CN108830282A - A kind of the breast lump information extraction and classification method of breast X-ray image - Google Patents
A kind of the breast lump information extraction and classification method of breast X-ray image Download PDFInfo
- Publication number
- CN108830282A CN108830282A CN201810530455.2A CN201810530455A CN108830282A CN 108830282 A CN108830282 A CN 108830282A CN 201810530455 A CN201810530455 A CN 201810530455A CN 108830282 A CN108830282 A CN 108830282A
- Authority
- CN
- China
- Prior art keywords
- breast
- classification
- task
- lump
- ray image
- 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
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/243—Classification techniques relating to the number of classes
- G06F18/2431—Multiple classes
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/048—Activation functions
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/40—Extraction of image or video features
- G06V10/44—Local feature extraction by analysis of parts of the pattern, e.g. by detecting edges, contours, loops, corners, strokes or intersections; Connectivity analysis, e.g. of connected components
-
- G—PHYSICS
- G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
- G16H—HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA
- G16H30/00—ICT specially adapted for the handling or processing of medical images
- G16H30/20—ICT specially adapted for the handling or processing of medical images for handling medical images, e.g. DICOM, HL7 or PACS
-
- G—PHYSICS
- G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
- G16H—HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA
- G16H50/00—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics
- G16H50/20—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics for computer-aided diagnosis, e.g. based on medical expert systems
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V2201/00—Indexing scheme relating to image or video recognition or understanding
- G06V2201/03—Recognition of patterns in medical or anatomical images
- G06V2201/032—Recognition of patterns in medical or anatomical images of protuberances, polyps nodules, etc.
Abstract
The invention discloses a kind of breast lump information extraction of breast X-ray image and classification methods, include the following steps：S1, breast X-ray image is inputted into four parallel-convolution neural networks；S2, the high-level semantics features that breast X-ray image is extracted based on four parallel-convolution neural networks；S3, multi-tag multi-task learning network training is carried out to the high-level semantics features of extraction, and obtains the classification information of breast lump.The breast lump information extraction and classification method of breast X-ray image provided by the invention, the redundancy feature of convolutional neural networks extraction can be effectively removed, four classification tasks can to tangle mutually by multi-tag multitask network to constrain and mutually promote, clear breast lump classification information is provided for doctor, the auxiliary diagnosis of breast lump related disease is provided.
Description
Technical field
The invention belongs to breast X-ray technical field of image processing, and in particular to a kind of breast lump letter of breast X-ray image
Breath extracts and classification method.
Background technique
Breast cancer is one of the most common malignant tumors in women, shows just to have a women in every two minutes according to related data
It is diagnosed with breast cancer.In China, breast cancer is averagely every 12 minutes the main reason for leading to 40~50 years old women die
Just there is a women to die of breast cancer；In Urban Women, every about just thering are 3~4 people to suffer from breast cancer in 10,000 people, not only such as
This, the disease incidence of breast cancer is just steeply risen with annual 3%~4% speed.Therefore, early detection, early diagnosis and early stage
Treatment is acknowledged as the effective measures of control breast cancer development, and has to the survival rate and quality of life that improve patient important
Meaning, wherein early detection top the list the position wanted.
Breast lump is the sings and symptoms that breast cancer must have, and generally individually, form is indefinite, and matter is harder tough.In recent years,
In order to improve the diagnosis efficiency of breast cancer, breast lump develops towards intelligent direction always in early stage research, with present
It continues to develop scientific and technically, artificial intelligence technology is constantly progressive, and artificial neural network technology is also increasingly mature, its classification energy
Power is also increasingly stronger, and has intelligence, provides a kind of new method for tumor of breast identification.
Convolutional neural networks (Convolutional Neural Network, CNN) are a kind of feedforward neural networks, it
Artificial neuron can respond the surrounding cells in a part of coverage area, have outstanding performance for large-scale image procossing；It is wrapped
Include convolutional layer (convolutional layer) and pond layer (pooling layer).
At present in medical image analysis field, convolutional neural networks are mainly used for the realization of single learning tasks.Such as it is right
Simple lump information extraction in breast X-ray image carries out simple benign from malignant tumors judgement, is difficult the medical diagnosis with doctor
Authority combines, it is difficult to relatively reliable data foundation is provided during the diagnosis of doctor.
