CN107945181A - Treating method and apparatus for breast cancer Lymph Node Metastasis pathological image - Google Patents
Treating method and apparatus for breast cancer Lymph Node Metastasis pathological image Download PDFInfo
- Publication number
- CN107945181A CN107945181A CN201711492182.9A CN201711492182A CN107945181A CN 107945181 A CN107945181 A CN 107945181A CN 201711492182 A CN201711492182 A CN 201711492182A CN 107945181 A CN107945181 A CN 107945181A
- Authority
- CN
- China
- Prior art keywords
- image
- image example
- probability
- happening
- block
- 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; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/0002—Inspection of images, e.g. flaw detection
- G06T7/0012—Biomedical image inspection
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/11—Region-based segmentation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/136—Segmentation; Edge detection involving thresholding
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/187—Segmentation; Edge detection involving region growing; involving region merging; involving connected component labelling
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/194—Segmentation; Edge detection involving foreground-background segmentation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/20—Special algorithmic details
- G06T2207/20081—Training; Learning
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/20—Special algorithmic details
- G06T2207/20084—Artificial neural networks [ANN]
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/30—Subject of image; Context of image processing
- G06T2207/30004—Biomedical image processing
- G06T2207/30068—Mammography; Breast
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/30—Subject of image; Context of image processing
- G06T2207/30004—Biomedical image processing
- G06T2207/30096—Tumor; Lesion
Abstract
The present invention provides a kind for the treatment of method and apparatus for breast cancer Lymph Node Metastasis pathological image.The processing method for breast cancer Lymph Node Metastasis pathological image of the present invention includes：Step S1：Image example is obtained, then obtains the corresponding image example calibration information of image example；Step S2：According to image example and image example calibration information generation training sample set；Step S3：Training and the parameter optimization of deep learning network are carried out according to training sample set, obtains convolutional neural networks model；Step S4：Test image is tested using convolutional neural networks model, detects target area.The treating method and apparatus for breast cancer Lymph Node Metastasis pathological image of the present invention realizes the automatic classification of image using the technology of deep learning, the technical problems such as the uniformity that solves the classification work of artificial diagosis is poor, accuracy is low, improve work efficiency, reduce probability of failure.
Description
Technical field
The present invention relates to field of computer technology, more particularly to a kind of processing for breast cancer Lymph Node Metastasis pathological image
Method and apparatus.
Background technology
With medical image development in science and technology, the importance of medical imaging increasingly increases.Medical personnel
Greater number more high-definition medical image can be obtained, these medical images are observed and analyzed, is finally made point
Class is evaluated.
The prior art mainly carries out artificial diagosis analysis data by veteran doctor.Specific practice is doctor couple
Pathological section is observed under the microscope made of pathological tissues, manual by the amplification factor for ceaselessly manually adjusting object lens
Progressively scan slice, using Pathomorphology inspection method, judges the specific location of lesion region and identifies each lesion
The corresponding lesion degree in region or lesion type.From the foregoing, it will be observed that the prior art has the shortcomings that time-consuming and laborious, the other prior art
The experience of diagosis person is relied primarily on, with subjectivity is strong, uniformity is poor, accuracy is low, accuracy is poor (for example, the leaching of breast cancer
The small transfer fawned on is difficult to be found, and is easily failed to pinpoint a disease in diagnosis) the shortcomings of.
The content of the invention
In view of this, the embodiment of the present invention provides a kind of processing method and dress for breast cancer Lymph Node Metastasis pathological image
Put, the technical problems such as the uniformity that can solve the classification work of artificial diagosis is poor, accuracy is low.
To achieve the above object, the first aspect according to embodiments of the present invention, there is provided one kind is used for breast cancer lymph
The processing method of pathological image is shifted, including：Step S1：Image example is obtained, then obtains the corresponding model of the image example
Example image calibration information；Step S2：According to the image example and image example calibration information generation training sample set；Step
Rapid S3：Training and the parameter optimization of deep learning network are carried out according to the training sample set, obtains convolutional neural networks model；
Step S4：Test image is tested using the convolutional neural networks model, detects target area.
Alternatively, step S2 includes：Step S21：The image example is divided into the image example of multiple pre-set dimensions
Block；Step S22：Information is demarcated with reference to the image example, each image example block is determined using characteristic point determining method
The positive and negative attribute of sample, to obtain the training sample set comprising image example block positive sample and image example block negative sample.
Alternatively, step S4 includes：Step S41：The test image is divided into multiple test image blocks, then to each
A test image block is classified using convolutional neural networks model, and it is general to obtain the corresponding event of each test image block
Rate；Step S42：Probability of happening thermodynamic chart is drawn according to each test image block and its corresponding probability of happening；Step
S43：Target area is confirmed according to the probability of happening thermodynamic chart.
Alternatively, step S43 includes：The probability of happening is extracted in the probability of happening thermodynamic chart and is more than predetermined probabilities threshold value
Region, as candidate target region；To each feature vector of the candidate target region extraction based on statistics and form simultaneously
Classified using random forest method, finally determine whether the candidate target region is the target area.
