CN114088605A - Blood cell classification method and related equipment - Google Patents
Blood cell classification method and related equipment Download PDFInfo
- Publication number
- CN114088605A CN114088605A CN202010744431.4A CN202010744431A CN114088605A CN 114088605 A CN114088605 A CN 114088605A CN 202010744431 A CN202010744431 A CN 202010744431A CN 114088605 A CN114088605 A CN 114088605A
- Authority
- CN
- China
- Prior art keywords
- pulse signal
- blood
- target
- neural network
- classification
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 210000000601 blood cell Anatomy 0.000 title claims abstract description 120
- 238000000034 method Methods 0.000 title claims abstract description 56
- 210000004369 blood Anatomy 0.000 claims abstract description 107
- 239000008280 blood Substances 0.000 claims abstract description 107
- 238000013145 classification model Methods 0.000 claims abstract description 101
- 238000013528 artificial neural network Methods 0.000 claims abstract description 99
- 238000012549 training Methods 0.000 claims abstract description 76
- 238000003860 storage Methods 0.000 claims abstract description 9
- 238000001514 detection method Methods 0.000 claims description 36
- 238000004422 calculation algorithm Methods 0.000 claims description 24
- 238000004590 computer program Methods 0.000 claims description 16
- 210000001772 blood platelet Anatomy 0.000 claims description 11
- 238000009826 distribution Methods 0.000 claims description 9
- 210000003743 erythrocyte Anatomy 0.000 claims description 9
- 238000010801 machine learning Methods 0.000 claims description 8
- 230000003993 interaction Effects 0.000 claims description 6
- 238000004364 calculation method Methods 0.000 claims description 3
- 230000000877 morphologic effect Effects 0.000 abstract description 14
- 238000000605 extraction Methods 0.000 abstract description 4
- 239000000523 sample Substances 0.000 description 55
- 238000012545 processing Methods 0.000 description 17
- 230000008569 process Effects 0.000 description 13
- 238000005070 sampling Methods 0.000 description 12
- 210000004027 cell Anatomy 0.000 description 9
- 230000006870 function Effects 0.000 description 9
- 238000004891 communication Methods 0.000 description 7
- 238000003062 neural network model Methods 0.000 description 7
- 239000003153 chemical reaction reagent Substances 0.000 description 6
- 238000010586 diagram Methods 0.000 description 6
- 238000004458 analytical method Methods 0.000 description 5
- 230000008901 benefit Effects 0.000 description 5
- 238000012986 modification Methods 0.000 description 4
- 230000004048 modification Effects 0.000 description 4
- 239000003085 diluting agent Substances 0.000 description 3
- 238000002360 preparation method Methods 0.000 description 3
- 230000000630 rising effect Effects 0.000 description 3
- 230000008859 change Effects 0.000 description 2
- 238000007635 classification algorithm Methods 0.000 description 2
- 239000004020 conductor Substances 0.000 description 2
- 230000014509 gene expression Effects 0.000 description 2
- 239000007788 liquid Substances 0.000 description 2
- 238000013178 mathematical model Methods 0.000 description 2
- 230000003287 optical effect Effects 0.000 description 2
- 239000003219 hemolytic agent Substances 0.000 description 1
- 239000012535 impurity Substances 0.000 description 1
- 210000000265 leukocyte Anatomy 0.000 description 1
- 239000004973 liquid crystal related substance Substances 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
- 244000005700 microbiome Species 0.000 description 1
- 210000002569 neuron Anatomy 0.000 description 1
- 239000012488 sample solution Substances 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N15/00—Investigating characteristics of particles; Investigating permeability, pore-volume or surface-area of porous materials
- G01N15/10—Investigating individual particles
- G01N15/1031—Investigating individual particles by measuring electrical or magnetic effects
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N15/00—Investigating characteristics of particles; Investigating permeability, pore-volume or surface-area of porous materials
- G01N15/06—Investigating concentration of particle suspensions
- G01N15/0656—Investigating concentration of particle suspensions using electric, e.g. electrostatic methods or magnetic methods
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N33/00—Investigating or analysing materials by specific methods not covered by groups G01N1/00 - G01N31/00
- G01N33/48—Biological material, e.g. blood, urine; Haemocytometers
- G01N33/483—Physical analysis of biological material
- G01N33/487—Physical analysis of biological material of liquid biological material
- G01N33/48707—Physical analysis of biological material of liquid biological material by electrical means
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N15/00—Investigating characteristics of particles; Investigating permeability, pore-volume or surface-area of porous materials
- G01N15/10—Investigating individual particles
- G01N2015/1028—Sorting particles
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2218/00—Aspects of pattern recognition specially adapted for signal processing
- G06F2218/12—Classification; Matching
Landscapes
- Engineering & Computer Science (AREA)
- Life Sciences & Earth Sciences (AREA)
- Health & Medical Sciences (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Chemical & Material Sciences (AREA)
- Theoretical Computer Science (AREA)
- General Health & Medical Sciences (AREA)
- Biomedical Technology (AREA)
- Data Mining & Analysis (AREA)
- Analytical Chemistry (AREA)
- Artificial Intelligence (AREA)
- Evolutionary Computation (AREA)
- Pathology (AREA)
- Immunology (AREA)
- Biophysics (AREA)
- General Engineering & Computer Science (AREA)
- Biochemistry (AREA)
- Molecular Biology (AREA)
- Computational Linguistics (AREA)
- Computing Systems (AREA)
- Software Systems (AREA)
- Mathematical Physics (AREA)
- Dispersion Chemistry (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Bioinformatics & Computational Biology (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Evolutionary Biology (AREA)
- Hematology (AREA)
- Urology & Nephrology (AREA)
- Food Science & Technology (AREA)
- Medicinal Chemistry (AREA)
- Investigating Or Analyzing Materials By The Use Of Electric Means (AREA)
Abstract
本申请提供了血细胞分类方法，将待分类血液样本的阻抗脉冲信号输入到训练完成的目标神经网络分类模型中，目标神经网络分类模型是由大量具有分类结果标签信息的阻抗脉冲信号样本训练而成，其具有较强的脉冲形态特征提取能力，从而可以根据待分类血液样本的阻抗脉冲信号的形态特征确定阻抗脉冲信号的分类结果，相比现有技术而言分类结果准确度更高，并且其可以识别出干扰因素导致的阻抗脉冲信号，以及可以基于阻抗脉冲信号进行识别分类，实时性也更高。另外，本申请还提供了血细胞分析仪以及存储介质。
The present application provides a blood cell classification method. The impedance pulse signal of the blood sample to be classified is input into the target neural network classification model after training. The target neural network classification model is trained from a large number of impedance pulse signal samples with label information of the classification result. , which has a strong pulse morphological feature extraction capability, so that the classification result of the impedance pulse signal can be determined according to the morphological feature of the impedance pulse signal of the blood sample to be classified. Compared with the prior art, the classification result is more accurate, and its The impedance pulse signal caused by the interference factor can be identified, and the identification and classification can be performed based on the impedance pulse signal, and the real-time performance is also higher. In addition, the present application also provides a blood cell analyzer and a storage medium.
