CN110390671A - A kind of method and device of Breast Calcifications detection - Google Patents
A kind of method and device of Breast Calcifications detection Download PDFInfo
- Publication number
- CN110390671A CN110390671A CN201910620175.5A CN201910620175A CN110390671A CN 110390671 A CN110390671 A CN 110390671A CN 201910620175 A CN201910620175 A CN 201910620175A CN 110390671 A CN110390671 A CN 110390671A
- Authority
- CN
- China
- Prior art keywords
- calcification
- types
- calcified regions
- breast image
- detection
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
-
- 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
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/10—Image acquisition modality
- G06T2207/10116—X-ray image
-
- 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
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Data Mining & Analysis (AREA)
- General Physics & Mathematics (AREA)
- Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- Health & Medical Sciences (AREA)
- Evolutionary Computation (AREA)
- Bioinformatics & Computational Biology (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Artificial Intelligence (AREA)
- Life Sciences & Earth Sciences (AREA)
- Evolutionary Biology (AREA)
- General Health & Medical Sciences (AREA)
- Medical Informatics (AREA)
- Nuclear Medicine, Radiotherapy & Molecular Imaging (AREA)
- Radiology & Medical Imaging (AREA)
- Quality & Reliability (AREA)
- Apparatus For Radiation Diagnosis (AREA)
Abstract
The invention discloses a kind of method and devices of Breast Calcifications detection, this method comprises: obtaining breast image；The breast image is input in N number of calcification detection model, determines the confidence level for each calcification types that the breast image is determined in N number of calcification detection model；N number of calcification detection model is to be respectively trained according to each calcification types classified in advance；N is positive integer；The confidence level for each calcification types that the breast image is determined in N number of calcification detection model determines the calcification detection result in the breast image according to the merging of each calcification types rule.
Description
Technical field
A kind of method and dress detected the present embodiments relate to machine learning techniques field more particularly to Breast Calcifications
It sets.
Background technique
Currently, the breast of the X-ray examination mankind using low dosage, breast image is obtained, by the identification to breast image,
The lesions such as various mammary tumors, tumour can be detected, facilitate early detection breast cancer, and reduce its death rate.Breast image is examined
Disconnected is a kind of effective detection method, can be used for diagnosing the relevant disease of a variety of female mammary glands.Certainly, wherein most important make
With or in the screening of breast cancer, especially early-stage breast cancer.If therefore can effectively detect various creams on breast image
Gland cancer early stage shows, and the help to doctor is huge.In general lesion is often based upon to the diagnosis of lesion on breast image
The detection of sign.Lesion sign is generally divided into calcification, lump/asymmetry and structural distortion.For same lesion, these signs
As may be simultaneously present, cause the degree of difficulty of diagnosis very big.
Summary of the invention
The embodiment of the present invention provides a kind of method and device of Breast Calcifications detection, to improve the efficiency of calcification detection,
And improve the accuracy rate of calcification detection.
A kind of method of Breast Calcifications detection provided in an embodiment of the present invention, comprising:
Obtain breast image；
The breast image is input in N number of calcification detection model, determines the breast image in N number of calcification
Detect the confidence level for each calcification types that model determines；N number of calcification detection model is according to each calcification classified in advance
What type was respectively trained；N is positive integer；
The confidence level for each calcification types that the breast image is determined in N number of calcification detection model, according to each
The merging rule of calcification types determines the calcification detection result in the breast image.
In above-mentioned technical proposal, being respectively trained for model is detected by N number of calcification, effectively increases different calcification types
Recall rate improves the possibility for diagnosing different calcification types, also, according to the merging of different calcification types rule, greatly
Improve calcification detection accuracy rate, effectively increase the accuracy of the detection of lesion sign.
A kind of possible implementation, it is described that the breast image is input in N number of calcification detection model, determine institute
State the confidence level for each calcification types that breast image is determined in N number of calcification detection model, comprising:
The breast image is input in the detection model of target area, determines the candidate calcification area of the breast image
Domain；
By the candidate corresponding breast image of calcified regions and the coordinate of the candidate calcified regions, it is input to calcification
In the fisrt feature extraction module for detecting model, the fisrt feature image of the candidate calcified regions is determined；The calcification inspection
Model is that N number of calcification detects any of model out；
The characteristic image that the fisrt feature extraction module exports is input in categorization module, determines the candidate calcification
The confidence level of the calcification types in region；The calcification types are that the calcification detects the corresponding calcification types of model.
In above-mentioned technical proposal, different calcification types, the different detection model of training are diagnosed, and then identify different calcifications
The calcified regions of type improve the possibility for diagnosing different calcification types.
A kind of possible implementation, N number of calcification detect model, including following one or more: putting blocky calcification
Detect model, cluster-shaped calcification detection model, benign calcification detection model；It is described by the breast image in N number of calcification
The confidence level for detecting each calcification types that model determines determines the mammary gland shadow according to the merging of each calcification types rule
Calcification as in detects result, comprising:
The confidence level of the calcification types is greater than to the candidate calcified regions of preset threshold, is determined as the calcification types
Calcified regions；The calcification types are a little blocky calcification types, cluster-shaped calcification types, any calcification in benign calcification types
Type；
If it is determined that the calcified regions for putting blocky calcification types are overlapped with the calcified regions of benign calcification types, then block will be put
The calcified regions of shape calcification types merge into the calcified regions of benign calcification types；
If it is determined that the calcified regions of benign calcification types are overlapped with the calcified regions of cluster-shaped calcification types, then it will be benign
The calcified regions of calcification types are incorporated into the calcified regions of the cluster-shaped calcification types.