To achieve the above object, according to embodiments of the present invention second aspect, it is proposed that one kind is used for breast cancer lymph
The processing unit of pathological image is shifted, including：First processing module, for obtaining image example, then obtains the example figure
As corresponding image example demarcates information；Second processing module, for according to the image example and image example calibration
Information generates training sample set；3rd processing module, for carrying out the training of deep learning network according to the training sample set
And parameter optimization, obtain convolutional neural networks model；Fourth processing module, for utilizing the convolutional neural networks model to surveying
Attempt, as being tested, to detect target area.
Alternatively, the Second processing module is additionally operable to：The image example is divided into the example of multiple pre-set dimensions
Image block；Information is demarcated with reference to the image example, the sample of each image example block is determined using characteristic point determining method
Positive and negative attribute, to obtain the training sample set comprising image example block positive sample and image example block negative sample.
Alternatively, the fourth processing module is additionally operable to：The test image is divided into multiple test image blocks, then
Classified to each test image block using convolutional neural networks model, obtain the corresponding event of each test image block
Probability；Probability of happening thermodynamic chart is drawn according to each test image block and its corresponding probability of happening；According to the event
Probability thermodynamic chart confirms target area.
Alternatively, the fourth processing module is additionally operable to：The probability of happening is extracted in the probability of happening thermodynamic chart to be more than
The region of predetermined probabilities threshold value, as candidate target region；To each candidate target region extraction based on statistics and form
Feature vector and classified using random forest method, finally determine the candidate target region whether be the target area.
To achieve the above object, according to embodiments of the present invention the 3rd aspect, it is proposed that a kind of electronic equipment, including：
One or more processors；Storage device, for storing one or more programs, when one or more of programs are by described one
A or multiple processors perform so that one or more of processors realize that the present invention's is used for breast cancer Lymph Node Metastasis pathology
The processing method of image.
To achieve the above object, according to embodiments of the present invention the 4th aspect, it is proposed that a kind of computer-readable medium,
Computer program is stored thereon with, realizes that the present invention's is used for breast cancer Lymph Node Metastasis pathology when described program is executed by processor
The processing method of image.
Any one embodiment in foregoing invention realizes the automatic classification of image using the technology of deep learning, improves work
Make efficiency, reduce probability of failure.
Further effect adds hereinafter in conjunction with embodiment possessed by above-mentioned non-usual optional mode
With explanation.
Brief description of the drawings
Attached drawing is used to more fully understand the present invention, does not form inappropriate limitation of the present invention.Wherein：
Fig. 1 is the key step of the processing method for breast cancer Lymph Node Metastasis pathological image according to embodiments of the present invention
Schematic diagram；
Fig. 2 is the main modular of the processing unit for breast cancer Lymph Node Metastasis pathological image according to embodiments of the present invention
Schematic diagram；
Fig. 3 is the processing method being used for for breast cancer Lymph Node Metastasis pathological image for realizing the embodiment of the present invention
The hardware architecture diagram of electronic equipment.
Embodiment
Explain below in conjunction with attached drawing to the one exemplary embodiment of the present invention, 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
Arrive, various changes and modifications can be made to the embodiments described herein, without departing from scope and spirit of the present invention.Together
Sample, for clarity and conciseness, eliminates the description to known function and structure in following description.
The present invention is directed to propose a kind of artificial intelligence for breast cancer Lymph Node Metastasis pathological image processing method and
Device, solves the problems, such as that the subjectivity in artificial treatment of the prior art is strong, uniformity is poor, accuracy is low, accuracy is poor, tool
Have objective and fair, reproducible, accuracy is high, and accuracy is good, it is time saving and energy saving the advantages of.
Fig. 1 is the key step of the processing method for breast cancer Lymph Node Metastasis pathological image according to embodiments of the present invention
Schematic diagram.As shown in Figure 1, the processing method for breast cancer Lymph Node Metastasis pathological image of the embodiment is mainly including as follows
Step.
Step S1：Image example is obtained, then obtains the corresponding image example calibration information of the image example.Need
Bright, the processing method for breast cancer Lymph Node Metastasis pathological image of the embodiment of the present invention can be used for handling pathology figure
Picture, but can be used for the common non-pathological image of processing.I other words：Image example can be example pathological image, can also
For example normal image.
Wherein, image example is then to remove image through degree of supersaturation Threshold segmentation binary conversion treatment by original image
Simultaneously extract what the optimization afterwards of display foreground region obtained in background area.Such as：Original image is empty from RGB (R-G-B) color
Between conversion to HSV (tone-saturation degree-lightness) color space, then on the saturation degree S components in image hsv color space, adopt
Tissue part and background parts are separated by image with threshold segmentation method, obtain image segmentation binary map.Then only retain
Numerical value is higher than the parts of images of default saturation degree threshold value, that is, retains tissue regions.Finally by a series of Image erosion and figure
As the operation such as expansion, the small connected region in the corresponding image of tissue regions is removed, and fills up present in tissue regions
Duck eye.The effect of wherein burn into expansive working is the obvious maximum region found in image；Remove the effect in small unicom region
Fruit is to remove noise；It is maximum unicom region in order to obtain to fill up the effect that duck eye present in tissue regions operates.So far, it is complete
Into image optimization process, the image example after optimization processing has been obtained.
