CN112101418B - Mammary tumor type identification method, system, medium and equipment - Google Patents

Mammary tumor type identification method, system, medium and equipment Download PDF

Info

Publication number
CN112101418B
CN112101418B CN202010823773.5A CN202010823773A CN112101418B CN 112101418 B CN112101418 B CN 112101418B CN 202010823773 A CN202010823773 A CN 202010823773A CN 112101418 B CN112101418 B CN 112101418B
Authority
CN
China
Prior art keywords
breast tumor
neural network
layer
dynamic convergence
sample
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.)
Active
Application number
CN202010823773.5A
Other languages
Chinese (zh)
Other versions
CN112101418A (en
Inventor
张智军
陈博钊
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
South China University of Technology SCUT
Original Assignee
South China University of Technology SCUT
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by South China University of Technology SCUT filed Critical South China University of Technology SCUT
Priority to CN202010823773.5A priority Critical patent/CN112101418B/en
Publication of CN112101418A publication Critical patent/CN112101418A/en
Application granted granted Critical
Publication of CN112101418B publication Critical patent/CN112101418B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/24—Classification techniques
    • G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
    • G06F18/2415—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches based on parametric or probabilistic models, e.g. based on likelihood ratio or false acceptance rate versus a false rejection rate
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
    • G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/04—Architecture, e.g. interconnection topology
    • G06N3/045—Combinations of networks
    • 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
    • G06N3/06—Physical realisation, i.e. hardware implementation of neural networks, neurons or parts of neurons
    • G06N3/061—Physical realisation, i.e. hardware implementation of neural networks, neurons or parts of neurons using biological neurons, e.g. biological neurons connected to an integrated circuit
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/08—Learning methods
    • G—PHYSICS
    • G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
    • G16H—HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA
    • G16H50/00—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics
    • G16H50/20—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics for computer-aided diagnosis, e.g. based on medical expert systems
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V2201/00—Indexing scheme relating to image or video recognition or understanding
    • G06V2201/03—Recognition of patterns in medical or anatomical images
    • G06V2201/032—Recognition of patterns in medical or anatomical images of protuberances, polyps nodules, etc.
    • Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
    • Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
    • Y02A—TECHNOLOGIES FOR ADAPTATION TO CLIMATE CHANGE
    • Y02A90/00—Technologies having an indirect contribution to adaptation to climate change
    • Y02A90/10—Information and communication technologies [ICT] supporting adaptation to climate change, e.g. for weather forecasting or climate simulation

Landscapes

  • Engineering & Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • Theoretical Computer Science (AREA)
  • Data Mining & Analysis (AREA)
  • Health & Medical Sciences (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Biomedical Technology (AREA)
  • Artificial Intelligence (AREA)
  • Biophysics (AREA)
  • Evolutionary Computation (AREA)
  • General Engineering & Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • Molecular Biology (AREA)
  • General Health & Medical Sciences (AREA)
  • Computational Linguistics (AREA)
  • Computing Systems (AREA)
  • Mathematical Physics (AREA)
  • Software Systems (AREA)
  • Medical Informatics (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Bioinformatics & Computational Biology (AREA)
  • Bioinformatics & Cheminformatics (AREA)
  • Public Health (AREA)
  • Evolutionary Biology (AREA)
  • Neurology (AREA)
  • Microelectronics & Electronic Packaging (AREA)
  • Databases & Information Systems (AREA)
  • Pathology (AREA)
  • Probability & Statistics with Applications (AREA)
  • Epidemiology (AREA)
  • Primary Health Care (AREA)
  • Medical Treatment And Welfare Office Work (AREA)
  • Image Analysis (AREA)

Abstract

The invention discloses a breast tumor type identification method, a breast tumor type identification system, a breast tumor type identification medium and breast tumor type identification equipment, wherein the breast tumor type identification method comprises the following steps: inputting biometric data of the sample; constructing a power excitation dynamic convergence differential neural network, setting each connection weight component between an input layer and an hidden layer and keeping unchanged, and randomly initializing each connection weight component between the hidden layer and an output layer; training a plurality of different power excitation dynamic convergence differential neural network models by adopting a plurality of different types of mapping functions; and placing the trained multiple power excitation dynamic convergence differential neural network into an integration framework, carrying out model integration treatment, carrying out type prediction on the biological characteristic data of the input sample, synthesizing multiple type prediction results, and obtaining a final breast tumor type recognition result by adopting a voting principle based on minority compliance. The invention solves the problem of diagnosing the breast tumor type, and achieves the effects of high diagnosis speed, high diagnosis accuracy and reliable diagnosis result.

Description

Mammary tumor type identification method, system, medium and equipment
Technical Field
The invention relates to the technical field of artificial intelligence identification, in particular to a breast tumor type identification method, a breast tumor type identification system, a breast tumor type identification medium and breast tumor type identification equipment.
Background
The statistical learning method plays a crucial role in intelligent diagnosis and evaluation of breast tumors, and in all statistical learning models, the neural network is suitable for being widely applied due to the excellent ability of learning fit samples, however, most traditional neural network training methods in the past adopt the thinking that gradient descent optimizes interlayer weights, and the training process possibly falls into the phenomena of local optimal solution, gradient disappearance or explosion along with the increase of the number of layers of the neural network; in addition, only a single neural network model is adopted in the diagnosis process of the breast tumor type, the reliability of the diagnosis result is difficult to ensure, and the defects are greatly restricted in the intelligent diagnosis process.
Disclosure of Invention
In order to overcome the defects and shortcomings in the prior art, the invention provides a breast tumor type identification method, which is used for diagnosing breast tumor types by training a power-excited dynamic convergence differential neural network model based on a neural dynamics method, does not involve complex operations such as gradient calculation and the like, greatly improves model training efficiency, and can effectively ensure reliability of breast tumor type diagnosis results by applying an integrated model.
A second object of the present invention is to provide a breast tumor type recognition system.
A third object of the present invention is to provide a storage medium.
It is a fourth object of the present invention to provide a computing device.
In order to achieve the above purpose, the present invention adopts the following technical scheme:
a method for identifying a breast tumor type, comprising the steps of:
inputting biometric data of the sample;
constructing a power excitation dynamic convergence differential neural network, setting the number of input neurons as the biological characteristic number of a single breast tumor sample, wherein the power excitation dynamic convergence differential neural network comprises an input layer, an hidden layer and an output layer, setting all connection weight components between the input layer and the hidden layer and keeping unchanged, and randomly initializing all connection weight components between the hidden layer and the output layer;
training a plurality of different power excitation dynamic convergence differential neural network models by adopting a plurality of different types of mapping functions;
and placing the trained multiple power excitation dynamic convergence differential neural network into an integration framework, carrying out model integration treatment, carrying out type prediction on the biological characteristic data of the input sample, synthesizing multiple type prediction results, and obtaining a final breast tumor type recognition result by adopting a voting principle based on minority compliance.
In order to achieve the second object, the present invention adopts the following technical scheme:
a breast tumor type identification system, comprising: the system comprises a biological characteristic input module, a neural network construction module, a neural network training module, an integrated model construction module and a comprehensive judgment module;
the biological characteristic input module is used for inputting biological characteristic data of a sample;
the neural network construction module is used for constructing a power excitation dynamic convergence differential neural network, and setting the number of input neurons as the number of biological characteristics of a single breast tumor sample;
the power excitation dynamic convergence differential neural network comprises an input layer, an implicit layer and an output layer, wherein each connection weight component between the input layer and the implicit layer is set and kept unchanged, and each connection weight component between the implicit layer and the output layer is randomly initialized;
the neural network training module is used for training a plurality of different power excitation dynamic convergence differential neural network models by adopting a plurality of different types of mapping functions;
the integrated model construction module is used for putting the trained multiple power excitation dynamic convergence difference neural networks into an integrated framework and carrying out model integration treatment to construct an integrated model, and the integrated model is used for carrying out type prediction on biological characteristic data of an input sample;
the comprehensive judgment module is used for integrating a plurality of type prediction results and obtaining a final breast tumor type recognition result by adopting a voting principle based on a minority compliance majority.
As an optimal technical scheme, the mapping function adopts three kinds of linear mapping function, sin mapping function and Sin mapping function.
As an preferable technical scheme, the specific training method of the neural network training module adopts the following expression:
E(k+1)-E(k)＝-αΦ(E(k)),α＞0
wherein E (k) represents the result of subtracting the expected value from the power-excited dynamic convergence differential neural network output after the kth learning sample, α represents the thermodynamic coefficient, and Φ represents the mapping function.
As a preferred solution, the excitation functions of the neurons of the hidden layer form a power function sequence, and the excitation functions of the neurons of the output layer adopt Softsign functions.
As a preferred technical solution, the specific implementation steps of the integrated model include:
let the input breast tumor sample be matrix X, the weight matrix between the hidden layer and the output layer after the kth learning sample be W (k), power excitation dynamic convergence differential neural network output Y (k):
Y(k)＝g(H(XI)W(k))
wherein:
setting the actual attribution type of the breast tumor sample into an expected value matrixThen->
And solving a training error epsilon (k) after a kth learning sample, outputting a power excitation dynamic convergence difference neural network judgment result output Y (k), obtaining a probability matrix P (k) through a Softmax layer, obtaining the attribution degree of each breast tumor sample for each type, and taking the type corresponding to the maximum degree as a prediction judgment result.
As an optimal technical scheme, the neural network training module adopts a cross entropy loss function formula to obtain a training error epsilon (k), and specifically calculates:
wherein L is sr And P sr Respectively representing the r elements of the s th row and the r th column in L and P (k);
setting a training error threshold value as epsilon ', and stopping iterative solution of a weight matrix W between an implicit layer and an output layer if epsilon (k) is less than epsilon';
solving a weight matrix between the hidden layer and the output layer to be W (k+1), wherein the weight matrix is expressed as:
wherein (H (XI)) + Moore-Penrose pseudo-inverse, function representing matrix H (XI)Representing the inverse of function g.
In order to achieve the third object, the present invention adopts the following technical scheme:
a storage medium storing a program which when executed by a processor implements the breast tumor type identification method described above.
In order to achieve the fourth object, the present invention adopts the following technical scheme:
a computing device comprising a processor and a memory for storing a program executable by the processor, the processor implementing the method of breast tumour type identification described above when executing the program stored by the memory.
Compared with the prior art, the invention has the following advantages and beneficial effects:
(1) The invention adopts the technical scheme of the integrated power excitation dynamic convergence differential neural network, solves the technical problem of breast tumor type diagnosis, and achieves the technical effects of high diagnosis speed, high diagnosis accuracy and reliable diagnosis result.
Drawings
FIG. 1 is a flow chart of a method for identifying breast tumor types according to the present invention;
FIG. 2 is a schematic diagram of a breast tumor type recognition system according to the present invention;
FIG. 3 is a schematic diagram of a power-excited dynamic convergence differential neural network of the present invention;
FIG. 4 is a schematic diagram of an integrated model of the present invention.
Detailed Description
The present invention will be described in further detail with reference to the drawings and examples, in order to make the objects, technical solutions and advantages of the present invention more apparent. It should be understood that the specific embodiments described herein are for purposes of illustration only and are not intended to limit the scope of the invention.
Examples
As shown in fig. 1, the present embodiment provides a breast tumor type identification method, which includes the following steps:
s1: after inputting biological characteristic data of a breast tumor sample, the power excitation dynamic convergence differential neural network construction steps specifically comprise:
s11: setting the number of input neurons as the biological characteristic number of a single breast tumor sample when constructing a power excitation dynamic convergence differential neural network, setting the number of output neurons as the total number of types of breast tumors, wherein the excitation functions of all neurons in an hidden layer form a power function sequence, the number needs to be manually adjusted and tested, and the excitation functions of all neurons in an output layer are Softsign functions;
s12: setting each connection weight component between the input layer and the hidden layer to be 1 and keeping unchanged, and randomly initializing each connection weight component between the hidden layer and the output layer, wherein training and updating are required;
s2: the power excitation dynamic convergence differential neural network training steps are specifically as follows:
s21: constructing an expression of a neural dynamics training method, and setting mapping function types to be three types of linear mapping functions, sin mapping functions and Sinh mapping functions;
the expression of the thermodynamic training method of this embodiment is: e (k+1) -E (k) = -alpha phi (E (k)), alpha > 0, wherein E (k) represents the intelligent doctor at the thThe result of subtracting the expected value from the output of the power-excited dynamic convergence differential neural network after the second learning of the sample is equivalent to the deviation between the prediction diagnosis result of the breast tumor sample and the actual attribution typeThe method comprises the steps of carrying out a first treatment on the surface of the Alpha > 0 represents the thermodynamic coefficient, which is equivalent to the learning and understanding speed of intelligent doctors on input breast tumor samples in the system establishment process; phi represents a mapping function, which is equivalent to a learning method;
s22: according to different types of mapping functions, applying a neural dynamics training method to a training power excitation dynamic convergence differential neural network, iteratively updating each connection weight component between an implicit layer and an output layer, and training out one excitation dynamic convergence differential neural network according to one type of mapping function, wherein three excitation dynamic convergence differential neural networks are trained out at the same time;
s3: the three trained power excitation dynamic convergence difference neural networks are introduced into an integration framework to form an integration model, and the method specifically comprises the following steps:
s31: placing the trained three power excitation dynamic convergence difference neural network into an integrated framework for predicting and judging unknown breast tumor sample types;
s32: carrying out model integration processing on three network models in an integration framework, and setting voting rules based on minority compliance and majority compliance for comprehensively evaluating unknown breast tumor sample types;
s4: the specific steps of comprehensive evaluation include:
s41: for the same unknown breast tumor sample, each power excitation dynamic convergence difference neural network in the integrated framework respectively predicts and judges the same, and three network models obtain three results;
s42: comprehensively judging the prediction judgment results by using voting rules based on a few obediences and majority to obtain the attribution of the final type of the sample;
s43: repeating S41 and S42 for prediction types for a plurality of unknown breast tumor samples one by one;
s5: the tumor type outputs the result.
The tumor biological characteristics input in this embodiment include: breast tumor mass thickness, cell size uniformity, cell shape uniformity, cell marginal adhesion, epithelial cell size, cell nucleus exposure, chromatin sparseness, cell nucleus normalization, nuclear division stage and the like, and outputting the output of the cell nucleus to be the type of breast tumor, namely benign or malignant;
as shown in fig. 2, this embodiment further provides a breast tumor type identification system, including: the system comprises a biological characteristic input module, a neural network construction module, a neural network training module, an integrated model construction module and a comprehensive judgment module;
in this embodiment, the biometric input module is configured to input biometric data of a sample;
in this embodiment, the neural network construction module is configured to construct a power-excited dynamic convergence differential neural network, and set the number of input neurons as the number of biological features of a single breast tumor sample;
in this embodiment, the power excitation dynamic convergence differential neural network includes an input layer, an hidden layer, and an output layer, each connection weight component between the input layer and the hidden layer is set and kept unchanged, and each connection weight component between the hidden layer and the output layer is randomly initialized;
in this embodiment, the neural network training module is configured to train out a plurality of different power-excited dynamic convergence differential neural network models by using a plurality of different types of mapping functions;
in this embodiment, the integrated model building module is configured to put the trained multiple power-excited dynamic convergence differential neural networks into an integrated framework and perform model integration processing, so as to build an integrated model, where the integrated model is used to perform type prediction on biological feature data of an input sample;
in this embodiment, the comprehensive judgment module is configured to integrate a plurality of type prediction results, and obtain a final breast tumor type recognition result by adopting a voting principle.
The expression of the thermodynamic training method of this embodiment is: e (k+1) -E (k) = -alpha phi (E (k)), alpha > 0, wherein E (k) represents the intelligent doctor at the thThe result of subtracting the expected value from the output of the power-excited dynamic convergence difference neural network after the second learning of the sample is equivalent to that of the breast tumor sampleDeviation between the predicted diagnostic result and the actual attribution type; alpha > 0 represents the thermodynamic coefficient, which is equivalent to the learning and understanding speed of intelligent doctors on input breast tumor samples in the system establishment process; Φ represents a mapping function, which corresponds to a learning method.
In the embodiment, three different power excitation dynamic convergence differential neural network models, namely a linear mapping function, a Sin mapping function and a Sinh mapping function, are trained by adopting three different types of mapping functions, so that three intelligent doctors are formed, and knowledge samples are formed through different learning modes, so that tumor type analysis and judgment are respectively carried out on input samples of unknown types.
As shown in fig. 3, for the intelligent doctor involved in the neural network construction module and the neural network training module, namely, the power excitation dynamic convergence differential neural network model, the network comprises three layers: the input layer, the hidden layer and the output layer, wherein each connecting weight component between the input layer and the hidden layer is 1 and is kept unchanged, namely, the weight matrix between the input layer and the hidden layer is a unit matrix I, so that each biological characteristic data of the input breast tumor sample is summarized; the excitation functions of the neurons of the hidden layers form a power function sequence, and the excitation function of the neurons of the first hidden layer from top to bottom is h 0 (z) =1, the second is h 1 (z) =z, the third is h 2 (z)＝z 2 … …, and so on, to obtain the nth value of h n (z)＝z n-1 Wherein z is an independent variable, thereby performing preliminary learning on the sample; the weight matrix W between the hidden layer and the output layer is randomly initialized to wait for training, which is equivalent to carrying out specific analysis on input sample data; the excitation function g of each neuron of the output layer is a Softsign function, and the analysis result is converted into a prediction judgment result in a nonlinear mapping mode. The expression of the Softsign function is as follows:
let the input breast tumor sample be matrix X, the weight matrix between the hidden layer and the output layer after the kth learning sample be W (k), power excitation dynamic convergence differential neural network output Y (k), namely the prediction judgment result of the intelligent doctor on the sample can be deduced by the following formula:
Y(k)＝g(H(XI)W(k)) (2)
wherein:
setting the actual attribution type of the breast tumor sample into an expected value matrixThen->Firstly, a training error epsilon (k) after the kth learning sample is obtained, wherein the training error is equivalent to the total deviation between the prediction judging result of a plurality of breast tumor samples and the actual attribution type of the samples.
The power excitation dynamic convergence difference neural network judgment result output Y (k) can obtain a probability matrix P (k) through a Softmax layer, the attribution degree of each breast tumor sample for each type is obtained, and the type corresponding to the maximum degree is taken as a prediction judgment result.
Let the scales of matrices Y (k) and P (k) be l×q, for the first of Y (k)Prediction judgment result y corresponding to each breast tumor sample s ＝[y s1 y s2 … y sq ]Line vector P of the s-th line in P (k) s I.e., the degree of attribution of the corresponding s-th sample to each type, is obtained by the following equation:
meanwhile, as a sample actual category label, an expected value matrixThe scale is set to l×q, the inner element consisting of-1 and 1. Therefore, the coding layer is required to change the element with the value of-1 into the value of 0, and the other elements are unchanged, so that the coding matrix L is obtained, and the L is a coding mode of the actual attribution type of the breast tumor sample. The training error ε (k) can thus be derived from the cross entropy loss function formula:
wherein L is sr And P sr Respectively represent the s-th row and the s-th row in L and P (k)Column elements.
And setting the training error threshold as epsilon' so as to take care of the intelligent doctor on learning the breast tumor sample degree. And if epsilon (k) < epsilon', stopping the iterative solution of the weight matrix W between the hidden layer and the output layer, and ending the learning process of the intelligent doctor on the sample. Otherwise, E (k+1) =E (k) -alpha phi (E (k)) can be deduced through an expression of the neuro-dynamics method, namely, a result that the power excitation dynamic convergence difference neural network output minus an expected value is due after the intelligent doctor (k+1) th learning sample is obtained. On this basis, a weight matrix between the hidden layer and the output layer needs to be found as W (k+1). Analog type (2)It is possible that E (k+1) and Y (k+1) satisfy the following equation simultaneously:
Y(k+1)＝g(H(XI)W(k+1)) (7)
from equation (6) and equation (7), a solution expression for W (k+1) can be derived:
it is further possible to derive the relationship of W (k+1) to W (k) after the (k+1) th learning sample. The iterative solution expression of W (k+1) is:
in the formulae (8) and (9), (H (XI)) + Moore-Penrose pseudo-inverse representing matrix H (XI); function ofTheoretically, the inverse of the function g should be represented. However, g is a Softsign function (as shown in equation (1)) and does not have an inverse function. Therefore, the equation (1) is segmented according to the value interval of the independent variable z, the inverse function of each segment is obtained, and then the inverse functions of the value intervals are spliced together to be approximately replaced.
Thus (2)The expression of (2) is:
as shown in fig. 4, for the integrated model related to the integrated model construction module and the comprehensive judgment module in fig. 2, the trained three power excitation dynamic convergence difference neural network is put into an integrated framework and is subjected to model integration processing, and voting rules based on a minority compliance majority are set for comprehensive judgment of unknown breast tumor sample types. The method is that a plurality of intelligent doctors for learning breast tumor samples by different methods gather together to analyze and judge the breast tumor samples of unknown types, and the obtained prediction results are used for comprehensive analysis. In the comprehensive analysis process, first, the three intelligent doctors respectively predict the types of the breast tumor samples. After all intelligent doctors complete the prediction, the integration framework adopts a voting principle based on a few obediences and majority to the prediction results, and comprehensively judges the final type attribution of the breast tumor sample. For example, for a breast tumor sample of an unknown type in fig. 4, the first intelligent doctor predicts that the result is malignant and the other two intelligent doctors predict that the result is benign, the integrated model ultimately determines that the test sample is attributed to benign.
The present embodiment also provides a storage medium, which may be a storage medium such as a ROM, a RAM, a magnetic disk, or an optical disk, and the storage medium stores one or more programs that when executed by a processor, implement the breast tumor type identification method described above.
The embodiment also provides a computing device, which may be a desktop computer, a notebook computer, a smart phone, a PDA handheld terminal, a tablet computer or other terminal devices with display function, and the computing device includes a processor and a memory, where the memory stores one or more programs, and when the processor executes the programs stored in the memory, the method for identifying breast tumor types is implemented.
The above examples are preferred embodiments of the present invention, but the embodiments of the present invention are not limited to the above examples, and any other changes, modifications, substitutions, combinations, and simplifications that do not depart from the spirit and principle of the present invention should be made in the equivalent manner, and the embodiments are included in the protection scope of the present invention.

Claims (7)

1. A method for identifying a breast tumor type, comprising the steps of:
inputting biometric data of the sample;
constructing a power excitation dynamic convergence differential neural network, setting the number of input neurons as the biological characteristic number of a single breast tumor sample, wherein the power excitation dynamic convergence differential neural network comprises an input layer, an hidden layer and an output layer, setting all connection weight components between the input layer and the hidden layer and keeping unchanged, and randomly initializing all connection weight components between the hidden layer and the output layer;
training a plurality of different power excitation dynamic convergence differential neural network models by adopting a plurality of different types of mapping functions;
the specific training method adopts the expression:
E(k+1)-E(k)＝-αΦ(E(k)),α＞0
wherein E (k) represents the result of subtracting the expected value from the output of the power-excited dynamic convergence differential neural network after the kth learning sample, alpha represents the thermodynamic coefficient, and phi represents the mapping function;
placing the trained multiple power excitation dynamic convergence differential neural network into an integration framework, carrying out model integration treatment, carrying out type prediction on biological characteristic data of an input sample, synthesizing multiple type prediction results, and obtaining a final breast tumor type recognition result by adopting a voting principle based on minority compliance;
the specific implementation steps of the model integration processing include:
let the input breast tumor sample be matrix X, the weight matrix between the hidden layer and the output layer after the kth learning sample be W (k), power excitation dynamic convergence differential neural network output Y (k):
Y(k)＝g(H(XI)W(k))
wherein:
setting the actual attribution type of the breast tumor sample into an expected value matrixThen->
The power excitation dynamic convergence difference neural network judgment result output Y (k) is subjected to a Softmax layer to obtain a probability matrix P (k), the attribution degree of each breast tumor sample to each type is obtained, and the type corresponding to the maximum degree is taken as a prediction judgment result.
2. A breast tumor type recognition system, comprising: the system comprises a biological characteristic input module, a neural network construction module, a neural network training module, an integrated model construction module and a comprehensive judgment module;
the biological characteristic input module is used for inputting biological characteristic data of a sample;
the neural network construction module is used for constructing a power excitation dynamic convergence differential neural network, and setting the number of input neurons as the number of biological characteristics of a single breast tumor sample;
the power excitation dynamic convergence differential neural network comprises an input layer, an implicit layer and an output layer, wherein each connection weight component between the input layer and the implicit layer is set and kept unchanged, and each connection weight component between the implicit layer and the output layer is randomly initialized;
the neural network training module is used for training a plurality of different power excitation dynamic convergence differential neural network models by adopting a plurality of different types of mapping functions;
the specific training method of the neural network training module adopts the following expression:
E(k+1)-E(k)＝-αΦ(E(k)),α＞0
wherein E (k) represents the result of subtracting the expected value from the output of the power-excited dynamic convergence differential neural network after the kth learning sample, alpha represents the thermodynamic coefficient, and phi represents the mapping function;
the integrated model construction module is used for putting the trained multiple power excitation dynamic convergence difference neural networks into an integrated framework and carrying out model integration treatment to construct an integrated model, and the integrated model is used for carrying out type prediction on biological characteristic data of an input sample;
the integrated model specifically executing steps comprise:
let the input breast tumor sample be matrix X, the weight matrix between the hidden layer and the output layer after the kth learning sample be W (k), power excitation dynamic convergence differential neural network output Y (k):
Y(k)＝g(H(XI)W(k))
wherein:
H(z)＝{h 0 (z)；h 1 (z)；h 2 (z)…h n-1 (z)}
＝{1；z；z 2 …z n-1 }
setting the actual attribution type of the breast tumor sample into an expected value matrixThen->
The power excitation dynamic convergence difference neural network judgment result output Y (k) obtains a probability matrix P (k) through a Softmax layer, obtains the attribution degree of each breast tumor sample for each type, and takes the type corresponding to the maximum degree as a prediction judgment result;
the comprehensive judgment module is used for integrating a plurality of type prediction results and obtaining a final breast tumor type recognition result by adopting a voting principle based on a minority compliance majority.
3. The breast tumor type recognition system of claim 2, wherein the mapping function employs three of a linear mapping function, a Sin mapping function, and a Sinh mapping function.
4. The breast tumor type recognition system of claim 2, wherein the excitation functions of the neurons of the hidden layer form a power function sequence and the excitation functions of the neurons of the output layer employ Softsign functions.
5. The breast tumor type recognition system of claim 2, wherein the neural network training module derives a training error epsilon (k) using a cross entropy loss function formula, specifically calculated as:
wherein L is sr And P sr Respectively representing the r elements of the s th row and the r th column in L and P (k);
setting a training error threshold value as epsilon ', and stopping iterative solution of a weight matrix W between an implicit layer and an output layer if epsilon (k) is less than epsilon';
solving a weight matrix between the hidden layer and the output layer to be W (k+1), wherein the weight matrix is expressed as:
wherein (H (XI)) + Moore-Penrose pseudo-inverse, function representing matrix H (XI)Representing the inverse of function g.
6. A storage medium storing a program which, when executed by a processor, implements the breast tumor type recognition method according to claim 1.
7. A computing device comprising a processor and a memory for storing a program executable by the processor, wherein the processor, when executing the program stored in the memory, implements the breast tumor type identification method of claim 1.
CN202010823773.5A 2020-08-17 2020-08-17 Mammary tumor type identification method, system, medium and equipment Active CN112101418B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202010823773.5A CN112101418B (en) 2020-08-17 2020-08-17 Mammary tumor type identification method, system, medium and equipment

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202010823773.5A CN112101418B (en) 2020-08-17 2020-08-17 Mammary tumor type identification method, system, medium and equipment

Publications (2)

Publication Number Publication Date
CN112101418A CN112101418A (en) 2020-12-18
CN112101418B true CN112101418B (en) 2023-08-22

Family

ID=73753881

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202010823773.5A Active CN112101418B (en) 2020-08-17 2020-08-17 Mammary tumor type identification method, system, medium and equipment

Country Status (1)

Country Link
CN (1) CN112101418B (en)

Families Citing this family (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN113012133A (en) * 2021-03-22 2021-06-22 华南理工大学 New coronary pneumonia diagnosis system, medium and equipment
CN113673566A (en) * 2021-07-19 2021-11-19 华南理工大学 New coronary pneumonia diagnosis system, medium and equipment based on CT image
CN113476065B (en) * 2021-09-06 2022-04-22 华南理工大学 Multiclass pneumonia diagnostic system

Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN104392143A (en) * 2014-12-09 2015-03-04 北京四方继保自动化股份有限公司 Method for predicting fault trends of steam turbines by aid of adaptive quantum neural networks
CN105678422A (en) * 2016-01-11 2016-06-15 广东工业大学 Empirical mode neural network-based chaotic time series prediction method
CN108921049A (en) * 2018-06-14 2018-11-30 华东交通大学 Tumour cell pattern recognition device and equipment based on quantum gate transmission line neural network

Patent Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN104392143A (en) * 2014-12-09 2015-03-04 北京四方继保自动化股份有限公司 Method for predicting fault trends of steam turbines by aid of adaptive quantum neural networks
CN105678422A (en) * 2016-01-11 2016-06-15 广东工业大学 Empirical mode neural network-based chaotic time series prediction method
CN108921049A (en) * 2018-06-14 2018-11-30 华东交通大学 Tumour cell pattern recognition device and equipment based on quantum gate transmission line neural network

Also Published As

Publication number Publication date
CN112101418A (en) 2020-12-18

Similar Documents

Publication Publication Date Title
CN112101418B (en) Mammary tumor type identification method, system, medium and equipment
CN105335756B (en) A kind of image classification method and image classification system based on Robust Learning model
CN110909926A (en) TCN-LSTM-based solar photovoltaic power generation prediction method
Yun Prediction model of algal blooms using logistic regression and confusion matrix
CN111046961B (en) Fault classification method based on bidirectional long-time and short-time memory unit and capsule network
CN112232407A (en) Neural network model training method and device for pathological image sample
CN115201686B (en) Lithium ion battery health state assessment method under incomplete charge and discharge data
CN112466468B (en) Immune level prediction system, device and storage medium based on sparse coding
Aghasafari et al. A deep learning algorithm to translate and classify cardiac electrophysiology
CN112164462B (en) Breast cancer risk assessment method, system, medium and equipment
CN113988357A (en) High-rise building wind-induced response prediction method and device based on deep learning
CN116580848A (en) Multi-head attention mechanism-based method for analyzing multiple groups of chemical data of cancers
CN113345564B (en) Early prediction method and device for patient hospitalization duration based on graph neural network
CN114578234A (en) Lithium ion battery degradation and capacity prediction model considering causality characteristics
CN109063750B (en) SAR target classification method based on CNN and SVM decision fusion
CN113688770A (en) Long-term wind pressure missing data completion method and device for high-rise building
CN113151842B (en) Method and device for determining conversion efficiency of wind-solar complementary water electrolysis hydrogen production
CN111461308B (en) Memristor neural network and weight training method
CN112733724A (en) Relativity relationship verification method and device based on discrimination sample meta-digger
CN115631847B (en) Early lung cancer diagnosis system, storage medium and equipment based on multiple groups of chemical characteristics
US20230223099A1 (en) Predicting method of cell deconvolution based on a convolutional neural network
CN113476065B (en) Multiclass pneumonia diagnostic system
Azam et al. Classification of COVID-19 symptoms using multilayer perceptron
CN110609971B (en) Calibration multiple regression network construction method
CN115083511A (en) Peripheral gene regulation and control feature extraction method based on graph representation learning and attention

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