Summary of the invention
For above-mentioned deficiency in the prior art, the breast lump information extraction of breast X-ray image provided by the invention and
Classification method solves when being analyzed in the prior art using convolutional neural networks breast X-ray image, can only carry out simple
The good pernicious information extraction of lump, it is difficult to for doctor provide accurately diagnosis authority the problem of.
In order to achieve the above object of the invention, the technical solution adopted by the present invention is：
A kind of the breast lump information extraction and classification method of breast X-ray image, include the following steps：
S1, breast X-ray image is inputted into four parallel-convolution neural networks；
S2, the high-level semantics features that breast X-ray image is extracted based on four parallel-convolution neural networks；
S3, multi-tag multi-task learning network training is carried out to the high-level semantics features of extraction, and obtains breast lump
Classification information.
Further, in the step S2：
Each parallel-convolution neural network includes two convolutional layers, two maximum pond layers and a full connection
Layer；Network-order is to be followed successively by convolutional layer, pond layer, convolutional layer, pond layer and full articulamentum；
Each convolutional layer includes 65 × 5 convolution kernels, and each pond layer includes 12 × 2 core, described
The size of full articulamentum is 192 × 1；
Each convolutional neural networks extract the feature of 192 breast X-ray images, by each convolutional Neural net
192 character representations that network extracts are one group of high-level semantics features vector, respectively X1,X2,X3And X4。
Further, the building method of the multi-tag multitask network is specially：
By high-level semantics features vector X described in 4 groups1,X2,X3And X4, respectively as in multi-tag multi-task learning for obtaining
Must classify discreet value 4 excitation functions input value, the relationship between discreet value and physical tags that will classify is as multi-tag
The objective function of multi-task learning carries out objective function to the objective function minimizing, and using stochastic gradient descent method
Network training obtains the value and deviant of all kernel functions of multi-tag multitask network, realizes multi-tag multitask network
Build；
The excitation function is softmax function or ReLU function.
Further, in the step S3：
Classification task in multi-tag multitask network includes the classification task of breast lump shape, breast lump edge
The classification task of classification task, the classification task of breast lump BI-BADS classification and Diagnosis of Breast tumors；
In the classification task of the breast lump shape, the label of classification is：
Y1=[- 7, -6, -5, -4, -3, -2, -1,1,2,3,4,5,6,7]
Respectively correspond 14 kinds of shapes of breast lump；
In the classification task at the breast lump edge, the label of classification is：
Y2=[- 9, -8, -7, -6, -5, -4, -3, -2, -1,1,2,3,4,5,6,7,8,9]
Respectively correspond 18 kinds of different edges of breast lump；
In the classification task of breast lump BI-BADS classification, the label of classification is：
Y3=[0,1,2,3,4,5,6]
Respectively correspond 7 kind different score values of the breast lump in BI-BADS standard；
In the good pernicious classification task of the breast lump, tag along sort is：
Y4=[1,2,3]
Respectively correspond benign, the pernicious and suspicious three classes undetermined of breast lump.
It further, include total losses function L (W, b) and j-th in the objective function of the multi-tag multitask network
The loss function L of taskI, j(W,b)；
Total losses function L (W, b) is：
In formula, LjFor the loss function of j-th of task, λjFor the loss weight of the loss function of j-th of task, wherein j
=1,2,3,4；Respectively lump Shape Classification task, mass edge classification task, lump BI-BADS classify task and swollen
The good pernicious classification task of block；
The loss function L of j-th of taskI, j(W, b) is defined as：
Wherein, N is the total number of training sample；
I indicates i-th of sample；
Indicate the output for being predicted as k-th of label value in j-th of task after the full articulamentum of convolutional neural networks
Value；
CjIndicate the number of tags of j-th of task.
Further, in the total losses function L, the loss weight λ of the loss function of j-th of taskjRespectivelyWith 1；
The loss function L of j-th of taskI, jIn, the number of tags C of j-th of taskjRespectively 14,18,7,3.
Beneficial effects of the present invention are：The breast lump information extraction of breast X-ray image provided by the invention and classification side
Method can effectively remove the redundancy feature of convolutional neural networks extraction, can make four points by multi-tag multitask network
Generic task, which tangles mutually, to constrain and mutually promotes, and provides clear breast lump classification information for doctor, provides breast lump phase
The auxiliary diagnosis of related disorders.
Detailed description of the invention
Fig. 1 is that the breast lump information extraction of breast X-ray image and classification method are realized in embodiment provided by the invention
Flow chart.
