CN108596200A - The method and apparatus of Medical Images Classification - Google Patents
The method and apparatus of Medical Images Classification Download PDFInfo
- Publication number
- CN108596200A CN108596200A CN201810006166.2A CN201810006166A CN108596200A CN 108596200 A CN108596200 A CN 108596200A CN 201810006166 A CN201810006166 A CN 201810006166A CN 108596200 A CN108596200 A CN 108596200A
- Authority
- CN
- China
- Prior art keywords
- feature
- input
- medical image
- input example
- neural networks
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; 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
- G06F18/2155—Generating training patterns; Bootstrap methods, e.g. bagging or boosting characterised by the incorporation of unlabelled data, e.g. multiple instance learning [MIL], semi-supervised techniques using expectation-maximisation [EM] or naïve labelling
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/243—Classification techniques relating to the number of classes
- G06F18/24323—Tree-organised classifiers
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; 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
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Physics & Mathematics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Artificial Intelligence (AREA)
- General Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- Evolutionary Computation (AREA)
- Bioinformatics & Computational Biology (AREA)
- Evolutionary Biology (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Computational Linguistics (AREA)
- Biomedical Technology (AREA)
- Biophysics (AREA)
- Health & Medical Sciences (AREA)
- General Health & Medical Sciences (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Image Analysis (AREA)
Abstract
The present invention provides a kind of method and apparatus of Medical Images Classification, can solve the problems, such as that the automatic classification accuracy of medical image is low.This method includes：Medical image is instantiated, input example is obtained；Learn and extract the feature of the input example automatically by convolutional neural networks；According to the feature, the input example is classified by grader, obtains input Exemplary classes result；According to the input Exemplary classes as a result, classifying by medical image described in flexible maximum delivered function pair, final classification result is obtained.
Description
Technical field
The present invention relates to field of computer technology more particularly to a kind of method and apparatus of Medical Images Classification.
Background technology
With the rapid development of medical imaging technology, various medical images are in explosive growth, become medical clinic applications
With the essential tool of teaching research.At the same time, the diagnosis identification of medical image, becomes the research hotspot of medical domain.
The classification for how accurately and efficiently determining medical image, also becomes the key point in numerous researchs.
Existing Medical Images Classification algorithm is broadly divided into design two parts of feature learning and grader.Current feature
Learning process is summed up by artificially observing come than more significant mainly by doctor according to the Clinical symptoms of medical image complexity
Feature, be used in combination mathematical linguistics to express；Then corresponding grader is designed according to different characteristic；It is learned finally by full supervision
The mode of habit trains to obtain object classifiers.
In realizing process of the present invention, inventor has found that at least there are the following problems in the prior art：
There is centainly incomprehensive by the feature of doctor's manual extraction, express the medicine of input with being unable to Comprehensive
Image；The training process of grader based on full monitor mode needs a large amount of artificial samples marked in detail, and this process takes
Effort and have certain subjectivity.
Invention content
In view of this, the embodiment of the present invention provides a kind of method, apparatus and device of Medical Images Classification, can solve to cure
Learn the low problem of classification of images accuracy rate.
To achieve the above object, one side according to the ... of the embodiment of the present invention provides a kind of side of Medical Images Classification
Method.
A kind of method of Medical Images Classification of the embodiment of the present invention includes：Medical image is instantiated, is inputted
Example；Learn and extract the feature of the input example automatically by convolutional neural networks；According to the feature, pass through grader
The input example is classified, input Exemplary classes result is obtained；According to the input Exemplary classes as a result, passing through flexibility
Medical image is classified described in maximum delivered function pair, obtains final classification result.
Optionally, it is described to medical image carry out instantiation include：The medical image is divided in a manner of sliding window.
Optionally, the feature for being learnt automatically by convolutional neural networks and extracting the input example includes：Pass through
The convolutional neural networks learn the feature of the input example automatically, are carried by the full articulamentum in the convolutional neural networks
Take out the feature of the input example.
