lymphocyte to be recognized. It is assumed that the remaining lymphocytes to be recognized for the first time that are in contact with the first lymphocyte to be recognized are the third lymphocytes to be recognized. Therefore, both the third lymphocyte to be recognized and the first lymphocyte to be recognized are the remaining lymphocytes to be recognized for the first time, and the third lymphocyte to be recognized and the first lymphocyte to be recognized are not the same lymphocyte to be recognized. Then, since the third lymphocyte to be recognized must be within the preset range around the first lymphocyte to be recognized. Therefore, a part of the second lymphocytes to be recognized may be the third lymphocytes to be recognized.

[0104] The number of the remaining lymphocytes to be recognized for the first time that are in contact with the first lymphocyte to be recognized is obtained, that is, the number of the third lymphocytes to be recognized that are in contact with the first lymphocyte to be recognized, and is defined as the second quantity.

[0105] In an exemplary embodiment, since the outer edge of the cytoplasm, that is, the edge of the cell membrane, is an approximately circular line, at the part where it is in contact with and squeezed by other cells, the color of the edge will deepen, while at the part where it is not in contact with and squeezed by other cells, the color of the edge will be lighter, such asFigure 8 As shown. Therefore, according to the color difference at the outer edge of the cytoplasm, the second quantity of the remaining lymphocytes to be recognized that are in contact with the first lymphocyte to be recognized is obtained. Thus, as Figure 9 shown, the process of obtaining the second quantity includes:

[0106] Step 3-1-2-1: Segment the outer edge of the cytoplasm of the first lymphocyte to be recognized according to the gray value to obtain a plurality of line segments.

[0107] For the outer edge of the cytoplasm of the first lymphocyte to be recognized, taking any one of the edge pixel points as the starting point, the gray values of each edge pixel point are obtained in a clockwise direction, so as to obtain a gray value sequence. Then, the difference sequence of the gray value sequence is obtained. Each element in the difference sequence is the absolute value of the difference between the subsequent gray value and the previous gray value in the gray value sequence. Then, the difference sequence is curve-fitted to obtain each maximum value (i.e., each peak) in the difference sequence curve. Taking each maximum value as the segmentation point of the outer edge of the cytoplasm, the outer edge of the cytoplasm of the first lymphocyte to be recognized is segmented to obtain a plurality of line segments.

[0108] Due to whether it is in contact with and squeezed by other cells, the colors of the corresponding edges are different. Therefore, through segmentation, each line segment is divided into two categories, one with a high gray value and the other with a low gray value. The category with a low gray value represents the part in contact with other cells, and the number of line segments in the category with a low gray value can represent the aggregation degree of surrounding cells.

[0109] Therefore, the number of contacts between the outer edge of the cytoplasm of the first lymphocyte to be recognized and the remaining lymphocytes to be recognized for the first time is different, and the outer edge of the cytoplasm of the first lymphocyte to be recognized will be divided into line segments of different depths and shallows with unequal quantities.

[0110] Step 3-1-2-2: Obtain the second quantity, where the second quantity is the number of the first line segments, and the first line segments are the line segments with a gray value less than a preset threshold.

[0111] Obtain the number of line segments with a smaller gray value among the line segments corresponding to the outer edge of the cytoplasm of the first lymphocyte to be recognized. In an exemplary embodiment, a threshold can be set, the average gray value of each line segment is obtained, the average gray value of each line segment is compared with the preset threshold, and the line segments corresponding to the average gray value less than the preset threshold are defined as the first line segments. Then, obtain the number of the first line segments as the second quantity of the third lymphocytes to be recognized that are in contact with the first lymphocyte to be recognized.

[0112] Using the above process, the second quantity of each of the remaining lymphocytes to be recognized for the first time is obtained.