Description
Technical Field
The application relates to the technical field of medical equipment, in particular to a blood cell classification method and related equipment.
Background
A blood cell analyzer is a commonly used medical device that can detect information about blood cells in blood, such as the types of blood cells and the number of each blood cell. At present, the blood cell analyzer for three-classification or five-classification mainly adopts a small hole impedance method for detecting blood cells.
The main process of the detection method is that blood cells to be detected pass through an impedance detection channel shaped like a small hole one by one, conductive diluent is filled in the channel, and the blood cells are bad conductors and can cause the resistance change at two ends of the channel when passing through the small hole channel, so that a blood cell analyzer converts the resistance change into pulse signals, effective pulse signals are identified by a pulse identification algorithm, and the amplitude and pulse width information of the effective pulses are analyzed to determine the blood cell types included in the blood to be detected. However, the classification accuracy of this method is low.
Disclosure of Invention
In view of the above, the present application provides a blood cell analysis method for improving the classification accuracy of blood cells. In order to achieve the above object, the present application provides the following technical solutions:
in a first aspect, the embodiments of the present application provide a cell classification method, including:
obtaining a target impedance pulse signal of a blood sample to be classified;
obtaining a pre-trained target neural network classification model, wherein the target neural network classification model is obtained by training a plurality of groups of blood training samples through a machine learning algorithm, and each group of blood training samples comprises impedance pulse signals and label information used for representing classification results of the impedance pulse signals;
inputting the target impedance pulse signal into the target neural network classification model to obtain a classification result of the target impedance pulse signal output by the target neural network classification model, wherein the classification result comprises blood cell types or interference.
In a second aspect, embodiments of the present application provide a blood analyzer, including:
a sample collection module for obtaining a blood sample;
a detection module for obtaining a target impedance pulse signal of the blood sample;
a processor configured to obtain at least a pre-trained target neural network classification model, wherein the target neural network classification model is obtained by training a plurality of groups of blood training samples by a machine learning algorithm, and each group of the blood training samples comprises impedance pulse signals of blood cells and label information used for representing classification results of the impedance pulse signals; inputting the target impedance pulse signal into the target neural network classification model to obtain a classification result of the target impedance pulse signal output by the target neural network classification model, wherein the classification result comprises blood cell types or interference;
a human-computer interaction module configured to output the classification result.
In a third aspect, the present embodiments provide a readable storage medium, on which a computer program is stored, which when loaded and executed by a processor, implements the blood cell sorting method according to any one of the first aspect.
According to the blood cell classification method provided by the embodiment of the application, the impedance pulse signals of the blood samples to be classified are input into the trained target neural network classification model, the target neural network classification model is formed by training a large number of impedance pulse signal samples with classification result label information, and the target neural network classification model has strong pulse morphological feature extraction capacity, so that the classification results of the impedance pulse signals can be determined according to the morphological features of the impedance pulse signals of the blood samples to be classified, and compared with the prior art, the classification result accuracy is higher. In addition, the impedance pulse signals caused by the interference factors can be identified, the identification and classification can be carried out on the basis of the impedance pulse signals, and the real-time performance is higher.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the embodiments or the prior art will be briefly described below, it is obvious that the drawings in the following description are only some embodiments of the present invention, and for those skilled in the art, other drawings can be obtained according to the drawings without creative efforts.
FIG. 1 is a schematic flow chart of a blood cell sorting method according to an embodiment of the present disclosure;
FIG. 2 is a diagram illustrating an exemplary method for classifying blood cells according to an embodiment of the present disclosure;
FIGS. 3a-3b are schematic diagrams illustrating two application scenarios of the blood cell classification method according to the embodiment of the present application;
FIG. 4 is a schematic flow chart illustrating a process for training a target neural network classification model according to an embodiment of the present application;
FIGS. 5a-5c are exemplary graphs of three impedance detection signals of a blood training sample provided by an embodiment of the present application;
FIG. 6 is a schematic structural diagram of a blood cell analyzer according to an embodiment of the present application;
FIG. 7 is a schematic view of another structure of a blood cell analyzer according to an embodiment of the present disclosure;
fig. 8 is a schematic structural diagram of a control module of a blood cell analyzer according to an embodiment of the present disclosure.
Detailed Description
The technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the drawings in the embodiments of the present application, and it is obvious that the described embodiments are only a part of the embodiments of the present application, and not all of the embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present application.
Human blood contains various cells such as red blood cells, white blood cells, and platelets, and a blood cell analyzer can detect the types of blood cells contained in a blood sample. At present, a blood cell analyzer, such as a three-class blood cell analyzer or a five-class blood cell analyzer, is usually used for classifying cell types in blood by a small-hole impedance method. The method is that a blood sample passes through a small hole with two electrodes, because blood cells are poor conductors, resistance changes at two ends can be caused when the blood cells pass through, and a detection circuit converts a resistance signal into a voltage signal so as to generate a pulse signal. One blood cell represents one pulse signal.
