[go: up one dir, main page]
More Web Proxy on the site http://driver.im/

CN112086174A - Three-dimensional knowledge diagnosis model construction method and system - Google Patents

Three-dimensional knowledge diagnosis model construction method and system Download PDF

Info

Publication number
CN112086174A
CN112086174A CN202011015038.8A CN202011015038A CN112086174A CN 112086174 A CN112086174 A CN 112086174A CN 202011015038 A CN202011015038 A CN 202011015038A CN 112086174 A CN112086174 A CN 112086174A
Authority
CN
China
Prior art keywords
dimensional knowledge
medical image
image data
training
model
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
Application number
CN202011015038.8A
Other languages
Chinese (zh)
Other versions
CN112086174B (en
Inventor
秦文健
田引黎
刘磊
张志诚
陈实富
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.)
Shenzhen Institute of Advanced Technology of CAS
Original Assignee
Shenzhen Institute of Advanced Technology of CAS
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 Shenzhen Institute of Advanced Technology of CAS filed Critical Shenzhen Institute of Advanced Technology of CAS
Priority to CN202011015038.8A priority Critical patent/CN112086174B/en
Publication of CN112086174A publication Critical patent/CN112086174A/en
Application granted granted Critical
Publication of CN112086174B publication Critical patent/CN112086174B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Classifications

    • 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
    • G16H30/00—ICT specially adapted for the handling or processing of medical images
    • G16H30/20—ICT specially adapted for the handling or processing of medical images for handling medical images, e.g. DICOM, HL7 or PACS
    • 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
    • G06N3/08—Learning methods
    • G06N3/088—Non-supervised learning, e.g. competitive learning
    • 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
    • G16H50/50—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics for simulation or modelling of medical disorders
    • 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
    • Y02T—CLIMATE CHANGE MITIGATION TECHNOLOGIES RELATED TO TRANSPORTATION
    • Y02T10/00—Road transport of goods or passengers
    • Y02T10/10—Internal combustion engine [ICE] based vehicles
    • Y02T10/40—Engine management systems

Landscapes

  • Engineering & Computer Science (AREA)
  • Health & Medical Sciences (AREA)
  • General Health & Medical Sciences (AREA)
  • Theoretical Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • Public Health (AREA)
  • Medical Informatics (AREA)
  • Biomedical Technology (AREA)
  • Data Mining & Analysis (AREA)
  • Computational Linguistics (AREA)
  • Epidemiology (AREA)
  • Molecular Biology (AREA)
  • Computing Systems (AREA)
  • General Engineering & Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • Mathematical Physics (AREA)
  • Software Systems (AREA)
  • Biophysics (AREA)
  • Primary Health Care (AREA)
  • Artificial Intelligence (AREA)
  • Evolutionary Computation (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Pathology (AREA)
  • Databases & Information Systems (AREA)
  • Nuclear Medicine, Radiotherapy & Molecular Imaging (AREA)
  • Radiology & Medical Imaging (AREA)
  • Image Analysis (AREA)
  • Measuring And Recording Apparatus For Diagnosis (AREA)

Abstract

The invention discloses a method and a system for constructing a three-dimensional knowledge diagnosis model. The method comprises the following steps: building a training set based on the known medical image data, the training set comprising
Figure DDA0002698772200000011
And
Figure DDA0002698772200000012
two moieties of which XmRepresents the m-th medical image data, YmLabel, X, representing mth example medical image dataiRepresenting the ith sample of unlabeled medical image data, wherein M and N are the number of corresponding samples respectively; at a set lossA neural network-based three-dimensional knowledge diagnosis model constructed by training a loss function as a target, wherein Xm，YmSupervised trained inputs and labels, X, as the principal branches of the three-dimensional knowledge diagnostic modeliAs input for unsupervised training of the main branches of the three-dimensional knowledge diagnostic model,
Figure DDA0002698772200000013
as an input of the unsupervised training of the auxiliary branch of the three-dimensional knowledge diagnostic model,
Figure DDA0002698772200000014
is to XiAnd carrying out various kinds of disturbed data. The invention can realize high-efficiency and intelligent diagnosis on the medical image and is used for clinical indication.