[0113] Since the shape of the nucleus of atypical lymphocytes is relatively irregular and it is easy to be adjacent to the cell membrane, the color on the outer edge of its cytoplasm is darker than that of other edges. Therefore, in the lymphocyte image, there is also a single edge with a lower gray value, such as Figure 10 shown. Affected by its nucleus, there may also be differences in the depth of the outer edge of the cytoplasm of atypical lymphocytes, which is likely to interfere with the screening of large granular lymphocytes. However, due to the uncertainty of the nucleus of atypical lymphocytes, the length of the edge with a lower gray value on the outer edge of the cytoplasm is uncertain. Moreover, generally speaking, the gray value of the outer edge of the cytoplasm of atypical lymphocytes is not low, or there is only a single line segment with a lower gray value. Therefore, it is also necessary to obtain the extrusion condition of the outer edge of the cytoplasm of the lymphocytes to be recognized remaining after the first time. In an exemplary embodiment, as Figure 11 shown, the following gives a specific obtaining process:

[0114] Step 3-1-2-3: Obtain the lengths of the line segments in each of the lymphocytes to be recognized remaining after the first time with the second quantity greater than 1, and obtain the average length.

[0115] To reduce the interference caused by the differences in the depth of the outer edge of the cytoplasm of atypical lymphocytes to the screening of large granular lymphocytes, lymphocytes to be recognized remaining after the first time with the second quantity greater than 1 are obtained from each of the lymphocytes to be recognized remaining after the first time. Then, obtain the lengths of the line segments in each of the lymphocytes to be recognized remaining after the first time with the second quantity greater than 1. Finally, average the lengths of the line segments in all the lymphocytes to be recognized remaining after the first time with the second quantity greater than 1 to obtain the average length.

[0116] Step 3-1-2-4: Obtain the standard line segment length according to the average length.

[0117] The average length reflects the overall level of the lengths of the line segments in all the lymphocytes to be recognized remaining after the first time with the second quantity greater than 1. Therefore, the standard line segment length is obtained according to the average length. In an exemplary embodiment, the obtained average length is used as the standard line segment length.

[0118] Step 3-1-2-5: Obtain the length differences between the lengths of the line segments in the first lymphocyte to be recognized and the standard line segment length, and obtain the average length difference corresponding to the first lymphocyte to be recognized.

[0119] Obtain the length differences between the lengths of the line segments in the first lymphocyte to be recognized and the standard line segment length. In this embodiment, the length difference is the absolute value of the difference in length. Then calculate the average length difference corresponding to the first lymphocyte to be recognized.

[0120] Step 3-1-3: Obtain the second possibility that the first lymphocyte to be recognized belongs to large granular lymphocytes according to the first quantity, the second quantity, and the extrusion condition of the outer edge of the cytoplasm.

[0121] The larger the first quantity and the second quantity are, the higher the possibility that the first lymphocyte to be recognized belongs to large granular lymphocytes; the smaller the average value of the length difference corresponding to the first lymphocyte to be recognized is, the higher the possibility that the first lymphocyte to be recognized belongs to large granular lymphocytes. In an exemplary embodiment, the following gives the calculation formula for the second possibility:

[0122]

[0123] where D i is the second possibility that the i-th remaining lymphocyte to be recognized for the first time belongs to large granular lymphocytes, N i is the second quantity corresponding to the i-th remaining lymphocyte to be recognized for the first time, M i is the first quantity corresponding to the i-th remaining lymphocyte to be recognized for the first time, L i,n is the length of the n-th first line segment in the i-th remaining lymphocyte to be recognized for the first time, and L0 is the standard line segment length.

[0124] Through N i and M i The quantities reflect the number of first line segments and the number of surrounding cells in the i-th remaining lymphocyte to be recognized for the first time. The larger the quantity is, the greater the influence of the i-th remaining lymphocyte to be recognized for the first time by the remaining cells, and the higher the possibility that the i-th remaining lymphocyte to be recognized for the first time belongs to large granular lymphocytes.

[0125] By calculating the difference from the standard line segment length, the morphological change of the edge can be further quantified. is the negative correlation normalization of . Since the edge of atypical lymphocytes may be more irregular, resulting in a larger difference in line segment length, therefore, the greater the difference from the standard line segment length, the less likely it belongs to large granular lymphocytes. The negative correlation normalization method in this embodiment can be: obtain the maximum and minimum values of the average values of the length differences corresponding to each remaining lymphocyte to be recognized for the first time, and then use the maximum and minimum value normalization method to normalize the corresponding to the i-th remaining lymphocyte to be recognized for the first time to 0-1, and finally calculate the value 1 minus the normalized result to achieve negative correlation normalization.