Pulse recognition algorithms are currently used to recognize pulse signals in order to determine the classification result of the pulse signals. Specifically, effective pulse signals are determined according to the rising and falling trends of the pulse signals within a period of time, then the amplitude and the pulse width of the effective pulse signals are calculated, and the blood cell types corresponding to the amplitude and the pulse width of the effective pulse signals are determined according to the set range because different blood cell types are provided with different amplitude and pulse width range values, so that the blood cell classification is realized.
The inventor of the present application has studied the above method and found that the prior art has at least the following problems:
firstly, the identification of effective pulse signals only depends on the rising and falling trends of signals, however, the detection environment of the orifice impedance method is complex, the bubbles in the diluent flow in the orifice channel, the electric signals received by the detection circuit and other interference factors can generate pulse signals with a high probability, and if the rising and falling trends of the pulse signals meet the identification standard, the pulse signals can be identified as effective pulse signals, so that the accuracy of classification results is influenced.
Secondly, the classification standard of the effective pulse signal is artificially set to be two items of information of pulse amplitude and pulse width. However, the two items of information are not comprehensive enough, and the classification of the pulse signals is not accurate enough. The pulse signals formed by many interference factors may also meet the classification criteria, resulting in an inability to distinguish the interference pulse signals from the blood cell pulse signals. In addition, some blood cells in a critical state may not be classified accurately, and the accuracy of the classification result is low.
In order to solve at least one of the above technical problems, the inventors of the present application propose a blood cell classification method, specifically, a trained neural network model is used to classify blood cell samples. The neural network model has stronger capability of feature extraction and feature analysis, so that more accurate classification results can be obtained. The method of the present application is further described below by way of specific embodiments in conjunction with the accompanying drawings.
Referring to fig. 1, a flow diagram of an embodiment of a blood cell sorting method is shown, which specifically includes steps 101-103.
S101: a target impedance pulse signal of a blood sample to be classified is obtained.
Wherein a blood sample to be classified is processed by a reagent containing a hemolytic agent and then sent to an impedance channel of a blood analyzer. And an impedance detection circuit of the impedance channel detects the blood sample to be classified according to the small-hole impedance method so as to obtain an impedance pulse signal of the blood sample to be classified. One specific example of an impedance pulse signal can be seen in fig. 2, which includes a plurality of impedance pulse signals.
It should be noted that, for the convenience of distinguishing from other impedance pulse signals in the following, the impedance pulse signal of the blood sample to be classified is referred to as a target impedance pulse signal. The impedance pulse signal is not limited to that obtained by the pinhole impedance method, and any method may be used as long as the impedance pulse signal of the blood sample can be detected.
Since the impedance pulse signal may be generated by interference factors, and in addition, impedance pulse signals of different expressions may correspond to the same or different types of blood cells, accurate identification by a neural network model is required.
S102: and obtaining a pre-trained target neural network classification model, wherein the target neural network classification model is obtained by training a plurality of groups of blood training samples through a machine learning algorithm, and each group of blood training samples comprise impedance pulse signals of blood cells and label information used for representing classification results of the impedance pulse signals.
The blood training sample for training comprises impedance pulse signals of blood cells and interference factors, each impedance pulse signal has corresponding label information, and the label information is used for representing the classification result of the impedance pulse signals, namely the impedance pulse signals are generated by which type of blood cells. In one embodiment, the label information may include Interference (Interference) and blood cell types such as Red Blood Cells (RBCs), Platelets (PLTs), etc. Note that, if the tag information is interference, it indicates that the impedance pulse signal is caused by an interference factor. The target neural network classification model is trained by the initial neural network model in advance, and the specific training mode can be referred to the following description.
S103: and inputting the target impedance pulse signal into the target neural network classification model to obtain a classification result of the target impedance pulse signal output by the target neural network classification model, wherein the classification result comprises blood cell types and interference.
The target impedance pulse signals are input into a target neural network classification model, and the model performs classification identification on the target impedance pulse signals to determine a classification result of each impedance pulse signal. The target neural network classification model identifies each target impedance pulse signal.
It can be understood that, the label information of the blood training sample includes which types, and the classification result output by the finally trained target neural network classification model includes which types. In one embodiment, the classification result of the target impedance pulse signal includes: interference and blood cell types such as red blood cells, platelets, etc.
The disturbance may be caused by various reasons, such as disturbance of the impedance detection circuit by an external electrical signal, failure of components of the impedance detection circuit itself, disturbance of microorganisms in the detection channel, impurities incorporated into the blood sample to be classified during sampling, bubbles contained in the diluent stream in the detection channel, and the like. The interference factors are more and the expression forms of the impedance pulse signals of the interference factors generated by different reasons may be different, and the existing identification algorithm may identify some interference factors as blood cell types, thereby affecting the accuracy of blood cell classification. However, the blood training samples of the target neural network classification model take the existence of the interference factors into consideration, and the impedance pulse signals caused by the interference factors can be identified, so that the accuracy of the classification result is improved.
In addition, the impedance pulse signals have pulse form characteristics, the pulse form characteristics of different types of impedance pulse signals are different, and the target neural network classification model obtains the classification result of the target impedance pulse signals by analyzing the pulse form characteristics of the target impedance pulse signals. In one embodiment, the pulse morphology features include a combination of any of: pulse position, pulse peak value, pulse width, pulse peak-to-width ratio, pulse rise time, pulse fall time, pulse line slope, pulse baseline information, total pulse area information, difference in area between and before the pulse peak value, and noise information.
