A method, system and storage medium for turbidity evaluation and processing of blood separation plasma samples

Abstract

The present invention discloses a turbidity evaluation processing method, system and storage medium for blood separation plasma samples. First, a first image containing target cells is processed to obtain a second image, and a third image containing only the target cells is cropped from the second image, thereby reducing the error caused by subsequent identification of the attribute information of the target cells due to factors such as the surrounding environment of the target cells. Furthermore, the third image is input into an image recognition model trained through operations such as convolutional layers, pooling layers, and activation functions. After detecting and recognizing the five-level association feature information of the target cells, the association recognition is performed with a preset control feature information table, and then the attribute information of the target cells is output. By ensuring the recognition accuracy of the sample image, the present invention realizes the turbidity evaluation monitoring and processing of plasma samples.

Description

Technical Field

[0001] The present invention relates to the field of image recognition, and particularly to a turbidity evaluation processing method, system and storage medium for a blood separation plasma sample. Background Art

[0002] A blood component separator (usually referred to as a blood separator or blood component separation device) is a medical device used to separate whole blood into different components. Its main purpose is to separate blood into multiple useful components, such as red blood cells, plasma, platelets and other blood derivatives, according to the physical and chemical properties of each component in the blood through specific separation techniques. These separated components are widely used in clinical treatment, research and the pharmaceutical industry and other fields.

[0003] During the operation of a blood component separator, turbidity is an important reference index; turbidity refers to the degree of light scattering and absorption by the tiny particulate matter suspended in the blood component separation liquid, and is usually used to measure the clarity or transparency of the blood component separation liquid. The higher the turbidity, the more turbid the liquid of the blood component separation liquid looks. When using a blood component separator for blood separation, the turbidity of the plasma sample may be affected by the following reasons:

[0004] For example: Reason for fat emulsion (suspension of fat particles): There are more lipid substances in the plasma of hyperlipidemic patients, or fat particles are not completely removed during blood collection and separation. The suspension of fat particles will increase the turbidity of the plasma, making the sample look milky white. At the same time, there may be some parasites or microorganisms in the blood sample: for example, certain bacteria (such as streptococcus), spirochetes (such as treponema pallidum) or plasmodium. For example: Reasons for cell debris and hemolysis: During the blood separation process, if the operation is improper (such as excessive centrifugal force or too long time), it may cause red blood cells, white blood cells, etc. to rupture and release intracellular components. Cell debris and the released cellular substances make the plasma turbid and affect the sample quality.

[0005] With the development of high-definition electron microscope technology, this technical means can not only apply image processing technology to identify various types of image features, but also further identify the attribute information of the image; however, further research finds that when sampling high-definition images of samples, due to the influence of the fat particle suspensions existing around the target cells, it will further lead to recognition errors; for example, when taking a photo to recognize the target cells in an image, there will be fat particle suspensions around the target cells, and even there will be various textures on the surface of the fat particle suspensions, resulting in recognition results such as recognizing the textures on the fat particle suspensions, thus mistakenly thinking that they are target cells (i.e., red blood cells and white blood cells), which is very different from the real recognition purpose;

[0006] Therefore, how to ensure the accurate identification of the types and quantities of interfering substances (i.e., malaria parasites and fat particle suspensions) in the blood component separation fluid is an important basis for the turbidity evaluation of plasma samples. Summary of the Invention

[0007] The purpose of the present invention is to provide a turbidity evaluation processing method, system, and storage medium for blood-separated plasma samples, which solves the above-mentioned technical problems pointed out in the prior art.

[0008] A turbidity evaluation processing method for blood-separated plasma samples includes the following operating steps:

[0009] Obtain a first image containing target cells;

[0010] Perform recognition processing on the first image to obtain a second image in the first image;

[0011] The second image is a target cell region image;

[0012] After performing recognition and analysis on the second image through an edge detection algorithm, intercept to obtain a third image;

[0013] The third image is the target cell image;

[0014] Input the third image into a pre-trained image recognition model to determine the attribute information of the target cells;

[0015] The attribute information of the target cells includes the name information of the target cells;

[0016] Determine the turbidity rating information of the sample according to the number of target cells recognized per unit area and output the turbidity rating information.

[0017] Preferably, the inputting the third image into a pre-trained image recognition model to determine the attribute information of the target cells includes:

[0018] Use the image recognition model to perform initial recognition on the third image, and the initial recognition obtains the five-level associated feature information of the current third image;

[0019] The five-level associated feature information includes color feature information, shape feature information, structural feature information, local feature information, and global feature information;

[0020] Calculate the similarity between each of the five-level associated feature information and all the control feature information in a preset feature information comparison table;

[0021] Determine that the control attribute information corresponding to the highest similarity is the attribute information of the target cells.

[0022] Preferably, the initial recognition of the third image using the image recognition model obtains the five - level associated feature information of the current third image, including the following operation steps:

[0023] Establish an initial image recognition model;

[0024] After inputting the third image into the trained image recognition model obtained by training based on the initial image recognition model, the five - level associated feature information corresponding to the third image is output.

[0025] Preferably, the step of inputting the third image into the trained image recognition model obtained by training based on the initial image recognition model and outputting the five - level associated feature information corresponding to the third image includes the following operation steps:

[0026] Obtain a training set, where the training set includes multiple training data and the annotation data corresponding to each training data; the training data are training images; the annotation data are the attribute information of the target cells corresponding to the training data for annotating the training data;

[0027] Input the training data into the pre - established initial image recognition model, and an image recognition vector is output;

[0028] Based on the image recognition vector and the annotation data, calculate the loss value through a loss function;

[0029] Back - propagate the loss value to update the parameters of the initial image recognition model, and perform iterative training through a preset maximum number of iterations and a maximum loss value threshold to obtain a trained image recognition model;

[0030] Input the third image into the trained image recognition model, and the five - level associated feature information corresponding to the third image is output.

[0031] Preferably, the step of inputting the third image into the trained image recognition model and outputting the five - level associated feature information corresponding to the third image includes the following operation steps:

[0032] Perform forward propagation on the third image through the trained image recognition model, and an image vector is output;

[0033] Decode the image vector to obtain the five - level associated feature information.

