CN106682451A - Formula proportion determining method for biological tissue simulation material and system - Google Patents
Formula proportion determining method for biological tissue simulation material and system Download PDFInfo
- Publication number
- CN106682451A CN106682451A CN201611123843.6A CN201611123843A CN106682451A CN 106682451 A CN106682451 A CN 106682451A CN 201611123843 A CN201611123843 A CN 201611123843A CN 106682451 A CN106682451 A CN 106682451A
- Authority
- CN
- China
- Prior art keywords
- formula
- composition
- biological tissue
- neural network
- dielectric property
- 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.)
- Granted
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/004—Artificial life, i.e. computing arrangements simulating life
- G06N3/006—Artificial life, i.e. computing arrangements simulating life based on simulated virtual individual or collective life forms, e.g. social simulations or particle swarm optimisation [PSO]
-
- 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
- G06N3/048—Activation functions
-
- G—PHYSICS
- G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
- G16B—BIOINFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR GENETIC OR PROTEIN-RELATED DATA PROCESSING IN COMPUTATIONAL MOLECULAR BIOLOGY
- G16B5/00—ICT specially adapted for modelling or simulations in systems biology, e.g. gene-regulatory networks, protein interaction networks or metabolic networks
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Health & Medical Sciences (AREA)
- Life Sciences & Earth Sciences (AREA)
- General Health & Medical Sciences (AREA)
- Biophysics (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- Artificial Intelligence (AREA)
- Data Mining & Analysis (AREA)
- Computational Linguistics (AREA)
- Biomedical Technology (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Evolutionary Computation (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Physiology (AREA)
- Bioinformatics & Computational Biology (AREA)
- Biotechnology (AREA)
- Evolutionary Biology (AREA)
- Medical Informatics (AREA)
- Spectroscopy & Molecular Physics (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
- Investigating Or Analyzing Materials By The Use Of Electric Means (AREA)
Abstract
The invention discloses a formula proportion determining method for biological tissue simulation material and a system. The system comprises an obtaining unit and a processing unit. The method comprises the following steps: firstly, obtaining the dielectric property fitting function corresponding to each ingredient in the formula and dielectric property fitting function of the biological tissue simulation material to be prepared, and inputting the obtained dielectric property fitting functions to a BP neural network for data processing, and outputting the proportion value of each ingredient in the formula. Through using the method and system, the excessive manual operation does not need to be involved in the process of preparing the biological tissue simulation material, and multiple times of the experiments are not needed, so the preparation has the characteristics of cheap cost, less time, high accuracy, convenience and efficiency. The method and system can be extensively applied to the preparation field of the biological tissue simulation material.
Description
Technical field
It is the present invention relates to the technology of preparing of biological tissue's simulation material more particularly to a kind of in biological tissue's simulation material
Formula proportion in preparation process determines method and system.
Background technology
The simulation material of imitated biological tissue's dielectric property (such as dielectric constant, electrical conductivity) is in bioelectromagnetic dosimetry, magnetic
Vital effect is served in resonance image-forming, ultra sonic imaging and other medical domain development.Meanwhile, with science and technology it is fast
Exhibition is hailed, the demand of biological tissue's simulation material is increasing, therefore, the preparation efficiency of biological microstructure modeling material is required
More and more higher.However, at present for the method for preparing biological tissue's simulation material, it mainly using test method(s) is repeated, that is, leads to
After crossing preparation several samples, by many experiments the ratio for preparing each composition in formula of biological tissue's simulation material is just can determine that
Example, so can then have that preparation cost is too high, experimental period length and cause that preparation efficiency is low, easily caused essence by environmental disturbances
Spend the defect such as low, it is difficult to meet increasing demand, and have impact on the further development of biological tissue's simulation material
And application.
The content of the invention
In order to solve above-mentioned technical problem, it is an object of the invention to provide a kind of recipe ratio of biological tissue's simulation material
Example determines method, make biological tissue's simulation material preparation can have with low cost, required time less, precision is high, convenience is fast
Prompt the advantages of.
It is a further object of the present invention to provide a kind of formula proportion of biological tissue's simulation material determines system, biological group is made
Knit the preparation of simulation material can have the advantages that with low cost, high precision, required time less, it is convenient and swift.
The technical solution used in the present invention is:A kind of formula proportion of biological tissue's simulation material determines method, the party
The step of method, includes:
The biological tissue for obtaining the dielectric property fitting function in formula corresponding to each composition and required preparation simulates
The dielectric property fitting function of material;
The dielectric property fitting function of acquisition is input into into BP neural network carries out data processing, so as to export formula in
The ratio value of each composition.
Further, the described dielectric property fitting function of acquisition is input into into BP neural network carries out data processing, from
And export formula in each composition ratio value the step for before be provided with the step for setting up BP neural network.
Further, described the step for set up BP neural network, it includes:
The training data of BP neural network is obtained, wherein, the training data includes training input data and training output
Data, the training input data includes the dielectric property fitting function in formula corresponding to each composition and required preparation
The dielectric property fitting function of biological tissue's simulation material, the training output data includes the ratio of each composition in formula most
The figure of merit, then using the initial weight of the BP neural network for obtaining, initial threshold and training data, so as to BP neural network
It is trained, terminates to require until meeting training;
BP neural network of the BP neural network obtained after training is terminated as required foundation.
Further, the initial weight and initial threshold of the BP neural network is obtained using fish-swarm algorithm.
Further, the described dielectric property fitting function of acquisition is input into into BP neural network carries out data processing, from
And export formula in each composition ratio value the step for after be provided with judgment processing steps, the judgment processing steps are concrete
For:
Whether the ratio value for judging each composition in the formula for exporting meets default judgment criteria, if so, then terminates, instead
It, then return and re-execute the step for setting up BP neural network.
Further, the judgment processing steps are specifically included:
The input factor is obtained, wherein, the input factor includes the dielectric property fitting in formula corresponding to each composition
The ratio value of each composition in function and the formula of output;
Obtain the output factor, wherein, biological tissue's simulation material of the output factor including required preparation with it is real
Dielectric property error between biological tissue;
According to the input factor and the output factor that obtain, the Judgement Matrix from the input factor to the output factor is set up;
According to maximum subjection principle, the maximum degree of membership of numerical value is selected from Judgement Matrix, then selected described in judgement
Whether degree of membership is less than preset value, if, then it represents that the ratio value of each composition meets default judge and marks in the formula of output
Standard, now terminates, otherwise, it means that the ratio value of each composition does not meet default judgment criteria in the formula of output, returns
Re-execute the step for setting up BP neural network.