Specific embodiment
A specific embodiment of the invention is described below, in order to facilitate understanding by those skilled in the art this hair
It is bright, it should be apparent that the present invention is not limited to the ranges of specific embodiment, for those skilled in the art,
As long as various change is in the spirit and scope of the present invention that the attached claims limit and determine, these variations are aobvious and easy
See, all are using the innovation and creation of present inventive concept in the column of protection.
As shown in Figure 1, the breast lump information extraction and classification method of a kind of breast X-ray image, include the following steps：
S1, breast X-ray image is inputted into four parallel-convolution neural networks；
S2, the high-level semantics features that breast X-ray image is extracted based on four parallel-convolution neural networks；
In above-mentioned steps S2：
Each parallel-convolution neural network includes two convolutional layers, two maximum pond layers and a full connection
Layer；Network-order is to be followed successively by convolutional layer, pond layer, convolutional layer, pond layer and full articulamentum；
Each convolutional layer includes 65 × 5 convolution kernels, and each pond layer includes 12 × 2 core, described
The size of full articulamentum is 192 × 1；
Each convolutional neural networks extract the feature of 192 breast X-ray images, by each convolutional Neural net
192 character representations that network extracts are one group of high-level semantics features vector, respectively X1,X2,X3And X4。
S3, multi-tag multi-task learning network training is carried out to the high-level semantics features of extraction, and obtains breast lump
Classification information.
The building method of above-mentioned multi-tag multitask network is specially：
Swash high-level semantics features described in 4 groups as 4 for being used to obtain classification discreet value in multi-tag multi-task learning
The input value for encouraging function, the relationship between discreet value and physical tags that will classify is as the target letter of multi-tag multi-task learning
Number carries out network training to objective function to the objective function minimizing, and using stochastic gradient descent method, obtains multi-tag
The value and deviant of all kernel functions of multitask network realize building for multi-tag multitask network；Excitation function is
Softmax function or ReLU function.
Classification task in above-mentioned multi-tag multitask network includes the classification task of breast lump shape, breast lump side
Classification task, the classification task of breast lump BI-BADS classification and the classification task of Diagnosis of Breast tumors of edge；Four component
The input of excitation function in generic task respectively corresponds 4 groups of high-level semantics features vectors X1, X2, X3 of convolutional neural networks extraction
And X4, i.e. the high-level semantics features that the different training objective of four classes causes each convolutional neural networks to extract are that have significant difference
's.
In the classification task of the breast lump shape, the label of classification is：
Y1=[- 7, -6, -5, -4, -3, -2, -1,1,2,3,4,5,6,7]
Respectively correspond 14 kinds of shapes of breast lump；
In the classification task at the breast lump edge, the label of classification is：
Y2=[- 9, -8, -7, -6, -5, -4, -3, -2, -1,1,2,3,4,5,6,7,8,9]
Respectively correspond 18 kinds of different edges of breast lump；
In the classification task of breast lump BI-BADS classification, the label of classification is：
Y3=[0,1,2,3,4,5,6]
Respectively correspond 7 kind different score values of the breast lump in BI-BADS standard；
In the good pernicious classification task of the breast lump, tag along sort is：
Y4=[1,2,3]
Respectively correspond benign, the pernicious and suspicious three classes undetermined of breast lump.
The objective function of the network of multi-tag multitask include total losses function L (W, b) be and j-th of task loss
Function LI, j(W,b)；
Total losses function L (W, b) is：
In formula, LjFor the loss function of j-th of task；
λjFor the loss weight of the loss function of j-th of task, wherein j=1,2,3,4；Respectively lump Shape Classification
Task, mass edge classification task, lump BI-BADS classification task and the good pernicious classification task of lump；Lose weight λj
RespectivelyWith 1；
The loss function L of j-th of taskI, j(W, b) is defined as：
Wherein, N is the total number of training sample；
I indicates i-th of training sample；
Indicate the output for being predicted as k-th of label value in j-th of task after the full articulamentum of convolutional neural networks
Value；
CjFor the number of tags of j-th of task, respectively 14,18,7,3.
In one embodiment of the invention, in the build process of above-mentioned multi-tag multitask network, boarding steps are utilized
Spending the process that descent method carries out network training to objective function is specially：
Total losses function L (W, b) is sought respectively be about the gradient of kernel function value WThe gradient of deviant b
For
According to the gradient of W, the iteration more new formula for obtaining W is：
The iteration of b more new formula is：
In formula, ρ is step-length, may be configured as the arbitrary value between 0-0.1；
Update is iterated to W and b respectively using formula (3) and formula (4), when reaching pre-set decision condition
When, iteration stopping completes multi-tag multitask network to obtain all kernel function value W and deviant b of multi-tag multitask
Build.Stochastic gradient descent training is to carry out gradient calculating with identical sample, therefore the speed of network training is non-every time
It is often fast.
The breast lump information extraction and classification method of breast X-ray image provided by the invention, can effectively remove convolution
The redundancy feature that neural network is extracted can make four classification tasks tangle constraint mutually simultaneously by multi-tag multitask network
It mutually promotes, provides clear breast lump classification information for doctor, the auxiliary diagnosis of breast lump related disease is provided.
Claims (6)
1. the breast lump information extraction and classification method of a kind of breast X-ray image, which is characterized in that include the following steps：
S1, breast X-ray image is inputted into four parallel-convolution neural networks；
S2, the high-level semantics features that breast X-ray image is extracted based on four parallel-convolution neural networks；
S3, multi-tag multi-task learning network training is carried out to the high-level semantics features of extraction, and obtains the classification of breast lump
Information.
2. the breast lump information extraction and classification method of breast X-ray image according to claim 1, which is characterized in that
In the step S2：
Each parallel-convolution neural network includes two convolutional layers, two maximum pond layers and a full articulamentum；Net
Network sequence is to be followed successively by convolutional layer, pond layer, convolutional layer, pond layer and full articulamentum；
Each convolutional layer includes 65 × 5 convolution kernels, and each pond layer includes 12 × 2 core, described to connect entirely
The size for connecing layer is 192 × 1；
Each convolutional neural networks extract the feature of 192 breast X-ray images, and each convolutional neural networks are mentioned
192 character representations taken are one group of high-level semantics features vector, respectively X1,X2,X3And X4。
3. the breast lump information extraction and classification method of breast X-ray image according to claim 2, which is characterized in that
The building method of the multi-tag multitask network is specially：
By high-level semantics features vector X described in 4 groups1,X2,X3And X4, respectively as in multi-tag multi-task learning for being divided
The input value of 4 excitation functions of class discreet value, the relationship between discreet value and physical tags that will classify is as multi-tag more
The objective function of business study carries out network to objective function to the objective function minimizing, and using stochastic gradient descent method
Training obtains the value and deviant of all kernel functions of multi-tag multitask network, realizes taking for multi-tag multitask network
It builds；
The excitation function is softmax function or ReLU function.
4. the breast lump information extraction and classification method of breast X-ray image according to claim 3, which is characterized in that
In the step S3：
Classification task in multi-tag multitask network includes the classification of the classification task of breast lump shape, breast lump edge
The classification task of task, the classification task of breast lump BI-BADS classification and Diagnosis of Breast tumors；
In the classification task of the breast lump shape, the label of classification is：
Y1=[- 7, -6, -5, -4, -3, -2, -1,1,2,3,4,5,6,7]
Respectively correspond 14 kinds of shapes of breast lump；
In the classification task at the breast lump edge, the label of classification is：
Y2=[- 9, -8, -7, -6, -5, -4, -3, -2, -1,1,2,3,4,5,6,7,8,9]
Respectively correspond 18 kinds of different edges of breast lump；
In the classification task of breast lump BI-BADS classification, the label of classification is：
Y3=[0,1,2,3,4,5,6]
Respectively correspond 7 kind different score values of the breast lump in BI-BADS standard；
In the good pernicious classification task of the breast lump, tag along sort is：
Y4=[1,2,3]
Respectively correspond benign, the pernicious and suspicious three classes undetermined of breast lump.
5. the breast lump information extraction and classification method of breast X-ray image according to claim 3, which is characterized in that
In the loss function L that the objective function of the multi-tag multitask network includes total losses function L (W, b) and j-th of taskI, j
(W,b)；
Total losses function L (W, b) is：
In formula, LjFor the loss function of j-th of task, λjFor the loss weight of the loss function of j-th of task, wherein j=1,
2,3,4；Respectively lump Shape Classification task, mass edge classification task, lump BI-BADS classification task and lump are good
Pernicious classification task；
The loss function L of j-th of taskI, j(W, b) is defined as：
Wherein, N is the total number of training sample；
I indicates i-th of sample；
xykIndicate the output valve for being predicted as k-th of label value in j-th of task after the full articulamentum of convolutional neural networks；
CjIndicate the number of tags of j-th of task.
6. the breast lump information extraction and classification method of breast X-ray image according to claim 5, which is characterized in that
In the total losses function L, the loss weight λ of the loss function of j-th of taskjRespectivelyWith 1；
The loss function L of j-th of taskI, jIn, the number of tags C of j-th of taskjRespectively 14,18,7,3.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810530455.2A CN108830282A (en) | 2018-05-29 | 2018-05-29 | A kind of the breast lump information extraction and classification method of breast X-ray image |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810530455.2A CN108830282A (en) | 2018-05-29 | 2018-05-29 | A kind of the breast lump information extraction and classification method of breast X-ray image |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN108830282A true CN108830282A (en) | 2018-11-16 |
Family
ID=64146109
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810530455.2A Pending CN108830282A (en) | 2018-05-29 | 2018-05-29 | A kind of the breast lump information extraction and classification method of breast X-ray image |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN108830282A (en) |
Cited By (11)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109886143A (en) * | 2019-01-29 | 2019-06-14 | 上海鹰瞳医疗科技有限公司 | Multi-tag disaggregated model training method and equipment |
| CN110232396A (en) * | 2019-04-09 | 2019-09-13 | 贵州大学 | X-ray breast image deep learning classification method |
| CN110414539A (en) * | 2019-08-05 | 2019-11-05 | 腾讯科技（深圳）有限公司 | A kind of method and relevant apparatus for extracting characterization information |
| CN110414562A (en) * | 2019-06-26 | 2019-11-05 | 平安科技（深圳）有限公司 | Classification method, device, terminal and the storage medium of X-ray |
| CN110428007A (en) * | 2019-08-01 | 2019-11-08 | 科大讯飞(苏州)科技有限公司 | X-ray image object detection method, device and equipment |
| CN110599476A (en) * | 2019-09-12 | 2019-12-20 | 腾讯科技（深圳）有限公司 | Disease grading method, device, equipment and medium based on machine learning |
| CN110728674A (en) * | 2019-10-21 | 2020-01-24 | 清华大学 | Image processing method and device, electronic equipment and computer readable storage medium |
| CN110738231A (en) * | 2019-07-25 | 2020-01-31 | 太原理工大学 | Method for classifying mammary gland X-ray images by improving S-DNet neural network model |
| CN110827296A (en) * | 2019-11-01 | 2020-02-21 | 南京信息工程大学 | Mammary gland X-ray image analysis method of multi-target integrated deep neural network |
| CN111709950A (en) * | 2020-08-20 | 2020-09-25 | 成都金盘电子科大多媒体技术有限公司 | Mammary gland molybdenum target AI auxiliary screening method |
| CN112614096A (en) * | 2020-12-17 | 2021-04-06 | 北京医准智能科技有限公司 | Ordinal number regression-based breast molybdenum target lesion benign and malignant analysis method |
Citations (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103425986A (en) * | 2013-08-31 | 2013-12-04 | 西安电子科技大学 | Breast lump image feature extraction method based on edge neighborhood weighing |
| CN105975915A (en) * | 2016-04-28 | 2016-09-28 | 大连理工大学 | Front vehicle parameter identification method based on multitask convolution nerve network |
| CN106326931A (en) * | 2016-08-25 | 2017-01-11 | 南京信息工程大学 | Mammary gland molybdenum target image automatic classification method based on deep learning |
| CN106339591A (en) * | 2016-08-25 | 2017-01-18 | 汤平 | Breast cancer prevention self-service health cloud service system based on deep convolutional neural network |
| CN106997475A (en) * | 2017-02-24 | 2017-08-01 | 中国科学院合肥物质科学研究院 | A kind of insect image-recognizing method based on parallel-convolution neutral net |
| CN107220506A (en) * | 2017-06-05 | 2017-09-29 | 东华大学 | Breast cancer risk assessment analysis system based on depth convolutional neural networks |
| CN107341506A (en) * | 2017-06-12 | 2017-11-10 | 华南理工大学 | A kind of Image emotional semantic classification method based on the expression of many-sided deep learning |
| CN107748900A (en) * | 2017-11-08 | 2018-03-02 | 山东财经大学 | Tumor of breast sorting technique and device based on distinction convolutional neural networks |
