Auxiliary diagnosis method and biological sample analysis method, device, equipment and medium thereof

Abstract

The application provides an auxiliary diagnosis method and a biological sample analysis method, device, equipment and medium thereof. The sample analysis method acquires pulse data of a biological sample of a to-be-detected object collected in each different detection channel, quantizes a plurality of pulse features of a set type carried by the pulse data of each different detection channel, obtains a corresponding first array, fuses the first arrays corresponding to the pulse features of the set type, obtains a second array, extracts features from the second array, and outputs a recognition result of a target object in the biological sample carried by the second array. The second array contains a plurality of pulse features of a set type of pulse data of each different detection channel. Thus, the accuracy of the recognition result obtained based on the second array as input data of a recognition model is higher.

Description

Technical Field

[0001] This application relates to the field of medical device technology, and in particular to an auxiliary diagnostic method and its biological sample analysis method, device, equipment and medium. Background Technology

[0002] Currently, blood cell analysis equipment detects and analyzes blood cells by converting the different biological characteristics of blood cells into corresponding pulse signals through a built-in detection channel. These pulse signals are then converted into image data for analysis, allowing for the extraction of relevant cell features and identification. However, during the conversion of pulse signals into image data, some information is lost. For example, after converting the pulse height into image data, pulse features such as full-width, half-width, and pulse area are lost. Clearly, these lost pulse features do not contribute to the analysis of blood cells based on this image data.

[0003] However, prior knowledge indicates that the aforementioned lost pulse features are related to the classification of blood cells. Therefore, the accuracy of analyzing blood cells based on image data converted from pulse signals corresponding to the biological characteristics of blood cells is not high. Summary of the Invention

[0004] To address the existing technical problems, this application provides a highly accurate auxiliary diagnostic method and its biological sample analysis method, apparatus, equipment, and medium.

[0005] A biological sample analysis method, comprising:

[0006] Acquire pulse data from biological samples of the object to be tested in different detection channels;

[0007] The pulse features of multiple preset types carried by the pulse data of each of the different detection channels are quantized to obtain the corresponding first arrays;

[0008] The first array corresponding to the pulse features of the specified type is fused to obtain the second array;

[0009] Feature extraction is performed on the second array, and the identification results of the target objects in the biological samples contained in the second array are output.

[0010] An auxiliary diagnostic method, comprising:

[0011] Obtain the identification result of the target object in the biological sample obtained by analyzing the biological sample of the object to be tested using the biological sample analysis method described above;

[0012] Obtain the clinical information of the subject to be tested;

[0013] Based on the identification results and the clinical information, abnormal target objects in the biological sample of the object to be tested are identified, and auxiliary diagnostic decision information corresponding to the abnormal target objects is output.

[0014] A biometric sample identification device, characterized in that it comprises:

[0015] The acquisition module is used to acquire pulse data collected from biological samples of the object to be tested in different detection channels;

[0016] The quantization module is used to quantize the pulse features of multiple preset types carried by the pulse data of each of the different detection channels to obtain the corresponding first arrays.

[0017] The fusion module is used to fuse the first array corresponding to the pulse features of the specified type to obtain a second array;

[0018] The identification module is used to extract features from the second array and output the identification results of the target objects in the biological samples carried by the second array.

[0019] A biological sample analysis device includes a processor and a memory, wherein the memory stores a computer program that can be executed by the processor, and the computer program, when executed by the processor, implements the biological sample analysis method.

[0020] A computer-readable storage medium storing a computer program that, when executed by a controller, implements the biological sample analysis method.

[0021] An auxiliary diagnostic device includes a processor and a memory, wherein the memory stores a computer program that can be executed by the processor, and the computer program, when executed by the processor, implements the auxiliary diagnostic method.

[0022] A computer-readable storage medium storing a computer program that, when executed by a controller, implements the aforementioned auxiliary diagnostic method.

[0023] As can be seen from the above, in one aspect, the sample analysis method provided in this application acquires pulse data collected from biological samples of the object to be detected in various detection channels, quantifies multiple preset types of pulse features carried by the pulse data of each different detection channel to obtain corresponding first arrays, then merges the first arrays corresponding to the preset types of pulse features to obtain a second array, performs feature extraction on the second array, and outputs the identification result of the target object in the biological sample carried by the second array. The second array contains multiple preset types of pulse features from the pulse data of each different detection channel; thus, the accuracy of obtaining the identification result based on the second array as input data for the identification model is high.

[0024] On the other hand, the auxiliary diagnostic method provided in this application determines abnormal target objects in the biological sample of the test subject based on the identification results obtained by the biological sample analysis method and the clinical information of the test subject, and outputs auxiliary diagnostic decision information corresponding to the abnormal target objects. In this way, the abnormal target objects contained in the biological sample can be used as identification objects, and the abnormal target objects in the biological sample can be determined by combining the clinical information of the test subject. Based on the abnormal target objects, the corresponding auxiliary diagnostic decision information can be output. Accurate auxiliary diagnostic decisions can be obtained without relying on the personal experience level of the laboratory physician, thereby improving the efficiency of testing, making the test results more accurate, and making the diagnostic conclusions more interpretable. Attached Figure Description

[0025] The accompanying drawings are for illustrative purposes only and are not intended to limit the invention. Furthermore, the same reference numerals denote the same parts throughout the drawings. In the drawings:

[0026] Figure 1 This is a schematic diagram of the process flow for the biological sample analysis method provided in Embodiment 1 of this application;

[0027] Figure 2 This is a schematic flowchart of the biological sample analysis method provided in Embodiment 2 of this application;

[0028] Figure 3 This is a schematic flowchart of the biological sample analysis method provided in Embodiment 3 of this application;

[0029] Figure 4 This is a schematic flowchart of the biological sample analysis method provided in Embodiment 4 of this application;

[0030] Figure 5 This is a schematic flowchart of the biological sample analysis method provided in Embodiment 5 of this application;

[0031] Figure 6This is a schematic flowchart of the biological sample analysis method provided in Embodiment Six of this application;

[0032] Figure 7 This is a schematic flowchart of the biological sample analysis method provided in Embodiment 7 of this application;

[0033] Figure 8 This is a schematic flowchart of the biological sample analysis method provided in Embodiment 8 of this application;

[0034] Figure 9 This is a schematic flowchart of the biological sample analysis method provided in Embodiment 9 of this application;

[0035] Figure 10 This is a schematic flowchart of the biological sample analysis method provided in Embodiment 10 of this application;

[0036] Figure 11 This is a schematic flowchart of the biological sample analysis method provided in Embodiment Eleven of this application;

[0037] Figure 12 This is a schematic diagram illustrating the second array configuration in a biological sample analysis method provided according to some embodiments of this application;

[0038] Figure 13 This is a schematic diagram of the structure of the biological sample analysis device provided according to Embodiment 1 of this application;

[0039] Figure 14 This is a schematic diagram of the structure of the biological sample analysis device provided according to Embodiment 1 of this application;

[0040] Figure 15 This is a schematic flowchart of an auxiliary diagnostic method provided according to some embodiments of this application. Detailed Implementation

[0041] The technical solution of this application will be further described in detail below with reference to the accompanying drawings and specific embodiments.