Pulse position (Pos), representing the position of the target impedance pulse signal in the complete detection signal of the blood sample to be classified, 10s for 10e for impedance detection signal acquisition, e.g. at a sampling rate of 1MHz6A sampling point, the peak value of a target impedance pulse signal is located at 1e6At each sampling point, its position is 1e6/10e60.1. A pulse Peak value (Peak) indicating a Peak magnitude of the target impedance pulse signal; taking an impedance pulse signal collected by a 12-bit impedance detection circuit as an example, the peak value is between 0 and 4095. The pulse Width (Width), which represents the Width of the target impedance pulse signal, may be in units of the number of sampling points. The pulse aspect Ratio (Ratio) represents a Ratio of a peak value divided by a width of the target impedance pulse signal. The pulse rise time (R-time) represents the time information that the target impedance pulse signal reaches the peak value from a steady rise. And the pulse falling time (D-time) represents the information of the time when the target impedance pulse signal falls from the peak value to the stable state. The pulse line slope (K) represents the slope of the peak of the target impedance pulse signal. Pulse baseline information (Base), which indicates the magnitude of the baseline taken by the target impedance pulse signal at the time of recognition. Total pulse Area information (Area), which represents the integral of the difference between the target impedance pulse signal and the baseline. Area difference (AreaDiff) between front and rear of pulse peak, first half pulse representing peak position of target impedance pulse signalThe difference between the area and the area of the second half pulse, and the ratio of the total area. Noise information (Noise) indicating an effective value of Noise of a data segment of the target-impedance-free pulse signal. It should be noted that the above pulse shape features are merely exemplary, and the target neural network classification model may use other forms of features as the classification criteria, and the embodiments of the present application are not limited in particular.
The target neural network classification model not only comprises the types of the pulse form features to be extracted, but also comprises the respective weight of each type of pulse form features and the same number of classification threshold values as the types of the classification results (each classification threshold value represents one classification result). It should be noted that these model parameters are determined by the training process of the target neural network classification model, and after the training process is finished, each model parameter of the target neural network classification model is determined. The model parameters may construct a mathematical model for classification. The mathematical model represents the correlation between the pulse morphology features and the final classification results.
In one embodiment, the target neural network classification model achieves classification by performing the following steps: extracting a plurality of pulse morphological characteristics from the target impedance pulse signal; weighting calculation is carried out on each pulse form characteristic and the weight corresponding to the pulse form characteristic, and a characteristic value weighting result is obtained; comparing the feature value weighting result with a preset classification threshold (i.e., a classification threshold obtained after training is completed) to determine a classification threshold corresponding to the feature value weighting result, where for convenience of description, the classification threshold may be referred to as a target classification threshold; and identifying the classification result of the target classification threshold as the classification result of the target impedance pulse signal. It should be noted that each classification threshold may be a numerical range, and after obtaining the feature value weighting result, it is determined which numerical range the feature value weighting result corresponds to, and the numerical range is the target classification threshold. Because each classification threshold value represents a classification result, the classification result of the target impedance pulse signal is determined according to the classification result corresponding to the target classification threshold value. For example, the classification threshold [ a1-a2] represents red blood cells, the classification threshold [ a3-a4] represents platelets, and the classification threshold [ a5-a6] represents interference; assuming that the target classification threshold corresponding to the target impedance pulse signal is [ a3-a4], the classification result of the target impedance pulse signal can be determined to be platelets.
The target neural network classification model is a neural network model, and has the advantages that morphological characteristics of impedance pulse signals corresponding to each classification result can be mined from a large number of blood training samples, and the optimal morphological characteristics for distinguishing different classification results are found.
In addition, the signal input into the classification model of the target neural network in the embodiment of the present application is an impedance pulse signal, that is, after the blood to be classified is sent into the cell analyzer and the impedance detection signal of the blood to be classified is obtained by the impedance detection circuit, the impedance detection signal can be input into the classification model of the target neural network in real time, and the model identifies and classifies the impedance pulse signal in the impedance detection signal. However, the existing identification algorithm needs to obtain a complete impedance detection signal, identify a valid pulse signal from the complete impedance detection signal, generate a distribution histogram according to the amplitude of the valid pulse signal, and analyze the distribution histogram to determine the blood cell type corresponding to the valid pulse signal. Compared with the prior art, the blood cell classification method does not need to obtain a complete impedance detection signal, and when a certain impedance pulse signal is detected in the primary blood sample processing process, the type of the impedance pulse signal generated before the impedance pulse signal is identified, so that the real-time performance is higher.
According to the technical scheme, the blood cell classification method provided by the embodiment of the application inputs the impedance pulse signals of the blood samples to be classified into the trained target neural network classification model, the target neural network classification model is formed by training a large number of impedance pulse signal samples with classification result label information, and the target neural network classification model has strong pulse form characteristic extraction capability, so that the classification results of the impedance pulse signals can be determined according to the form characteristics of the impedance pulse signals of the blood samples to be classified, and the pulse form characteristics are richer and more comprehensive, and compared with the prior art, the classification result accuracy is higher. In addition, the impedance pulse signals caused by the interference factors can be identified, the identification and classification can be carried out on the basis of the impedance pulse signals, and the real-time performance is higher.
See fig. 2, which shows a schematic flow chart of an application of the cell sorting method. As shown in fig. 2, assuming that a partial impedance detection signal of a blood sample to be classified includes a plurality of target impedance pulse signals, the partial impedance detection signal is input to a target neural network classification model for identification. The classification model of the target neural network can extract the pulse morphological characteristics x of the target impedance pulse signal from the classification modeliIt should be noted that the pulse shape characteristics may be various, and although only one x is shown in the figureiBut does not mean that there is only one pulse shape feature, which is just one example of a feature. The target neural network classification model calculates the pulse morphological characteristics according to the classification mode described above to obtain the corresponding classification result yiAnd output. As shown in fig. 2, some of the impedance detection signals are labeled as blood cells, some as platelets, and some as interference.
In practical applications, the classification result can be further used in various scenarios related to blood cell processing, and two specific application scenarios are provided as examples for illustration.
And in the first scene, prompting the interference condition according to the classification result. As shown in fig. 3a, after the classification result is obtained by the classification model of the target neural network, statistics may be performed on the target impedance pulse signal of the interference type, and corresponding prompt information is output according to relevant statistical information.
Specifically, if the classification result of the target impedance pulse signal includes interference, the relevant information of the target impedance pulse signal of the interference type is counted; the related information comprises distribution information and/or morphological information; and outputting prompt information corresponding to the relevant information according to the relevant information.