[0034] Preferably, the step of performing forward propagation on the third image through the trained image recognition model and outputting an image vector includes the following operation steps:

[0035] Obtain image and model parameter data; in the initial state, set p convolution kernels of size a×b; based on the convolution kernels, the image and the model parameter data, perform a convolution operation on all pixel points in the third image to obtain the signal value corresponding to each pixel point;

[0036] The image and model parameter data include all pixel points in the third image, the coordinates (x, y) corresponding to all pixel points, the intensity value t corresponding to the pixel points i and the weight value corresponding to the trained image recognition model; p is greater than or equal to 1;

[0037] The signal value corresponding to the pixel point is expressed as:

[0038]

[0039] Perform a non-linear mapping operation on the signal value through an activation function to obtain the activation signal corresponding to the signal value;

[0040] Perform downsampling on the activation signal through a pooling layer to obtain a target activation signal;

[0041] Convert the target activation signal through a fully connected layer to obtain an image vector.

[0042] Preferably, calculating the similarity between the five-level associated feature information and all reference feature information in the preset feature information comparison table includes the following operation steps:

[0043] Set header data for the five-level associated feature information;

[0044] The header data includes color feature header data O, shape feature header data P, structure feature header data Q, local feature header data R, and global feature header data S;

[0045] Traverse all reference feature information in the feature information comparison table to obtain the reference header data in the reference feature information;

[0046] The reference header data includes reference color feature header data O', reference shape feature header data P', reference structure feature header data Q', reference local feature header data R', and reference global feature header data S';

[0047] Establish an association relationship between the reference header data and the header data to obtain a reference association matrix D;

[0048] The reference association matrix is expressed as:

[0049]

[0050] Calculate and obtain the similarity of each factor in the control association matrix.

[0051] The present invention also provides a turbidity evaluation processing system for a blood separation plasma sample, including a first image acquisition module, a second image acquisition module, a third image acquisition module, and an image recognition module;

[0052] Among them, the first image acquisition module is used to acquire a first image containing target cells;

[0053] The second image acquisition module is used to perform recognition processing on the first image to obtain a second image in the first image;

[0054] The second image is a target cell region image;

[0055] The third image acquisition module is used to intercept a third image after performing recognition and analysis on the second image through an edge detection algorithm;

[0056] The third image is the target cell image;

[0057] The image recognition module is used to input the third image into a pre-trained image recognition model to determine the attribute information of the target cells;

[0058] The attribute information of the target cells includes the name information of the target cells.

[0059] Preferably, the image recognition module includes an initial recognition sub-module, a calculation sub-module, and a screening and acquisition sub-module;

[0060] The initial recognition sub-module is used to perform initial recognition on the third image by using the image recognition model, and initially recognize and obtain the five-level association feature information of the current third image;

[0061] The five-level association feature information includes color feature information, shape feature information, structural feature information, local feature information, and global feature information;

[0062] The calculation sub-module is used to calculate and obtain the similarities of the five-level association feature information with all control feature information in a preset feature information comparison table;

[0063] The screening and acquisition sub-module is used to determine that the control attribute information corresponding to the highest similarity is the attribute information of the target cells.

[0064] Preferably, the initial recognition sub-module is specifically used to establish an initial image recognition model;

[0065] After inputting the third image into the trained image recognition model obtained by training based on the initial image recognition model, the five-level associated feature information corresponding to the third image is output.

[0066] Preferably, in specific implementation, the initial recognition sub-module is used to obtain a training set, which includes a plurality of training data and the annotation data corresponding to each training data; the training data is a training image; the annotation data is the attribute information corresponding to the target cells in the training data for annotation training.

[0067] Input the training data into the pre-established initial image recognition model, and output an image recognition vector.

[0068] Based on the image recognition vector and the annotation data, calculate the loss value through a loss function.

[0069] Backpropagate the loss value to update the parameters of the initial image recognition model, and perform iterative training through a preset maximum number of iterations and a maximum threshold of the loss value to obtain a trained image recognition model.

[0070] Input the third image into the trained image recognition model, and output the five-level associated feature information corresponding to the third image.

[0071] Preferably, in specific implementation, the initial recognition sub-module is further used to perform forward propagation on the third image through the trained image recognition model, and output an image vector.

[0072] Perform a decoding operation on the image vector to obtain five-level associated feature information.

[0073] Preferably, in specific implementation, the initial recognition sub-module is further used to obtain an image and model parameter data; in the initial state, p convolution kernels of size a×b are set; based on the convolution kernels, the image, and the model parameter data, perform a convolution operation on all pixel points in the third image to obtain the signal value corresponding to each pixel point.

[0074] The image and model parameter data includes all pixel points in the third image, the coordinates (x, y) corresponding to all pixel points, the intensity value t corresponding to the pixel points i and the weight value corresponding to the trained image recognition model; p is greater than or equal to 1.

[0075] The signal value corresponding to the pixel point is expressed as:

[0076]

[0077] Perform a non - linear mapping operation on the signal value through an activation function to obtain an activation signal corresponding to the signal value;

[0078] Perform downsampling on the activation signal through a pooling layer to obtain a target activation signal;

[0079] Convert the target activation signal through a fully - connected layer to obtain an image vector;

[0080] Preferably, the calculation sub - module is specifically used to set header data for the five - level associated feature information;

[0081] The header data includes color feature header data O, shape feature header data P, structure feature header data Q, local feature header data R, and global feature header data S;

[0082] Traverse all the control feature information in the feature information comparison table to obtain the control header data in the control feature information;

[0083] The control header data includes control color feature header data O', control shape feature header data P', control structure feature header data Q', control local feature header data R', and control global feature header data S';

[0084] Establish an association relationship between the control header data and the header data to obtain a control association matrix D;

[0085] The control association matrix is expressed as:

[0086]

[0087] Calculate and obtain the similarity of each factor in the control association matrix;

[0088] The present invention also provides a storage medium, in which a computer program is stored. When the computer program is executed by a processor, the steps of the above - mentioned turbidity evaluation processing method for a blood - separated plasma sample are implemented.