Description

Three-dimensional knowledge diagnosis model construction method and system
Technical Field
The invention relates to the technical field of medical image processing, in particular to a method and a system for constructing a three-dimensional knowledge diagnosis model.
Background
In recent years, with the rapid development of deep learning, students at home and abroad simulate the diagnosis of diseases by doctors through computers, and extract and diagnose and analyze intelligent diagnosis indexes of diseases by utilizing the deep learning. The related scientific achievements have reached expert-level disease diagnosis, and even have better performance than experienced clinicians in some aspects. While deep learning based disease diagnosis techniques have achieved great success, these deep learning based diagnosis methods rely on medical image data of a large number of authentic labeled samples, and it is very difficult to acquire a large number of clinical medical image data with authentic labeled samples. The most common method for solving the deficiency is the regularization technology, and the L1 norm is the currently mainstream regularization technology, because the global optimal solution can be effectively calculated. However, the L1 norm has unbiased properties, which prevent consistency in variable selection. To ensure statistical performance of variable selection, strongly non-representable conditions are required. In addition, due to imaging equipment, individual differences and the like, the acquired clinical medical image data may have quality problems such as data loss and the like. The data is generally removed by means of data cleaning and the like to ensure the normalization of the data, which undoubtedly wastes a large amount of clinical data and is not in accordance with the real clinical diagnosis condition.
Through analysis, the defects of the prior art are mainly as follows:
1) the data volume of the real label in clinical medicine is very limited, and when the dimension P is far larger than the sample volume, an overfitting phenomenon occurs when the model parameters are solved. To prevent overfitting, a certain regularization approach is usually adopted to limit the parameter space to a certain range. The L1 norm regularization is a relatively common regularization method. However, the L1 norm regularization is unbiased, which prevents consistency in variable selection. To ensure statistical performance of variable selection, strongly non-representable conditions are required.
2) Due to the obtained data quality problem and the limited real label data, the accuracy of the probability value generated by predicting the sample data by the network model to generate the real standard is not high.
3) The network architecture commonly used at present is based on a two-dimensional convolutional neural network, and the two-dimensional convolutional neural network does not fully utilize spatial information.
Disclosure of Invention
The invention aims to overcome the defects of the prior art and provides a method and a system for constructing a three-dimensional knowledge diagnosis model, which is a new technical scheme for realizing efficient and intelligent diagnosis.
According to a first aspect of the invention, a method for constructing a three-dimensional knowledge diagnosis model is provided. The method comprises the following steps:
building a training set based on the known medical image data, the training set comprising
Figure BDA0002698772180000021
And
Figure BDA0002698772180000022
two moieties of which XmRepresents the m-th medical image data, YmLabel, X, representing mth example medical image dataiRepresenting the ith sample of unlabeled medical image data, wherein M and N are the number of corresponding samples respectively;
at a set lossA three-dimensional knowledge diagnosis model based on neural network and constructed by training a function for a target, wherein Xm，YmSupervised trained inputs and labels, X, as the principal branches of the three-dimensional knowledge diagnostic modeliAs input for unsupervised training of the main branches of the three-dimensional knowledge diagnostic model,
Figure BDA0002698772180000025
as an input of the unsupervised training of the auxiliary branch of the three-dimensional knowledge diagnostic model,
Figure BDA0002698772180000026
is to XiAnd carrying out various kinds of disturbed data.
According to a second aspect of the present invention, there is provided a three-dimensional knowledge diagnosis model construction system. The system comprises:
a data acquisition unit: for creating a training set based on known medical image data, the training set comprising
Figure BDA0002698772180000023
And
Figure BDA0002698772180000024
two moieties of which XmRepresents the m-th medical image data, YmLabel, X, representing mth example medical image dataiRepresenting the ith sample of unlabeled medical image data, wherein M and N are the number of corresponding samples respectively;
a model training unit: it is used for training a built three-dimensional knowledge diagnosis model based on a neural network by taking a set loss function as a target, wherein Xm，YmSupervised trained inputs and labels, X, as the principal branches of the three-dimensional knowledge diagnostic modeliAs input for unsupervised training of the main branches of the three-dimensional knowledge diagnostic model,
Figure BDA0002698772180000027
as an input of the unsupervised training of the auxiliary branch of the three-dimensional knowledge diagnostic model,
Figure BDA0002698772180000028
is to XiAnd carrying out various kinds of disturbed data.
According to a third aspect of the invention, a medical image recognition system is provided. The system comprises:
a data acquisition unit: the system is used for acquiring medical image data to be detected;
an image recognition unit: and the medical image data is input into the main branch of the three-dimensional knowledge diagnosis model constructed by the invention, and the classification result is obtained.
Compared with the prior art, the method has the advantages that the sample data is input into the three-dimensional knowledge diagnosis model based on the convolutional neural network, and the probability value of the real standard is generated through prediction by the network model. Then, the probability value of the initial prediction is optimized and calculated, the problem of high-dimensional learning caused by limited real label data is solved, and the identification efficiency and the identification precision of the medical image are improved.
Other features of the present invention and advantages thereof will become apparent from the following detailed description of exemplary embodiments thereof, which proceeds with reference to the accompanying drawings.
Drawings
The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments of the invention and together with the description, serve to explain the principles of the invention.
FIG. 1 is a flow diagram of a method of building a three-dimensional knowledge diagnostic model according to one embodiment of the invention;
FIG. 2 is a specific example of a three-dimensional knowledge diagnostic model building process according to one embodiment of the present invention;
FIG. 3 is a schematic diagram of a convolutional neural network-based three-dimensional knowledge diagnostic model according to one embodiment of the present invention;
FIG. 4 is a diagram illustrating an optimization process based on maximum expectation theory, according to an embodiment of the present invention.
Detailed Description
Various exemplary embodiments of the present invention will now be described in detail with reference to the accompanying drawings. It should be noted that: the relative arrangement of the components and steps, the numerical expressions and numerical values set forth in these embodiments do not limit the scope of the present invention unless specifically stated otherwise.
The following description of at least one exemplary embodiment is merely illustrative in nature and is in no way intended to limit the invention, its application, or uses.
Techniques, methods, and apparatus known to those of ordinary skill in the relevant art may not be discussed in detail but are intended to be part of the specification where appropriate.
In all examples shown and discussed herein, any particular value should be construed as merely illustrative, and not limiting. Thus, other examples of the exemplary embodiments may have different values.
It should be noted that: like reference numbers and letters refer to like items in the following figures, and thus, once an item is defined in one figure, further discussion thereof is not required in subsequent figures.
Referring to fig. 1 and 2, the provided method for constructing a three-dimensional knowledge diagnosis model includes the following steps:
step S110, medical image data is acquired.
In this step, known medical image data is acquired for subsequent training. For example, medical image data containing a tumor and normal expert annotations and non-expert-annotated medical image data may be gathered, and the image data may be labeled normal, cancer, or other types.
Step S120, a training set and a test set are established based on the medical image data.
A data set is created from the collected medical image data, for example, the data set is divided into a training set and a test set. The training set comprises
Figure BDA0002698772180000041
And
Figure BDA0002698772180000042
two parts, wherein M + N ═ K, XmObtaining the m-th medical image data; y ismLabeling the m-th medical image data label for the expert; xiThe obtained ith example is labeled-free medical image data. The test set is
Figure BDA0002698772180000043
And S130, constructing a three-dimensional knowledge diagnosis model, and training by using a training set to obtain a classification probability value of input data.
For example, a convolutional neural network-based three-dimensional knowledge diagnosis model is constructed, as shown in fig. 3, and the model structurally comprises a main branch and an auxiliary branch from the whole, and functionally comprises a feature extraction function (i.e. a three-dimensional convolutional neural network feature extractor) and a classification function (i.e. a full connection layer & softmax), wherein the main branch and the auxiliary branch share the three-dimensional convolutional neural network feature extractor.
Medical image data X in training setm，Ym，XiAnd are and
Figure BDA0002698772180000044
and j is 1, … p, inputting the model to perform network training, and acquiring the probability value through softmax. Wherein Xm，YmSupervised trained inputs and labels as principal branches of a model, XiAs input for unsupervised training of the main branches of the model,
Figure BDA0002698772180000045
is to XiAnd carrying out various kinds of disturbed data as input of unsupervised training of the auxiliary branch of the model. In this context, X ismObtaining probability value p through networkm，XiObtaining probability value p through networki，
Figure BDA0002698772180000051
Recording of probability values obtained over a network
Figure BDA0002698772180000052
In a preferred embodiment, step S130 specifically includes:
step S131: a three-dimensional knowledge diagnosis model or three-dimensional convolution network is designed based on the Incepton v 3.
For example, a three-dimensional knowledge diagnosis model designed based on the existing inclusion v3 includes 5 3D convolutional layers, 9 block structures, 2 3D pooling layers, one fully connected layer and one softmax. The 3D convolutional layer and the block structure are used for feature extraction, and the pooling layer is used for data compression and can adopt maximum pooling or average pooling. The output channel is set to 2, e.g. normal and cancer, corresponding to the constructed training set.
Step S132: medical image data Xm，Xi，
Figure BDA0002698772180000053
And inputting a model, and acquiring the characteristics of the medical image through the 3D convolution layer and the block structure.
Step S133: mixing Xm，XiAnd inputting the full connection layer of the main branch through the characteristics obtained in the step S132 to perform characteristic attribute combination. Will be provided with
Figure BDA0002698772180000054
And performing feature attribute combination through the full connection layer of the feature input auxiliary branch obtained in the step S132.
Step S134: inputting the characteristics of the full connection layer to construct a three-dimensional knowledge diagnosis classifier, and obtaining the predicted probability value p through softmaxm，piAnd
Figure BDA0002698772180000055
Figure BDA0002698772180000056
is the output result of the jth auxiliary branch of the ith image, which is equivalent to the diagnostic result of the jth doctor of the ith image.
In this step S130, a three-dimensional convolutional neural network structure having a branch structure is constructed, and a plurality of doctors can be simulated. In addition, the data with the labels, the data without the labels and the disturbed data are learned, and the semi-supervised learning mode can reduce data labeling and improve the prediction accuracy of learning.
And step S140, in the training process, the obtained probability value is optimized by using the maximum expectation theory so as to obtain a more accurate predicted probability value.
For example, the initial probability value p obtained in step S130m，pi，
Figure BDA0002698772180000057
Inputting a convergence layer (representing a layer optimized by a maximum expectation theory), and acquiring a probability value mu by the convergence layer through a maximum voting methodiUsing the maximum expectation theory on μiOptimizing to obtain more accurate predicted value mu'i. The content of step S140 will be described in more detail below.
And S150, in the training process, solving parameters of the three-dimensional knowledge diagnosis model by using the set loss function and the set propagation mechanism to obtain the optimized three-dimensional knowledge diagnosis model.
Preferably, the aggregation layer loss function is combined with a supervision loss function and a folding concave regularization loss function back propagation, and the model parameters are solved through chain rule derivation back propagation to obtain the optimized three-dimensional knowledge diagnosis model.
For example, YmAs a label for model supervised training, and the predicted value p obtained in step S130mA supervised loss function calculation is performed, expressed as:
Figure BDA0002698772180000061
wherein J (w, b) is a supervised classification loss function, wm,bmAre weights and offsets. L (. mu.)i,pi) To aggregate the loss functions of the layers, to overcome the high-dimensional learning problem, fold-concave regularization is preferably added to the loss functions, and the total loss function is expressed as：
Figure BDA0002698772180000062
Pλ(. to) is a fold-valley regularization penalty function,
Figure BDA0002698772180000063
a and lambda are fine tuning parameters.
In the embodiment, by adding a fold-valley regularization penalty function as a strong irreducible condition to the loss function, the statistical performance of variable selection is ensured, training errors and model complexity can be balanced, and overfitting is avoided. In addition, the back propagation algorithm can conveniently calculate the derivative of the loss function to each parameter, and the obtained derivative is used for the model training optimization by a gradient descent method.
The trained three-dimensional knowledge diagnosis model or the optimized three-dimensional knowledge diagnosis model is obtained after the model convergence through the training of the steps S130-S150.
Step S160, evaluating the trained three-dimensional knowledge diagnostic model using the test set.
To evaluate the training effect, the test set may be further processed
Figure BDA0002698772180000065
Inputting the trained three-dimensional knowledge diagnosis model to evaluate the prediction accuracy of the model.
In a preferred embodiment, as shown in fig. 4, the step S140 includes:
step S141: determining an initial decision probability value mu by maximum votingiThrough μiCalculation of initial sensitivity αjAnd specific betaj(where j ∈ P is the jth doctor).
Figure BDA0002698772180000064
Wherein
Figure BDA0002698772180000071
Step S142: calculation of conditional expectation (E-step):
Figure BDA0002698772180000072
wherein D is the set of observations and the set of observations,
Figure BDA0002698772180000073
Figure BDA0002698772180000074
g is an implicit variable (true tag value),
Figure BDA0002698772180000075
is the currently estimated parameter.
Step S143: inferring μ using Bayesian theoryiAnd alphajAnd betajThe relationship, expressed as:
Figure BDA0002698772180000076
step S144: sensitivity and specificity (M-step) were calculated.
For example, μ obtained by the given observation set D and step S143iThe above desired value is differentiated and the derivative is made 0, and alpha is calculatedjAnd betajExpressed as:
Figure BDA0002698772180000077
step S145: binding sensitivity alphajAnd specific betajLearning a softmax classifier by using parameters to obtain a predicted probability value mu'i。
In one embodiment, the loss function of an aggregation layer is defined as:
Figure BDA0002698772180000078
further, after obtaining the trained three-dimensional knowledge diagnosis model, the medical image data to be detected is used as input, for example, input to the main branch, and then corresponding output, for example, belonging to normal or cancer, can be obtained, and the output result can be used for clinical indication.
Correspondingly, the invention also provides a three-dimensional knowledge diagnosis model construction system and a medical image recognition system, which are used for realizing one or more aspects of the method.
For example, a medical image recognition system includes: a data acquisition unit for acquiring medical image data to be detected; an image recognition unit for inputting the medical image data to a main branch of the optimized three-dimensional knowledge diagnostic model, thereby obtaining a classification result.
For example, the three-dimensional knowledge diagnosis model construction system includes: a data acquisition unit for establishing a training set based on known medical image data, the training set comprising
Figure BDA0002698772180000079
And
Figure BDA00026987721800000710
two moieties of which XmRepresents the m-th medical image data, YmLabel, X, representing mth example medical image dataiRepresenting the ith sample of unlabeled medical image data, wherein M and N are the number of corresponding samples respectively; a model training unit for training the constructed neural network-based three-dimensional knowledge diagnosis model with the set loss function as a target, wherein Xm，YmSupervised trained inputs and labels, X, as the principal branches of the three-dimensional knowledge diagnostic modeliAs input for unsupervised training of the main branches of the three-dimensional knowledge diagnostic model,
Figure BDA0002698772180000081
as an aid to the three-dimensional knowledge diagnostic modelThe input of unsupervised training of the aid branches,
Figure BDA0002698772180000082
is to XiAnd carrying out various kinds of disturbed data.
In conclusion, the three-dimensional knowledge diagnosis model provided by the invention effectively overcomes the problem of high-dimensional learning of the diagnosis model caused by limited real label data, and overcomes the defect of low accuracy of the predicted value of the network model caused by quality problems such as data loss and the like by adopting a maximum expectation theory and a semi-supervised algorithm. And the knowledge diagnosis model is optimized based on the folding concave regularization, and compared with L1 norm regularization, the folding concave regularization can obtain better statistical performance by only needing fewer theoretical rules. Compared with a common deep learning framework, the method provided by the invention adopts the maximum expectation theory to simulate doctor consultation to optimize judgment accuracy, so that the defect that the initial predicted value is not accurate enough due to quality problems such as data loss is solved. In addition, in order to overcome the problem of insufficient two-dimensional spatial information, all calculation processes of the constructed model adopt a three-dimensional convolutional neural network structure, and efficient and intelligent diagnosis is realized.
It should be noted that the present invention is applicable to analysis and recognition of various medical images, and the recognition result is not limited to tumors, cancers, and the like. Those skilled in the art may make appropriate modifications or alterations to the above-described embodiments, for example, to design more or fewer 3D convolutional layers, block structures, or to employ other sorting methods than softmax, without departing from the spirit and scope of the invention.
The present invention may be a system, method and/or computer program product. The computer program product may include a computer-readable storage medium having computer-readable program instructions embodied therewith for causing a processor to implement various aspects of the present invention.
The computer readable storage medium may be a tangible device that can hold and store the instructions for use by the instruction execution device. The computer readable storage medium may be, for example, but not limited to, an electronic memory device, a magnetic memory device, an optical memory device, an electromagnetic memory device, a semiconductor memory device, or any suitable combination of the foregoing. More specific examples (a non-exhaustive list) of the computer readable storage medium would include the following: a portable computer diskette, a hard disk, a Random Access Memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), a Static Random Access Memory (SRAM), a portable compact disc read-only memory (CD-ROM), a Digital Versatile Disc (DVD), a memory stick, a floppy disk, a mechanical coding device, such as punch cards or in-groove projection structures having instructions stored thereon, and any suitable combination of the foregoing. Computer-readable storage media as used herein is not to be construed as transitory signals per se, such as radio waves or other freely propagating electromagnetic waves, electromagnetic waves propagating through a waveguide or other transmission medium (e.g., optical pulses through a fiber optic cable), or electrical signals transmitted through electrical wires.
The computer-readable program instructions described herein may be downloaded from a computer-readable storage medium to a respective computing/processing device, or to an external computer or external storage device via a network, such as the internet, a local area network, a wide area network, and/or a wireless network. The network may include copper transmission cables, fiber optic transmission, wireless transmission, routers, firewalls, switches, gateway computers and/or edge servers. The network adapter card or network interface in each computing/processing device receives computer-readable program instructions from the network and forwards the computer-readable program instructions for storage in a computer-readable storage medium in the respective computing/processing device.
The computer program instructions for carrying out operations of the present invention may be assembler instructions, Instruction Set Architecture (ISA) instructions, machine-related instructions, microcode, firmware instructions, state setting data, or source or object code written in any combination of one or more programming languages, including an object oriented programming language such as Smalltalk, C + + or the like and conventional procedural programming languages, such as the "C" programming language or similar programming languages. The computer-readable program instructions may execute entirely on the user's computer, partly on the user's computer, as a stand-alone software package, partly on the user's computer and partly on a remote computer or entirely on the remote computer or server. In the case of a remote computer, the remote computer may be connected to the user's computer through any type of network, including a Local Area Network (LAN) or a Wide Area Network (WAN), or the connection may be made to an external computer (for example, through the Internet using an Internet service provider). In some embodiments, aspects of the present invention are implemented by personalizing an electronic circuit, such as a programmable logic circuit, a Field Programmable Gate Array (FPGA), or a Programmable Logic Array (PLA), with state information of computer-readable program instructions, which can execute the computer-readable program instructions.
Aspects of the present invention are described herein with reference to flowchart illustrations and/or block diagrams of methods, apparatus (systems) and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and/or block diagrams, and combinations of blocks in the flowchart illustrations and/or block diagrams, can be implemented by computer-readable program instructions.
These computer-readable program instructions may be provided to a processor of a general purpose computer, special purpose computer, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create means for implementing the functions/acts specified in the flowchart and/or block diagram block or blocks. These computer-readable program instructions may also be stored in a computer-readable storage medium that can direct a computer, programmable data processing apparatus, and/or other devices to function in a particular manner, such that the computer-readable medium storing the instructions comprises an article of manufacture including instructions which implement the function/act specified in the flowchart and/or block diagram block or blocks.
The computer readable program instructions may also be loaded onto a computer, other programmable data processing apparatus, or other devices to cause a series of operational steps to be performed on the computer, other programmable apparatus or other devices to produce a computer implemented process such that the instructions which execute on the computer, other programmable apparatus or other devices implement the functions/acts specified in the flowchart and/or block diagram block or blocks.
The flowchart and block diagrams in the figures illustrate the architecture, functionality, and operation of possible implementations of systems, methods and computer program products according to various embodiments of the present invention. In this regard, each block in the flowchart or block diagrams may represent a module, segment, or portion of instructions, which comprises one or more executable instructions for implementing the specified logical function(s). In some alternative implementations, the functions noted in the block may occur out of the order noted in the figures. For example, two blocks shown in succession may, in fact, be executed substantially concurrently, or the blocks may sometimes be executed in the reverse order, depending upon the functionality involved. It will also be noted that each block of the block diagrams and/or flowchart illustration, and combinations of blocks in the block diagrams and/or flowchart illustration, can be implemented by special purpose hardware-based systems which perform the specified functions or acts, or combinations of special purpose hardware and computer instructions. It is well known to those skilled in the art that implementation by hardware, by software, and by a combination of software and hardware are equivalent.
Having described embodiments of the present invention, the foregoing description is intended to be exemplary, not exhaustive, and not limited to the embodiments disclosed. Many modifications and variations will be apparent to those of ordinary skill in the art without departing from the scope and spirit of the described embodiments. The terminology used herein is chosen in order to best explain the principles of the embodiments, the practical application, or improvements made to the technology in the marketplace, or to enable others of ordinary skill in the art to understand the embodiments disclosed herein. The scope of the invention is defined by the appended claims.

Claims (10)

1. A three-dimensional knowledge diagnosis model construction method comprises the following steps:
building a training set based on the known medical image data, the training set comprising
Figure FDA0002698772170000011
And
Figure FDA0002698772170000012
two moieties of which XmRepresents the m-th medical image data, YmLabel, X, representing mth example medical image dataiRepresenting the ith sample of unlabeled medical image data, wherein M and N are the number of corresponding samples respectively;
training a constructed three-dimensional knowledge diagnosis model based on a neural network by taking a set loss function as a target, wherein Xm，YmSupervised trained inputs and labels, X, as the principal branches of the three-dimensional knowledge diagnostic modeliAs input for unsupervised training of the main branches of the three-dimensional knowledge diagnostic model,
Figure FDA0002698772170000013
as an input of the unsupervised training of the auxiliary branch of the three-dimensional knowledge diagnostic model,
Figure FDA0002698772170000014
is to XiAnd carrying out various kinds of disturbed data.
2. The method of claim 1, wherein training the constructed neural network-based three-dimensional knowledge diagnostic model with the set loss function as a target comprises:
designing a three-dimensional knowledge diagnosis model, wherein a main branch and an auxiliary branch of the three-dimensional knowledge diagnosis model share the same feature extractor, and the main branch and the auxiliary branch respectively comprise an independent full-connection layer and a classifier;
medical image data Xm，Xi，
Figure FDA0002698772170000015
Inputting the three-dimensional knowledge diagnosis model, and extracting the features of the medical image through the feature extractor;
extracting Xm，XiInputting corresponding features into the full connection layer of the main branch of the three-dimensional knowledge diagnosis model for feature attribute combination, and extracting the features
Figure FDA0002698772170000016
Inputting the corresponding features into a full connection layer of an auxiliary branch of the three-dimensional knowledge diagnosis model to perform feature attribute combination;
inputting the features passing through the full connection layer into a corresponding classifier to obtain a predicted probability value pm，piAnd
Figure FDA0002698772170000017
wherein,
Figure FDA0002698772170000018
representing the output result, p, of the ith auxiliary branch of the ith imagemRepresents XmCorresponding probability value, piRepresents XiThe corresponding probability value is then used to determine,
Figure FDA0002698772170000019
to represent
Figure FDA00026987721700000110
A corresponding probability value;
obtaining the initial probability value pm，pi，
Figure FDA00026987721700000111
Inputting the aggregation layer to obtain an initial judgment probability value muiAnd optimized to obtain a more accurate predicted value mu'i；
And (3) obtaining the optimized three-dimensional knowledge diagnosis model by taking the set loss function as a target and conducting derivation back propagation through a chain rule.
3. The method of claim 2, wherein the obtained initial probability value p is usedm，pi，
Figure FDA00026987721700000112
Inputting the aggregation layer to obtain an initial judgment probability value muiAnd optimized to obtain a more accurate predicted value mu'iThe method comprises the following steps:
determining an initial decision probability value mu by maximum votingiThrough μiCalculation of initial sensitivity αjAnd specific betajWhere j ∈ P is considered the jth doctor:
Figure FDA0002698772170000021
wherein,
Figure FDA0002698772170000022
the calculation conditions expect that:
Figure FDA0002698772170000023
wherein D is the set of observations,
Figure FDA0002698772170000024
Figure FDA0002698772170000025
g is an implicit variable that is used to indicate,
Figure FDA00026987721700000211
is the current estimated parameter;
inferring mu using Bayesian theoryiAnd alphaj、βjIs expressed as:
Figure FDA0002698772170000026
by a given observation set D and obtained μiThe desired value is derived and the derivative is made 0, and alpha is calculatedjAnd betajExpressed as:
Figure FDA0002698772170000027
binding sensitivity alphajAnd specific betajA parameter learning classifier for obtaining a predicted value mu'i。
4. The method of claim 2, wherein obtaining the optimized three-dimensional knowledge diagnostic model by chain rule derivative back propagation with the set loss function as a target comprises:
and combining the polymerization layer loss function with a supervision loss function and a folding concave regularization loss function for back transmission, and solving the parameters of the three-dimensional knowledge diagnosis model through chain rule derivation and back transmission to obtain the optimized three-dimensional knowledge diagnosis model.
5. The method of claim 4, wherein the loss function is set to:
Figure FDA0002698772170000028
Pλ(. to) is a fold-valley regularization penalty function,
Figure FDA0002698772170000029
a, lambda are fine tuning parameters, J (w, b) are supervised classification loss functions,
Figure FDA00026987721700000210
wm,bmfor weight and bias, L (. mu.)i,pi) As a function of the loss of the polymeric layer.
6. The method of claim 5, wherein the loss function of the polymeric layer is expressed as:
Figure FDA0002698772170000031
7. the method of claim 2, wherein the three-dimensional knowledge diagnostic model is designed based on inclusion v3, wherein the feature extractor comprises a plurality of 3D convolutional layers and a plurality of block structures, and wherein the classifier is a softmax classifier.
8. A three-dimensional knowledge diagnostic model building system, comprising:
a data acquisition unit: for creating a training set based on known medical image data, the training set comprising
Figure FDA0002698772170000032
And
Figure FDA0002698772170000033
two moieties of which XmRepresents the m-th medical image data, YmLabel, X, representing mth example medical image dataiRepresenting the ith sample of unlabeled medical image data, wherein M and N are the number of corresponding samples respectively;
a model training unit: it is used for training a built three-dimensional knowledge diagnosis model based on a neural network by taking a set loss function as a target, wherein Xm，YmSupervised trained inputs and labels, X, as the principal branches of the three-dimensional knowledge diagnostic modeliAs input for unsupervised training of the main branches of the three-dimensional knowledge diagnostic model,
Figure FDA0002698772170000034
as an input of the unsupervised training of the auxiliary branch of the three-dimensional knowledge diagnostic model,
Figure FDA0002698772170000035
is to XiAnd carrying out various kinds of disturbed data.
9. A medical image recognition system, comprising:
a data acquisition unit: the system is used for acquiring medical image data to be detected;
an image recognition unit: a main branch for inputting the medical image data into a three-dimensional knowledge diagnosis model constructed according to the method of any one of claims 1 to 7, obtaining classification results.
10. A computer-readable storage medium, on which a computer program is stored which, when being executed by a processor, carries out the steps of the method according to any one of claims 1 to 7.
CN202011015038.8A 2020-09-24 2020-09-24 Three-dimensional knowledge diagnosis model construction method and system Active CN112086174B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202011015038.8A CN112086174B (en) 2020-09-24 2020-09-24 Three-dimensional knowledge diagnosis model construction method and system

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202011015038.8A CN112086174B (en) 2020-09-24 2020-09-24 Three-dimensional knowledge diagnosis model construction method and system

Publications (2)

Publication Number Publication Date
CN112086174A true CN112086174A (en) 2020-12-15
CN112086174B CN112086174B (en) 2023-12-12

Family

ID=73738700

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202011015038.8A Active CN112086174B (en) 2020-09-24 2020-09-24 Three-dimensional knowledge diagnosis model construction method and system

Country Status (1)

Country Link
CN (1) CN112086174B (en)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN112734906A (en) * 2020-12-30 2021-04-30 华东师范大学 Three-dimensional reconstruction method of ultrasonic or CT medical image based on knowledge distillation

Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN109034205A (en) * 2018-06-29 2018-12-18 西安交通大学 Image classification method based on the semi-supervised deep learning of direct-push
KR20190117969A (en) * 2018-04-09 2019-10-17 주식회사 뷰노 Method for semi supervised reinforcement learning using data with label and data without label together and apparatus using the same
US20200160177A1 (en) * 2018-11-16 2020-05-21 Royal Bank Of Canada System and method for a convolutional neural network for multi-label classification with partial annotations

Patent Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
KR20190117969A (en) * 2018-04-09 2019-10-17 주식회사 뷰노 Method for semi supervised reinforcement learning using data with label and data without label together and apparatus using the same
CN109034205A (en) * 2018-06-29 2018-12-18 西安交通大学 Image classification method based on the semi-supervised deep learning of direct-push
US20200160177A1 (en) * 2018-11-16 2020-05-21 Royal Bank Of Canada System and method for a convolutional neural network for multi-label classification with partial annotations

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN112734906A (en) * 2020-12-30 2021-04-30 华东师范大学 Three-dimensional reconstruction method of ultrasonic or CT medical image based on knowledge distillation
CN112734906B (en) * 2020-12-30 2022-08-19 华东师范大学 Three-dimensional reconstruction method of ultrasonic or CT medical image based on knowledge distillation

Also Published As

Publication number Publication date
CN112086174B (en) 2023-12-12

Similar Documents

Publication Publication Date Title
Fuhrman et al. A review of explainable and interpretable AI with applications in COVID‐19 imaging
Ozdemir et al. A 3D probabilistic deep learning system for detection and diagnosis of lung cancer using low-dose CT scans
Lin et al. Comparison of handcrafted features and convolutional neural networks for liver MR image adequacy assessment
Nichols et al. Machine learning: applications of artificial intelligence to imaging and diagnosis
CN111488921B (en) Intelligent analysis system and method for panoramic digital pathological image
US20220328189A1 (en) Systems, methods, and apparatuses for implementing advancements towards annotation efficient deep learning in computer-aided diagnosis
Wu et al. Cascaded fully convolutional networks for automatic prenatal ultrasound image segmentation
Bi et al. Automatic liver lesion detection using cascaded deep residual networks
Fang et al. Automatic breast cancer detection based on optimized neural network using whale optimization algorithm
Abdelsamea et al. A survey on artificial intelligence in histopathology image analysis
Wankhade et al. A novel hybrid deep learning method for early detection of lung cancer using neural networks
Li et al. Hierarchical deep network with uncertainty-aware semi-supervised learning for vessel segmentation
Sreenivasu et al. [Retracted] Dense Convolutional Neural Network for Detection of Cancer from CT Images
CN114530222B (en) Cancer patient classification system based on multiunit science and image data fusion
CN113724185B (en) Model processing method, device and storage medium for image classification
CN115115570A (en) Medical image analysis method and apparatus, computer device, and storage medium
Escobar et al. Voxel‐wise supervised analysis of tumors with multimodal engineered features to highlight interpretable biological patterns
Tian et al. Radiomics and its clinical application: artificial intelligence and medical big data
Tang et al. Lesion segmentation and RECIST diameter prediction via click-driven attention and dual-path connection
CN114580501A (en) Bone marrow cell classification method, system, computer device and storage medium
CN114565919A (en) Tumor microenvironment spatial relationship modeling system and method based on digital pathological image
CN112086174B (en) Three-dimensional knowledge diagnosis model construction method and system
Gu et al. Efficient echocardiogram view classification with sampling-free uncertainty estimation
US20240028831A1 (en) Apparatus and a method for detecting associations among datasets of different types
CN115831356B (en) Auxiliary prediction diagnosis method based on artificial intelligence algorithm

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