CN115578309A - Method, system, electronic device and storage medium for acquiring lung cancer characteristic information - Google Patents
Method, system, electronic device and storage medium for acquiring lung cancer characteristic information Download PDFInfo
- Publication number
- CN115578309A CN115578309A CN202210932005.2A CN202210932005A CN115578309A CN 115578309 A CN115578309 A CN 115578309A CN 202210932005 A CN202210932005 A CN 202210932005A CN 115578309 A CN115578309 A CN 115578309A
- Authority
- CN
- China
- Prior art keywords
- image
- lung
- image data
- dcgans
- lung cancer
- 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
- 238000000034 method Methods 0.000 title claims abstract description 76
- 208000020816 lung neoplasm Diseases 0.000 title claims abstract description 72
- 206010058467 Lung neoplasm malignant Diseases 0.000 title claims abstract description 70
- 201000005202 lung cancer Diseases 0.000 title claims abstract description 70
- 210000004072 lung Anatomy 0.000 claims abstract description 135
- 238000013527 convolutional neural network Methods 0.000 claims abstract description 48
- 238000012549 training Methods 0.000 claims abstract description 28
- 230000011218 segmentation Effects 0.000 claims abstract description 26
- 206010061289 metastatic neoplasm Diseases 0.000 claims abstract description 19
- 238000007781 pre-processing Methods 0.000 claims abstract description 18
- 101710122751 Sphingomyelinase C Proteins 0.000 claims abstract 6
- 238000009499 grossing Methods 0.000 claims description 24
- 230000002708 enhancing effect Effects 0.000 claims description 20
- 238000012545 processing Methods 0.000 claims description 19
- 238000001914 filtration Methods 0.000 claims description 18
- 238000004590 computer program Methods 0.000 claims description 13
- 230000000877 morphologic effect Effects 0.000 claims description 12
- 230000002146 bilateral effect Effects 0.000 claims description 10
- 230000006870 function Effects 0.000 claims description 8
- 230000003042 antagnostic effect Effects 0.000 claims description 6
- 238000003709 image segmentation Methods 0.000 claims description 6
- 238000013507 mapping Methods 0.000 claims description 6
- 238000007619 statistical method Methods 0.000 claims description 6
- 238000012512 characterization method Methods 0.000 claims 1
- 238000010586 diagram Methods 0.000 description 7
- 238000004891 communication Methods 0.000 description 6
- 208000009119 Giant Axonal Neuropathy Diseases 0.000 description 2
- 230000003044 adaptive effect Effects 0.000 description 2
- 238000013473 artificial intelligence Methods 0.000 description 2
- 230000015572 biosynthetic process Effects 0.000 description 2
- 238000002591 computed tomography Methods 0.000 description 2
- 201000003382 giant axonal neuropathy 1 Diseases 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 230000001360 synchronised effect Effects 0.000 description 2
- 238000003786 synthesis reaction Methods 0.000 description 2
- 206010050017 Lung cancer metastatic Diseases 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/0002—Inspection of images, e.g. flaw detection
- G06T7/0012—Biomedical image inspection
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- 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
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T5/00—Image enhancement or restoration
- G06T5/70—Denoising; Smoothing
-
- 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
- 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
- G06T2207/20—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
- G06T2207/30061—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
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Computer Vision & Pattern Recognition (AREA)
- General Health & Medical Sciences (AREA)
- Health & Medical Sciences (AREA)
- Artificial Intelligence (AREA)
- Data Mining & Analysis (AREA)
- Radiology & Medical Imaging (AREA)
- Nuclear Medicine, Radiotherapy & Molecular Imaging (AREA)
- Life Sciences & Earth Sciences (AREA)
- Medical Informatics (AREA)
- Biomedical Technology (AREA)
- Biophysics (AREA)
- Computational Linguistics (AREA)
- Quality & Reliability (AREA)
- Evolutionary Computation (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- General Engineering & Computer Science (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Image Processing (AREA)
Abstract
The application provides a method, a system, an electronic device and a storage medium for acquiring lung cancer characteristic information, comprising the following steps: preprocessing image data to be classified into image data to be segmented; segmenting a lung area image and a non-lung tissue image from the image data to be segmented by a segmentation method based on a Deep Convolutional Neural Network (DCNN); and training the lung area image through an anti-network DCGANS based on the generation of the deep convolutional neural network to obtain lung cancer characteristic information capable of determining the homologous lung cancer SMPLC or the lung metastatic tumor IM. The process of image classification SMPLC and IM is more accurate and faster.
Description
Technical Field
The present application belongs to the medical field, and in particular, relates to a method, a system, an electronic device, and a storage medium for acquiring lung cancer characteristic information.
Background
According to the method, a doctor often cannot rapidly and accurately distinguish two Lung cancers by a judgment method mainly depending on personal ability and experience of the treating doctor, and further the doctor is influenced to use different treatment schemes for different Lung cancers.
Disclosure of Invention
The embodiment of the invention mainly aims to provide a method, a system, electronic equipment and a storage medium for acquiring lung cancer characteristic information, so that the process of classifying images into SMPLC and IM is more accurate and faster.
In a first aspect, a method for acquiring lung cancer characteristic information is provided, and the method includes:
preprocessing image data to be classified into image data to be segmented;
segmenting a lung area image and a non-lung tissue image from the image data to be segmented by a segmentation method based on a Deep Convolutional Neural Network (DCNN);
and training the lung area image through an anti-network DCGANS based on the generation of the deep convolutional neural network to obtain lung cancer characteristic information capable of determining the homologous lung cancer SMPLC or the lung metastatic tumor IM.
In one possible implementation, the
Training the lung area image through an anti-network DCGANS based on generation of a deep convolutional neural network to obtain lung cancer characteristic information capable of determining homologous lung cancer SMPLC or lung metastatic tumor IM, wherein the lung cancer characteristic information comprises the following steps:
generating DCGANS by a multi-layer perceptron with DCNN set as GANS, and classifying the lung region images through the DCGANSThe method comprises the following steps: defining the maximum minimum form of the objective function as
Acquiring, selecting a maximum logD (G (z)) to train a discriminator of the DCGANS, setting a Convolutional Neural Network (CNN) as a multi-layer perceptron of the DCGANS, pre-training a Deep Convolutional Neural Network (DCNN) as a discriminator of the DCGANS, and detecting a false CT image and a real image generated by the DCGANS through the DCNN.
In another possible implementation manner, the preprocessing the image data to be classified into the image data to be segmented includes:
smoothing the image data to be classified into first image data through bilateral filtering based on a spatial domain and a value domain;
denoising the first image data into second image data through a non-local mean denoising algorithm NL-means;
enhancing the second image data into image data to be segmented by the second image data through a dark channel algorithm, wherein the enhancing process comprises the following steps: and enhancing the contrast of the second image data and improving the visibility of the second image data.
In another possible implementation manner, the segmenting the lung region image and the non-lung tissue image from the image data to be segmented by the DCNN-based segmentation method includes:
segmenting an initial lung region from the image data to be segmented by a segmentation method based on DCNN, comprising: dividing the image data to be divided into a plurality of image blocks, removing lung area image blocks and background image blocks with similar densities to obtain pleural tissue image blocks, and subtracting the background image blocks extracted by binarization and morphological operations to obtain lung area image blocks;
performing local contour refinement on the initial lung region by linear iterative clustering, SLIC, comprising: generating a super-pixel image with the boundary aligned with the local edge of the image and the approximate size, mapping the super-pixel image to the initial lung region, endowing different labels to the super-pixel image according to whether the super-pixel image is positioned in the initial lung region, and classifying super-pixel points with different labels by an adjacent point statistical method to obtain a local contour thinning result;
correcting the initial lung region image refined by the local contour through edge tracking to obtain a lung region image, wherein the method comprises the following steps: according to the similarity between the edge point and the adjacent edge point, starting from an initial point, recording connectable edge lines to form a closed curve, and performing contour smoothing on the closed curve according to a morphological closing operation and a moving average filtering method.
In a second aspect, a system for acquiring lung cancer feature information is provided, the system comprising:
the pre-processing module is used for pre-processing the image data to be classified into image data to be segmented;
the image segmentation module is used for segmenting a lung region image and a non-lung tissue image from the image data to be segmented by a segmentation method based on a Deep Convolutional Neural Network (DCNN);
and the lung cancer characteristic information acquisition module is used for training the lung area image through an antagonistic network DCGANS generated based on a deep convolutional neural network to acquire lung cancer characteristic information capable of determining the homologous lung cancer SMPLC or the lung metastatic tumor IM.
In one possible implementation, the training of the lung region image through the antagonistic network DCGANS based on the generation of the deep convolutional neural network to obtain lung cancer feature information that can determine the homologous lung cancer smclc or the lung metastatic tumor IM includes:
generating DCGANS by a multi-layer perceptron with DCNN set as GANS, and classifying the lung region images through the DCGANS, wherein the DCGANS comprises the following steps: defining the maximum minimum form of the objective function as
Acquiring, selecting a maximum logD (G (z)) to train a discriminator of the DCGANS, setting a convolutional neural network CNN as a multi-layer perceptron of the DCGANS, pre-training a deep convolutional neural network DCNN as a discriminator of the DCGANS, and detecting the DCG through the DCNNFalse CT images and real images generated by the ANS.
In another possible implementation manner, the obtaining module further includes:
the smoothing unit is used for smoothing the image data to be classified into first image data through bilateral filtering based on a spatial domain and a value domain;
the denoising processing unit is used for denoising the first image data into second image data through a non-local mean denoising algorithm NL-means;
the enhancement processing unit is used for enhancing the second image data into image data to be segmented through the dark channel algorithm, and the enhancement processing comprises the following steps: and enhancing the contrast of the second image data and improving the visibility of the second image data.
In another possible implementation manner, the image segmentation module includes:
an initial lung region segmentation unit, configured to segment an initial lung region from the image data to be segmented by a DCNN-based segmentation method, including: dividing the image data to be divided into a plurality of image blocks, removing lung area image blocks and background image blocks with similar densities to obtain pleural tissue image blocks, and subtracting the background image blocks extracted by binarization and morphological operations to obtain lung area image blocks;
a local contour refinement unit, configured to perform local contour refinement on the initial lung region through linear iterative clustering SLIC, including: generating a super-pixel image with the boundary aligned with the local edge of the image and the approximate size, mapping the super-pixel image to the initial lung region, endowing different labels to the super-pixel image according to whether the super-pixel image is positioned in the initial lung region, and classifying super-pixel points with different labels by an adjacent point statistical method to obtain a local contour thinning result;
the lung area image acquisition unit is used for correcting the initial lung area image refined by the local contour through edge tracking to acquire a lung area image, and comprises: according to the similarity between the edge point and the adjacent edge point, starting from an initial point, recording connectable edge lines to form a closed curve, and performing contour smoothing on the closed curve according to a morphological closing operation and a moving average filtering method.
In a third aspect, an electronic device is provided, which includes a memory, a processor, and a computer program stored in the memory and executable on the processor, and when the processor executes the computer program, the method for acquiring lung cancer characteristic information as provided in the first aspect is implemented.
In a fourth aspect, a non-transitory computer readable storage medium is provided, on which a computer program is stored, which computer program, when executed by a processor, implements the method of lung cancer characteristic information acquisition as provided in the first aspect.
Drawings
In order to more clearly illustrate the technical solutions in the embodiments of the present application, the drawings used in the description of the embodiments of the present application will be briefly described below.
Fig. 1 is a flowchart of a method for acquiring lung cancer feature information according to an embodiment of the present invention;
FIG. 2 is a flowchart of a method for obtaining lung cancer feature information according to another embodiment of the present invention;
FIG. 3 is a flowchart of a method for acquiring lung cancer feature information according to another embodiment of the present invention
FIG. 4 is a block diagram of a system for acquiring lung cancer feature information according to an embodiment of the present invention;
FIG. 5 is a block diagram of a system for acquiring lung cancer feature information according to another embodiment of the present invention;
FIG. 6 is a block diagram of a system for acquiring lung cancer feature information according to another embodiment of the present invention;
fig. 7 is a schematic physical structure diagram of an electronic device according to the present invention.
Detailed description of the invention
Reference will now be made in detail to embodiments of the present application, examples of which are illustrated in the accompanying drawings, wherein like or similar reference numerals refer to the same or similar modules or modules having the same or similar functionality throughout. The embodiments described below with reference to the drawings are exemplary only for the purpose of explaining the present application and are not to be construed as limiting the present invention.
As used herein, the singular forms "a", "an", "the" and "the" are intended to include the plural forms as well, unless the context clearly indicates otherwise. It will be further understood that the terms "comprises" and/or "comprising," when used in this specification, specify the presence of stated features, integers, steps, operations, modules, and/or components, but do not preclude the presence or addition of one or more other features, integers, steps, operations, modules, components, and/or groups thereof. It will be understood that when a module is referred to as being "connected" or "coupled" to another module, it can be directly connected or coupled to the other module or intervening modules may also be present. Further, "connected" or "coupled" as used herein may include wirelessly connected or wirelessly coupled. As used herein, the term "and/or" includes all or any module and all combinations of one or more of the associated listed items.
To make the objectives, technical solutions and advantages of the present application more clear, the following detailed description of the implementations of the present application will be made with reference to the accompanying drawings.
The technical solutions of the present application and the technical solutions of the present application, for example, and solving the above technical problems, will be described in detail with specific examples below. The following several specific embodiments may be combined with each other, and details of the same or similar concepts or processes may not be repeated in some embodiments. Embodiments of the present application will be described below with reference to the accompanying drawings.
Fig. 1 is a flowchart illustrating a method for acquiring lung cancer characteristic information according to an embodiment of the present invention, where the method includes:
102, segmenting a lung region image and a non-lung tissue image from the image data to be segmented by a segmentation method based on a Deep Convolutional Neural Network (DCNN);
and 103, training the lung area image through an antagonistic network DCGANS generated based on the deep convolutional neural network to acquire lung cancer characteristic information capable of determining the homologous lung cancer SMPLC or the lung metastatic tumor IM.
In the embodiment of the invention, the image data to be classified is processed into the image data to be segmented through a preprocessing program comprising smoothing processing, denoising processing and enhancing processing. Segmenting image data to be segmented by a preset segmentation method based on DCNN (Deep convolutional Neural Network), and segmenting a lung region and other non-lung regions from the image data to be segmented. The method comprises the steps of applying DCNN to a GANS (generic adaptive Networks, generation of countermeasure Networks) to generate a DCGANS model, training a Lung area through the DCGANS model to obtain Lung Cancer feature information which can be used for distinguishing SMPLC (Synchronous Multiple Primary Lung Cancer) or IM (Lung metastatic tumor), and distinguishing a Lung area image into the SMPLC or IM according to the Lung Cancer feature information.
The method for training the lung area image through the generation of the countermeasure network DCGANS based on the deep convolutional neural network to obtain the lung cancer characteristic information capable of determining the homologous lung cancer SMPLC or the lung metastatic tumor IM comprises the following steps:
generating DCGANS by a multi-layer perceptron with DCNN set as GANS, and classifying the lung region images through the DCGANS, wherein the DCGANS comprises the following steps: defining the maximum and minimum form of the objective function as
Acquiring, selecting a maximum logD (G (z)) to train a discriminator of the DCGANS, setting a convolutional neural network CNN as a multi-layer perceptron of the DCGANS, pre-training a deep convolutional neural network DCNN as a discriminator of the DCGANS, and detecting a false CT image and a real image generated by the DCGANS through the DCNN.
In the embodiment of the invention, the DCGANS model combines DCNN and GANS, a multi-layer perceptron of the GANS model is replaced by the DCNN, the model is an unsupervised representation learning model, labels are not used, therefore, the cost of model training is lower, and deep and high-level features are fully learned in both the component part of an object and the scene in a generator and a discriminator. The DCNN is pre-trained as a discriminator in the GANS to detect false CT images and true images produced by the generative model. In this way, the discriminator will learn the structure of the CT image, while robust features of the CT scan can be extracted. The classification of smmplc and IM is achieved by training a pre-trained DCNN on the master data set and replacing the last fully-connected layer of the GANs discriminator with a SoftMax layer for classification purposes.
According to the embodiment of the invention, image data to be classified is preprocessed into image data to be segmented, a lung region image and a non-lung tissue image are segmented from the image data to be segmented by a segmentation method based on DCNN, the lung region image is trained by an anti-network DCGANS generated based on a deep convolutional neural network, and lung cancer characteristic information capable of determining homologous lung cancer SMPLC or lung metastatic tumor IM is acquired. Through the intelligent judgment of artificial intelligence to image data for the process of image classification SMPLC and IM is faster, and the classification result is more accurate.
As shown in fig. 2, a flowchart of a method for classifying an SMPLC and an IM according to another embodiment of the present invention is provided, where the preprocessing of the image data to be classified into image data to be segmented includes:
202, denoising the first image data into second image data through a non-local mean denoising algorithm NL-means;
In the embodiment of the present invention, the preprocessing of the image data to be classified includes, but is not limited to: smoothing, denoising and enhancing. For smoothing, bilateral filtering considering a spatial domain and a value domain is used for smoothing image data to be classified, and the bilateral filtering can carry out smoothing on images while protecting edges by carrying out nonlinear synthesis on pixel values of two adjacent frames of images; for denoising, denoising image data to be classified by using an NL-means algorithm, wherein the NL-means algorithm can fully utilize redundant information in an image and keep the detail characteristics of the image while denoising; for enhancement processing, a dark channel algorithm is used for enhancing the image data to be classified, and the dark channel algorithm can enhance the contrast and the visibility of the image data to be classified by combining with a guide filtering function.
Fig. 3 is a flowchart of a method for classifying SMPLC and IM according to still another embodiment of the present invention, where the method for segmenting the image data to be segmented into the lung region image and the non-lung tissue image by the DCNN-based segmentation method includes:
step 301, segmenting an initial lung region from the image data to be segmented by a DCNN-based segmentation method, including: dividing the image data to be divided into a plurality of image blocks, removing lung area image blocks and background image blocks with similar densities to obtain pleural tissue image blocks, and subtracting the background image blocks extracted by binarization and morphological operations to obtain lung area image blocks;
step 302, performing local contour refinement on the initial lung region through linear iterative clustering SLIC, including: generating a super-pixel image with the boundary aligned with the local edge of the image and the approximate size, mapping the super-pixel image to the initial lung region, endowing different labels to the super-pixel image according to whether the super-pixel image is positioned in the initial lung region, and classifying super-pixel points with different labels by an adjacent point statistical method to obtain a local contour thinning result;
step 303, correcting the initial lung region image refined by the local contour through edge tracking to obtain a lung region image, including: according to the similarity between the edge point and the adjacent edge point, starting from an initial point, recording connectable edge lines to form a closed curve, and performing contour smoothing on the closed curve according to a morphological closing operation and a moving average filtering method.
Fig. 4 is a flowchart of a system for acquiring lung cancer feature information according to an embodiment of the present invention, where the system includes:
the preprocessing module 401 is configured to preprocess the image data to be classified into image data to be segmented;
an image segmentation module 402, configured to segment a lung region image and a non-lung tissue image from the image data to be segmented by a segmentation method based on a deep convolutional neural network DCNN;
a lung cancer characteristic information obtaining module 403, configured to train the lung region image through an antagonistic network DCGANS generated based on a deep convolutional neural network, so as to obtain lung cancer characteristic information that may determine a homologous lung cancer SMPLC or a lung metastatic tumor IM.
In the embodiment of the invention, the image data to be classified is processed into the image data to be segmented through a preprocessing program comprising smoothing processing, denoising processing and enhancing processing. Segmenting image data to be segmented by a preset segmentation method based on DCNN (Deep convolutional Neural Network), and segmenting a lung region and other non-lung regions from the image data to be segmented. The DCNN is applied to a GANS (generic adaptive Networks, generation of countermeasure Networks) to generate a DCGANS model, the Lung area is trained through the DCGANS model, lung Cancer characteristic information which can be used for distinguishing an SMPLC (Synchronous Multiple Primary Lung Cancer) or an IM (metastatic Lung Cancer) is obtained, and then the Lung area image is distinguished into the SMPLC or the IM according to the Lung Cancer characteristic information.
The method for training the lung area image through the generation of the countermeasure network DCGANS based on the deep convolutional neural network to obtain the lung cancer characteristic information capable of determining the homologous lung cancer SMPLC or the lung metastatic tumor IM comprises the following steps:
generating DCGANS by a multi-layer perceptron with DCNN set as GANS, and classifying the lung region images through the DCGANS, wherein the DCGANS comprises the following steps: defining the maximum and minimum form of the objective function as
Obtaining, selectingAnd training a discriminator of the DCGANS by maximizing logD (G (z)), setting a Convolutional Neural Network (CNN) as a multi-layer perceptron of the DCGANS, pre-training a Deep Convolutional Neural Network (DCNN) as an identifier of the DCGANS, and detecting false CT images and real images generated by the DCGANS through the DCNN.
In the embodiment of the invention, the DCGANS model combines DCNN and GANS, a multi-layer perceptron of the GANS model is replaced by the DCNN, the model is an unsupervised representation learning model, labels are not used, therefore, the cost of model training is lower, and deep and high-level features are fully learned in both the component part of an object and the scene in a generator and a discriminator. The DCNN is pre-trained as a discriminator in the GANS to detect false CT images and true images produced by the generative model. In this way, the discriminator will learn the structure of the CT image, while robust features of the CT scan can be extracted. By training the pre-trained DCNN on the master data set, the last fully-connected layer of the GANs discriminator is replaced by a SoftMax layer for classification purposes, and the classification of sMPLC and IM is realized.
According to the embodiment of the invention, image data to be classified is preprocessed into image data to be segmented, a lung region image and a non-lung tissue image are segmented from the image data to be segmented by a segmentation method based on DCNN, the lung region image is trained by an anti-network DCGANS generated based on a deep convolutional neural network, and lung cancer characteristic information capable of determining homologous lung cancer SMPLC or lung metastatic tumor IM is obtained. The image data are intelligently judged through artificial intelligence, so that the process of image classification of the SMPLC and the IM is quicker, and the classification result is more accurate.
As shown in fig. 5, which is a block diagram of a system for classifying smcplc and IM according to another embodiment of the present invention, the preprocessing module 401 includes:
a smoothing processing unit 4011, configured to smooth the image data to be classified into first image data through bilateral filtering based on a spatial domain and a value domain;
the denoising processing unit 4012 is configured to denoise the first image data into second image data through a non-local mean denoising algorithm NL-means;
an enhancement processing unit 4013, which performs enhancement processing on the second image data into image data to be segmented by using a dark channel algorithm, where the enhancement processing includes: and enhancing the contrast of the second image data and improving the visibility of the second image data.
In the embodiment of the present invention, the preprocessing of the image data to be classified includes, but is not limited to: smoothing, denoising and enhancing. For smoothing, bilateral filtering considering a spatial domain and a value domain is used for smoothing image data to be classified, and the bilateral filtering can carry out smoothing on images while protecting edges by carrying out nonlinear synthesis on pixel values of two adjacent frames of images; for denoising, denoising image data to be classified by using an NL-means algorithm, wherein the NL-means algorithm can fully utilize redundant information in an image and keep the detail characteristics of the image while denoising; for enhancement processing, a dark channel algorithm is used for enhancing the image data to be classified, and the dark channel algorithm can enhance the contrast and the visibility of the image data to be classified by combining with a guide filtering function.
As shown in fig. 6, which is a block diagram of a system for classifying an SMPLC and an IM according to still another embodiment of the present invention, the image segmentation module 402 includes:
the initial lung region segmentation unit 4021 is configured to segment an initial lung region from the image data to be segmented by a DCNN-based segmentation method, and includes: dividing the image data to be divided into a plurality of image blocks, removing lung area image blocks and background image blocks with similar densities to obtain pleural tissue image blocks, and subtracting the background image blocks extracted by binarization and morphological operations to obtain lung area image blocks;
a local contour refinement unit 4022, configured to perform local contour refinement on the initial lung region through linear iterative clustering SLIC, and including: generating a super-pixel image with the boundary aligned with the local edge of the image and approximate size, mapping the super-pixel image to the initial lung region, endowing different labels to the super-pixel image according to whether the super-pixel image is positioned in the initial lung region, and classifying super-pixel points of different labels by an adjacent point statistical method to obtain a local contour thinning result;
the lung region image obtaining unit 4023 is configured to correct the initial lung region image refined by the local contour by edge tracking, and obtain a lung region image, including: according to the similarity between the edge point and the adjacent edge point, starting from an initial point, recording connectable edge lines to form a closed curve, and performing contour smoothing on the closed curve according to a morphological closing operation and a moving average filtering method.
Fig. 7 illustrates a physical structure diagram of an electronic device, and as shown in fig. 7, the electronic device may include: a processor (processor) 701, a communication Interface (Communications Interface) 702, a memory (memory) 703 and a communication bus 704, wherein the processor, the communication Interface and the memory complete communication with each other through the communication bus. The processor may invoke logic instructions in the memory to perform a method of lung cancer characteristic information acquisition, the method comprising: preprocessing image data to be classified into image data to be segmented; segmenting a lung region image and a non-lung tissue image from the image data to be segmented by a segmentation method based on a Deep Convolutional Neural Network (DCNN); and training the lung area image through an anti-network DCGANS generated based on a deep convolutional neural network to obtain lung cancer characteristic information capable of determining the homologous lung cancer SMPLC or the lung metastatic tumor IM.
In addition, the logic instructions in the memory may be implemented in the form of software functional units and may be stored in a computer readable storage medium when sold or used as a stand-alone product. Based on such understanding, the technical solution of the present invention may be embodied in the form of a software product, which is stored in a storage medium and includes instructions for causing a computer device (which may be a personal computer, a server, or a network device) to execute all or part of the steps of the method according to the embodiments of the present invention. And the aforementioned storage medium includes: a U-disk, a removable hard disk, a Read-Only Memory (ROM), a Random Access Memory (RAM), a magnetic disk or an optical disk, and other various media capable of storing program codes.
In another aspect, an embodiment of the present invention further provides a computer program product, where the computer program product includes a computer program stored on a non-transitory computer-readable storage medium, the computer program includes program instructions, and when the program instructions are executed by a computer, the computer can execute the method for acquiring lung cancer feature information provided by the above-mentioned method embodiments, where the method includes: preprocessing image data to be classified into image data to be segmented; segmenting a lung region image and a non-lung tissue image from the image data to be segmented by a segmentation method based on a Deep Convolutional Neural Network (DCNN); and training the lung area image through an anti-network DCGANS generated based on a deep convolutional neural network to obtain lung cancer characteristic information capable of determining the homologous lung cancer SMPLC or the lung metastatic tumor IM.
In yet another aspect, an embodiment of the present invention further provides a non-transitory computer-readable storage medium, on which a computer program is stored, where the computer program is implemented by a processor to perform the method for acquiring lung cancer feature information provided by the above embodiments, where the method includes: preprocessing image data to be classified into image data to be segmented; segmenting a lung region image and a non-lung tissue image from the image data to be segmented by a segmentation method based on a Deep Convolutional Neural Network (DCNN); and training the lung area image through an anti-network DCGANS based on the generation of the deep convolutional neural network to obtain lung cancer characteristic information capable of determining the homologous lung cancer SMPLC or the lung metastatic tumor IM.
It should be understood that, although the steps in the flowcharts of the figures are shown in order as indicated by the arrows, the steps are not necessarily performed in order as indicated by the arrows. The steps are not performed in the exact order shown and may be performed in other orders unless explicitly stated herein. Moreover, at least a portion of the steps in the flow chart of the figure may include multiple sub-steps or multiple stages, which are not necessarily performed at the same time, but may be performed at different times, and the order of execution is not necessarily sequential, but may be performed alternately or alternately with other steps or at least a portion of the sub-steps or stages of other steps.
The above description is only a partial implementation of the present invention, and it should be noted that, for those skilled in the art, a plurality of modifications and embellishments can be made without departing from the principle of the present invention, and these modifications and embellishments should also be regarded as the protection scope of the present invention.
Claims (10)
1. A method for acquiring lung cancer characteristic information, which is characterized by comprising the following steps:
preprocessing image data to be classified into image data to be segmented;
segmenting a lung region image and a non-lung tissue image from the image data to be segmented by a segmentation method based on a Deep Convolutional Neural Network (DCNN);
and training the lung area image through an anti-network DCGANS based on the generation of the deep convolutional neural network to obtain lung cancer characteristic information capable of determining the homologous lung cancer SMPLC or the lung metastatic tumor IM.
2. The method of claim 1, wherein the training of the lung region image through the generation of the antagonistic network based on the deep convolutional neural network DCGANS to obtain lung cancer feature information that can determine homologous lung cancer SMPLC or lung metastatic tumor IM comprises:
generating DCGANS by a multi-layer perceptron with DCNN set as GANS, and classifying the lung region images through the DCGANS, wherein the DCGANS comprises the following steps: defining the maximum and minimum form of the objective function as
Acquiring, selecting and training a discriminator of the DCGANS by maximizing logD (G (z)), and setting a Convolutional Neural Network (CNN) as a multi-layer perceptron of the DCGANSAnd pre-training a Deep Convolutional Neural Network (DCNN) into a discriminator of the DCGANS, and detecting a false CT image and a real image generated by the DCGANS through the DCNN.
3. The method of any one of claims 1-2, wherein preprocessing the image data to be classified into image data to be segmented comprises:
smoothing the image data to be classified into first image data through bilateral filtering based on a spatial domain and a value domain;
denoising the first image data into second image data by a non-local mean denoising algorithm NL-means;
enhancing the second image data into image data to be segmented by the second image data through a dark channel algorithm, wherein the enhancing process comprises the following steps: and enhancing the contrast of the second image data and improving the visibility of the second image data.
4. The method according to any one of claims 1-2, wherein the segmenting lung region images and non-lung tissue images from the image data to be segmented by a DCNN-based segmentation method comprises:
segmenting an initial lung region from the image data to be segmented by a DCNN-based segmentation method, comprising: dividing the image data to be divided into a plurality of image blocks, removing lung area image blocks and background image blocks with similar densities to obtain pleural tissue image blocks, and subtracting the background image blocks extracted by binarization and morphological operations to obtain lung area image blocks;
performing local contour refinement on the initial lung region by linear iterative clustering, SLIC, comprising: generating a super-pixel image with the boundary aligned with the local edge of the image and the approximate size, mapping the super-pixel image to the initial lung region, endowing different labels to the super-pixel image according to whether the super-pixel image is positioned in the initial lung region, and classifying super-pixel points with different labels by an adjacent point statistical method to obtain a local contour thinning result;
correcting the initial lung region image refined by the local contour through edge tracking to obtain a lung region image, wherein the method comprises the following steps: according to the similarity between the edge point and the adjacent edge point, starting from an initial point, recording connectable edge lines to form a closed curve, and performing contour smoothing on the closed curve according to a morphological closing operation and a moving average filtering method.
5. A system for acquiring lung cancer characteristic information, the system comprising:
the pre-processing module is used for pre-processing the image data to be classified into image data to be segmented;
the image segmentation module is used for segmenting a lung region image and a non-lung tissue image from the image data to be segmented by a segmentation method based on a Deep Convolutional Neural Network (DCNN);
and the lung cancer characteristic information acquisition module is used for training the lung area image through an antagonistic network DCGANS generated based on a deep convolutional neural network to acquire lung cancer characteristic information capable of determining the homologous lung cancer SMPLC or the lung metastatic tumor IM.
6. The system of claim 5, wherein the training of the lung region image through the DCGANS based on the deep convolutional neural network generation to obtain lung cancer characterization information that can determine homologous lung cancer SMPLCs or lung metastatic tumors IM comprises:
generating DCGANS by a multi-layer perceptron with DCNN set as GANS, and classifying the lung region images through the DCGANS, wherein the DCGANS comprises the following steps: defining the maximum minimum form of the objective function as
Acquiring, selecting a maximum logD (G (z)) to train a discriminator of the DCGANS, setting a convolutional neural network CNN as a multi-layer perceptron of the DCGANS, pre-training a deep convolutional neural network DCNN as a discriminator of the DCGANS, and detecting a false CT image and a real image generated by the DCGANS through the DCNN.
7. The system of any of claims 5-6, wherein the acquisition module further comprises:
the smoothing unit is used for smoothing the image data to be classified into first image data through bilateral filtering based on a spatial domain and a value domain;
the denoising processing unit is used for denoising the first image data into second image data through a non-local mean denoising algorithm NL-means;
the enhancement processing unit is used for enhancing the second image data into image data to be segmented through the dark channel algorithm, and the enhancement processing comprises the following steps: and enhancing the contrast of the second image data and improving the visibility of the second image data.
8. The system of claim 7, wherein the image segmentation module comprises:
an initial lung region segmentation unit, configured to segment an initial lung region from the image data to be segmented by a DCNN-based segmentation method, including: dividing the image data to be divided into a plurality of image blocks, removing lung area image blocks and background image blocks with similar densities to obtain pleural tissue image blocks, and subtracting the background image blocks extracted by binarization and morphological operations to obtain lung area image blocks;
a local contour refinement unit, configured to perform local contour refinement on the initial lung region through linear iterative clustering SLIC, including: generating a super-pixel image with the boundary aligned with the local edge of the image and the approximate size, mapping the super-pixel image to the initial lung region, endowing different labels to the super-pixel image according to whether the super-pixel image is positioned in the initial lung region, and classifying super-pixel points with different labels by an adjacent point statistical method to obtain a local contour thinning result;
the lung area image acquisition unit is used for correcting the initial lung area image refined by the local contour through edge tracking to acquire a lung area image, and comprises: according to the similarity between the edge point and the adjacent edge point, starting from an initial point, recording connectable edge lines to form a closed curve, and performing contour smoothing on the closed curve according to a morphological closing operation and a moving average filtering method.
9. An electronic device comprising a memory, a processor and a computer program stored on the memory and executable on the processor, wherein the processor when executing the program implements the method of lung cancer characteristic information acquisition according to any one of claims 1 to 4.
10. A non-transitory computer readable storage medium having stored thereon a computer program, wherein the computer program, when executed by a processor, implements the method of lung cancer feature information acquisition according to any one of claims 1-4.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210932005.2A CN115578309A (en) | 2022-08-04 | 2022-08-04 | Method, system, electronic device and storage medium for acquiring lung cancer characteristic information |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210932005.2A CN115578309A (en) | 2022-08-04 | 2022-08-04 | Method, system, electronic device and storage medium for acquiring lung cancer characteristic information |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN115578309A true CN115578309A (en) | 2023-01-06 |
Family
ID=84580162
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202210932005.2A Pending CN115578309A (en) | 2022-08-04 | 2022-08-04 | Method, system, electronic device and storage medium for acquiring lung cancer characteristic information |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115578309A (en) |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109447998A (en) * | 2018-09-29 | 2019-03-08 | 华中科技大学 | Based on the automatic division method under PCANet deep learning model |
| CN110706234A (en) * | 2019-10-08 | 2020-01-17 | 浙江工业大学 | Automatic fine segmentation method for image |
| CN113838026A (en) * | 2021-09-22 | 2021-12-24 | 中南大学 | Non-small cell lung cancer detection method, non-small cell lung cancer detection device, computer equipment and storage medium |
| WO2022063199A1 (en) * | 2020-09-24 | 2022-03-31 | 上海健康医学院 | Pulmonary nodule automatic detection method, apparatus and computer system |
-
2022
- 2022-08-04 CN CN202210932005.2A patent/CN115578309A/en active Pending
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109447998A (en) * | 2018-09-29 | 2019-03-08 | 华中科技大学 | Based on the automatic division method under PCANet deep learning model |
| CN110706234A (en) * | 2019-10-08 | 2020-01-17 | 浙江工业大学 | Automatic fine segmentation method for image |
| WO2022063199A1 (en) * | 2020-09-24 | 2022-03-31 | 上海健康医学院 | Pulmonary nodule automatic detection method, apparatus and computer system |
| CN113838026A (en) * | 2021-09-22 | 2021-12-24 | 中南大学 | Non-small cell lung cancer detection method, non-small cell lung cancer detection device, computer equipment and storage medium |
Non-Patent Citations (4)
| Title |
|---|
| 于明浩: "基于深度卷积对抗生成网络的肺结节分类和分割方法", 《中国优秀硕士学位论文全文数据库 医药卫生科技辑》, pages 15 - 18 * |
| 宋丽梅 等: "《人工智能与船海工程》", 31 July 2020, 上海：上海科学技术出版社, pages: 117 - 118 * |
| 宋丽梅 等: "《机器视觉与机器学习》", 北京：机械工业出版社, pages: 30 * |
| 王耀南 等: "《移动作业机器人感知、规划与控制》", 30 September 2020, 北京：国防工业出版社, pages: 45 - 46 * |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109522908B (en) | Image significance detection method based on region label fusion | |
| Gangwar et al. | IrisSeg: A fast and robust iris segmentation framework for non-ideal iris images | |
| CN108154519A (en) | Dividing method, device and the storage medium of eye fundus image medium vessels | |
| CN111340824B (en) | Image feature segmentation method based on data mining | |
| CN109241973B (en) | Full-automatic soft segmentation method for characters under texture background | |
| CN111681198A (en) | Morphological attribute filtering multimode fusion imaging method, system and medium | |
| Unni et al. | Tumour detection in double threshold segmented mammograms using optimized GLCM features fed SVM | |
| CN108010035A (en) | Finger vena image segmentation method and its system, terminal based on the detection of direction paddy shape | |
| CN112801031A (en) | Vein image recognition method and device, electronic equipment and readable storage medium | |
| Omer et al. | Preprocessing of digital mammogram image based on otsu’s threshold | |
| CN113421210B (en) | Surface point Yun Chong construction method based on binocular stereoscopic vision | |
| CN116843581B (en) | Image enhancement method, system, device and storage medium for multi-scene graph | |
| CN101847260B (en) | Image processing apparatus, and image processing method | |
| CN117877085A (en) | Psychological analysis method based on micro-expression recognition | |
| CN109145875B (en) | Method and device for removing black frame glasses in face image | |
| CN116524269A (en) | Visual recognition detection system | |
| CN115578309A (en) | Method, system, electronic device and storage medium for acquiring lung cancer characteristic information | |
| Lu et al. | Directed assistance for ink-bleed reduction in old documents | |
| Mathiyalagan et al. | Image fusion using convolutional neural network with bilateral filtering | |
| Manchanda et al. | An improved multifocus image fusion algorithm using deep learning and adaptive fuzzy filter | |
| Shaik et al. | Unsupervised segmentation of image using novel curve evolution method | |
| CN117974706B (en) | Rock slice particle pit segmentation method based on dynamic threshold and local search | |
| CN117237619B (en) | Water rescue detection system and method based on machine vision technology | |
| CN113487496B (en) | Image denoising method, system and device based on pixel type inference | |
| CN111832460B (en) | Face image extraction method and system based on multi-feature fusion |
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 |