[0126] The norm is a normalization function. The normalization method here can also be: obtaining the maximum and minimum values of the sum of the first quantity and the second quantity of the lymphocytes to be recognized remaining for the first time, and then normalizing the first quantity and the second quantity to 0-1 by using the maximum and minimum value normalization method.

[0127] Step 3-1-4: Determine whether the first lymphocyte to be recognized belongs to large granular lymphocytes according to the second possibility.

[0128] The greater the second possibility, the greater the possibility of belonging to large granular lymphocytes. Therefore, in an exemplary embodiment, a first large granular lymphocyte possibility threshold is preset. The value range of the first large granular lymphocyte possibility threshold is 0-1, which is set according to the actual situation. The larger the first large granular lymphocyte possibility threshold is set, the higher the strictness of screening large granular lymphocytes with aggregation is. In this embodiment, 0.7 is taken as an example.

[0129] Determine the first lymphocyte to be recognized corresponding to the second possibility greater than or equal to the first large granular lymphocyte possibility threshold as large granular lymphocytes, and realize the first screening of large granular lymphocytes. Set the lymphocytes to be recognized remaining after the first screening as the lymphocytes to be recognized remaining for the second time. Then, the lymphocytes to be recognized remaining for the second time in the lymphocyte image are large granular lymphocytes and abnormal lymphocytes.

[0130] Step 3-2: Based on the similarity of the nucleus shapes of the lymphocytes to be recognized remaining for the second time and the large granular lymphocytes obtained by the first screening, and the cell distance, screen out large granular lymphocytes for the second time.

[0131] The reason for screening large granular lymphocytes for the second time is that in the lymphocyte image, in addition to the aggregation phenomenon of large granular lymphocytes, there are also some free large granular lymphocytes, that is, large granular lymphocytes without aggregation phenomenon. As Figure 12 shown, the number of the remaining lymphocytes around these free large granular lymphocytes is small. Therefore, there is no aggregation situation in the lymphocyte image, and there is no situation of the first line segment mentioned above, and the large granular lymphocytes cannot be screened by the screening method given in step 3-1. However, since the second possibility that the first lymphocyte to be recognized belongs to large granular lymphocytes is obtained in step 3-1, because the nucleus shape of abnormal lymphocytes is relatively irregular, while the nucleus of large granular lymphocytes is relatively round and regular. For details, please refer to Figure 3, it is possible to obtain the probability that each free lymphocyte to be recognized belongs to large granular lymphocytes according to the similarity of the cell nucleus in each free lymphocyte to the cell nucleus in the large granular lymphocytes that have been recognized, and in combination with its distance from the surrounding lymphocytes. Therefore, based on the similarity of the cell nucleus shapes of the lymphocytes to be recognized remaining in the second time and the large granular lymphocytes obtained by the first screening, as well as the cell distance, large granular lymphocytes are obtained by the second screening.

[0132] In an exemplary embodiment, as Figure 13 shown, the specific implementation process of the second screening is given as follows:

[0133] Step 3-2-1: Obtain the difference in the roundness of the cell nuclei of the fourth lymphocyte to be recognized and the reference large granular lymphocyte.

[0134] For the convenience of description, it is assumed that the fourth lymphocyte to be recognized is any one of the lymphocytes to be recognized remaining in the second time. It is assumed that the reference large granular lymphocyte is any one of the large granular lymphocytes obtained by the first screening.

[0135] From the shapes of large granular lymphocytes and atypical lymphocytes, it can be seen that the rounder the cell nucleus, the more likely it is to be a large granular lymphocyte. Therefore, it is necessary to obtain the difference in the roundness of the cell nuclei of the fourth lymphocyte to be recognized and the reference large granular lymphocyte for participation in the subsequent screening. In an exemplary embodiment, as Figure 14 shown, the process of obtaining the difference in roundness includes:

[0136] Step 3-2-1-1: Obtain the circumscribed circle of the cell nucleus of the fourth lymphocyte to be recognized, and obtain the ratio of the area of the cell nucleus of the fourth lymphocyte to be recognized to the area of its corresponding circumscribed circle as the roundness of the fourth lymphocyte to be recognized.

[0137] Step 3-2-1-2: Obtain the circumscribed circle of the cell nucleus of the reference large granular lymphocyte, and obtain the ratio of the area of the cell nucleus of the reference large granular lymphocyte to the area of its corresponding circumscribed circle as the roundness of the reference large granular lymphocyte.

[0138] Step 3-2-1-3: Obtain the difference between the roundness of the fourth lymphocyte to be recognized and the roundness of the reference large granular lymphocyte.

[0139] Among them, circumcircles are constructed for the nuclei of the fourth lymphocyte to be recognized and the reference large granular lymphocytes, and the area of the nucleus of the fourth lymphocyte to be recognized, the area of the circumcircle of the nucleus of the fourth lymphocyte to be recognized, the area of the nucleus of the reference large granular lymphocytes, and the area of the circumcircle of the nucleus of the reference large granular lymphocytes are obtained. Then, the ratio of the area of the nucleus of the fourth lymphocyte to be recognized to the area of the circumcircle of the nucleus of the fourth lymphocyte to be recognized, and the ratio of the area of the nucleus of the reference large granular lymphocytes to the area of the circumcircle of the nucleus of the reference large granular lymphocytes are calculated.

[0140] It should be understood that the larger the ratio of the areas, the smaller the difference in area between the nucleus and its circumcircle, the closer the two areas are, indicating that the nucleus is more circular and the roundness of the nucleus is greater. Therefore, the ratio of the area of the nucleus of the fourth lymphocyte to be recognized to the area of its corresponding circumcircle is used as the roundness of the shape of the fourth lymphocyte to be recognized; the ratio of the area of the nucleus of the reference large granular lymphocytes to the area of its corresponding circumcircle is used as the roundness of the shape of the reference large granular lymphocytes.

[0141] The difference in the roundness of the shape between the fourth lymphocyte to be recognized and the reference large granular lymphocytes is specifically the absolute value of the difference between the two roundnesses of the shapes.

[0142] Step 3-2-2: Obtain the cell distance between the fourth lymphocyte to be recognized and the reference large granular lymphocytes.

[0143] In an exemplary embodiment, the distance between each edge pixel point of the cytoplasmic outer edge of the fourth lymphocyte to be recognized and each edge pixel point of the cytoplasmic outer edge of the reference large granular lymphocytes is obtained, and then the minimum distance is obtained from the multiple distances obtained as the cell distance between the fourth lymphocyte to be recognized and the reference large granular lymphocytes.

[0144] Step 3-2-3: Obtain the reference weight of the reference large granular lymphocytes for the fourth lymphocyte to be recognized according to the difference in roundness of the shape and the cell distance.

[0145] The smaller the difference in the roundness of the shape, the more similar the roundness between the fourth lymphocyte to be identified and the reference large granular lymphocyte, that is, the more similar the shape. The more the fourth lymphocyte to be identified and the reference large granular lymphocyte indicate the same type of lymphocyte, that is, the fourth lymphocyte to be identified is more likely to be a large granular lymphocyte. Since the reference weight of the reference large granular lymphocyte for the fourth lymphocyte to be identified is the weight corresponding to the possibility that the fourth lymphocyte to be identified is a large granular lymphocyte, then the reference weight of the reference large granular lymphocyte for the fourth lymphocyte to be identified is greater; the smaller the cell distance between the fourth lymphocyte to be identified and the reference large granular lymphocyte, the greater the aggregation of other lymphocytes around the fourth lymphocyte to be identified may be, and the fourth lymphocyte to be identified is more likely to be a large granular lymphocyte. Then, the reference weight of the reference large granular lymphocyte for the fourth lymphocyte to be identified is greater. Therefore, the reference weight is inversely proportional to the difference in the roundness of the shape and inversely proportional to the cell distance.

[0146] Through the reference weight, the smaller the difference in the roundness of the shape, the greater the impact on determining the fourth lymphocyte to be identified as a large granular cell. The smaller the cell distance, the closer the reference large granular lymphocyte and the fourth lymphocyte to be identified are in space, and the greater the impact on determining the fourth lymphocyte to be identified as a large granular cell, enhancing the impact of spatial aggregation.

[0147] Step 3-2-4: Weight and average the second possibility that the reference large granular lymphocyte belongs to a large granular lymphocyte according to the reference weight to obtain the third possibility that the fourth lymphocyte to be identified belongs to a large granular lymphocyte.

[0148] The greater the second possibility that the reference large granular lymphocyte belongs to a large granular lymphocyte, the greater the weight it occupies when obtaining the possibility that the fourth lymphocyte to be identified belongs to a large granular lymphocyte. Therefore, weight and average the second possibility that the reference large granular lymphocyte belongs to a large granular lymphocyte according to the reference weight to obtain the third possibility that the fourth lymphocyte to be identified belongs to a large granular lymphocyte.

[0149] In an exemplary embodiment, a specific calculation formula for the third possibility is given as follows:

[0150]

[0151] Among them, Y j is the third possibility that the jth second remaining lymphocyte to be identified belongs to a large granular lymphocyte, D k is the second possibility that the kth reference large granular lymphocyte belongs to a large granular lymphocyte, K is the number of reference large granular lymphocytes, d j,kis the cell distance between the j-th lymphocyte to be recognized remaining after the second time and the k-th reference large granular lymphocyte, M j is the roundness of the j-th lymphocyte to be recognized remaining after the second time, M k is the roundness of the k-th reference large granular lymphocyte.

[0152] f(d j,k *|M j -M k |) represents the negative correlation normalization of d j,k *|M j -M k |. f(d j,k *|M j -M k |) is the reference weight of the k-th reference large granular lymphocyte for the j-th lymphocyte to be recognized remaining after the second time. The negative correlation normalization method here can be: obtain the maximum and minimum values among all reference weights, and then based on the maximum and minimum values, use the maximum-minimum normalization method to normalize d j,k *|M j -M k |, and then subtract the normalized d j,k *|M j -M k | from the value 1 to achieve the negative correlation normalization of d j,k *|M j -M k |.

[0153] Step 3-2-5: Determine whether the fourth lymphocyte to be recognized belongs to large granular lymphocytes according to the third possibility.

[0154] The greater the third possibility, the greater the possibility of belonging to large granular lymphocytes. Therefore, in an exemplary embodiment, a second large granular lymphocyte possibility threshold is preset. The value range of the second large granular lymphocyte possibility threshold is 0-1, which is set according to the actual situation. The larger the second large granular lymphocyte possibility threshold is set, the higher the strictness of screening large granular lymphocytes with aggregation. In this embodiment, 0.8 is taken as an example.

[0155] Determine the fourth lymphocyte to be recognized corresponding to the third possibility greater than or equal to the second large granular lymphocyte possibility threshold as large granular lymphocytes, and achieve the second screening of large granular lymphocytes.

[0156] Step 4: Obtain abnormal lymphocytes according to the screened large granular lymphocytes.

[0157] Ordinary lymphocytes are obtained through screening in step 2, and large granular lymphocytes are obtained through screening in step 3. Then, the remaining lymphocytes in the lymphocyte image are atypical lymphocytes, thereby obtaining atypical lymphocytes. The classification of lymphocytes in the lymphocyte image is achieved.

[0158] In subsequent applications, the numbers of ordinary lymphocytes, large granular lymphocytes, and atypical lymphocytes in the lymphocyte image can be counted, integrated with other data analyzed by an automated microscope blood analyzer, and uploaded to the background system for storage for subsequent calls.

[0159] It should be noted that the above sequence of embodiments of the present invention is only for description and does not represent the superiority or inferiority of the embodiments. The processes depicted in the accompanying drawings do not necessarily require the specific order or continuous order shown to achieve the desired results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0160] Each embodiment in this specification is described in a progressive manner. The same or similar parts among the embodiments can be referred to each other, and each embodiment focuses on the differences from other embodiments.

Claims

1. A method for peripheral blood cell morphology-assisted extraction under microscopic imaging, characterized in that: include: Acquiring a lymphocyte image, wherein the lymphocytes to be identified in the lymphocyte image include common lymphocytes, large granular lymphocytes and atypical lymphocytes; Based on the distance between the nucleus and cytoplasm of the lymphocytes to be identified, ordinary lymphocytes are screened out; According to the density of lymphocytes around the remaining lymphocytes to be identified and the squeezing of the outer edges of the cytoplasm of the remaining lymphocytes to be identified, large granular lymphocytes are screened and obtained; Based on the large granular lymphocytes obtained by screening, atypical lymphocytes are obtained.

2. The method for peripheral blood cell morphology-assisted extraction under microscopic imaging according to claim 1, characterized in that: The screening of common lymphocytes based on the distance between the nucleus and the cytoplasm of the lymphocytes to be identified includes: Obtain the distances from the outer edge of the cell nucleus to the outer edge of the cytoplasm in multiple directions starting from the center of the cell nucleus of the lymphocyte to be identified, and obtain the distance mean and distance variance; According to the area of ​​the lymphocyte to be identified, the distance mean and the distance variance, a first possibility that the lymphocyte to be identified belongs to a common lymphocyte is obtained; the first possibility is inversely proportional to the area, inversely proportional to the distance variance, and inversely proportional to the distance mean; According to the first possibility, ordinary lymphocytes are screened and obtained.

3. The method for peripheral blood cell morphology-assisted extraction under microscopic imaging according to claim 1, characterized in that: The large granular lymphocytes are screened and obtained according to the density of lymphocytes around the remaining lymphocytes to be identified and the squeezing of the outer edges of the cytoplasm of the remaining lymphocytes to be identified, including: According to the density of lymphocytes around the remaining lymphocytes to be identified in the first time and the squeezing of the outer edges of the cytoplasm of the remaining lymphocytes to be identified in the first time, large granular lymphocytes are obtained in the first screening; Based on the similarity of the nuclear shapes of the remaining lymphocytes to be identified in the second screening and the large granular lymphocytes obtained in the first screening, as well as the cell distance, large granular lymphocytes were obtained in the second screening.

4. The method for peripheral blood cell morphology-assisted extraction under microscopic imaging as claimed in claim 3, characterized in that: The large granular lymphocytes obtained by the first screening according to the density of lymphocytes around the first remaining lymphocytes to be identified and the squeezing of the outer edges of the cytoplasm of the first remaining lymphocytes to be identified include: Obtaining a first number of second lymphocytes to be identified within a preset range around the first lymphocyte to be identified; the first lymphocyte to be identified is any lymphocyte to be identified remaining in the first time, and the second lymphocyte to be identified is other lymphocytes to be identified remaining in the first time except the first lymphocyte to be identified; Obtaining a second number of third lymphocytes to be identified that are in contact with the first lymphocytes to be identified; the third lymphocytes to be identified are the remaining lymphocytes to be identified except for the first lymphocytes to be identified; According to the first number, the second number and the squeezing condition of the outer edge of the cytoplasm, a second possibility is obtained that the first lymphocyte to be identified belongs to a large granular lymphocyte; According to the second possibility, it is determined whether the first lymphocyte to be identified is a large granular lymphocyte.

5. The method for peripheral blood cell morphology-assisted extraction under microscopic imaging according to claim 4, characterized in that: The process of obtaining the second quantity includes: Segmenting the outer edge of the cytoplasm of the first lymphocyte to be identified according to the grayscale value to obtain a plurality of line segments; The second number is obtained, where the second number is the number of first line segments, and the first line segments are line segments whose grayscale values ​​are less than a preset threshold.

6. The method for peripheral blood cell morphology-assisted extraction under microscopic imaging according to claim 5, characterized in that: The process of obtaining the squeezing condition of the outer edges of the cytoplasm of the first remaining lymphocytes to be identified comprises: Obtaining the length of each line segment in each of the first remaining lymphocytes to be identified whose second number is greater than 1, and obtaining a length mean; Obtain the standard line segment length according to the average length value; The length difference between the length of each line segment in the first lymphocyte to be identified and the length of the standard line segment is obtained, and the mean value of the length difference corresponding to the first lymphocyte to be identified is obtained.

7. The method for peripheral blood cell morphology-assisted extraction under microscopic imaging according to claim 6, characterized in that: The calculation formula of the second possibility is as follows: Among them, D i is the second possibility that the remaining lymphocyte to be identified in the first time belongs to a large granular lymphocyte, N i is the second number of lymphocytes to be identified corresponding to the first remaining i-th lymphocyte, M i is the first number of lymphocytes to be identified that remain for the first time, L i,n is the length of the nth first line segment in the i-th remaining lymphocyte to be identified, L0 is the standard line segment length, For The negative correlation normalization of , norm is the normalization function.

8. The method for peripheral blood cell morphology-assisted extraction under microscopic imaging as claimed in claim 3, characterized in that: Based on the similarity of the nucleus shapes of the remaining lymphocytes to be identified in the second screening and the large granular lymphocytes obtained in the first screening, as well as the cell distance, the large granular lymphocytes obtained in the second screening include: Obtaining a difference in the roundness of the shape of the nucleus of a fourth lymphocyte to be identified and a reference large granular lymphocyte; the fourth lymphocyte to be identified is any one of the lymphocytes to be identified remaining in the second screening, and the reference large granular lymphocyte is any one of the large granular lymphocytes obtained in the first screening; Obtaining the cell distance between the fourth lymphocyte to be identified and the reference large granular lymphocyte; According to the shape roundness difference and the cell distance, obtaining a reference weight of the reference large granular lymphocyte to the fourth lymphocyte to be identified, wherein the reference weight is inversely proportional to the shape roundness difference and inversely proportional to the cell distance; The second possibility that the reference large granular lymphocyte belongs to the large granular lymphocyte is weighted according to the reference weight and the average is calculated to obtain a third possibility that the fourth lymphocyte to be identified belongs to the large granular lymphocyte; It is determined whether the fourth lymphocyte to be identified is a large granular lymphocyte based on the third possibility.

9. The method for peripheral blood cell morphology-assisted extraction under microscopic imaging according to claim 8, characterized in that: The process of obtaining the difference in shape roundness includes: Obtaining a circumscribed circle of a cell nucleus of a fourth lymphocyte to be identified, and obtaining a ratio of an area of ​​the cell nucleus of the fourth lymphocyte to be identified to an area of ​​a corresponding circumscribed circle thereof, as a roundness of a shape of the fourth lymphocyte to be identified; Obtaining the circumscribed circle of the cell nucleus of a reference large granular lymphocyte, and obtaining the ratio of the area of ​​the cell nucleus of the reference large granular lymphocyte to the area of ​​the corresponding circumscribed circle as the roundness of the shape of the reference large granular lymphocyte; Obtain a difference between the roundness of the shape of the fourth lymphocyte to be identified and the roundness of the shape of the reference large granular lymphocyte.

10. The method for peripheral blood cell morphology-assisted extraction under microscopic imaging according to claim 8, characterized in that: The calculation formula of the third possibility is as follows: Among them, Y j is the third possibility that the jth remaining lymphocyte to be identified belongs to a large granular lymphocyte, D k is the second possibility that the kth reference large granular lymphocyte belongs to large granular lymphocytes, K is the number of reference large granular lymphocytes, d j,k is the cell distance between the jth remaining lymphocyte to be identified and the kth reference large granular lymphocyte, M j is the roundness of the shape of the jth remaining lymphocyte to be identified, M k is the roundness of the kth reference large granular lymphocyte, f(d j,k *|M j -M k |) indicates the j,k *|M j -M k |Normalized negative correlation.

Images