[0042] To make the objectives, technical solutions, and advantages of this application clearer, the application will be further described in detail below with reference to the accompanying drawings. The described embodiments should not be regarded as limitations on this application. All other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0043] In the following description, the phrase "some embodiments" refers to a subset of all possible embodiments. It should be noted that "some embodiments" can be the same subset or different subsets of all possible embodiments, and can be combined with each other without conflict.

[0044] In the following description, the terms "first, second, and third" are used merely to distinguish similar objects and do not represent a specific ordering of objects. It is understood that "first, second, and third" may be interchanged in a specific order or sequence where permitted, so that the embodiments of this application described herein can be implemented in an order other than that illustrated or described herein.

[0045] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application belongs. The terminology used herein is for the purpose of describing embodiments of this application only and is not intended to limit this application.

[0046] Please see Figure 1 The biological sample analysis method provided in Embodiment 1 of this application can be applied to biological sample analysis equipment. The biological sample analysis method includes, but is not limited to, S02, S04, S06, and S08, as described in detail below:

[0047] S02: Acquire pulse data of the biological sample of the object to be tested in different detection channels.

[0048] S02 can be derived from Figure 13 The acquisition module 101 in the biological sample analysis device shown is used to implement this, or is implemented by... Figure 14 The biological sample analysis device shown has a memory 202 that stores the corresponding acquisition program, which is then processed by... Figure 14 The processor 201 in the biological sample analysis device shown is implemented when executing the acquisition program stored in the memory 202.

[0049] The subject to be tested refers to the owner of the biological sample to be tested. Taking a human blood sample as an example, the subject to be tested is usually the patient who provides the blood sample. The biological sample of the subject to be tested can be a sample containing various biological cell information or other biological information, such as blood samples, urine samples, other body fluids (pleural effusion, ascites, cerebrospinal fluid, serous cavity effusion, synovial fluid), etc. The biological cell type can be at least one of the group consisting of neutrophils, lymphocytes, monocytes, eosinophils, and basophils; it can also be immature granulocytes, tumor cells, lymphoblasts, plasma cells, atypical lymphocytes, proerythroblasts, basal erythroblasts and polychromatic erythroblasts, normal erythroblasts, promegalyte cells, basal giant cells, polychromatic giant cells, and nucleated erythroblasts selected from normal giant erythroblasts and macronucleated globules. The detection channels include a counting channel for counting cells in a biological sample based on the Coulter impedance principle, a concentration measurement channel for determining the concentration of various components in the biological sample based on the colorimetric principle, and a classification and identification channel for identifying various components of the biological sample based on laser heat dissipation and nucleic acid fluorescence staining technology. The counting channels further include white blood cell counting channels, red blood cell counting channels, and platelet counting channels. The concentration measurement channel, based on the colorimetric principle, is used to determine the concentration of components such as hemoglobin, albumin, and globulin. The classification and identification channel is mainly used for white blood cell classification, reticulocyte identification, nucleated red blood cell identification, basophil identification, low platelet count identification, and primitive cell identification. The pulse data are electrical pulse signals generated by each of the detection channels detecting the biological characteristics of the biological sample, converting the corresponding biological characteristics into electrical pulse signals. Taking blood as an example, the biological characteristics of blood include the size of blood cells, the complexity of cell contents, and nucleic acid content.

[0050] Specifically, in some embodiments, taking the biological sample as an example (blood sample), each of the detection channels includes, but is not limited to, an SFL (side fluorescence) channel for determining DNA / RNA content, an FSC (forward scattering) channel for determining cell volume, an SSC (side scattering) channel for determining cell complexity, a WBC channel for white blood cell counting, an RBC channel for red blood cell counting, and a PLT channel for platelet counting. Furthermore, in some embodiments, each of the detection channels can be set as follows: Figure 14The internal components of the illustrated biosample analysis device, specifically the pulse data, represent the raw data collected by the biosample analysis device. In other embodiments, each detection channel is located within a biosample detection device that is communicatively connected to the biosample analysis device, and the biosample analysis device performs corresponding analysis based on the pulse data collected by the biosample detection device. However, it should be noted that although the description of the embodiments in this application uses blood samples as an example of biological samples, this should not be construed as limiting the scope of protection of this application.

[0051] S04: Quantize the pulse features of multiple preset types carried by the pulse data of each of the different detection channels to obtain the corresponding first array.

[0052] S04 can be derived from Figure 13 The quantization module 102 in the biological sample analysis device shown is used to implement, or is implemented by, the quantization module 102 in the biological sample analysis device shown. Figure 14 The memory 202 in the biological sample analysis device shown stores the corresponding quantification program, which is then processed by... Figure 14 The processor 201 in the illustrated biological sample analysis device executes the quantization program stored in the memory 202.

[0053] The pulse features of the specified type refer to the pulse features related to the corresponding analysis task of the biological sample, that is, the pulse feature types obtained by quantization according to the corresponding analysis task requirements. The multiple pulse features of the specified types include, but are not limited to, pulse height, full peak pulse width, first half peak pulse width, second half peak pulse width, and pulse area. Quantizing the multiple pulse features of the specified types refers to the process of expressing the changes in the values ​​of each corresponding type of pulse feature, or the values ​​obtained after processing them, in the form of numerical values. Each first array in each first array corresponding to each different set type of pulse feature contains information about the pulse feature of that specified type in the pulse data of each different detection channel.

[0054] S06: Merge the first array corresponding to the pulse features of the specified type to obtain the second array.

[0055] S06 can be made by Figure 13 The fusion module 103 in the illustrated biological sample analysis device is implemented, or is achieved by... Figure 14 The memory 202 in the biological sample analysis device shown stores the corresponding fusion program, which is then processed by... Figure 14 The processor 201 in the illustrated biological sample analysis device is implemented when executing the fusion program stored in the memory 202.

[0056] Fusing the first array corresponding to the pulse features of the specified type means superimposing the first arrays corresponding to the pulse features of each specified type in a specified dimension to obtain a second array containing pulse feature information of multiple different specified types.

[0057] S08: Perform feature extraction on the second array and output the identification result of the target object in the biological sample carried by the second array.

[0058] S08 can be... Figure 13 The identification module 104 in the shown biological sample analysis device is implemented, or is achieved by... Figure 14 The biological sample analysis device shown has a memory 202 that stores the corresponding identification program, which is then processed by... Figure 14 The processor 201 in the illustrated biological sample analysis device executes the identification program stored in the memory 202.