[0089] Compared with the prior art, the embodiments of the present invention have at least the following technical advantages:

[0090] Analyzing the above-mentioned turbidity evaluation processing method, system and storage medium for blood separation plasma samples provided by the present invention, it can be seen that in specific applications, first, the first image containing target cells is processed to obtain a second image, and a third image containing only target cells is cropped from the second image, which can reduce errors caused by factors such as the surrounding environment of the target cells in subsequent identification of the attribute information of the target cells; then, the third image is input into the trained image recognition model to detect and recognize the feature information of the third image. Through the image recognition model, the input image can be converted into corresponding feature vectors or feature descriptors, and these feature information can accurately represent the content and features of the image;

[0091] By comparing the feature information of the third image with the preset feature information comparison table, and calculating the similarity between the obtained feature information and all the comparison feature information in the preset feature information comparison table, the feature information most similar to the input image can be quickly found; in this way, feature matching and image recognition analysis tasks can be effectively carried out, calculating the similarity between the feature information of the third image and the comparison feature information in the feature information comparison table, so as to determine the comparison feature information with the highest similarity as the target feature information, and then determine the comparison attribute information corresponding to the target feature information as the attribute information of the third image, that is, the attribute information of the target cells. By determining the comparison feature information with the highest similarity, the feature information corresponding to the target image can be accurately found, and the comparison attribute information corresponding to the comparison feature information can be used to describe the attribute characteristics of the target cells, such as color, shape, texture, etc.; finally, the target cells in the image can be quickly and effectively recognized, and misidentification caused by other factors can be prevented. BRIEF DESCRIPTION OF THE DRAWINGS

[0092] In order to more clearly illustrate the specific embodiments of the present invention or the technical solutions in the prior art, the following will briefly introduce the drawings required for use in the description of the specific embodiments or the prior art. Obviously, the following drawings are some embodiments of the present invention. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts.

[0093] Figure 1 It is a schematic diagram of the overall operation steps of the turbidity evaluation processing method for blood separation plasma samples provided in Embodiment 1 of the present invention;

[0094] Figure 2 It is a schematic diagram of the operation steps for determining the attribute information of target cells in the turbidity evaluation processing method for blood separation plasma samples provided in Embodiment 1 of the present invention;

[0095] Figure 3Schematic diagram of the operation steps for obtaining five - level associated feature information of a turbidity evaluation processing method for a blood - separated plasma sample provided in the first embodiment of the present invention;

[0096] Figure 4 Schematic diagram of the more detailed operation steps for obtaining five - level associated feature information of a turbidity evaluation processing method for a blood - separated plasma sample provided in the first embodiment of the present invention;

[0097] Figure 5 Schematic diagram for further explaining the operation steps for obtaining five - level associated feature information of a turbidity evaluation processing method for a blood - separated plasma sample provided in the first embodiment of the present invention;

[0098] Figure 6 Schematic diagram of the operation steps for obtaining an image vector of a turbidity evaluation processing method for a blood - separated plasma sample provided in the first embodiment of the present invention;

[0099] Figure 7 Schematic diagram of the operation steps for calculating the similarity between five - level associated feature information and control feature information of a turbidity evaluation processing method for a blood - separated plasma sample provided in the first embodiment of the present invention;

[0100] Figure 8 Schematic diagram of the overall architecture of a turbidity evaluation processing system for a blood - separated plasma sample provided in the second embodiment of the present invention.

[0101] Reference numerals: First image acquisition module 10; Second image acquisition module 20; Third image acquisition module 30; Image recognition module 40; Initial recognition sub - module 41; Calculation sub - module 42; Screening and acquisition sub - module 43. Detailed implementation manners

[0102] Next, the technical solutions of the present invention will be described clearly and completely with reference to the accompanying drawings. Obviously, the described embodiments are some, but not all, of the embodiments of the present invention. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0103] Next, the present invention will be further described in detail through specific embodiments in combination with the accompanying drawings.

[0104] Embodiment 1

[0105] As Figure 1 shown, correspondingly, the present invention also proposes a turbidity evaluation processing method for a blood - separated plasma sample, including the following operation steps:

[0106] Step S10: Obtain a first image containing target cells;

[0107] The first image is an initial image, which is a plurality of initial images for identifying target cells in the image, obtained by photographing with a camera or manually collecting;

[0108] Step S20: Perform recognition processing on the first image to obtain a second image in the first image;

[0109] The second image is a target cell region image;

[0110] It should be noted that in the above embodiments of the present application, the first image can be recognized and processed by a YOLO model or an SSD model, so as to frame the image of the area where the target cells are located in the first image (i.e., the above-mentioned second image), and screen out the areas outside the area where the target cells are located in the first image, only retaining the target cell region that is of great significance for the recognition of target cells, reducing the amount of data processed by the subsequent server, and thus improving the data processing efficiency.

[0111] Step S30: Perform recognition and analysis on the second image through an edge detection algorithm to obtain the edge information of the target cell; intercept a third image according to the edge information;

[0112] The third image is the target cell image;

[0113] It should be noted that the above-mentioned recognition of the target cell region through the edge detection algorithm to obtain the edge information of the target cell, and then the target cell image can be intercepted according to the edge information in the subsequent steps, which can reduce the recognition error caused by the influence of the pixel points of the image outside the target cell image when recognizing the attribute information of the target cell in the subsequent steps; the above-mentioned edge detection algorithms include Canny edge detection (detecting the edges in the image through Gaussian filtering, calculating gradients, non-maximum suppression, and double-threshold processing), Sobel edge detection (calculating the gradients in the horizontal and vertical directions of the image, and then merging the two gradient results to obtain an edge image), Laplacian edge detection (detecting the edges in the image by calculating the Laplacian operator of the image), etc., which can realize the operation of automatic recognition and detection. The above-mentioned edge detection algorithms are all prior arts, and the embodiments of the present application will not elaborate;

[0114] The above-mentioned edge information refers to the contour of the outermost part of the target cell in the target cell region; according to this contour, an image containing only the target cell can be intercepted from the target region image.

[0115] Step S40: Input the third image into a pre-trained image recognition model to determine the attribute information of the target cell;

[0116] The attribute information of the target cells includes the name information and type information of the target cells; the above-mentioned target cells can be red blood cells or white blood cells, etc. However, what affects the identification of the target cells in the blood component separation solution is the interference substances in the collected images; for example, the interference substances in the blood component separation solution include parasites or microorganisms such as malaria parasites and fat granule suspensions;

[0117] During specific operations, the separated blood components are made into smears and stained (such as Giemsa staining). Whether there are cells or parasites in the shape of long strips is checked through a high-precision electron image microscope. It should be noted that in the technical solution adopted in the embodiment of the present application, the target cell image can be directly intercepted from the target cell area through the edge information, so as to obtain an image containing only the target cells, ensuring that when using the image recognition model to identify and analyze the attribute information of the target cells subsequently, it will not be affected by other factors (these factors will include the pixel points remaining in the target cell area that do not belong to the target cell image), thereby ensuring the improvement of the recognition accuracy;

[0118] In the embodiment of the present application, first, the first image containing the target cells is processed to obtain a second image, and a third image containing only the target cells is intercepted from the second image, which can reduce the error caused by factors such as the surrounding environment of the target cells when subsequently identifying the attribute information of the target cells; then the third image is input into the trained image recognition model to detect and recognize the feature information of the third image. Through the image recognition model, the input image can be converted into a corresponding feature vector or feature descriptor, and these feature information can accurately represent the content and features of the image; by comparing the feature information of the third image with the preset feature information comparison table, and calculating the similarity between the obtained feature information and all the comparison feature information in the preset feature information comparison table, the feature information most similar to the input image can be quickly found; in this way, the feature matching and image recognition analysis tasks can be effectively carried out, calculating the similarity between the feature information of the third image and the comparison feature information in the feature information comparison table, so as to determine the comparison feature information with the highest similarity as the target feature information, and then determine the comparison attribute information corresponding to the target feature information as the attribute information of the third image, that is, the attribute information of the target cells. By determining the comparison feature information with the highest similarity, the feature information corresponding to the target image can be accurately found, and the comparison attribute information corresponding to the comparison feature information can be used to describe the attribute characteristics of the target cells, such as color, shape, texture, etc.; finally, the target cells in the image can be quickly and effectively recognized, and the recognition error caused by other factors can be prevented.

[0119] Then, according to the number of target cells recognized per unit area, the turbidity rating information of the sample is determined and the turbidity rating information is output.

[0120] In cell analysis, the number of cells and their distribution can reflect the turbidity level of the sample. The rating criteria can be set according to specific application scenarios and are generally divided into the following levels: Low Turbidity: The number of cells per unit area is small, indicating that the sample is relatively clear. Moderate Turbidity: The number of cells per unit area is moderate, indicating that the sample has a certain degree of turbidity. High Turbidity: The number of cells per unit area is large, indicating that the sample is relatively turbid.

[0121] For example, calculate the number of target cells per unit area

[0122] Determine the analysis area. First, determine the correspondence between the image size and the actual area: First, it is necessary to know the actual physical area corresponding to each pixel in the image (such as square micrometers, square millimeters, etc.). This requires pre-calibrating the magnification and resolution of the microscope or imaging device. For example: If the magnification of the microscope is 1000 times and the actual area corresponding to each pixel is 0.1 square micrometers. The image size is 1024x1024 pixels, then the actual analysis area is 1024 * 0.1 = 102.4 micrometers × 102.4 micrometers.

[0123] Cell counting, identification and marking of target cells: Through a pre-trained image recognition model, the system has identified and marked all target cells in the second image (image of the target cell area).

[0124] Calculate the cell density per unit area. Use the following formula to calculate the density of target cells per unit area: Cell density = total number of target cells divided by the area of the analysis region; For example: Total number of target cells: 500, Area of the analysis region: 100 square micrometers; Cell density = 500 / 100 = 5 cells / square micrometer;

[0125] Determine the turbidity rating: Based on the calculated cell density, map it to a predefined turbidity rating standard. The following is an exemplary rating mechanism: Cell density range (cells / square micrometer) Turbidity level 0 - 1 Extremely low turbidity; 1 - 5 Low turbidity; 5 - 10 Moderate turbidity; 10 - 20 High turbidity; > 20 Extremely high turbidity.

[0126] Specifically, as Figure 2 shown, in step S40, the third image is input into a pre-trained image recognition model to determine the attribute information of the target cells, including:

[0127] Step S41: Use the image recognition model to perform an initial recognition on the third image, and the initial recognition obtains the five-level association feature information of the current third image;

[0128] The five-level associated feature information includes color feature information, shape feature information, structural feature information, local feature information, and global feature information;

[0129] It should be noted that after the color feature information, shape feature information, structural feature information, local feature information, and global feature information of the third image are recognized and output by the image recognition model, the corresponding attribute information of the target cell in the third image can be recognized by the preset feature information comparison table, which is the attribute information of the target cell, that is, the type information and name information of the target cell;

[0130] For example: The color feature information output by the image recognition model for the third image is C.01 - red; the shape feature information is S.314 - biconcave disc shape, the structural feature information is St.152 - no cell nucleus and St.02 - having a cell membrane main body; the local feature information is L.166 - no edges; the global feature information is G.152 - red blood cells in human blood; thus, it can be recognized from the feature information comparison table that the object in the current third image is {C.01, S.314, St.152, St.02, L.166, G.152} - red blood cells.

[0131] Step S42: Calculate and obtain the similarity between each of the five-level associated feature information and all the comparison feature information in the preset feature information comparison table;

[0132] Step S43: Determine the comparison feature information corresponding to the highest similarity as the target comparison feature information; determine the comparison attribute information corresponding to the target feature information as the attribute information of the target cell.

[0133] It should be noted that for the technical solution adopted in the embodiment of the present application above, the feature information of the third image is detected and recognized through a pre-trained image recognition model. Through the image recognition model, the input image can be converted into a corresponding feature vector or feature descriptor, and these feature information can accurately represent the content and features of the image; by comparing the preset feature information comparison table with the feature information of the third image, by calculating the similarity between the obtained feature information and all the comparison feature information in the preset feature information comparison table, the feature information most similar to the input image can be quickly found; in this way, feature matching and image recognition analysis tasks can be effectively carried out, calculating the similarity between the feature information of the third image and the comparison feature information in the feature information comparison table, so as to determine the comparison feature information with the highest similarity as the target feature information, and further determine the comparison attribute information corresponding to the target feature information as the attribute information of the third image, that is, the attribute information of the target cell. By determining the comparison feature information with the highest similarity, the feature information corresponding to the target image can be accurately found, and the comparison attribute information corresponding to the comparison feature information can be used to describe the attribute characteristics of the target cell, such as color, shape, texture, etc.

[0134] Specifically, as Figure 3 shown, in step S41, the image recognition model is used to perform an initial recognition on the third image, and the five-level associated feature information of the current third image is obtained through the initial recognition, including the following operation steps:

[0135] Step S411: Establish an initial image recognition model;

[0136] Step S412: After inputting the third image into the trained image recognition model trained based on the initial image recognition model, the five-level associated feature information corresponding to the third image is output;

[0137] It should be noted that after establishing the initial image recognition model in the embodiment of the present application above, the initial image recognition model is trained to obtain a trained image recognition model, and then the third image is input into the trained image recognition model, and the five-level associated feature information corresponding to the third image is output, that is, the third image containing only the target cell is input into the trained image recognition model, and the color feature information, shape feature information, structural feature information, local feature information, and global feature information corresponding to the target cell in the third image are output, so as to identify the attribute information of the target cell in the third image according to the five-level feature information in the third image;

[0138] In this embodiment, after inputting the image containing only the above-mentioned red blood cells into the trained image recognition model, the color feature information of the output red blood cell image (i.e., the third image) is C.01-red; the shape feature information is S.314-biconcave disc-shaped, the structural feature information is St.152-no cell nucleus and St.02-with cell membrane body; the local feature information is L.166-no edges; and the global feature information is G.152-red blood cells in human blood.