It is point blocky calcification types, cluster-shaped calcification types, benign calcium according to the calcification types in above-mentioned technical proposal
The characteristics of changing type, it is determined that the merging mode of different calcified regions further improves the accuracy rate of calcification detection.
A kind of possible implementation, the method also includes:
According to the priority of each calcification types, the determination priority of the calcification types of the candidate calcified regions is determined；Institute
State the priority orders of each calcification types are as follows: cluster-shaped calcification types, benign calcification types put blocky calcification types.
In above-mentioned technical proposal, by the way that the priority of a calcification types is arranged, the more of candidate calcified regions are effectively reduced
Secondary calculating, reduces calculation amount, improves the detector efficiency of calcification.
Second aspect, the embodiment of the present invention provide a kind of display methods of Breast Calcifications, are applied to appoint in above-described embodiment
Method described in one, this method comprises:
Show the galactophore image obtained；
On the galactophore image, for each calcified regions in the breast image, each calcification area is shown
The calcification in domain detects result；The calcification detection result includes calcification types and the corresponding calcification area of calcification types of calcification detection
Domain.
The third aspect, the embodiment of the present invention provide a kind of device of Breast Calcifications detection, this method comprises:
Acquiring unit, for obtaining breast image；
Processing unit determines the breast image for the breast image to be input in N number of calcification detection model
In the confidence level for each calcification types that N number of calcification detection model determines；N number of calcification detection model is according in advance
What each calcification types of classification were respectively trained；N is positive integer；The breast image is determined in N number of calcification detection model
The confidence levels of each calcification types the calcification inspection in the breast image is determined according to the merging of each calcification types rule
Result out.
A kind of possible implementation, the processing unit are specifically used for:
The breast image is input in the detection model of target area, determines the candidate calcification area of the breast image
Domain；By the candidate corresponding breast image of calcified regions and the coordinate of the candidate calcified regions, it is input to calcification detection
In the fisrt feature extraction module of model, the fisrt feature image of the candidate calcified regions is determined；The calcification detects mould
Type is that N number of calcification detects any of model；The characteristic image that the fisrt feature extraction module exports is input to
In categorization module, the confidence level of the calcification types of the candidate calcified regions is determined；The calcification types are calcification detection
The corresponding calcification types of model.
A kind of possible implementation, N number of calcification detect model, including following one or more: putting blocky calcification
Detect model, cluster-shaped calcification detection model, benign calcification detection model；The processing unit, is specifically used for:
The confidence level of the calcification types is greater than to the candidate calcified regions of preset threshold, is determined as the calcification types
Calcified regions；The calcification types are a little blocky calcification types, cluster-shaped calcification types, any calcification in benign calcification types
Type；If it is determined that the calcified regions for putting blocky calcification types are overlapped with the calcified regions of benign calcification types, then it will point bulk
The calcified regions of calcification types merge into the calcified regions of benign calcification types；If it is determined that the calcified regions of benign calcification types with
The calcified regions of cluster-shaped calcification types have coincidence, then the calcified regions of benign calcification types are incorporated into the cluster-shaped calcification
The calcified regions of type.
A kind of possible implementation, the processing unit are also used to:
According to the priority of each calcification types, the determination priority of the calcification types of the candidate calcified regions is determined；Institute
State the priority orders of each calcification types are as follows: cluster-shaped calcification types, benign calcification types put blocky calcification types.
Fourth aspect, the embodiment of the present invention provide a kind of display device of Breast Calcifications, are applied to appoint in above-described embodiment
Method described in one, the device include:
Display unit, for showing the galactophore image obtained；
Processing unit, for showing institute for each calcified regions in the breast image in the galactophore image
State the calcification detection result of each calcified regions；The calcification detection result includes the calcification types and calcification types of calcification detection
Corresponding calcified regions.
5th aspect, it is described computer-readable to deposit the embodiment of the invention also provides a kind of computer readable storage medium
Storage media is stored with computer executable instructions, and the computer executable instructions are for making the computer execute above-mentioned mammary gland
The method of calcification detection.
6th aspect, the embodiment of the invention also provides a kind of calculating equipment, comprising:
Memory, for storing program instruction；
Processor executes above-mentioned mammary gland according to the program of acquisition for calling the program instruction stored in the memory
The method of calcification detection.
Detailed description of the invention
To describe the technical solutions in the embodiments of the present invention more clearly, make required in being described below to embodiment
Attached drawing is briefly introduced, it should be apparent that, drawings in the following description are only some embodiments of the invention, for this
For the those of ordinary skill in field, without creative efforts, it can also be obtained according to these attached drawings other
Attached drawing.
Fig. 1 is a kind of schematic diagram of system architecture provided in an embodiment of the present invention；
Fig. 2 is a kind of flow diagram of the method for Breast Calcifications detection provided in an embodiment of the present invention；
Fig. 3 is a kind of schematic diagram of breast image provided in an embodiment of the present invention；
Fig. 4 is a kind of display schematic diagram of Breast Calcifications provided in an embodiment of the present invention；
Fig. 5 is a kind of structural schematic diagram of Breast Calcifications detection device provided in an embodiment of the present invention；
Fig. 6 is a kind of structural schematic diagram of Breast Calcifications detection device provided in an embodiment of the present invention.
Specific embodiment
To make the objectives, technical solutions, and advantages of the present invention clearer, below in conjunction with attached drawing to the present invention make into
It is described in detail to one step, it is clear that described embodiments are only a part of the embodiments of the present invention, rather than whole implementation
Example.Based on the embodiments of the present invention, obtained by those of ordinary skill in the art without making creative efforts
All other embodiment, shall fall within the protection scope of the present invention.