[0059] In some embodiments, an AI recognition model can be used to extract features from the second array and output the recognition result of the target object in the biological sample carried by the second array. The AI ​​recognition model is an Artificial Intelligence (AI) model that uses machine learning to analyze the second array to extract features related to the target object, thereby outputting the recognition result of the target object. The target object refers to an object in the biological sample that is relevant to the current analysis task. Taking a blood sample as an example, if the current analysis needs to determine whether the Ig (immature granulocytes) in the blood to be tested is abnormal, then the target object is the content, volume, and / or distribution of immature granulocytes. In some embodiments, depending on the current analysis task, the target object may include, but is not limited to, the number of different types of cells such as white blood cells, red blood cells, and platelets, or the content and distribution of biological sample components such as hemoglobin, albumin, globulin, and DNA / RNA.

[0060] In the above embodiments, the sample analysis method acquires pulse data collected from biological samples of the target object in various detection channels, quantifies multiple preset types of pulse features carried by the pulse data in each of the different detection channels to obtain corresponding first arrays, and then merges the first arrays corresponding to the preset types of pulse features to obtain a second array. By extracting features from the second array, the identification result of the target object in the biological sample carried by the second array is output. The second array contains multiple preset types of pulse features from the pulse data of each different detection channel; therefore, the accuracy of obtaining the identification result based on the second array as input data for the identification model is high.

[0061] In other embodiments, features can be extracted from the second array using conventional algorithm models to output the identification results of the target objects in the biological samples carried by the second array. These conventional algorithms include image morphology algorithms, image classification algorithms, clustering algorithms, or threshold segmentation algorithms.

[0062] Please see Figure 2 The biological sample analysis method provided in Embodiment 2 of this application differs from Embodiment 1 in that, in Embodiment 2, S04: quantifying the multiple preset types of pulse features carried by the pulse data of each of the different detection channels to obtain corresponding first arrays includes: quantifying the multiple preset types of pulse features carried by the pulse data of each of the different detection channels to obtain each two-dimensional first array corresponding to each detection channel, wherein the first dimension value of each two-dimensional first array in the first dimension direction is determined according to the value range of the multiple preset types of pulse features, and the second dimension value of each two-dimensional first array in the second dimension direction is determined according to the number of channels of the detection channel.

[0063] Furthermore, in Embodiment 2, S06: fusing the first array corresponding to the first pulse feature and the first array corresponding to the second pulse feature to obtain a second array of multi-feature fusion includes: superimposing and fusing the first array corresponding to the first pulse feature and the two-dimensional first array corresponding to the second pulse feature in the third dimension direction to obtain a three-dimensional second array of multi-feature fusion, wherein the third dimension value of the three-dimensional second array in the third dimension direction is determined according to the number of features of the multiple set types of pulse features.

[0064] Specifically, the structural diagram of the second array is as follows: Figure 12As shown, H two-dimensional first arrays P1 to PH corresponding to H pulse features of a given type are superimposed and fused in the third dimension to obtain a three-dimensional second array. The first dimension values ​​L corresponding to each two-dimensional first array P1 to PH in the first dimension are all equal, and the second dimension values ​​W corresponding to each two-dimensional first array P1 to PH in the second dimension are also all equal. In some embodiments, the first dimension values ​​of each two-dimensional first array P1 to PH in the first dimension are determined according to the value range of the H pulse features of the given type, such as being set to the maximum value within the value range. The second dimension values ​​of each two-dimensional first array P1 to PH in the second dimension are equal to the total number of channels W of each detection channel C1 to CW. Each row of the two-dimensional first array corresponding to a pulse feature of a given type represents the L quantized values ​​of the corresponding pulse feature of the given type carried in the pulse data of the corresponding detection channel, such as F1, F2 to FL. The first dimension of the three-dimensional second array is the same as the first dimension of each two-dimensional first array and its corresponding dimension value. The second dimension of the three-dimensional second array is the same as the second dimension of each two-dimensional first array and its corresponding dimension value. The third dimension of the three-dimensional second array is perpendicular to the first dimension and the second dimension, respectively, and the corresponding third dimension value is the number H of multiple set types of pulses.

[0065] Please see Figure 3 The biological sample analysis method provided in Example 3 differs from that in Example 1 in that, in Example 3, S04: the pulse features of different preset types carried by the pulse data of each of the different detection channels are quantified according to a preset method to obtain corresponding first arrays, including S041 and S042, which are described in detail below:

[0066] S041: Perform frequency statistics on the first pulse feature of a set type carried by the pulse data of each detection channel to obtain a first array of set length values ​​corresponding to the first pulse feature.

[0067] S042: Based on the first array corresponding to the first pulse feature and the second pulse feature of a set type, calculate the mean value corresponding to each frequency position of the second pulse feature and the first pulse feature, and determine the first array corresponding to the second pulse feature.

[0068] The frequency statistics refer to counting the number of times each value of the first pulse feature occurs, and then arranging the frequencies corresponding to different values ​​in order to form a first array corresponding to the first pulse feature. To facilitate the training of the AI ​​recognition model, the dimension values ​​of each of the first arrays constituting the second array need to be set to be the same and fixed in different dimensional directions. The first dimension direction of the first array corresponding to the first pulse feature is the length direction, and its corresponding first dimension value is the length value. The second dimension direction of the first array corresponding to the first pulse feature is the width direction, and its corresponding second dimension value is the width value, which is equal to the number of different detection channels.

[0069] The second pulse feature of a given type may include one second pulse feature of a given type, or it may include multiple second pulse features of different given types. The frequency positions of the first pulse feature refer to the positions corresponding to each element in the first array corresponding to the first pulse feature, i.e., the positions where the same first pulse feature value is located. Calculating the mean value corresponding to each frequency position of the second pulse feature and the first pulse feature means averaging the second pulse features corresponding to the same first pulse feature value to obtain the average value of the second pulse feature at each frequency position. For example, in some embodiments, the first pulse feature is the pulse height, and the second pulse feature is the peak width (full peak width or half peak width), then each element in the first array corresponding to the second pulse feature is the average peak width corresponding to pulses with the same pulse height.

[0070] Please see Figure 4 According to the biological sample analysis method provided in Embodiment 4 of this application, S041: performing frequency statistics on the first pulse feature of a set type carried by the pulse data of each detection channel to obtain a first array of set length values ​​corresponding to the first pulse feature, including: performing frequency statistics on the pulse height carried by the pulse data of each detection channel to obtain a first array corresponding to the pulse height.

[0071] Further, S042: Based on the first array corresponding to the first pulse feature and the second pulse feature of a set type, calculate the mean value corresponding to each frequency position of the second pulse feature and the first pulse feature, and determine the first array corresponding to the second pulse feature, including: based on the first array corresponding to the pulse height of the pulse data and the full peak width of the pulse data, calculate the mean value corresponding to each frequency position of the full peak width and the pulse height, and determine the first array corresponding to the full peak width.

[0072] The second array, obtained by fusing the first array corresponding to the pulse height and the first array corresponding to the full peak width, contains pulse height information and corresponding pulse width information from multiple different detection channels. Therefore, inputting the second array into the AI ​​recognition model can extract more features that are beneficial to the analysis task, thus improving the accuracy of the analysis.

[0073] Please see Figure 5 As shown, the difference between the biological sample analysis method provided in Example 5 and Example 4 is that in Example 5, the full peak width is decomposed into the first half peak width and the second half peak width, that is, the second pulse feature includes the first half peak width and the second half peak width. Then, S042: Based on the first array corresponding to the first pulse feature and the second pulse feature of a set type, calculate the mean value corresponding to each frequency position of the second pulse feature and the first pulse feature, and determine the first array corresponding to the second pulse feature, including: based on the first array corresponding to the pulse height of the pulse data and the first half peak width of the pulse data, calculate the mean value corresponding to each frequency position of the first half peak width and the pulse height, and determine the first array corresponding to the first half peak width; and based on the first array corresponding to the pulse height of the pulse data and the second half peak width of the pulse data, calculate the mean value corresponding to each frequency position of the second half peak width and the pulse height, and determine the first array corresponding to the second half peak width.

[0074] Furthermore, in other embodiments according to this application, S042: based on the first array corresponding to the first pulse feature and the second pulse feature of a set type, calculating the mean value corresponding to each frequency position of the second pulse feature and the first pulse feature, and determining the first array corresponding to the second pulse feature, may further include: based on the first array corresponding to the pulse height of the pulse data and the pulse area of ​​the pulse data, calculating the mean value corresponding to each frequency position of the pulse area and the pulse height, and determining the first array corresponding to the pulse area.

[0075] The second array obtained in S06 not only includes pulse height information that contributes to the analysis and identification of biological samples, but also information such as full peak width, first half peak width, second half peak width, or pulse area that can also contribute to the analysis and identification of biological samples, which helps to improve the accuracy of the identification results in S08.

[0076] Please refer to Figure 6As shown, in Embodiment Six provided by this application, before S041: performing frequency statistics on the first pulse feature of a set type carried by the pulse data of each detection channel to obtain a first array of set length values ​​corresponding to the first pulse feature, the biological sample analysis method further includes S03: setting the length value of the first array corresponding to the first pulse feature in the length dimension according to the height range of the pulse height.

[0077] Specifically, in some embodiments, setting the length value of the first array corresponding to the first pulse feature in the length dimension according to the height range of the pulse height includes: selecting the maximum value of the pulse height as the length value of the first array corresponding to the first pulse feature in the length dimension according to the height range of the pulse height.

[0078] For example, in some embodiments, the pulse height range is [0, 4096], with a maximum value of 4096. Therefore, 4069 can be set as the length of the first array. If the number of detection channels is 6, then each first array corresponding to a pulse feature of different settings is a 6×4096 two-dimensional array, meaning each first array has 6 rows and 4096 columns. In Example 4, the number of pulse features of different settings is 2, so in S06, the second array is a 2×6×4096 three-dimensional array, where the first dimension is 4096, the second dimension is 6, and the third dimension is 2. In Example 4, the number of pulse features of different settings is 3, so in S06, the second array is a 3×6×4096 three-dimensional array, where the first dimension is 4096, the second dimension is 6, and the third dimension is 3.

[0079] Please see Figure 7 As shown, in the biological sample analysis method provided according to Embodiment 7 of this application, taking a blood sample as an example, S02: acquiring the pulse data collected by the biological sample of the object to be tested in various detection channels includes: acquiring cell pulse signals collected by multiple detection channels of the biological sample cells of the object to be tested. Each cell pulse signal is an electrical pulse signal converted from the biological characteristics of the corresponding cell.

[0080] Please see Figure 8 As shown, in the biological sample analysis method provided according to Embodiment 8 of this application, S02: acquiring cell pulse signals collected by multiple detection channels of the biological sample cells of the object to be tested, further including S021a and S022a, which are described in detail below:

[0081] S021a: Obtain device information file data for the biological sample analyzer.

[0082] S022a: Based on the device information file data, determine multiple target detection channels for the biological sample cells of the object to be detected, and acquire the cell pulse signals collected by the multiple target detection channels.

[0083] The biosample analyzer refers to an instrument that analyzes the detection data corresponding to the biological sample and outputs corresponding analysis results. The Device Information File (INF) data, by acquiring this data, allows for the determination of channel information, such as the channel number, related to the current analysis task within the biosample analyzer. This identifies multiple target detection channels for the biological sample cells to be tested, and the cell pulse signals collected by each identified detection channel are obtained. Acquiring the biosample analyzer's device information file data includes receiving and parsing the device information file data to determine the target detection channels related to the analysis task based on the parsed device information file data.

[0084] Please refer to Figure 9 As shown, unlike Embodiment 8, in the biological sample analysis method provided according to Embodiment 9 of this application, S02: acquiring pulse data collected from the biological sample of the object to be tested in various detection channels, further includes S021b and S022b, which are described in detail below:

[0085] S021b: Acquire sample detection data output by the biological sample analyzer; the biological sample analyzer is equipped with a combination of two or more detection channels as follows: SFL channel, FSC channel, SSC channel, WBC channel, PLT channel, and RBC channel.

[0086] S022b: Based on the sample detection data, acquire the pulse data collected by the biological sample of the object to be detected in the SFL channel, FSC channel, SSC channel, WBC channel, PLT channel and RBC channel.

[0087] Please see Figure 10 The biological sample analysis method provided in Embodiment 10 of this application includes:

[0088] S021b: Acquire sample detection data output by the biological sample analyzer; the biological sample analyzer is equipped with a combination of two or more detection channels as follows: SFL channel, FSC channel, SSC channel, WBC channel, PLT channel, and RBC channel.

[0089] S022b: Based on the sample detection data, acquire the pulse data collected by the biological sample of the object to be detected in the SFL channel, FSC channel, SSC channel, WBC channel, PLT channel and RBC channel.

[0090] S041: Perform frequency statistics on the first pulse feature of a set type carried by the pulse data of each detection channel to obtain a first array of set length values ​​corresponding to the first pulse feature.

[0091] S0421: Based on the first array corresponding to the pulse height of the pulse data and the first half peak width of the pulse data, calculate the mean value of the first half peak width corresponding to each frequency position of the pulse height, and determine the first array corresponding to the first half peak width.

[0092] S0422: Based on the first array corresponding to the pulse height of the pulse data and the second half peak width of the pulse data, calculate the mean value of the second half peak width corresponding to each frequency position of the pulse height, and determine the first array corresponding to the second half peak width.

[0093] S06: Merge the first array corresponding to the first pulse feature and the first array corresponding to the second pulse feature to obtain a second array of multi-feature fusion.

[0094] S08: Extract features from the second array using a convolutional neural network model, and output the identification results of abnormal target objects in the biological samples carried by the second array.

[0095] S09: Based on the identification result of the abnormal target object, an alarm for the abnormal target object is generated.

[0096] In Embodiment 10, the AI ​​recognition model is a convolutional neural network model such as ResNet or VGG. In other embodiments, the AI ​​recognition model can also be other models constructed using one-dimensional convolution.

[0097] Before inputting the second array into the AI ​​recognition model, the AI ​​recognition model needs to be trained. Training the AI ​​recognition model includes acquiring sample data. The sample data includes a second array of classification labels carrying the characteristics of the target object. The classification labels of the second array include positive sample labels for abnormal target objects (such as IG, bimodal red blood cells, etc.) and negative sample labels for normal target objects.

[0098] In some embodiments, the AI ​​recognition model is used as the target object (IG, immature granulocytes) alarm model, and the detection channels are SFL channel, FSC channel, SSC channel, WBC channel, PLT channel, and RBC channel. The first pulse feature is the pulse height, and the second pulse feature is the first half-peak width and the second half-peak width. The steps for obtaining the second array are as follows:

[0099] S11: Data preprocessing to obtain pulse data of each detection channel of the biological sample.

[0100] S12: Obtain the pulse height of the above 6 channels and count their frequencies. Obtain the first 6x4096 array corresponding to the pulse height;

[0101] S13: Obtain the first half peak width and the second half peak width of the above 6 channels, calculate the average value of the first half peak width and the second half peak width according to the frequency position of the pulse height, and obtain the 6x4096 first array corresponding to the first half peak width and the second half peak width.

[0102] S14: By superimposing the three first arrays mentioned above, a three-dimensional (3x6x4096) second array can be obtained.

[0103] After obtaining the sample data, the positive and negative samples are mixed and divided into training, validation and test sets according to a preset ratio to train the AI ​​recognition model. For example, the mixed positive and negative samples are divided into training, validation and test sets in a ratio of 8:1:1 to train the AI ​​recognition model.

[0104] After the AI ​​recognition model is trained, the second array is input into the trained target object alarm model, which outputs an alarm indicating whether the target object is abnormal. If the output result is a binary classification result, the alarm model will only output alarms for "target object" and "target object is not abnormal".

[0105] As can be seen from the above, the sample analysis method provided in this application acquires pulse data collected from biological samples of the target object in various detection channels, quantifies multiple preset types of pulse features carried by the pulse data of each different detection channel to obtain corresponding first arrays, and then merges the first arrays corresponding to the preset types of pulse features to obtain a second array. An AI recognition model is then used to extract features from the second array, and the recognition result of the target object in the biological sample carried by the second array is output. Since the second array contains multiple preset types of pulse features from the pulse data of each different detection channel, the accuracy of obtaining the recognition result based on the second array as input data for the AI ​​recognition model is high.

[0106] In addition, please see Figure 11 Example 11 provides a biological sample analysis method for obtaining pulse data from the SFL, FSC, SSC, WBC, PLT, and RBC channels, further illustrating the biological sample analysis method provided in this application in detail. In Example 11, the biological sample analysis method includes first acquiring the INF (Device Information File) data of the biological sample analyzer; then parsing the INF data; determining the SFL, FSC, SSC, WBC, PLT, and RBC channels based on the parsed INF data; then acquiring the pulses from the SFL, FSC, SSC, WBC, PLT, and RBC channels; and then acquiring the pulse height and peak width data of all pulse data. A multi-channel fused pulse frequency matrix (second array) is obtained based on the pulse height and peak width data. Finally, an AI recognition model composed of a convolutional neural network is used to extract and recognize features from the multi-channel fused pulse frequency matrix to obtain the corresponding recognition result. The model alarm prediction of the AI ​​recognition model is output, such as an IG alarm or a biphasic red blood cell alarm.

[0107] Please see Figure 13 In some embodiments, this application also provides a biological sample analysis device. The biological sample analysis device includes an acquisition module 101, a quantization module 102, a fusion module 103, and an identification module 104. The acquisition module 101 is used to acquire pulse data collected from the biological sample of the object to be tested in various detection channels; the quantization module 102 is used to quantize multiple preset pulse features carried by the pulse data of each of the different detection channels to obtain corresponding first arrays; the fusion module 103 is used to fuse the first arrays corresponding to the preset pulse features to obtain a second array; the identification module 104 is used to extract features from the second array using an AI identification model and output the identification result of the target object in the biological sample carried by the second array.

[0108] Please see Figure 14 In some embodiments, this application also provides a biological sample analysis device, which includes a memory 202 and a processor 201. When the processor 201 executes computer program instructions stored in the memory 202, it performs the steps of the biological sample analysis method according to any embodiment of this application.

[0109] In addition, this application also provides a computer-readable storage medium storing computer program instructions; when the computer program instructions are executed by a processor, they implement the steps of the biological sample analysis method provided according to any embodiment of this application.

[0110] The aforementioned processor may be a CPU (Central Processing Unit), an ASIC (Application Specific Integrated Circuit), or one or more integrated circuits configured to implement embodiments of the present invention. The moving target detection device includes one or more processors, which may be processors of the same type, such as one or more CPUs; or processors of different types, such as one or more CPUs and one or more ASICs.

[0111] The aforementioned memory may include high-speed RAM (Random Access Memory) and may also include NVM (Non-Volatile Memory), such as at least one disk storage device.

[0112] In addition, please see Figure 15 This is a flowchart illustrating the auxiliary diagnostic method provided in this application. The auxiliary diagnostic method provided in this application determines the abnormal target object in the biological sample based on the identification results obtained by the biological sample analysis method provided in this application and the clinical information of the object to be tested, and outputs the corresponding auxiliary diagnostic information.

[0113] Specifically, such as Figure 15 As shown, the auxiliary diagnostic method includes the following steps:

[0114] S22: Obtain the identification result of the target object in the biological sample obtained by analyzing the biological sample of the object to be tested using the biological sample analysis method provided in any embodiment of this application.

[0115] The auxiliary diagnostic method is applied to the auxiliary diagnostic device provided in this application. The auxiliary diagnostic device, from, for example... Figure 14 The identification results are obtained in the biological sample analysis device.

[0116] S24: Obtain the clinical information of the subject to be tested.

[0117] The clinical information is primarily stored in a Laboratory Information System (LIS), and the auxiliary diagnostic device retrieves this clinical information from the LIS. The clinical information varies depending on the individual being tested, including age, gender, test sample type, and past medical history. In one specific example, the clinical information of the individual being tested includes their age, gender, test sample type, and past medical history.

[0118] S26: Based on the identification results and the clinical information, determine the abnormal target object in the biological sample of the object to be tested, and output auxiliary diagnostic decision information corresponding to the abnormal target object.