It has been mentioned above that the classification result obtained in the embodiment of the present application may include interference, and in the case that the classification result indicates that interference really exists, the relevant information of the interference may be counted. Illustratively, distribution information such as the number and the positions of the disturbance in the complete impedance detection signal of the blood sample to be classified can be counted; or the morphological information, such as amplitude, of the target impedance pulse signal corresponding to the interference can be counted. And prompting the interference situation according to the counted related information. For example, it may be determined whether the distribution position of the interference meets a preset distance criterion, and if so, it indicates whether the interference distribution is too dense; and if the number of the interferences exceeds a preset alarm number threshold, the interferences are excessive. If the interference is too much or too dense, the interference factor exists in the detection environment of the blood sample to be classified, and then relevant prompt information can be output to prompt medical personnel to perform interference detection on the detection environment of the blood cell analysis instrument for detection so as to eliminate the relevant interference factor.
And in the second scenario, different types of blood cells are counted according to the classification result. After the classification result is obtained by the classification model of the target neural network, counting each type of blood cells, so that the measured value of each type of blood cells can be given in the analysis report of the blood cells, as shown in fig. 3 b.
Specifically, in addition to the interference, the embodiment of the present application may identify different types of blood cells, and if the classification result of the target impedance pulse signal includes a blood cell type, count the number of target impedance pulse signals of the same blood cell type in the blood to be classified; and determining the number of the target impedance pulse signals as the number of the blood cells of the same blood cell type. For example, the number of erythrocytes is A, the number of platelets is B, etc.
It should be noted that the above two application scenarios are only exemplary, and the blood cell classification result can also be applied in any other blood cell processing scenario that can be expected by those skilled in the art.
The above describes the usage of the target neural network classification model, and the following describes the training process of the target neural network classification model.
As shown in FIG. 4, the training step of the target neural network classification model includes S401-S403.
S401: obtaining a plurality of sets of blood training samples, each set of blood training samples including an impedance pulse signal and label information representing a classification result of the impedance pulse signal, the classification result including a blood cell type or an interference.
In order to make the trained target neural network classification model more accurate, a large number of blood training samples can be collected in advance to obtain a learning sample library. Similar to step S201 in the previous embodiment, each set of blood training samples includes impedance pulse signals and label information of the classification result. Note that the label information may be manually labeled or automatically identified by an algorithm.
Illustratively, the impedance pulse signals of the blood training samples include three categories, respectively: the impedance pulse signal generated by the blood sample without the interference factor (for convenience of description, referred to as a first type pulse signal sample), the impedance pulse signal generated by the interference factor (for convenience of description, referred to as a second type pulse signal sample), and the impedance pulse signal generated by the blood sample with the interference factor (for convenience of description, referred to as a third type pulse signal sample).
As shown in fig. 5a-5c, there are examples of the first type pulse signal sample, the second type pulse signal sample and the third type pulse signal sample. Each type of pulse signal sample has a label, i.e. label information of the classification result of the impedance pulse signal. For the first class of pulse signal samples, the existing pulse recognition and blood cell classification algorithm can be used for positioning the position of the impedance pulse signal and calculating the classification result of the impedance pulse signal; for the second type of pulse signal samples, the existing pulse recognition and blood cell classification algorithm can be used, and the samples are directly marked as interference types after being positioned to impedance pulse signals; for the third class of pulse signal samples, the classification result of the impedance pulse signal may be manually marked.
S402: an initial neural network classification model is obtained.
The initial neural network model may be a neural network model with a classification function, such as a BP (back propagation) network, an LSTM (Long-Short Term Memory), and the like. The model parameters of the initial neural network classification model are determined by the training step of step S403. It should be noted that the present step is not limited to be executed after step S401, and may be executed before or simultaneously with step S401.
S403: and inputting a plurality of groups of blood training samples into the initial neural network classification model to obtain label information output by the initial neural network classification model, and stopping training when the relation between the output label information and the label information of the blood training samples meets a convergence condition to obtain a target neural network classification model.
The blood training samples can be input into the initial neural network classification model in batches, and the model can extract the characteristic data of the blood training samples. For example, a plurality of sampling points are cut from the vicinity of the impedance pulse signal position of a blood training sample to form a pulse waveform, and feature data of the pulse waveform including, but not limited to, position information Pos, Type information Type, pulse Peak value Peak, pulse Width, pulse baseline information Base, pulse total Area information Area, Area difference before and after the pulse Peak value Area, Noise information Noise, etc. may be extracted, and the feature data din (i) of any one pulse waveform i may be recorded as [ Pos (i), Type (i), Peak (i), Width (i), (ii), (iii).
After the label information of a batch of blood training samples is predicted and output according to the characteristic data, the difference between the predicted label information and the labeled label information is compared, the model parameters of the initial neural network classification model are adjusted according to the difference, then the next batch of blood training samples are input, the training process is repeated until the initial neural network classification model is obtainedAnd the classification model is converged, so that the target neural network classification model is obtained. For example, in the training process of the BP (Back Propagation) network, the learning algorithm uses the Steepest Descent BP (SDBP) algorithm, the output of each neuron is calculated backward from the first layer of the network, the influence of the model parameter values on the total error is calculated forward from the last layer, and the model parameters of the network are continuously adjusted through the Back Propagation algorithm, so that the sum of squares of the total error of the network is minimized, and the convergence of the network is realized. It should be noted that the Learning algorithm may be other algorithms, such as LM (Leverberg-Marquardt) algorithm, Momentum BP (MOBP) algorithm, Variable Learning rate BP (VLBP) algorithm, elastic back prediction (RPROP) algorithm, Variable gradient algorithm, quasi-newton algorithm, and so on; the error function for adjustment can be a mean square error function or the like; the number of hidden layer nodes is set to 50; the maximum training frequency is set to 10000, and the target error is set to 0.001; minimum gradient 1e-6。
By the training method provided by the embodiment, the target neural network classification model for classification can be obtained, and the model can identify interference in blood samples, so that the influence of the interference on blood cell type identification is avoided, and the accuracy of blood cell type identification results is improved.
While the embodiments of the blood cell sorting method have been described above, in order to ensure the practical application and implementation of the method, the present application provides several embodiments of blood cell sorting devices as follows.