Wherein, image example calibration information is on the somewhere tissue regions in image example for record extraneous input
The details of certain lesion degree.Such as：Can be in the position of the input terminal input tumor region of computer by medical practitioner
Confidence ceases (such as irising out tumor region in the image example that display screen is shown using mouse), the pixel in these tumor regions
The coordinate information of point can be recorded with xml document form and preserved on computers, thus obtain the example of xml document form
Image information.
Step S2：According to image example and image example calibration information generation training sample set.This step is specifically wrapped
Include steps S21 and step S22.
Step S21：Image example is divided into the image example block of multiple pre-set dimensions.For example, can be an example
Image is divided into the image example block of multiple 512 × 512 pixels.
Step S22：Information is demarcated with reference to image example, the sample of each image example block is determined using characteristic point determining method
Positive and negative attribute, to obtain the training sample set comprising image example block positive sample and image example block negative sample.
Specifically：For several geometric properties points of each image example block, respectively extraction internal preset position.Such as
The position that some image example block of fruit has more than half purpose geometric properties points falls in the calibration range of image example calibration information
Portion, then it is image example block positive sample to illustrate the image example block.If some image example block has the several spies of more than half purposes
The position of sign point falls outside the calibration range of image example calibration information, then it is that image example block is born to illustrate the image example block
Sample.So as to obtain the training sample set comprising image example block positive sample and image example block negative sample.From the foregoing, it will be observed that feature
Point determining method can carry out the positive and negative property of judgement sample only by the data of a small amount of sampled point, so as to avoid to complete in image block
The traversal of portion's pixel, improves operation efficiency.To more fully understand those skilled in the art, below with example pathological image block
Exemplified by.For the example pathological image block, center, upper left, lower-left, the geometric properties of five positions of upper right and bottom right can be extracted
Point.Then judge this five points whether in the calibration cancerous region in image example calibration information respectively.If 3 to 5
A point is inside calibration cancerous region, then it is assumed that the example pathological image block is positive sample.Otherwise, the example pathological image block
For negative sample.
Step S3：Training and the parameter optimization of deep learning network are carried out according to training sample set, obtains convolutional Neural net
Network model.
Specifically, Caffe Open Frameworks can be used and calculated carrying Nvidia GTX TITAN X (Pascal) GPU
Calculated on the platform of card, and accelerate storehouse to accelerate calculating process using CUDA and cuDNN GPU.It can use
GoogleNet models, the training sample set obtained using step S2 carry out training and the parameter optimization of network.Preferably, may be used also
Will verify that false positive (False Positive) sample concentrated and obtained is placed in the negative sample of training sample concentration, net is carried out
The second training and parameter optimization of network, so as to obtain final trained network model.
Step S4：Test image is tested using convolutional neural networks model, detects target area.For example, can be with
Test pathological image is tested using convolutional neural networks model.Cancerous region in test result, that is, test image.This
Step specifically includes steps S41 to step S43.
Step S41：Test image is divided into multiple test image blocks, convolution god then is utilized to each test image block
Classify through network model, obtain the probability of happening of each test image block.Such as：Test pathological image is divided into multiple
Then image block is classified using convolutional neural networks model, obtain the corresponding canceration probability of each test pathological image block.
Step S42：Probability of happening thermodynamic chart is drawn according to all test image blocks and its corresponding probability of happening.Example
Such as：The test pathological image block of obtained known canceration probability is put into corresponding position in original test image, is drawn whole
The canceration probability thermodynamic chart opened.
Step S43：Target area is confirmed according to probability of happening thermodynamic chart.Wherein, probability of happening thermodynamic chart can be canceration
Probability thermodynamic chart, target area can be cancerous regions.Specifically, can probability of happening thermodynamic chart (such as canceration probability heat
Try hard to) in extraction the probability of happening (for example, canceration probability) be more than predetermined probabilities threshold value region, as candidate target region (example
Such as, candidate's cancerous region).Then to each candidate target region (for example, candidate cancer region) extraction based on statistics (for example,
Probable value variance in region) and form (for example, the area in region) feature vector and classified using random forest method, most
Determine whether the candidate target region (for example, candidate cancer region) is target area (for example, cancerous area) eventually, so as to obtain
Realize the detection of cancerous area.
The processing method for breast cancer Lymph Node Metastasis pathological image of the embodiment of the present invention uses the technology of deep learning
Realize the automatic classification of image, at least have the following advantages that：The uniformity that solves the classification work of artificial diagosis is poor, accuracy is low etc.