[0119] The abnormal target objects in the biological sample refer to the metrics that can be used to characterize whether the subject under test has potential disease risks. Taking blood samples as an example, abnormal target objects in blood samples refer to target objects in the blood sample that can be used to measure whether the subject under test has potential disease risks. The auxiliary diagnostic decision information corresponding to the abnormal target objects refers to the auxiliary diagnostic conclusion determined based on the biological sample of the subject under test, which is formed by comprehensively analyzing the identification results of the target objects in the biological sample and the clinical information of the subject under test. It can include the basis for identifying the abnormal target objects. The basis for the basis can be text or labeled images. The auxiliary diagnostic decision information can reflect the information on which the decision is based, so that the basis for identifying the abnormal target objects in the current test results can be intuitively understood, improving the readability of the auxiliary diagnostic results.

[0120] Furthermore, in some embodiments, the auxiliary diagnostic device can further acquire other analytical results from the biological sample analysis device besides the identification results, such as report parameters obtained based on the analysis of the biological sample. Then, by combining the other analytical results, the identification results, and the clinical information, the device can determine the abnormal target objects in the biological sample data and output a corresponding auxiliary diagnostic report. The auxiliary diagnostic decision information obtained through the auxiliary diagnostic device allows for rapid screening of samples with abnormal target objects in the corresponding test results of the biological sample, facilitating laboratory physicians to make more efficient next-step diagnostic and treatment decisions based on the auxiliary diagnostic decision information.

[0121] The analytical results obtained by the biosample analysis equipment, i.e., biosample analysis data, refer to the corresponding analytical data obtained by detecting various biological samples containing target objects (biological cell information or other biological information). The target object can be the healthy or unhealthy characteristics presented in biological samples corresponding to different cell types, such as abnormal cell types, cell numbers, cell sizes, cell composition ratios, cell contents, and nucleic acid content in the biological sample. The target object refers to the object determined according to the detection requirements of different biological samples. For example, in the case of cell analysis data, the target object refers to one or several cells determined according to the detection requirements of different biological samples. For instance, in the PLT histogram of cell analysis data for blood samples, the target cell refers to the distribution characteristics of platelet numbers of different sizes. The biosample analysis data also includes sample data of other biological information. The target object refers to information characterizing the characteristics of the corresponding biological information, such as information characterizing whether there are abnormal peaks in the CRP reaction curve output by the immunoassay analyzer, whether there is blockage in the impedance detection channel of the blood cell analyzer, or whether immature granulocytes are found in the DIFF detection channel of the blood cell analyzer.

[0122] The auxiliary diagnostic method provided in this application obtains the identification result of the target object based on the sample analysis method, and then determines the abnormal target object in the biological sample by combining the clinical information of the subject to be tested. It outputs auxiliary diagnostic decision information corresponding to the abnormal object. The auxiliary diagnostic decision information can more intuitively understand whether there is an abnormal target object in the biological sample of the subject to be tested. When it is determined that there is an abnormal target object, the decision basis can be known through the auxiliary diagnostic decision information. Accurate auxiliary diagnostic results can be obtained without relying on the personal experience level of the laboratory physician. On the one hand, it can improve the efficiency of testing and make the test results more accurate. On the other hand, the diagnostic conclusions are more interpretable. It transforms the numerical test reports in the biological sample analysis equipment and laboratory information system into descriptive auxiliary diagnostic decision information, helps laboratory physicians reduce the workload of screening test data, and can also better match the results of the auxiliary diagnostic report with the hierarchical medical treatment policy.

[0123] In some embodiments, determining the abnormal features of the sample to be detected of the object to be detected and outputting auxiliary diagnostic decision information corresponding to the abnormal features includes one of the following:

[0124] Determine the abnormal features of the test sample of the object to be detected, correct the parameter alarm state formed by the identification result based on the target cell features according to the abnormal features, and output the corrected auxiliary diagnostic decision information corresponding to the abnormal features;

[0125] Determine the abnormal characteristics of the test sample of the object to be tested, and output auxiliary diagnostic decision information containing the analysis of the non-compliance causes corresponding to the abnormal characteristics;

[0126] Identify the abnormal characteristics of the test sample of the object to be tested, determine the supporting evidence and / or treatment recommendations corresponding to the abnormal characteristics, and output an auxiliary diagnostic report containing the supporting evidence and / or treatment recommendations.

[0127] Biosample analyzers analyze samples and output biological sample image data, parameter reports, alarm parameters, etc. For example, cell analyzers analyze samples and output cell analysis data, blood parameter reports, and alarm parameters. Auxiliary diagnostic devices can use auxiliary diagnostic knowledge graphs to integrate the analysis results from biosample analyzers and the clinical information of the tested object in the laboratory information system as inputs for auxiliary diagnostic solutions. This simulates the auxiliary diagnostic decisions made by laboratory physicians based on test results data and outputs the key information upon which these decisions are based. This reduces reliance on the individual experience level of laboratory physicians, decreases their workload, and makes auxiliary diagnostic decisions more readable, accurate, and efficient. One optional approach to auxiliary diagnostic decision information is to correct the parameter alarm status formed by the identification results based on the target features according to the abnormal characteristics, and output the corrected auxiliary diagnostic decision information corresponding to the abnormal characteristics. The key information contained in the auxiliary diagnostic decision information highlights the basis for correcting the parameter alarm status. Thus, by integrating multiple data sources to make auxiliary diagnostic decisions, the bias of a single data source is avoided. Another option for auxiliary diagnostic decision-making information is to include an analysis of the reasons for non-compliance corresponding to the abnormal characteristics. From the key information included in the auxiliary diagnostic decision-making information, the decision-making logic for non-compliant specimens can be highlighted. This allows for the differentiation of non-compliance reasons such as sample abnormalities (chylous blood, hemolysis, cold agglutination of red blood cells, etc.) and bioanalytical instrument abnormalities (clogging, voltage instability, etc.). Yet another option for auxiliary diagnostic decision-making information is to determine supporting evidence and / or treatment recommendations corresponding to the abnormal characteristics, and output an auxiliary diagnostic report containing such supporting evidence and / or treatment recommendations, forming the auxiliary diagnostic report based on the key information used to make the auxiliary diagnostic decision.

[0128] In the above embodiments, the auxiliary diagnostic device, in conjunction with the biosample analyzer and the testing information system, integrates the biosample analysis data, the identification results, and the clinical information to make auxiliary diagnostic decisions. Based on the key information on which the auxiliary diagnostic decisions are made, different forms of auxiliary diagnostic decision information are formed, resulting in more readable auxiliary diagnostic results and easier and faster judgment of the reliability of the results.

[0129] In some embodiments, determining the abnormal target object of the biological sample of the object to be detected and outputting auxiliary diagnostic decision information corresponding to the abnormal target object includes:

[0130] Identify the abnormal target objects in the biological sample of the object to be tested, determine the corresponding diagnostic and treatment recommendations, and output an auxiliary diagnostic report containing a textual description of the diagnostic and treatment recommendations; or,

[0131] Identify the abnormal target object of the biological sample of the object to be tested, determine the supporting evidence corresponding to the abnormal target object, and output an auxiliary diagnostic report containing textual descriptions, labeled image data, and numerical descriptions of the supporting evidence.

[0132] Taking blood samples as an example, the abnormal targets in the biological samples being tested can be correlated with the diagnostic results of historical cases in the diagnostic case database, such as low platelet count or increased white blood cell count with immature granulocytes. Treatment recommendations for low platelet count could include testing for autoantibodies and platelet-related immature granulocytes to clarify the diagnosis. Treatment recommendations for increased white blood cell count with immature granulocytes could include recommending bone marrow cytology if necessary. The auxiliary diagnostic report provides a textual description of the abnormal targets in the biological samples and their corresponding treatment recommendations, making the auxiliary diagnostic results more readable. Supporting evidence corresponding to the abnormal targets refers to determining the cause of the abnormal characteristics. For example, supporting evidence for low platelet count can include a comparison of the platelet count value with the corresponding reference value; supporting evidence for increased white blood cell count with immature granulocytes can include a comparison of the white blood cell count with the corresponding reference value, a comparison of the immature granulocyte count with the corresponding reference value, a comparison of the immature granulocyte ratio with the corresponding reference value, and a comparison of a scatter plot of patients with abnormal sites with a reference scatter plot confirming the presence of immature granulocytes under microscopic examination. The auxiliary diagnostic report provides textual descriptions of the abnormal characteristics of the blood sample and their corresponding supporting evidence, along with labeled image data and numerical descriptions. This allows laboratory physicians to efficiently and accurately understand the specific reasons for the abnormalities in the current test results, grasp the basis for the conclusions in the auxiliary diagnostic report, and make decisions accordingly.

[0133] In the above embodiments, the auxiliary diagnostic report can provide the basis for identifying abnormal target objects in the biological sample test results of the object to be tested. Furthermore, by describing the auxiliary diagnostic conclusions and their formation through text descriptions or a combination of text and images, it can also provide explanations of supporting evidence for the auxiliary diagnostic conclusions in the auxiliary diagnostic report, thereby enabling the output of more readable auxiliary diagnostic results.

[0134] In some embodiments, determining the abnormal target object of the biological sample of the object to be detected and outputting an auxiliary diagnostic report containing a description of the abnormal target object includes:

[0135] Get instructions on whether to obtain a simplified or detailed diagnostic report;

[0136] Based on the selection instruction for the simplified diagnostic report, an auxiliary diagnostic report containing a textual description of the cause of the abnormal target object is output; or based on the selection instruction for the detailed diagnostic report, an auxiliary diagnostic report containing a textual description of the supporting evidence, labeled image data, and numerical description is output.

[0137] The auxiliary diagnostic device provides a selection button for the type of auxiliary diagnostic report through the application interface. Users can click the selection button to choose between a simplified diagnostic report and a detailed diagnostic report. Based on the user's selection of the auxiliary diagnostic report type, if the user chooses to output a simplified diagnostic report, the device identifies the abnormal target in the blood sample of the subject to be tested, determines the corresponding treatment recommendations, and outputs an auxiliary diagnostic report containing a textual description of the treatment recommendations. If the user chooses to output a detailed diagnostic report, the device identifies the abnormal target in the blood sample of the subject to be tested, determines the supporting evidence corresponding to the abnormal characteristics, and outputs an auxiliary diagnostic report containing a textual description of the supporting evidence, labeled image data, and numerical descriptions.

[0138] In the above embodiments, users can choose the type of auxiliary diagnostic report they receive, obtain a simplified diagnostic report, quickly browse the current auxiliary diagnostic conclusions, and choose to obtain a detailed diagnostic report if they need to understand the specific basis for making the corresponding auxiliary diagnostic conclusions, so as to meet more personalized usage needs.

[0139] In some embodiments, S26: Based on the identification result and the clinical information, determine the abnormal target object in the biological sample of the object to be tested, and output auxiliary diagnostic decision information corresponding to the abnormal target object, further including: based on the biological sample analysis data corresponding to the biological sample, the identification result, and the clinical information, determine the abnormal target object in the biological sample of the object to be tested, and output auxiliary diagnostic decision information corresponding to the abnormal target object. The specific steps of this scheme are as follows:

[0140] The report parameters, the identification results, and the clinical information in the biological sample analysis data are input into the auxiliary diagnostic model;

[0141] The auxiliary diagnostic model determines the abnormal feature type of the sample of the object to be tested based on the report parameters, the identification results, and the clinical information, and outputs an auxiliary diagnostic report corresponding to the abnormal feature type.

[0142] The sample to be tested is a blood sample. An auxiliary diagnostic model is established, using blood report parameters from cell analysis data, the identification results, and the clinical information of the subject as inputs. The model learns features from a large amount of multimodal data, analogous to the brain of a professional laboratory physician, exhibiting no forgetting curve, and its output results can be progressively optimized as the amount of data increases. Optionally, the auxiliary diagnostic model can be obtained by training a neural network model or by constructing an auxiliary diagnostic knowledge graph.

[0143] In some embodiments, this application also provides an auxiliary diagnostic device, the auxiliary diagnostic device including a processor and a memory, the memory storing a computer program executable by the processor, the computer program implementing the auxiliary diagnostic method when executed by the processor.

[0144] In addition, this application also provides another computer-readable storage medium storing computer program instructions, which, when executed by a processor, implement the steps of the auxiliary diagnostic method provided according to any embodiment of this application.

[0145] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the technical scope disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

Claims

1. A method for analyzing biological samples, characterized in that, include: Acquire pulse data from biological samples of the object to be tested in different detection channels; The pulse features of multiple preset types carried by the pulse data of each of the different detection channels are quantized to obtain the corresponding first arrays; The first array corresponding to the pulse features of the specified type is fused to obtain the second array; Feature extraction is performed on the second array, and the identification results of the target objects in the biological samples contained in the second array are output; The step of quantizing the pulse features of multiple predetermined types carried by the pulse data of each of the different detection channels to obtain corresponding first arrays includes: Frequency statistics are performed on the first pulse feature of a set type carried by the pulse data of each detection channel to obtain a first array of set length values ​​corresponding to the first pulse feature; Based on the first array corresponding to the first pulse feature and the second pulse feature of a set type, calculate the mean value corresponding to each frequency position of the second pulse feature and the first pulse feature, and determine the first array corresponding to the second pulse feature.

2. The biological sample analysis method according to claim 1, characterized in that, The step of extracting features from the second array and outputting the identification result of the target object in the biological sample carried by the second array includes: The second array is used to extract features using an AI recognition model, and the recognition results of the target objects in the biological samples contained in the second array are output; or, The second array is used to extract features using a traditional algorithm model, and the identification results of the target objects in the biological samples contained in the second array are output.

3. The biological sample analysis method according to claim 1, characterized in that, The step of quantizing the pulse features of multiple preset types carried by the pulse data of each of the different detection channels to obtain corresponding first arrays includes: The pulse data of each of the different detection channels are quantized to carry multiple preset types of pulse features to obtain each two-dimensional first array corresponding to each detection channel. The first dimension value of each two-dimensional first array in the first dimension direction is determined according to the value range of multiple preset types of pulse features, and the second dimension value of each two-dimensional first array in the second dimension direction is determined according to the number of channels of the detection channel. The step of fusing the first array corresponding to the pulse features of the specified type to obtain the second array includes: The two-dimensional first arrays corresponding to the pulse features of each set type are superimposed and fused in the third dimension to obtain a three-dimensional second array of multi-feature fusion. The third dimension value of the three-dimensional second array in the third dimension direction is determined according to the number of features of the multiple set types of pulse features.

4. The biological sample analysis method according to claim 1, characterized in that, The step of fusing the first array corresponding to the pulse features of the specified type to obtain the second array includes: The first array corresponding to the first pulse feature and the first array corresponding to the second pulse feature are merged to obtain the second array.

5. The biological sample analysis method according to claim 1, characterized in that, The step of performing frequency statistics on the first pulse feature of a predetermined type carried by the pulse data of each of the detection channels to obtain a first array corresponding to the first pulse feature includes: The frequency of the pulse height carried by the pulse data of each detection channel is counted to obtain the first array corresponding to the pulse height.

6. The biological sample analysis method according to claim 5, characterized in that, The step of calculating the mean value of each frequency position of the second pulse feature corresponding to the first pulse feature and the second pulse feature of a set type based on the first array corresponding to the first pulse feature and determining the first array corresponding to the second pulse feature includes at least one of the following: Based on the first array corresponding to the pulse height of the pulse data and the full peak width of the pulse data, calculate the mean value of the full peak width corresponding to each frequency position of the pulse height, and determine the first array corresponding to the full peak width; Based on the first array corresponding to the pulse height of the pulse data and the first half peak width of the pulse data, calculate the mean value of the first half peak width corresponding to each frequency position of the pulse height, and determine the first array corresponding to the first half peak width; Based on the first array corresponding to the pulse height of the pulse data and the second half peak width of the pulse data, calculate the mean value of the second half peak width corresponding to each frequency position of the pulse height, and determine the first array corresponding to the second half peak width; Based on the first array corresponding to the pulse height of the pulse data and the pulse area of ​​the pulse data, the mean value of the pulse area corresponding to each frequency position of the pulse height is calculated, and the first array corresponding to the pulse area is determined.

7. The biological sample analysis method according to claim 5, characterized in that, Before performing frequency statistics on the pulse heights carried by the pulse data of each of the detection channels to obtain the first array corresponding to the pulse heights, the biological sample analysis method further includes: Based on the height range of the pulse height, the length value of the first array corresponding to the first pulse feature in the length dimension is set; The step of performing frequency counting on the pulse heights carried by the pulse data of each of the detection channels to obtain a first array corresponding to the pulse height includes: The frequency of the pulse height carried by the pulse data of each detection channel is counted, and a first array corresponding to the pulse height is obtained based on the result of the frequency count and the length value.

8. The biological sample analysis method according to claim 7, characterized in that, The step of setting the length value of the first array corresponding to the first pulse feature in the length dimension according to the height range of the pulse height includes: Based on the height range of the pulse height, the maximum value of the pulse height is selected as the length value of the first array corresponding to the first pulse feature in the length dimension.

9. The biological sample analysis method according to claim 1, characterized in that, The acquisition of pulse data from biological samples of the object to be tested collected in various detection channels includes: Cell pulse signals collected from multiple detection channels of the biological sample cells of the object to be tested are obtained.

10. The biological sample analysis method according to claim 9, wherein acquiring the cell pulse signals collected by multiple detection channels of the biological sample cells of the object to be tested includes: Obtain device information files from the biological sample analyzer; Based on the device information file data, multiple target detection channels for the biological sample cells of the object to be detected are determined, and cell pulse signals collected by the multiple target detection channels are acquired.

11. The biological sample analysis method according to claim 1, characterized in that, The acquisition of pulse data from biological samples of the object to be tested collected in various detection channels includes: Acquire sample detection data output by a biosample analyzer; the biosample analyzer is equipped with a combination of two or more detection channels as follows: SFL detection channel, FSC detection channel, SSC detection channel, WBC detection channel, PLT detection channel, and RBC detection channel; Based on the sample detection data, pulse data of the biological sample of the object to be detected are acquired in the SFL channel, FSC channel, SSC channel, WBC channel, PLT channel, and RBC channel.

12. The biological sample analysis method according to any one of claims 1, characterized in that, After performing feature extraction on the second array and outputting the identification results of the target object in the biological sample carried by the second array, the process includes: Based on the identification results of the target object, a target object is formed.

13. The biological sample analysis method according to any one of claims 2, characterized in that, The step of extracting features from the second array using an AI recognition model and outputting the recognition result of the target object in the biological sample carried by the second array includes: The second array is used to extract features using an AI recognition model built on a one-dimensional convolutional approach, and the recognition result of the target object in the biological sample contained in the second array is output; or The second array is used to extract features by a convolutional neural network model, and the identification results of the target objects in the biological samples contained in the second array are output.

14. An auxiliary diagnostic method, characterized in that, include: To obtain the identification result of the target object in the biological sample obtained by analyzing the biological sample of the object to be tested using the biological sample analysis method as described in any one of claims 1 to 13; Obtain the clinical information of the subject to be tested; Based on the identification results and the clinical information, abnormal target objects in the biological sample of the object to be tested are identified, and auxiliary diagnostic decision information corresponding to the abnormal target objects is output.

15. A biological sample analysis device, characterized in that, include: The acquisition module is used to acquire pulse data collected from biological samples of the object to be tested in different detection channels; The quantization module is used to quantize the pulse features of multiple preset types carried by the pulse data of each of the different detection channels to obtain the corresponding first arrays. The fusion module is used to fuse the first array corresponding to the pulse features of the specified type to obtain a second array; The identification module is used to extract features from the second array and output the identification result of the target object in the biological sample carried by the second array; The quantization module is used to perform frequency statistics on the first pulse feature of a set type carried by the pulse data of each detection channel to obtain a first array of set length values ​​corresponding to the first pulse feature; based on the first array corresponding to the first pulse feature and the second pulse feature of a set type, it calculates the mean value corresponding to each frequency position of the second pulse feature and the first pulse feature to determine the first array corresponding to the second pulse feature.

16. A biological sample analysis device, characterized in that, It includes a processor and a memory, wherein the memory stores a computer program that can be executed by the processor, and the computer program, when executed by the processor, implements the biological sample analysis method as described in any one of claims 1 to 13.

17. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program that, when executed by a controller, implements the biological sample analysis method as described in any one of claims 1 to 13.

18. An auxiliary diagnostic device, characterized in that, It includes a processor and a memory, wherein the memory stores a computer program that can be executed by the processor, and the computer program, when executed by the processor, implements the auxiliary diagnostic method as described in claim 14.

19. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program that, when executed by the controller, implements the auxiliary diagnostic method as described in claim 14.

Images