| CN107886514A (en) * | 2017-11-22 | 2018-04-06 | 浙江中医药大学 | Breast molybdenum target image lump semantic segmentation method based on depth residual error network |
| CN108052977A (en) * | 2017-12-15 | 2018-05-18 | 福建师范大学 | Breast molybdenum target picture depth study classification method based on lightweight neutral net |
-
2018
- 2018-05-29 CN CN201810530455.2A patent/CN108830282A/en active Pending
Patent Citations (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103425986A (en) * | 2013-08-31 | 2013-12-04 | 西安电子科技大学 | Breast lump image feature extraction method based on edge neighborhood weighing |
| CN105975915A (en) * | 2016-04-28 | 2016-09-28 | 大连理工大学 | Front vehicle parameter identification method based on multitask convolution nerve network |
| CN106326931A (en) * | 2016-08-25 | 2017-01-11 | 南京信息工程大学 | Mammary gland molybdenum target image automatic classification method based on deep learning |
| CN106339591A (en) * | 2016-08-25 | 2017-01-18 | 汤平 | Breast cancer prevention self-service health cloud service system based on deep convolutional neural network |
| CN106997475A (en) * | 2017-02-24 | 2017-08-01 | 中国科学院合肥物质科学研究院 | A kind of insect image-recognizing method based on parallel-convolution neutral net |
| CN107220506A (en) * | 2017-06-05 | 2017-09-29 | 东华大学 | Breast cancer risk assessment analysis system based on depth convolutional neural networks |
| CN107341506A (en) * | 2017-06-12 | 2017-11-10 | 华南理工大学 | A kind of Image emotional semantic classification method based on the expression of many-sided deep learning |
| CN107748900A (en) * | 2017-11-08 | 2018-03-02 | 山东财经大学 | Tumor of breast sorting technique and device based on distinction convolutional neural networks |
| CN107886514A (en) * | 2017-11-22 | 2018-04-06 | 浙江中医药大学 | Breast molybdenum target image lump semantic segmentation method based on depth residual error network |
| CN108052977A (en) * | 2017-12-15 | 2018-05-18 | 福建师范大学 | Breast molybdenum target picture depth study classification method based on lightweight neutral net |
Non-Patent Citations (4)
| Title |
|---|
| SIHONG CHEN等: "Automatic Scoring of Multiple Semantic Attributes With Multi-Task Feature Leverage: A Study on Pulmonary Nodules in CT Images", 《 IEEE TRANSACTIONS ON MEDICAL IMAGING》 * |
| SIHONG CHEN等: "Bridging Computational Features Toward Multiple Semantic Features with Multi-task Regression: A Study of CT Pulmonary Nodules", 《MICCAI 2016》 * |
| XIAOGUANG TU等: "Automatic Categorization and Scoring of Solid, Part-Solid and Non-Solid Pulmonary Nodules in CT Images with Convolutional Neural Network", 《SCIENTIFIC REPORTS》 * |
| 薛迪秀: "基于卷积神经网络的医学图像癌变识别研究", 《中国博士学位论文全文数据库 信息科技辑》 * |
Cited By (17)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109886143A (en) * | 2019-01-29 | 2019-06-14 | 上海鹰瞳医疗科技有限公司 | Multi-tag disaggregated model training method and equipment |
| CN110232396A (en) * | 2019-04-09 | 2019-09-13 | 贵州大学 | X-ray breast image deep learning classification method |
| CN110232396B (en) * | 2019-04-09 | 2022-07-01 | 贵州大学 | X-ray mammary gland image deep learning classification method |
| CN110414562B (en) * | 2019-06-26 | 2023-11-24 | 平安科技（深圳）有限公司 | X-ray film classification method, device, terminal and storage medium |
| CN110414562A (en) * | 2019-06-26 | 2019-11-05 | 平安科技（深圳）有限公司 | Classification method, device, terminal and the storage medium of X-ray |
| WO2020258507A1 (en) * | 2019-06-26 | 2020-12-30 | 平安科技（深圳）有限公司 | X-ray film classification method and apparatus, terminal, and storage medium |
| CN110738231A (en) * | 2019-07-25 | 2020-01-31 | 太原理工大学 | Method for classifying mammary gland X-ray images by improving S-DNet neural network model |
| CN110428007A (en) * | 2019-08-01 | 2019-11-08 | 科大讯飞(苏州)科技有限公司 | X-ray image object detection method, device and equipment |
| CN110414539A (en) * | 2019-08-05 | 2019-11-05 | 腾讯科技（深圳）有限公司 | A kind of method and relevant apparatus for extracting characterization information |
| CN110599476A (en) * | 2019-09-12 | 2019-12-20 | 腾讯科技（深圳）有限公司 | Disease grading method, device, equipment and medium based on machine learning |
| CN110599476B (en) * | 2019-09-12 | 2023-05-23 | 腾讯科技（深圳）有限公司 | Disease grading method, device, equipment and medium based on machine learning |
| CN110728674A (en) * | 2019-10-21 | 2020-01-24 | 清华大学 | Image processing method and device, electronic equipment and computer readable storage medium |
| CN110728674B (en) * | 2019-10-21 | 2022-04-05 | 清华大学 | Image processing method and device, electronic equipment and computer readable storage medium |
| CN110827296A (en) * | 2019-11-01 | 2020-02-21 | 南京信息工程大学 | Mammary gland X-ray image analysis method of multi-target integrated deep neural network |
| CN111709950B (en) * | 2020-08-20 | 2020-11-06 | 成都金盘电子科大多媒体技术有限公司 | Mammary gland molybdenum target AI auxiliary screening method |
| CN111709950A (en) * | 2020-08-20 | 2020-09-25 | 成都金盘电子科大多媒体技术有限公司 | Mammary gland molybdenum target AI auxiliary screening method |
| CN112614096A (en) * | 2020-12-17 | 2021-04-06 | 北京医准智能科技有限公司 | Ordinal number regression-based breast molybdenum target lesion benign and malignant analysis method |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108830282A (en) | A kind of the breast lump information extraction and classification method of breast X-ray image | |
| Zhang et al. | DeepPap: deep convolutional networks for cervical cell classification | |
| Ghyar et al. | Computer vision based approach to detect rice leaf diseases using texture and color descriptors | |
| Swasono et al. | Classification of tobacco leaf pests using VGG16 transfer learning | |
| Rothe et al. | Cotton leaf disease identification using pattern recognition techniques | |
| Song et al. | Simultaneous cell detection and classification in bone marrow histology images | |
| CN107016405A (en) | A kind of insect image classification method based on classification prediction convolutional neural networks | |
| Chanda et al. | Plant disease identification and classification using back-propagation neural network with particle swarm optimization | |
| Khoshdeli et al. | Detection of nuclei in H&E stained sections using convolutional neural networks | |
| CN109271895A (en) | Pedestrian's recognition methods again based on Analysis On Multi-scale Features study and Image Segmentation Methods Based on Features | |
| Karthikeyan et al. | Feature selection and parameters optimization of support vector machines based on hybrid glowworm swarm optimization for classification of diabetic retinopathy | |
| Sertel et al. | Texture classification using nonlinear color quantization: Application to histopathological image analysis | |
| Ji et al. | Research on urine sediment images recognition based on deep learning | |
| Ishengoma et al. | Hybrid convolution neural network model for a quicker detection of infested maize plants with fall armyworms using UAV-based images | |
| CN109740734A (en) | A kind of method of neuron spatial arrangement in optimization convolutional neural networks | |
| CN109344856A (en) | A kind of off-line signature verification method based on multilayer discriminate feature learning | |
| Wang et al. | Neutrophils identification by deep learning and voronoi diagram of clusters | |
| Ma et al. | A novel two-stage deep method for mitosis detection in breast cancer histology images | |
| Al-Hadidi et al. | Glioblastomas brain tumour segmentation based on convolutional neural networks. | |
| Raj et al. | Computer aided agriculture development for crop disease detection by segmentation and classification using deep learning architectures | |
| Krawczyk et al. | Breast thermogram analysis using a cost-sensitive multiple classifier system | |
| Tsai et al. | Deep learning techniques for colorectal cancer tissue classification | |
| Sharmila et al. | Convolution Neural Networks based lungs disease detection and Severity classification | |
| Mamatha et al. | Lung Cancer Prediction from CT Images and using Deep Learning Techniques | |
| Saleh et al. | Classification of melanoma images using modified teaching learning based artificial bee colony |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| RJ01 | Rejection of invention patent application after publication | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20181116 |