In one embodiment, the blood cell classification device may be a processing device in which a pre-trained target neural network classification model is built in advance, and a target impedance pulse signal detected by the blood analyzer is input to the processing device. The processing equipment comprises a processor, wherein the processor is configured to obtain a target neural network classification model, the target neural network classification model is obtained by training a plurality of groups of blood training samples through a machine learning algorithm, and each group of the blood training samples comprises impedance pulse signals of blood cells and label information used for representing classification results of the impedance pulse signals; and inputting the target impedance pulse signal to the target neural network classification model to obtain a classification result of the target impedance pulse signal output by the target neural network classification model, wherein the classification result comprises blood cell types or interference. The processing device further comprises a human-computer interaction module configured to output the classification result.
In another embodiment, the blood cell sorting apparatus may be a blood analyzer. The blood analyzer can detect a blood sample to be classified to obtain a target impedance pulse signal, and can classify the target impedance pulse signal to obtain a classification result and output the classification result.
The present application provides a blood analyzer embodiment, see fig. 6, specifically includes: the sample acquisition module 601, the detection module 602, the processor 603 and the human-computer interaction module 604:
a sample collection module 601 for obtaining a blood sample;
a detection module 602, configured to obtain a target impedance pulse signal of a blood sample to be classified;
a processor 603 configured to obtain at least a pre-trained target neural network classification model, where the target neural network classification model is obtained by training, by a machine learning algorithm, a plurality of sets of blood training samples, and each set of the blood training samples includes an impedance pulse signal of a blood cell and label information used for representing a classification result of the impedance pulse signal; inputting the target impedance pulse signal into the target neural network classification model to obtain a classification result of the target impedance pulse signal output by the target neural network classification model, wherein the classification result comprises blood cell types or interference;
a human-computer interaction module 604 configured to output the classification result.
In a specific implementation manner, the target neural network classification model obtains a classification result of the target impedance pulse signal by analyzing a pulse form characteristic of the target impedance pulse signal.
In this particular implementation, the pulse morphology features include a combination of any of: pulse position, pulse peak value, pulse width, pulse peak-to-width ratio, pulse rise time, pulse fall time, pulse line slope, pulse baseline information, total pulse area information, difference in area between and before the pulse peak value, and noise information.
In a specific implementation manner, when the target impedance pulse signal is input to the target neural network classification model to obtain a classification result of the target impedance pulse signal output by the target neural network classification model, the processor 603 is specifically configured to:
inputting the target impedance pulse signal to the target neural network classification model to cause the target neural network classification model to perform the following classification steps: extracting a plurality of pulse morphological characteristics from the target impedance pulse signal; weighting calculation is carried out on each pulse form characteristic and the weight corresponding to the pulse form characteristic, and a characteristic value weighting result is obtained; comparing the characteristic value weighting result with a preset classification threshold value to determine a corresponding target classification threshold value; and identifying the classification result of the target classification threshold as the classification result of the target impedance pulse signal.
In one specific implementation, the processor 603 is further configured to: if the classification result of the target impedance pulse signal comprises interference, counting relevant information of the target impedance pulse signal of the interference type; the related information comprises distribution information and/or morphological information; generating prompt information corresponding to the relevant information according to the relevant information; and the human-computer interaction module is also used for outputting the prompt information.
In one specific implementation, the processor 603 is further configured to: counting the number of target impedance pulse signals of the same blood cell type if the classification result of the target impedance pulse signals comprises the blood cell type; and determining the number of the target impedance pulse signals as the number of the blood cells of the same blood cell type.
In this particular implementation, the blood cell types include: red blood cells or platelets.
In one specific implementation, the processor 603 is further configured to: obtaining a plurality of groups of blood training samples, wherein each group of blood training samples comprises impedance pulse signals and label information used for representing classification results of the impedance pulse signals, and the classification results comprise blood cell types or interferences; obtaining an initial neural network classification model; and inputting a plurality of groups of the blood training samples into the initial neural network classification model to obtain label information output by the initial neural network classification model, and stopping training when the relation between the output label information and the label information of the blood training samples meets a convergence condition to obtain a target neural network classification model.
In this particular implementation, the impedance pulse signal of the blood training sample includes: the impedance pulse signal generated by the blood sample without the interference factor, the impedance pulse signal generated by the interference factor and the impedance pulse signal generated by the blood sample with the interference factor.
The present application further provides a readable storage medium having a computer program stored thereon, wherein the computer program is loaded into and executed by a processor to perform the steps of any of the above-described embodiments of the blood cell sorting method.
Referring to fig. 7, the embodiment of the present application further provides a schematic structural diagram of a blood cell analyzer. As shown in fig. 7, the blood analyzer 700 includes at least a sampling device 710, a sample preparation device 720, a detection device 730, a control device 740, and a display device 750.
The sampling device 710 has a pipette (e.g., a sampling needle) with a pipette nozzle and a driving unit for driving the pipette to quantitatively draw a blood sample to be tested through the pipette nozzle, for example, the sampling needle is moved by the driving unit to draw the blood sample to be tested from a sample container containing the blood sample.
The sample preparation device 720 has at least one reaction cell and a reagent supply device (not shown). The at least one reaction cell is used for receiving a blood sample to be tested, which is drawn by the sampling device 710, and the reagent supply device provides a processing reagent to the at least one reaction cell, so that the blood sample to be tested, which is drawn by the sampling device 710, and the processing reagent provided by the reagent supply device are mixed in the reaction cell to prepare a sample solution to be tested.
The detecting device 730 is used for detecting the sample liquid to be detected prepared by the sample preparation device 720 to obtain the data related to the blood cells. In some embodiments, the detection device 730 includes an impedance detection section and the blood cell related data includes a target impedance pulse signal. The target impedance pulse signal may be amplified by the amplifier and then transmitted to the control device 740, and the control device 740 may perform signal processing on the target impedance pulse signal to obtain a classification result of the sample liquid to be detected, such as a blood cell type (red blood cells or platelets) or interference.
As shown in fig. 8, the control device 740 includes at least a processing component 741, a RAM742, a ROM743, a communication interface 744, a memory 746, and an I/O interface 745. The processing component 741, RAM742, ROM743, communication interface 744, memory 746, and I/O interface 745 communicate via the bus 747. The processing component may be a CPU, GPU or other chip with computing capabilities. The memory 746 contains therein various computer programs such as an operating system and an application program to be executed by the processor component 741, and data necessary for executing the computer programs. In addition, data stored locally during blood sample analysis, if desired, can be stored in memory 746. The I/O interface 745 is constituted by a serial interface such as USB, IEEE1394, or RS-232C, a parallel interface such as SCSI, IDE, or IEEE1284, and an analog signal interface composed of a D/a converter and an a/D converter. An input device including a keyboard, a mouse, a touch panel, or other control buttons is connected to the I/O interface 745, and a user can directly input data to the control apparatus 740 using the input device. Further, a display device 740 having a display function, for example: liquid crystal screens, touch screens, LED display screens and the like. The control device 740 may output the processed data as image display data to the display device 740 for display, for example: classification result data, prompt information, and the like. Communication interface 744 is an interface that may be any communication protocol currently known. The communication interface 744 communicates with the outside through a network. Control device 740 may communicate data with any device connected through the network via communication interface 744 using a communications protocol.
The control apparatus 740 comprises a processor and a storage medium storing a computer program, the control apparatus 740 being configured to perform the following steps when the computer program is executed by the processor: obtaining a pre-trained target neural network classification model, wherein the target neural network classification model is obtained by training a plurality of groups of blood training samples through a machine learning algorithm, and each group of blood training samples comprises impedance pulse signals of blood cells and label information used for representing classification results of the impedance pulse signals; and inputting the target impedance pulse signal to the target neural network classification model to obtain a classification result of the target impedance pulse signal output by the target neural network classification model, wherein the classification result comprises blood cell types or interference.
In some embodiments, the target neural network classification model obtains the classification result of the target impedance pulse signal by analyzing the pulse morphology features of the target impedance pulse signal. The pulse morphology features include a combination of any of: pulse position, pulse peak value, pulse width, pulse peak-to-width ratio, pulse rise time, pulse fall time, pulse line slope, pulse baseline information, total pulse area information, difference in area between and before the pulse peak value, and noise information.
In some embodiments, the control device 740 is configured such that when the computer program is executed by the processor, it further performs the steps of: if the classification result of the target impedance pulse signal comprises interference, counting relevant information of the target impedance pulse signal of the interference type; the related information comprises distribution information and/or morphological information; and generating prompt information corresponding to the related information according to the related information. In some embodiments, the control device 740 is configured such that when the computer program is executed by the processor, it further performs the steps of: counting the number of target impedance pulse signals of the same blood cell type if the classification result of the target impedance pulse signals comprises the blood cell type; and determining the number of the target impedance pulse signals as the number of the blood cells of the same blood cell type.
In some embodiments, the control device 740 is configured such that when the computer program is executed by the processor, it further performs the steps of: obtaining a plurality of groups of blood training samples, wherein each group of blood training samples comprises impedance pulse signals and label information used for representing classification results of the impedance pulse signals, and the classification results comprise blood cell types or interferences; obtaining an initial neural network classification model; and inputting a plurality of groups of the blood training samples into the initial neural network classification model to obtain label information output by the initial neural network classification model, and stopping training when the relation between the output label information and the label information of the blood training samples meets a convergence condition to obtain a target neural network classification model.
In the above embodiment, the impedance pulse signal of the blood training sample includes: the impedance pulse signal generated by the blood sample without the interference factor, the impedance pulse signal generated by the interference factor and the impedance pulse signal generated by the blood sample with the interference factor.
The display device 750 is used for displaying the classification result obtained by the control device 740. The display device 750 is configured as a user interface, for example. In some embodiments, the display device 750 is also used for displaying the prompt information obtained by the control device 740.
Reference is made herein to various exemplary embodiments. However, those skilled in the art will recognize that changes and modifications may be made to the exemplary embodiments without departing from the scope hereof. For example, the various operational steps, as well as the components used to perform the operational steps, may be implemented in differing ways depending upon the particular application or consideration of any number of cost functions associated with operation of the system (e.g., one or more steps may be deleted, modified or incorporated into other steps).
The terms "first," "second," and the like in the description and claims herein and in the above-described drawings are used for distinguishing between different objects and not necessarily for describing a particular sequential order. Furthermore, the terms "include" and "have," as well as any variations thereof, are intended to cover non-exclusive inclusions. For example, a process, method, system, article, or apparatus that comprises a list of steps or elements is not limited to only those steps or elements listed, but may alternatively include other steps or elements not listed, or inherent to such process, method, or apparatus.
Additionally, as will be appreciated by one skilled in the art, the principles herein may be reflected in a computer program product on a computer readable storage medium, which is pre-loaded with computer readable program code. Any tangible, non-transitory computer-readable storage medium may be used, including magnetic storage devices (hard disks, floppy disks, etc.), optical storage devices (CD-ROMs, DVDs, Blu Ray disks, etc.), flash memory, and/or the like. These computer program instructions may be loaded onto a general purpose computer, special purpose computer, or other programmable data processing apparatus to produce a machine, such that the instructions which execute on the computer or other programmable data processing apparatus create means for implementing the functions specified. These computer program instructions may also be stored in a computer-readable memory that can direct a computer or other programmable data processing apparatus to function in a particular manner, such that the instructions stored in the computer-readable memory produce an article of manufacture including means for implementing the function specified. The computer program instructions may also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer implemented process such that the instructions which execute on the computer or other programmable apparatus provide steps for implementing the functions specified.
The foregoing detailed description has been described with reference to various embodiments. However, one skilled in the art will recognize that various modifications and changes may be made without departing from the scope of the present disclosure. Accordingly, the disclosure is to be considered in an illustrative and not a restrictive sense, and all such modifications are intended to be included within the scope thereof. Also, advantages, other advantages, and solutions to problems have been described above with regard to various embodiments. However, the benefits, advantages, solutions to problems, and any element(s) that may cause any element(s) to occur or become more pronounced are not to be construed as a critical, required, or essential feature or element of any or all the claims. As used herein, the terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements but may include other elements not expressly listed or inherent to such process, method, system, article, or apparatus. Furthermore, the term "coupled," and any other variation thereof, as used herein, refers to a physical connection, an electrical connection, a magnetic connection, an optical connection, a communicative connection, a functional connection, and/or any other connection.
The above examples only show some embodiments, and the description thereof is more specific and detailed, but not construed as limiting the scope of the invention. It should be noted that, for a person skilled in the art, several variations and modifications can be made without departing from the inventive concept, which falls within the scope of the present invention. Therefore, the protection scope of the present patent shall be subject to the appended claims.
Claims (11)
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010744431.4A CN114088605A (en) | 2020-07-29 | 2020-07-29 | Blood cell classification method and related equipment |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010744431.4A CN114088605A (en) | 2020-07-29 | 2020-07-29 | Blood cell classification method and related equipment |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN114088605A true CN114088605A (en) | 2022-02-25 |
Family
ID=80294942
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010744431.4A Pending CN114088605A (en) | 2020-07-29 | 2020-07-29 | Blood cell classification method and related equipment |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN114088605A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116698692A (en) * | 2023-08-03 | 2023-09-05 | 深圳市灵明光子科技有限公司 | Particle detector and particle detection method |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| FR2733596A1 (en) * | 1995-04-28 | 1996-10-31 | Hycel Groupe Lisabio | Method for blood sample analysis using a neural network. |
| CN101226133A (en) * | 2008-01-28 | 2008-07-23 | 宁波大学 | A Classification and Recognition Method of Blood Cell Pulse Signal |
| CN103364452A (en) * | 2013-06-03 | 2013-10-23 | 南昌大学 | Blood cell classification and identification circuit and method |
| CN106769698A (en) * | 2016-12-29 | 2017-05-31 | 长春迪瑞医疗科技股份有限公司 | A kind of haemocyte abnormal pulsers signal identification processing method based on theory of electrical impedance |
| CN108021903A (en) * | 2017-12-19 | 2018-05-11 | 南京大学 | The error calibrating method and device of artificial mark leucocyte based on neutral net |
| US20190347467A1 (en) * | 2018-05-10 | 2019-11-14 | Juntendo Educational Foundation | Image analysis method, apparatus, non-transitory computer readable medium, and deep learning algorithm generation method |
-
2020
- 2020-07-29 CN CN202010744431.4A patent/CN114088605A/en active Pending
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| FR2733596A1 (en) * | 1995-04-28 | 1996-10-31 | Hycel Groupe Lisabio | Method for blood sample analysis using a neural network. |
| CN101226133A (en) * | 2008-01-28 | 2008-07-23 | 宁波大学 | A Classification and Recognition Method of Blood Cell Pulse Signal |
| CN103364452A (en) * | 2013-06-03 | 2013-10-23 | 南昌大学 | Blood cell classification and identification circuit and method |
| CN106769698A (en) * | 2016-12-29 | 2017-05-31 | 长春迪瑞医疗科技股份有限公司 | A kind of haemocyte abnormal pulsers signal identification processing method based on theory of electrical impedance |
| CN108021903A (en) * | 2017-12-19 | 2018-05-11 | 南京大学 | The error calibrating method and device of artificial mark leucocyte based on neutral net |
| US20190347467A1 (en) * | 2018-05-10 | 2019-11-14 | Juntendo Educational Foundation | Image analysis method, apparatus, non-transitory computer readable medium, and deep learning algorithm generation method |
Non-Patent Citations (3)
| Title |
|---|
| 张时民;: "五分类法血细胞分析仪测定原理和散点图特征", 中国医疗器械信息, no. 12, 25 December 2008 (2008-12-25) * |
| 贾丹丹;李宏;: "基于小波包和神经网络的血细胞识别方法的研究", 中国医疗器械杂志, no. 04, 30 July 2008 (2008-07-30) * |
| 贾丹丹;李宏;: "基于神经网络和小波分析的血细胞识别算法", 计算机应用与软件, no. 08, 15 August 2009 (2009-08-15) * |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116698692A (en) * | 2023-08-03 | 2023-09-05 | 深圳市灵明光子科技有限公司 | Particle detector and particle detection method |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109117730A (en) | Electrocardiogram auricular fibrillation real-time judge method, apparatus, system and storage medium | |
| CN102177427A (en) | Detecting and handling coincidence in particle analysis | |
| CN109656366B (en) | Emotional state identification method and device, computer equipment and storage medium | |
| CN109891508A (en) | Single cell type detection method, device, equipment and storage medium | |
| WO2022005586A1 (en) | Managing defects in a model training pipeline using synthetic data sets associated with defect types | |
| CN107785066B (en) | Method, device and system for modifying heartbeat type | |
| CN107389536B (en) | Flow cell particle classification counting method based on density-distance center algorithm | |
| CN115184244A (en) | Blood analysis system | |
| CN108648765A (en) | A kind of method, apparatus and terminal of voice abnormality detection | |
| CN103892830A (en) | Emotion detection method and system based on human skin resistance changes | |
| CN104515725B (en) | A kind of method and system and its cytoanalyze for recognizing abnormal particle | |
| CN116030958A (en) | Auxiliary diagnosis method, device, equipment and system and storage medium | |
| Xiao et al. | A deep feature fusion network for fetal state assessment | |
| CN116029960A (en) | Auxiliary diagnosis method, image processing model and storage medium | |
| CN114088605A (en) | Blood cell classification method and related equipment | |
| CN109508649A (en) | A kind of pulse signal analysis recognition method of cellanalyzer | |
| EP2920573B1 (en) | Particle data segmentation result evaluation methods and flow cytometer | |
| CN114707608A (en) | Medical quality control data processing method, apparatus, device, medium, and program product | |
| US20220292855A1 (en) | Analysis method and analyzer | |
| CN117331047A (en) | Human behavior data analysis method and system based on millimeter wave radar | |
| CN110136826A (en) | Intelligent medical assistant diagnosis system based on deep learning | |
| CN112022142B (en) | Electrocardiosignal type identification method, device and medium | |
| CN114496196A (en) | Automatic auditing system for clinical biochemical inspection in medical laboratory | |
| CN113808088A (en) | A pollution detection method and system | |
| JPH11258296A (en) | Automatic discrimination method for partial discharge |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |