CN111488796A - Cell canceration recognition method based on AFM and SVM classifiers - Google Patents
Cell canceration recognition method based on AFM and SVM classifiers Download PDFInfo
- Publication number
- CN111488796A CN111488796A CN202010165002.1A CN202010165002A CN111488796A CN 111488796 A CN111488796 A CN 111488796A CN 202010165002 A CN202010165002 A CN 202010165002A CN 111488796 A CN111488796 A CN 111488796A
- Authority
- CN
- China
- Prior art keywords
- cell
- afm
- svm
- data
- svm classifier
- 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
- 238000000034 method Methods 0.000 title claims abstract description 55
- 238000012706 support-vector machine Methods 0.000 claims abstract description 57
- 238000007373 indentation Methods 0.000 claims abstract description 19
- 238000002474 experimental method Methods 0.000 claims abstract description 11
- 238000012545 processing Methods 0.000 claims abstract description 11
- 238000012876 topography Methods 0.000 claims abstract description 10
- 238000012549 training Methods 0.000 claims description 19
- 206010028980 Neoplasm Diseases 0.000 claims description 12
- 201000011510 cancer Diseases 0.000 claims description 12
- 238000012360 testing method Methods 0.000 claims description 12
- 238000006073 displacement reaction Methods 0.000 claims description 11
- 238000013507 mapping Methods 0.000 claims description 4
- 238000010606 normalization Methods 0.000 claims description 4
- 230000009977 dual effect Effects 0.000 claims description 3
- 230000000694 effects Effects 0.000 claims description 3
- 238000005457 optimization Methods 0.000 claims description 3
- 238000012163 sequencing technique Methods 0.000 claims description 3
- 208000005623 Carcinogenesis Diseases 0.000 claims 1
- 230000036952 cancer formation Effects 0.000 claims 1
- 231100000504 carcinogenesis Toxicity 0.000 claims 1
- 238000012795 verification Methods 0.000 claims 1
- 238000001514 detection method Methods 0.000 abstract description 3
- 239000000523 sample Substances 0.000 description 11
- 208000005718 Stomach Neoplasms Diseases 0.000 description 10
- 206010017758 gastric cancer Diseases 0.000 description 10
- 201000011549 stomach cancer Diseases 0.000 description 10
- 230000002496 gastric effect Effects 0.000 description 8
- 101150086731 ges-1 gene Proteins 0.000 description 8
- 238000004458 analytical method Methods 0.000 description 3
- 239000006059 cover glass Substances 0.000 description 3
- 230000001575 pathological effect Effects 0.000 description 3
- 238000002790 cross-validation Methods 0.000 description 2
- 238000012258 culturing Methods 0.000 description 2
- 238000005034 decoration Methods 0.000 description 2
- 238000002575 gastroscopy Methods 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 239000000853 adhesive Substances 0.000 description 1
- 230000001070 adhesive effect Effects 0.000 description 1
- 238000004630 atomic force microscopy Methods 0.000 description 1
- 229910052788 barium Inorganic materials 0.000 description 1
- DSAJWYNOEDNPEQ-UHFFFAOYSA-N barium atom Chemical compound [Ba] DSAJWYNOEDNPEQ-UHFFFAOYSA-N 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000005859 cell recognition Effects 0.000 description 1
- 238000005119 centrifugation Methods 0.000 description 1
- 230000008859 change Effects 0.000 description 1
- 238000007796 conventional method Methods 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000029087 digestion Effects 0.000 description 1
- 201000010099 disease Diseases 0.000 description 1
- 208000037265 diseases, disorders, signs and symptoms Diseases 0.000 description 1
- 210000001156 gastric mucosa Anatomy 0.000 description 1
- 239000011521 glass Substances 0.000 description 1
- 238000003384 imaging method Methods 0.000 description 1
- 238000002955 isolation Methods 0.000 description 1
- 239000007788 liquid Substances 0.000 description 1
- 229920002521 macromolecule Polymers 0.000 description 1
- 239000000463 material Substances 0.000 description 1
- 235000012054 meals Nutrition 0.000 description 1
- 238000005259 measurement Methods 0.000 description 1
- 230000004962 physiological condition Effects 0.000 description 1
- 238000003825 pressing Methods 0.000 description 1
- 230000008569 process Effects 0.000 description 1
- 210000002784 stomach Anatomy 0.000 description 1
- 230000003746 surface roughness Effects 0.000 description 1
- 210000001519 tissue Anatomy 0.000 description 1
- 239000000439 tumor marker Substances 0.000 description 1
- 239000002699 waste material Substances 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/60—Type of objects
- G06V20/69—Microscopic objects, e.g. biological cells or cellular parts
- G06V20/698—Matching; Classification
-
- 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
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- Life Sciences & Earth Sciences (AREA)
- General Physics & Mathematics (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Evolutionary Biology (AREA)
- Evolutionary Computation (AREA)
- Computer Vision & Pattern Recognition (AREA)
- General Engineering & Computer Science (AREA)
- Bioinformatics & Computational Biology (AREA)
- Artificial Intelligence (AREA)
- Health & Medical Sciences (AREA)
- Biomedical Technology (AREA)
- General Health & Medical Sciences (AREA)
- Molecular Biology (AREA)
- Multimedia (AREA)
- Investigating Or Analysing Biological Materials (AREA)
Abstract
The invention discloses a cell canceration identification method based on an AFM (atomic force microscope) and an SVM (support vector machine) classifier, and aims to provide a cell canceration identification method which combines mechanical data with an SVM classifier, is quick and accurate, and is low in cost. The method comprises the following steps: using AFM equipment to obtain a topography of a cell to be identified; performing a nano indentation experiment on the cell to be identified by using AFM equipment to obtain a mechanical characteristic curve and micro-morphology structure information of the cell to be identified; processing the obtained mechanical characteristic curve data and the micro-morphology structure information to obtain related mechanical characteristic parameters; and (3) taking the obtained relevant mechanical characteristic parameter data as input data of an SVM classifier, identifying whether the cells are cancerated or not through a classification identification model of the SVM classifier, and outputting the cells to be identified by the SVM classifier. The method has the advantages of short detection period, high identification efficiency, high accuracy, high automation degree, low cost and simple steps.
Description
Technical Field
The invention relates to the technical field of biology, in particular to a cell canceration identification method.
Background
In the present society, cancer has become a disease with high morbidity, which brings great threat to human life and huge economic loss to society.
At present, cancer identification methods are mainly based on the combination of various detection means and clinical experience of doctors. Taking gastric cancer as an example, a general gastric cancer identification method is to perform gastroscopy, X-ray barium meal, tumor marker examination and the like, wherein the gastroscopy can directly observe the pathological part and range of gastric mucosa and can acquire pathological tissues for pathological examination, and the method is the most effective method for identifying gastric cancer, but the method needs complicated steps, takes long time, often needs 2-7 days, and identification is performed by doctors with the help of clinical experience, so that certain identification errors exist, and the accuracy cannot be guaranteed.
With the development of technology, cancer cell identification using image images in combination with SVM linear classifiers is currently being applied. The existing method mainly comprises the steps of collecting a large number of single cell pictures through a microscope, extracting relevant characteristics of cells according to image recognition, and finally training an SVM linear classifier to recognize the cells. The method needs to collect a large number of single-cell photos, and can spend a large amount of time, energy, materials and financial resources; the cell characteristic data extracted through image recognition is often high-dimensional and redundant, and when the SVM classifier is trained, the training recognition speed is low, the main components are unknown, and the recognition accuracy is low.
Aiming at the problems of the existing method, a cell canceration identification method which is rapid, accurate and low in cost is needed to be provided.
Disclosure of Invention
The invention aims to provide a cell canceration identification method which combines mechanical data with an SVM classifier, is rapid and accurate and has low cost aiming at the technical defects in the prior art.
The technical scheme adopted for realizing the purpose of the invention is as follows:
a cell canceration recognition method based on AFM and SVM classifiers comprises the following steps:
(1) using AFM equipment to obtain a topography of a cell to be identified;
(2) performing a nano indentation experiment on the cell to be identified by using AFM equipment to obtain a mechanical characteristic curve and micro-morphology structure information of the cell to be identified;
(3) processing the mechanical characteristic curve data and the micro-morphology structure information obtained in the step (2) to obtain related mechanical characteristic parameters;
(4) and (4) taking the relevant mechanical characteristic parameter data obtained in the step (3) as input data of an SVM classifier, identifying whether the cells are cancerated or not through a classification identification model of the SVM classifier, and outputting the SVM classifier as the types of the cells to be identified.
The classification recognition model is obtained by the following method:
(1) respectively acquiring the topography maps of normal cells and cancerated cells by using AFM equipment;
(2) performing a nano indentation experiment on the normal cells and the cancer cells respectively by using AFM equipment to obtain mechanical characteristic curves and micro-morphology structure information of the normal cells and the cancer cells;
(3) processing the mechanical characteristic curve data and the micro-morphology structure information obtained in the step (2) to obtain related mechanical characteristic parameters;
(4) randomly sequencing the related mechanical characteristic parameter data obtained in the step (3), dividing the data into training data and testing data, using the training data and the testing data as input data of an SVM classifier, and carrying out normalization processing on the input data;
(5) using the optimized penalty factor C and the width parameter sigma of the Gaussian kernel function2Training the SVM classifier by adopting a cross validation method according to the training data, obtaining an identification model containing an optimal hyperplane after training, and storing the identification model of the trained SVM classifier;
(6) testing the classification recognition effect of the recognition model of the SVM classifier by using the recognition model of the SVM classifier saved in the step (5) and the test data, and outputting the tested classification result;
(7) if the tested classification result meets the requirement, the classification result is used as the classification recognition model; and if the requirement is not met, repeating the steps (4) to (5).
The mechanical characteristic curve is a force-displacement curve.
The indentation force when carrying out the nanoindentation experiment is 2nN, Z L ength is 3uM, and the indentation speed is 2.0 uM/s.
The penalty factor C and the width parameter sigma of the Gaussian kernel function2The optimization method comprises the steps of setting a kernel function as a Gaussian kernel by using an L ibSVM software package, and searching a penalty factor C and a Gaussian kernel function parameter sigma which enable the recognition rate to be maximum by adopting a violent search traversal method and combining the training data and a L ibSVM software package2Then, the penalty factor C and the width parameter sigma of the Gaussian kernel function are refined2Repeating the steps to obtain an optimized penalty factor C and a width parameter sigma of the Gaussian kernel function2(ii) a Wherein, the penalty factor C [ -5, -3, -1,1,3,5,7,9,11,13,15 [ -5 [ -3 [ -1 [ -3 [ -5 [ -7 [ -9 [ -11 [ -13 [ -]Parameter of Gaussian Kernel function σ2＝[-15,-13,-11,-9,-7,-5,-3,-1,1,3]。
Defining the optimal hyperplane:
s.t.yi[wTφ(xi)+b]≥1-ζi,ζi≥0 (2)；
combining the optimal hyperplane original problem with the dual problem, and solving the formula (1) and the formula (2) through an (SMO) algorithm to obtain the optimal hyperplane original problem
Wherein: k (x)iX) represents a kernel function αiRepresenting a lagrange multiplier; w ═ w (w)1；w1；....wd(ii) a ) Is a normal vector; b is a displacement term; phi (x)i) Representing the feature vector after mapping the sample feature vector x; c is a penalty factor; zetaiIs a relaxation variable; y isiRepresenting an actual target value; i and j respectively represent the ith and jth samples; n represents the total number of samples; d represents the dimension of w.
The recognition model function of the SVM classifier is
If g (x) is 1, the cell is a normal cell; if g (x) is-1, the cancer cell is obtained.
And when the nano indentation experiment is carried out, the indentation position is selected according to the obtained topography.
Compared with the prior art, the invention has the beneficial effects that:
1. the cell canceration recognition method obtains the mechanical characteristic curve and the micro-morphology structure information of the cell by means of AFM, processes the information to obtain the relevant mechanical parameters of the cell, uses the obtained relevant mechanical parameters as the input data of the SVM separator, and recognizes whether the cell is cancerated or not by the SVM separator, and has the advantages of short detection period, high recognition efficiency, high accuracy, high automation degree, low cost and simple steps.
2. The cell canceration identification method adopts AFM to extract mechanical data information, has simple requirements on samples, is simple to operate, has low requirements on environment and has strong applicability.
3. The cell canceration recognition method provided by the invention adopts normalization processing on the input data of the SVM classifier, so that mechanical parameters with different orders of magnitude can be unified to the same order of magnitude, the dimensional influence among the parameters is eliminated, and meanwhile, the solving speed and precision of the SVM classifier are improved.
Drawings
FIG. 1 is a graphical representation of SGC cells according to an embodiment of the present invention;
FIG. 2 shows an SGC cell force-displacement curve according to an embodiment of the present invention;
FIG. 3 illustrates a ROC curve for an embodiment of the present invention;
FIG. 4 is a flow chart of the method for identifying cell canceration based on AFM and SVM classifiers according to the present invention.
Detailed Description
The present invention will be described in detail below with reference to the drawings, in which the identification of cancer of gastric cancer cells is taken as an example.
The cell canceration identification method based on AFM (atomic Force Microcopy) and SVM classifiers comprises the following steps:
1. obtaining a classification recognition model required by cell recognition:
1-1, respectively culturing the gastric cell GES-1 and the gastric cancer cell SGC in a medium dish for 48 hours to allow the two cells to divide, grow and adhere to the wall, and then respectively transferring the two cells to a small dish with a cover glass by enzymolysis digestion, centrifugation, liquid change and the like;
1-2, culturing two types of cells in the step 1-1 for 24 hours respectively, clamping the edge of the cover glass in the step 1-1 by using tweezers after the cells adhere to the wall, putting the cover glass on a dish cover of a small dish, adding culture solution into the dish cover, and finally putting the dish cover at the center of a glass slide on a sample platform;
1-3, installing an AFM spherical probe on an AFM head, pressing a power button and an isolation button, starting the AFM, enabling the AFM probe to enter a culture solution which is covered by a dish and is loaded with the gastric cell GES-1 through a stepping motor of the AFM head, adjusting the position of AFM laser and four quadrants, enabling a laser red point to be located in the middle of the four quadrants, enabling a reflected SUM value to reach about 2.5V maximum, calibrating the rigidity k of the probe through a thermal noise method of AFM self-contained software, enabling the calibrated rigidity to be close to a factory value of 0.06N/m of a merchant, selecting a QI mode of the AFM, scanning the gastric cell GES-1 to obtain a topography map of the gastric cell GES-1, then using a contact mode of the AFM to carry out a nano indentation experiment on the gastric cell GES-1, setting an indentation force to be 2nN, setting a Z L ength to be 3uM, an indentation speed to be 2.0uM/s, selecting an indentation position to carry out a micro-structure of each gastric cell through the selected topography map of the gastric cell GES-1, and recording the micro displacement of each gastric cell by AFM;
an AFM probe enters culture solution which is filled with gastric cancer cells SGC and is covered by a dish, and the force-displacement curve and the micro-morphology structure information of the gastric cancer cells are obtained by adopting the same method;
the obtained topography and force-displacement curves of the SGC are shown in fig. 3 and 4;
and 1-4, processing the force-displacement curve and microstructure information in the step 1-3 by using cell mechanics knowledge, and acquiring the relevant mechanical parameters of the Young modulus, the adhesive force, the adhesion work, the surface roughness and the height of the cell by adopting a conventional method. For example, the Young modulus can be calculated by the Hertz contact formula (1) of the spherical indenter;
wherein FsphericalThe contact force between the needle tip and the sample, R is the effective radius of the needle tip and is the indentation depth of the sample;
and 1-5, classifying and identifying the obtained mechanical characteristic parameters by using an SVM classifier to obtain cell classification, and acquiring a classification identification model required by the cell identification.
The classification identification method in the embodiment is as follows:
1-5-1, taking the relevant mechanical characteristic parameters in the step 1-4 as input data of an SVM classifier, randomly sequencing the input data, taking 1500 groups of data as training data and 500 groups of data as test data, wherein the label of a GES-1 cell is 1, the label of an SGC cell is-1, and then carrying out normalization processing on the data;
1-5-2, defining an optimal hyperplane of a classification recognition model:
s.t.yi[wTφ(xi)+b]≥1-ζi,ζi≥0 (3)；
wherein: w ═ w (w)1；w1；....wd(ii) a ) Is a normal vector; b is a displacement term; phi (x)i) Representing the feature vector after mapping the sample feature vector x; c is a penalty factor; zetaiIs a relaxation variable; i and j respectively represent the ith and jth samples; n represents the total number of samples; d represents the dimension of w.
Combining the hyperplane original problem with the dual problem, and solving the formula (2) and the formula (3) through an SMO (sequential minimumization) algorithm to obtain a formula (4):
wherein, k (x)iX) represents a kernel function αiRepresenting a lagrange multiplier; w ═ w (w)1；w1；....wd(ii) a ) Is a normal vector; b is a displacement term; phi (x)i) Representing the feature vector after mapping the sample feature vector x; c is a penalty factor; zetaiIs a relaxation variable; y isiRepresenting an actual target value;
the SVM model is obtained by adopting a method of L ibSVM software package, wherein a kernel function is set as a Gaussian kernel, and the Gaussian kernel function is as follows:
wherein x and y are respectively a feature vector and an actual output value of an input sample;
firstly, roughly searching a penalty factor C (C [ -5, -3, -1,1,3,5,7,9,11,13, 15) which maximizes the recognition rate by adopting a violent search traversal method and combining the training data of the step 1-5-1 and the L ibSVM software package]) And a Gaussian kernel parameter σ2(σ2＝[-15,-13,-11,-9,-7,-5,-3,-1,1,3]) Then refine CAnd σ2Repeating the steps to obtain optimized C and sigma2；
1-5-3, Using C and σ optimized in step 1-5-22And step 1-5-1, training an SVM classifier by using the training data, training an SVM by using a cross validation method, and storing a recognition model function of the trained SVM classifier, wherein the SVM classifier at the moment contains the required optimal hyperplane
If g (x) is 1([ w ]Tφ(xi)+b]Not less than 0), the stomach cell GES-1; if g (x) is-1 ([ w ]Tφ(xi)+b]< -1)), gastric cancer cell SGC;
1-5-4, substituting the test data in the step 1-5-1 into the step 1-5-3 to test the actual effect of the trained recognition model of the SVM classifier, if the recognition precision and the recognition efficiency both reach the pre-assumed result, not training the model, namely the model is finally applied to the actual classification recognition model, and if the recognition precision and the recognition efficiency do not reach the ideal state, repeating the steps 1-5-1 to 1-5-3 until the ideal state is reached to obtain the classification recognition model.
The obtained classification recognition model
The ROC curve plotted in conjunction with the test data is shown in fig. 3, and the classification results are as follows:
optimization finished，#iter＝742
nu＝0.988000
obj＝-30.091246，rho＝-0.277842
nSV＝1482，nBSV＝1482
Total nSV＝1482
Accuracy＝100％(500/500)(classificaticn)
as can be seen from fig. 3 and the above precision analysis, the method has high classification precision reaching 100%, and can realize precise identification of gastric cancer cells.
2. The flow chart of the identification method for classifying and identifying the cells to be identified is shown in FIG. 4:
2-1, referring to the method in the step 1-3, using AFM equipment to obtain a topography of the cell to be identified;
2-2, referring to the method in the step 1-3, performing a nano indentation experiment on the cell to be identified by using AFM equipment to obtain a mechanical characteristic curve and micro-morphology structure information of the cell to be identified;
2-3, processing the obtained mechanical characteristic curve data to be identified and the micro-morphology structure information by referring to the method in the step 1-3 to obtain relevant mechanical characteristic parameters of the cells to be identified;
and 2-4, taking the relevant mechanical property parameter data of the to-be-recognized cells obtained in the step 2-3 as input data of an SVM classifier, recognizing by using the classification recognition model obtained in the step 1, and outputting the output of the SVM classifier as the types of the to-be-recognized cells.
The cell canceration identification method provided by the invention overcomes the problems of time consumption, labor waste, high cost, incapability of guaranteeing accuracy and complicated steps in the traditional cancer identification method, and provides a gastric cancer cell identification method based on AFM (atomic force microscopy) measurement of cell mechanical properties in combination with SVM (support vector machine). The method is automatically identified by a program, so that the method has the advantages of high efficiency, high accuracy, high automation degree and low cost on the whole, and an effective supplementary method is provided for identifying whether cells are cancerated. Moreover, AFM (atomic Force Microcopy) is used as a surface information analysis tool with extremely high resolution and simple requirements on samples, the operation is simple, the operation flow is easy to learn, the requirements on environment are not high, the result is fast, the tool can be used for carrying out surface imaging, mechanical property analysis and microstructure observation on cancer cells and other biological macromolecules under the physiological condition with the atomic resolution, and further mechanical data for identifying whether the cells are cancerated is obtained, the steps are simple, and the cost is low.
The foregoing is only a preferred embodiment of the present invention, and it should be noted that, for those skilled in the art, various modifications and decorations can be made without departing from the principle of the present invention, and these modifications and decorations should also be regarded as the protection scope of the present invention.
Claims (8)
1. A cell canceration recognition method based on an AFM (atomic force microscope) classifier and an SVM (support vector machine) classifier is characterized by comprising the following steps of:
(1) using AFM equipment to obtain a topography of a cell to be identified;
(2) performing a nano indentation experiment on the cell to be identified by using AFM equipment to obtain a mechanical characteristic curve and micro-morphology structure information of the cell to be identified;
(3) processing the mechanical characteristic curve data and the micro-morphology structure information obtained in the step (2) to obtain related mechanical characteristic parameters;
(4) and (4) taking the relevant mechanical characteristic parameter data obtained in the step (3) as input data of an SVM classifier, identifying whether the cells are cancerated or not through a classification identification model of the SVM classifier, and outputting the SVM classifier as the types of the cells to be identified.
2. The AFM and SVM classifier based cell canceration recognition method according to claim 1, wherein the classification recognition model is obtained by:
(1) respectively acquiring the topography maps of normal cells and cancerated cells by using AFM equipment;
(2) performing a nano indentation experiment on the normal cells and the cancer cells respectively by using AFM equipment to obtain mechanical characteristic curves and micro-morphology structure information of the normal cells and the cancer cells;
(3) processing the mechanical characteristic curve data and the micro-morphology structure information obtained in the step (2) to obtain related mechanical characteristic parameters;
(4) randomly sequencing the related mechanical characteristic parameter data obtained in the step (3), dividing the data into training data and testing data, using the training data and the testing data as input data of an SVM classifier, and carrying out normalization processing on the input data;
(5) using the optimized penalty factor C and the width parameter sigma of the Gaussian kernel function2And the training data starts to train the SVM by adopting a cross verification methodThe classifier obtains an identification model containing an optimal hyperplane after training, and stores the identification model of the trained SVM classifier;
(6) testing the classification recognition effect of the recognition model of the SVM classifier by using the recognition model of the SVM classifier saved in the step (5) and the test data, and outputting the tested classification result;
(7) if the tested classification result meets the requirement, the classification result is used as the classification recognition model; and if the requirement is not met, repeating the steps (4) to (5).
3. The method for identifying cell canceration based on AFM and SVM classifiers according to claim 1 or 2, wherein the mechanical property curve is a force-displacement curve.
4. The method for identifying cell canceration based on AFM and SVM classifiers according to claim 1 or 2, wherein the nano-indentation experiment is performed with an indentation force of 2nN, Z L ength of 3uM, and an indentation speed of 2.0 uM/s.
5. The AFM and SVM classifier based cell canceration recognition method of claim 2, wherein the penalty factor C and the width parameter σ of the Gaussian kernel function2The optimization method comprises the steps of setting a kernel function as a Gaussian kernel by using an L ibSVM software package, and searching a penalty factor C and a Gaussian kernel function parameter sigma which enable the recognition rate to be maximum by adopting a violent search traversal method and combining the training data and a L ibSVM software package2Then, the penalty factor C and the width parameter sigma of the Gaussian kernel function are refined2Repeating the steps to obtain an optimized penalty factor C and a width parameter sigma of the Gaussian kernel function2(ii) a Wherein, the penalty factor C [ -5, -3, -1,1,3,5,7,9,11,13,15 [ -5 [ -3 [ -1 [ -3 [ -5 [ -7 [ -9 [ -11 [ -13 [ -]Parameter of Gaussian Kernel function σ2＝[-15,-13,-11,-9,-7,-5,-3,-1,1,3]。
6. The AFM and SVM classifier based cell carcinogenesis recognition method according to claim 2, characterized in that the optimal hyperplane is defined:
s.t.yi[wTφ(xi)+b]≥1-ζi,ζi≥0 (2)；
combining the optimal hyperplane original problem with the dual problem, and solving the formula (1) and the formula (2) through an SMO algorithm to obtain
Wherein: k (x)iX) represents a kernel function αiRepresenting a lagrange multiplier; w ═ w (w)1；w1；....wd(ii) a ) Is a normal vector; b is a displacement term; phi (x)i) Representing the feature vector after mapping the sample feature vector x; c is a penalty factor; zetaiIs a relaxation variable; y isiRepresenting an actual target value; i and j respectively represent the ith and jth samples; n represents the total number of samples; d represents the dimension of w.
7. The method of claim 6, wherein the SVM classifier is based on an AFM classifier and an SVM classifier, and the recognition model function is
If g (x) is 1, the cell is a normal cell; if g (x) is-1, the cancer cell is obtained.
8. The method for identifying cell canceration based on AFM and SVM classifiers according to claim 1 or 2, wherein the indentation position is selected according to the obtained topographic map when performing the nanoindentation experiment.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010165002.1A CN111488796A (en) | 2020-03-11 | 2020-03-11 | Cell canceration recognition method based on AFM and SVM classifiers |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010165002.1A CN111488796A (en) | 2020-03-11 | 2020-03-11 | Cell canceration recognition method based on AFM and SVM classifiers |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN111488796A true CN111488796A (en) | 2020-08-04 |
Family
ID=71811684
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010165002.1A Pending CN111488796A (en) | 2020-03-11 | 2020-03-11 | Cell canceration recognition method based on AFM and SVM classifiers |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111488796A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113516061A (en) * | 2021-06-23 | 2021-10-19 | 南开大学 | PC12 cell state identification method based on machine learning |
Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109270390A (en) * | 2018-09-14 | 2019-01-25 | 广西电网有限责任公司电力科学研究院 | Diagnosis Method of Transformer Faults based on Gaussian transformation Yu global optimizing SVM |
-
2020
- 2020-03-11 CN CN202010165002.1A patent/CN111488796A/en active Pending
Patent Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109270390A (en) * | 2018-09-14 | 2019-01-25 | 广西电网有限责任公司电力科学研究院 | Diagnosis Method of Transformer Faults based on Gaussian transformation Yu global optimizing SVM |
Non-Patent Citations (4)
| Title |
|---|
| 张婕: "基于级联多分类器融合的宫颈癌细胞识别技术的研究" * |
| 张艳;吴玲;: "基于支持向量机和交叉验证的变压器故障诊断" * |
| 文永康;谭红芳;马勇;: "基于支持向量机的战术电台网络通信信号调制识别" * |
| 李密;席宁;王越超;刘连庆;: "基于多参数成像AFM的细胞及分子力学特性探测研究进展" * |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113516061A (en) * | 2021-06-23 | 2021-10-19 | 南开大学 | PC12 cell state identification method based on machine learning |
| CN113516061B (en) * | 2021-06-23 | 2022-04-29 | 南开大学 | PC12 cell state identification method based on machine learning |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| DK2973397T3 (en) | Tissue-object-based machine learning system for automated assessment of digital whole-slide glass | |
| WO2019084517A1 (en) | System and methods of image-based assay using crof and machine learning | |
| Tania et al. | Intelligent image-based colourimetric tests using machine learning framework for lateral flow assays | |
| CN107209092A (en) | System and method for centre dissection | |
| JP2010540931A (en) | Method and system using multiple wavelengths for processing biological specimens | |
| ES2692346T3 (en) | Systems and methods for the automated extraction of information of high content of complete organisms | |
| Mao et al. | A deep convolutional neural network trained on representative samples for circulating tumor cell detection | |
| CN111488796A (en) | Cell canceration recognition method based on AFM and SVM classifiers | |
| CN112784767A (en) | Cell example segmentation algorithm based on leukocyte microscopic image | |
| Park et al. | Deep learning predicts the differentiation of kidney organoids derived from human induced pluripotent stem cells | |
| Zhang et al. | MSHT: Multi-stage hybrid transformer for the ROSE image analysis of pancreatic cancer | |
| Roy et al. | A semi-automated positioning system for contact-mode atomic force microscopy (AFM) | |
| CN103268494B (en) | Parasite egg recognition methods based on rarefaction representation | |
| CN111462082A (en) | Focus picture recognition device, method and equipment and readable storage medium | |
| Suganya et al. | Ultrasound ovary cyst image classification with deep learning neural network with Support vector machine | |
| Xia et al. | AI-enabled microscopic blood analysis for microfluidic COVID-19 hematology | |
| TW201819913A (en) | Magnetophorisis measuring system for determining moving status of object and quantifying amount of magnetic particles contained therein | |
| Nigam et al. | Revisiting HEp-2 cell image classification | |
| CN109491409B (en) | Method for automatically extracting mitochondria of target cell in vitro | |
| CN116757998A (en) | Screening method and device for CTC cells and CTC-like cells based on AI | |
| Alkrimi et al. | Classification of red blood cells using principal component analysis technique | |
| Le Minh et al. | Deep feature fusion for breast cancer diagnosis on histopathology images | |
| Scholler et al. | Automatic diagnosis and classification of breast surgical samples with dynamic full-field OCT and machine learning | |
| Liu et al. | 3D Rotation-Invariant description from tensor operation on spherical HOG field | |
| Bykanov et al. | Artificial Intelligence Technologies in the Microsurgical Operating Room |
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 | ||
| WD01 | Invention patent application deemed withdrawn after publication | ||
| WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20200804 |