Optionally, the grader is single layer decision tree classifier.
To achieve the above object, one side according to the ... of the embodiment of the present invention provides a kind of dress of Medical Images Classification
It sets.
A kind of device of Medical Images Classification of the embodiment of the present invention includes：Module is instantiated, for medical image example
Change to obtain input example；Feature learning and extraction module learn and extract described defeated for passing through convolutional neural networks automatically
Enter the feature of example；Exemplary classes module, for according to the feature, the input example to be classified by grader,
Obtain input Exemplary classes result；Image classification module, for being passed according to the input Exemplary classes as a result, passing through flexible maximum
Delivery function classifies to the medical image, obtains final classification result.
Optionally, the instantiation module includes：The medical image is divided in a manner of sliding window.
Optionally, the feature learning and extraction module include：Learnt automatically by the convolutional neural networks described defeated
The feature for entering example extracts the feature of the input example by the full articulamentum in the convolutional neural networks.
Optionally, in the Exemplary classes module, using single layer decision tree as grader.
To achieve the above object, according to the ... of the embodiment of the present invention in another aspect, providing a kind of realization Medical Images Classification
Method electronic equipment.
The a kind of electronic equipment of the embodiment of the present invention includes：One or more processors；Storage device, for storing one
Or multiple programs, when one or more of programs are executed by one or more of processors so that one or more of
The method that processor realizes the Medical Images Classification of the embodiment of the present invention.
To achieve the above object, another aspect according to the ... of the embodiment of the present invention, provides a kind of computer-readable medium.
A kind of computer-readable medium of the embodiment of the present invention, is stored thereon with computer program, and described program is handled
For realizing the method for the Medical Images Classification for making the computer execution embodiment of the present invention when device executes.
One embodiment in foregoing invention has the following advantages that or advantageous effect：Compared to incomplete manual extraction
Feature more accurate can input figure by the automated characterization study of neural fusion and extracting method with comprehensively expression
The feature of picture, to improve the classification accuracy of medical image；By using more case-based learning algorithms, can weaken to sample
The dependence manually marked learn fine locality information automatically using rough markup information of overall importance, avoid
The artificial mark sample processes for taking time and effort and having certain subjectivity, in the minimum artificial markup information of guarantee and most
While the feature representation got well, the automatic classification for carrying out medical image.
Further effect possessed by above-mentioned non-usual optional mode adds hereinafter in conjunction with specific implementation mode
With explanation.
Description of the drawings
Attached drawing does not constitute inappropriate limitation of the present invention for more fully understanding the present invention.Wherein：
Fig. 1 is the schematic diagram of the key step of the method for Medical Images Classification according to the ... of the embodiment of the present invention；
Fig. 2 is the flow diagram of the method for Medical Images Classification according to the ... of the embodiment of the present invention；
Fig. 3 is the schematic diagram of the main modular of the device of Medical Images Classification according to the ... of the embodiment of the present invention；
Fig. 4 is adapted for the structural representation of the computer system for the terminal device or server of realizing the embodiment of the present application
Figure.
Specific implementation mode
It explains to the exemplary embodiment of the present invention below in conjunction with attached drawing, including the various of the embodiment of the present invention
Details should think them only exemplary to help understanding.Therefore, those of ordinary skill in the art should recognize
It arrives, various changes and modifications can be made to the embodiments described herein, without departing from scope and spirit of the present invention.Together
The description to known function and structure is omitted for clarity and conciseness in sample in following description.
The technical solution of the embodiment of the present invention inputs one group of medical image first, and instantiation processing is carried out to it；Then lead to
Cross the feature that convolutional neural networks learn and extract the input example automatically；It, will be described by grader according to the feature
Input example is classified, and input Exemplary classes result is obtained；According to the input Exemplary classes as a result, passing through flexible maximum biography
Delivery function classifies to the medical image, obtains final classification result.