The applicable system architecture of the method that Fig. 1 is detected by Breast Calcifications provided in an embodiment of the present invention.With reference to Fig. 1 institute
Show, which can be server 100, including processor 110, communication interface 120 and memory 130.
Wherein, the terminal device that communication interface 120 is applicable in for doctor communicates, and receives and dispatches the letter of terminal device transmission
Breath realizes communication.
In the embodiment of the present application, memory 130 is stored with the instruction that can be executed by least one processor 110, at least
The instruction that one processor 110 is stored by executing memory 130 can execute institute in the method for Breast Calcifications detection above-mentioned
Include the steps that.
Wherein, processor 110 is the control centre of the equipment of Breast Calcifications detection, can use various interfaces and route connects
The various pieces for connecing the equipment of Breast Calcifications detection, by running or executing the instruction and calling that are stored in memory 130
The data being stored in memory 130, to realize that Breast Calcifications detect.Optionally, processor 110 may include one or more
Processing unit, processor 110 can integrate application processor and modem processor, wherein application processor mainly handles behaviour
Make system, user interface and application program etc., modem processor mainly handles wireless communication.It is understood that above-mentioned
Modem processor can not also be integrated into processor 110.In some embodiments, processor 110 and memory 130 can
To realize on the same chip, in some embodiments, they can also be realized respectively on independent chip.
Processor 110 can be general processor, such as central processing unit (CPU), digital signal processor, dedicated integrated
Circuit (Application Specific Integrated Circuit, ASIC), field programmable gate array or other can
Perhaps transistor logic, discrete hardware components may be implemented or execute the application implementation for programmed logic device, discrete gate
Each method, step and logic diagram disclosed in example.General processor can be microprocessor or any conventional processor
Deng.The step of method in conjunction with disclosed in the embodiment of the present application, can be embodied directly in hardware processor and execute completion, Huo Zheyong
Hardware and software module combination in processor execute completion.
Memory 130 is used as a kind of non-volatile computer readable storage medium storing program for executing, can be used for storing non-volatile software journey
Sequence, non-volatile computer executable program and module.Memory 130 may include the storage medium of at least one type,
It such as may include flash memory, hard disk, multimedia card, card-type memory, random access storage device (Random Access
Memory, RAM), static random-access memory (Static Random Access Memory, SRAM), may be programmed read-only deposit
Reservoir (Programmable Read Only Memory, PROM), read-only memory (Read Only Memory, ROM), band
Electrically erasable programmable read-only memory (Electrically Erasable Programmable Read-Only Memory,
EEPROM), magnetic storage, disk, CD etc..Memory 902 can be used for carrying or storing have instruction or data
The desired program code of structure type and can by any other medium of computer access, but not limited to this.The application is real
Applying the memory 130 in example can also be circuit or other devices that arbitrarily can be realized store function, for storing program
Instruction and/or data.
It should be noted that above-mentioned structure shown in FIG. 1 is only a kind of example, it is not limited in the embodiment of the present invention.
Based on foregoing description, Fig. 2 illustratively shows a kind of stream of Breast Calcifications detection provided in an embodiment of the present invention
Journey, the device which can be detected by Breast Calcifications execute.
As shown in Fig. 2, the process specifically includes:
Step 201, breast image is obtained.
Breast image can be two dimensional image, or 3-D image, it is not limited here.Calcification includes but is not limited to
Punctate clacification, benign calcification, the calcification of the cluster village.Breast image can be x-ray image, computed tomography (Computed
Tomography, abbreviation CT) image, magnetic resonance imaging (Magnetic Resonance Imaging, abbreviation MRI) image etc.
Deng for clearer description breast image, Fig. 3 illustrates the breast image of the Breast Calcifications of a patient.
Step 202, the breast image is input in N number of calcification detection model, determines the breast image in institute
State the confidence level for each calcification types that N number of calcification detection model determines；N number of calcification detection model is according to classification in advance
Each calcification types be respectively trained；N is positive integer.
Calcification types can be distinguished according to benign and malignant degree, for example, being directed to benign calcification types, can also be divided
Are as follows: cutaneous calcification, angiosteosis, thin thick rodlike calcification, circle or punctate clacification, malnutritive calcification, suture calcification, annular or
Egg-shell calcification, hollow form calcification, punctate clacification.
The possible calcification of high malignancy can be with are as follows: particle punctate clacification, line sample or line sample branch shape calcification etc..
According to distribution mode, tufted can be divided into, section shape, section sample, area-shaped, diffuse shape etc..
Due to calcification types with the grade malignancy of calcification be not it is directly related, diagnosis there are bigger difficulty degree, because
This, can be arranged different calcification types according to needs such as the detections of diagnosis and calcification, and then different calcification inspections is respectively trained
Model out, and then can preferably learn the feature of the calcification types of mammary gland, the accuracy rate of calcification detection is improved, and then improve calcium
Change pernicious.
Wherein, to improve accuracy of identification, the interference of non-mammary region is reduced, the breast image can be input to mammary gland
In body of gland parted pattern, the body of gland region of the breast image is determined；The body of gland region of the breast image is input to N number of
Calcification detects in model.
Step 203, the confidence of each calcification types breast image determined in N number of calcification detection model
Degree determines the calcification detection result in the breast image according to the merging of each calcification types rule.
In above-mentioned technical proposal, being respectively trained for model is detected by N number of calcification, effectively increases different calcification types
Recall rate improves the possibility for diagnosing different calcification types, also, according to the merging of different calcification types rule, greatly
Improve calcification detection accuracy rate, effectively increase the accuracy of the detection of lesion sign.
A kind of possible implementation, it is described that the breast image is input in N number of calcification detection model, determine institute
State the confidence level for each calcification types that breast image is determined in N number of calcification detection model, comprising:
The breast image is input in the detection model of target area, determines the candidate calcification area of the breast image
Domain；
By the candidate corresponding breast image of calcified regions and the coordinate of the candidate calcified regions, it is input to calcification
In the fisrt feature extraction module for detecting model, the fisrt feature image of the candidate calcified regions is determined；The calcification inspection
Model is that N number of calcification detects any of model out；
The characteristic image that the fisrt feature extraction module exports is input in categorization module, determines the candidate calcification
The confidence level of the calcification types in region；The calcification types are that the calcification detects the corresponding calcification types of model.
In above-mentioned technical proposal, different calcification types, the different detection model of training are diagnosed, and then identify different calcifications
The calcified regions of type improve the possibility for diagnosing different calcification types.
In embodiments of the present invention, depth residual error network model can be respectively adopted to marked in N number of calcification detection model
What multiple breast images of calcification candidate region obtained after being trained.Therefore, breast image is being input to calcification detection mould
Before in type, it is also necessary to be trained to multiple breast images of marked calcification candidate region.
Specifically, several breast images that can be will acquire are directly as training sample, it can also be to several creams of acquisition
Gland image carries out enhancing operation, expands the data volume of training sample, enhancing operation includes but is not limited to: translating up and down at random
Set pixel (such as 0~20 pixel), Random-Rotation set angle (such as -15~15 degree), random scaling set multiple (such as
0.85~1.15 times).
It may then pass through the professionals such as doctor the calcification in training sample is marked, the content of label includes calcium
The centre coordinate of change and the diameter of calcification.Specifically, calcification can be labeled by several doctors, and is voted by more people
The mode of synthesis determines final calcification and calcification parameter, is as a result saved with the mode of mask figure.It should be noted that artificial
Mark in training sample calcification and the enhancing of training sample operation in no particular order, it can calcium in first handmarking's training sample
Change, then the training sample of calcification will be marked to carry out enhancing operation again, training sample first can also be subjected to enhancing operation, then
Manually the training sample after enhancing operation is marked.
Training sample is finally input to depth residual error network model again to be trained, by the calcification of output and preparatory label
The mask figure of training sample carry out loss function calculating, then decay using back-propagation algorithm and with momentum and ladder
Sgd optimization algorithm iterate, determine calcification detect model.
After training calcification detection model, so that it may breast image is input in calcification detection model, determine mammary gland
The candidate calcification of image.
Calcification detection model may include N number of continuous convolution feature extraction block and a full articulamentum, and N is greater than 0.It will
Breast image is handled through N number of continuous convolution feature extraction block, obtains the characteristics of image of breast image.Wherein, the convolution feature
Extracting block includes L convolution module, and L is greater than 0, as shown in figure 5, including convolutional layer, BN layers and sharp in any one convolution module
Encourage layer；It is mentioned for the continuous first convolution feature extraction block of any two and the second convolution feature in N number of convolution feature extraction block
Block is taken, the second characteristics of image of the second convolution feature extraction block output and the first image of the first convolution feature extraction block output are special
As the input of third product feature extraction block or the output of N number of continuous convolution feature extraction block after sign addition；Third convolution
Feature extraction block is to mention after the second convolution feature extraction block and with the second continuous convolution feature of convolution feature extraction block
Take block.The number of the characteristic image of first convolution feature extraction block output is less than the characteristic pattern that the first convolution feature extraction block inputs
The number of the number of picture, the characteristic image of the second convolution feature extraction block output is greater than the spy of the first convolution feature extraction block output
Levy the number of image.
Convolution feature extraction block can be the characteristic extracting module in 2D convolutional neural networks in the embodiment of the present invention, accordingly
, the convolution kernel size of the first convolution feature extraction block can be with for the convolution kernel size of m*m, the second convolution feature extraction block
For n*n；M and n can be the same or different, it is not limited here；Wherein, m, n are the integer more than or equal to 1.By setting
Different convolution kernels is set, the perception that can effectively improve feature extraction is wild, is conducive to the accuracy for improving calcification detection.?
To after the characteristics of image of breast image, so that it may the characteristics of image of breast image be inputted full articulamentum, to export mammary gland shadow
The candidate calcification of picture.
In embodiments of the present invention, can be divided into each calcification types: cluster-shaped calcification types, benign calcification types, point
It is illustrated for blocky calcification types, in step 204, a kind of possible implementation, N number of calcification detects model,
Including following one or more: putting blocky calcification detection model, cluster-shaped calcification detection model, benign calcification detection model；Institute
The confidence level for stating each calcification types for determining the breast image in N number of calcification detection model, according to each calcification class
The merging rule of type determines the calcification detection result in the breast image, comprising:
The confidence level of the calcification types is greater than to the candidate calcified regions of preset threshold, is determined as the calcification types
Calcified regions；The calcification types are a little blocky calcification types, cluster-shaped calcification types, any calcification in benign calcification types
Type；
If it is determined that the calcified regions for putting blocky calcification types are overlapped with the calcified regions of benign calcification types, then block will be put
The calcified regions of shape calcification types merge into the calcified regions of benign calcification types；
If it is determined that the calcified regions of benign calcification types are overlapped with the calcified regions of cluster-shaped calcification types, then it will be benign
The calcified regions of calcification types are incorporated into the calcified regions of the cluster-shaped calcification types.
