CN110246578A - A kind of lung cancer sample rapid detection method - Google Patents
A kind of lung cancer sample rapid detection method Download PDFInfo
- Publication number
- CN110246578A CN110246578A CN201910508037.8A CN201910508037A CN110246578A CN 110246578 A CN110246578 A CN 110246578A CN 201910508037 A CN201910508037 A CN 201910508037A CN 110246578 A CN110246578 A CN 110246578A
- Authority
- CN
- China
- Prior art keywords
- image
- lung cancer
- images
- convolutional neural
- feature
- 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
- 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
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/11—Region-based segmentation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/70—Determining position or orientation of objects or cameras
- G06T7/73—Determining position or orientation of objects or cameras using feature-based methods
-
- 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
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/10—Image acquisition modality
- G06T2207/10072—Tomographic images
- G06T2207/10081—Computed x-ray tomography [CT]
-
- 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
- G06T2207/30061—Lung
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/30—Subject of image; Context of image processing
- G06T2207/30004—Biomedical image processing
- G06T2207/30096—Tumor; Lesion
Abstract
The present invention relates to a kind of lung cancer sample rapid detection methods, by the segmented image template for extracting lung areas from CT images image, and the segmented image is handled, using Lung Cancer Images sophisticated category method, textural characteristics are merged with shape feature, by distributing the weight of two kinds of features, carry out template matching with fusion feature, and the database based on big data in processing result and healthy cloud service platform is subjected to classification contrast judgement, the efficiency and accuracy of pulmonary cancer diagnosis will be greatly improved.
Description
Technical field
The present invention relates to lung cancer detection technical field, specifically a kind of lung cancer sample rapid detection method.
Background technique
Lung cancer is the most common malignant tumour of countries in the world today, and the death rate occupy the first place of various tumours, to the mankind
Health and lives constitute great threat.In China, lung cancer about causes 500,000 deaths every year, accounts for entire cases of cancer
28%, and 5 annual survival rates of lung cancer patient only have 14%.However, researches show that the postoperative 10 years survival rates of I phase lung cancer can reach
92%.Therefore, the key for reducing the patients with lung cancer death rate is to early diagnose and early treatment, the Lung neoplasm of the early stage of lung cancer detect
As key, cooperate treatment appropriate, the survival rate of patient can be improved to 50%.
5 kinds of basic methods of sieving and diagnosis lung cancer at present, x-ray chest radiograph are preferred screening means, secondly be exactly CT, MIR and
PET.CT is considered as the best approach --- " goldstandard " for detecting Lung neoplasm.However, because economic, convenient and radiological dose is fitted
Medium reason, x-ray chest radiograph are more often used, in fact, the almost all of early stage of lung cancer is found by rabat, but to dept. of radiology
It is a highly difficult task based on the rabat discovery early stage of lung cancer for doctor.It is nowadays accepted that the most universal in the world, warp
Ji and traditional method of lung cancer diagnosis is by chest X- light emission line image (the mainly number caused by CR/DR technology
The X- light emission line image of change) diagnose the early stage of lung cancer.At present in Economic contrast developed regions, chest is all carried out in health screening
X-ray technology checks whether there is pulmonary disease.However a large amount of generaI investigation chest pictures are diagnosed, for radiologist
It is a challenge.
With the arrival of cloud era, big data is also by originally more concerns, the image list of previous hospital's pulmonary cancer diagnosis
Solely storage, each hospital can only be trained doctor according to the influence that oneself stores, improve the diagnosis capability to lung cancer, effect
Slow effect, and since sample is few, it is difficult to achieve the purpose that quick, Accurate Diagnosis, if big data can be based on by all diagnostic graphs
As carry out image procossing, and by cloud computing concentrate carry out Classification and Identification, will greatly improve pulmonary cancer diagnosis efficiency and accurately
Degree.
Summary of the invention
Technical problem to be solved by the invention is to provide a kind of lung cancer sample rapid detection methods, to solve the prior art
Present in defect.
The technical scheme to solve the above technical problems is that
A kind of lung cancer sample rapid detection method, by the segmented image mould for extracting lung areas from CT images image
Plate, and the segmented image is handled according to following steps:
Step 1: carrying out lesion localization to original input picture using the focal area detection method based on grey scale change；
Step 2: the focal part navigated to is generated the different template of a large amount of scales at random, by the lesion of back
The generating unit control of formwork is directly reduced the generation of redundancy template, each sample is random in lesions position by positioning
Generate template of a large amount of image block as image；
Step 3: input picture to be carried out to the scaling of different scale, the image block of scaling and formwork are carried out respectively special
Sign is extracted, texture feature extraction MB-LBP and shape feature wavelet moment, preferable to medical image recognition effect in view of textural characteristics,
The textural characteristics of different weights are merged the shape spy of wavelet moment description by the parameter of experiment adjustment weight distribution by we respectively
It levies, the fusion feature under two kinds of weights regards two class textural characteristics as；
Step 4: matching by the different location to image, matched result is expressed as to the shape of characteristic response figure
Formula；
Step 5: using improved mean value spatial pyramid model by characteristic response figure be converted into the feature of an octuple to
Amount:
Step 6: feature vector realizes Lung Cancer Images sophisticated category as input, using support vector machines；Pass through above-mentioned step
Suddenly image classification that treated is sent to healthy cloud service platform by network, and the healthy cloud service platform mainly includes
The image reading module for receiving and reading chest x-ray piece or CT images image that user sends over, by textural characteristics with
Shape feature fusion carries out template matching with fusion feature, with the dress of user's access platform by distributing the weight of two kinds of features
Standby user name or number is the file generation module of folder name, based on depth convolutional neural networks to the lung after segmentation
The doubtful Lung Cancer Types categorization module that area image is classified, storage has is consulted with the generation health that doubtful Lung Cancer Types are index
The early prevention of file and the healthy file generating module for the treatment of are ask, is used for the health consultation file of user to be fed back to access
The automatic transmission module of the file at family takes for the healthy file of early prevention and treatment to be supplied to user to the healthy cloud
The downloading service module downloaded on the website of business platform.
Further, segmented image module lung from CT images image by way of based on full convolutional neural networks
Portion region is split, and the convolutional neural networks is changed to full convolutional neural networks, in the convolutional neural networks
Full articulamentum is changed to warp lamination, directly obtains dense prediction, that is, each picture in output end after input piece image in this way
Class belonging to element, to obtain an end-to-end method to realize lung's object images semantic segmentation；
Further, the depth convolutional neural networks are the 8th layer of the full connections in the convolutional neural networks
It is connected to a Softmax classifier after layer, for carrying out Classification and Identification to doubtful Lung Cancer Types；
The beneficial effects of the present invention are: merged textural characteristics with shape feature using Lung Cancer Images sophisticated category method,
By distributing the weight of two kinds of features, template matching is carried out with fusion feature, and by processing result and based on the data of big data
Library carries out classification comparison, will greatly improve the efficiency and accuracy of pulmonary cancer diagnosis.
Detailed description of the invention
Fig. 1 is schematic structural view of the invention；
Specific embodiment
The principle and features of the present invention will be described below with reference to the accompanying drawings, and the given examples are served only to explain the present invention, and
It is non-to be used to limit the scope of the invention.
As shown in Figure 1, a kind of lung cancer sample rapid detection method, by point for extracting lung areas from CT images image
Image template is cut, and the segmented image is handled according to following steps:
Step 1: carrying out lesion localization to original input picture using the focal area detection method based on grey scale change；
Step 2: the focal part navigated to is generated the different template of a large amount of scales at random, by the lesion of back
The generating unit control of formwork is directly reduced the generation of redundancy template, each sample is random in lesions position by positioning
Generate template of a large amount of image block as image；
Step 3: input picture to be carried out to the scaling of different scale, the image block of scaling and formwork are carried out respectively special
Sign is extracted, texture feature extraction MB-LBP and shape feature wavelet moment, preferable to medical image recognition effect in view of textural characteristics,
The textural characteristics of different weights are merged the shape spy of wavelet moment description by the parameter of experiment adjustment weight distribution by we respectively
It levies, the fusion feature under two kinds of weights regards two class textural characteristics as；
Step 4: matching by the different location to image, matched result is expressed as to the shape of characteristic response figure
Formula；
Step 5: using improved mean value spatial pyramid model by characteristic response figure be converted into the feature of an octuple to
Amount:
Step 6: feature vector realizes Lung Cancer Images sophisticated category as input, using support vector machines；
Image classification that treated through the above steps is sent to healthy cloud service platform, the healthy cloud by network
Service platform mainly includes the image reading for receiving and reading chest x-ray piece or CT images image that user sends over
Module merges textural characteristics with shape feature, by the weight of two kinds of features of distribution, carries out template matching with fusion feature,
Using the user name of the equipment of user's access platform or number as the file generation module of folder name, it is based on depth convolutional Neural
The doubtful Lung Cancer Types categorization module that network classifies to the lung areas image after segmentation, storage have with doubtful Lung Cancer Types
For index generation health consultation file early prevention and treatment healthy file generating module, for by user health consult
The automatic transmission module of file that file feeds back to access user is ask, for early prevention and the healthy file for the treatment of to be supplied to use
The downloading service module downloaded on family to the website of the healthy cloud service platform.
More specifically, segmented image module lung from CT images image by way of based on full convolutional neural networks
Portion region is split, and the convolutional neural networks is changed to full convolutional neural networks, in the convolutional neural networks
Full articulamentum is changed to warp lamination, directly obtains dense prediction, that is, each picture in output end after input piece image in this way
Class belonging to element, to obtain an end-to-end method to realize lung's object images semantic segmentation；
More specifically, the depth convolutional neural networks are the 8th layer of the full connections in the convolutional neural networks
It is connected to a Softmax classifier after layer, for carrying out Classification and Identification to doubtful Lung Cancer Types；
Concrete operating principle:
The chest x-ray piece or CT images that the self-service healthy cloud service system of the prevention lung cancer is sended over according to user
Image carries out CT images image using the dividing method of lung areas in the slave CT images image based on full convolutional neural networks
Then the segmentation of lung's object carries out image procossing；Then according to doubtful Lung Cancer Types classification specification depth convolutional Neural net
Network carries out identification classification to the lung images after segmentation；If the user has history chest x-ray piece or CT images image, just again
It is compared with the history chest x-ray piece of the user or CT images image, compares its difference；If the user has pathology
Expert clinical diagnosis report just combines these information to carry out comprehensive analysis, proposes diagnosing and treating suggestion, automatically generates self-service strong
Then the report of health detection result is submitted to senior radiologist and confirmed by health test results report, finally will be healthy
Test results report information feeds back to user；
Compare accurately doubtful Lung Cancer Types accuracy of identification in order to obtain, it is desirable that every kind of classification including have assemblage characteristic
Classification doubtful Lung Cancer Types characteristic image at least at 3000 or more, can be used data enhancing converter technique come increase input
The amount of data；
The foregoing is merely presently preferred embodiments of the present invention, is not intended to limit the invention, it is all in spirit of the invention and
Within principle, any modification, equivalent replacement, improvement and so on be should all be included in the protection scope of the present invention.
Claims (2)
1. a kind of lung cancer sample rapid detection method, it is characterised in that: by point for extracting lung areas from CT images image
Image template is cut, and the segmented image is handled according to following steps:
Step 1: carrying out lesion localization to original input picture using the focal area detection method based on grey scale change；
Step 2: the focal part navigated to is generated the different template of a large amount of scales at random, by the lesion localization of back,
The generating unit control of formwork is directly reduced into the generation of redundancy template, each sample is randomly generated in lesions position
A large amount of template of the image block as image；
Step 3: input picture to be carried out to the scaling of different scale, feature is carried out with formwork to the image block of scaling respectively and is mentioned
It taking, texture feature extraction MB-LBP and shape feature wavelet moment are preferable to medical image recognition effect in view of textural characteristics, we
By the parameter of experiment adjustment weight distribution, the textural characteristics of different weights are merged to the shape feature of wavelet moment description respectively,
Fusion feature under two kinds of weights regards two class textural characteristics as；
Step 4: matching by the different location to image, matched result is expressed as to the form of characteristic response figure；
Step 5: converting characteristic response figure to using improved mean value spatial pyramid model the feature vector of an octuple；
Step 6: feature vector realizes Lung Cancer Images sophisticated category as input, using support vector machines；
Image classification that treated through the above steps is sent to healthy cloud service platform, the healthy cloud service by network
Platform mainly includes the image reading module for receiving and reading chest x-ray piece or CT images image that user sends over,
Textural characteristics are merged with shape feature, by distributing the weight of two kinds of features, template matching are carried out with fusion feature, with user
The user name or number of the equipment of access platform are the file generation module of folder name, are based on depth convolutional neural networks pair
The doubtful Lung Cancer Types categorization module that lung areas image after segmentation is classified, storage have with doubtful Lung Cancer Types for index
Generation health consultation file early prevention and treatment healthy file generating module, for by the health consultation file of user
The automatic transmission module of file for feeding back to access user, for the healthy file of early prevention and treatment to be supplied to user to institute
The downloading service module downloaded on the website for the healthy cloud service platform stated；
Wherein, the segmented image module by way of based on full convolutional neural networks from CT images image lung areas into
The convolutional neural networks are changed to full convolutional neural networks, in the full articulamentum of the convolutional neural networks by row segmentation
It is changed to warp lamination, directly obtains dense prediction in output end after input piece image in this way, that is, belonging to each pixel
Class, to obtain an end-to-end method to realize lung's object images semantic segmentation.
2. a kind of lung cancer sample rapid detection method according to claim 1, it is characterised in that: the depth convolution mind
It is to be connected to a Softmax classifier after the 8th layer of the convolutional neural networks of full articulamentum through network, is used for
Classification and Identification is carried out to doubtful Lung Cancer Types.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910508037.8A CN110246578A (en) | 2019-06-12 | 2019-06-12 | A kind of lung cancer sample rapid detection method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910508037.8A CN110246578A (en) | 2019-06-12 | 2019-06-12 | A kind of lung cancer sample rapid detection method |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN110246578A true CN110246578A (en) | 2019-09-17 |
Family
ID=67886835
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910508037.8A Pending CN110246578A (en) | 2019-06-12 | 2019-06-12 | A kind of lung cancer sample rapid detection method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110246578A (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112967293A (en) * | 2021-03-04 | 2021-06-15 | 首都师范大学 | Image semantic segmentation method and device and storage medium |
| CN113450899A (en) * | 2021-06-22 | 2021-09-28 | 上海市第一人民医院 | Intelligent diagnosis guiding method based on artificial intelligence cardiopulmonary examination images |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105718947A (en) * | 2016-01-21 | 2016-06-29 | 吉林大学 | Lung cancer image fine classification method based on fusion of LBP and wavelet moment features |
| 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 |
| CN106372390A (en) * | 2016-08-25 | 2017-02-01 | 姹ゅ钩 | Deep convolutional neural network-based lung cancer preventing self-service health cloud service system |
| CN109003672A (en) * | 2018-07-16 | 2018-12-14 | 北京睿客邦科技有限公司 | A kind of early stage of lung cancer detection classification integration apparatus and system based on deep learning |
-
2019
- 2019-06-12 CN CN201910508037.8A patent/CN110246578A/en active Pending
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105718947A (en) * | 2016-01-21 | 2016-06-29 | 吉林大学 | Lung cancer image fine classification method based on fusion of LBP and wavelet moment features |
| 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 |
| CN106372390A (en) * | 2016-08-25 | 2017-02-01 | 姹ゅ钩 | Deep convolutional neural network-based lung cancer preventing self-service health cloud service system |
| CN109003672A (en) * | 2018-07-16 | 2018-12-14 | 北京睿客邦科技有限公司 | A kind of early stage of lung cancer detection classification integration apparatus and system based on deep learning |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112967293A (en) * | 2021-03-04 | 2021-06-15 | 首都师范大学 | Image semantic segmentation method and device and storage medium |
| CN113450899A (en) * | 2021-06-22 | 2021-09-28 | 上海市第一人民医院 | Intelligent diagnosis guiding method based on artificial intelligence cardiopulmonary examination images |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Zhu et al. | Multi-scale coarse-to-fine segmentation for screening pancreatic ductal adenocarcinoma | |
| Becker et al. | Deep learning in mammography: diagnostic accuracy of a multipurpose image analysis software in the detection of breast cancer | |
| US7529394B2 (en) | CAD (computer-aided decision) support for medical imaging using machine learning to adapt CAD process with knowledge collected during routine use of CAD system | |
| Reiser et al. | Computerized mass detection for digital breast tomosynthesis directly from the projection images | |
| US9392986B2 (en) | Method and apparatus for cone beam breast CT image-based computer-aided detection and diagnosis | |
| Sahiner et al. | Computer‐aided detection of clustered microcalcifications in digital breast tomosynthesis: a 3D approach | |
| US20220254023A1 (en) | System and Method for Interpretation of Multiple Medical Images Using Deep Learning | |
| Bogoni et al. | Impact of a computer-aided detection (CAD) system integrated into a picture archiving and communication system (PACS) on reader sensitivity and efficiency for the detection of lung nodules in thoracic CT exams | |
| Costaridou | Medical image analysis methods | |
| Dong et al. | One step further into the blackbox: a pilot study of how to build more confidence around an AI-based decision system of breast nodule assessment in 2D ultrasound | |
| Abe et al. | Computer-aided diagnosis in chest radiology | |
| Lai et al. | DBT masses automatic segmentation using U-net neural networks | |
| Xi et al. | Deep learning for differentiation of benign and malignant solid liver lesions on ultrasonography | |
| Ahmed et al. | Automated pulmonary nodule classification and detection using deep learning architectures | |
| CN110246578A (en) | A kind of lung cancer sample rapid detection method | |
| JP5048233B2 (en) | Method and system for anatomical shape detection in a CAD system | |
| Dolejsi et al. | The lung time: Annotated lung nodule dataset and nodule detection framework | |
| Arzhaeva et al. | Computer‐aided detection of interstitial abnormalities in chest radiographs using a reference standard based on computed tomography | |
| Gebre et al. | Detecting hip osteoarthritis on clinical CT: a deep learning application based on 2-D summation images derived from CT | |
| Roy et al. | Automated detection of lung nodules in CT scans: False‐positive reduction with the radial‐gradient index | |
| Xu et al. | Improved cascade R-CNN for medical images of pulmonary nodules detection combining dilated HRNet | |
| Jaglan et al. | Escalate the efficacy of breast tumor detection through magnetic resonance imaging: a framework | |
| He et al. | Medical image analysis of multiple myeloma based on convolutional neural network | |
| Bhowmik et al. | Automated triage of screening breast mri examinations in high-risk women using an ensemble deep learning model | |
| Suárez-Cuenca et al. | Automated detection of pulmonary nodules in CT: false positive reduction by combining multiple classifiers |
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: 20190917 |