[0139] St.152 - Absence of nucleus: Mature red blood cells lack a nucleus, a characteristic that distinguishes them from other blood cell types, such as white blood cells. St.02 - Cell membrane: Red blood cells possess a flexible and tough cell membrane, enabling them to circulate through narrow capillaries. Local characteristics (L.166 - Absence of edges): Red blood cells have a smooth surface with no distinct edges, which aids their flexibility.

[0140] When blood smears are treated with Giemsa stain and examined under a light microscope, the colors of red blood cells and white blood cells appear as follows: Red blood cell (erythrocyte) color: Pink to red. This is because the dye in the Giemsa stain binds to the hemoglobin within the red blood cells, giving them their characteristic pink to red hue.

[0141] Similarly, if the color feature information of (i.e. the third image) is C.04-green; the shape feature information is S.316-arched, the structural feature information is St.196-with cell nucleus and St.02-with cell membrane body; the local feature information is L.199-with edges; the global feature information is G.187-Plasmodium gametocytes in human blood.

[0142] Blood smears were processed using Giemsa stain and observed under an optical microscope. Plasmodium gametocytes are arched or crescent-shaped, which is their most notable morphological feature. In immunofluorescence staining, Plasmodium gametocytes are usually labeled green. This is because one of the commonly used fluorescent markers is fluorescein isothiocyanate (FITC), which emits green fluorescence. When FITC is used to link specific antibodies to label pathogens, P. falciparum gametocytes will show green fluorescence under a fluorescence microscope, thereby improving the sensitivity and specificity of the detection.

[0143] Specifically, if Figure 4 As shown, in step S412, the third image is input into the trained image recognition model obtained by training the initial image recognition model, and then the five-level correlation feature information corresponding to the third image is output, which includes the following steps:

[0144] Step S4121: Obtain a training set, where the training set includes multiple training data and the annotation data corresponding to each piece of the training data; the training data is obtained from training images through web crawling technology; the annotation data is the attribute information corresponding to the target cells in the training data for annotating the training data;

[0145] Step S4122: Input the training data into a pre-established initial image recognition model, and output an image recognition vector;

[0146] Step S4123: Calculate and obtain a loss value based on the image recognition vector and the annotation data through a loss function;

[0147] Step S4124: Backpropagate the loss value to update the parameters of the initial image recognition model, complete the first iteration, set an iteration counter, and increment the iteration count of the iteration counter by one; and repeat the above operations, recording the loss value and the iteration count;

[0148] The iteration counter is initially set to 0 and is 1 when the first iteration is completed;

[0149] Step S4125: Preset a maximum number of iterations and a maximum threshold for the loss value; stop training when the loss value is less than or equal to the maximum threshold of the loss value or the number of iterations reaches the maximum number of iterations, and obtain a trained image recognition model;

[0150] Step S4126: Input the third image into the trained image recognition model, and output the five-level association feature information corresponding to the third image.

[0151] It should be noted that in the above embodiments of the present application, a large amount of training data and annotation data in the training set obtained by web crawling are used to train the image recognition model. Multiple images containing target cells are crawled from the plasma sample database through web crawling technology as training data, and these data are annotated, that is, attribute information is added to the target cells in each image, providing annotated training data for subsequent model training; further, an image recognition vector is obtained through the forward calculation of the model, and the image data is converted into a vector representation that can be used to calculate the loss value through the model; then the loss value is calculated through the defined loss function. The loss value is a metric for measuring the gap between the model prediction result and the true label, evaluating the performance of the current model on the training set, and providing a feedback signal for the next parameter update; and through repeated parameter updates and iterations, the model is gradually optimized to improve the accuracy of image recognition. Finally, the termination conditions for training are set to avoid overfitting or unlimited training; through cyclic iterations, the model can be gradually optimized to enable accurate recognition of images, and finally a trained image recognition model is obtained;

[0152] In the embodiment of the present application above, a large number of training images (i.e., the above-mentioned training set, which can also be considered as images containing only target cells) are labeled with five-level associated feature information (color feature information, shape feature information, structural feature information, local feature information) of the target cells in the training images. Then, the training images are input into the initial image recognition model to output prediction data (i.e., the above-mentioned image recognition vector). Furthermore, a loss value is calculated based on the prediction data and the labeled data (i.e., five-level associated feature information). Further, a finally trained image recognition model is obtained according to a preset loss value threshold and iteration number threshold. Finally, the current third image is input into the trained image recognition model, so as to output the five-level associated feature information corresponding to the current third image.

[0153] Preferably, as Figure 5 shown, in step S4126, inputting the third image into the trained image recognition model and outputting the five-level associated feature information corresponding to the third image includes the following operation steps:

[0154] Step S41261: Perform forward propagation on the third image through the trained image recognition model to output an image vector;

[0155] Step S41262: Perform a decoding operation on the image vector to obtain five-level associated feature information;

[0156] It should be noted that for the above decoding operation, operations such as feature visualization and deconvolution can be used to obtain the five-level associated feature information corresponding to the third image, so that the attribute information of the target cells in the image can be analyzed by using the five-level associated feature information and the reference feature information.

[0157] It should be noted that when the embodiment of the present application obtains the five-level associated feature information in the third image (i.e., the color feature information, shape feature information, structural feature information, local feature information of the above-mentioned red blood cells), in the trained image recognition model, forward propagation is first performed to output an image vector, and then the image vector is decoded according to a preset decoder to obtain color feature information, shape feature information, structural feature information, and local feature information.

[0158] Specifically, as Figure 6 shown, in step S41261, performing forward propagation on the third image through the trained image recognition model to output an image vector includes the following operation steps:

[0159] Step S412611: Obtain the image and model parameter data; in the initial state, set p convolution kernels of size a×b; based on the convolution kernels, the image, and the model parameter data, perform a convolution operation on all pixel points in the third image to obtain the signal value corresponding to each pixel point;

[0160] The image and model parameter data include all pixel points in the third image, the coordinates (x, y) corresponding to all pixel points, the intensity value t corresponding to the pixel point i and the weight value corresponding to the trained image recognition model; p is greater than or equal to 1;

[0161] The signal value corresponding to the pixel point is expressed as (i.e., the signal value corresponding to the current pixel point i):

[0162]

[0163] The technical solution adopted in the above embodiments of the present application can scan and recognize an image through one or more (i.e., the above p convolution kernels) to extract p features in the image; by using one or more convolution kernels to perform a convolution operation on the input image, a series of feature maps are generated; each feature map represents the distribution of a certain local feature in the image, that is, the above embodiments of the present application finally output five feature information (i.e., the above five-level associated feature information), so the minimum number of types of convolution kernels in the embodiments of the present application is five; by using at least five convolution kernels to perform convolution on each pixel point in the image respectively, the signal value corresponding to each pixel point is obtained (the signal value can be a color signal value, a shape signal value, a structure signal value, a local signal value, a global signal value, etc.), so that the signal value of each pixel point in the current third image can be finally obtained, and then in the following operation process, the feature vector of the third image is finally recognized and analyzed through operations such as activation functions and pooling layers, so that in the operation of the above step S41262, the five-level associated feature information is decoded from the feature vector through a decoder; the result of the convolution operation is a two-dimensional matrix, that is, the signal value corresponding to the above pixel point. Explanation: Among them, the five-level associated feature information (which can also be understood as multi-level associated features) refers to five important levels of parameter feature information. However, the reason for choosing five levels for the multi-level associated features is the above-mentioned parameter feature information extracted by sampling. If the number of types is too large (7-8 types), it will cause the pressure of convolution operations, but if the number of types of parameter feature information is less (for example, 2-3 types), it will cause a decrease in recognition accuracy;

[0164] In addition, the selection of the above five-level associated feature information is for five important parameter feature information selected for a specific detailed texture image.

[0165] Step S412612: Perform a non-linear mapping operation on the signal value through an activation function to obtain the activation signal corresponding to the signal value;

[0166] It should be noted that the above activation functions include ReLU (Rectified Linear Unit), Sigmoid, Tanh, etc. They perform non-linear mapping on each element in the feature map (i.e., the signal value corresponding to the above pixel points), generating new matrix information (i.e., the above activation signal), which can enhance the expression ability of the image recognition model.

[0167] Step S412613: Perform downsampling operation on the activation signal through a pooling layer to obtain a target activation signal;

[0168] It should be noted that the above pooling layer usually uses a max-pooling layer or an average-pooling layer for downsampling operation, which can reduce the spatial dimension of the feature signal matrix and retain important feature information; max-pooling selects the maximum value of each region as the pooling result, while average-pooling calculates the average value of each region;

[0169] Taking the max-pooling layer as an example for the above downsampling operation: Input a feature signal matrix with a size of 4x4: And perform a 2×2 max-pooling operation to obtain the maximum value of each divided local area as the output value (i.e., the above target activation signal): In this way, the output value 6-8-14-16 can be obtained after the pooling operation;

[0170] Step S412614: Convert the target activation signal through a fully connected layer to obtain an image vector;

[0171] It should be noted that the above fully connected layer converts the target activation signal to obtain a feature vector with a fixed length (i.e., the above image vector can obtain five-level associated feature information after decoding operation);

[0172] In the above embodiments of the present application, the convolutional layer extracts image features, the activation function introduces non-linear transformation to enhance the image expressiveness, the pooling layer reduces the spatial dimension, improves the processing speed, and reduces the computational processing pressure. The fully connected layer converts the target activation signal to obtain a feature vector with a fixed length, and finally obtains a result with an abstract feature representation, providing the trained image recognition model with the ability to effectively process and analyze images.

[0173] Specifically, as Figure 7 shown, in step S42, calculate the similarity between the five-level associated feature information and all the reference feature information in the preset feature information comparison table, including the following operation steps:

[0174] Step S421: Set header data for the five-level associated feature information;

[0175] The header data includes color feature header data O, shape feature header data P, structural feature header data Q, local feature header data R, and global feature header data S;

[0176] Step S422: Traverse all the reference feature information in the feature information comparison table to obtain the reference header data in the reference feature information;

[0177] The reference header data includes reference color feature header data O', reference shape feature header data P', reference structural feature header data Q', reference local feature header data R', and reference global feature header data S';

[0178] Step S423: Establish an association relationship between the reference header data and the header data to obtain a reference association matrix D;

[0179] The reference association matrix is expressed as:

[0180]

[0181] Step S424: Calculate the similarity of each factor in the reference association matrix (and repeat the above operations until the similarities between the above five-level associated feature information and all the reference feature information in the preset feature information comparison table are obtained after traversing all the reference feature information in the feature information comparison table);

[0182] It should be noted that in the above embodiments of the present application, by setting header data for the five-level associated feature information, the color feature information in the above embodiments is obtained as O#C.01, the shape feature information is P#S.314, the structure feature information is Q#St.152 and Q#St.02 - with a cell membrane main body; the local feature information is R#L.166; the global feature information is S#G.152; and then an association relationship is established according to the control header data of the control feature information in the feature information control table and the header data of the five-level associated feature information. For example: the control header data of the control feature information in the feature information control table are: red blood cells = {O'#C.01, P'#S.314, Q'#St.152, Q'#St.02, R'#L.166, S'#G.152}, fat granule suspension = {O'#C.02, P'#S.04, Q'#St.152, Q'#St.02, R'#L.25, S'#G.33}, etc. Then, the association relationship is established by respectively corresponding the header data of the five-level associated feature information to the control header data of the control feature information in the feature information control table, that is, O#C.01 - O'#C.01, P#S.314 - P'#S.314, Q#St.152 - Q'#St.152, Q#St.02 - Q'#St.02, R#L.166 - R'#L.166, S#G.152 - S'#G.152 and O#C.01 - O'#C.02, P#S.314 - P'#S.04, Q#St.152 - Q'#St.152, Q#St.02 - Q'#St.02, R#L.166 - R'#L.25, S#G.152 - S'#G.33, which is expressed as:

[0183] Furthermore, calculate the similarity of each factor in each matrix, and repeat the above operations to obtain the similarities of the above five-level associated feature information with all control feature information in the preset feature information control table.

[0184] In summary, a turbidity evaluation and processing method for blood-separated plasma samples provided by the present application first processes a first image containing target cells to obtain a second image, and extracts a third image containing only target cells from the second image, which can reduce errors caused by factors such as the surrounding environment of the target cells in subsequent identification of the attribute information of the target cells; then, the third image is input into a trained image recognition model to detect and recognize the feature information of the third image. Through the image recognition model, the input image can be converted into corresponding feature vectors or feature descriptors, and these feature information can accurately represent the content and features of the image; by comparing the obtained feature information with all the reference feature information in the preset feature information comparison table, the similarity between the calculated feature information and the reference feature information in the preset feature information comparison table can be calculated, and the feature information most similar to the input image can be quickly found; in this way, feature matching and image recognition analysis tasks can be effectively carried out, the similarity between the feature information of the third image and the reference feature information in the feature information comparison table is calculated, so as to determine the reference feature information with the highest similarity as the target feature information, and then determine the reference attribute information corresponding to the target feature information as the attribute information of the third image, that is, the attribute information of the target cells. By determining the reference feature information with the highest similarity, the feature information corresponding to the target image can be accurately found, and the reference attribute information corresponding to the reference feature information can be used to describe the attribute characteristics of the target cells, such as color, shape, texture, etc.; finally, the target cells in the image can be quickly and effectively recognized, and recognition errors caused by other factors can be prevented;

[0185] Specifically, when detecting and outputting the feature information of the third image through a trained image recognition model, first establish an initial image recognition model, then train the initial image recognition model to obtain a trained image recognition model, and then input the third image into the trained image recognition model. The convolutional kernel in the trained image recognition model performs a convolution operation on the pixel points in the third image to obtain color signal values, shape signal values, structure signal values, local signal values, global signal values, etc. of each pixel point. Then, through operations such as activation functions and pooling layers, by enhancing the image expressiveness and reducing the data volume, the effect of extracting the five-level feature information of the image is achieved. Then, after obtaining a reference correlation matrix through feature matching with the preset feature information comparison table, the attribute information of the target cells corresponding to the five-level correlation feature information is recognized by calculating the similarity;

[0186] Embodiment 2

[0187] In addition, based on the same concept of the above method embodiments, the embodiments of the present application further provide a turbidity evaluation processing system for blood-separated plasma samples to implement the above method of the present application. Since the principle of solving problems in the method embodiments is similar to that of the system, it has at least all the beneficial effects brought by the technical solutions of the above embodiments, which will not be elaborated here one by one.

[0188] As Figure 8 shown, the present invention provides a turbidity evaluation processing system for blood-separated plasma samples, including a first image acquisition module 10, a second image acquisition module 20, a third image acquisition module 30, and an image recognition module 40;

[0189] Among them, the first image acquisition module 10 is used to acquire a first image containing target cells;

[0190] The first image is an initial image, which is a plurality of initial images for identifying target cells in the image obtained by shooting with a camera or manually collected;

[0191] The second image acquisition module 20 is used to perform recognition processing on the first image to obtain a second image in the first image;

[0192] The second image is a target cell region image;

[0193] The third image acquisition module 30 is used to perform recognition and analysis on the second image through an edge detection algorithm to obtain the edge information of the target cell; and intercept a third image according to the edge information;

[0194] The third image is the target cell image;

[0195] The image recognition module 40 is used to input the third image into a pre-trained image recognition model to determine the attribute information of the target cell;

[0196] The attribute information of the target cell includes the name information and type information of the target cell;

[0197] Preferably, the image recognition module 40 includes an initial recognition sub-module 41, a calculation sub-module 42, and a screening and acquisition sub-module 43;

[0198] The initial recognition sub-module 41 is used to perform initial recognition on the third image by using the image recognition model, and initially recognize the five-level association feature information of the current third image;

[0199] The five-level association feature information includes color feature information, shape feature information, structural feature information, local feature information, and global feature information;

[0200] The calculation sub-module 42 is configured to calculate and obtain the similarities between the five-level associated feature information and all the reference feature information in a preset feature information comparison table;

[0201] The screening and obtaining sub-module 43 is configured to determine the reference feature information corresponding to the highest similarity as the target reference feature information; and determine the reference attribute information corresponding to the target feature information as the attribute information of the target cell.

[0202] Preferably, the initial recognition sub-module 41 is specifically configured to establish an initial image recognition model;

[0203] After inputting the third image into the trained image recognition model obtained by training based on the initial image recognition model, the five-level associated feature information corresponding to the third image is output.

[0204] Preferably, in specific implementation, the initial recognition sub-module 41 is configured to obtain a training set, where the training set includes a plurality of training data and the annotation data corresponding to each training data; the training data is obtained from training images through web crawler technology; the annotation data is the attribute information of the target cell corresponding to the training data for annotating the training data;

[0205] Input the training data into a pre-established initial image recognition model, and output an image recognition vector;

[0206] Based on the image recognition vector and the annotation data, calculate and obtain a loss value through a loss function;

[0207] Backpropagate the loss value to update the parameters of the initial image recognition model, complete the first iteration, and set an iteration counter, and increment the iteration counter by one; and repeat the above operations, recording the loss value and the iteration times;

[0208] The iteration counter is initially set to 0 and is 1 when the first iteration is completed;

[0209] Preset a maximum iteration number and a maximum loss value threshold; stop training when the loss value is less than or equal to the maximum loss value threshold or the iteration number reaches the maximum iteration number, and obtain a trained image recognition model;

[0210] Input the third image into the trained image recognition model, and output the five-level associated feature information corresponding to the third image.

[0211] Preferably, in specific implementation, the initial recognition sub-module 41 is further configured to perform forward propagation on the third image through the trained image recognition model, and output an image vector;

[0212] Perform a decoding operation on the image vector to obtain five-level associated feature information;

[0213] Preferably, the initial recognition sub-module 41, in specific implementation, is further configured to obtain an image and model parameter data; in the initial state, set p convolution kernels of size a×b; based on the convolution kernels, the image, and the model parameter data, perform a convolution operation on all pixel points in the third image to obtain a signal value corresponding to each pixel point;

[0214] The image and model parameter data include all pixel points in the third image, the coordinates (x, y) corresponding to all pixel points, the intensity value t corresponding to the pixel points i , and the weight value corresponding to the trained image recognition model; p is greater than or equal to 1;

[0215] The signal value corresponding to the pixel point is expressed as:

[0216]

[0217] Perform a non-linear mapping operation on the signal value through an activation function to obtain an activation signal corresponding to the signal value;

[0218] Perform a downsampling operation on the activation signal through a pooling layer to obtain a target activation signal;

[0219] Perform a conversion on the target activation signal through a fully connected layer to obtain an image vector;

[0220] Preferably, the calculation sub-module 42 is specifically configured to set header data for the five-level associated feature information;

[0221] The header data includes color feature header data O, shape feature header data P, structure feature header data Q, local feature header data R, and global feature header data S;

[0222] Traverse all the comparison feature information in the feature information comparison table to obtain the comparison header data in the comparison feature information;

[0223] The comparison header data includes comparison color feature header data O', comparison shape feature header data P', comparison structure feature header data Q', comparison local feature header data R', and comparison global feature header data S';

[0224] Establish an association relationship between the comparison header data and the header data to obtain a comparison association matrix D;

[0225] The comparison association matrix is expressed as:

[0226]

[0227] Calculate and obtain the similarity of each factor in the control association matrix (and after repeating the above operations until all the control feature information in the feature information comparison table is traversed, the similarities between the above five-level association feature information and all the control feature information in the preset feature information comparison table are obtained);

[0228] Embodiment III

[0229] According to another aspect of the embodiments of the present application, there is also provided a storage medium having non-volatile program code executable by a processor. The storage medium stores a computer program, and when the computer program is executed by the processor, the steps of a turbidity evaluation processing method for a blood separation plasma sample described in any one of the above embodiments are implemented.

[0230] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit them; those of ordinary skill in the art can modify the technical solutions described in the foregoing embodiments, or perform equivalent replacements for some or all of the technical features; and these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present invention.

Claims

1. A turbidity evaluation and processing method for blood-separated plasma samples, characterized in that The operation steps are as follows: Obtain a first image containing target cells; Perform recognition processing on the first image to obtain a second image in the first image; The second image is a target cell region image; After performing recognition and analysis on the second image through an edge detection algorithm, intercept a third image; The third image is the target cell image; Input the third image into a pre-trained image recognition model to determine the attribute information of the target cells; The attribute information of the target cells includes the name information of the target cells; Identify and label all target cells in the target cell region image, and calculate the number of target cells recognized per unit area based on the total number of target cells; Determine the turbidity rating information of the sample based on the number of target cells recognized per unit area and output the turbidity rating information; The inputting the third image into a pre-trained image recognition model to determine the attribute information of the target cells includes: Use the image recognition model to perform initial recognition on the third image, and initially recognize and obtain five-level associated feature information of the current third image; The five-level associated feature information includes color feature information, shape feature information, structural feature information, local feature information, and global feature information; Calculate and obtain the similarity between each of the five-level associated feature information and all reference feature information in a pre-set feature information comparison table; Determine that the reference attribute information corresponding to the highest similarity is the attribute information of the target cells.

2. The turbidity evaluation processing method of a blood-separated plasma sample according to claim 1, wherein The using the image recognition model to perform initial recognition on the third image, and initially recognize and obtain five-level associated feature information of the current third image includes the following operation steps: Establish an initial image recognition model; Input the third image into a trained image recognition model trained based on the initial image recognition model, and output to obtain five-level associated feature information corresponding to the third image.

3. The turbidity evaluation processing method for a blood-separated plasma sample according to claim 2, wherein The inputting the third image into a trained image recognition model trained based on the initial image recognition model, and output to obtain five-level associated feature information corresponding to the third image includes the following operation steps: Obtain a training set, where the training set includes multiple training data and the annotation data corresponding to each training data; the training data are training images; the annotation data are the attribute information corresponding to the target cells in the training data for annotation training of the training data; Input the training data into a pre-established initial image recognition model, and output to obtain an image recognition vector; Calculate and obtain a loss value based on the image recognition vector and the annotation data through a loss function; Backpropagate the loss value to update the parameters of the initial image recognition model, and perform iterative training through a preset maximum number of iterations and a maximum loss value threshold to obtain a trained image recognition model; Input the third image into the trained image recognition model, and output to obtain five-level associated feature information corresponding to the third image.

4. A turbidity evaluation processing method for blood separation plasma samples according to claim 3, characterized in that, The inputting the third image into the trained image recognition model, and output to obtain five-level associated feature information corresponding to the third image includes the following operation steps: Forward propagate the third image through the trained image recognition model to output an image vector; Perform a decoding operation on the image vector to obtain five-level associated feature information.

5. A turbidity evaluation and processing method for a blood-separated plasma sample according to claim 4, characterized in that, The forward propagation of the third image through the trained image recognition model to output an image vector includes the following operation steps: Obtain image and model parameter data; in the initial state, set p convolution kernels of size a×b; based on the convolution kernels, the image, and the model parameter data, perform a convolution operation on all pixel points in the third image to obtain a signal value corresponding to each pixel point; The image and model parameter data include all pixel points in the third image, the coordinates (x, y) corresponding to all pixel points, and the intensity value t corresponding to the pixel points i , and the weight value corresponding to the trained image recognition model; p is greater than or equal to 1; The signal value corresponding to the pixel point is expressed as: ； Perform a non-linear mapping operation on the signal value through an activation function to obtain an activation signal corresponding to the signal value; Perform a downsampling operation on the activation signal through a pooling layer to obtain a target activation signal; Convert the target activation signal through a fully connected layer to obtain an image vector.

6. A turbidity evaluation processing method for a blood-separated plasma sample according to claim 5, characterized in that The calculation of obtaining the similarity between the five-level associated feature information and all reference feature information in the preset feature information comparison table includes the following operation steps: Set header data for the five-level associated feature information; Traverse all reference feature information in the feature information comparison table to obtain the reference header data in the reference feature information; Establish an association relationship between the reference header data and the header data to obtain a reference association matrix D; Calculate the similarity of each factor in the reference association matrix.

7. A turbidity evaluation processing method for a blood-separated plasma sample according to claim 6, characterized in that, The header data includes color feature header data O, shape feature header data P, structure feature header data Q, local feature header data R, and global feature header data S; The reference header data includes reference color feature header data O', reference shape feature header data P', reference structure feature header data Q', reference local feature header data R', and reference global feature header data S'; The reference association matrix is expressed as: 。 8. A storage medium, in which a computer program is stored, and when the computer program is executed by a processor, the steps of a turbidity evaluation processing method for a blood separation plasma sample according to any one of claims 1-7 above are implemented.

Images