Fig. 1 is the schematic diagram of the key step of the method for Medical Images Classification according to the ... of the embodiment of the present invention；
As shown in Figure 1, the method for the Medical Images Classification of the embodiment of the present invention mainly includes the following steps：
Step S11：Medical image is instantiated, input example is obtained.It in this step, can be by the medicine figure
As being divided in a manner of sliding window.
Step S12：Learn and extract the feature of the input example automatically by convolutional neural networks.In this step,
Neural network can be designed first, and the training process that may then pass through on given data set obtains the optimal of neural network
Weight, the neural network of optimization can carry out the operation of convolution sum maximum pondization or progress convolution sum by the example to input
Average pondization operation, study arrive optimal feature representation, full articulamentum finally may be used, by the feature of learn 1000 dimensions
It extracts.
Step S13：According to the feature, the input example is classified by grader, obtains input example point
Class result.In this step, gradient can be passed through using single layer decision tree as grader based on the frame of more case-based learnings
Descent algorithm iteratively trains grader, and optimal grader is finally obtained when object function is restrained.The grader can root
According to the feature of the extraction by the Exemplary classes.
Step S14：According to the input Exemplary classes as a result, by medical image described in flexible maximum delivered function pair into
Row classification, obtains final classification result.In this step, General Mean algorithms may be used as flexible maximum delivered letter
Number obtains the classification results of described image according to the classification results of the input example.
The implementation of specific Medical Images Classification is specific as follows：
Fig. 2 is the flow diagram of the method for Medical Images Classification according to the ... of the embodiment of the present invention；
As shown in Fig. 2, the present invention carries out instantiation processing to it first, then passes through after inputting one group of input picture
Convolutional neural networks learn and extract the feature of the input example automatically, and according to the feature, will be described by grader
Input example is classified, and obtains input Exemplary classes as a result, finally according to the input Exemplary classes as a result, most by flexibility
Big transmission function classifies to the medical image, obtains final classification result.
With reference to specific embodiment, the technological means of the present invention is illustrated：
1. instantiation is handled
Original medical image size is 10000x10000 pixels.Herein by the window of 200x200 pixels with 100 pixels
It is overlapped step-length slip scan on the original image, to which original image to be cut to the image subblock of 9801 200x200 pixels
(inputting example).
2. feature learning and extraction
The input example includes large amount of complex, redundancy information, therefore the input is examples translating for can be high
The characteristic of degree expression example information excavated, is medical image analysis necessary step, the process i.e. characterology
It practises and extracts.The convolutional neural networks for possessing powerful feature learning and feature representation ability are feature learning and extraction process
Only choosing.Here, design neural network 3x200x200-32C5-MP2-32C5-MP2-32C5-MP2-64C5-MP2-
1000N learns and extracts the feature of the input example automatically.The input example of one group of 3 channel 200x200 pixel is given, is passed through
After core size is the convolutional layer of 5x5 pixels, shallow-layer characteristic pattern is obtained；It is 2x2 pixels to be then input to core size
Maximum pond layer, further extracts main feature.After three identical convolutional layers and maximum pond layer, further feature is obtained
Figure.Finally by full articulamentum, the feature extraction of 1000 dimensions learnt is come out.Wherein, 3x200x200 indicates that input is real
Example is the image subblock of triple channel 200x200 pixels.C (convolutional layer) indicates convolutional layer.32 and 64 difference tables
Show the port number of the quantity of the convolution kernel of the convolutional layer namely the characteristic pattern of convolutional layer output.MP (Max pooling) is indicated
Maximum pond layer.In deep neural network, the convolutional layer extraction low-level features of bottom, such as edge, angle point, and higher
Convolutional layer extracts the more advanced feature for containing clearer semantic information.The process of the deep learning of neural network, be exactly from
Low-level features are gradually inferred to the process of advanced features.Finally, all offices with class discrimination are integrated using full articulamentum
Portion's information extracts the foundation as classification.Wherein, design pond layer is after convolutional layer, it is therefore an objective under certain
Sampling rule carries out down-sampling to characteristic pattern, to keep to characteristic pattern dimensionality reduction and to a certain extent the Scale invariant of feature
Property, while simplifying the computation complexity of network and preventing over-fitting to a certain extent.
3. Exemplary classes
It according to the feature, needs that the input example is classified by grader, obtains input Exemplary classes knot
Fruit.But the training process of the grader based on full monitor mode needs the sample largely marked in detail, and sample is detailed
Artificial mark not only takes time and effort, and has certain subjectivity.Thus, it is intended that utilize rough mark of overall importance
Information learns fine locality information automatically.The central idea of this namely more case-based learning.Defining a medical image is
One is wrapped, and the image subblock in image is example, and it is positive example to have the image subblock of cancer, and the image subblock of no cancer is negative example.
When at least one positive example in packet, this coating is labeled as positive closure.Define Xi={ xi1, xi2..., ximIt is instruction
Practice i-th of the packet concentrated, m is the label of example in packet, { xi1, xi2..., ximFor all examples in i-th of packet.yi∈{-
1 ,+1 } be image level i.e. package level label, that is, the rough markup information of overall importance.- 1 represents negative packet ,+
1 represents positive closure.yij∈ { -1 ,+1 } is the label of interior j-th of the example of i-th of packet namely the fine locality information.If
Count H (x)：X → [0,1] refers to the grader of package level, h (x)：X → [0,1] refers to the other grader of instance-level.Here,
We select single layer decision tree as grader.Then, using cross entropy loss function as object function.It is defined as follows：
Wherein, 1 (yi=1) and 1 (yi=-1) all it is indicator function.Using gradient descent algorithm, iteratively training divides for we
Class device h ' (x)：Then grader h is updated according to equation h (x) ← h (x)+α h ' (x)
(x).Here, the value of factor alpha depends on minimizing the line search algorithm of object function.After the iteration of enough times, mesh
Scalar functions are restrained, to obtain optimal Exemplary classes device h (x).Exemplary classes device classifies the example, to obtain
Exemplary classes result yij。
4. image classification
According to the input Exemplary classes as a result, point of the medical image can be obtained by flexible maximum delivered function
Class result.Here, using General Mean (GM) function as flexible maximum delivered function, by Exemplary classes result yijIt calculates
Obtain image classification result yi.Wherein, GM functions are defined as follows：
The method of Medical Images Classification according to the ... of the embodiment of the present invention can be seen that compared to incomplete manual extraction
Feature more accurate can input figure by the automated characterization study of neural fusion and extracting method with comprehensively expression
The feature of picture, to improve the classification accuracy of medical image；By using more case-based learning algorithms, can weaken to sample
The dependence manually marked learn fine locality information automatically using rough markup information of overall importance, avoid
The artificial mark sample processes for taking time and effort and having certain subjectivity, in the minimum artificial markup information of guarantee and most
While the feature representation got well, the automatic classification for carrying out medical image.
Fig. 3 is the schematic diagram of the main modular of the device of Medical Images Classification according to the ... of the embodiment of the present invention；
As shown in figure 3, the device 300 of the Medical Images Classification of the embodiment of the present invention includes mainly：Instantiation module 301,
Feature learning and extraction module 302, Exemplary classes module 303, image classification module 304.Wherein：
Instantiation module 301 can be used for instantiating medical image to obtain input example；Feature learning and extraction module
302 can be used for learning automatically by convolutional neural networks and extracting the feature of the input example；Exemplary classes module 303 can be used
According to the feature, the input example is classified by grader, obtains input Exemplary classes result；Image classification
Module 304 can be used for according to the input Exemplary classes as a result, being carried out by medical image described in flexible maximum delivered function pair
Classification, obtains final classification result.
From the above, it can be seen that compared to the feature of incomplete manual extraction, by neural fusion from
Dynamic feature learning and extracting method, the feature of input picture more can accurately and be comprehensively expressed, to improve medicine figure
The classification accuracy of picture；By using more case-based learning algorithms, the dependence manually marked to sample can be weakened, using thick
Markup information of overall importance slightly, learns fine locality information automatically, avoids and takes time and effort and have certain subjectivity
Property artificial mark sample processes, to while ensureing minimum artificial markup information and best feature representation, automatically
Carry out the classification of medical image.
According to an embodiment of the invention, the present invention also provides a kind of electronic equipment and a kind of readable medium.
The present invention electronic equipment include：One or more processors；Storage device, for storing one or more journeys
Sequence, when one or more of programs are executed by one or more of processors so that one or more of processors are real
The method of the Medical Images Classification of the existing embodiment of the present invention.
The computer-readable medium of the present invention, is stored thereon with computer program, is used when described program is executed by processor
In the method for realizing the Medical Images Classification for making the computer execute the embodiment of the present invention.
Fig. 4 is adapted for the structural representation of the computer system for the terminal device or server of realizing the embodiment of the present application
Figure.
As shown in figure 4, it illustrates the computer systems 400 suitable for the terminal device for realizing the embodiment of the present application
Structural schematic diagram.Terminal device shown in Fig. 4 is only an example, should not be to the function and use scope of the embodiment of the present application
Bring any restrictions.
As shown in figure 4, computer system 400 includes central processing unit (CPU) 401, it can be read-only according to being stored in
Program in memory (ROM) 402 or be loaded into the program in random access storage device (RAM) 403 from storage section 408 and
Execute various actions appropriate and processing.In RAM 403, also it is stored with system 400 and operates required various programs and data.
CPU 401, ROM 402 and RAM 403 are connected with each other by bus 404.Input/output (I/O) interface 405 is also connected to always
Line 404.
It is connected to I/O interfaces 405 with lower component：Importation 406 including keyboard, mouse etc.；It is penetrated including such as cathode
The output par, c 407 of spool (CRT), liquid crystal display (LCD) etc. and loud speaker etc.；Storage section 408 including hard disk etc.；
And the communications portion 409 of the network interface card including LAN card, modem etc..Communications portion 409 via such as because
The network of spy's net executes communication process.Driver 410 is also according to needing to be connected to I/O interfaces 405.Detachable media 411, such as
Disk, CD, magneto-optic disk, semiconductor memory etc. are mounted on driver 410, as needed in order to be read from thereon
Computer program be mounted into storage section 408 as needed.
Particularly, it according to embodiment disclosed by the invention, may be implemented as counting above with reference to the process of flow chart description
Calculation machine software program.For example, embodiment disclosed by the invention includes a kind of computer program product comprising be carried on computer
Computer program on readable medium, the computer program include the program code for method shown in execution flow chart.
In such embodiment, which can be downloaded and installed by communications portion 409 from network, and/or from can
Medium 411 is dismantled to be mounted.When the computer program is executed by central processing unit (CPU) 401, the system that executes the application
The above-mentioned function of middle restriction.
It should be noted that computer-readable medium shown in the application can be computer-readable signal media or meter
Calculation machine readable storage medium storing program for executing either the two arbitrarily combines.Computer readable storage medium for example can be --- but not
Be limited to --- electricity, magnetic, optical, electromagnetic, infrared ray or semiconductor system, device or device, or arbitrary above combination.Meter
The more specific example of calculation machine readable storage medium storing program for executing can include but is not limited to：Electrical connection with one or more conducting wires, just
It takes formula computer disk, hard disk, random access storage device (RAM), read-only memory (ROM), erasable type and may be programmed read-only storage
Device (EPROM or flash memory), optical fiber, portable compact disc read-only memory (CD-ROM), light storage device, magnetic memory device,
Or above-mentioned any appropriate combination.In this application, can be any include computer readable storage medium or storage journey
The tangible medium of sequence, the program can be commanded the either device use or in connection of execution system, device.And at this
In application, computer-readable signal media may include in a base band or as the data-signal that a carrier wave part is propagated,
Wherein carry computer-readable program code.Diversified forms may be used in the data-signal of this propagation, including but unlimited
In electromagnetic signal, optical signal or above-mentioned any appropriate combination.Computer-readable signal media can also be that computer can
Any computer-readable medium other than storage medium is read, which can send, propagates or transmit and be used for
By instruction execution system, device either device use or program in connection.Include on computer-readable medium
Program code can transmit with any suitable medium, including but not limited to：Wirelessly, electric wire, optical cable, RF etc. or above-mentioned
Any appropriate combination.
Flow chart in attached drawing and block diagram, it is illustrated that according to the system of the various embodiments of the application, method and computer journey
The architecture, function and operation in the cards of sequence product.In this regard, each box in flowchart or block diagram can generation
A part for a part for one module, program segment, or code of table, above-mentioned module, program segment, or code includes one or more
Executable instruction for implementing the specified logical function.It should also be noted that in some implementations as replacements, institute in box
The function of mark can also occur in a different order than that indicated in the drawings.For example, two boxes succeedingly indicated are practical
On can be basically executed in parallel, they can also be executed in the opposite order sometimes, this is depended on the functions involved.Also it wants
It is noted that the combination of each box in block diagram or flow chart and the box in block diagram or flow chart, can use and execute rule
The dedicated hardware based systems of fixed functions or operations is realized, or can use the group of specialized hardware and computer instruction
It closes to realize.
Being described in module involved in the embodiment of the present application can be realized by way of software, can also be by hard
The mode of part is realized.Described module can also be arranged in the processor, for example, can be described as：A kind of processor packet
Include instantiation module, feature learning and extraction module, Exemplary classes module, image classification module.Wherein, the title of these units
The restriction to the unit itself is not constituted under certain conditions, for example, feature learning and extraction module are also described as
" automatic study and extraction input example aspects module ".
As on the other hand, present invention also provides a kind of computer-readable medium, which can be
Included in equipment described in above-described embodiment；Can also be individualism, and without be incorporated the equipment in.Above-mentioned calculating
Machine readable medium carries one or more program, when said one or multiple programs are executed by the equipment, makes
Obtaining the equipment includes：This method includes：Medical image is instantiated, input example is obtained；Certainly by convolutional neural networks
The feature of the dynamic medical image for learning and extracting the input example；It is by grader that the input is real according to the feature
Example is classified, and input Exemplary classes result is obtained；According to the input Exemplary classes as a result, passing through flexible maximum delivered function
Classify to the medical image, obtains final classification result.
The said goods can perform the method that the embodiment of the present invention is provided, and has the corresponding function module of execution method and has
Beneficial effect.The not technical detail of detailed description in the present embodiment, reference can be made to the method that the embodiment of the present invention is provided.
Technical solution according to the ... of the embodiment of the present invention passes through neural network compared to the feature of incomplete manual extraction
The automated characterization of realization learns and extracting method, the feature of input picture more can accurately and be comprehensively expressed, to improve
The classification accuracy of medical image；By using more case-based learning algorithms, the dependence manually marked to sample can be weakened
Property, using rough markup information of overall importance, learn fine locality information automatically, avoids and take time and effort and have one
The artificial mark sample processes of fixed subjectivity, to ensure the same of minimum artificial markup information and best feature representation
When, the automatic classification for carrying out medical image.
Above-mentioned specific implementation mode, does not constitute limiting the scope of the invention.Those skilled in the art should be bright
It is white, design requirement and other factors are depended on, various modifications, combination, sub-portfolio and replacement can occur.It is any
Modifications, equivalent substitutions and improvements made by within the spirit and principles in the present invention etc., should be included in the scope of the present invention
Within.