Technical problem, improves work efficiency, and reduces probability of failure.
Fig. 2 is the main modular of the processing unit for breast cancer Lymph Node Metastasis pathological image according to embodiments of the present invention
Schematic diagram.As shown in Fig. 2, the processing unit for breast cancer Lymph Node Metastasis pathological image of the embodiment includes the first processing
Module 21, Second processing module 22, the 3rd processing module 23, fourth processing module 24.
First processing module 21 is used to obtain image example, then obtains the corresponding image example calibration of the image example
Information.It should be noted that the processing unit for breast cancer Lymph Node Metastasis pathological image of the embodiment of the present invention can be used for
Pathological image is handled, but can be used for the common non-pathological image of processing.I other words：Image example can be example pathology figure
Picture, or example normal image.
Wherein, image example is then to remove image through degree of supersaturation Threshold segmentation binary conversion treatment by original image
Simultaneously extract what the optimization afterwards of display foreground region obtained in background area.Such as：Original image is empty from RGB (R-G-B) color
Between conversion to HSV (tone-saturation degree-lightness) color space, then on the saturation degree S components in image hsv color space, adopt
Tissue part and background parts are separated by image with threshold segmentation method, obtain image segmentation binary map.Then only retain
Numerical value is higher than the parts of images of default saturation degree threshold value, that is, retains tissue regions.Finally by a series of Image erosion and figure
As the operation such as expansion, the small connected region in the corresponding image of tissue regions is removed, and fills up present in tissue regions
Duck eye.The effect of wherein burn into expansive working is the obvious maximum region found in image；Remove the effect in small unicom region
Fruit is to remove noise；It is maximum unicom region in order to obtain to fill up the effect that duck eye present in tissue regions operates.So far, it is complete
Into image optimization process, the image example after optimization processing has been obtained.
Wherein, image example calibration information is on the somewhere tissue regions in image example for record extraneous input
The details of certain lesion degree.Such as：Can be in the position of the input terminal input tumor region of computer by medical practitioner
Confidence ceases (such as irising out tumor region in the image example that display screen is shown using mouse), the pixel in these tumor regions
The coordinate information of point can be recorded with xml document form and preserved on computers, thus obtain the example of xml document form
Image information.
Second processing module 22 is used for according to image example and image example calibration information generation training sample set.Tool
Body, Second processing module 22 is used for following two-step pretreatment process：
(1) image example is divided into the image example block of multiple pre-set dimensions.For example, can be an image example point
It is segmented into the image example block of multiple 512 × 512 pixels.
(2) with reference to image example calibration information, determine that the sample of each image example block is positive and negative using characteristic point determining method
Attribute, to obtain the training sample set comprising image example block positive sample and image example block negative sample.
Specifically：For several geometric properties points of each image example block, respectively extraction internal preset position.Such as
The position that some image example block of fruit has more than half purpose geometric properties points falls in the calibration range of image example calibration information
Portion, then it is image example block positive sample to illustrate the image example block.If some image example block has the several spies of more than half purposes
The position of sign point falls outside the calibration range of image example calibration information, then it is that image example block is born to illustrate the image example block
Sample.So as to obtain the training sample set comprising image example block positive sample and image example block negative sample.From the foregoing, it will be observed that feature
Point determining method can carry out the positive and negative property of judgement sample only by the data of a small amount of sampled point, so as to avoid to complete in image block
The traversal of portion's pixel, improves operation efficiency.To more fully understand those skilled in the art, below with example pathological image block
Exemplified by.For the example pathological image block, center, upper left, lower-left, the geometric properties of five positions of upper right and bottom right can be extracted
Point.Then judge this five points whether in the calibration cancerous region in image example calibration information respectively.If 3 to 5
A point is inside calibration cancerous region, then it is assumed that the example pathological image block is positive sample.Otherwise, the example pathological image block
For negative sample.
3rd processing module 23 is used for training and the parameter optimization that deep learning network is carried out according to training sample set, obtains
Convolutional neural networks model.
Specifically, Caffe Open Frameworks can be used and calculated carrying Nvidia GTX TITAN X (Pascal) GPU
Calculated on the platform of card, and accelerate storehouse to accelerate calculating process using CUDA and cuDNN GPU.It can use
GoogleNet models, the training sample set obtained using step S2 carry out training and the parameter optimization of network.Preferably, may be used also
Will verify that false positive (False Positive) sample concentrated and obtained is placed in the negative sample of training sample concentration, net is carried out
The second training and parameter optimization of network, so as to obtain final trained network model.
Fourth processing module 24 is used to test test image using convolutional neural networks model, detects target area
Domain.For example, test pathological image can be tested using convolutional neural networks model.In test result, that is, test image
Cancerous region.Specifically, fourth processing module 24 is used for following three step process process：
(1) test image is divided into multiple test image blocks, convolutional Neural net then is utilized to each test image block
Network model is classified, and obtains the probability of happening of each test image block.Such as：Test pathological image is divided into multiple images
Then block is classified using convolutional neural networks model, obtain the corresponding canceration probability of each test pathological image block.
