CN105701799B - Divide pulmonary vascular method and apparatus from lung's mask image - Google Patents
Divide pulmonary vascular method and apparatus from lung's mask image Download PDFInfo
- Publication number
- CN105701799B CN105701799B CN201511030492.XA CN201511030492A CN105701799B CN 105701799 B CN105701799 B CN 105701799B CN 201511030492 A CN201511030492 A CN 201511030492A CN 105701799 B CN105701799 B CN 105701799B
- Authority
- CN
- China
- Prior art keywords
- lung
- voxel
- vessel
- image
- blood vessel
- 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.)
- Active
Links
- 210000004072 lung Anatomy 0.000 title claims abstract description 213
- 230000002685 pulmonary effect Effects 0.000 title claims abstract description 123
- 238000000034 method Methods 0.000 title claims abstract description 53
- 230000002792 vascular Effects 0.000 title claims abstract description 27
- 210000004204 blood vessel Anatomy 0.000 claims abstract description 222
- 230000004044 response Effects 0.000 claims abstract description 75
- 230000004927 fusion Effects 0.000 claims abstract description 5
- 210000001147 pulmonary artery Anatomy 0.000 claims description 46
- 230000011218 segmentation Effects 0.000 claims description 21
- 238000013507 mapping Methods 0.000 claims description 15
- 238000005457 optimization Methods 0.000 claims description 11
- 238000009499 grossing Methods 0.000 claims description 8
- 238000007781 pre-processing Methods 0.000 claims description 7
- 238000004364 calculation method Methods 0.000 claims description 3
- 230000008569 process Effects 0.000 claims description 3
- 238000012545 processing Methods 0.000 claims description 3
- 239000003550 marker Substances 0.000 claims description 2
- 239000008280 blood Substances 0.000 abstract description 3
- 210000004369 blood Anatomy 0.000 abstract description 3
- 230000006870 function Effects 0.000 description 69
- 238000010586 diagram Methods 0.000 description 14
- 230000000694 effects Effects 0.000 description 11
- 238000002591 computed tomography Methods 0.000 description 6
- 238000001514 detection method Methods 0.000 description 5
- 238000009826 distribution Methods 0.000 description 5
- 238000004458 analytical method Methods 0.000 description 4
- 239000011159 matrix material Substances 0.000 description 3
- 238000003860 storage Methods 0.000 description 3
- 230000003190 augmentative effect Effects 0.000 description 2
- 238000003745 diagnosis Methods 0.000 description 2
- 230000003902 lesion Effects 0.000 description 2
- 239000000463 material Substances 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 230000035945 sensitivity Effects 0.000 description 2
- 238000000638 solvent extraction Methods 0.000 description 2
- 238000012800 visualization Methods 0.000 description 2
- 206010058467 Lung neoplasm malignant Diseases 0.000 description 1
- 208000010378 Pulmonary Embolism Diseases 0.000 description 1
- 206010056342 Pulmonary mass Diseases 0.000 description 1
- 230000004075 alteration Effects 0.000 description 1
- 230000003416 augmentation Effects 0.000 description 1
- 238000005520 cutting process Methods 0.000 description 1
- 238000002924 energy minimization method Methods 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 238000000605 extraction Methods 0.000 description 1
- 238000001914 filtration Methods 0.000 description 1
- PCHJSUWPFVWCPO-UHFFFAOYSA-N gold Chemical compound [Au] PCHJSUWPFVWCPO-UHFFFAOYSA-N 0.000 description 1
- 238000003709 image segmentation Methods 0.000 description 1
- 238000003384 imaging method Methods 0.000 description 1
- 230000003993 interaction Effects 0.000 description 1
- 201000005202 lung cancer Diseases 0.000 description 1
- 208000020816 lung neoplasm Diseases 0.000 description 1
- 210000003097 mucus Anatomy 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
- 210000003437 trachea Anatomy 0.000 description 1
Classifications
-
- 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
- G06T2207/20212—Image combination
- G06T2207/20221—Image fusion; Image merging
-
- 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
- Apparatus For Radiation Diagnosis (AREA)
Abstract
The present invention proposes that one kind dividing pulmonary vascular method and apparatus from lung's mask image, this includes calculating the response of blood vessel similarity function of each voxel in blood vessel range scale under different blood vessel scale in lung mask image in multiple voxels from pulmonary vascular method is divided in lung's mask image, and the pulmonary vascular image using response as each voxel under different blood vessel scale；The target blood scale of each voxel in lung's mask image, and the corresponding pulmonary vascular image of target blood scale for obtaining each voxel are obtained, using the target pulmonary vascular image as each voxel；The target pulmonary vascular image of multiple voxels is merged, and obtains the target pulmonary vascular image after fusion.It can effectively be promoted through the invention and divide pulmonary vascular accuracy from lung's mask image, and promoted and divide pulmonary vascular effect from lung's mask image.
Description
Technical Field
The invention relates to the technical field of segmentation of pulmonary blood vessels, in particular to a method and a device for segmenting pulmonary blood vessels from a pulmonary mask image.
Background
The accurate lung vessel segmentation is an important step in a lung computer aided detection and diagnosis (CAD) system, lung vessel tissues need to be extracted firstly in the automatic identification and detection of the pulmonary embolism to narrow the range of lesion detection, the lung vessel interference is removed in the early detection and diagnosis of the lung cancer to reduce the false positive of the lung nodule detection, and the distribution of the lung vessels is used for guiding the lung lobe segmentation in the operation navigation. However, the pulmonary vessels present a complex tree structure with 23 branches in the lung and varying tube diameters from 20 microns to 15 mm. On the CT image, the pulmonary vessels generally show high-density images due to the internal filling of blood, but the gray scale distribution is not uniform, and particularly, the small vessels are greatly affected by partial volume effect. Mucus filled trachea around pulmonary vessels, pulmonary nodules and some dense lesions all interfere with the accuracy of pulmonary vessel extraction. Geometric Models of pulmonary vessels (Geometry Models) refer to a priori knowledge of the shape characteristics of blood vessels in elongated, tubular, tree-like distributions. The Hessian matrix characteristic analysis method can effectively identify spherical objects, cylindrical objects and sheet objects, and is a typical method for segmenting blood vessels by using a geometric model of the blood vessels.
In the multiscale vascular filter based on Hessian matrix eigenvalue analysis in the related technology, the maximum value of all scale vascular similarity functions is used as the result of scale selection. In this way, the response is weak at the bifurcation junction of the pulmonary vessels, and vessel discontinuity is easily caused after thresholding segmentation.
Disclosure of Invention
The present invention is directed to solving, at least to some extent, one of the technical problems in the related art.
Therefore, an object of the present invention is to provide a method for segmenting lung blood vessels from a lung mask image, which can effectively improve the accuracy of segmenting lung blood vessels from the lung mask image and improve the effect of segmenting lung blood vessels from the lung mask image.
Another objective of the present invention is to provide an apparatus for segmenting pulmonary blood vessels from a mask image of the lungs.
In order to achieve the above object, a method for segmenting lung blood vessels from a lung mask image according to an embodiment of the present invention includes: calculating a response value of a vessel similarity function of each voxel in a plurality of voxels in the lung mask image under different vessel scales in a vessel scale range, and taking the response value as the lung vessel image of each voxel under different vessel scales; acquiring a target blood vessel scale of each voxel in the lung mask image, and acquiring a lung blood vessel image corresponding to the target blood vessel scale of each voxel to serve as the target lung blood vessel image of each voxel; and fusing the target pulmonary blood vessel images of the multiple voxels, and acquiring the fused target pulmonary blood vessel images.
According to the method for segmenting the pulmonary blood vessels from the pulmonary mask image, provided by the embodiment of the first aspect of the invention, the accuracy of segmenting the pulmonary blood vessels from the pulmonary mask image can be effectively improved, and the effect of segmenting the pulmonary blood vessels from the pulmonary mask image can be improved by calculating the response value of the blood vessel similarity function of each voxel in the pulmonary mask image under different blood vessel scales in the blood vessel scale range, taking the response value as the pulmonary blood vessel image, acquiring the target pulmonary blood vessel image of each voxel in the pulmonary mask image, and fusing the target pulmonary blood vessel images of a plurality of voxels.
In order to achieve the above object, a device for segmenting lung blood vessels from a lung mask image according to a second aspect of the present invention comprises: the calculation module is used for calculating a response value of a blood vessel similarity function of each voxel in a plurality of voxels in the lung mask image under different blood vessel scales in a blood vessel scale range, and taking the response value as the lung blood vessel image of each voxel under different blood vessel scales; an obtaining module, configured to obtain a target blood vessel scale of each voxel in the lung mask image, and obtain a lung blood vessel image corresponding to the target blood vessel scale of each voxel, as a target lung blood vessel image of each voxel; and the fusion module is used for fusing the target pulmonary blood vessel images of the voxels and acquiring the fused target pulmonary blood vessel images.
According to the device for segmenting the pulmonary blood vessels from the pulmonary mask image, which is provided by the embodiment of the second aspect of the invention, the accuracy of segmenting the pulmonary blood vessels from the pulmonary mask image can be effectively improved, and the effect of segmenting the pulmonary blood vessels from the pulmonary mask image can be improved by calculating the response value of the blood vessel similarity function of each voxel in the pulmonary mask image under different blood vessel scales in the blood vessel scale range, taking the response value as the pulmonary blood vessel image, acquiring the target pulmonary blood vessel image of each voxel in the pulmonary mask image, and fusing the target pulmonary blood vessel images of a plurality of voxels.
Additional aspects and advantages of the invention will be set forth in part in the description which follows and, in part, will be obvious from the description, or may be learned by practice of the invention.
Drawings
The foregoing and/or additional aspects and advantages of the present invention will become apparent and readily appreciated from the following description of the embodiments, taken in conjunction with the accompanying drawings of which:
FIG. 1 is a flowchart illustrating a method for segmenting pulmonary blood vessels from a mask image of a lung according to an embodiment of the present invention;
FIG. 2 is a flowchart illustrating a method for segmenting lung vessels from a mask image of a lung according to another embodiment of the present invention;
FIG. 3a is a schematic diagram of an original image of a mask image of a local lung;
FIG. 3b is a schematic diagram of a filtered pulmonary vessel image at a vessel dimension σ of 0.5 mm;
FIG. 3c is a schematic diagram of a filtered pulmonary vessel image at a vessel dimension σ of 0.75 mm;
FIG. 4 is a three-dimensional visualization effect diagram of the pulmonary vessel segmentation in the embodiment of the present invention;
FIG. 5 is a schematic structural diagram of an apparatus for segmenting pulmonary blood vessels from a mask image of a lung according to another embodiment of the present invention;
FIG. 6 is a schematic structural diagram of an apparatus for segmenting pulmonary blood vessels from a mask image of a lung according to another embodiment of the present invention;
FIG. 7 is a schematic structural diagram of an apparatus for segmenting pulmonary blood vessels from a mask image of a lung according to another embodiment of the present invention;
fig. 8 is a schematic structural diagram of an apparatus for segmenting pulmonary blood vessels from a mask image of a lung according to another embodiment of the present invention.
Detailed Description
Reference will now be made in detail to embodiments of the present invention, examples of which are illustrated in the accompanying drawings, wherein like or similar reference numerals refer to the same or similar elements or elements having the same or similar function throughout. The embodiments described below with reference to the accompanying drawings are illustrative only for the purpose of explaining the present invention, and are not to be construed as limiting the present invention. On the contrary, the embodiments of the invention include all changes, modifications and equivalents coming within the spirit and terms of the claims appended hereto.
Fig. 1 is a flowchart illustrating a method for segmenting lung blood vessels from a lung mask image according to an embodiment of the present invention, the method for segmenting lung blood vessels from a lung mask image includes:
s101: and calculating a response value of the vessel similarity function of each voxel in a plurality of voxels in the lung mask image under different vessel scales in the vessel scale range, and taking the response value as the lung vessel image of each voxel under different vessel scales.
Alternatively, the lung mask image may be obtained by preprocessing the lung image by using the prior art, which is not described herein again.
In the embodiment of the invention, the lung image can be a three-dimensional CT image of the lung, a lung mask image is obtained by preprocessing the lung image, and the pulmonary blood vessel is segmented according to a plurality of voxels in the lung mask image, so that the interference of voxels irrelevant to the pulmonary blood vessel in the lung image can be avoided, and the efficiency of segmenting the pulmonary blood vessel is improved.
The three-dimensional CT image of the lung is an image of the lung of a human body scanned by Computed Tomography (CT).
The voxel is a volume element for short, and is the minimum unit of digital data on three-dimensional space segmentation, and the voxel is used in the fields of three-dimensional imaging, scientific data, medical images and the like.
In an embodiment of the invention, a voxel refers to a volume element of a pulmonary vessel in a mask image of the lung.
In an embodiment of the present invention, the vessel scale range is a possible scale of the pulmonary vessel of the position where each voxel in the pulmonary mask image is located, and the vessel scale range may be pre-specified by the user, for example, the vessel scale range is σ e { σ ∈ [1,...,σi,...,σkI belongs to 1-k, the value of k is 1-N, and N is a positive integer.
Alternatively, the response value of the vessel similarity function of each voxel at each vessel scale in the vessel scale range may be calculated, and the response value may be used as the pulmonary vessel image of each voxel at each vessel scale.
For example, if the vessel scale range is (a, B, C), it is necessary to calculate the response values of the vessel similarity function of voxel 1 in three vessel scales of a, B, and C, and calculate the response values of the vessel similarity function of voxel 2 in three vessel scales of a, B, and C, and so on, calculate the response values of the vessel similarity function of each voxel in the lung mask image in three vessel scales of a, B, and C.
The vessel similarity function may also be referred to as a vessel filter, for example, a Frangi filter, or a fractional order derivative filter, which is not limited by the embodiments of the present invention.
S102: and acquiring the target blood vessel scale of each voxel in the lung mask image, and acquiring the lung blood vessel image corresponding to the target blood vessel scale of each voxel to serve as the target lung blood vessel image of each voxel.
Alternatively, the target vessel dimension for each voxel in the lung mask image may be obtained according to a Markov Random Field (MRF) based optimization method.
In an embodiment of the invention, the target vessel scale for each voxel, i.e. the vessel scale selection, is obtained for each voxel over a vessel scale range σ e { σ }1,...,σi,...,σkSelecting a blood vessel scale closest to the true scale of the pulmonary vessel where the voxel is located as a target blood vessel scale, wherein i is 1-k, the value of k is 1-N, and N is a positive integer.
The blood vessel scale selection is carried out on each voxel in the lung mask image based on a multi-label MRF optimization method, the blood vessel scale selection problem is regarded as a graph partitioning problem based on the multi-label MRF optimization method, and the blood vessel scale selection problem can be solved by a graph theory method (such as minimum partitioning).
In the embodiment of the invention, the multi-label-based MRF optimization method is a graph theory-based energy minimization method, and the main idea is to map voxels of a lung mask image to vertexes of a weighted graph, map the relation between adjacent voxels to edges of the weighted graph, and map the similarity between adjacent voxels to the weights of the edges to obtain the weighted graph of the lung mask image, establish an energy model of the weighted graph, and complete the cutting of the graph by minimizing the energy model.
In the embodiment of the present invention, the target blood vessel scale is a blood vessel scale in a blood vessel scale range specified by a user in advance, the target blood vessel scale of each voxel is closest to the real scale of the pulmonary vessel at the position of the current voxel, each voxel in the pulmonary mask image corresponds to one target blood vessel scale, and different voxels in the pulmonary mask image may correspond to different or the same target blood vessel scale.
In the embodiment of the present invention, the target pulmonary blood vessel image of each voxel corresponds to the target blood vessel scale of the voxel, and the target pulmonary blood vessel image corresponding to the target blood vessel scale of each voxel is obtained in step S101.
S103: and fusing the target pulmonary vessel images of the multiple voxels, and acquiring the fused target pulmonary vessel image.
In an embodiment of the present invention, the target pulmonary blood vessel image of each of a plurality of voxels in the pulmonary mask image may be fused to obtain a fused target pulmonary blood vessel image.
For example, if the lung mask image includes voxel 1, voxel 2, and voxel 3, and the target pulmonary blood vessel image of voxel 1 is P1, the target pulmonary blood vessel image of voxel 2 is P2, and the target pulmonary blood vessel image of voxel 3 is P3, the target pulmonary blood vessel image P1, the target pulmonary blood vessel image P2, and the target pulmonary blood vessel image P3 are fused to obtain the fused target pulmonary blood vessel image P, and so on.
In this embodiment, by calculating a response value of a blood vessel similarity function of each voxel in the lung mask image at different blood vessel scales in a blood vessel scale range, taking the response value as the lung blood vessel image, obtaining a target lung blood vessel image of each voxel in the lung mask image, and fusing the target lung blood vessel images of a plurality of voxels, the accuracy of segmenting lung blood vessels from the lung mask image can be effectively improved, and the effect of segmenting lung blood vessels from the lung mask image is improved.
Fig. 2 is a flowchart illustrating a method for segmenting lung blood vessels from a lung mask image according to another embodiment of the present invention, the method for segmenting lung blood vessels from a lung mask image includes:
s201: the lung image is preprocessed to obtain a mask image of the lung.
Alternatively, the lung image may be preprocessed using known techniques to obtain a mask image of the lung.
In the embodiment of the invention, the lung image can be a three-dimensional CT image of the lung, a lung mask image is obtained by preprocessing the lung image, and the pulmonary blood vessel is segmented according to a plurality of voxels in the lung mask image, so that the interference of voxels irrelevant to the pulmonary blood vessel in the lung image can be avoided, and the efficiency of segmenting the pulmonary blood vessel is improved.
S202: and calculating a response value of the vessel similarity function of each voxel in a plurality of voxels in the lung mask image under different vessel scales in the vessel scale range, and taking the response value as the lung vessel image of each voxel under different vessel scales.
Alternatively, the position parameter p of a gray-scale function of a mask image of the lung at a certain voxel can be determined by the prior art0And analyzing the neighborhood taylor expansion to obtain a Hessian matrix eigenvalue, calculating a response value of a vessel similarity function of each voxel in different vessel scales in the vessel scale range according to the vessel similarity function, and using the response value as a pulmonary vessel image of each voxel in different vessel scales, where the vessel similarity function may also be referred to as a vessel filter, and the vessel filter is, for example, a Frangi filter or a fractional order differential filter, which is not limited in this embodiment of the present invention.
Fig. 3 is a diagram of the filtering effect of pulmonary blood vessel images under different blood vessel scales, where fig. 3a is a schematic diagram of an original image of a local pulmonary mask image, fig. 3b is a schematic diagram of a filtered pulmonary blood vessel image when a blood vessel scale σ is 0.5mm, and fig. 3c is a schematic diagram of a filtered pulmonary blood vessel image when the blood vessel scale σ is 0.75mm, and it can be seen from fig. 3b and fig. 3c that when a smaller blood vessel scale is adopted, a blood vessel similarity function can detect a thinner pulmonary blood vessel, and when a larger blood vessel scale is adopted, a blood vessel similarity function can detect a thicker pulmonary blood vessel, and within a certain blood vessel scale range, the blood vessel similarity function has response values for pulmonary blood vessels with different thicknesses.
S203: the voxels of the lung mask image are mapped to the vertexes of the weighted graph, the relationship between the adjacent voxels is mapped to the edges of the weighted graph, and the similarity between the adjacent voxels is mapped to the weight of the edges, so that the weighted graph of the lung mask image is obtained.
In an embodiment of the invention, the lung mask image I (x, y, z) may be mapped to a weighted graph G (V, E), where the vertex V ∈ V in the weighted graph corresponds to a voxel in the lung mask image I (x, y, z), and the edge in the weighted graph isThe adjacency of any two voxels (p, q) is identified, and the weights of the edges identify the differences between the voxels, such as gray scale, location, etc.
The lung mask image is mapped into a weighted graph, a target blood vessel scale selection problem is regarded as a division problem of the weighted graph, and the local region attribute of the lung mask image is subjected to space constraint, so that the globally optimized blood vessel scale distribution can be obtained.
S204: and establishing an energy model of the weighted graph, and acquiring a mark when the energy model obtains the energy minimum value.
In an embodiment of the invention, the expression of the energy model is as follows:
wherein,in order to be able to perform the data item,for the smoothing term, η is the weight for adjusting the voxel's own energy and the connected energy of neighboring voxels during segmentation, fpThe vector f corresponds to the segmentation result of the lung mask image for the label assigned to the voxel with position parameter p in the lung mask image.
In an embodiment of the invention, the data items are arranged as:
wherein σiFor the ith vessel scale in the range of vessel scales,among the response values of the vascular similarity function of the voxels having the position parameter p, the response value of the vascular similarity function which is the largest, j being 0, …, k, k having values from 1 to N, N being a positive integer, L (p, σ), is the response value of the vascular similarity function which is the largest, j being 0i) For the voxel with position parameter p at vessel scale σiThe response value of the vessel similarity function of (a);
the smoothing term is set as:
wherein f ispFor the label assigned to the voxel with position parameter p in the mask image of the lungs, fqIs a label assigned to a voxel with position parameter q in the lung mask image.
In the energy model, the smoothing termDefining the interaction between adjacent voxels, which may be a function of the labelOr may be a function of the underlying dataAlternatively, it may be a function of the label and the underlying dataSince neighboring voxels are usually assigned similar, or equal, labels, the usual smoothing terms have a labeled function, e.g. | fp-fq|，(fp-fq)2，min{T,|fp-fq|}，min{T,(fp-fq)2Etc., functions of the underlying data such as exp (- (I (p) -I (q))2) At min { T, | fp-fq|}，min{T,(fp-fq)2In the present invention, T is a threshold value, which is selected by a user according to experience, and this is not limited in the embodiment of the present invention.
Specifically, according to the response value of the blood vessel similarity function of the voxel with the position parameter p in the lung mask image under a plurality of blood vessel scales, the response value of the blood vessel similarity function is the largest, and the response value of the voxel with the position parameter p under the blood vessel scale sigmaiObtaining response value of vessel similarity function and vessel scale sigmaiCorresponding data item and according to the label f assigned to the voxel with position parameter p in the lung mask imagepAnd a label f assigned to a voxel with a position parameter qqAnd acquiring a smooth term, and calculating E (f) to acquire a mark when the energy model acquires the energy minimum value, wherein i is an index of a blood vessel scale, i is 0, …, k, k takes values from 1 to N, and N is a positive integer.
S205: and taking the blood vessel scale corresponding to the mark as the target blood vessel scale of each voxel in the lung mask image.
In an embodiment of the invention, the target vessel scale for each voxel, i.e. the vessel scale selection, is obtained for each voxel over a vessel scale range σ e { σ }1,...,σi,...,σkSelecting a blood vessel scale closest to the true scale of the pulmonary vessel where the voxel is located as a target blood vessel scale, wherein i is 1-k, the value of k is 1-N, and N is a positive integer.
The multi-label-based MRF optimization method is to solve the energy minimum value based on the maximum flow minimum cut method, so that the cost of a cut set of a graph is exactly equal to a given energy model. The maximum flow minimum cut method mainly includes two categories: the embodiment of the present invention provides a Pushrelabel (Pushrelabel) method and an augmented path (augmentation paths) method, and an improved augmented path method proposed by Boykov can approximate an optimal solution of a multi-label problem within polynomial time, which is not limited in the embodiment of the present invention.
In the embodiment of the present invention, the energy model E (f) defined in step S204, solving the minimum value of the energy model E (f) is converted into solving the minimum cut set of the graph G (V, E). The segmentation result determined by the binary vector corresponding to the minimum cut set is the lung mask image segmentation result.
In the embodiment of the invention, the vessel scale of the energy minimum value of the energy model of the lung mask image is obtained in different vessel scales by obtaining each voxel based on the multi-label MRF optimization method to obtain the target lung vessel image of each voxel, and modeling is performed aiming at the relation of the neighboring voxels, so that the response value of the vessel similarity function of the neighboring voxels with similar structures and the target vessel scale are possibly different, and the accuracy of the segmentation of the pulmonary vessels is effectively improved.
S206: and acquiring a pulmonary vessel image corresponding to the target vessel scale of each voxel to serve as the target pulmonary vessel image of each voxel.
In the embodiment of the invention, the pulmonary blood vessel image corresponding to the target blood vessel scale can be extracted from the pulmonary blood vessel images of each voxel under different blood vessel scales, so as to obtain the target pulmonary blood vessel image of each voxel.
S207: and fusing the target pulmonary vessel images of the multiple voxels, and acquiring the fused target pulmonary vessel image.
Specifically, the target pulmonary blood vessel image of each voxel extracted in step S206 is fused to obtain a fused target pulmonary blood vessel image.
S208: and judging whether the response value of the blood vessel similarity function of each voxel in the fused target pulmonary vessel image is smaller than a preset threshold value, if so, executing the step S209, otherwise, executing the step S210.
In the step, the fused target pulmonary vessel image is processed based on threshold value and connectivity analysis according to the response threshold value of the vessel similarity function, and when the response values of the vessel similarity functions of low-density voxels and pulmonary vessels are low, a lower response threshold value of the vessel similarity function is adopted, so that more thin-branch pulmonary vessels can be segmented.
The preset threshold is obtained through calculation of a second function, wherein the second function is as follows:
where p is the location parameter for each voxel, σpTarget vessel scale, T (σ), for a voxel with a location parameter pp) Response threshold, t, for vessel similarity functionσIs a vessel scale threshold value, n is the vessel scale in the vessel scale range is smaller than tσI is the index of the vessel scale, i is 0, …, k, k takes values from 1 to N, N is a positive integer, T is a positive integerminAnd TmaxTubular feature response intensity threshold, T, specified for userminFor example, may be set to 0.05, TmaxFor example, it may be set to 0.15.
S209: the first function value of the voxel is set to 0.
Wherein the first function is:
where p is the location parameter for each voxel, σpTarget vessel dimension, V, for a voxel with a position parameter pseg(L(p,σp),σp) Is a discriminant function for pulmonary vessel segmentation, L (p, σ)p) For a voxel with position parameter p at a target vessel scale σpResponse value of vessel similarity function of (1), T (σ)p) Is a preset threshold.
S210: the first function value of a voxel is set to 1.
In an embodiment of the present invention, the lung vessel segmentation results are compared to the gold standard, as shown in table 1 below, and the accuracy of the segmentation is described by three quantitative assessment indicators: area under ROC (Receiver Operating characteristics) curve (AZ), Specificity (Specificity) and Sensitivity (Sensitivity).
TABLE 1
The segmentation effect of the pulmonary vessels in the embodiment of the invention is shown in fig. 4, fig. 4 is a three-dimensional visualization effect diagram of the segmentation of the pulmonary vessels in the embodiment of the invention, and fig. 4 shows that the invention can better segment the detailed parts in the image, is not easily interfered by small noise, and presents the anatomical features of fewer vessels in the lung fissure part.
In the embodiment, the lung image is preprocessed to obtain the lung mask image, and the pulmonary blood vessel is segmented according to a plurality of voxels in the lung mask image, so that the interference of voxels irrelevant to the pulmonary blood vessel in the lung image can be avoided, and the efficiency of segmenting the pulmonary blood vessel is improved. The method comprises the steps of obtaining an energy model of each voxel in different vessel scales by using a multi-label MRF optimization method to obtain a vessel scale with the minimum energy value, obtaining a target pulmonary vessel image of each voxel, mapping the pulmonary mask image into a weighted graph, regarding the target vessel scale selection problem as a division problem of the weighted graph, performing space constraint on the local region attribute of the pulmonary mask image, and obtaining the globally optimized vessel scale distribution. And performing threshold-based and connectivity analysis on the fused target pulmonary vessel image according to the response threshold of the vessel similarity function, and adopting a lower response threshold of the vessel similarity function when the response values of the vessel similarity functions of low-density voxels and pulmonary vessels are low, so that more thin-branch pulmonary vessels can be segmented. By calculating the response value of the blood vessel similarity function of each voxel in the lung mask image in different blood vessel scales in the blood vessel scale range, taking the response value as the lung blood vessel image, acquiring the target lung blood vessel image of each voxel in the lung mask image, and fusing the target lung blood vessel images of a plurality of voxels, the accuracy of segmenting the lung blood vessels from the lung mask image can be effectively improved, and the effect of segmenting the lung blood vessels from the lung mask image is improved.
Fig. 5 is a schematic structural diagram of an apparatus for segmenting pulmonary vessels from a pulmonary mask image according to another embodiment of the present invention, in which the apparatus 50 for segmenting pulmonary vessels from a pulmonary mask image includes a calculating module 501, configured to calculate a response value of a vessel similarity function of each of a plurality of voxels in the pulmonary mask image at a different vessel scale in a vessel scale range, and use the response value as a pulmonary vessel image of each voxel at the different vessel scale; an obtaining module 502, configured to obtain a target blood vessel scale of each voxel in the lung mask image, and obtain a lung blood vessel image corresponding to the target blood vessel scale of each voxel, as a target lung blood vessel image of each voxel; and a fusion module 503, configured to fuse the target pulmonary blood vessel images of multiple voxels, and obtain a fused target pulmonary blood vessel image.
A calculating module 501, configured to calculate a response value of a blood vessel similarity function of each voxel in a plurality of voxels in the lung mask image at different blood vessel scales in a blood vessel scale range, and use the response value as the lung blood vessel image of each voxel at different blood vessel scales.
An obtaining module 502, configured to obtain a target blood vessel scale of each voxel in the lung mask image, and obtain a lung blood vessel image corresponding to the target blood vessel scale of each voxel, as a target lung blood vessel image of each voxel.
Optionally, as shown in fig. 6, the obtaining module 502 specifically includes:
the target blood vessel scale obtaining sub-module 5021 is used for obtaining the target blood vessel scale of each voxel in the lung mask image according to a multi-marker Markov random field optimization method.
Optionally, as shown in fig. 7, the target vessel dimension acquisition submodule 5021 includes:
the weighted map mapping unit 50211 is configured to perform weighted map mapping on the lung mask image to obtain a weighted map of the lung mask image.
Optionally, the weighted graph mapping unit 50211 is specifically configured to:
the voxels of the lung mask image are mapped to the vertexes of the weighted graph, the relationship between the adjacent voxels is mapped to the edges of the weighted graph, and the similarity between the adjacent voxels is mapped to the weight of the edges, so that the weighted graph of the lung mask image is obtained.
The energy model building unit 50212 is used for building an energy model of the weighted graph and obtaining a mark when the energy model obtains the energy minimum value.
Optionally, the energy model of the weighted graph is:
wherein,in order to be able to perform the data item,for the smoothing term, η is the weight for adjusting the voxel's own energy and the connected energy of neighboring voxels during segmentation, fpThe vector f corresponds to the segmentation result of the lung mask image for the label assigned to the voxel with position parameter p in the lung mask image.
Optionally, the data items are arranged as:
wherein σiFor the ith vessel scale in the range of vessel scales,among the response values of the vascular similarity function of the voxels having the position parameter p, the response value of the vascular similarity function which is the largest, j being 0, …, k, k having values from 1 to N, N being a positive integer, L (p, σ), is the response value of the vascular similarity function which is the largest, j being 0i) For the voxel with position parameter p at vessel scale σiThe response value of the vessel similarity function of (a);
the smoothing term is set as:
wherein f ispFor the label assigned to the voxel with position parameter p in the mask image of the lungs, fqIs a label assigned to a voxel with position parameter q in the lung mask image.
An obtaining unit 50213 is used for taking the corresponding vessel scale of the label as the target vessel scale of each voxel in the lung mask image.
The target pulmonary blood vessel image obtaining sub-module 5022 is configured to obtain a pulmonary blood vessel image corresponding to the target blood vessel scale of each voxel, and use the pulmonary blood vessel image as the target pulmonary blood vessel image of each voxel.
And a fusion module 503, configured to fuse the target pulmonary blood vessel images of multiple voxels, and obtain a fused target pulmonary blood vessel image.
Optionally, as shown in fig. 8, the apparatus 50 for segmenting pulmonary blood vessels from a pulmonary mask image further includes:
a comparing module 504, configured to compare the response value of the blood vessel similarity function of each voxel in the fused target pulmonary blood vessel image with a preset threshold.
Optionally, the preset threshold is obtained by calculating a second function, where the second function is:
where p is the location parameter for each voxel, σpTarget vessel scale, T (σ), for a voxel with a location parameter pp) Response threshold, t, for vessel similarity functionσIs a vessel scale threshold value, n is the vessel scale in the vessel scale range is smaller than tσI is the index of the vessel scale, i is 0, …, k, k takes values from 1 to N, N is a positive integer, T is a positive integerminAnd TmaxA tubular feature response intensity threshold is specified for the user.
A setting module 505, configured to set the first function value of each voxel in the target pulmonary vascular image to 0 when the response value of the vascular similarity function of each voxel in the target pulmonary vascular image is smaller than a preset threshold, and set the first function value of each voxel in the target pulmonary vascular image to 1 when the response value of the vascular similarity function of each voxel in the target pulmonary vascular image is greater than or equal to the preset threshold.
Optionally, the first function is:
where p is the location parameter for each voxel, σpTarget vessel dimension, V, for a voxel with a position parameter pseg(L(p,σp),σp) Is a discriminant function for pulmonary vessel segmentation, L (p, σ)p) For a voxel with position parameter p at a target vessel scale σpResponse value of vessel similarity function of (1), T (σ)p) Is a preset threshold.
Optionally, as shown in fig. 8, the apparatus 50 for segmenting pulmonary blood vessels from a pulmonary mask image further includes:
a preprocessing module 506, configured to preprocess the lung image to obtain a mask image of the lung.
It should be noted that the foregoing explanation of the embodiment of the method for segmenting a pulmonary blood vessel is also applicable to the device 50 for segmenting a pulmonary blood vessel of this embodiment, and the implementation principle is similar, and will not be described herein again.
In this embodiment, by calculating a response value of a blood vessel similarity function of each voxel in the lung mask image at different blood vessel scales in a blood vessel scale range, taking the response value as the lung blood vessel image, obtaining a target lung blood vessel image of each voxel in the lung mask image, and fusing the target lung blood vessel images of a plurality of voxels, the accuracy of segmenting lung blood vessels from the lung mask image can be effectively improved, and the effect of segmenting lung blood vessels from the lung mask image is improved.
It should be noted that the terms "first," "second," and the like in the description of the present invention are used for descriptive purposes only and are not to be construed as indicating or implying relative importance. In addition, in the description of the present invention, "a plurality" means two or more unless otherwise specified.
Any process or method descriptions in flow charts or otherwise described herein may be understood as representing modules, segments, or portions of code which include one or more executable instructions for implementing specific logical functions or steps of the process, and alternate implementations are included within the scope of the preferred embodiment of the present invention in which functions may be executed out of order from that shown or discussed, including substantially concurrently or in reverse order, depending on the functionality involved, as would be understood by those reasonably skilled in the art of the present invention.
It should be understood that portions of the present invention may be implemented in hardware, software, firmware, or a combination thereof. In the above embodiments, the various steps or methods may be implemented in software or firmware stored in memory and executed by a suitable instruction execution system. For example, if implemented in hardware, as in another embodiment, any one or combination of the following techniques, which are known in the art, may be used: a discrete logic circuit having a logic gate circuit for implementing a logic function on a data signal, an application specific integrated circuit having an appropriate combinational logic gate circuit, a Programmable Gate Array (PGA), a Field Programmable Gate Array (FPGA), or the like.
It will be understood by those skilled in the art that all or part of the steps carried by the method for implementing the above embodiments may be implemented by hardware related to instructions of a program, which may be stored in a computer readable storage medium, and when the program is executed, the program includes one or a combination of the steps of the method embodiments.
In addition, functional units in the embodiments of the present invention may be integrated into one processing module, or each unit may exist alone physically, or two or more units are integrated into one module. The integrated module can be realized in a hardware mode, and can also be realized in a software functional module mode. The integrated module, if implemented in the form of a software functional module and sold or used as a stand-alone product, may also be stored in a computer readable storage medium.
The storage medium mentioned above may be a read-only memory, a magnetic or optical disk, etc.
In the description herein, references to the description of the term "one embodiment," "some embodiments," "an example," "a specific example," or "some examples," etc., mean that a particular feature, structure, material, or characteristic described in connection with the embodiment or example is included in at least one embodiment or example of the invention. In this specification, the schematic representations of the terms used above do not necessarily refer to the same embodiment or example. Furthermore, the particular features, structures, materials, or characteristics described may be combined in any suitable manner in any one or more embodiments or examples.
Although embodiments of the present invention have been shown and described above, it is understood that the above embodiments are exemplary and should not be construed as limiting the present invention, and that variations, modifications, substitutions and alterations can be made to the above embodiments by those of ordinary skill in the art within the scope of the present invention.
Claims (16)
1. A method for segmenting pulmonary vessels from a pulmonary mask image, comprising the steps of:
calculating a response value of a vessel similarity function of each voxel in a plurality of voxels in the lung mask image under different vessel scales in a vessel scale range, and taking the response value as the lung vessel image of each voxel under different vessel scales;
acquiring a target blood vessel scale of each voxel in the lung mask image, and acquiring a lung blood vessel image corresponding to the target blood vessel scale of each voxel to serve as the target lung blood vessel image of each voxel;
and fusing the target pulmonary blood vessel images of the multiple voxels, and acquiring the fused target pulmonary blood vessel images.
2. The method of segmenting pulmonary vessels from a pulmonary mask image of claim 1, wherein the target vessel dimension for each voxel in the pulmonary mask image is obtained according to a multi-marker markov random field optimization-based method.
3. The method for segmenting pulmonary vessels from a pulmonary mask image as claimed in claim 2, wherein the obtaining of the target vessel dimension for each voxel in the pulmonary mask image according to the multi-label-based markov random field optimization method comprises:
carrying out weighted image mapping processing on the lung mask image to obtain a weighted image of the lung mask image;
establishing an energy model of the weighted graph, and acquiring a mark when the energy model obtains an energy minimum value;
and taking the blood vessel scale corresponding to the mark as the target blood vessel scale of each voxel in the lung mask image.
4. The method for segmenting lung blood vessels from lung mask image as claimed in claim 3, wherein said performing weighted map mapping process on said lung mask image to obtain weighted map of lung mask image comprises:
and mapping the voxels of the lung mask image as the vertexes of the weighted graph, mapping the relation between adjacent voxels as the edges of the weighted graph, and mapping the similarity between adjacent voxels as the weight of the edges to obtain the weighted graph of the lung mask image.
5. The method for segmenting lung vessels from lung mask images as claimed in claim 3, wherein the energy model of the weighted graph is:
wherein,in order to be able to perform the data item,for the smoothing term, η is the weight for adjusting the voxel's own energy and the connected energy of neighboring voxels during segmentation, fpThe vector f corresponds to the segmentation result of the lung mask image for the label assigned to the voxel with position parameter p in the lung mask image.
6. The method for segmenting lung vessels from lung mask images as claimed in claim 5, wherein the data items are set as:
wherein σiFor the ith vessel scale in the range of vessel scales,among the response values of the vascular similarity function of the voxels having the position parameter p, the response value of the vascular similarity function which is the largest, j being 0, …, k, k having values from 1 to N, N being a positive integer, L (p, σ), is the response value of the vascular similarity function which is the largest, j being 0i) For the voxel with position parameter p at vessel scale σiThe response value of the vessel similarity function of (a);
the smoothing term is set as:
wherein f ispFor a label assigned to a voxel with a position parameter p in the lung mask image, fqIs a label assigned to a voxel with position parameter q in the lung mask image.
7. The method for segmenting lung vessels from lung mask images as claimed in claim 1, further comprising, after obtaining the fused target lung vessel image:
comparing the response value of the blood vessel similarity function of each voxel in the fused target pulmonary blood vessel image with a preset threshold value;
if the value of the first function of the voxel is smaller than the preset threshold, setting the first function value of the voxel as 0;
and if the first function value is larger than or equal to the preset threshold, setting the first function value of the voxel as 1.
8. The method for segmenting lung vessels from a lung mask image as set forth in claim 7, wherein the first function is:
where p is the location parameter of each voxel, σpTarget vessel dimension, V, for a voxel with a position parameter pseg(L(p,σp),σp) Is a discriminant function for pulmonary vessel segmentation, L (p, σ)p) For a voxel with position parameter p at a target vessel scale σpResponse value of vessel similarity function of (1), T (σ)p) And the preset threshold value is used.
9. The method of segmenting lung vessels from lung mask images as claimed in claim 7, wherein the preset threshold is calculated by a second function, the second function is:
where p is the location parameter of each voxel, σpTarget vessel scale, T (σ), for a voxel with a location parameter pp) Is the response threshold, t, of the vessel similarity functionσIs a vessel scale threshold value, n is the vessel scale in the vessel scale range is smaller than tσI is the index of the vessel scale, i is 0, …, k, k takes values from 1 to N, N is a positive integer, T is a positive integerminAnd TmaxA tubular feature response intensity threshold is specified for the user.
10. The method of segmenting lung vessels from a lung mask image as set forth in claim 1, further comprising, prior to said computing response values of the vessel similarity function at different vessel scales in a range of vessel scales for each of a plurality of voxels in the lung mask image:
and preprocessing the lung image to obtain the lung mask image.
11. An apparatus for segmenting pulmonary blood vessels from a mask image of a lung, comprising:
the calculation module is used for calculating a response value of a blood vessel similarity function of each voxel in a plurality of voxels in the lung mask image under different blood vessel scales in a blood vessel scale range, and taking the response value as the lung blood vessel image of each voxel under different blood vessel scales;
an obtaining module, configured to obtain a target blood vessel scale of each voxel in the lung mask image, and obtain a lung blood vessel image corresponding to the target blood vessel scale of each voxel, as a target lung blood vessel image of each voxel;
and the fusion module is used for fusing the target pulmonary blood vessel images of the voxels and acquiring the fused target pulmonary blood vessel images.
12. The apparatus for segmenting lung vessels from lung mask images as claimed in claim 11, wherein the acquiring module specifically comprises:
the target blood vessel scale obtaining submodule is used for obtaining the target blood vessel scale of each voxel in the lung mask image according to a multi-mark Markov random field optimization method;
and the target pulmonary vessel image acquisition submodule is used for acquiring a pulmonary vessel image corresponding to the target vessel scale of each voxel to serve as the target pulmonary vessel image of each voxel.
13. The apparatus for segmenting pulmonary vessels from a pulmonary mask image as set forth in claim 12, wherein the target vessel scale acquisition submodule includes:
the weighting map mapping unit is used for carrying out weighting map mapping processing on the lung mask image so as to obtain a weighting map of the lung mask image;
an energy model establishing unit, configured to establish an energy model of the weighted graph, and obtain a flag when the energy model obtains an energy minimum value;
and the acquisition unit is used for taking the blood vessel scale corresponding to the mark as the target blood vessel scale of each voxel in the lung mask image.
14. The apparatus for segmenting lung vessels from a lung mask image as claimed in claim 13, wherein the weighted map mapping unit is specifically configured to:
and mapping the voxels of the lung mask image as the vertexes of the weighted graph, mapping the relation between adjacent voxels as the edges of the weighted graph, and mapping the similarity between adjacent voxels as the weight of the edges to obtain the weighted graph of the lung mask image.
15. The apparatus for segmenting lung blood vessels from a lung mask image as set forth in claim 11, further comprising:
a comparison module, configured to compare a response value of the blood vessel similarity function of each voxel in the fused target pulmonary blood vessel image with a preset threshold;
a setting module, configured to set a first function value of each voxel in the target pulmonary blood vessel image to 0 when the response value of the blood vessel similarity function of each voxel in the target pulmonary blood vessel image is smaller than the preset threshold, and set the first function value of each voxel in the target pulmonary blood vessel image to 1 when the response value of the blood vessel similarity function of each voxel in the target pulmonary blood vessel image is greater than or equal to the preset threshold.
16. The apparatus for segmenting lung blood vessels from a lung mask image as set forth in claim 11, further comprising:
and the preprocessing module is used for preprocessing the lung image so as to obtain the lung mask image.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201511030492.XA CN105701799B (en) | 2015-12-31 | 2015-12-31 | Divide pulmonary vascular method and apparatus from lung's mask image |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201511030492.XA CN105701799B (en) | 2015-12-31 | 2015-12-31 | Divide pulmonary vascular method and apparatus from lung's mask image |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN105701799A CN105701799A (en) | 2016-06-22 |
| CN105701799B true CN105701799B (en) | 2018-10-30 |
Family
ID=56226060
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201511030492.XA Active CN105701799B (en) | 2015-12-31 | 2015-12-31 | Divide pulmonary vascular method and apparatus from lung's mask image |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN105701799B (en) |
Families Citing this family (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| EP3547252A4 (en) * | 2016-12-28 | 2019-12-04 | Shanghai United Imaging Healthcare Co., Ltd. | Multi-modal image processing system and method |
| CN106875398B (en) * | 2017-01-04 | 2020-06-19 | 北京数科网维技术有限责任公司 | Method, device and terminal for realizing interactive image segmentation |
| CN107633514B (en) * | 2017-09-19 | 2020-07-31 | 北京大学第三医院 | Pulmonary nodule peripheral blood vessel quantitative evaluation system and method |
| CN108010041B (en) * | 2017-12-22 | 2020-08-11 | 数坤（北京）网络科技有限公司 | Human heart coronary artery extraction method |
| CN108364297B (en) * | 2018-03-19 | 2020-08-14 | 青岛海信医疗设备股份有限公司 | Blood vessel image segmentation method, terminal and storage medium |
| CN109872328B (en) | 2019-01-25 | 2021-05-07 | 腾讯科技（深圳）有限公司 | Brain image segmentation method, device and storage medium |
| CN111383191B (en) * | 2019-12-11 | 2024-03-08 | 北京深睿博联科技有限责任公司 | Image processing method and device for vascular fracture repair |
| CN112465749B (en) * | 2020-11-11 | 2024-03-22 | 沈阳东软智能医疗科技研究院有限公司 | Method and device for extracting pulmonary embolism image, storage medium and electronic equipment |
| CN113361509B (en) * | 2021-08-11 | 2021-11-09 | 西安交通大学医学院第一附属医院 | Image processing method for facial paralysis detection |
| CN114202469B (en) * | 2021-11-11 | 2022-08-19 | 北京医准智能科技有限公司 | Method and device for selecting hyper-parameters of Frangi filter, electronic equipment and storage medium |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101393644A (en) * | 2008-08-15 | 2009-03-25 | 华中科技大学 | Hepatic portal vein tree modeling method and system thereof |
| CN101425186A (en) * | 2008-11-17 | 2009-05-06 | 华中科技大学 | Liver subsection method based on CT image and system thereof |
| CN102243759A (en) * | 2010-05-10 | 2011-11-16 | 东北大学 | Three-dimensional lung vessel image segmentation method based on geometric deformation model |
Family Cites Families (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP4248399B2 (en) * | 2001-10-16 | 2009-04-02 | コーニンクレッカ フィリップス エレクトロニクス エヌ ヴィ | Automatic branch labeling method |
| US7058210B2 (en) * | 2001-11-20 | 2006-06-06 | General Electric Company | Method and system for lung disease detection |
| US7020316B2 (en) * | 2001-12-05 | 2006-03-28 | Siemens Corporate Research, Inc. | Vessel-feeding pulmonary nodule detection by volume projection analysis |
-
2015
- 2015-12-31 CN CN201511030492.XA patent/CN105701799B/en active Active
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101393644A (en) * | 2008-08-15 | 2009-03-25 | 华中科技大学 | Hepatic portal vein tree modeling method and system thereof |
| CN101425186A (en) * | 2008-11-17 | 2009-05-06 | 华中科技大学 | Liver subsection method based on CT image and system thereof |
| CN102243759A (en) * | 2010-05-10 | 2011-11-16 | 东北大学 | Three-dimensional lung vessel image segmentation method based on geometric deformation model |
Non-Patent Citations (5)
| Title |
|---|
| 《3D INTERACTIVE CORONARY ARTERY SEGMENTATION USING RANDOM FORESTS AND MARKOV RANDOM FIELD OPTIMIZATION》;Jingjing Deng等;《2014 IEEE International Conference on Image Processing》;20141231;全文 * |
| 《A Novel Approach for Lung Nodules Segmentation in Chest CT Using Level Sets》;Amal A.Farag等;《IEEE TRANSACTIONS ON IMAGE PROCESSING》;20131231;第22卷(第12期);全文 * |
| 《一种基于Canny算子的level-set肺部血管分割算法》;高齐新等;《系统仿真学报》;20081031;第20卷(第20期);全文 * |
| 《一种基于水平集的肺部血管快速分割方法》;高齐新等;《东北大学学报（自然科学版）》;20080630;第29卷(第6期);全文 * |
| 《基于Markov随机场的脑部三维磁共振血管造影数据的分割》;周寿军等;《集成技术》;20140131;第3卷(第1期);全文 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN105701799A (en) | 2016-06-22 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN105701799B (en) | Divide pulmonary vascular method and apparatus from lung's mask image | |
| CN111528871B (en) | Methods and systems for whole body bone removal and vessel visualization in medical images | |
| US10111713B2 (en) | Surgery assistance apparatus, surgery assistance method and non-transitory computer-readable recording medium having stored therein surgery assistance program | |
| JP6877868B2 (en) | Image processing equipment, image processing method and image processing program | |
| JP4950071B2 (en) | Method for automatic extraction of pulmonary artery tree from 3D medical images | |
| JP4999163B2 (en) | Image processing method, apparatus, and program | |
| US9406146B2 (en) | Quantitative perfusion analysis | |
| CN107633514B (en) | Pulmonary nodule peripheral blood vessel quantitative evaluation system and method | |
| US9142029B2 (en) | Region extraction apparatus, method and program | |
| US8165376B2 (en) | System and method for automatic detection of rib metastasis in computed tomography volume | |
| CN110738652B (en) | Pulmonary artery and vein separation method and device | |
| Singadkar et al. | Automatic lung segmentation for the inclusion of juxtapleural nodules and pulmonary vessels using curvature based border correction | |
| WO2011046511A1 (en) | A method and system for segmenting a liver object in an image | |
| US7480401B2 (en) | Method for local surface smoothing with application to chest wall nodule segmentation in lung CT data | |
| US8050470B2 (en) | Branch extension method for airway segmentation | |
| Yim et al. | Correction of segmented lung boundary for inclusion of pleural nodules and pulmonary vessels in chest CT images | |
| Alirr et al. | Survey on liver tumour resection planning system: steps, techniques, and parameters | |
| US7835555B2 (en) | System and method for airway detection | |
| Gemme et al. | Automatic MPST-cut for segmentation of carpal bones from MR volumes | |
| Geng et al. | 3D multi-scale pulmonary vascular segmentation algorithm based on multi-label MRF optimization | |
| Prażuch et al. | BRONCO: Automated modelling of the bronchovascular bundle using the Computed Tomography Images | |
| Rani et al. | 3D Tumor segmentation using surface rendering technique | |
| Zai et al. | Automatic and Effective Delineation of Coronary Arteries from CTA Data Using Two-Way Active Contour Model | |
| Kiraly et al. | Parsing of the Lungs and Airways | |
| Rodrigues et al. | A region-based algorithm for automatic bone segmentation in volumetric CT |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |