EBV antibody determination method and system applied to auxiliary screening and diagnosis of nasopharyngeal carcinoma
Through pixel point clustering and contour search technology, EBV antibody detection is automatically processed, which solves the difficulty and inconsistency of manual interpretation, and improves the accuracy of early screening and diagnosis of nasopharyngeal carcinoma.
Patent Information
- Application Number
- CN202411001717.8
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-07-25
- Publication Date
- 2024-11-15
AI Technical Summary
The existing EBV antibody detection methods rely on manual interpretation, which are difficult, time-consuming, inconsistency and misjudgment risks, affecting the accuracy of early screening and diagnosis of nasopharyngeal carcinoma.
Cluster analysis is performed using grayscale differences and gradient differences of pixel points, and the specific cells in fluorescent pictures are automatically identified, and the target screening is interfered with. The target number is counted to determine whether the antibody is positive or negative, which improves detection efficiency and accuracy. .
Automatic EBV antibody detection is realized, which reduces errors in manual interpretation, improves the diagnostic efficiency and accuracy of nasopharyngeal carcinoma and EB virus infection-related diseases, and reduces the risk of misjudgment.
Smart Images

Figure CN118969243A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the field of medical detection technology, and in particular to an EBV antibody determination method, system, device and storage medium for auxiliary screening and diagnosis of nasopharyngeal carcinoma. Background Art
[0002] Nasopharyngeal carcinoma is a malignant tumor originating from the nasopharynx, which is highly invasive and metastatic. At present, the diagnosis of nasopharyngeal carcinoma mainly relies on methods such as pathological biopsy and imaging examination. Although these methods are accurate, they have the disadvantages of being highly invasive, expensive, and having poor sensitivity. Studies have shown that infection with Epstein-Barr Virus (EBV or EB virus) is the main risk factor for the occurrence of nasopharyngeal carcinoma. Among patients with nasopharyngeal carcinoma, more than 90% of EB virus tests are positive; in the years before the diagnosis of nasopharyngeal carcinoma, the level of immunoglobulin A (IgA) antibodies to EB virus in patients with nasopharyngeal carcinoma is significantly increased, especially in the years before the diagnosis. High levels of EBV IgA antibodies are significantly associated with the risk of developing nasopharyngeal carcinoma in the future. Therefore, EBV antibody determination is of great significance in the early screening and diagnosis of nasopharyngeal carcinoma.
[0003] Epstein-Barr virus capsid antigen (VCA) and early antigen (EA) are two key Epstein-Barr virus antigens used in the diagnosis of nasopharyngeal carcinoma. The current methods for determining EBV antibodies mainly include enzyme-linked immunosorbent assay (ELISA), indirect immunofluorescence (IIF) and immunoblotting. Among them, indirect immunofluorescence (IIF) is considered to be the gold standard for detecting VCA-IgA and EA-IgA antibodies (CACA guidelines for holisticintegrative management of nasopharyngeal carcinoma|Holistic IntegrativeOncology (springer.com)). The basic principle of this method is as follows: the patient's serum sample is dripped onto a glass slide containing Epstein-Barr virus antigens (VCA or EA), and then fluorescently labeled anti-human immunoglobulin (referred to as labeled antibody) is added. If the patient's serum sample contains target antibodies (VCA-IgA or EA-IgA), these target antibodies will bind to the inherent antigens (VCA or EA) on the glass slide. After the labeled antibody is added, it will bind to the target antibody that has already bound to the antigen. At this time, it will show specific fluorescence characteristics with a significantly increased cell volume and significantly enhanced fluorescence intensity (these antibody cells bound to the antigen are usually recorded as target cells or specific cells). By identifying the specific fluorescence characteristics of the cells, the positive and negative of the patient's serum sample antibodies can be determined, thereby achieving the purpose of EBV antibody determination.
[0004] At present, in clinical applications, the indirect immunofluorescence images of VCA-IgA and EA-IgA used for EBV antibody determination are mainly interpreted manually by inspectors, or interpreted under a microscope by experienced technicians. This manual interpretation method has the following shortcomings:
[0005] First, manual interpretation is difficult and complex. Due to the diversity of fluorescence signal intensity and cell morphology, nonspecific cells may also change their morphology and fluorescence intensity in actual operation, and these cells are recorded as interfering cells. The existence of these interfering cells poses a great challenge to clinical laboratory personnel and is prone to the risk of misjudgment.
[0006] Secondly, the inconsistency of interpretation results. Different inspectors may have large differences in their interpretation of the same sample, resulting in inconsistent results, affecting the accuracy and reliability of the diagnosis.
[0007] In addition, manual interpretation is time-consuming and labor-intensive. Manual interpretation requires a lot of time and energy, and is prone to fatigue in a high-load working environment, further increasing the risk of misjudgment.
[0008] In addition, manual interpretation can easily lead to a waste of medical resources. Since EBV infection is universal, more than 90% of adults have positive serum EBV antibodies, which will inevitably lead to false positive results in the diagnosis of nasopharyngeal carcinoma, thus resulting in a waste of medical resources.
[0009] Based on the above shortcomings, although indirect immunofluorescence (IIF) is the gold standard for detecting VCA-IgA and EA-IgA antibodies, and provides an effective way to measure EBV antibodies associated with NPC, the manual interpretation of the results is not only difficult and time-consuming, but also susceptible to subjective factors, which increases the risk of misjudgment. Therefore, there is an urgent need for a method to solve the above problems in order to improve the efficiency and accuracy of EBV antibody detection. Summary of the invention
[0010] In view of the technical problems existing in the prior art, the present invention proposes an EBV antibody determination method, system, device and storage medium for auxiliary screening and diagnosis of nasopharyngeal carcinoma, which improves the efficiency and accuracy of screening and diagnosis when the indirect immunofluorescence method is used to determine EBV antibodies.
[0011] In order to solve the above technical problems, according to one aspect of the present invention, the present invention provides an EBV antibody determination method for auxiliary screening and diagnosis of nasopharyngeal carcinoma, comprising the following steps:
[0012] Cluster analysis is performed on the pixels of the fluorescent image of the test sample prepared based on the indirect immunofluorescence method according to the grayscale difference and gradient difference of the pixels, and the pixels of the fluorescent image are clustered into three clusters, wherein the antigen of the test sample is VCA or EA;
[0013] The three clusters are respectively determined as a first foreground target cluster, a second foreground target cluster and a third background cluster based on the number of pixels and grayscale mean of the clusters;
[0014] The pixel points constituting the first type of foreground target cluster are assigned a first gray value, and the remaining pixel points are assigned a second gray value to obtain a first type of foreground target mask;
[0015] Performing contour search on the target in the first-category foreground target mask to obtain a target image, wherein the plurality of target images constitute a candidate first-category foreground target set;
[0016] Determining interference targets from a plurality of targets in the candidate first-category foreground target set based on grayscale values and target contours respectively and screening them out to obtain a first-category foreground target set; and
[0017] The number of targets in the first type of foreground target set is counted, and when the number of targets is greater than or equal to a first number threshold, the antibody of the test sample is determined to be positive, and when the number of targets is less than the first number threshold, the antibody of the test sample is determined to be negative.
[0018] According to another aspect of the present invention, the present invention also provides an EBV antibody determination system for auxiliary screening and diagnosis of nasopharyngeal carcinoma, comprising a clustering module, a mask acquisition module, a target screening module and a judgment module, wherein the clustering module is configured to perform cluster analysis on the pixel points of the detection sample fluorescence image prepared based on the indirect immunofluorescence method according to the grayscale difference and gradient difference of the pixel points, and cluster the pixels of the fluorescence image into three cluster clusters, wherein the antigen of the detection sample is VCA or EA; the mask acquisition module is connected to the clustering module, and is configured to determine the three cluster clusters as a first foreground cluster cluster, a second foreground cluster cluster and a picture background cluster cluster based on the number of pixels and the grayscale mean of the cluster clusters; the pixel points constituting the first foreground cluster cluster are assigned as the first The target screening module is connected to the mask acquisition module, and is configured to perform cell contour search on the first-category foreground target mask to obtain a target image, wherein a plurality of target images constitute a candidate first-category foreground target set, and interference targets are determined from a plurality of targets in the candidate first-category foreground target set based on grayscale values and target contours and screened out to obtain a first-category foreground target set; the judgment module is connected to the target screening module, and is configured to count the number of targets in the first-category foreground target set, and when the number of targets is greater than or equal to a first number threshold, it is determined that the antibody of the test sample is positive, and when the number of targets is less than the first number threshold, it is determined that the antibody of the test sample is negative.
[0019] According to another aspect of the present invention, the present invention also provides an electronic device, which includes a processor and a memory, wherein a computer program instruction set is stored on the memory, and when the processor executes the computer program instruction set on the memory, the aforementioned EBV antibody determination method for auxiliary screening and diagnosis of nasopharyngeal carcinoma is implemented.
[0020] According to another aspect of the present invention, the present invention also provides a computer-readable storage medium, wherein a computer program instruction set is stored on the computer-readable storage medium, and when the computer program instruction set is executed by a processor, the aforementioned EBV antibody determination method for auxiliary screening and diagnosis of nasopharyngeal carcinoma is implemented.
[0021] According to another aspect of the present invention, the present invention also provides a computer program product, which includes a computer program instruction set, and when the computer program instruction set is executed by a processor, the aforementioned EBV antibody determination method for auxiliary screening and diagnosis of nasopharyngeal carcinoma is implemented.
[0022] The method and system provided by the present invention replace the manual interpretation of the fluorescent images of the test samples, effectively solve the defects caused by manual interpretation, have fast processing speed and high detection accuracy, and effectively improve the screening and diagnosis efficiency of patients with nasopharyngeal carcinoma and Epstein-Barr virus infection-related diseases. BRIEF DESCRIPTION OF THE DRAWINGS
[0023] The preferred embodiments of the present invention will be further described in detail below with reference to the accompanying drawings, wherein:
[0024] Figure 1 This is a schematic diagram of equipment connection of an EBV antibody assay system for auxiliary screening and diagnosis of nasopharyngeal carcinoma according to one embodiment of the present invention;
[0025] Figure 2 is a flow chart of an EBV antibody determination method for auxiliary screening and diagnosis of nasopharyngeal carcinoma according to one embodiment of the present invention;
[0026] Figure 3 is a flow chart of a method for performing unsupervised clustering processing on an original image of a detection sample fluorescence image according to an embodiment of the present invention;
[0027] Figure 4 This is a schematic diagram of a fluorescent image of a test sample with the antigen CA prepared based on an indirect immunofluorescence method according to an embodiment of the present invention;
[0028] Figure 5 is a schematic diagram of an image after clustering and re-assigning pixel values according to an embodiment of the present invention;
[0029] Figure 6 is a schematic diagram of a first type of foreground object mask according to an embodiment of the present invention;
[0030] Figure 7 is a schematic diagram of a second type of foreground object mask according to an embodiment of the present invention;
[0031] Figure 8 is a flow chart of a processing method for searching a target contour using a Canny edge detection method according to an embodiment of the present invention;
[0032] Fig. 9 is a schematic diagram of an image obtained after searching for a target contour according to an embodiment of the present invention;
[0033] Fig.10is a flow chart of a method for determining an interference target based on a gray value according to an embodiment of the present invention;
[0034] Fig.11 is a schematic diagram of an image after ellipse fitting is performed on a target according to an embodiment of the present invention;
[0035] Fig.12 is a schematic diagram of an image after ellipse fitting is performed on another target according to an embodiment of the present invention;
[0036] Fig.13 is a flow chart of a titer level determination method according to one embodiment of the present invention;
[0037] Fig.14 This is a principle block diagram of an EBV antibody determination system for auxiliary screening and diagnosis of nasopharyngeal carcinoma according to an embodiment of the present invention;
[0038] Fig.15 is a schematic diagram of a ROC curve obtained by testing the method flow of the present invention through data set 1 according to an embodiment of the present invention;
[0039] Fig.16 is a schematic diagram of a ROC curve obtained by testing the method flow of the present invention through data set 2 according to an embodiment of the present invention; and
[0040] Fig.17 It is a schematic diagram of the hardware structure principle of an electronic device according to an embodiment of the present invention. DETAILED DESCRIPTION
[0041] In order to make the purpose, technical solution and advantages of the embodiments of the present invention clearer, the technical solution in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.
[0042] In the following detailed description, reference may be made to the various specification drawings that are part of the present application and are used to illustrate specific embodiments of the present application. In the accompanying drawings, similar reference numerals describe substantially similar components in different figures. The various specific embodiments of the present application are described below in sufficient detail so that a person of ordinary skill in the art with relevant knowledge and skills in the art can implement the technical solutions of the present application. It should be understood that other embodiments may also be used or structural, logical or electrical changes may be made to the embodiments of the present application.
[0043] See also Figure 1 , Figure 11 is a schematic diagram of the device connection of an EBV antibody determination system for nasopharyngeal carcinoma auxiliary screening and diagnosis according to an embodiment of the present invention. In this embodiment, the EBV antibody determination system is implemented by one or more computing devices 101, such as a desktop computer, a laptop computer or a tablet computer, and the computing device 101 is connected to a fluorescence microscope 102 with a camera, or connected to other storage devices 103 or connected to other devices via a network 104, and the computing device 101 obtains a fluorescence image of a test sample prepared based on an indirect immunofluorescence method from these devices. For example, when the computing device 101 is connected to a fluorescence microscope 102 connected to a camera, a fluorescence image of the test sample can be obtained in real time. The specific operation steps are, for example: first, drip the patient serum onto a glass slide containing EB virus antigen (VCA or EA), and then add fluorescently labeled anti-human immunoglobulin thereto to obtain a test sample slide; then place the test sample slide under the fluorescence microscope 102 connected to a camera; then adjust the optical path; then control the camera to randomly select 1 to 8 fields of view for fluorescence film shooting, and then obtain 1 to 8 fluorescence images of the same test sample. The fluorescence microscope 102 sends the obtained fluorescence image to the computing device 101. In addition, the fluorescence image of the test sample can also be taken by a fluorescence microscope connected to other computing devices and stored in a storage device 103, and then the storage device 103 is connected to the computing device 101, and the computing device 101 reads the test sample fluorescence image from the storage device 103, or the computing device 101 is connected to other devices through the network 104, and reads or requests the test sample fluorescence image to be processed. The present invention implements an EBV antibody determination method for nasopharyngeal carcinoma auxiliary screening and diagnosis provided by the present invention through an EBV antibody determination system implemented by a computing device 101, and determines whether the antibody of the test sample is negative or positive by processing the fluorescent image of the test sample prepared based on the indirect immunofluorescence method. The processing method provided by the present invention replaces the manual interpretation of the fluorescent image of the test sample, effectively solves the defects caused by manual interpretation, has a fast processing speed and high detection accuracy, and effectively improves the diagnostic efficiency of patients with nasopharyngeal carcinoma and EB virus infection-related diseases.
[0044] Figure 2 is a flow chart of an EBV antibody determination method for nasopharyngeal carcinoma auxiliary screening and diagnosis according to an embodiment of the present invention, combined with Figure 1 , Figure 2 The method shown can be Figure 1 The method is implemented by the computing device 101 in the embodiment of the present invention, and includes:
[0045] Step S1, clustering analysis is performed on the pixels of the fluorescence image of the test sample prepared based on the indirect immunofluorescence method according to the grayscale difference and gradient difference of the pixels, and the pixels of the fluorescence image are clustered into three clusters. The three clusters correspond to positive (or specific) cell clusters, negative (or non-specific) cell clusters and image background, respectively.
[0046] Step S2: assigning the same grayscale value to the pixels in each cluster to obtain a corresponding cluster mask, wherein the grayscale values of the pixels in each cluster in the cluster mask are different from each other.
[0047] Step S3, based on the number of pixels and grayscale mean, the three clusters in the cluster mask are respectively determined as a first foreground target cluster mask, a second foreground target cluster mask and a third background cluster mask. In the present invention, unless otherwise specified, the target refers to a cell in an image.
[0048] Step S4, obtaining corresponding target masks based on the clustering masks. Specifically, based on the original fluorescent image or the first / second foreground target clustering mask, the pixel points of the first foreground target clustering mask are assigned to the first grayscale value, and the remaining pixels are assigned to the second grayscale value to obtain the first foreground target mask. In addition, the pixel points of the second foreground target clustering mask can also be assigned to the first grayscale value, and the remaining pixels can be assigned to the second grayscale value to obtain the second foreground target clustering mask.
[0049] Step S5, performing contour search on the target in the first-category foreground target mask to obtain a plurality of target images, wherein the plurality of target images constitute a candidate first-category foreground target set.
[0050] Step S6, determining interference targets from the multiple target images based on grayscale values and target contours respectively and screening out interference targets to obtain a first category foreground target set.
[0051] Step S7, counting the number of targets in the first category of foreground target set.
[0052] Step S8, determining whether the target quantity is greater than or equal to a first quantity threshold. When the target quantity is greater than or equal to the first quantity threshold, determining that the antibody of the test sample is positive in step S9; when the target quantity is less than the first quantity threshold, determining that the antibody of the test sample is negative in step S10.
[0053] In step S1, unsupervised clustering processing is performed on the original image of the detection sample fluorescence image. According to the detection sample, it can be known that the detection sample includes non-specific cells, such as various cells in the patient's serum (or simply referred to as negative cells), and may also include antigen-antibody cells combined with labeled antibodies (or simply referred to as positive cells). For the detection sample fluorescence image, the image may contain images of normal cells, images of antigen-antibody cells labeled with antibodies, and images of blank areas without cells. Therefore, when unsupervised clustering processing is performed on the original image of the detection sample fluorescence image, the attribution of the pixels in the image is set to 3 categories, representing specific cells, non-specific cells and image background, respectively.
[0054] See also Figure 3 , Figure 3 The flowchart of performing unsupervised clustering processing on the original image of the detection sample fluorescence image according to an embodiment of the present invention specifically includes:
[0055] Step S11, randomly select 3 points in the original image as the centroid of the set.
[0056] Step S12, respectively calculate the distance between each pixel in the image and the centroids of the three sets, where grayscale difference and gradient difference are used as the distance.
[0057] Step S13, dividing all pixel points into three groups according to distance.
[0058] Step S14, calculating the new centroid of each set respectively.
[0059] Step S15, respectively calculating the distance difference between the new centroid and the original centroid of each set.
[0060] Step S16, determine whether the distance difference is less than the centroid distance threshold. If the distance difference is less than the centroid distance threshold, then in step S17, determine that an ideal cluster is obtained, terminate the iteration, and end the clustering process. If the distance difference is greater than or equal to the centroid distance threshold, return to step S12.
[0061] In one embodiment, three ideal clusters are obtained after a limited number of iterations (such as 100), which respectively represent specific cells, nonspecific cells, and image background. The present invention can effectively distinguish specific cells, nonspecific cells, and image background from images by setting appropriate grayscale differences, thereby obtaining ideal clusters.
[0062] In order to facilitate subsequent data processing, Figure 1Before or after step S1, the image is further grayed and filtered. Specifically, the original fluorescent image is grayed to obtain a single-channel grayscale image, for example, the grayscale value of the G channel of the original image is extracted as the grayscale value after graying, thereby obtaining the G channel grayscale image. Since noise may be left in the original image due to dust or other reasons in the slide, the single-channel grayscale image can be filtered to remove the noise in the image, and then the image after graying and filtering and the corresponding grayscale value and other data are stored.
[0063] In one embodiment, a median filter is used to remove noise from the image. The grayscale calculation formula for filtering is shown in the following formula 1-1:
[0064] g(x,y)=med{f(xk,yl),(k,l∈W)} 1-1
[0065] Among them, f(x, y), g(x, y) correspond to the original image and the processed image respectively, k is the convolution center value, and l is the sliding window size.
[0066] Step S12 is an optional step, and the pixel points in each cluster can be assigned values based on the grayscale values of the pixel points in the image after the grayscale and filtering processing. For example, the maximum grayscale value is determined from the cluster with the least number of pixels, and all the pixels in the cluster are assigned the maximum grayscale value. For the cluster ranked second in terms of the number of pixels, the grayscale value with a larger number of pixels is determined, and all the pixels in the cluster are assigned the grayscale value. For the third cluster with the largest number of pixels and a lower grayscale value, the grayscale value of all the pixels therein is assigned to 0, so that the clusters can be clearly distinguished. See Figure 4 and Figure 5 , Figure 4 is a schematic diagram of a fluorescence image of a test sample in which the antigen is VCA prepared based on an indirect immunofluorescence method according to an embodiment of the present invention, Figure 5 1 is a schematic diagram of a picture after clustering and re-assigning pixel values according to an embodiment of the present invention. In this embodiment, a clustering mask is obtained through step S12, which is convenient for data backtracking and convenient for checking clustering results. When the clustering mask is obtained without processing, subsequent steps S3 and S4 perform corresponding processing based on the cluster clusters.
[0067] Generally, the fluorescence intensity of positive cells in the fluorescence image of the detected sample is high and the number is generally small, so the number of pixels occupied is also small, while the number of negative cells is relatively large, and the fluorescence intensity is higher than the background and lower than the fluorescence intensity of positive cells. Based on the above characteristics, the three clusters determined above can be preliminarily determined as the first type of foreground target clusters, the second type of foreground target clusters, and the third type of background clusters according to the number of pixels and grayscale mean corresponding to different clusters. Specifically, the cluster with the least number of pixels is determined as the first type of foreground target cluster, the cluster with the largest number of pixels and the lowest grayscale mean is determined as the third type of background cluster, and the cluster ranked second and with a grayscale mean significantly higher than the third type of background cluster is determined as the second type of foreground target cluster.
[0068] In order to facilitate subsequent pixel processing, in one embodiment, a first label is set for all pixels in the first category of foreground target clusters and recorded as 0, a second label is set for all pixels in the second category of foreground target clusters and recorded as 1, and a third label is set for all pixels in the third category of background clusters and recorded as 2. Although the label content in the current embodiment is a specific number 0-2, it can of course be other numbers or symbols, or only the first category of foreground target clusters and the second category of foreground target clusters are marked, or only the first category of foreground target clusters are marked according to the number and type of target masks.
[0069] In step S4, specifically, in an original image or the aforementioned first-category foreground target clustering mask or the second-category foreground target clustering mask, the grayscale values of all pixels with a label of 0 are assigned to 0, and the grayscale values of the remaining pixels are assigned to 255. The image obtained after the assignment is recorded as the first-category foreground target mask, such as Figure 6 As shown; then copy another original image or another first-category foreground target cluster mask or second-category foreground target cluster mask, assign the grayscale value of all pixels with label 1 to 0, and assign the grayscale value of the remaining pixels to 255. The image obtained after the assignment is recorded as the second-category foreground target mask, as shown Figure 7 As shown. After obtaining the second-category foreground target mask, the grayscale value of the pixel point corresponding to the second-category foreground target mask is obtained from the grayscale image of the aforementioned filtered fluorescence image according to the second-category foreground target mask, and the grayscale mean of all the pixel points in the second-category foreground target mask is calculated, that is, the grayscale mean of all the pixel points with a grayscale value of 0 in the second-category foreground target mask in the original image is calculated, and recorded as the second-category foreground target grayscale mean. It should be noted that Figure 7The second type of foreground target mask shown is not necessary. The second type of foreground target mask obtained by processing can be convenient for viewing the intermediate processing process or facilitating data backtracking. Therefore, the corresponding grayscale value can be obtained from the original image based on the second type of foreground target cluster obtained above and the second type of foreground target grayscale mean value can be calculated.
[0070] In step S5, the first type of foreground target mask obtained in step S4 is used to perform target contour search. In one embodiment, an edge detection algorithm is used to perform contour search. For example, Figure 8 1 is a flow chart of a processing method for searching a target contour using a Canny edge detection method according to an embodiment of the present invention, the method comprising the following steps:
[0071] Step S51, the current image (such as Figure 6 The first type of foreground target mask shown in the figure) uses the Sobel operator to calculate the first-order derivatives in the horizontal and vertical directions to obtain a gradient map, where the first-order derivative is the image gradient.
[0072] Step S52, determining the gradient and direction of the boundary according to the obtained gradient map.
[0073] Step S53, checking each pixel point, and determining the maximum grayscale value from the pixels points with the same gradient direction around each pixel point.
[0074] Step S54, determining the maximum and minimum gradient thresholds according to the grayscale maximum value, and determining the true edge according to the gradient thresholds.
[0075] Step S55, setting the grayscale value of the pixel points belonging to the edge to 255, and setting the grayscale value of other pixel points to 0.
[0076] Step S56, traverse all current pixel points, and when a pixel point with a grayscale value of 255 is found in the 4-neighborhood or 8-neighborhood of a pixel point, determine the pixel point with a grayscale value of 255 as a contour point.
[0077] Step S57, connect all the contour points to obtain the complete contour of a target (cell) in the first type of foreground target mask. After the above processing, the specific pixel points that constitute the contour of a cell are determined from the first type of foreground target mask. Fig. 9 As shown, Fig. 9 is a schematic diagram of an image obtained after searching for a target contour according to an embodiment of the present invention. Fig. 9 It can be seen that after Figure 8The process shown identifies multiple cells, and these cells are specific cells determined based on the aforementioned clustering. However, due to the complexity of the cells themselves, some non-specific cells may also undergo changes in morphology and fluorescence intensity. Therefore, among the specific cells determined based on clustering, there may be individual non-specific cells with larger fluorescence intensity, which are referred to as interference targets in the following description. Therefore, in order to be able to screen out non-specific cells that are mistakenly clustered together during clustering from the currently obtained cell population, they are first screened out according to brightness (specifically grayscale values are used in this embodiment), and then further screened out according to morphology, thereby screening out non-specific cells to the maximum extent.
[0078] See also Fig.10 , Fig.10 The method for determining an interference target based on grayscale value according to an embodiment of the present invention is a flow chart. Specifically, the method comprises the following steps:
[0079] Step S61, based on Fig. 9 The pixel points of each target determined in , calculate the minimum bounding rectangle of each target image.
[0080] Step S62 , according to the minimum bounding rectangle of each target image, the corresponding target image is intercepted on the filtered single-channel grayscale image as the target image A to be processed.
[0081] Step S63, after performing adaptive threshold segmentation and binarization processing on each target image A to be processed, a segmentation mask of the target image A to be processed is obtained, wherein the grayscale values in the mask A of the target image to be processed are respectively the first grayscale value and the second grayscale value after the binarization processing, wherein the first grayscale value is less than the second grayscale value. In one embodiment, the first grayscale value is 0 and the second grayscale value is 255.
[0082] The process of adaptive threshold segmentation includes:
[0083] First, grayscale the target image A to be processed; then traverse the grayscale values of the pixels in the grayscale image to establish a grayscale histogram; then count the number of pixels in each grayscale level to obtain the probability of the pixel being at different grayscale levels; then traverse the grayscale values of the pixels in the image from the minimum to the maximum as the threshold, and all pixels with grayscale values higher than the current threshold will be recorded as category A, otherwise they will be recorded as category B; calculate the variance σ between category A and category B under different thresholds respectively 2 Variance σ 2 The calculation formula 1-2 is as follows:
[0084] σ 2 =P A (t)P B (t)(M A (t)-MB (t)) 2 1-2
[0085] Among them, P A (t) is the probability that the pixel is classified as category A when the threshold is t; P B (t) is the probability that the pixel is classified as category B when the threshold is t; M A (t) is the grayscale mean of the pixels in class A when the threshold is t, M B (t) is the grayscale mean of the pixels in class B when the threshold is t.
[0086] When the threshold t makes σ 2 When the value is the largest, it is considered that the difference between categories A and B is the most obvious. At this time, the threshold is positioned as the optimal segmentation threshold.
[0087] After obtaining the optimal segmentation threshold, the original grayscale image of the target image A to be processed is binarized using the optimal segmentation threshold, that is, the pixels with grayscale values higher than the optimal segmentation threshold are assigned a value of 0, and the pixels with grayscale values lower than the threshold are assigned a value of 255. At this time, the segmentation mask of the target image A to be processed is obtained.
[0088] Step S64, calculating the grayscale mean of the target image A to be processed. Specifically, by comparing all the pixels whose grayscale values in the segmentation mask of the target image A to be processed are the first grayscale value (such as 0), the grayscale values of the corresponding pixels are read from the target image A to be processed and the grayscale mean of the target image A to be processed is calculated.
[0089] Step S65, obtaining the grayscale mean of the second-category foreground target. If the grayscale mean of the second-category foreground target has not been calculated before, the grayscale values of all pixels in the second-category foreground target cluster are obtained from the filtered single-channel grayscale image, or the grayscale values of all pixels of the first grayscale value in the second-category foreground target mask are obtained from the filtered single-channel grayscale image, and then the grayscale mean of all pixels is calculated as the grayscale mean of the second-category foreground target.
[0090] Step S66, calculating the difference between the grayscale mean of the target image A to be processed and the grayscale mean of the second type of foreground target.
[0091] Step S67, judging whether the difference is less than the gray mean threshold, if the difference is less than the gray mean threshold, indicating that the brightness of the current target image A to be processed is not much different from the brightness of the second type of foreground target representing non-specific cells, then in step S68, determining that the target corresponding to the target image A to be processed is an interference target. If the difference is greater than or equal to the gray mean threshold, indicating that the brightness of the current target image A to be processed is significantly different from the brightness of the second type of foreground target representing non-specific cells, then in step S69, determining that the target image A to be processed is a target representing specific cells.
[0092] right Fig. 9 The above process is repeated for each target in the image, and a portion of non-specific cells are screened out. Then, interference is determined from the multiple targets currently remaining based on the target contour. Specifically, an ellipse fitting is performed on the contour of each target to obtain a target ellipse and its major axis and minor axis. Among them,
[0093] Among them, the ellipse fitting equation is shown in the following formula 1-3:
[0094]
[0095] Among them, point (h, k) is the center of the ellipse, a and b are the semi-axis lengths along the x-axis and y-axis respectively, and a ≥ b, θ is the rotation angle of the ellipse.
[0096] When fitting an ellipse to the contour of a target, randomly select points on the contour of the target and substitute them into the ellipse fitting equation, and obtain the optimal solution of the equation through iterative calculation, so as to fit the ellipse of the target and its major and minor axes. Fig.11 and Fig.12 As shown, Fig.11 It is a schematic diagram of an image after ellipse fitting is performed on the contour of a target according to an embodiment of the present invention. Fig.12 It is a schematic diagram of an image after ellipse fitting is performed on the contour of another target according to an embodiment of the present invention.
[0097] The ratio of the major axis to the minor axis of the ellipse of each target is calculated, and the area corresponding to each target outline is calculated, wherein the area corresponding to each target outline is, for example, the number of all pixel points constituting the current target.
[0098] The target is determined to be an interference target when the ratio of the major axis to the minor axis is greater than or equal to the axis ratio threshold and the area corresponding to the target contour is less than the area threshold.
[0099] The above process is repeated to screen out interfering targets representing nonspecific cells based on cell morphology, and then the remaining targets are marked as specific cells.
[0100] In step S8, the first quantity threshold for judging whether the antibody of the test sample is positive varies according to the matrix (i.e., antigen) in the test sample. For example, in one embodiment, for the test sample whose antigen is VCA, the first quantity threshold is set to 5, that is, when there are 5 or more specific cells, the antibody of the test sample is determined to be positive; for the test sample whose antigen is EA, the first quantity threshold is set to 1, that is, as long as 1 specific cell is detected, the antibody of the test sample is determined to be positive. In another embodiment, the first quantity threshold is a quantity range, which is used to correspond to the different requirements of the user for the strictness of judgment. For example, for the test sample whose antigen is VCA, the first quantity threshold is set to [4,8], that is, any value from 4 to 8 can be used as the first quantity threshold, and the larger the value of the first quantity threshold, the higher the strictness of the judgment. According to clinical experience, the first quantity threshold cannot be set too low, and a false positive result will be caused when the value is too low.
[0101] Further, after a test sample is determined to be antibody positive based on the above method, the titer level classification can also be performed to determine the titer level of the test sample. In one embodiment, the fluorescent image of the test sample is input into a trained classification model, and the titer level of the test sample is obtained after processing by the classification model, wherein the classification model is trained to achieve titer level classification of a preset number of levels. The classification model is, for example, a machine learning model or a deep learning model, and 5 levels are set during training (corresponding to different dilution ratios of the sample to be tested), such as titer level 1 (1:10), titer level 2 (1:32), titer level 3 (1:100), titer level 4 (1:320), and titer level 5 (1:n, n>320). The ratio in the brackets is the corresponding test titer reported by the clinic at a specific sample dilution ratio. The input of the classification model is the fluorescent image of the test sample, and the output is any one of the 5 levels.
[0102] See also Fig.13 , Fig.13 : is a flow chart of a titer level determination method according to an embodiment of the present invention. First, the fluorescent image of the current positive test sample is resized, for example, the original size is adjusted to 512×512. Then the image is transformed in dimension, for example, the original dimension of the image 512×512 is stretched to 1×262144. Then it is input into a machine learning classifier, and the machine learning classifier is used to classify the titer level to obtain the final titer level.
[0103] The fluorescent images used in the above examples are all qualified images after quality screening. Figure 1When the fluorescence microscope 102 in the apparatus performs real-time photography, fluorescence images are usually taken for a plurality of randomly selected fields of view of a test sample slide, and then the one with the best quality is selected from the plurality of fluorescence images for the aforementioned processing. Specifically, each fluorescence image is tested according to the preset detection items, and when the fluorescence image meets the quality conditions, it is stored for the aforementioned processing. The detection items include one or more of contrast detection, exposure detection, light spot and bubble detection, etc.
[0104] For contrast detection, for example, the grayscale difference and change trend of pixels in an image are calculated, where the contrast calculation formula is shown in the following formula 1-4:
[0105]
[0106] Among them, C is the contrast, P Max is the maximum grayscale value of the image, P Min is the minimum grayscale value of the image.
[0107] The contrast C is then compared with a preset threshold range. If the contrast C is within the threshold range, the contrast of the image is considered to meet the requirements through contrast detection. If the contrast C exceeds the threshold range, for example, is greater than the maximum value of the threshold range or less than the minimum value of the threshold range, the image quality is confirmed to be unqualified.
[0108] For the exposure detection, for example, a grayscale histogram of the image is first established. In a specific embodiment, the grayscale histogram is used to divide the pixels into 26 grayscale levels according to the grayscale value, and the step length between each level is 10. Then the number of pixels exceeding the preset grayscale level is determined. When the number of pixels exceeding the preset grayscale level is greater than a threshold, or the number of pixels exceeding the preset grayscale level accounts for a proportion of the total number of pixels that exceeds a threshold (such as 1 / 5), the image is considered to be overexposed and unqualified. Otherwise, the image is confirmed to have passed the exposure detection and is qualified.
[0109] For light spot and bubble detection, in one embodiment, the step of performing light spot and bubble detection on the fluorescent image includes the following steps:
[0110] The fluorescence image is grayed out to obtain a single-channel grayscale image. For example, the grayscale value of the G channel of the original image is extracted as the grayscale value after graying to obtain the G channel grayscale image.
[0111] Perform adaptive threshold segmentation on the single-channel grayscale image to obtain the segmentation threshold. The specific process is as follows:
[0112] The grayscale values of pixels in the grayscale image are traversed to establish a grayscale histogram, and the number of pixels in each grayscale level is counted to obtain the probability of pixels being located at different pixel levels.
[0113] Then, the minimum to maximum grayscale values of the pixels in the grayscale image are traversed as the threshold range. All pixels with grayscale values higher than the current threshold will be recorded as category A, otherwise they will be recorded as category B.
[0114] Calculate the variance σ between category A and category B under different thresholds respectively 2 Variance σ 2 The calculation formula is shown in 1-2.
[0115] When the threshold t makes σ 2 When the value is the largest, it is considered that the difference between category A and category B is the most obvious. At this time, the threshold is positioned as the optimal segmentation threshold.
[0116] Then, the single-channel grayscale image is binarized based on the segmentation threshold to obtain an image mask. For example, the pixel points in the single-channel grayscale image whose grayscale value is higher than the optimal segmentation threshold are assigned a value of 0 (first grayscale value), and the pixel points whose grayscale value is lower than the threshold are assigned a value of 255 (second grayscale value). The image obtained after the assignment operation is recorded as the image mask. Among them, the pixel points with a grayscale value of 0 in the image mask are recorded as the foreground, and the pixel points with a grayscale value of 255 are recorded as the background.
[0117] Then, the image mask is judged. For example, the number and position of the pixels with a gray value of 0 in the image mask are counted, and when the pixels with a gray value of 0 are located in a continuous range in the image and the number is greater than or equal to a preset number threshold, it is confirmed that the light spot or bubble is detected.
[0118] If light spots or bubbles are detected, the current fluorescence image quality is considered unacceptable.
[0119] After the aforementioned quality inspection, the qualified images are stored for use in subsequent processing.
[0120] On the other hand, Fig.14 As shown, Fig.14 It is a principle block diagram of an EBV antibody determination system for auxiliary screening and diagnosis of nasopharyngeal carcinoma according to an embodiment of the present invention, wherein the EBV antibody determination system 10 includes an image acquisition module 11, a clustering module 12, a mask acquisition module 13, a target screening module 14 and a judgment module 15, wherein the image acquisition module 11 is used to obtain a fluorescent image of a test sample prepared based on an indirect immunofluorescence method, and the antigen of the test sample is VCA or EA. In one embodiment, the image acquisition module 11 can read the fluorescent image of a specified test sample from a storage device by data reading, or read the fluorescent image of a specified test sample from another device through a network, and can also be connected to a fluorescent microscope in real time to receive the fluorescent image of the test sample taken in real time by the fluorescent microscope.
[0121] The clustering module 12 is connected to the image acquisition module 11, and is used to perform cluster analysis on the pixels of the fluorescence image according to the grayscale difference and gradient difference of the pixels, and cluster the pixels of the fluorescence image into three clusters. Optionally, the pixels in each cluster can be further assigned the same grayscale value to obtain a corresponding cluster mask, wherein the grayscale values of the three cluster masks are different from each other. For the specific processing flow, please refer to the description of the aforementioned method part, which will not be repeated here.
[0122] The mask acquisition module 13 is connected to the clustering module 12, and is used to determine the three clusters as the first foreground target cluster, the second foreground target cluster and the third background cluster based on the number of pixels and grayscale mean of the clusters; the pixels constituting the first foreground target cluster are assigned the first grayscale value, and the remaining pixels are assigned the second grayscale value to obtain the first foreground target mask; optionally, the pixels constituting the second foreground target cluster can also be assigned the first grayscale value, and the remaining pixels can be assigned the second grayscale value to obtain the second foreground target mask. For the specific processing flow, please refer to the description of the aforementioned method part, which will not be repeated here.
[0123] The target screening module 14 is connected to the mask acquisition module 13, and is used to search for the target contour of the first-class foreground target mask to obtain a target image, wherein a plurality of target images constitute a candidate first-class foreground target set, and interference targets are determined and screened out from a plurality of targets in the candidate first-class foreground target set based on grayscale values and target contours to obtain a first-class foreground target set. For the specific processing flow, please refer to the description of the aforementioned method part, which will not be repeated here.
[0124] The judgment module 15 is connected to the target screening module 14, and is used to count the number of targets in the first category foreground target set, and determine that the antibody of the test sample is positive when the target number is greater than or equal to a first number threshold, and determine that the antibody of the test sample is negative when the target number is less than the first number threshold.
[0125] In another embodiment, the EBV antibody determination system 10 further includes a titer grade module 16, which is connected to the judgment module 15 and is used to classify the titer grade of the fluorescent image of the test sample determined to be positive to determine the titer grade of the test sample.
[0126] In another embodiment, the EBV antibody determination system 10 further includes a quality inspection module 17, which is respectively connected to the image acquisition module 11 and the clustering module 12, and is used to detect the fluorescent image acquired by the image acquisition module against preset detection items. When the fluorescent image meets the quality conditions, the clustering module 12 performs cluster analysis on the pixels of the fluorescent image.
[0127] In addition, in order to facilitate human-computer interaction, the EBV antibody assay system 10 includes an interactive interface and a database, through which the staff can issue image reading instructions or shooting instructions, and can view the processing data of the EBV antibody assay system 10, such as various intermediate data and final result data. After processing a fluorescent image of a test sample, the EBV antibody assay system 10 stores its intermediate processing data and results in the database for preservation.
[0128] Among them, the method provided by the present invention has the same processing flow for the test samples with VCA as the antigen and the test samples with EA as the antigen, but due to the fact that there are certain batch differences in the overall fluorescence intensity of the two different antigen matrices, the difference threshold (hereinafter referred to as the fluorescence threshold) used to represent the grayscale mean of the fluorescence difference threshold is not the same. When the fluorescence threshold is higher, the judgment condition of positive cells is more stringent, thereby reducing the positive results of the fluorescence image. In order to obtain the optimal fluorescence threshold, the present invention establishes corresponding evaluation data sets for the processing flow of the two matrices, for example, the evaluation data set of the VCA matrix is recorded as data set 1, and the evaluation data set of the EA matrix is recorded as data set 2. In order to test whether the two optimal fluorescence thresholds obtained are generalizable and whether they can adapt to new data, a test data set of EA matrix and VCA matrix pairing is established, which is recorded as data set 3. There is no duplicate data between each data set.
[0129] When the fluorescence threshold is set to 60-80, the ROC curve obtained by testing the method flow of the present invention through data set 1 is as follows: Fig.15 As shown, the ROC curve obtained by testing the method flow of the present invention through data set 2 is as follows Fig.16 As shown. The abscissa of the ROC curve is the false positive rate (FPR), and the ordinate is the true positive rate (TPR). In the ROC curve after the evaluation of the data set 1 of the VCA matrix, the AUC value can reach 0.992. In the ROC curve after the evaluation of the data set 2 of the EA matrix, the AUC value can reach 0.981, which shows that the performance of the method provided by the present invention is good.
[0130] The results of the performance evaluation on Dataset 3 based on the best threshold obtained by the evaluation are shown in Table 1:
[0131] Table 1: Performance test results of dataset 3
[0132]
[0133] Among them, TP (True Positive) represents the positive samples predicted to be positive; FN (False Negative) represents the positive samples predicted to be negative; FP (False Positive) represents the negative samples predicted to be positive; TN (True Negative) represents the negative samples predicted to be negative; Sensitivity represents the sensitivity of judgment, and its calculation method is: Sensitivity = TP / (TP+FN); Specificity represents the specificity of judging foreground cells, and its calculation method is: Specificity = TN / (TN+FP); PPV represents the positive accuracy rate, and its calculation method is: PPV = TP / (TP+FP); NPV represents the negative accuracy rate, and its calculation method is: NPV = TN / (TN+FN).
[0134] As shown in the evaluation results, after the optimal threshold is obtained based on data set 1 and data set 2, the final evaluation result on data set 3 has excellent performance. When the grayscale mean threshold used in the method and system provided by the present invention is set as an adjustable parameter, the user can adjust it through the interactive interface to meet the different requirements of different people for detection strictness.
[0135] In addition, after testing, the accuracy of the classification performance of titer levels was 95.9% for the VCA matrix and 93.5% for the EA matrix, and there were no more than two misclassified titer levels in the classification results.
[0136] Fig.17 6 is a schematic diagram of the hardware structure principle of an electronic device according to an embodiment of the present invention, the electronic device can be implemented as a server or other terminal devices, such as a desktop personal computer, a tablet computer, a laptop computer, etc., including a processor 601 and a memory 602, the memory 602 stores a program instruction set, and when the processor 601 executes the program instruction set on the memory 602, any of the aforementioned EBV antibody determination methods for nasopharyngeal carcinoma auxiliary screening and diagnosis is implemented. The electronic device in this embodiment is, for example, Figure 1 The computing device 101 in FIG.
[0137] Specifically, the processor 601 may include a central processing unit (CPU), or an application specific integrated circuit (ASIC), or may be configured to implement one or more integrated circuits of the embodiment of the present invention.
[0138] The memory 602 may include a large capacity memory for data or instructions. By way of example and not limitation, the memory 602 may include a hard disk drive (HDD), a floppy disk drive, a flash memory, an optical disk, a magneto-optical disk, a magnetic tape, or a universal serial bus (USB) drive or a combination of two or more of these. In appropriate cases, the memory 602 may include a removable or non-removable (or fixed) medium. In appropriate cases, the memory 602 may be inside or outside the integrated gateway disaster recovery device. In a specific embodiment, the memory 602 is a non-volatile solid-state memory.
[0139] The memory may include a read-only memory (ROM), a random access memory (RAM), a disk storage medium device, an optical storage medium device, a flash memory device, an electrical, optical or other physical / tangible memory storage device. Therefore, generally, the memory includes one or more tangible (non-transitory) computer-readable storage media (e.g., a memory device) encoded with software including computer executable instructions, and when the software is executed (e.g., by one or more processors), it is operable to perform the EBV antibody determination method for nasopharyngeal carcinoma auxiliary screening and diagnosis provided by the present invention.
[0140] In one example, the electronic device may further include a communication interface 603 and a bus 604. The processor 601, the memory 602, and the communication interface 603 are connected via the bus 604 and communicate with each other.
[0141] The communication interface 603 is mainly used to implement communication between various modules, systems, units and / or devices in the embodiment of the present invention.
[0142] Bus 604 includes hardware, software or both, and the components of online data flow billing equipment are coupled to each other. For example, but not limitation, bus may include accelerated graphics port (AGP) or other graphics bus, enhanced industrial standard architecture (EISA) bus, front side bus (FSB), hypertransport (HT) interconnection, industrial standard architecture (ISA) bus, infinite bandwidth interconnection, low pin count (LPC) bus, memory bus, micro channel architecture (MCA) bus, peripheral component interconnection (PCI) bus, PCI-Express (PCI-X) bus, serial advanced technology attachment (SATA) bus, video electronics standard association local (VLB) bus or other suitable bus or two or more of these combinations. In appropriate cases, bus 604 may include one or more buses. Although the embodiment of the present invention describes and shows a specific bus, the present invention considers any suitable bus or interconnection.
[0143] The present invention also provides a computer-readable storage medium having computer program instructions stored thereon, and the computer program instructions can be executed by a processor to implement any of the EBV antibody determination methods for nasopharyngeal carcinoma auxiliary screening and diagnosis in the aforementioned embodiments. The computer-readable storage medium can be any medium that can tangibly contain or store computer-executable instructions for use by or in combination with an instruction execution system, device, and equipment. The storage medium can be a transient computer-readable storage medium or a non-transient computer-readable storage medium. Non-transient computer-readable storage media may include, but are not limited to, magnetic storage devices, optical storage devices, and / or semiconductor storage devices. Embodiments corresponding to such storage devices include, for example, disks, optical disks based on CD, DVD, or Blu-ray technology, and persistent solid-state memories such as flash memory, solid-state drives, and the like.
[0144] The present invention also provides a computer program product, which includes a set of computer program instructions, and when the computer program instruction set is executed by a processor, any of the EBV antibody determination methods for nasopharyngeal carcinoma auxiliary screening and diagnosis in the aforementioned embodiments is implemented. The computer program product includes but is not limited to application installation packages published on websites and application stores, application plug-ins, and small programs that can be run in certain applications.
[0145] It should be clear that the present invention is not limited to the specific configuration and processing described above and shown in the figures. For the sake of simplicity, a detailed description of the known method is omitted here. In the above embodiments, several specific steps are described and shown as examples. However, the method process of the present invention is not limited to the specific steps described and shown, and those skilled in the art can make various changes, modifications and additions, or change the order between the steps after understanding the spirit of the present invention.
[0146] The above embodiments are only used to illustrate the present invention, but not to limit the present invention. Ordinary technicians in the relevant technical field can make various changes and modifications without departing from the scope of the present invention. Therefore, all equivalent technical solutions should also fall within the scope of the present invention.
Claims
1. A method for determining EBV antibodies for auxiliary screening and diagnosis of nasopharyngeal carcinoma, comprising: Cluster analysis is performed on the pixels of the fluorescent image of the test sample prepared based on the indirect immunofluorescence method according to the grayscale difference and gradient difference of the pixels, and the pixels of the fluorescent image are clustered into three clusters, wherein the antigen of the test sample is VCA or EA; The three clusters are respectively determined as a first foreground target cluster, a second foreground target cluster and a third background cluster based on the number of pixels and grayscale mean of the clusters; The pixel points constituting the first type of foreground target cluster are assigned a first gray value, and the remaining pixel points are assigned a second gray value to obtain a first type of foreground target mask; Performing contour search on the target in the first-category foreground target mask to obtain a target image, wherein the plurality of target images constitute a candidate first-category foreground target set; Determining interference targets from a plurality of targets in the candidate first-category foreground target set based on grayscale values and target contours respectively and screening them out to obtain a first-category foreground target set; and The number of targets in the first type of foreground target set is counted, and when the number of targets is greater than or equal to a first number threshold, the antibody of the test sample is determined to be positive, and when the number of targets is less than the first number threshold, the antibody of the test sample is determined to be negative.
2. The EBV antibody determination method for auxiliary screening and diagnosis of nasopharyngeal carcinoma according to claim 1, wherein: Before clustering analysis is performed on the pixels of the fluorescent image, the method further includes: detecting the fluorescent image according to preset detection items, and performing cluster analysis on the pixels of the fluorescent image when the fluorescent image meets the quality condition.
3. The EBV antibody determination method for auxiliary screening and diagnosis of nasopharyngeal carcinoma according to claim 2, wherein: The detection items include one or more of contrast detection, exposure level detection, light spot and bubble detection.
4. The EBV antibody determination method for auxiliary screening and diagnosis of nasopharyngeal carcinoma according to claim 3, wherein: The step of performing light spot and bubble detection on the fluorescent image includes: Performing single-channel grayscale conversion on the fluorescence image to obtain a single-channel grayscale image; Performing adaptive threshold segmentation processing on the single-channel grayscale image to obtain a segmentation threshold; Binarizing the single-channel grayscale image based on the segmentation threshold to obtain an image mask, wherein the grayscale values in the image mask are respectively the first grayscale value and the second grayscale value after the binarization, wherein the first grayscale value is less than the second grayscale value; Counting the number and positions of pixels whose grayscale value is the first grayscale value in the image mask; and When the pixel points of the first gray value are located in a continuous range in the image and the number of the pixel points is greater than or equal to the second number threshold, it is confirmed that the light spot or the bubble is detected.
5. The EBV antibody determination method for auxiliary screening and diagnosis of nasopharyngeal carcinoma according to claim 1, wherein: Before or after clustering analysis of the pixels of the fluorescence image, the following steps are also included: Performing single-channel grayscale conversion on the fluorescence image to obtain a single-channel grayscale image; and The single-channel grayscale image is filtered to remove noise in the image.
6. The EBV antibody determination method for auxiliary screening and diagnosis of nasopharyngeal carcinoma according to claim 5, wherein: The step of determining interference targets from a plurality of targets in the candidate first-category foreground target set based on grayscale values comprises: Obtain the grayscale values of all pixels in the second category foreground target cluster from the filtered single-channel grayscale image, and calculate the grayscale mean of all pixels as the grayscale mean of the second category foreground target; Calculating the minimum bounding rectangle of each target image in the candidate first-category foreground target set; According to the minimum circumscribed rectangle of each target image, the corresponding target image is intercepted on the filtered single-channel grayscale image as the target image to be processed; After performing adaptive threshold segmentation and binarization processing on the target image to be processed, a segmentation mask of the target to be processed is obtained, wherein the grayscale values in the segmentation mask of the target to be processed are respectively the first grayscale value and the second grayscale value after the binarization processing, wherein the first grayscale value is less than the second grayscale value; By comparing all the pixels whose grayscale values are the first grayscale value in the segmentation mask of the target to be processed, the grayscale values of the corresponding pixels are obtained from the target image to be processed, and the grayscale mean value of the target to be processed is calculated; Compare the difference between the grayscale mean of the target to be processed and the grayscale mean of the second type of foreground target; and When the difference is less than the grayscale mean threshold, the target to be processed is determined to be an interference target.
7. The EBV antibody determination method for nasopharyngeal carcinoma auxiliary screening and diagnosis according to claim 1 or 6, wherein the step of determining the interference target from the multiple targets in the candidate first-category foreground target set based on the target profile comprises: Perform ellipse fitting on the contour of each target to obtain the corresponding ellipse and its major axis and minor axis; Calculate the ratio of the major axis to the minor axis of the ellipse for each target; Calculate the area corresponding to the contour of each target; as well as When the ratio of the major axis to the minor axis of the ellipse of a target is greater than or equal to the axis ratio threshold and the area is less than the area threshold, the target is determined to be an interference target.
8. The EBV antibody assay method for nasopharyngeal carcinoma auxiliary screening and diagnosis according to claim 1, wherein: The steps of respectively determining the three clusters as a first type of foreground target cluster, a second type of foreground target cluster and a third type of background cluster based on the number of pixels and grayscale mean of the clusters include: Count the number of pixels in each cluster respectively; Calculate the grayscale mean of each cluster; Determine the cluster with the least number of pixels as the first type of foreground target cluster; and Among the remaining two clusters, the cluster with fewer pixels and larger grayscale mean is determined as the second type of foreground target cluster, and the remaining cluster is the third type of background cluster.
9. The EBV antibody determination method for auxiliary screening and diagnosis of nasopharyngeal carcinoma according to claim 8, wherein: When determining to obtain the first category of foreground target clusters, a first label is set for each pixel constituting the first category of foreground target clusters; correspondingly, the pixel points in the first category of foreground target clusters to which the first label is set are assigned a first grayscale value, and the remaining pixel points are assigned a second grayscale value to obtain a first category of foreground target mask, wherein the first grayscale value is less than the second grayscale value.
10. The EBV antibody determination method for auxiliary screening and diagnosis of nasopharyngeal carcinoma according to claim 1, further comprising: The pixel points in each cluster are assigned the same grayscale value to obtain a cluster mask, wherein the grayscale values of the pixel points in each cluster in the cluster mask are different from each other.
11. The EBV antibody assay method for auxiliary screening and diagnosis of nasopharyngeal carcinoma according to claim 1, further comprising: The pixels constituting the second type of foreground object cluster are assigned a first gray value, and the remaining pixels are assigned a second gray value to obtain a second type of foreground object mask, wherein the first gray value is less than the second gray value. 12 . The EBV antibody determination method for auxiliary screening and diagnosis of nasopharyngeal carcinoma according to claim 1 , 4 , 6 , 9 , or 11 , wherein the first grayscale value is 0 and the second grayscale value is 255.
13. The EBV antibody determination method for auxiliary screening and diagnosis of nasopharyngeal carcinoma according to claim 1, further comprising: The fluorescent images of the antibody-positive test samples are classified by titer level to determine the titer level of the test samples.
14. The EBV antibody determination method for nasopharyngeal carcinoma auxiliary screening and diagnosis according to claim 13, when the titer grade classification is performed on the fluorescent image of the antibody-positive test sample, the fluorescent image of the test sample is input into the trained classification model, and the titer grade of the test sample is obtained after being processed by the classification model, wherein: The classification model is trained to achieve titer level classification of a preset number of levels, the input of the classification model is a fluorescence image of the detection sample, and the output is a titer level in the preset number of levels.
15. An EBV antibody determination system for auxiliary screening and diagnosis of nasopharyngeal carcinoma, comprising: A clustering module is configured to perform cluster analysis on the pixels of the fluorescent image of the test sample prepared based on the indirect immunofluorescence method according to the grayscale difference and gradient difference of the pixels, and cluster the pixels of the fluorescent image into three clusters, wherein the antigen of the test sample is VCA or EA; A mask acquisition module, which is connected to the clustering module and is configured to determine the three clusters as a first-category foreground target cluster, a second-category foreground target cluster, and a third-category background cluster based on the number of pixels and grayscale mean values of the clusters; assigning pixels constituting the first-category foreground target cluster to a first grayscale value and assigning remaining pixels to a second grayscale value to obtain a first-category foreground target mask; a target screening module, connected to the mask acquisition module, configured to perform contour search on the target in the first-category foreground target mask to obtain a target image, wherein a plurality of target images constitute a candidate first-category foreground target set, and to determine interference targets from a plurality of targets in the candidate first-category foreground target set based on grayscale values and target contours and screen out interference targets to obtain a first-category foreground target set; and A judgment module is connected to the target screening module and is configured to count the number of targets in the first category of foreground target set, and determine that the antibody of the test sample is positive when the number of targets is greater than or equal to a first number threshold, and determine that the antibody of the test sample is negative when the number of targets is less than the first number threshold.
16. The EBV antibody determination system for auxiliary screening and diagnosis of nasopharyngeal carcinoma according to claim 15 further comprises a titer grade module, which is connected to the judgment module and is configured to classify the titer grade of the fluorescent image of the test sample determined to be positive to determine the titer grade of the test sample.
17. The EBV antibody determination system for auxiliary screening and diagnosis of nasopharyngeal carcinoma according to claim 15, further comprising a quality inspection module, which is connected to the clustering module and is configured to detect the fluorescent image to be processed against preset detection items. When the fluorescent image meets the quality conditions, the clustering module performs cluster analysis on the pixels of the fluorescent image.
18. An electronic device comprising a processor and a memory, wherein the memory stores a computer program instruction set, and when the processor executes the computer program instruction set on the memory, the EBV antibody determination method for auxiliary screening and diagnosis of nasopharyngeal carcinoma according to any one of claims 1 to 14 is implemented.
19. A computer-readable storage medium, wherein: The computer-readable storage medium stores a computer program instruction set, and when the computer program instruction set is executed by a processor, the EBV antibody determination method for auxiliary screening and diagnosis of nasopharyngeal carcinoma according to any one of claims 1 to 14 is implemented.
20. A computer program product comprising a computer program instruction set, wherein when the computer program instruction set is executed by a processor, the method for determining EBV antibodies for auxiliary screening and diagnosis of nasopharyngeal carcinoma according to any one of claims 1 to 14 is implemented.
Citation Information
Cited By
Deformity area identification method for paranasal sinus image under nasal endoscope
CN119831982A
A method for identifying deformed regions of nasal sinus images under nasal endoscopy
CN119831982B