(2) probability of happening thermodynamic chart is drawn according to all test image blocks and its corresponding probability of happening.Such as：Will
To the test pathological image block of known canceration probability be put into corresponding position in original test image, draw the canceration of whole
Probability thermodynamic chart.
(3) target area is confirmed according to probability of happening thermodynamic chart.Wherein, probability of happening thermodynamic chart can be canceration probability heat
Try hard to, target area can be cancerous region.Specifically, can be in probability of happening thermodynamic chart (such as canceration probability thermodynamic chart)
The region that the probability of happening (for example, canceration probability) is more than predetermined probabilities threshold value is extracted, as candidate target region (for example, candidate
Cancerous region).Then statistics is based on to each candidate target region (for example, candidate cancer region) extraction (for example, general in region
Rate value variance) and form (for example, the area in region) feature vector and classified using random forest method, it is final that determine should
Whether candidate target region (for example, candidate cancer region) is target area (for example, cancerous area), so as to obtain realizing cancer
The detection in disease region.
The processing unit for breast cancer Lymph Node Metastasis pathological image of the embodiment of the present invention uses the technology of deep learning
Realize the automatic classification of image, at least have the following advantages that：The uniformity that solves the classification work of artificial diagosis is poor, accuracy is low etc.
Technical problem, improves work efficiency, and reduces probability of failure.
According to an embodiment of the invention, present invention also offers a kind of electronic equipment and a kind of readable storage medium storing program for executing.
The electronic equipment of the present invention includes：At least one processor；And deposited with what at least one processor communication was connected
Reservoir；Wherein, memory storage has the instruction that can be performed by a processor, instructs and is performed by least one processor, so that
At least one processor performs the processing method provided by the present invention for breast cancer Lymph Node Metastasis pathological image.
The computer-readable recording medium of the present invention, computer-readable recording medium storage computer instruction, computer refer to
Order is used to make computer perform the processing method provided by the present invention for breast cancer Lymph Node Metastasis pathological image.
Below with reference to Fig. 3, it illustrates suitable for for realizing the structural representation of the electronic equipment 300 of the embodiment of the present application
Figure.Terminal shown in Fig. 3 is only an example, should not bring any limit to the function and use scope of the embodiment of the present application
System.
As shown in figure 3, terminal 300 includes central processing unit (CPU) 301, it can be according to being stored in read-only storage
(ROM) program in 302 or performed each from the program that storage part 308 is loaded into random access storage device (RAM) 303
Kind appropriate action and processing.In RAM 303, also it is stored with system 300 and operates required various programs and data.CPU
301st, ROM 302 and RAM 303 are connected with each other by bus 304.Input/output (I/O) interface 305 is also connected to bus
304。
I/O interfaces 305 are connected to lower component：Importation 306 including keyboard, mouse etc.；Penetrated including such as cathode
The output par, c 307 of spool (CRT), liquid crystal display (LCD) etc. and loudspeaker etc.；Storage part 308 including hard disk etc.；
And the communications portion 309 of the network interface card including LAN card, modem etc..Communications portion 309 via such as because
The network of spy's net performs communication process.Driver 310 is also according to needing to be connected to I/O interfaces 305.Detachable media 311, such as
Disk, CD, magneto-optic disk, semiconductor memory etc., are installed on driver 310, in order to read from it as needed
Computer program be mounted into as needed storage part 308.
Especially, disclosed embodiment, the process described above with reference to flow chart may be implemented as counting according to the present invention
Calculation machine software program.For example, embodiment disclosed by the invention includes a kind of computer program product, it includes being carried on computer
Computer program on computer-readable recording medium, the computer program include the program code for being used for the method shown in execution flow chart.
In such embodiment, which can be downloaded and installed by communications portion 309 from network, and/or from can
Medium 311 is dismantled to be mounted.When the computer program is performed by central processing unit (CPU) 301, the system that performs the application
The above-mentioned function of middle restriction.
It should be noted that the 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 any combination.Computer-readable recording medium for example can be --- but not
Be limited to --- electricity, magnetic, optical, electromagnetic, system, device or the device of infrared ray or semiconductor, or it is any more than combination.Meter
The more specifically 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
Take formula computer disk, hard disk, random access storage device (RAM), read-only storage (ROM), erasable type and may be programmed read-only storage
Device (EPROM or flash memory), optical fiber, portable compact disc read-only storage (CD-ROM), light storage device, magnetic memory device,
Or above-mentioned any appropriate combination.In this application, computer-readable recording medium can any include or store 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 can include in a base band or as carrier wave a part propagation data-signal,
Wherein carry computer-readable program code.The data-signal of this propagation can take various forms, 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 beyond storage medium is read, which, which can send, propagates or transmit, is used for
By instruction execution system, device either device use or program in connection.Included on computer-readable medium
Program code can be transmitted with any appropriate medium, be included but not limited to：Wirelessly, electric wire, optical cable, RF etc., or it is above-mentioned
Any appropriate combination.
Flow chart and block diagram in attached drawing, it is illustrated that according to the system of the various embodiments of the application, method and computer journey
Architectural framework in the cards, function and the operation of sequence product.At this point, each square frame in flow chart or block diagram can generation
The part of one module of table, program segment or code, a part for above-mentioned module, program segment or code include one or more
The executable instruction of logic function as defined in being used for realization.It should also be noted that some as replace realization in, institute in square frame
The function of mark can also be with different from the order marked in attached drawing generation.For example, two square frames succeedingly represented are actual
On can perform substantially in parallel, they can also be performed in the opposite order sometimes, this is depending on involved function.Also
It is noted that the combination of each square frame and block diagram in block diagram or flow chart or the square frame in flow chart, can use and perform rule
The dedicated hardware based systems of fixed functions or operations is realized, or can use the group of specialized hardware and computer instruction
Close 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 set within a processor, for example, can be described as：A kind of processor bag
Include sending module, acquisition module, determining module and first processing module.Wherein, the title of these modules is under certain conditions simultaneously
The restriction in itself to the module is not formed, for example, sending module is also described as " sending picture to the server-side connected
Obtain the module of request ".
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 computer-readable recording medium carries one or more program, this hair is realized when said one or multiple programs are executed by processor
The processing method for breast cancer Lymph Node Metastasis pathological image of bright proposition.
Above-mentioned embodiment, does not form limiting the scope of the invention.Those skilled in the art should be bright
It is white, depending on design requirement and other factors, various modifications, combination, sub-portfolio and replacement can occur.It is any
Modifications, equivalent substitutions and improvements made within the spirit and principles in the present invention etc., should be included in the scope of the present invention
Within.
Claims (10)
- A kind of 1. processing method for breast cancer Lymph Node Metastasis pathological image, it is characterised in that including：Step S1：Image example is obtained, then obtains the corresponding image example calibration information of the image example；Step S2：According to the image example and image example calibration information generation training sample set；Step S3：Training and the parameter optimization of deep learning network are carried out according to the training sample set, obtains convolutional Neural net Network model；Step S4：Test image is tested using the convolutional neural networks model, detects target area.
- 2. according to the method described in claim 1, it is characterized in that, step S2 includes：Step S21：The image example is divided into the image example block of multiple pre-set dimensions；Step S22：Information is demarcated with reference to the image example, each image example block is determined using characteristic point determining method The positive and negative attribute of sample, to obtain the training sample set comprising image example block positive sample and image example block negative sample.
- 3. according to the method described in claim 1, it is characterized in that, step S4 includes：Step S41：The test image is divided into multiple test image blocks, volume then is utilized to each test image block Product neural network model is classified, and obtains the corresponding probability of happening of each test image block；Step S42：Probability of happening thermodynamic chart is drawn according to each test image block and its corresponding probability of happening；Step S43：Target area is confirmed according to the probability of happening thermodynamic chart.
- 4. according to the method described in claim 3, it is characterized in that, step S43 includes：The region that the probability of happening is more than predetermined probabilities threshold value is extracted in the probability of happening thermodynamic chart, as candidate target area Domain；Divided to each feature vector of the candidate target region extraction based on statistics and form and using random forest method Class, finally determines whether the candidate target region is the target area.
- A kind of 5. processing unit for breast cancer Lymph Node Metastasis pathological image, it is characterised in that including：First processing module, for obtaining image example, then obtains the corresponding image example calibration information of the image example；Second processing module, for according to the image example and image example calibration information generation training sample set；3rd processing module, for carrying out training and the parameter optimization of deep learning network according to the training sample set, obtains Convolutional neural networks model；Fourth processing module, for being tested using the convolutional neural networks model test image, detects target area.
- 6. the processing unit for breast cancer Lymph Node Metastasis pathological image stated according to claim 5, it is characterised in that described Two processing modules are additionally operable to：The image example is divided into the image example block of multiple pre-set dimensions；Information is demarcated with reference to the image example, determines that the sample of each image example block is positive and negative using characteristic point determining method Attribute, to obtain the training sample set comprising image example block positive sample and image example block negative sample.
- 7. the processing unit for breast cancer Lymph Node Metastasis pathological image stated according to claim 5, it is characterised in that described Four processing modules are additionally operable to：The test image is divided into multiple test image blocks, convolutional Neural net then is utilized to each test image block Network model is classified, and obtains the corresponding probability of happening of each test image block；Probability of happening thermodynamic chart is drawn according to each test image block and its corresponding probability of happening；Target area is confirmed according to the probability of happening thermodynamic chart.
- 8. the processing unit for breast cancer Lymph Node Metastasis pathological image stated according to claim 7, it is characterised in that described Four processing modules are additionally operable to：The region that the probability of happening is more than predetermined probabilities threshold value is extracted in the probability of happening thermodynamic chart, is made For candidate target region；To each feature vector of the candidate target region extraction based on statistics and form and using random gloomy Woods method is classified, and finally determines whether the candidate target region is the target area.