Further, the dielectric property fitting function corresponding to the composition is the dielectric property by composition at different frequencies
The curvilinear function for being fitted.
Further, the dielectric property includes dielectric constant and/or electrical conductivity.
Another technical scheme for being taken of the present invention is:A kind of formula proportion of biological tissue's simulation material determines system,
The system is included:
Acquiring unit, for obtaining formula in dielectric property fitting function corresponding to each composition and required preparation
The dielectric property fitting function of biological tissue's simulation material;
Processing unit, for the dielectric property fitting function of acquisition to be input into into BP neural network data processing is carried out,
The ratio value of each composition in so as to export formula.
Further, to be provided with before the processing unit and set up unit for set up BP neural network.
Further, the unit of setting up includes:
Training module, for obtaining the training data of BP neural network, wherein, the training data includes training input number
According to training output data, the training input data include dielectric property fitting function in formula corresponding to each composition with
And the dielectric property fitting function of biological tissue's simulation material of required preparation, the training output data include formula in each
The ratio optimal value of composition, then using the initial weight of the BP neural network for obtaining, initial threshold and training data, so as to
BP neural network is trained, terminates to require until meeting training;
BP neural network acquisition module, the BP neural network for obtaining after training is terminated is refreshing as the BP of required foundation
Jing networks.
Further, the initial weight and initial threshold of the BP neural network is obtained using fish-swarm algorithm.
Further, judgement processing unit is provided with after the processing unit, the judgement processing unit is used to judge output
Formula in the ratio value of each composition whether meet default judgment criteria, if so, then terminate, conversely, then return holding again
Row sets up the flow chart of data processing corresponding to unit.
Further, the judgement processing unit is specifically included:
Input factor acquisition module, for obtaining the input factor, wherein, the input factor includes each composition in formula
The ratio value of each composition in corresponding dielectric property fitting function and the formula of output;
Output factor acquisition module, for obtaining the output factor, wherein, the output factor includes the biology of required preparation
Dielectric property error between microstructure modeling material and real biological tissue;
Judgement Matrix sets up module, for according to the input factor and the output factor for obtaining, setting up from the input factor to defeated
Go out the Judgement Matrix of the factor;
Judging treatmenting module, for according to maximum subjection principle, the maximum degree of membership of numerical value being selected from Judgement Matrix, so
Whether the degree of membership selected described in judging afterwards is less than preset value, if, then it represents that the ratio value of each composition in the formula of output
Meet default judgment criteria, now terminate, otherwise, it means that output formula in each composition ratio value do not meet it is default
Judgment criteria, return re-execute the flow chart of data processing set up corresponding to unit.
Further, the dielectric property fitting function corresponding to the composition is the dielectric property by composition at different frequencies
The curvilinear function for being fitted.
Further, the dielectric property includes dielectric constant and/or electrical conductivity.
The invention has the beneficial effects as follows:By using the method for the present invention, in preparation process, it is only necessary to each in formula
Dielectric property fitting function corresponding to individual composition is acquired acquisition, and obtains the known required biological tissue's mould for preparing
After intending the dielectric property fitting function of material, they are input into carries out data processing into the BP neural network for training, and just can
The ratio value of each composition is obtained, such producer is just capable of achieving biological tissue according to the ratio value of each composition for being exported
The making of simulation material.As can be seen here, scheme is determined compared to traditional formula proportion, the method for the present invention need not be related to excessively
Manual operation, without many experiments, make biological tissue's simulation material preparation have with low cost, required time less, essence
Accuracy is high, it is convenient and swift the advantages of.
The present invention another beneficial effect be:By using the system of the present invention, in preparation process, it is only necessary to obtain single
Unit acquires the dielectric property fitting function in formula corresponding to each composition, and obtains the known required biology for preparing
The dielectric property fitting function of microstructure modeling material, then processing unit the dielectric property fitting function of acquisition is input into training
Carry out data processing in good BP neural network, so as to export formula in each composition ratio value, such producer is according to institute
The ratio value of each composition of output, is just capable of achieving the making of biological tissue's simulation material.As can be seen here, match somebody with somebody compared to traditional
Square ratio-dependent scheme, the system of the present invention need not be related to excessive manual operation, without many experiments, make biological tissue's mould
Intend material preparation have the advantages that with low cost, required time less, precision it is high, convenient and swift.
Description of the drawings
Fig. 1 is the step of a kind of formula proportion of biological tissue's simulation material of the invention determines method schematic flow sheet;
Fig. 2 is the specific embodiment step stream that a kind of formula proportion of biological tissue's simulation material of the invention determines method
Journey schematic diagram;
Fig. 3 is a specific embodiment steps flow chart schematic diagram of the step for BP neural network is set up in Fig. 2;
Fig. 4 is a specific embodiment steps flow chart schematic diagram of judgment processing steps in Fig. 2;
Fig. 5 is the structural frames schematic diagram that a kind of formula proportion of biological tissue's simulation material of the invention determines system;
Fig. 6 is the specific embodiment structural frames that a kind of formula proportion of biological tissue's simulation material of the invention determines system
Schematic diagram;
Fig. 7 is the specific embodiment structural frames schematic diagram that unit is set up described in Fig. 6;
Fig. 8 is the specific embodiment structural frames schematic diagram that processing unit is judged described in Fig. 6.
Specific embodiment
As shown in figure 1, a kind of formula proportion of biological tissue's simulation material is included the step of determining method, the method:
The biological tissue for obtaining the dielectric property fitting function in formula corresponding to each composition and required preparation simulates
The dielectric property fitting function of material;
The dielectric property fitting function of acquisition is input into into BP neural network carries out data processing, so as to export formula in
The ratio value of each composition.Wherein, biological tissue's simulation material of preparation prepares formula needed for described formula is referred to.
The preferred implementation of this method embodiment is further used as, for the dielectric property corresponding to described composition is intended
Function is closed, it is the curvilinear function that the dielectric property by composition at different frequencies is fitted, i.e., described composition is in difference
Different dielectric properties can be corresponded under frequency, institute after dielectric property corresponding at different frequencies carries out curve fitting to these
The curvilinear function for drawing, its dielectric property fitting function just corresponding to the composition.Equally, for the required preparation
The dielectric property fitting function of biological tissue's simulation material, it is in different frequency by biological tissue's simulation material of required preparation
Under the curvilinear function that fitted of dielectric property.
The preferred implementation of this method embodiment is further used as, the dielectric property includes dielectric constant and/or electricity
Conductance, i.e., for the dielectric property fitting function corresponding to the composition, its specifically included dielectric constant fitting function and/or
Electrical conductivity fitting function.From the foregoing, for the dielectric constant fitting function corresponding to the composition, it is specifically, described
Composition correspond to different dielectric constants at different frequencies, the dielectric constant march corresponding at different frequencies to these
The curvilinear function drawn after line fitting;Equally, for described electrical conductivity fitting function, it is specifically, described composition exists
The different electrical conductivity of correspondence under different frequency, gained after electrical conductivity corresponding at different frequencies carries out curve fitting to these
The curvilinear function for going out.In addition, the dielectric property fitting function of the biological tissue's simulation material for the required preparation, its is concrete
Including the dielectric constant fitting function and electrical conductivity fitting function of biological tissue's simulation material of required preparation.
The preferred implementation of this method embodiment is further used as, it is described by the dielectric property fitting function for obtaining input
Data processing is carried out into BP neural network, so as to export formula in each composition ratio value the step for before be provided with foundation
The step for BP neural network.And for it is described set up BP neural network the step for, and it is described acquisition formula in each into
The dielectric property fitting function of point corresponding dielectric property fitting function and biological tissue's simulation material of required preparation this
One step, the priority logical order between them can be adjusted according to the actual requirements, it is preferable that it is described set up BP neural network this
One step is arranged on the biology of the dielectric property fitting function in the acquisition formula corresponding to each composition and required preparation
Before the step for dielectric property fitting function of microstructure modeling material.
The preferred implementation of this method embodiment is further used as, it is described by the dielectric property fitting function for obtaining input
Data processing is carried out into BP neural network, so as to export formula in each composition ratio value the step for after be provided with judgement
Process step, the judgment processing steps are specially:
Whether the ratio value for judging each composition in the formula for exporting meets default judgment criteria, if so, then terminates, instead
It, then return and re-execute the step for setting up BP neural network.
The acquisition of the dielectric property fitting function in embodiment 1, formula corresponding to each composition
In the present embodiment, for the formula for preparing biological tissue's simulation material, the included composition of its preferred determination has
Deionized water, gelatin, agar, Sodium Chloride, aluminium powder, edible oil, abluent, totally 7 kinds of compositions.Wherein, for described 7 kinds into
Point, they have respectively other functions in addition to participating in adjusting the dielectric property of biological tissue's simulation material, also, for example:1st, go
Ionized water, it is used to dissolve gelatin, agar and Sodium Chloride;2nd, gelatin, it plays a part of gel;3rd, agar, it is used to carry
The fusing point of high biological tissue's simulation material;4th, Sodium Chloride and aluminium powder, they can significantly change the electricity of biological tissue's simulation material
Conductance;5th, edible oil, it can significantly change the dielectric constant of biological tissue's simulation material;6th, abluent, it is used for reduction and goes
Surface tension between ionized water and edible oil, promotes the fusion between each composition in formula.In addition, in the present embodiment, institute
Electrical characteristics fitting function is given an account of including dielectric constant fitting function and electrical conductivity fitting function.
For 7 kinds of compositions in above-mentioned formula, the dielectric constant fitting function and electrical conductivity fitting corresponding to each of which
Function, its obtaining step is comprised preferably:
The first step, gather each composition dielectric constant at different frequencies and electrical conductivity;
For each composition in formula, their dielectric property, i.e. dielectric constant and electrical conductivity, impedance analysis can be passed through
Measuring, concrete measuring method includes for instrument and coaxial probe:1st, for deionized water, edible oil and abluent, due to these three
Composition is liquid, therefore coaxial probe can be immersed in these three compositions respectively, is then obtained by electric impedance analyzer
Its dielectric constant and electrical conductivity at different frequencies;2nd, for Sodium Chloride and aluminium powder, because its form is powder, therefore can
Directly coaxial probe is pressed against above powder, so as to obtain its dielectric constant and electrical conductivity at different frequencies;3rd, for
Gelatin and agar, because it is solid, and its rough surface, therefore need to be pulverized for after powder according to above-mentioned for chlorination
The measuring method of sodium and aluminium powder measuring, so as to obtain their dielectric constants and electrical conductivity at different frequencies;
Preferably, in above-mentioned gatherer process, every kind of composition dielectric constant at different frequencies and electrical conductivity should repeat
Measurement more than 20 times, and the number of data point acquired every time is 800;
Second step, each composition dielectric constant at different frequencies for obtaining and electrical conductivity are fitted, so as to obtain
Obtain the dielectric constant fitting function and electrical conductivity fitting function corresponding to each composition;
Wherein, it is preferred to use method of least square to each composition dielectric constant at different frequencies and electrical conductivity carrying out
Fitting, concrete fit procedure is as follows:
Step 1. draws rough figure scatterplot according to the data obtained by electric impedance analyzer, is dissipated according to figure
Point diagram determines the frequency n of polynomial fitting;
Step 2. is according to the data point (x that need to be fittedi,yj), wherein i=0,1 ..m, m are the number of the data point of collection,
Such as 800, then calculateOrder
Even the first formulaFor minima;
Step 3. it can be seen from above-mentioned first formula,For a0,a1,a2...anThe function of many variables, root
The condition of extreme value is sought according to the function of many variables, to a0,a1,a2...anCarry out seeking derivative operation, the second formula can be obtained as follows:
There is the 3rd formula as follows:
It can be seen from above-mentioned 3rd formula, the 3rd formula is with regard to a to step 4.0,a1,a2...anSystem of linear equations, lead to
Cross and equation group is solved, just can respectively obtain a0,a1,a2...anValue;
Step 5. is according to a0,a1,a2...anValue, just can obtain dielectric constant fitting function and the electricity corresponding to each composition
Conductance fitting function, as shown in the 4th formula:
It can be seen that, according to above-mentioned steps 1-5 and the 7 kinds of compositions for being obtained dielectric constant at different frequencies and conductance
Rate, just can respectively to the dielectric constant of seven kinds of compositions such as deionized water, gelatin, agar, Sodium Chloride, aluminium powder, edible oil, abluent
Least square fitting is carried out with electrical conductivity, corresponding fitting function is obtained.
In addition, for the dielectric constant fitting function and electrical conductivity of biological tissue's simulation material of the required preparation are fitted
Function, they are known parameters, and direct access just may be used.
Embodiment 2, the formula proportion of biological tissue's simulation material determine method
As shown in Fig. 2 a kind of formula proportion of biological tissue's simulation material determines method, it comprises the following steps that shown.
Step S1, set up BP neural network.
As shown in figure 3, step S1 includes:
The setting of S101, BP neural network;
The BP neural network includes input layer, hidden layer and output layer, and input signal is passed successively from input layer
To each hidden layer node, output layer output is finally passed to;
In the present embodiment, the nodes of input layer are set as 16, deionized water, gelatin, agar, chlorine is corresponded respectively
Change sodium, aluminium powder, edible oil, abluent, each corresponding dielectric constant fitting function of this 7 kinds of compositions and electrical conductivity fitting letter
Number, and the dielectric constant fitting function and electrical conductivity fitting function of biological tissue's simulation material of required preparation, i.e., described BP
The training input data of neutral net includes the dielectric constant fitting function and electrical conductivity fitting in formula corresponding to each composition
Function, and the dielectric constant fitting function and electrical conductivity fitting function of biological tissue's simulation material of required preparation;Setting is defeated
The node for going out layer is 1, and its value is formula proportion optimum, i.e., the training output data of described BP neural network be in formula each
The ratio optimal value of composition;The number of plies of hidden layer is set as 1, its interstitial content is then preferably determined using below equation:
J=log2(I)
Wherein, I is the interstitial content of input layer, and in the present embodiment, its value is the interstitial content that 16, J is hidden layer, because
This, J's is sized to 4;
S102, the initial weight and initial threshold that obtain BP neural network, then using the first of the BP neural network for obtaining
Beginning weights, initial threshold and training data, so as to be trained to the BP neural network of above-mentioned setting, until meeting training knot
Till beam request;
Wherein, for the initial weight and initial threshold of the BP neural network, it is preferably obtained using fish-swarm algorithm,
Concrete obtaining step is included:
S1021, the dimension for determining Artificial Fish, the dimension of the Artificial Fish includes the weights and threshold value, i.e. H of BP neural network
=H (v11,...vI1,u11,...,uk1,w1k,...,wpk,θk);
S1022, the parameter to fish-swarm algorithm are initialized, wherein, fish-swarm algorithm needs initialized parameter to include kind
Group's size fish_size, the perceived distance fish_dis tange of Artificial Fish, maximum step-length fish_step of Artificial Fish movement,
Crowding factor delta, maximum iteration time max_gen and desired value G;
S1023, primary iteration number of times k=0 is set, and the individual of fish_size Artificial Fish is randomly generated in feasible zone
Body, forms the initial shoal of fish, and each component be (- 1, the 1) random number in interval;
Food concentration value Y of each Artificial Fish individuality current location in S1024, the initial shoal of fish of calculating, and compare size, retain
Maximum enters bulletin board;
S1025, simulate foraging behavior respectively to each Artificial Fish, bunch behavior and behavior of knocking into the back, so as to obtain Artificial Fish point
Food concentration value that Mo Ni be not resulting after foraging behavior, bunch behavior and behavior of knocking into the back, then selects the maximum food of numerical value
Concentration value, and using the behavior corresponding to the food concentration value that this is selected as actual execution;
Wherein, the foraging behavior of Artificial Fish, bunch behavior and behavior of knocking into the back computational methods it is as follows:
1., the computational methods of foraging behavior
If the current state of Artificial Fish is Fi, in its (d within sweep of the eyei,j≤ fish_dis tange) random selection
One state FjIf, Yi<Yj, then take a step forward to the direction, following 5th formula is performed, the 5th formula is as follows:
fi,j(k+1)=fi,j(k)+rand·step(Fi,j-xi,j(k))
Conversely, then randomly choosing state F againi, judge whether to meet advance condition;After repeating default number of attempt, such as
Fruit is still unsatisfactory for condition, then the 6th formula below execution formula, and the 6th formula is as follows:
xi,j(k+1)=xi,j(k)+rand·step
2., bunch the computational methods of behavior
If the current state of Artificial Fish is Fi, it is n to explore the number of partners in visible domainf, form set ΦiIf, Φi≠
φ, then explore center X in setcenter, calculate food concentration value F of the centercenterIf metBehavior formula of bunching then is performed, as shown in following 7th formula:
xi,j(k+1)=xi,j(k)+rand·step(xcenter,j-xi,j(k))
3., knock into the back the computational methods of behavior
If the current state of Artificial Fish is Fi, it is n to explore the number of partners in visible domainf, form set ΦiIf, Φi≠
φ, then gather the interior partner X for exploring food concentration maximummax, calculate food concentration value F at thismaxIf met
The behavior formula that knocks into the back then is performed, as shown in following 8th formula:
xi,j(k+1)=xi,j(k)+rand·step(xmax,j-xi,j(k))
S1026, judge that food concentration value elected, whether more than the food concentration value on bulletin board, if so, then will
Food concentration value elected replaces the food concentration value on bulletin board, conversely, the food concentration value on bulletin board is then constant;
S1027, judge whether to meet termination condition, specifically, judge iterationses whether reach maximum iteration time or
Person judges whether the trueness error for having met solution, if so, then using the individual information corresponding to current each Artificial Fish as obtaining
The initial weight for obtaining and initial threshold, conversely, then return that step S1025 is re-executed, until meeting termination condition;
S103, training is terminated after the BP neural network that obtains as required foundation BP neural network.
Step S2, in the biological tissue's simulation material needed for preparing, dielectric is carried out to current ready each composition
The acquisition of fitting of constant function and electrical conductivity fitting function, and obtain the known required simulation material institute of biological tissue for preparing
Corresponding dielectric constant fitting function and electrical conductivity fitting function, then intend the dielectric constant fitting function for obtaining and electrical conductivity
Conjunction function is input into and carries out exporting the ratio value of each composition after data processing into the above-mentioned BP neural network built up.
Further, since in the formula that exported of BP neural network each composition ratio value, i.e. formula proportion may have
The situation of irrationality, the therefore in order that ratio value of each composition for finally giving is more accurate, more closing to reality, in institute
State and be provided with after step S2 judgment processing steps S3, step S3 is specially:
Whether the ratio value for judging each composition in the formula for exporting meets default judgment criteria, if so, then terminates, institute
The formula proportion for stating output is finally to determine the formula proportion for obtaining, conversely, then return re-execute set up BP neural network this
Step S102 in one step, is recalculated using fish-swarm algorithm and draws new initial weight and initial threshold, is then based on new
Initial weight and initial threshold BP neural network is trained again, until meet training terminate require till, then will
The dielectric property fitting function of acquisition is input into carries out data processing into the BP neural network for newly training;
Or, whether the ratio value for judging each composition in the formula for exporting meets default judgment criteria, if so, then ties
Beam, the formula proportion of the output is finally to determine the formula proportion for obtaining, conversely, then return to re-execute setting up BP nerve net
The step for network, the training data of BP neural network is reacquired, recalculated using fish-swarm algorithm and draw new initial weight
And initial threshold, new training data, initial weight and initial threshold are then based on, BP neural network is trained again,
Till until meeting training end requirement, then the dielectric property fitting function of acquisition is input into the BP nerve net for newly training
Data processing is carried out in network.
As the preferred implementation of above-mentioned steps S3, as shown in figure 4, it is specifically included:
S301, the acquisition input factor, wherein, the input factor includes the dielectric constant in formula corresponding to each composition
The ratio value of each composition in the formula of fitting function, electrical conductivity fitting function and output;Specifically, the input factor can table
It is shown as:
U={ u1,u2,u3,u4,u5,u6,u7,u8,u9,u10,u11,u12,u13,u14,u15,u16,u17,u18,u19,u20,u21}
Wherein, { u1,u2,u3Be respectively dielectric constant fitting function corresponding to deionized water, electrical conductivity fitting function and
Ratio value shared by it, { u4,u5,u6Be respectively dielectric constant fitting function corresponding to gelatin, electrical conductivity fitting function and its
Shared ratio value, { u7,u8,u9It is respectively dielectric constant fitting function corresponding to agar, electrical conductivity fitting function and its institute
The ratio value for accounting for, { u10,u11,u12Be respectively dielectric constant fitting function corresponding to Sodium Chloride, electrical conductivity fitting function and its
Shared ratio value, { u13,u14,u15Be respectively dielectric constant fitting function corresponding to aluminium powder, electrical conductivity fitting function and its
Shared ratio value, { u16,u17,u18Be respectively dielectric constant fitting function corresponding to edible oil, electrical conductivity fitting function and
Ratio value shared by it, { u19,u20,u21The dielectric constant fitting function, the electrical conductivity fitting function that are respectively corresponding to abluent
And its shared ratio value;
S302, obtain the output factor, wherein, biological tissue's simulation material of the output factor including required preparation with it is true
Dielectric property error between real biological tissue;Specifically, the output factor is represented by:
V={ v1,v2,v3,v4,v5,v6,v7,v8,v9,v10}
Wherein, v1Representative errors are less than 10%, v2Representative errors are represented less than 20%, v3Representative errors are less than 30%, according to this
Analogize;
S303, the input factor according to acquisition and the output factor, set up the Judgement Matrix from the input factor to the output factor,
The Judgement Matrix is as follows:
S304, judge whether to need to be normalized Judgement Matrix, if so, then Judgement Matrix is normalized
Execution step S305 after process, conversely, then direct execution step S305;
S305, according to maximum subjection principle, the maximum degree of membership of numerical value is selected from Judgement Matrix, then judge the choosing
Whether the degree of membership for going out is less than preset value, and in the present embodiment, the preset value is 20%, if, then it represents that in the formula of output
The ratio value of each composition meets default judgment criteria, i.e., the ratio of each composition in the formula that described BP neural network is exported
Example value is rational, is now terminated, and the formula proportion of the output is finally to determine the formula proportion for obtaining, and producer is according to this
One formula proportion come the biological tissue's simulation material needed for preparing just can, otherwise, it means that each composition in the formula of output
Ratio value does not meet default judgment criteria, now, then returns and re-executes step S102, is recalculated using fish-swarm algorithm
Go out new initial weight and initial threshold, be then based on new initial weight and initial threshold is instructed to BP neural network again
Practice, train till terminating requirement until meeting, process is re-started using the BP neural network for newly training, or, then return
Return and re-execute the step for setting up BP neural network S1, the training data of BP neural network is reacquired, using fish-swarm algorithm
Recalculate and draw new initial weight and initial threshold, be then based on new training data, initial weight and initial threshold, weight
Newly BP neural network is trained, is trained till terminating requirement until meeting, using the BP neural network for newly training come weight
Newly processed.
Technical characteristic described in said method embodiment is suitable for following system embodiment.
Embodiment 3, the formula proportion of biological tissue's simulation material determine system
Method is determined based on the formula proportion of above-mentioned biological tissue's simulation material, system corresponding thereto is set up, i.e., one
The determination system of biological tissue's simulation material is planted, as shown in figure 5, it is included:
Acquiring unit 401, for obtaining formula in dielectric property fitting function corresponding to each composition and required system
The dielectric property fitting function of standby biological tissue's simulation material;
Processing unit 402, is carried out at data for the dielectric property fitting function of acquisition to be input into into BP neural network
Reason, so as to export formula in each composition ratio value.For described acquiring unit 401 and processing unit 402, they can lead to
Cross processor to realize its function.
The preferred implementation of the system embodiment is further used as, the dielectric property fitting function corresponding to the composition
It is curvilinear function that the dielectric property by composition at different frequencies is fitted.
The preferred implementation of the system embodiment is further used as, the dielectric property includes dielectric constant and/or electricity
Conductance.
The preferred implementation of the system embodiment is further used as, as shown in fig. 6, setting before the processing unit 402
Having set up unit 400 for set up BP neural network.For the order set up between unit 400 and acquiring unit 401
Relation, their priority logical order can set according to practical situation, and preferably, the unit 400 of setting up is located at acquisition
Before unit 401.
The preferred implementation of the system embodiment is further used as, as shown in fig. 7, the unit 400 of setting up includes:
Training module 4001, for obtaining the training data of BP neural network, wherein, the training data includes that training is defeated
Enter data and training output data, the training input data includes the dielectric property fitting letter in formula corresponding to each composition
The dielectric property fitting function of biological tissue's simulation material of several and required preparation, the training output data is included in formula
The ratio optimal value of each composition, then using the initial weight of the BP neural network for obtaining, initial threshold and training data,
So as to be trained BP neural network, terminate to require until meeting training;
BP neural network acquisition module 4002, for the BP neural network that obtains after training is terminated as required foundation
BP neural network.
It is further used as the preferred implementation of the system embodiment, the initial weight of the BP neural network and initial threshold
Value is obtained using fish-swarm algorithm.
The preferred implementation of the system embodiment is further used as, as shown in fig. 6, setting after the processing unit 402
Have and judge processing unit 403, it is described judge processing unit 403 for each composition in the formula for judging output ratio value whether
Meet default judgment criteria, if so, then terminate, conversely, then return re-executing the data processing set up corresponding to unit 400
Flow process.
The preferred implementation of the system embodiment is further used as, as shown in figure 8, the judgement processing unit 403 has
Body includes:
Input factor acquisition module 4031, for obtaining the input factor, wherein, the input factor includes in formula each
Dielectric property fitting function corresponding to composition and in the formula of output each composition ratio value;
Output factor acquisition module 4032, for obtaining the output factor, wherein, the output factor includes required preparation
Dielectric property error between biological tissue's simulation material and real biological tissue;
Judgement Matrix sets up module 4033, for according to the input factor and the output factor for obtaining, setting up from the input factor
To the Judgement Matrix of the output factor;
Judging treatmenting module 4034, for according to maximum subjection principle, selecting from Judgement Matrix, numerical value is maximum to be subordinate to
Whether degree, the degree of membership then selected described in judgement is less than preset value, if, then it represents that the ratio of each composition in the formula of output
Example value meets default judgment criteria, now terminates, otherwise, it means that the ratio value of each composition does not meet in the formula of output
Default judgment criteria, return re-executes the flow chart of data processing set up corresponding to unit 400.
For above-mentioned judging treatmenting module 4034, it is specifically for judging whether to need to be normalized Judgement Matrix at place
Reason, performs after being if so, then normalized to Judgement Matrix and passes judgment on step, conversely, then directly perform that step is passed judgment on, it is described
Step is passed judgment on specifically, according to maximum subjection principle, the maximum degree of membership of numerical value is selected from Judgement Matrix, then judges described
Whether the degree of membership selected is less than preset value, if, then it represents that the ratio value of each composition meets default in the formula of output
Judgment criteria, now terminates, otherwise, it means that the ratio value of each composition does not meet default judge mark in the formula of output
Standard, return re-executes the flow chart of data processing set up corresponding to unit 400, recalculated using fish-swarm algorithm draw it is new just
Beginning weights and initial threshold, are then based on new initial weight and initial threshold is trained to BP neural network again, until
Till meeting training end requirement, or, the training data of BP neural network is reacquired, recalculated using fish-swarm algorithm
Go out new initial weight and initial threshold, be then based on new training data, initial weight and initial threshold, it is again neural to BP
Network is trained, and trains till terminating requirement until meeting.
Obtained by above-mentioned, by using the method and system of the invention described above, during biological tissue's simulation material,
Producer only needs the dielectric property fitting function in the formula that will be collected corresponding to each composition, and the life of required preparation
The dielectric property fitting function of thing microstructure modeling material, be input into is carried out after data processing, just into the BP neural network for training
The ratio value of each composition can be obtained, such producer is just capable of achieving biological tissue's simulation material according to the formula proportion for being obtained
The making of material.In addition, the dielectric constant fitting function and electrical conductivity fitting function corresponding to each composition is present invention employs, with
And the dielectric constant fitting function and electrical conductivity fitting function of biological tissue's simulation material of required preparation, it is used as BP nerve net
The training input data of network, when can so make subsequent applications, the ratio value that BP neural network is exported, should closer to practical situation
It is higher with practicality, and the obtaining step of the dielectric constant fitting function corresponding to each composition and electrical conductivity fitting function
Fairly simple, it is high that convenience is implemented in operation;Further, the initial threshold and weights of BP neural network are obtained using fish-swarm algorithm,
The training process of BP neural network can be made more rapidly;The ratio value of each composition for exporting is entered using this mode of Judgement Matrix
Row reasonability judges, not only step is simple for it, be easily achieved, and the accuracy for judging is higher.
It is more than that the preferable enforcement to the present invention is illustrated, but the enforcement is not limited to the invention
Example, those of ordinary skill in the art can also make a variety of equivalent variations on the premise of without prejudice to spirit of the invention or replace
Change, the deformation or replacement of these equivalents are all contained in the application claim limited range.
Claims (16)
1. a kind of formula proportion of biological tissue's simulation material determines method, it is characterised in that:The step of the method, includes:
Biological tissue's simulation material of dielectric property fitting function and required preparation in acquisition formula corresponding to each composition
Dielectric property fitting function;
The dielectric property fitting function of acquisition is input into into BP neural network carries out data processing, so as to export formula in each
The ratio value of composition.
2. according to claim 1 a kind of formula proportion of biological tissue's simulation material determines method, it is characterised in that:It is described
The dielectric property fitting function of acquisition is input into into BP neural network carries out data processing, so as to export formula in each composition
Ratio value the step for before be provided with the step for setting up BP neural network.
3. according to claim 2 a kind of formula proportion of biological tissue's simulation material determines method, it is characterised in that:It is described
The step for setting up BP neural network, it includes:
The training data of BP neural network is obtained, wherein, the training data includes training input data and training output data,
The training input data includes the biology of the dielectric property fitting function in formula corresponding to each composition and required preparation
The dielectric property fitting function of microstructure modeling material, the training output data includes that the ratio of each composition in formula is optimum
Value, then using the initial weight of the BP neural network for obtaining, initial threshold and training data, so as to enter to BP neural network
Row training;
BP neural network of the BP neural network obtained after training is terminated as required foundation.
4. according to claim 3 a kind of formula proportion of biological tissue's simulation material determines method, it is characterised in that:It is described
The initial weight and initial threshold of BP neural network is obtained using fish-swarm algorithm.
5. a kind of formula proportion of biological tissue's simulation material determines method according to any one of claim 2-4, its feature
It is:The described dielectric property fitting function of acquisition is input into into BP neural network carries out data processing, so as to export formula
In each composition ratio value the step for after be provided with judgment processing steps, the judgment processing steps are specially:
Whether the ratio value for judging each composition in the formula for exporting meets default judgment criteria, if so, then terminates, conversely,
Then return and re-execute the step for setting up BP neural network.
6. according to claim 5 a kind of formula proportion of biological tissue's simulation material determines method, it is characterised in that:It is described
Judgment processing steps are specifically included:
The input factor is obtained, wherein, the input factor includes the dielectric property fitting function in formula corresponding to each composition
And output formula in each composition ratio value;
The output factor is obtained, wherein, the output factor includes biological tissue's simulation material and real biology of required preparation
Dielectric property error between tissue;
According to the input factor and the output factor that obtain, the Judgement Matrix from the input factor to the output factor is set up;
According to maximum subjection principle, the maximum degree of membership of numerical value is selected from Judgement Matrix, that what is then selected described in judgement is subordinate to
Whether degree is less than preset value, if, then it represents that the ratio value of each composition meets default judgment criteria in the formula of output, this
When terminate, otherwise, it means that the ratio value of each composition does not meet default judgment criteria in the formula of output, return is held again
The step for row sets up BP neural network.
7. a kind of formula proportion of biological tissue's simulation material determines method according to any one of claim 1-4, its feature
It is:Dielectric property fitting function corresponding to the composition is that the dielectric property by composition at different frequencies is fitted
Curvilinear function.
8. according to claim 7 a kind of formula proportion of biological tissue's simulation material determines method, it is characterised in that:It is described
Dielectric property includes dielectric constant and/or electrical conductivity.
9. a kind of formula proportion of biological tissue's simulation material determines system, it is characterised in that:The system is included:
Acquiring unit, for obtaining formula in dielectric property fitting function corresponding to each composition and the biology of required preparation
The dielectric property fitting function of microstructure modeling material;
Processing unit, for the dielectric property fitting function of acquisition to be input into into BP neural network data processing is carried out, so as to
The ratio value of each composition in output formula.
10. according to claim 9 a kind of formula proportion of biological tissue's simulation material determines system, it is characterised in that:Institute
State to be provided with before processing unit and set up unit for set up BP neural network.
11. according to claim 10 a kind of formula proportion of biological tissue's simulation material determine system, it is characterised in that:Institute
State and set up unit and include:
Training module, for obtaining the training data of BP neural network, wherein, the training data include training input data and
Training output data, the training input data includes the dielectric property fitting function corresponding to each composition and institute in formula
The dielectric property fitting function of the biological tissue's simulation material that need to be prepared, the training output data includes each composition in formula
Ratio optimal value, then using the initial weight of BP neural network, initial threshold and training data for obtaining, so as to BP
Neutral net is trained;
BP neural network acquisition module, for the BP neural network that obtains after training is terminated as required foundation BP nerve net
Network.
A kind of 12. formula proportion of biological tissue's simulation material according to claim 11 determine system, it is characterised in that:Institute
The initial weight and initial threshold for stating BP neural network is obtained using fish-swarm algorithm.
A kind of 13. formula proportion of biological tissue's simulation material according to any one of claim 10-12 determine system, and it is special
Levy and be:Judgement processing unit is provided with after the processing unit, the judgement processing unit is used in the formula for judging output
Whether the ratio value of each composition meets default judgment criteria, if so, then terminates, conversely, then return re-executing foundation list
Flow chart of data processing corresponding to unit.
A kind of 14. formula proportion of biological tissue's simulation material according to claim 13 determine system, it is characterised in that:Institute
State and judge that processing unit is specifically included:
Input factor acquisition module, for obtaining the input factor, wherein, the input factor includes that each composition institute is right in formula
The ratio value of each composition in the dielectric property fitting function answered and the formula of output;
Output factor acquisition module, for obtaining the output factor, wherein, the output factor includes the biological tissue of required preparation
Dielectric property error between simulation material and real biological tissue;
Judgement Matrix sets up module, for according to obtain the input factor and output the factor, set up from input the factor to output because
The Judgement Matrix of son;
Judging treatmenting module, for according to maximum subjection principle, the maximum degree of membership of numerical value being selected from Judgement Matrix, then sentences
Whether the disconnected degree of membership selected is less than preset value, if, then it represents that the ratio value of each composition meets in the formula of output
Default judgment criteria, now terminates, otherwise, it means that the ratio value of each composition does not meet default commenting in the formula of output
Sentence standard, return re-executes the flow chart of data processing set up corresponding to unit.
A kind of 15. formula proportion of biological tissue's simulation material according to any one of claim 9-12 determine system, and it is special
Levy and be:Dielectric property fitting function corresponding to the composition is that the dielectric property by composition at different frequencies is fitted
Curvilinear function.
A kind of 16. formula proportion of biological tissue's simulation material according to claim 15 determine system, it is characterised in that:Institute
Electrical characteristics are given an account of including dielectric constant and/or electrical conductivity.
Priority Applications (2)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201611123843.6A CN106682451B (en) | 2016-12-08 | 2016-12-08 | A kind of formula rate of biological tissue's simulation material determines method and system |
PCT/CN2016/110365 WO2018103131A1 (en) | 2016-12-08 | 2016-12-16 | Method and system for determining formula proportions of biological tissue simulation material |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201611123843.6A CN106682451B (en) | 2016-12-08 | 2016-12-08 | A kind of formula rate of biological tissue's simulation material determines method and system |
Publications (2)
Publication Number | Publication Date |
---|---|
CN106682451A true CN106682451A (en) | 2017-05-17 |
CN106682451B CN106682451B (en) | 2019-05-14 |
Family
ID=58868255
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201611123843.6A Active CN106682451B (en) | 2016-12-08 | 2016-12-08 | A kind of formula rate of biological tissue's simulation material determines method and system |
Country Status (2)
Country | Link |
---|---|
CN (1) | CN106682451B (en) |
WO (1) | WO2018103131A1 (en) |
Cited By (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN109558664A (en) * | 2018-11-22 | 2019-04-02 | 广东工业大学 | A kind of compound material formula formulating method of injection molding manufacture |
CN117373582A (en) * | 2023-12-06 | 2024-01-09 | 小米汽车科技有限公司 | Material composition prediction method and device based on simulation experiment model and storage medium |
Families Citing this family (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN114527567B (en) * | 2021-12-13 | 2023-10-27 | 中国北方车辆研究所 | Design method of large-view-field lens protection glass |
Citations (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102258960A (en) * | 2011-05-26 | 2011-11-30 | 浙江迦南科技股份有限公司 | Method for automatically generating formula of fluidized bed based on neural network system |
CN104897449A (en) * | 2015-06-09 | 2015-09-09 | 南方医科大学 | Method for configuring human tissue dielectric property simulating material |
Family Cites Families (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN1876172A (en) * | 2005-06-02 | 2006-12-13 | 北京中医药大学 | BP neural net model method for preparing pharmaceutical composition |
US7945392B2 (en) * | 2005-11-14 | 2011-05-17 | Mclean Hospital Corporation | Methods and systems for drug screening and computational modeling based on biologically realistic neurons |
CN101620249B (en) * | 2009-07-27 | 2012-05-23 | 张莉 | Neural net method for measuring electromagnetic parameters of artificial electromagnetic material |
CN102508042B (en) * | 2011-10-19 | 2014-04-09 | 中国人民解放军第四军医大学 | Open-ended coaxial probe and method for measuring dielectric spectrum property of biological tissues |
CN104330642B (en) * | 2014-11-12 | 2017-03-22 | 中国人民解放军国防科学技术大学 | Probe for measuring anisotropy dielectric spectrum characteristics of biological tissues and measuring and calculating method thereof |
CN104573273B (en) * | 2015-01-27 | 2017-05-24 | 江西理工大学 | Building method of modified bioadsorbent structure-activity relationship model and application thereof |
-
2016
- 2016-12-08 CN CN201611123843.6A patent/CN106682451B/en active Active
- 2016-12-16 WO PCT/CN2016/110365 patent/WO2018103131A1/en active Application Filing
Patent Citations (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102258960A (en) * | 2011-05-26 | 2011-11-30 | 浙江迦南科技股份有限公司 | Method for automatically generating formula of fluidized bed based on neural network system |
CN104897449A (en) * | 2015-06-09 | 2015-09-09 | 南方医科大学 | Method for configuring human tissue dielectric property simulating material |
Non-Patent Citations (1)
Title |
---|
郭栋,等;: "基于人工神经网络的BATIO3陶瓷配方研究", 《硅酸盐学报》 * |
Cited By (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN109558664A (en) * | 2018-11-22 | 2019-04-02 | 广东工业大学 | A kind of compound material formula formulating method of injection molding manufacture |
CN109558664B (en) * | 2018-11-22 | 2023-04-18 | 广东工业大学 | Method for formulating composite material manufactured by injection molding |
CN117373582A (en) * | 2023-12-06 | 2024-01-09 | 小米汽车科技有限公司 | Material composition prediction method and device based on simulation experiment model and storage medium |
CN117373582B (en) * | 2023-12-06 | 2024-04-26 | 小米汽车科技有限公司 | Material composition prediction method and device based on simulation experiment model and storage medium |
Also Published As
Publication number | Publication date |
---|---|
CN106682451B (en) | 2019-05-14 |
WO2018103131A1 (en) | 2018-06-14 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN103955270B (en) | Character high-speed input method of brain-computer interface system based on P300 | |
CN108564611A (en) | A kind of monocular image depth estimation method generating confrontation network based on condition | |
CN110413924A (en) | A kind of Web page classification method of semi-supervised multiple view study | |
CN106682451A (en) | Formula proportion determining method for biological tissue simulation material and system | |
CN106897570A (en) | A kind of COPD test system based on machine learning | |
CN106570516A (en) | Obstacle recognition method using convolution neural network | |
CN109167996A (en) | It is a kind of based on convolutional neural networks without reference stereo image quality evaluation method | |
CN102496156A (en) | Medical image segmentation method based on quantum-behaved particle swarm cooperative optimization | |
CN105205504B (en) | A kind of image attention regional quality evaluation index learning method based on data-driven | |
CN109871124A (en) | Emotion virtual reality scenario appraisal procedure based on deep learning | |
CN113657449A (en) | Traditional Chinese medicine tongue picture greasy classification method containing noise labeling data | |
Xin et al. | Three‐dimensional reconstruction of Vitis vinifera (L.) cvs Pinot Noir and Merlot grape bunch frameworks using a restricted reconstruction grammar based on the stochastic L‐system | |
CN109388227A (en) | A method of user experience is predicted using eye movement data recessiveness | |
CN109358185A (en) | Rubber compounding performance prediction model and prediction technique based on extreme learning machine | |
CN112579708A (en) | Link prediction similarity index method based on improved local path | |
CN105513058B (en) | A kind of brain active region detection method and device | |
Gao et al. | Physical fitness clustering analysis based on self-organizing feature maps network | |
CN109408947A (en) | A kind of infringement webpage judgment method based on machine learning | |
CN103868821A (en) | Ultrasonic-based portable fish meat freshness detection device and evaluation method | |
CN108074240A (en) | Recognition methods, identification device, computer readable storage medium and program product | |
CN114519722A (en) | Carotid artery extraction method based on convolutional neural network | |
CN108921827A (en) | Stereo image quality evaluation method is referred to based on independent component analysis and the complete of binocular fusion characteristic | |
Lu et al. | Fatigue detection technology for online learning | |
Archambeault et al. | Quantifying the quality of agreement between simulation and validation data for multiple data sets | |
CN112150605B (en) | Knee joint model construction method for MRI local SAR estimation |
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 | ||
GR01 | Patent grant | ||
GR01 | Patent grant |