CN113538415A - Segmentation method and device for pulmonary blood vessels in medical image and electronic equipment - Google Patents
Segmentation method and device for pulmonary blood vessels in medical image and electronic equipment Download PDFInfo
- Publication number
- CN113538415A CN113538415A CN202110935661.3A CN202110935661A CN113538415A CN 113538415 A CN113538415 A CN 113538415A CN 202110935661 A CN202110935661 A CN 202110935661A CN 113538415 A CN113538415 A CN 113538415A
- Authority
- CN
- China
- Prior art keywords
- pulmonary
- vessel
- result
- medical image
- outputting
- 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.)
- Withdrawn
Links
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
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/213—Feature extraction, e.g. by transforming the feature space; Summarisation; Mappings, e.g. subspace methods
- G06F18/2132—Feature extraction, e.g. by transforming the feature space; Summarisation; Mappings, e.g. subspace methods based on discrimination criteria, e.g. discriminant analysis
-
- 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
-
- 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
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/12—Edge-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
- G06T2207/10—Image acquisition modality
- G06T2207/10072—Tomographic images
- G06T2207/10088—Magnetic resonance imaging [MRI]
-
- 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
- G06T2207/30—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
- G06T2207/30101—Blood vessel; Artery; Vein; Vascular
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Data Mining & Analysis (AREA)
- Artificial Intelligence (AREA)
- Evolutionary Computation (AREA)
- General Engineering & Computer Science (AREA)
- Life Sciences & Earth Sciences (AREA)
- Health & Medical Sciences (AREA)
- General Health & Medical Sciences (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Biomedical Technology (AREA)
- Computing Systems (AREA)
- Molecular Biology (AREA)
- Computational Linguistics (AREA)
- Biophysics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Quality & Reliability (AREA)
- Radiology & Medical Imaging (AREA)
- Nuclear Medicine, Radiotherapy & Molecular Imaging (AREA)
- Medical Informatics (AREA)
- Evolutionary Biology (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Image Analysis (AREA)
Abstract
The application provides a segmentation method of a pulmonary blood vessel in a medical image, which comprises the following steps: acquiring an original medical image; training an original medical image by adopting a deep learning network model, and outputting a pre-segmentation result by utilizing the model obtained by training; extracting the vessel characteristics of each pixel point in the original medical image, judging the vessel characteristics according to a vessel characteristic judging function to obtain a vessel characteristic judging result, processing the original medical image according to the vessel characteristic judging result, and outputting a first pulmonary vessel enhancement result; and taking the pulmonary artery and pulmonary vein region of the first pulmonary blood vessel enhancement result as an initial growth point to perform region growth, outputting a second pulmonary blood vessel enhancement result after the region growth, fusing the second pulmonary blood vessel enhancement result with the pre-segmentation result, and outputting a final segmentation result. The application also provides a segmentation device of the pulmonary blood vessel in the medical image and electronic equipment. Therefore, the pulmonary blood vessel segmentation method and device can rapidly and accurately segment and split the pulmonary blood vessel, and efficiency and accuracy of the pulmonary blood vessel segmentation are improved.
Description
Technical Field
The present application relates to the field of medical image processing technologies, and in particular, to a method and an apparatus for segmenting a pulmonary blood vessel in a medical image, and an electronic device.
Background
Lung cancer has become the first cancer killer threatening the health of China, surgical operation plays a decisive role in the treatment of removing lung cancer, and minimally invasive thoracoscopic surgery improves the postoperative life quality and survival rate of patients. The anatomical lung resection is safe and feasible through the thoracoscope, and the selective early lung cancer treatment is basically determined. The clear and accurate pulmonary vascular structure obtained before the operation has a crucial influence on the success of the operation, and the rapid and accurate segmentation, splitting and reconstruction of pulmonary artery and vein vessels through a pulmonary medical image is a very effective method.
The current main methods include a region growing method based on image gray scale and a deep learning method using a large number of labeled samples. However, since pulmonary vessels are very complex, neither simple region growing methods nor deep learning methods can produce satisfactory results, often requiring extensive post-manual repair of broken vessel branches, or correction of misclassified pulmonary arterial vessels, pulmonary venous vessels.
In view of the above, the prior art is obviously inconvenient and disadvantageous in practical use, and needs to be improved.
Disclosure of Invention
In view of the foregoing defects, an object of the present application is to provide a method, an apparatus and an electronic device for segmenting a pulmonary blood vessel in a medical image, which can rapidly and accurately segment and segment the pulmonary blood vessel, thereby improving the efficiency and accuracy of segmenting the pulmonary blood vessel.
In order to solve the technical problem, the present application is implemented as follows:
in a first aspect, an embodiment of the present application provides a method for segmenting a pulmonary blood vessel in a medical image, the method including:
acquiring original medical images of pulmonary artery blood vessels and pulmonary vein blood vessels to be segmented;
training the original medical image by adopting a deep learning network model, and outputting a pre-segmentation result by utilizing the model obtained by training;
extracting the vessel characteristics of each pixel point in the original medical image, judging the vessel characteristics according to a preset vessel characteristic judging function to obtain a vessel characteristic judging result, processing the original medical image according to the vessel characteristic judging result, and outputting a first pulmonary blood vessel enhancement result;
and performing region growth by taking the pulmonary artery and pulmonary vein region of the first pulmonary blood vessel enhancement result as an initial growth point, outputting a second pulmonary blood vessel enhancement result after the region growth, fusing the second pulmonary blood vessel enhancement result with the pre-segmentation result, and outputting a final segmentation result.
According to the segmentation method of the pulmonary vessels in the medical images, the training of the original medical images by adopting the deep learning network model and the output of the pre-segmentation result comprise the following steps:
outputting downsampling data after downsampling the original medical image, training the downsampling data by adopting a neural network model, and outputting a first pulmonary vascular network model with low resolution;
respectively obtaining low-resolution prediction results of the pulmonary artery blood vessel and the pulmonary vein blood vessel through the first pulmonary vascular network model, performing up-sampling processing on the prediction results, and outputting up-sampled data;
training the up-sampling data and the original medical image by using the neural network model, and outputting a second pulmonary vascular network model with high resolution;
and training the original medical image by respectively adopting the first pulmonary vascular network model and the second pulmonary vascular network model, and outputting the pre-segmentation results of the pulmonary artery blood vessels and the pulmonary vein blood vessels.
According to the segmentation method of pulmonary vessels in medical images, after the outputting the pre-segmentation results of the pulmonary artery vessels and pulmonary vein vessels, the method further comprises:
and respectively carrying out connected region identification on the pre-segmentation results of the pulmonary artery blood vessel and the pulmonary vein blood vessel, only reserving a maximum connected region, and removing broken branches.
According to the segmentation method of the pulmonary vessels in the medical image, the acquiring of the original medical image of the pulmonary artery vessels and the pulmonary vein vessels to be segmented comprises:
acquiring original medical images of the pulmonary artery blood vessels and the pulmonary vein blood vessels to be segmented and corresponding original segmentation data;
the outputting of the downsampled data after the downsampling processing of the original medical image comprises:
outputting the down-sampling data after down-sampling processing is carried out on the original medical image and the original segmentation data, wherein the down-sampling data are low-resolution images and marking data;
the training the up-sampled data and the raw medical image using the neural network model includes:
training the up-sampled data, the raw medical image, and the raw segmentation data using the neural network model.
According to the segmentation method of the pulmonary blood vessel in the medical image, the original medical image is a CT image or an MRI image; or
The neural network model is a 3D-UNET neural network model.
According to the method for segmenting the pulmonary blood vessel in the medical image, the steps of extracting the vessel feature of each pixel point in the original medical image, judging the vessel feature according to a preset vessel feature judging function to output a vessel feature judging result, processing the original medical image according to the vessel feature judging result, and outputting a first pulmonary blood vessel enhancement result include:
calculating the vessel characteristics of each pixel point in the original medical image;
extracting and constructing a Hessian matrix according to the vessel characteristics of the pixel points;
constructing the vessel feature discriminant function;
judging the vessel characteristics of each pixel point of the Hessian matrix according to the vessel characteristic judging function, and outputting a vessel characteristic judging result;
and processing the original medical image according to the vessel feature judgment result of each pixel point, and outputting the first pulmonary vessel enhancement result.
According to the segmentation method for the pulmonary blood vessel in the medical image, the calculating the vessel characteristics of each pixel point in the original medical image comprises:
constructing a Gaussian filter GσFor the Gaussian filter GσCalculating a second derivative, and convolving the second derivative with each pixel point of the original medical image one by one to obtain the corresponding vessel characteristic Ixx(σ)、Iyy(σ)、Izz(σ)、Ixy(σ)、Ixz(sigma) and Iyz(σ)；
The extracting and constructing Hessian matrix according to the vessel characteristics of the pixel points comprises the following steps:
extracting and constructing the Hessian matrix according to the second derivative in the three-dimensional neighborhood of each pixel point:
the constructing the vessel feature discriminant function includes:
constructing the vessel feature discriminant function Vs(λ)；
Wherein the content of the first and second substances,
H(λ)exp(h(λ1))|exp(h(λ2，λ3))|h(λ2，λ3)h(λ1) And h (λ)2，λ3) Is an adjustable characteristic function;
the distinguishing the vessel characteristics of each pixel point of the Hessian matrix according to the vessel characteristic distinguishing function and outputting the vessel characteristic distinguishing result includes:
the vessel features of the pixel points of the Hessian matrix are judged according to the vessel feature judgment function, the pixel points which are judged to be positive are located in the areas of the pulmonary artery and the pulmonary vein, and the vessel feature judgment result is output;
the processing the original medical image according to the vessel feature discrimination result of each pixel point and outputting the first pulmonary vessel enhancement result includes:
processing the original medical image according to the vessel feature discrimination result of each pixel point, and outputting the first pulmonary vessel enhancement result as follows:
according to the segmentation method of the pulmonary vessels in the medical image, the method comprises the steps of performing region growing by using the pulmonary artery and pulmonary vein region of the first pulmonary vessel enhancement result as an initial growing point, outputting a second pulmonary vessel enhancement result after the region growing, fusing the second pulmonary vessel enhancement result with the pre-segmentation result, and outputting a final segmentation result, wherein the method comprises the following steps:
extracting a lung parenchymal region of the first pulmonary vascular enhancement result through a predetermined threshold segmentation algorithm;
intercepting the first pulmonary vessel enhancement result by taking the lung parenchymal region as a region of interest;
performing region growing by taking the intercepted pulmonary artery and pulmonary vein region of the first pulmonary artery and pulmonary vein enhancement result as an initial growing point, if any neighborhood point in the pulmonary artery and pulmonary vein region does not belong to pulmonary vessel data but is contained in the pulmonary vessel data of the lung parenchymal region, marking the neighborhood point as a pulmonary artery vessel and a pulmonary vein vessel, and performing the process circularly until all neighborhood points of the pulmonary artery and pulmonary vein region do not meet the condition, and outputting a second pulmonary vessel enhancement result after region growing;
and fusing the second pulmonary vessel enhancement result and the pre-segmentation result, and outputting the final segmentation result.
In a second aspect, an embodiment of the present application provides an apparatus for segmenting a pulmonary blood vessel in a medical image, for implementing a method for segmenting a pulmonary blood vessel in a medical image according to any one of the foregoing embodiments, the apparatus includes:
the image acquisition module is used for acquiring original medical images of pulmonary artery blood vessels and pulmonary vein blood vessels to be segmented;
the deep learning module is used for training the original medical image by adopting a deep learning network model and outputting a pre-segmentation result;
the vessel enhancement module is used for extracting vessel features of each pixel point in the original medical image, distinguishing the vessel features according to a preset vessel feature distinguishing function to obtain a vessel feature distinguishing result, processing the original medical image according to the vessel feature distinguishing result and outputting a first pulmonary vessel enhancement result;
and the growth fusion module is used for performing region growth by taking the pulmonary artery and pulmonary vein region of the first pulmonary blood vessel enhancement result as an initial growth point, outputting a second pulmonary blood vessel enhancement result after the region growth, fusing the second pulmonary blood vessel enhancement result with the pre-segmentation result, and outputting a final segmentation result.
In a third aspect, an embodiment of the present application provides an electronic device, including: storage medium, a processor and a computer program stored on the memory and executable on the processor, wherein the processor when executing the computer program implements a method for segmentation of pulmonary vessels in medical images as described in any of the above.
After acquiring original medical images of pulmonary artery blood vessels and pulmonary vein blood vessels to be segmented, training the original medical images by adopting a deep learning network model, and outputting a pre-segmentation result; extracting vessel characteristics of each pixel point in the original medical image, judging the vessel characteristics according to a preset vessel characteristic judging function to obtain a vessel characteristic judging result, processing the original medical image according to the vessel characteristic judging result, and outputting a first pulmonary blood vessel enhancement result; and taking the pulmonary artery and pulmonary vein region of the first pulmonary blood vessel enhancement result as an initial growth point to perform region growth, outputting a second pulmonary blood vessel enhancement result after the region growth, fusing the second pulmonary blood vessel enhancement result with the pre-segmentation result, and outputting a final segmentation result. Therefore, the lung blood vessel segmentation method and the lung blood vessel segmentation device have the advantages that the lung blood vessel can be rapidly and accurately segmented and split by combining the advantages of the deep learning segmentation algorithm and the region growing segmentation algorithm, the efficiency and the accuracy of the segmentation of the lung blood vessel are improved, and the postoperative life quality and the survival rate of a patient can be further improved.
Drawings
Fig. 1 is a flowchart illustrating a segmentation method for a pulmonary vessel in a medical image according to an embodiment of the present application;
FIG. 2 is a flowchart illustrating a deep learning step provided by an embodiment of the present application;
FIG. 3 is a schematic flow chart illustrating a pulmonary vessel enhancement procedure provided by an embodiment of the present application;
FIG. 4 is a schematic flow chart of a growth fusion step provided in an embodiment of the present application;
FIG. 5 is a graph of predicted results of a low resolution first pulmonary vascular network model provided by an embodiment of the present application;
FIG. 6 is a graph of predicted results of a second high resolution pulmonary vascular network model provided by an embodiment of the present application;
FIG. 7 is a graph of predicted results after pulmonary vascular enhancement provided by an embodiment of the present application;
FIG. 8 is a graph of the predicted results after growth fusion provided by the examples of the present application;
FIG. 9 is a schematic structural diagram of a segmentation apparatus for pulmonary blood vessels in a medical image according to an embodiment of the present application;
fig. 10 is a schematic structural diagram of an electronic device provided in an embodiment of the present application;
fig. 11 is a schematic hardware structure diagram of an electronic device according to an embodiment of the present application.
Detailed Description
In order to make the objects, technical solutions and advantages of the present application more apparent, the present application is described in further detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the present application and are not intended to limit the present application.
It should be noted that references in the specification to "one embodiment," "an example embodiment," etc., indicate that the embodiment described may include a particular feature, structure, or characteristic, but every embodiment may not necessarily include the particular feature, structure, or characteristic. Moreover, such phrases are not intended to refer to the same embodiment. Further, when a particular feature, structure, or characteristic is described in connection with an embodiment, it is submitted that it is within the knowledge of one skilled in the art to effect such feature, structure, or characteristic in connection with other embodiments whether or not explicitly described.
Moreover, where certain terms are used throughout the description and following claims to refer to particular components or features, those skilled in the art will understand that manufacturers may refer to a component or feature by different names or terms. This specification and the claims that follow do not intend to distinguish between components or features that differ in name but not function. In the following description and in the claims, the terms "include" and "comprise" are used in an open-ended fashion, and thus should be interpreted to mean "include, but not limited to. In addition, the term "connected" as used herein includes any direct and indirect electrical connection. Indirect electrical connection means include connection by other means.
The method for segmenting a pulmonary vessel in a medical image provided by the embodiment of the present application is described in detail below with reference to the accompanying drawings by specific embodiments and application scenarios thereof.
Fig. 1 is a schematic flowchart of a segmentation method for a pulmonary blood vessel in a medical image according to an embodiment of the present application, including the steps of:
s101, obtaining original medical images of pulmonary artery blood vessels and pulmonary vein blood vessels to be segmented.
Alternatively, the original medical image may be a CT image or an MRI image or the like.
And S102, training the original medical image by adopting a deep learning network model, and outputting a pre-segmentation result.
The method comprises a deep learning step, wherein global information and local information of an original medical image can be comprehensively considered through a deep learning network model, so that a pre-segmentation result of the pulmonary artery and the pulmonary vein with complete whole structure of the blood vessel and rich and continuous fine branch endings is obtained.
Optionally, the deep learning network model may be a neural network model, preferably a 3D-UNET neural network model.
S103, extracting the vessel characteristics of each pixel point in the original medical image, judging the vessel characteristics according to a preset vessel characteristic judging function to obtain a vessel characteristic judging result, processing the original medical image according to the vessel characteristic judging result, and outputting a first pulmonary vessel enhancement result.
This step is a pulmonary vessel enhancement step, and by means of a vessel segmentation method that is specific to the enhancement of tubular structures, vessels can be obtained that are more abundant and complete than the pre-segmentation result, although this result does not distinguish between pulmonary artery and pulmonary vein.
S104, taking the pulmonary artery and pulmonary vein region of the first pulmonary blood vessel enhancement result as an initial growth point to perform region growth, outputting a second pulmonary blood vessel enhancement result after the region growth, fusing the second pulmonary blood vessel enhancement result with a pre-segmentation result, and outputting a final segmentation result.
The step is a growth fusion step, and the blood vessel segmentation results of the two steps (S102 and S103) can be fused through a specially-arranged fusion growth and automatic repair module of the pulmonary artery and the pulmonary vein, so that a richer and more accurate final segmentation result is generated.
Fig. 2 is a schematic flowchart of a deep learning step provided in an embodiment of the present application, where the deep learning step further includes:
s201, outputting down-sampling data after down-sampling processing is carried out on the original medical image, training the down-sampling data by adopting a neural network model, and outputting a first pulmonary vascular network model with low resolution.
Optionally, an original medical image of a pulmonary artery blood vessel and a pulmonary vein blood vessel to be segmented and corresponding original segmentation data thereof are acquired. And respectively carrying out downsampling processing on the original medical image and the original segmentation data and then outputting downsampled data, wherein the downsampled data are low-resolution images and marking data.
Optionally, the neural network model is a 3D-UNET neural network model.
Alternatively, the original medical image may be a CT image or an MRI image or the like.
S202, respectively obtaining low-resolution prediction results of pulmonary artery blood vessels and pulmonary vein blood vessels through the first pulmonary vascular network model, performing up-sampling processing on the prediction results, outputting up-sampled data, and performing up-sampling processing to obtain up-sampled data to serve as a new input channel.
Fig. 5 shows the prediction result of the low-resolution first pulmonary vascular network model, with many discontinuities in vessel twigs.
S203, training the up-sampling data and the original medical image by adopting a neural network model, and outputting a second pulmonary vascular network model with high resolution.
Optionally, the up-sampled data, the raw medical image and the raw segmentation data are trained using a neural network model.
Optionally, the neural network model is a 3D-UNET neural network model.
Fig. 6 shows the predicted result of the high-resolution second pulmonary vascular network model, which has more abundant and continuous bronchioles than the predicted result of the low-resolution first pulmonary vascular network model.
S204, training the original medical image by respectively adopting the first pulmonary vascular network model and the second pulmonary vascular network model, and respectively outputting pre-segmentation results of pulmonary artery blood vessels and pulmonary vein blood vessels.
Optionally, after step S204, the method may further include:
and respectively carrying out connected region identification on the pre-segmentation results of the pulmonary artery blood vessel and the pulmonary vein blood vessel, only reserving the maximum connected region, and removing the broken branch.
In this example, the deep learning step is used for deep learning network model training and prediction, and global information and local information of the original medical image can be comprehensively considered through a two-stage deep learning network model with low resolution and high resolution, so that a pre-segmentation result of the pulmonary artery and the pulmonary vein with complete whole blood vessel structure and rich and continuous fine branch endings is obtained.
Fig. 3 is a schematic flow chart of a pulmonary vascular enhancement step provided in an embodiment of the present application, further including:
s301, calculating the vessel characteristics of each pixel point in the original medical image.
Optionally, the vessel feature discriminant function is mostly formed by a Gaussian function, and the Gaussian filter G is formedσFor Gaussian filter GσCalculating a second derivative, and convolving the second derivative with each pixel point of the original medical image one by one to obtain corresponding vessel characteristics Ixx(σ)、Iyy(σ)、Izz(σ)、Ixy(σ)、Ixz(sigma) and Iyz(σ)。
S302, extracting and constructing a Hessian matrix according to the vessel characteristics of each pixel point.
Optionally, a Hessian matrix is extracted and constructed according to the second derivative in the three-dimensional neighborhood of each pixel:
wherein three eigenvalues of H are calculated and ranked as desired, e.g. the vessels should be ranked as λ1＜λ2＜λ3；
S303, aiming at the characteristic of more small blood vessels in the lung, a vessel characteristic discrimination function is constructed.
Optionally, a vessel feature discriminant function V is constructedS(λ)。
Wherein the content of the first and second substances,
H(λ)exp(h(λ1))|exp(h(λ2，λ3))|h(λ2，λ3)h(λ1) And h (λ)2，λ3) Is a tunable characteristic function.
S304, vessel characteristics of each pixel point of the Hessian matrix are judged according to the vessel characteristic judging function, and vessel characteristic judging results are output.
Optionally, vessel features of each pixel point of the Hessian matrix are discriminated according to a vessel feature discrimination function, the pixel point which is determined to be positive is in the region of the pulmonary artery and the pulmonary vein, and a vessel feature discrimination result is output.
S305, processing the original medical image according to the vessel feature discrimination result of each pixel point, and outputting a first pulmonary vessel enhancement result.
Optionally, processing the original medical image according to the vessel feature discrimination result of each pixel point, and outputting a first pulmonary vessel enhancement result as follows:
fig. 7 shows the results of vessel segmentation after pulmonary vessel enhancement, and it can be seen that the vessel segmentation after pulmonary vessel enhancement has more abundant fine branch endings than the segmentation results of fig. 5 and 6, but cannot distinguish the arteriovenous.
In this embodiment, the pulmonary vessel enhancement step can obtain a richer and more complete vessel than the pre-segmentation result by a vessel segmentation method specifically directed to the enhancement of the tubular structure, although this result cannot distinguish between the pulmonary artery and the pulmonary vein. The tubular structure enhancement technology is to construct a vessel characteristic discrimination function taking a Hessian matrix characteristic value as a variable to discriminate the extracted vessel characteristic, the scheme firstly calculates the Hessian matrix formed by second order derivatives in a three-dimensional neighborhood of each pixel point in an image, then discriminates the characteristic value of the Hessian matrix by using the vessel characteristic discrimination function, and the pixel point which is discriminated to be positive is in a vessel region (namely a pulmonary artery region and a pulmonary vein region). By the thin blood vessel enhancement algorithm, thin blood vessels which are difficult to distinguish in lung image data can be enhanced and segmented.
Fig. 4 is a schematic flowchart of a growth fusion step provided in an embodiment of the present application, where the growth fusion step further includes:
s401, extracting a lung parenchymal region of the first pulmonary blood vessel enhancement result through a preset threshold segmentation algorithm.
S402, using the lung parenchymal Region as ROI (Region Of Interest), intercepting the first pulmonary blood vessel enhancement result, that is, extracting only all blood vessel data Of the lung parenchymal Region in the first pulmonary blood vessel enhancement result.
And S403, performing region growing by taking the pulmonary artery and pulmonary vein region of the intercepted first pulmonary artery and pulmonary vein enhancement result as an initial growing point, if any neighborhood point in the pulmonary artery and pulmonary vein region does not belong to the pulmonary artery blood vessel data but is contained in the pulmonary blood vessel data of the lung parenchymal region, marking the neighborhood point as a pulmonary artery blood vessel and a pulmonary vein blood vessel, circularly performing the process until all neighborhood points of the pulmonary artery and pulmonary vein region do not meet the condition, and outputting a second pulmonary vein enhancement result after region growing.
S404, fusing the second pulmonary vessel enhancement result and the pre-segmentation result, and outputting a final segmentation result. In the step, the missing blood vessel data in the initial pulmonary artery data and pulmonary vein data can be supplemented into the pulmonary artery data or pulmonary vein data according to the second pulmonary blood vessel enhancement result, so that abundant pulmonary artery data and pulmonary vein data can be obtained, and the purposes of fusion growth and automatic repair are achieved.
Fig. 8 is a graph of the prediction result after growth fusion provided in the embodiment of the present application, and the result is fused with the final segmentation result after growth and automatic repair, so that the detail of the prediction result after growth fusion is more abundant and accurate, and arteriovenous resolution is realized.
In the present example, an automatic pulmonary vessel separation function is provided to solve the problem that the pulmonary artery and the pulmonary vein are adhered together and cannot be separated after the pulmonary vessel enhancement. In the growth fusion step, the growth fusion and the automatic repair of the pulmonary artery and the pulmonary vein are realized through a specially-arranged fusion growth and automatic repair module of the pulmonary artery and the pulmonary vein. In particular, the blood vessel segmentation results of the deep learning step and the pulmonary vessel enhancement step can be fused, so that a richer and more accurate final segmentation result can be generated.
According to the segmentation method for the pulmonary blood vessels in the medical images, the pulmonary blood vessels can be rapidly and accurately segmented and split by combining the advantages of the deep learning segmentation algorithm and the region growing segmentation algorithm, the efficiency and the accuracy of segmentation of the pulmonary blood vessels are improved, and the postoperative life quality and the survival rate of patients can be further improved.
It should be noted that, according to the segmentation method for a pulmonary blood vessel in a medical image provided by the embodiment of the present application, the execution subject may be an electronic device, a segmentation apparatus for a pulmonary blood vessel in a medical image, or a control module in the segmentation apparatus for a pulmonary blood vessel in a medical image, for executing the segmentation method for a pulmonary blood vessel in a medical image. The embodiment of the present application takes the example that the segmentation apparatus for a pulmonary blood vessel in a medical image performs the segmentation method for a pulmonary blood vessel in a medical image, and the segmentation apparatus for a pulmonary blood vessel in a medical image provided by the embodiment of the present application is described.
The segmentation device for the pulmonary blood vessel in the medical image in the embodiment of the present application may be a device, or may be a component, an integrated circuit, or a chip in a terminal. The device can be mobile electronic equipment or non-mobile electronic equipment. By way of example, the mobile electronic device may be a mobile phone, a tablet computer, a notebook computer, a palm top computer, a vehicle-mounted electronic device, a wearable device, an Ultra Mobile Personal Computer (UMPC), a netbook or a Personal Digital Assistant (PDA), and the like, and the non-mobile electronic device may be a server, a Network Attached Storage (NAS), a Personal Computer (PC), a Television (TV), a teller machine or a self-service machine, and the like, and the embodiments of the present application are not particularly limited.
The segmentation device for pulmonary vessels in medical images in the embodiments of the present application may be a device having an operating system. The operating system may be an Android (Android) operating system, an ios operating system, or other possible operating systems, and embodiments of the present application are not limited specifically.
The segmentation device for the pulmonary blood vessel in the medical image provided by the embodiment of the application can realize each process realized by the segmentation method embodiment of the pulmonary blood vessel in the medical image shown in fig. 1-6, and is not repeated here.
Fig. 9 is a schematic structural diagram of a segmentation system for a pulmonary blood vessel in a medical image according to an embodiment of the present application, for implementing a segmentation method for a pulmonary blood vessel in a medical image as shown in fig. 1 to 8, where the segmentation apparatus 100 for a pulmonary blood vessel in a medical image includes:
fig. 9 is a schematic structural diagram of a segmentation system for a pulmonary blood vessel in a medical image according to an embodiment of the present application, for implementing a segmentation method for a pulmonary blood vessel in a medical image as shown in fig. 1 to 8, where the segmentation apparatus 100 for a pulmonary blood vessel in a medical image includes:
an image acquisition module 801, configured to acquire an original medical image of a pulmonary artery blood vessel and a pulmonary vein blood vessel to be segmented.
And the deep learning module 802 is configured to train the original medical image by using a deep learning network model, and output a pre-segmentation result.
The vessel enhancement module 803 is configured to extract vessel features of each pixel point in the original medical image, discriminate the vessel features according to a predetermined vessel feature discrimination function to obtain a vessel feature discrimination result, process the original medical image according to the vessel feature discrimination result, and output a first pulmonary vessel enhancement result.
And the growth fusion module 804 is used for performing region growth by taking the pulmonary artery and pulmonary vein region of the first pulmonary blood vessel enhancement result as an initial growth point, outputting a second pulmonary blood vessel enhancement result after the region growth, fusing the second pulmonary blood vessel enhancement result with the pre-segmentation result, and outputting a final segmentation result.
Optionally, the deep learning module 802 includes:
and the down-sampling sub-module is used for outputting down-sampling data after down-sampling processing is carried out on the original medical image, training the down-sampling data by adopting a neural network model and outputting a first pulmonary vascular network model with low resolution.
And the up-sampling sub-module is used for respectively obtaining low-resolution prediction results of the pulmonary artery blood vessel and the pulmonary vein blood vessel through the first pulmonary vascular network model, performing up-sampling processing on the prediction results and outputting up-sampled data.
And the first training submodule is used for training the up-sampling data and the original medical image by adopting a neural network model and outputting a second pulmonary vascular network model with high resolution.
And the second training submodule is used for training the original medical image by respectively adopting the first pulmonary vascular network model and the second pulmonary vascular network model and outputting a pre-segmentation result of pulmonary artery blood vessels and pulmonary vein blood vessels.
Optionally, the deep learning module 802 further includes:
and the region identification submodule is used for respectively identifying the connected regions of the pre-segmentation results of the pulmonary artery blood vessels and the pulmonary vein blood vessels after the pre-segmentation results of the pulmonary artery blood vessels and the pulmonary vein blood vessels are output, only reserving the maximum connected region and removing the broken branches.
Optionally, the image obtaining module 801 is configured to obtain an original medical image of a pulmonary artery blood vessel and a pulmonary vein blood vessel to be segmented and corresponding original segmentation data.
Optionally, the down-sampling sub-module is configured to output down-sampled data after performing down-sampling processing on the original medical image and the original segmentation data, where the down-sampled data is an image with a low resolution and annotation data.
Training the up-sampled data and the original medical image by adopting a neural network model, comprising the following steps:
optionally, the first training sub-module is configured to train the up-sampled data, the original medical image, and the original segmentation data by using a neural network model.
Optionally, the raw medical image is a CT image or an MRI image.
Optionally, the neural network model is a 3D-UNET neural network model.
Optionally, the vessel enhancement module 803 comprises:
and the calculating submodule is used for calculating the vessel characteristics of each pixel point in the original medical image.
And the matrix construction submodule is used for extracting and constructing a Hessian matrix according to the vessel characteristics of each pixel point.
And the function construction submodule is used for constructing the vessel feature discriminant function.
And the discrimination submodule is used for discriminating the vessel characteristics of each pixel point of the Hessian matrix according to the vessel characteristic discrimination function and outputting a vessel characteristic discrimination result.
And the blood vessel enhancer module is used for processing the original medical image according to the vascular feature discrimination result of each pixel point and outputting a first pulmonary blood vessel enhancement result.
Optionally, the computation submodule is used to construct a gaussian filter GσFor Gaussian filter GσCalculating a second derivative, and convolving the second derivative with each pixel point of the original medical image one by one to obtain corresponding vessel characteristics Ixx(σ)、Iyy(σ)、Izz(σ)、Ixy(σ)、Ixz(sigma) and Iyz(σ)。
Optionally, the matrix construction submodule is configured to extract and construct a Hessian matrix according to the second derivative in the three-dimensional neighborhood of each pixel:
optionally, the function construction submodule is used for constructing a vessel feature discriminant function Vs(λ)：
Wherein the content of the first and second substances,
H(λ)exp(h(λ1))|exp(h(λ2，λ3))|h(λ2，λ3)h(λ1) And h (λ)2，λ3) Is a tunable characteristic function.
Optionally, the distinguishing submodule is configured to distinguish, according to the vessel feature distinguishing function, the vessel feature of each pixel point of the Hessian matrix, distinguish that the pixel point that is positive is located in the pulmonary artery and pulmonary vein region, and output a vessel feature distinguishing result.
Optionally, the blood vessel enhancement sub-module is configured to process the original medical image according to the vessel feature determination result of each pixel point, and output a first pulmonary blood vessel enhancement result as follows:
optionally, the growth fusion module 804 includes:
and the extraction region submodule is used for extracting the lung parenchyma region of the first pulmonary blood vessel enhancement result through a preset threshold segmentation algorithm.
And the truncation submodule is used for truncating the first pulmonary vessel enhancement result by taking the lung parenchymal region as the interested region.
And the growth sub-module is used for performing region growth by taking the pulmonary artery and pulmonary vein region of the intercepted first pulmonary vein enhancement result as an initial growth point, if any neighborhood point in the pulmonary artery and pulmonary vein region does not belong to the pulmonary vein data but is contained in the pulmonary vein data of the lung parenchymal region, marking the neighborhood point as a pulmonary artery blood vessel and a pulmonary vein blood vessel, and the process is circularly performed until all neighborhood points of the pulmonary artery and pulmonary vein region do not meet the condition, and outputting a second pulmonary vein enhancement result after region growth.
And the fusion submodule is used for fusing the second pulmonary vessel enhancement result and the pre-segmentation result and outputting a final segmentation result.
According to the segmentation device for the pulmonary vessels in the medical images, after the original medical images of the pulmonary artery vessels and the pulmonary vein vessels to be segmented are obtained, the original medical images are trained by adopting a deep learning network model, and a pre-segmentation result is output; extracting vessel characteristics of each pixel point in the original medical image, judging the vessel characteristics according to a preset vessel characteristic judging function to obtain a vessel characteristic judging result, processing the original medical image according to the vessel characteristic judging result, and outputting a first pulmonary blood vessel enhancement result; and taking the pulmonary artery and pulmonary vein region of the first pulmonary blood vessel enhancement result as an initial growth point to perform region growth, outputting a second pulmonary blood vessel enhancement result after the region growth, fusing the second pulmonary blood vessel enhancement result with the pre-segmentation result, and outputting a final segmentation result. Therefore, the lung blood vessel segmentation method and the lung blood vessel segmentation device have the advantages that the lung blood vessel can be rapidly and accurately segmented and split by combining the advantages of the deep learning segmentation algorithm and the region growing segmentation algorithm, the efficiency and the accuracy of the segmentation of the lung blood vessel are improved, and the postoperative life quality and the survival rate of a patient can be further improved.
Optionally, as shown in fig. 10, an electronic device 500 is further provided in an embodiment of the present application, and includes a processor 501, a memory 502, and a program or an instruction stored in the memory 502 and executable on the processor 501, where the program or the instruction is executed by the processor 501 to implement each process of the above-mentioned method for segmenting a pulmonary blood vessel in a medical image, and can achieve the same technical effect, and in order to avoid repetition, the details are not repeated here.
It should be noted that the electronic devices in the embodiments of the present application include the mobile electronic devices and the non-mobile electronic devices described above.
Fig. 11 is a hardware structure diagram of an electronic device implementing an embodiment of the present application.
The electronic device 600 includes, but is not limited to: a radio frequency unit 601, a network module 602, an audio output unit 603, an input unit 604, a sensor 605, a display unit 606, a user input unit 607, an interface unit 608, a memory 609, a processor 610, and the like.
It should be understood that, in the embodiment of the present application, the radio frequency unit 601 may be used for receiving and sending signals during a message sending and receiving process or a call process, and specifically, receives downlink data from a base station and then processes the received downlink data to the processor 610; in addition, the uplink data is transmitted to the base station. In general, radio frequency unit 601 includes, but is not limited to, an antenna, at least one amplifier, a transceiver, a coupler, a low noise amplifier, a duplexer, and the like. Further, the radio frequency unit 601 may also communicate with a network and other devices through a wireless communication system.
The electronic device provides wireless broadband internet access to the user via the network module 602, such as assisting the user in sending and receiving e-mails, browsing web pages, and accessing streaming media.
The audio output unit 603 may convert audio data received by the radio frequency unit 601 or the network module 602 or stored in the memory 609 into an audio signal and output as sound. Also, the audio output unit 603 may also provide audio output related to a specific function performed by the electronic apparatus 600 (e.g., a call signal reception sound, a message reception sound, etc.). The audio output unit 603 includes a speaker, a buzzer, a receiver, and the like.
The input unit 604 is used to receive audio or video signals. It is to be understood that, in the embodiment of the present application, the input Unit 604 may include a Graphics Processing Unit (GPU) 6041 and a microphone 6042, and the Graphics Processing Unit 6041 processes image data of a still picture or a video obtained by an image capturing apparatus (such as a camera) in a video capturing mode or an image capturing mode.
The electronic device 600 also includes at least one sensor 605, such as a light sensor, motion sensor, and other sensors. Specifically, the light sensor includes an ambient light sensor that can adjust the brightness of the display panel 6061 according to the brightness of ambient light, and a proximity sensor that can turn off the display panel 6061 and/or the backlight when the electronic apparatus 600 is moved to the ear. As one type of motion sensor, an accelerometer sensor can detect the magnitude of acceleration in each direction (generally three axes), detect the magnitude and direction of gravity when stationary, and can be used to identify the posture of an electronic device (such as horizontal and vertical screen switching, related games, magnetometer posture calibration), and vibration identification related functions (such as pedometer, tapping); the sensors 605 may also include fingerprint sensors, pressure sensors, iris sensors, molecular sensors, gyroscopes, barometers, hygrometers, thermometers, infrared sensors, etc., which are not described in detail herein.
The display unit 606 is used to display information input by the user or information provided to the user. The Display unit 606 may include a Display panel 6061, and the Display panel 6061 may be configured by a Liquid Crystal Display (LCD), an Organic Light-Emitting Diode (OLED), or the like.
The user input unit 607 may be used to receive input numeric or character information and generate key signal inputs related to user settings and function control of the electronic device. Specifically, the user input unit 607 includes a touch panel 6071 and other input devices 6072. Touch panel 6071, also referred to as a touch screen, may collect touch operations by a user on or near it (e.g., operations by a user on or near touch panel 6071 using a finger, stylus, or any suitable object or accessory). The touch panel 6071 may include two parts of a touch detection device and a touch controller. Other input devices 6072 may include, but are not limited to, a physical keyboard, function keys (e.g., volume control keys, switch keys, etc.), a trackball, a mouse, and a joystick, which are not described in detail herein.
The interface unit 608 is an interface for connecting an external device to the electronic apparatus 600. For example, the external device may include a wired or wireless headset port, an external power supply (or battery charger) port, a wired or wireless data port, a memory card port, a port for connecting a device having an identification module, an audio input/output (I/O) port, a video I/O port, an earphone port, and the like. The interface unit 608 may be used to receive input (e.g., data information, power, etc.) from external devices and transmit the received input to one or more elements within the electronic device 600 or may be used to transmit data between the electronic device 600 and external devices.
The memory 609 may be used to store software programs as well as various data. The memory 609 may mainly include a program storage area and a data storage area, wherein the program storage area may store an operating system, an application program required by at least one function (such as a sound playing function, an image playing function, etc.), and the like; the storage data area may store data (such as audio data, a phonebook, etc.) created according to the use of the cellular phone, and the like. Further, the memory 609 may include high speed random access memory, and may also include non-volatile memory, such as at least one magnetic disk storage device, flash memory device, or other volatile solid state storage device.
The processor 610 is a control center of the electronic device, connects various parts of the whole electronic device by using various interfaces and lines, performs various functions of the electronic device and processes data by running or executing software programs and/or modules stored in the memory 609, and calling data stored in the memory 609, thereby performing overall monitoring of the electronic device. Processor 610 may include one or more processing units; preferably, the processor 610 may integrate an application processor, which mainly handles operating systems, user interfaces, application programs, etc., and a modem processor, which mainly handles wireless communications. It will be appreciated that the modem processor described above may not be integrated into the processor 610.
Those skilled in the art will appreciate that the electronic device 600 may further comprise a power source (e.g., a battery) for supplying power to the various components, and the power source may be logically connected to the processor 610 through a power management system, so as to implement functions of managing charging, discharging, and power consumption through the power management system. The electronic device structure shown in fig. 9 does not constitute a limitation of the electronic device, and the electronic device may include more or less components than those shown, or combine some components, or arrange different components, and thus, the description is not repeated here. In the embodiment of the present application, the electronic device includes, but is not limited to, a mobile phone, a tablet computer, a notebook computer, a palm computer, a vehicle-mounted terminal, a wearable device (e.g., a bracelet, glasses), a pedometer, and the like.
The embodiment of the present application further provides a readable storage medium, where a program or an instruction is stored on the readable storage medium, and when the program or the instruction is executed by a processor, the program or the instruction implements each process of the above method for segmenting a pulmonary blood vessel in a medical image, and can achieve the same technical effect, and in order to avoid repetition, details are not repeated here.
The processor is the processor in the electronic device described in the above embodiment. The readable storage medium includes a computer readable storage medium, such as a Read-Only Memory (ROM), a Random Access Memory (RAM), a magnetic disk or an optical disk, and so on.
The embodiment of the present application further provides a chip, where the chip includes a processor and a communication interface, the communication interface is coupled to the processor, and the processor is configured to run a program or an instruction to implement each process of the above-mentioned method for segmenting a pulmonary blood vessel in a medical image, and can achieve the same technical effect, and in order to avoid repetition, the description is omitted here.
It should be understood that the chips mentioned in the embodiments of the present application may also be referred to as system-on-chip, system-on-chip or system-on-chip, etc.
It should be noted that, in this document, the terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements but may include other elements not expressly listed or inherent to such process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprising an … …" does not exclude the presence of other like elements in a process, method, article, or apparatus that comprises the element. Further, it should be noted that the scope of the methods and apparatus of the embodiments of the present application is not limited to performing the functions in the order illustrated or discussed, but may include performing the functions in a substantially simultaneous manner or in a reverse order based on the functions involved, e.g., the methods described may be performed in an order different than that described, and various steps may be added, omitted, or combined. In addition, features described with reference to certain examples may be combined in other examples.
Through the above description of the embodiments, those skilled in the art will clearly understand that the method of the above embodiments can be implemented by software plus a necessary general hardware platform, and certainly can also be implemented by hardware, but in many cases, the former is a better implementation manner. Based on such understanding, the technical solutions of the present application may be embodied in the form of a software product, which is stored in a storage medium (such as ROM/RAM, magnetic disk, optical disk) and includes instructions for enabling a terminal (such as a mobile phone, a computer, a server, an air conditioner, or a network device) to execute the method according to the embodiments of the present application.
There are, of course, many other embodiments of the invention that will be apparent to those skilled in the art to which the invention pertains without departing from its spirit and scope, and it is intended that all such modifications and variations be considered within the scope of the appended claims.
Claims (10)
1. A method for segmenting pulmonary vessels in medical images, comprising the steps of:
acquiring original medical images of pulmonary artery blood vessels and pulmonary vein blood vessels to be segmented;
training the original medical image by adopting a deep learning network model, and outputting a pre-segmentation result by utilizing the model obtained by training;
extracting the vessel characteristics of each pixel point in the original medical image, judging the vessel characteristics according to a preset vessel characteristic judging function to obtain a vessel characteristic judging result, processing the original medical image according to the vessel characteristic judging result, and outputting a first pulmonary blood vessel enhancement result;
and performing region growth by taking the pulmonary artery and pulmonary vein region of the first pulmonary blood vessel enhancement result as an initial growth point, outputting a second pulmonary blood vessel enhancement result after the region growth, fusing the second pulmonary blood vessel enhancement result with the pre-segmentation result, and outputting a final segmentation result.
2. The method of claim 1, wherein the training of the raw medical image using the deep learning network model and outputting the pre-segmentation result comprises:
outputting downsampling data after downsampling the original medical image, training the downsampling data by adopting a neural network model, and outputting a first pulmonary vascular network model with low resolution;
respectively obtaining low-resolution prediction results of the pulmonary artery blood vessel and the pulmonary vein blood vessel through the first pulmonary vascular network model, performing up-sampling processing on the prediction results, and outputting up-sampled data;
training the up-sampling data and the original medical image by using the neural network model, and outputting a second pulmonary vascular network model with high resolution;
and training the original medical image by respectively adopting the first pulmonary vascular network model and the second pulmonary vascular network model, and outputting the pre-segmentation results of the pulmonary artery blood vessels and the pulmonary vein blood vessels.
3. The method of claim 2, wherein after outputting the pre-segmentation results of the pulmonary artery and vein vessels, further comprising:
and respectively carrying out connected region identification on the pre-segmentation results of the pulmonary artery blood vessel and the pulmonary vein blood vessel, only reserving a maximum connected region, and removing broken branches.
4. The method of claim 2, wherein the obtaining of the original medical image of the pulmonary artery and vein vessels to be segmented comprises:
acquiring original medical images of the pulmonary artery blood vessels and the pulmonary vein blood vessels to be segmented and corresponding original segmentation data;
the outputting of the downsampled data after the downsampling processing of the original medical image comprises:
outputting the down-sampling data after down-sampling processing is carried out on the original medical image and the original segmentation data, wherein the down-sampling data are low-resolution images and marking data;
the training the up-sampled data and the raw medical image using the neural network model includes:
training the up-sampled data, the raw medical image, and the raw segmentation data using the neural network model.
5. The method of claim 2, wherein the raw medical image is a CT image or an MRI image; or
The neural network model is a 3D-UNET neural network model.
6. The method according to claim 1, wherein the extracting the vessel feature of each pixel point in the original medical image, discriminating the vessel feature according to a predetermined vessel feature discrimination function to output a vessel feature discrimination result, processing the original medical image according to the vessel feature discrimination result, and outputting a first pulmonary vessel enhancement result, comprises:
calculating the vessel characteristics of each pixel point in the original medical image;
extracting and constructing a Hessian matrix according to the vessel characteristics of the pixel points;
constructing the vessel feature discriminant function;
judging the vessel characteristics of each pixel point of the Hessian matrix according to the vessel characteristic judging function, and outputting a vessel characteristic judging result;
and processing the original medical image according to the vessel feature judgment result of each pixel point, and outputting the first pulmonary vessel enhancement result.
7. The method of claim 6, wherein said calculating the vessel features for each pixel point in the original medical image comprises:
constructing a Gaussian filter GσFor the Gaussian filter GσCalculating a second derivative, and convolving the second derivative with each pixel point of the original medical image one by one to obtain the corresponding vessel feature
The extracting and constructing Hessian matrix according to the vessel characteristics of the pixel points comprises the following steps:
extracting and constructing the Hessian matrix according to the second derivative in the three-dimensional neighborhood of each pixel point:
the constructing the vessel feature discriminant function includes:
constructing the vessel feature discriminant function VS(λ)；
Wherein the content of the first and second substances,
H(λ)＝exp(h(λ1))|exp(h(λ2，λ3))|h(λ2，λ3)h(λ1) And h (λ)2，λ3) Is an adjustable characteristic function;
the distinguishing the vessel characteristics of each pixel point of the Hessian matrix according to the vessel characteristic distinguishing function and outputting the vessel characteristic distinguishing result includes:
the vessel features of the pixel points of the Hessian matrix are judged according to the vessel feature judgment function, the pixel points which are judged to be positive are located in the areas of the pulmonary artery and the pulmonary vein, and the vessel feature judgment result is output;
the processing the original medical image according to the vessel feature discrimination result of each pixel point and outputting the first pulmonary vessel enhancement result includes:
processing the original medical image according to the vessel feature discrimination result of each pixel point, and outputting the first pulmonary vessel enhancement result as follows:
8. the method of claim 1, wherein the region growing with the pulmonary artery and pulmonary vein region of the first pulmonary vessel enhancement result as an initial growing point, outputting a second pulmonary vessel enhancement result after the region growing, fusing the second pulmonary vessel enhancement result with the pre-segmentation result, and outputting a final segmentation result, comprises:
extracting a lung parenchymal region of the first pulmonary vascular enhancement result through a predetermined threshold segmentation algorithm;
intercepting the first pulmonary vessel enhancement result by taking the lung parenchymal region as a region of interest;
performing region growing by taking the intercepted pulmonary artery and pulmonary vein region of the first pulmonary artery and pulmonary vein enhancement result as an initial growing point, if any neighborhood point in the pulmonary artery and pulmonary vein region does not belong to pulmonary vessel data but is contained in the pulmonary vessel data of the lung parenchymal region, marking the neighborhood point as a pulmonary artery vessel and a pulmonary vein vessel, and performing the process circularly until all neighborhood points of the pulmonary artery and pulmonary vein region do not meet the condition, and outputting a second pulmonary vessel enhancement result after region growing;
and fusing the second pulmonary vessel enhancement result and the pre-segmentation result, and outputting the final segmentation result.
9. A segmentation apparatus for pulmonary blood vessels in medical images, for implementing the segmentation method for pulmonary blood vessels in medical images as claimed in any one of claims 1 to 8, the apparatus comprising:
the image acquisition module is used for acquiring original medical images of pulmonary artery blood vessels and pulmonary vein blood vessels to be segmented;
the deep learning module is used for training the original medical image by adopting a deep learning network model and outputting a pre-segmentation result;
the vessel enhancement module is used for extracting vessel features of each pixel point in the original medical image, distinguishing the vessel features according to a preset vessel feature distinguishing function to obtain a vessel feature distinguishing result, processing the original medical image according to the vessel feature distinguishing result and outputting a first pulmonary vessel enhancement result;
and the growth fusion module is used for performing region growth by taking the pulmonary artery and pulmonary vein region of the first pulmonary blood vessel enhancement result as an initial growth point, outputting a second pulmonary blood vessel enhancement result after the region growth, fusing the second pulmonary blood vessel enhancement result with the pre-segmentation result, and outputting a final segmentation result.
10. An electronic device comprising a storage medium, a processor and a computer program stored on the memory and executable on the processor, wherein the processor implements the method for segmenting pulmonary blood vessels in medical images according to any one of claims 1 to 8 when executing the computer program.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110935661.3A CN113538415A (en) | 2021-08-16 | 2021-08-16 | Segmentation method and device for pulmonary blood vessels in medical image and electronic equipment |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110935661.3A CN113538415A (en) | 2021-08-16 | 2021-08-16 | Segmentation method and device for pulmonary blood vessels in medical image and electronic equipment |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN113538415A true CN113538415A (en) | 2021-10-22 |
Family
ID=78091134
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110935661.3A Withdrawn CN113538415A (en) | 2021-08-16 | 2021-08-16 | Segmentation method and device for pulmonary blood vessels in medical image and electronic equipment |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113538415A (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114240851A (en) * | 2021-11-25 | 2022-03-25 | 深圳市旭东数字医学影像技术有限公司 | Lung arteriovenous extraction method and device based on CT image and electronic equipment |
| CN114972859A (en) * | 2022-05-19 | 2022-08-30 | 推想医疗科技股份有限公司 | Pixel classification method, model training method, device, equipment and medium |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103295195A (en) * | 2013-05-16 | 2013-09-11 | 深圳市旭东数字医学影像技术有限公司 | Vessel enhancement method and vessel enhancement system for low-contrast images |
| DE102018222606A1 (en) * | 2018-12-20 | 2020-06-25 | Siemens Healthcare Gmbh | Method and device for the detection of an anatomical feature of a blood vessel section |
| CN111340812A (en) * | 2020-02-20 | 2020-06-26 | 中南大学 | Interactive liver image segmentation method based on deep neural network |
| CN111932554A (en) * | 2020-07-31 | 2020-11-13 | 青岛海信医疗设备股份有限公司 | Pulmonary blood vessel segmentation method, device and storage medium |
-
2021
- 2021-08-16 CN CN202110935661.3A patent/CN113538415A/en not_active Withdrawn
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103295195A (en) * | 2013-05-16 | 2013-09-11 | 深圳市旭东数字医学影像技术有限公司 | Vessel enhancement method and vessel enhancement system for low-contrast images |
| DE102018222606A1 (en) * | 2018-12-20 | 2020-06-25 | Siemens Healthcare Gmbh | Method and device for the detection of an anatomical feature of a blood vessel section |
| CN111340812A (en) * | 2020-02-20 | 2020-06-26 | 中南大学 | Interactive liver image segmentation method based on deep neural network |
| CN111932554A (en) * | 2020-07-31 | 2020-11-13 | 青岛海信医疗设备股份有限公司 | Pulmonary blood vessel segmentation method, device and storage medium |
Non-Patent Citations (1)
| Title |
|---|
| 刘宴丽: "基于Hessian矩阵和区域生长的肝血管树的分割算法研究", 《计算机与现代化》 * |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114240851A (en) * | 2021-11-25 | 2022-03-25 | 深圳市旭东数字医学影像技术有限公司 | Lung arteriovenous extraction method and device based on CT image and electronic equipment |
| CN114972859A (en) * | 2022-05-19 | 2022-08-30 | 推想医疗科技股份有限公司 | Pixel classification method, model training method, device, equipment and medium |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109086709B (en) | Feature extraction model training method and device and storage medium | |
| CN108594997B (en) | Gesture skeleton construction method, device, equipment and storage medium | |
| CN109558512B (en) | Audio-based personalized recommendation method and device and mobile terminal | |
| CN107895369B (en) | Image classification method, device, storage medium and equipment | |
| WO2019011098A1 (en) | Unlocking control method and relevant product | |
| CN108288032B (en) | Action characteristic acquisition method, device and storage medium | |
| CN113538415A (en) | Segmentation method and device for pulmonary blood vessels in medical image and electronic equipment | |
| CN110991457B (en) | Two-dimensional code processing method and device, electronic equipment and storage medium | |
| JP2021531589A (en) | Motion recognition method, device and electronic device for target | |
| CN113706447A (en) | Method, device and equipment for recognizing midline of brain and storage medium | |
| CN113918767A (en) | Video clip positioning method, device, equipment and storage medium | |
| CN110705438A (en) | Gait recognition method, device, equipment and storage medium | |
| CN110991445A (en) | Method, device, equipment and medium for identifying vertically arranged characters | |
| CN108647566B (en) | Method and terminal for identifying skin characteristics | |
| CN107798662B (en) | Image processing method and mobile terminal | |
| CN111626035B (en) | Layout analysis method and electronic equipment | |
| CN110490953B (en) | Text-based image generation method, terminal device and medium | |
| CN109819331B (en) | Video call method, device and mobile terminal | |
| CN111402271A (en) | Image processing method and electronic equipment | |
| CN113721827B (en) | Floating keyboard display method, electronic equipment and storage medium | |
| CN111563838B (en) | Image processing method and electronic equipment | |
| WO2019001253A1 (en) | Method for iris liveness detection and related product | |
| CN114140655A (en) | Image classification method and device, storage medium and electronic equipment | |
| CN111145083B (en) | Image processing method, electronic equipment and computer readable storage medium | |
| CN113407774A (en) | Cover determining method and device, computer equipment and storage medium |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| WW01 | Invention patent application withdrawn after publication |
Application publication date: 20211022 |
|
| WW01 | Invention patent application withdrawn after publication |