- 9. a kind of electronic equipment, it is characterised in that including：One or more processors；Storage device, for storing one or more programs,When one or more of programs are performed by one or more of processors so that one or more of processors are real The now method as described in any in Claims 1-4.
- 10. a kind of computer-readable medium, is stored thereon with computer program, it is characterised in that described program is held by processor The method as described in any in Claims 1-4 is realized during row.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201711492182.9A CN107945181A (en) | 2017-12-30 | 2017-12-30 | Treating method and apparatus for breast cancer Lymph Node Metastasis pathological image |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201711492182.9A CN107945181A (en) | 2017-12-30 | 2017-12-30 | Treating method and apparatus for breast cancer Lymph Node Metastasis pathological image |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN107945181A true CN107945181A (en) | 2018-04-20 |
Family
ID=61938191
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201711492182.9A Pending CN107945181A (en) | 2017-12-30 | 2017-12-30 | Treating method and apparatus for breast cancer Lymph Node Metastasis pathological image |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN107945181A (en) |
Cited By (11)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108961296A (en) * | 2018-07-25 | 2018-12-07 | 腾讯科技（深圳）有限公司 | Eye fundus image dividing method, device, storage medium and computer equipment |
| CN109009110A (en) * | 2018-06-26 | 2018-12-18 | 东北大学 | Axillary lymphatic metastasis forecasting system based on MRI image |
| CN110211117A (en) * | 2019-05-31 | 2019-09-06 | 广东世纪晟科技有限公司 | The processing system of identification line tube and the method for Optimized Segmentation in medical image |
| CN110517243A (en) * | 2019-08-23 | 2019-11-29 | 强联智创（北京）科技有限公司 | A kind of localization method and system based on DSA image |
| CN110689525A (en) * | 2019-09-09 | 2020-01-14 | 上海中医药大学附属龙华医院 | Method and device for recognizing lymph nodes based on neural network |
| CN111091527A (en) * | 2018-10-24 | 2020-05-01 | 华中科技大学 | Method and system for automatically detecting pathological change area in pathological tissue section image |
| WO2020118618A1 (en) * | 2018-12-13 | 2020-06-18 | 深圳先进技术研究院 | Mammary gland mass image recognition method and device |
| CN111680553A (en) * | 2020-04-29 | 2020-09-18 | 北京联合大学 | Pathological image identification method and system based on depth separable convolution |
| CN111986213A (en) * | 2020-08-21 | 2020-11-24 | 四川大学华西医院 | Processing method, training method and device of slice image and storage medium |
| CN113269752A (en) * | 2021-05-27 | 2021-08-17 | 中山大学孙逸仙纪念医院 | Image detection method, device terminal equipment and storage medium |
| CN113838558A (en) * | 2021-08-16 | 2021-12-24 | 电子科技大学 | Method and device for analyzing breast cancer pathological image based on convolutional neural network |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106530290A (en) * | 2016-10-27 | 2017-03-22 | 朱育盼 | Medical image analysis method and device |
| CN107358611A (en) * | 2017-06-28 | 2017-11-17 | 南京信息工程大学 | A kind of automatic division method of panoramic scanning pathological image transport zone |
| CN107451615A (en) * | 2017-08-01 | 2017-12-08 | 广东工业大学 | Thyroid papillary carcinoma Ultrasound Image Recognition Method and system based on Faster RCNN |
| CN107527069A (en) * | 2017-08-22 | 2017-12-29 | 京东方科技集团股份有限公司 | Image processing method, device, electronic equipment and computer-readable medium |
-
2017
- 2017-12-30 CN CN201711492182.9A patent/CN107945181A/en active Pending
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106530290A (en) * | 2016-10-27 | 2017-03-22 | 朱育盼 | Medical image analysis method and device |
| CN107358611A (en) * | 2017-06-28 | 2017-11-17 | 南京信息工程大学 | A kind of automatic division method of panoramic scanning pathological image transport zone |
| CN107451615A (en) * | 2017-08-01 | 2017-12-08 | 广东工业大学 | Thyroid papillary carcinoma Ultrasound Image Recognition Method and system based on Faster RCNN |
| CN107527069A (en) * | 2017-08-22 | 2017-12-29 | 京东方科技集团股份有限公司 | Image processing method, device, electronic equipment and computer-readable medium |
Cited By (18)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109009110A (en) * | 2018-06-26 | 2018-12-18 | 东北大学 | Axillary lymphatic metastasis forecasting system based on MRI image |
| CN108961296A (en) * | 2018-07-25 | 2018-12-07 | 腾讯科技（深圳）有限公司 | Eye fundus image dividing method, device, storage medium and computer equipment |
| CN111192285A (en) * | 2018-07-25 | 2020-05-22 | 腾讯医疗健康(深圳)有限公司 | Image segmentation method, image segmentation device, storage medium and computer equipment |
| CN111192285B (en) * | 2018-07-25 | 2022-11-04 | 腾讯医疗健康(深圳)有限公司 | Image segmentation method, image segmentation device, storage medium and computer equipment |
| CN111091527A (en) * | 2018-10-24 | 2020-05-01 | 华中科技大学 | Method and system for automatically detecting pathological change area in pathological tissue section image |
| CN111091527B (en) * | 2018-10-24 | 2022-07-05 | 华中科技大学 | Method and system for automatically detecting pathological change area in pathological tissue section image |
| WO2020118618A1 (en) * | 2018-12-13 | 2020-06-18 | 深圳先进技术研究院 | Mammary gland mass image recognition method and device |
| CN110211117A (en) * | 2019-05-31 | 2019-09-06 | 广东世纪晟科技有限公司 | The processing system of identification line tube and the method for Optimized Segmentation in medical image |
| CN110517243B (en) * | 2019-08-23 | 2022-03-25 | 强联智创（北京）科技有限公司 | Positioning method and system based on DSA image |
| CN110517243A (en) * | 2019-08-23 | 2019-11-29 | 强联智创（北京）科技有限公司 | A kind of localization method and system based on DSA image |
| CN110689525A (en) * | 2019-09-09 | 2020-01-14 | 上海中医药大学附属龙华医院 | Method and device for recognizing lymph nodes based on neural network |
| CN110689525B (en) * | 2019-09-09 | 2023-10-13 | 上海中医药大学附属龙华医院 | Method and device for identifying lymph nodes based on neural network |
| CN111680553A (en) * | 2020-04-29 | 2020-09-18 | 北京联合大学 | Pathological image identification method and system based on depth separable convolution |
| CN111986213A (en) * | 2020-08-21 | 2020-11-24 | 四川大学华西医院 | Processing method, training method and device of slice image and storage medium |
| CN111986213B (en) * | 2020-08-21 | 2022-12-23 | 四川大学华西医院 | Processing method, training method and device of slice image and storage medium |
| CN113269752A (en) * | 2021-05-27 | 2021-08-17 | 中山大学孙逸仙纪念医院 | Image detection method, device terminal equipment and storage medium |
| CN113838558A (en) * | 2021-08-16 | 2021-12-24 | 电子科技大学 | Method and device for analyzing breast cancer pathological image based on convolutional neural network |
| CN113838558B (en) * | 2021-08-16 | 2023-04-18 | 电子科技大学 | Method and device for analyzing breast cancer pathological image based on convolutional neural network |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN107945181A (en) | Treating method and apparatus for breast cancer Lymph Node Metastasis pathological image | |
| US11487995B2 (en) | Method and apparatus for determining image quality | |
| CN111161275B (en) | Method and device for segmenting target object in medical image and electronic equipment | |
| JP7198577B2 (en) | Image analysis method, device, program, and method for manufacturing trained deep learning algorithm | |
| JP7058373B2 (en) | Lesion detection and positioning methods, devices, devices, and storage media for medical images | |
| KR101623431B1 (en) | Pathological diagnosis classifying apparatus for medical image and pathological diagnosis system using the same | |
| US11967069B2 (en) | Pathological section image processing method and apparatus, system, and storage medium | |
| CN111488921B (en) | Intelligent analysis system and method for panoramic digital pathological image | |
| CN107491771A (en) | Method for detecting human face and device | |
| Jia et al. | A study on automated segmentation of blood regions in wireless capsule endoscopy images using fully convolutional networks | |
| US20180053296A1 (en) | Cytologic diagnosis support apparatus, cytologic diagnosis support method, remote diagnosis support system, service providing system, and image processing method | |
| US9773185B2 (en) | Image processing apparatus, image processing method, and computer readable recording device | |
| US11967181B2 (en) | Method and device for retinal image recognition, electronic equipment, and storage medium | |
| WO2019184851A1 (en) | Image processing method and apparatus, and training method for neural network model | |
| CN112989995B (en) | Text detection method and device and electronic equipment | |
| CN111951274A (en) | Image segmentation method, system, readable storage medium and device | |
| CN110246579B (en) | Pathological diagnosis method and device | |
| CN113269737B (en) | Fundus retina artery and vein vessel diameter calculation method and system | |
| CN112966792B (en) | Blood vessel image classification processing method, device, equipment and storage medium | |
| US20230051951A1 (en) | Method for training image processing model | |
| CN110473176B (en) | Image processing method and device, fundus image processing method and electronic equipment | |
| CN110838094A (en) | Pathological section staining style conversion method and electronic equipment | |
| CN112489062B (en) | Medical image segmentation method and system based on boundary and neighborhood guidance | |
| US20220358650A1 (en) | Systems and methods to process electronic images to provide localized semantic analysis of whole slide images | |
| CN113889238B (en) | Image identification method and device, electronic equipment and storage 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 | ||
| WD01 | Invention patent application deemed withdrawn after publication | ||
| WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20180420 |