Claims (10)
1. a kind of method that medical image is classified automatically, which is characterized in that including：
Medical image is instantiated, input example is obtained；
Learn and extract the feature of the input example automatically by convolutional neural networks；
According to the feature, the input example is classified by grader, obtains input Exemplary classes result；
According to the input Exemplary classes as a result, being classified by medical image described in flexible maximum delivered function pair, obtain
Final classification result.
2. according to the method described in claim 1, it is characterized in that, it is described to medical image carry out instantiation include：It will be described
Medical image is divided in a manner of sliding window.
3. according to the method described in claim 1, it is characterized in that, described learnt by convolutional neural networks and extract institute automatically
State input example feature include：The feature for learning the input example automatically by the convolutional neural networks, by described
Full articulamentum in convolutional neural networks extracts the feature of the input example.
4. according to the method described in claim 1, it is characterized in that, the grader is single layer decision tree classifier.
5. a kind of device that medical image is classified automatically, which is characterized in that including：
Module is instantiated, for being instantiated to medical image to obtain input example；
Feature learning and extraction module, the feature for learning and extracting the input example automatically by convolutional neural networks；
It is real to obtain input for according to the feature, the input example to be classified by grader for Exemplary classes module
Example classification results；
Image classification module is used for according to the input Exemplary classes as a result, passing through medicine described in flexible maximum delivered function pair
Image is classified, and final classification result is obtained.
6. according to the method described in claim 5, it is characterized in that, the instantiation module includes：By the medical image with
The mode of sliding window is divided.
7. according to the method described in claim 5, it is characterized in that, the feature learning and extraction module include：By described
Convolutional neural networks learn the feature of the input example automatically, are extracted by the full articulamentum in the convolutional neural networks
The feature of the input example.
8. according to the method described in claim 5, it is characterized in that, in the Exemplary classes module, made using single layer decision tree
For grader.
9. a kind of electronic equipment, which is characterized in that including：
One or more processors；
Storage device, for storing one or more programs,
When one or more of programs are executed by one or more of processors so that one or more of processors are real
The now method as described in claim 1-4 is any.
10. a kind of computer-readable medium, is stored thereon with computer program, which is characterized in that described program is held by processor
The method as described in any in claim 1-4 is realized when row.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810006166.2A CN108596200A (en) | 2018-01-03 | 2018-01-03 | The method and apparatus of Medical Images Classification |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810006166.2A CN108596200A (en) | 2018-01-03 | 2018-01-03 | The method and apparatus of Medical Images Classification |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN108596200A true CN108596200A (en) | 2018-09-28 |
Family
ID=63599778
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810006166.2A Pending CN108596200A (en) | 2018-01-03 | 2018-01-03 | The method and apparatus of Medical Images Classification |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN108596200A (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109378010A (en) * | 2018-10-29 | 2019-02-22 | 珠海格力电器股份有限公司 | Neural network model training method, voice denoising method and device |
| CN112116609A (en) * | 2019-06-21 | 2020-12-22 | 斯特拉克斯私人有限公司 | Machine learning classification method and system based on structure or material segmentation in image |
| CN115777107A (en) * | 2020-06-05 | 2023-03-10 | 第一百欧有限公司 | Neural network learning method through automatic encoder and multi-instance learning and computing system for executing neural network learning method |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20090080731A1 (en) * | 2007-09-26 | 2009-03-26 | Siemens Medical Solutions Usa, Inc. | System and Method for Multiple-Instance Learning for Computer Aided Diagnosis |
| CN104077599A (en) * | 2014-07-04 | 2014-10-01 | 西安电子科技大学 | Polarization SAR image classification method based on deep neural network |
| CN106682696A (en) * | 2016-12-29 | 2017-05-17 | 华中科技大学 | Multi-example detection network based on refining of online example classifier and training method thereof |
-
2018
- 2018-01-03 CN CN201810006166.2A patent/CN108596200A/en active Pending
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20090080731A1 (en) * | 2007-09-26 | 2009-03-26 | Siemens Medical Solutions Usa, Inc. | System and Method for Multiple-Instance Learning for Computer Aided Diagnosis |
| CN104077599A (en) * | 2014-07-04 | 2014-10-01 | 西安电子科技大学 | Polarization SAR image classification method based on deep neural network |
| CN106682696A (en) * | 2016-12-29 | 2017-05-17 | 华中科技大学 | Multi-example detection network based on refining of online example classifier and training method thereof |
Non-Patent Citations (1)
| Title |
|---|
| YAN XU等: "Deep learning of feature representation with multiple instance learning for medical image analysis", 《2014 IEEE INTERNATIONAL CONFERENCE ON ACOUSTICS, SPEECH AND SIGNAL PROCESSING (ICASSP)》 * |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109378010A (en) * | 2018-10-29 | 2019-02-22 | 珠海格力电器股份有限公司 | Neural network model training method, voice denoising method and device |
| CN112116609A (en) * | 2019-06-21 | 2020-12-22 | 斯特拉克斯私人有限公司 | Machine learning classification method and system based on structure or material segmentation in image |
| CN115777107A (en) * | 2020-06-05 | 2023-03-10 | 第一百欧有限公司 | Neural network learning method through automatic encoder and multi-instance learning and computing system for executing neural network learning method |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US11586851B2 (en) | Image classification using a mask image and neural networks | |
| EP3968179A1 (en) | Place recognition method and apparatus, model training method and apparatus for place recognition, and electronic device | |
| US12061991B2 (en) | Transfer learning with machine learning systems | |
| US10599978B2 (en) | Weighted cascading convolutional neural networks | |
| US10828000B2 (en) | Medical image data analysis | |
| CN108109152A (en) | Medical Images Classification and dividing method and device | |
| CN114580501B (en) | Bone marrow cell classification method, system, computer device and storage medium | |
| CN112950642B (en) | Training method, device, electronic device and medium for point cloud instance segmentation model | |
| CN111881944A (en) | Method, electronic device and computer readable medium for image authentication | |
| CN112215244A (en) | Cargo image detection method, device, equipment and storage medium | |
| CN115797637B (en) | Semi-supervised segmentation model based on inter-model and intra-model uncertainty | |
| Gehlot et al. | Ednfc-net: Convolutional neural network with nested feature concatenation for nuclei-instance segmentation | |
| US20240020531A1 (en) | System and Method for Transforming a Trained Artificial Intelligence Model Into a Trustworthy Artificial Intelligence Model | |
| CN108596200A (en) | The method and apparatus of Medical Images Classification | |
| CN109657056A (en) | Target sample acquisition methods, device, storage medium and electronic equipment | |
| CN114638973A (en) | Target image detection method and image detection model training method | |
| Cao et al. | Supervised contrastive pre-training for mammographic triage screening models | |
| CN119942206A (en) | Image abnormal state classification method, device, computer equipment and storage medium | |
| CN113537339A (en) | Method and system for identifying symbiotic or associated minerals based on multi-label image classification | |
| CN113642590A (en) | Cargo image processing method, device, equipment and storage medium | |
| Li et al. | Labour-saving detection of hybrid rice rows at the pollination stage based on a multi-perturbed semi-supervised model | |
| CN107341189A (en) | A kind of indirect labor carries out the method and system of examination, classification and storage to image | |
| CN108257081A (en) | For generating the method and apparatus of picture | |
| CN115375657A (en) | Training method, detection method, device, medium and equipment of polyp detection model | |
| CN118506040B (en) | Remote sensing image clustering method, device, equipment and medium |
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 | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20180928 |
|
| RJ01 | Rejection of invention patent application after publication |