It is point blocky calcification types, cluster-shaped calcification types, benign calcium according to the calcification types in above-mentioned technical proposal
The characteristics of changing type, it is determined that the merging mode of different calcified regions further improves the accuracy rate of calcification detection.
A kind of possible implementation, the method also includes:
According to the priority of each calcification types, the determination priority of the calcification types of the candidate calcified regions is determined, with
Determine that the breast image is input to the sequence of N number of calcification detection model；The priority orders of each calcification types are as follows:
Cluster-shaped calcification types, benign calcification types put blocky calcification types.
For example, may comprise steps of:
According to the calcified regions of determining cluster-shaped calcification types, the candidate mammary region of the breast image is screened；
The corresponding galactophore image of candidate calcified regions after screening is input to the corresponding calcification detection of benign calcification types
In model；
According to the calcified regions of determining benign calcification types, the candidate mammary region of breast image is implemented in screening；
The corresponding galactophore image of candidate calcified regions after screening is input to the corresponding calcification inspection of a bulk calcification types
Out in model, and then determine a blocky calcification.
By the above-mentioned means, by the way that the priority of a calcification types is arranged, can be effectively reduced in above-mentioned technical proposal
The multiple calculating of candidate calcified regions, can effectively reduce the calculation amount of model, improve the detector efficiency of calcification.
Certainly, different embodiments is chosen, it can be according to specific application scenarios, i.e. the identification situation choosing of breast image
It selects, to improve the detection precision and detector efficiency of calcification.
By above-mentioned detection mode, the candidate calcification of false positive can be filtered out, what is obtained is exactly final calcification, thus
The accuracy of calcification detection can be improved.
A kind of embodiment of the present invention provided in an embodiment of the present invention provides a kind of display methods of Breast Calcifications, is applied to upper
Method described in any one of embodiment is stated, this method comprises:
Show the galactophore image obtained；
On the galactophore image, for each calcified regions in the breast image, each calcification area is shown
The calcification in domain detects result；The calcification detection result includes calcification types and the corresponding calcification area of calcification types of calcification detection
Domain.
As shown in figure 3, being the schematic diagram of breast image provided in an embodiment of the present invention, as shown in figure 4, implementing for the present invention
The display schematic diagram for the Breast Calcifications that example provides, it can be seen that calcification in the breast image of Fig. 4 is cluster-shaped calcification, and root
The calcified regions of cluster-shaped calcification are determined according to model.
Based on the same technical idea, Fig. 5 illustratively shows a kind of Breast Calcifications inspection provided in an embodiment of the present invention
Device out, the device can execute the process of Breast Calcifications detection.As shown in figure 5, the device specifically includes:
Acquiring unit 401, for obtaining breast image；
Processing unit 402 determines the mammary gland for the breast image to be input in N number of calcification detection model
The confidence level for each calcification types that image is determined in N number of calcification detection model；According to N number of calcification detection model
What each calcification types classified in advance were respectively trained；N is positive integer；The breast image is detected into model in N number of calcification
The confidence level of determining each calcification types determines the calcium in the breast image according to the merging of each calcification types rule
Change detection result.
A kind of possible implementation, the processing unit 402, is specifically used for:
The breast image is input in the detection model of target area, determines the candidate calcification area of the breast image
Domain；By the candidate corresponding breast image of calcified regions and the coordinate of the candidate calcified regions, it is input to calcification detection
In the fisrt feature extraction module of model, the fisrt feature image of the candidate calcified regions is determined；The calcification detects mould
Type is that N number of calcification detects any of model；The characteristic image that the fisrt feature extraction module exports is input to
In categorization module, the confidence level of the calcification types of the candidate calcified regions is determined；The calcification types are calcification detection
The corresponding calcification types of model.
A kind of possible implementation, N number of calcification detect model, including following one or more: putting blocky calcification
Detect model, cluster-shaped calcification detection model, benign calcification detection model；The processing unit 402, is specifically used for:
The confidence level of the calcification types is greater than to the candidate calcified regions of preset threshold, is determined as the calcification types
Calcified regions；The calcification types are a little blocky calcification types, cluster-shaped calcification types, any calcification in benign calcification types
Type；If it is determined that the calcified regions for putting blocky calcification types are overlapped with the calcified regions of benign calcification types, then it will point bulk
The calcified regions of calcification types merge into the calcified regions of benign calcification types；If it is determined that the calcified regions of benign calcification types with
The calcified regions of cluster-shaped calcification types have coincidence, then the calcified regions of benign calcification types are incorporated into the cluster-shaped calcification
The calcified regions of type.
A kind of possible implementation, the processing unit 402, is also used to:
According to the priority of each calcification types, the determination priority of the calcification types of the candidate calcified regions is determined；Institute
State the priority orders of each calcification types are as follows: cluster-shaped calcification types, benign calcification types put blocky calcification types.
Based on identical inventive concept, as shown in fig. 6, the embodiment of the present invention provides a kind of display device of Breast Calcifications,
Applied to method described in any one of above-described embodiment, which includes:
Display unit 501, for showing the galactophore image obtained；
Processing unit 502, for being shown in the galactophore image for each calcified regions in the breast image
Show the calcification detection result of each calcified regions；The calcification detection result includes the calcification types and calcification of calcification detection
The corresponding calcified regions of type.
Based on the same technical idea, the embodiment of the invention also provides a kind of calculating equipment, comprising:
Memory, for storing program instruction；
Processor executes above-mentioned mammary gland according to the program of acquisition for calling the program instruction stored in the memory
The method of calcification detection.
Based on the same technical idea, the embodiment of the invention also provides a kind of computer-readable non-volatile memories to be situated between
Matter, including computer-readable instruction, when computer is read and executes the computer-readable instruction, so that computer executes
The method for stating Breast Calcifications detection.
The present invention be referring to according to the method for the embodiment of the present invention, the process of equipment (system) and computer program product
Figure and/or block diagram describe.It should be understood that every one stream in flowchart and/or the block diagram can be realized by computer program instructions
The combination of process and/or box in journey and/or box and flowchart and/or the block diagram.It can provide these computer programs
Instruct the processor of general purpose computer, special purpose computer, Embedded Processor or other programmable data processing devices to produce
A raw machine, so that being generated by the instruction that computer or the processor of other programmable data processing devices execute for real
The device for the function of being specified in present one or more flows of the flowchart and/or one or more blocks of the block diagram.
These computer program instructions, which may also be stored in, is able to guide computer or other programmable data processing devices with spy
Determine in the computer-readable memory that mode works, so that it includes referring to that instruction stored in the computer readable memory, which generates,
Enable the manufacture of device, the command device realize in one box of one or more flows of the flowchart and/or block diagram or
The function of being specified in multiple boxes.
These computer program instructions also can be loaded onto a computer or other programmable data processing device, so that counting
Series of operation steps are executed on calculation machine or other programmable devices to generate computer implemented processing, thus in computer or
The instruction executed on other programmable devices is provided for realizing in one or more flows of the flowchart and/or block diagram one
The step of function of being specified in a box or multiple boxes.
Although preferred embodiments of the present invention have been described, it is created once a person skilled in the art knows basic
Property concept, then additional changes and modifications may be made to these embodiments.So it includes excellent that the following claims are intended to be interpreted as
It selects embodiment and falls into all change and modification of the scope of the invention.
Obviously, various changes and modifications can be made to the invention without departing from essence of the invention by those skilled in the art
Mind and range.In this way, if these modifications and changes of the present invention belongs to the range of the claims in the present invention and its equivalent technologies
Within, then the present invention is also intended to include these modifications and variations.
Claims (10)
1. a kind of method of Breast Calcifications detection, which is characterized in that this method comprises:
Obtain breast image；
The breast image is input in N number of calcification detection model, determines that the breast image is detected in N number of calcification
The confidence level for each calcification types that model determines；N number of calcification detection model is according to each calcification types classified in advance
It is respectively trained；N is positive integer；
The confidence level for each calcification types that the breast image is determined in N number of calcification detection model, according to each calcification
The merging rule of type determines the calcification detection result in the breast image；The calcification detection result includes calcification inspection
The corresponding calcified regions of calcification types and calcification types out.
2. the method as described in claim 1, which is characterized in that described that the breast image is input to N number of calcification detection mould
In type, the confidence level for each calcification types that the breast image is determined in N number of calcification detection model is determined, comprising:
The breast image is input in the detection model of target area, determines the candidate calcified regions of the breast image；
By the candidate corresponding breast image of calcified regions and the coordinate of the candidate calcified regions, it is input to calcification detection
In the fisrt feature extraction module of model, the fisrt feature image of the candidate calcified regions is determined；The calcification detects mould
Type is that N number of calcification detects any of model；
The characteristic image that the fisrt feature extraction module exports is input in categorization module, determines the candidate calcified regions
Calcification types confidence level；The calcification types are that the calcification detects the corresponding calcification types of model.
3. method according to claim 2, which is characterized in that the N number of calcification detects model, including with the next item down or more
: the blocky calcification detection model of point, cluster-shaped calcification detection model, benign calcification detect model；It is described by the breast image
It is determined in the confidence level for each calcification types that N number of calcification detection model determines according to the merging of each calcification types rule
The calcification in the breast image detects result out, comprising:
The confidence level of the calcification types is greater than to the candidate calcified regions of preset threshold, is determined as the calcification of the calcification types
Region；The calcification types are a little blocky calcification types, cluster-shaped calcification types, any calcification class in benign calcification types
Type；
If it is determined that the calcified regions for putting blocky calcification types are overlapped with the calcified regions of benign calcification types, then blocky calcium will be put
The calcified regions for changing type merge into the calcified regions of benign calcification types；
If it is determined that the calcified regions of benign calcification types are overlapped with the calcified regions of cluster-shaped calcification types, then by benign calcification
The calcified regions of type are incorporated into the calcified regions of the cluster-shaped calcification types.
4. method according to claim 2, which is characterized in that the method also includes:
According to the priority of each calcification types, the determination priority of the calcification types of the candidate calcified regions is determined；It is described each
The priority orders of calcification types are as follows: cluster-shaped calcification types, benign calcification types put blocky calcification types.
5. a kind of display methods of Breast Calcifications, which is characterized in that it is applied to method described in any one of Claims 1-4,
This method comprises:
Show the galactophore image obtained；
On the galactophore image, for each calcified regions in the breast image, each calcified regions are shown
Calcification detects result；The calcification detection result includes the calcification types and the corresponding calcified regions of calcification types of calcification detection.
6. a kind of device of Breast Calcifications detection, which is characterized in that this method comprises:
Acquiring unit, for obtaining breast image；
Processing unit determines the breast image in institute for the breast image to be input in N number of calcification detection model
State the confidence level for each calcification types that N number of calcification detection model determines；N number of calcification detection model is according to classification in advance
Each calcification types be respectively trained；N is positive integer；By the breast image in each of N number of calcification detection model determination
The confidence level of a calcification types determines the calcification detection knot in the breast image according to the merging of each calcification types rule
Fruit；The calcification detection result includes the calcification types and the corresponding calcified regions of calcification types of calcification detection.
7. device as claimed in claim 6, which is characterized in that the processing unit is specifically used for:
The breast image is input in the detection model of target area, determines the candidate calcified regions of the breast image；It will
The coordinate of the corresponding breast image of candidate's calcified regions and the candidate calcified regions, is input to calcification detection model
In fisrt feature extraction module, the fisrt feature image of the candidate calcified regions is determined；The calcification detection model is institute
State any of N number of calcification detection model；The characteristic image that the fisrt feature extraction module exports is input to classification mould
In block, the confidence level of the calcification types of the candidate calcified regions is determined；The calcification types are that the calcification detects model pair
The calcification types answered.
8. device as claimed in claim 7, which is characterized in that the N number of calcification detects model, including with the next item down or more
: the blocky calcification detection model of point, cluster-shaped calcification detection model, benign calcification detect model；The processing unit, it is specific to use
In:
The confidence level of the calcification types is greater than to the candidate calcified regions of preset threshold, is determined as the calcification of the calcification types
Region；The calcification types are a little blocky calcification types, cluster-shaped calcification types, any calcification class in benign calcification types
Type；If it is determined that the calcified regions for putting blocky calcification types are overlapped with the calcified regions of benign calcification types, then blocky calcium will be put
The calcified regions for changing type merge into the calcified regions of benign calcification types；If it is determined that calcified regions and the group of benign calcification types
The calcified regions of Clustered Calcifications type have coincidence, then the calcified regions of benign calcification types are incorporated into the cluster-shaped calcification class
The calcified regions of type.
9. a kind of computer readable storage medium, which is characterized in that the computer-readable recording medium storage has computer can
It executes instruction, the computer executable instructions are for executing the computer as in Claims 1-4 or claim 5
Described in any item methods.
10. a kind of calculating equipment characterized by comprising
Memory, for storing program instruction；
Processor, for calling the program instruction stored in the memory, according to acquisition program execute as claim 1 to
Method described in any one of 4 or claim 5.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910620175.5A CN110390671B (en) | 2019-07-10 | 2019-07-10 | Method and device for detecting mammary gland calcification |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910620175.5A CN110390671B (en) | 2019-07-10 | 2019-07-10 | Method and device for detecting mammary gland calcification |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN110390671A true CN110390671A (en) | 2019-10-29 |
| CN110390671B CN110390671B (en) | 2021-11-30 |
Family
ID=68286446
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910620175.5A Active CN110390671B (en) | 2019-07-10 | 2019-07-10 | Method and device for detecting mammary gland calcification |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110390671B (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111986165A (en) * | 2020-07-31 | 2020-11-24 | 上海依智医疗技术有限公司 | Method and device for detecting calcification in breast image |
Citations (13)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20090297002A1 (en) * | 2005-01-04 | 2009-12-03 | Zhang Daoxian H | Computer aided detection of microcalcification clusters |
| CN104637044A (en) * | 2013-11-07 | 2015-05-20 | 中国科学院深圳先进技术研究院 | Ultrasonic image extracting system for calcified plaque and sound shadow thereof |
| CN104715254A (en) * | 2015-03-17 | 2015-06-17 | 东南大学 | Ordinary object recognizing method based on 2D and 3D SIFT feature fusion |
| CN104732242A (en) * | 2015-04-08 | 2015-06-24 | 苏州大学 | Multi-classifier construction method and system |
| CN104794708A (en) * | 2015-04-10 | 2015-07-22 | 浙江工业大学 | Atherosclerosis plaque composition dividing method based on multi-feature learning |
| CN105135591A (en) * | 2015-07-01 | 2015-12-09 | 西安理工大学 | Train air conditioning unit fault diagnosing method based on multi-classification strategy |
| CN105913086A (en) * | 2016-04-12 | 2016-08-31 | 福州大学 | Computer-aided mammary gland diagnosing method by means of characteristic weight adaptive selection |
| CN107330449A (en) * | 2017-06-13 | 2017-11-07 | 瑞达昇科技(大连)有限公司 | A kind of BDR sign detection method and device |
| CN108960322A (en) * | 2018-07-02 | 2018-12-07 | 河南科技大学 | A kind of coronary calcification patch automatic testing method based on cardiac CT image |
| CN108961207A (en) * | 2018-05-02 | 2018-12-07 | 上海大学 | Lymph node Malignant and benign lesions aided diagnosis method based on multi-modal ultrasound image |
| CN109308488A (en) * | 2018-08-30 | 2019-02-05 | 深圳大学 | Breast ultrasound image processing apparatus, method, computer equipment and storage medium |
| CN109363698A (en) * | 2018-10-16 | 2019-02-22 | 杭州依图医疗技术有限公司 | A kind of method and device of breast image sign identification |
| CN109949271A (en) * | 2019-02-14 | 2019-06-28 | 腾讯科技（深圳）有限公司 | A kind of detection method based on medical image, the method and device of model training |
-
2019
- 2019-07-10 CN CN201910620175.5A patent/CN110390671B/en active Active
Patent Citations (13)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20090297002A1 (en) * | 2005-01-04 | 2009-12-03 | Zhang Daoxian H | Computer aided detection of microcalcification clusters |
| CN104637044A (en) * | 2013-11-07 | 2015-05-20 | 中国科学院深圳先进技术研究院 | Ultrasonic image extracting system for calcified plaque and sound shadow thereof |
| CN104715254A (en) * | 2015-03-17 | 2015-06-17 | 东南大学 | Ordinary object recognizing method based on 2D and 3D SIFT feature fusion |
| CN104732242A (en) * | 2015-04-08 | 2015-06-24 | 苏州大学 | Multi-classifier construction method and system |
| CN104794708A (en) * | 2015-04-10 | 2015-07-22 | 浙江工业大学 | Atherosclerosis plaque composition dividing method based on multi-feature learning |
| CN105135591A (en) * | 2015-07-01 | 2015-12-09 | 西安理工大学 | Train air conditioning unit fault diagnosing method based on multi-classification strategy |
| CN105913086A (en) * | 2016-04-12 | 2016-08-31 | 福州大学 | Computer-aided mammary gland diagnosing method by means of characteristic weight adaptive selection |
| CN107330449A (en) * | 2017-06-13 | 2017-11-07 | 瑞达昇科技(大连)有限公司 | A kind of BDR sign detection method and device |
| CN108961207A (en) * | 2018-05-02 | 2018-12-07 | 上海大学 | Lymph node Malignant and benign lesions aided diagnosis method based on multi-modal ultrasound image |
| CN108960322A (en) * | 2018-07-02 | 2018-12-07 | 河南科技大学 | A kind of coronary calcification patch automatic testing method based on cardiac CT image |
| CN109308488A (en) * | 2018-08-30 | 2019-02-05 | 深圳大学 | Breast ultrasound image processing apparatus, method, computer equipment and storage medium |
| CN109363698A (en) * | 2018-10-16 | 2019-02-22 | 杭州依图医疗技术有限公司 | A kind of method and device of breast image sign identification |
| CN109949271A (en) * | 2019-02-14 | 2019-06-28 | 腾讯科技（深圳）有限公司 | A kind of detection method based on medical image, the method and device of model training |
Non-Patent Citations (2)
| Title |
|---|
| PENGCHENG XI 等: "Abnormality Detection in Mammography using Deep Convolutional Neural Networks", 《HTTPS://ARXIV.ORG/PDF/1803.01906.PDF》 * |
| 叶淑玲: "高频超声对乳腺肿块内微小钙化灶的诊断与临床意义", 《中国实用医药》 * |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111986165A (en) * | 2020-07-31 | 2020-11-24 | 上海依智医疗技术有限公司 | Method and device for detecting calcification in breast image |
| CN111986165B (en) * | 2020-07-31 | 2024-04-09 | 北京深睿博联科技有限责任公司 | Calcification detection method and device in breast image |
Also Published As
| Publication number | Publication date |
|---|---|
| CN110390671B (en) | 2021-11-30 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Abdelrahman et al. | Convolutional neural networks for breast cancer detection in mammography: A survey | |
| US10867384B2 (en) | System and method for automatically detecting a target object from a 3D image | |
| Duran et al. | ProstAttention-Net: A deep attention model for prostate cancer segmentation by aggressiveness in MRI scans | |
| CN111428709B (en) | Image processing method, device, computer equipment and storage medium | |
| Oliver et al. | A review of automatic mass detection and segmentation in mammographic images | |
| CN109447065A (en) | A kind of method and device of breast image identification | |
| Jony et al. | Detection of lung cancer from CT scan images using GLCM and SVM | |
| CN109363698A (en) | A kind of method and device of breast image sign identification | |
| George et al. | Computer assisted recognition of breast cancer in biopsy images via fusion of nucleus-guided deep convolutional features | |
| KR102320431B1 (en) | medical image based tumor detection and diagnostic device | |
| CN109919254B (en) | Breast density classification method, system, readable storage medium and computer device | |
| Ström et al. | Pathologist-level grading of prostate biopsies with artificial intelligence | |
| Gheonea et al. | Diagnosis system for hepatocellular carcinoma based on fractal dimension of morphometric elements integrated in an artificial neural network | |
| Kumar et al. | Mammogram image segmentation using SUSAN corner detection | |
| CN113011257A (en) | Breast cancer immunohistochemical artificial intelligence interpretation method | |
| Alsheh Ali et al. | Detection of potential microcalcification clusters using multivendor for‐presentation digital mammograms for short‐term breast cancer risk estimation | |
| Hassan et al. | A dilated residual hierarchically fashioned segmentation framework for extracting gleason tissues and grading prostate cancer from whole slide images | |
| Banerjee et al. | A CADe system for gliomas in brain MRI using convolutional neural networks | |
| Chen et al. | A deep learning model based on dynamic contrast-enhanced magnetic resonance imaging enables accurate prediction of benign and malignant breast lessons | |
| EP4081952A1 (en) | Systems and methods for analyzing two-dimensional and three-dimensional image data | |
| CN110390671A (en) | A kind of method and device of Breast Calcifications detection | |
| Kr Ghosh et al. | Development of intuitionistic fuzzy special embedded convolutional neural network for mammography enhancement | |
| Mercan et al. | Automated Diagnosis of Breast Cancer and Pre-invasive Lesions on Digital Whole Slide Images. | |
| Mahmoudi et al. | Differentiation between pancreatic ductal adenocarcinoma and normal pancreatic tissue for treatment response assessment using multi-scale texture analysis of CT images | |
| CN109523547A (en) | A kind of method and device of image tubercle detection |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |