CN108961278A - The method and its system of abdominal wall muscle segmentation based on image data - Google Patents
The method and its system of abdominal wall muscle segmentation based on image data Download PDFInfo
- Publication number
- CN108961278A CN108961278A CN201810638182.3A CN201810638182A CN108961278A CN 108961278 A CN108961278 A CN 108961278A CN 201810638182 A CN201810638182 A CN 201810638182A CN 108961278 A CN108961278 A CN 108961278A
- Authority
- CN
- China
- Prior art keywords
- muscle
- abdominal
- hernia
- image
- abdominal hernia
- 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.)
- Granted
Links
Classifications
-
- 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
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/155—Segmentation; Edge detection involving morphological operators
Landscapes
- Engineering & Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Image Processing (AREA)
- Apparatus For Radiation Diagnosis (AREA)
- Image Analysis (AREA)
Abstract
The present invention is suitable for technical field of image processing, provides a kind of method and system of abdominal wall muscle segmentation based on image data, which comprises the imaged image of the abdominal wall muscle of acquisition A, is carried out adaptive bilateral filtering pretreatment；B, according to abdominal hernia muscle image feature, pre-segmentation abdominal hernia muscle imagery zone is set to the imaged image of the pretreated abdominal wall muscle, extracts initial abdominal hernia muscle cut zone；C, the image feature function for setting the abdominal hernia muscle extracts abdominal hernia muscle edge contour in the initial abdominal hernia muscle cut zone；D, processing is optimized to the abdominal hernia muscle edge contour, obtains the final abdominal hernia muscle cut zone.Whereby, the present invention realizes the visible for automatically extracting the 3 D stereo multi-angle of the imagery zone for being partitioned into celiocele muscle and celiocele muscle region.
Description
Technical field
The present invention relates to technical field of image processing more particularly to a kind of sides of the abdominal wall muscle segmentation based on image data
Method and its system.
Background technique
In modern medical service, B ultrasound, CT (Computed Tomography Computed tomography) and MRI
(Nuclear Magnetic Resonance Imaging, Magnetic resonance imaging) is commonly used in diagnosis abdominal hernia disease
Video diagnostic technology, in process of clinical application, B ultrasound, CT and MRI image have following three points not to the diagnosis of abdominal hernia disease
Foot:
B ultrasound, CT and MRI image are unclear to the identification of abdominal hernia defect abdominal wall muscle.It is most in abdominal hernia patient
Will appear the lesion of abdominal wall muscle around defect, the CT value of this portion of tissue can reduce (normal bone flesh CT value 45~75, and
The CT value of abdominal wall muscle is not often in this 95% credibility interval around defect), local situations such as there are also steatosises.Therefore face
Two dimensional image, which is seen, in bed existing equipment (CT, MRI) inspection is not clear, even if part myodegeneration lesion can be seen, but
It is easy ignored.
B ultrasound, CT and MRI image are unclear to the measurement of abdominal hernia hernical sac space size and volume.Clinically, abdominal hernia disease
When people carries out operative treatment, the space dissection for needing to assess the first abdominal cavity (normal abdominal cavity) and the second abdominal cavity (hernical sac space) is closed
System needs to measure its respectively volume size shared by spatial volume size and content respectively, and then whether assess abdominal cavity can be complete
Accommodate all hernial content Hui Na.And the bidimensional images such as CT, MRI cannot accomplish clearly to assess respective spatial volume size and interior
Tolerant shared volume size.
B ultrasound, CT and MRI image are unclear to the measurement of abdominal hernia hernia ring defect location and size.B ultrasound or abdominal CT are unenhanced
Show hernia ring defect location and size (small-sized: 0~4cm, medium-sized: 4~8cm, large-scale: 8~12cm: epimegetic: > 12cm),
But bidimensional image cannot consider the comprehensive condition (hernia anchor ring product, hernical sac volume) of abdominal hernia, not account for abdomen around defect more
The lesion situation of wall muscle.
In summary, the prior art is in actual use, it is clear that there is inconvenient and defect, so it is necessary to be improved.
Summary of the invention
For above-mentioned defect, the side for the abdominal wall muscle segmentation based on image data that the purpose of the present invention is to provide a kind of
Method and its system, to realize the three-dimensional for automatically extracting the imagery zone and celiocele muscle region that are partitioned into celiocele muscle
The visible of stereo multi-angle.
To achieve the goals above, the present invention provides a kind of method of abdominal wall muscle segmentation based on image data, special
Sign is, comprising:
A, the imaged image of the abdominal wall muscle of acquisition is subjected to adaptive bilateral filtering pretreatment；
B, according to abdominal hernia muscle image feature, pre-segmentation is set to the imaged image of the pretreated abdominal wall muscle
Abdominal hernia muscle imagery zone extracts initial abdominal hernia muscle cut zone；
C, the image feature function for setting the abdominal hernia muscle, mentions in the initial abdominal hernia muscle cut zone
Take abdominal hernia muscle edge contour；
D, processing is optimized to the abdominal hernia muscle edge contour, obtains the final abdominal hernia muscle cut section
Domain.
According to the method, the step A includes:
A1, the noise level spatial domain standard deviation Sigma S that the imaged image is calculated using interative computation；
A2, the codomain Gaussian function standard deviation Sigma R for calculating the imaged image；
A3, according to the spatial domain standard deviation Sigma S and the codomain Gaussian function standard deviation Sigma R, to described
Imaged image carries out bilateral filtering pretreatment, obtains the imaged image of the pretreated abdominal wall muscle.
According to the method, the step B includes:
B1, according to the abdominal hernia muscle image feature, in the imaged image of the abdominal wall muscle after the pre-treatment
Serial cross-sectional image first, which is opened, delineates the first abdominal hernia muscle imagery zone profile with last；
B2, the series according to abdominal hernia muscle image feature, in the imaged image of the abdominal wall muscle after the pre-treatment
Coronal bit image first, which is opened, delineates the second abdominal hernia muscle imagery zone profile with last；
The first, second abdominal hernia muscle imagery zone profile that B3, basis are delineated, generates the abdominal muscles image
First segmentation result of image.
According to the method, the step C includes:
C1, according to first segmentation result, determine initial profile line；The initial profile line includes or does not include complete
Initial abdominal hernia muscle cut zone described in portion；
C2, the initial abdominal hernia muscle cut zone is divided by inside and outside two parts according to the initial profile line,
Measure the average gray value of inside and outside two parts；
C3, calculate inside and outside two parts region image grayscale apart from advantage；
C4, according to C-V Level Set Models, be iterated and calculate inside and outside two parts profile, it is described initial to form segmentation
Abdominal hernia muscle cut zone boundary, obtain the abdominal hernia muscle edge contour.
According to the method, the step D includes:
D1, to obtaining, the abdominal hernia muscle edge contour carries out free growth, burn into expands and the shape of cavity filling
State processing, keeps the integrality and flatness of the abdominal hernia muscle edge contour；
D2, to the smeared out boundary in the abdominal hernia muscle edge contour, to the profile point of the smeared out boundary along institute
Smeared out boundary sequence is stated, the final abdominal hernia muscle cut zone is obtained.
In order to realize another goal of the invention of the invention, the present invention also provides a kind of stomach wall flesh based on image data
The system of meat segmentation, comprising:
The imaged image of preprocessing module, the abdominal wall muscle for that will acquire carries out adaptive bilateral filtering pretreatment；
First extraction module is used for according to abdominal hernia muscle image feature, to the shadow of the pretreated abdominal wall muscle
As image setting pre-segmentation abdominal hernia muscle imagery zone, initial abdominal hernia muscle cut zone is extracted；
Second extraction module, for setting the image feature function of the abdominal hernia muscle, in the initial abdominal hernia
Abdominal hernia muscle edge contour is extracted in muscle cut zone；
Optimization module obtains the final stomach wall for optimizing processing to the abdominal hernia muscle edge contour
Hernia muscle cut zone.
According to the system, the preprocessing module includes:
First computational submodule, for calculating the noise level spatial domain standard deviation of the imaged image using interative computation
Sigma S；
Second computational submodule, for calculating the codomain Gaussian function standard deviation Sigma R of the imaged image；
Submodule is pre-processed, for according to the spatial domain standard deviation Sigma S and the codomain Gaussian function standard deviation
Sigma R carries out bilateral filtering pretreatment to the imaged image, obtains the striograph of the pretreated abdominal wall muscle
Picture.
According to the system, first extraction module includes:
First delineates submodule, is used for according to the abdominal hernia muscle image feature, the stomach wall flesh after the pre-treatment
Serial cross-sectional image first in the imaged image of meat, which is opened, delineates the first abdominal hernia muscle imagery zone profile with last；
Second delineates submodule, for according to abdominal hernia muscle image feature, the abdominal wall muscle after the pre-treatment
The coronal bit image first of series in imaged image, which is opened, delineates the second abdominal hernia muscle imagery zone profile with last；
Submodule is generated, for according to the first, second abdominal hernia muscle imagery zone profile for delineating, described in generation
First segmentation result of abdominal muscles imaged image.
According to the system, second extraction module includes:
Initial submodule, for determining initial profile line according to first segmentation result；The initial profile line includes
Or not comprising all initial abdominal hernia muscle cut zone；
Submodule is measured, for being divided into the initial abdominal hernia muscle cut zone according to the initial profile line
Inside and outside two parts measure the average gray value of inside and outside two parts；
Third computational submodule, for calculate inside and outside two parts region image grayscale apart from advantage；
4th computational submodule, for being iterated and calculating inside and outside two parts profile according to C-V Level Set Models,
The boundary for dividing the initial abdominal hernia muscle cut zone is formed, the abdominal hernia muscle edge contour is obtained.
According to the system, the optimization module includes:
First optimization submodule, for the abdominal hernia muscle edge contour to carry out free growth, burn into expands to obtaining
And the Morphological scale-space of cavity filling, keep the integrality and flatness of the abdominal hernia muscle edge contour；
Second optimization submodule, for the smeared out boundary in the abdominal hernia muscle edge contour, to the fuzzy side
The profile point on boundary sorts along the smeared out boundary, obtains the final abdominal hernia muscle cut zone.
The present invention after the imaged image of collected abdominal wall muscle is carried out adaptive bilateral filtering pretreatment by obtaining
Good image boundary is based on pretreated image, according to abdominal hernia muscle image feature, to the pretreated abdomen later
The imaged image of wall muscle sets pre-segmentation abdominal hernia muscle imagery zone, extracts initial abdominal hernia muscle cut zone, obtains
Obtain interested target area；After extracting interested target area, the image feature function of the abdominal hernia muscle is set,
Abdominal hernia muscle edge contour is extracted, in the initial abdominal hernia muscle cut zone to ultimately form segmentation object side
Boundary；Last application target regional morphology processing, optimizes processing to the abdominal hernia muscle edge contour, obtains final
The abdominal hernia muscle cut zone.It is partitioned into the Imaging Features of celiocele muscle automatically in image data as a result, automatically
The imagery zone for being partitioned into celiocele muscle is extracted, realizes the visible of the 3 D stereo multi-angle of hernia muscle region.It is quasi-
True segmentation result rebuilds the anatomical structure that can intuitively show human internal organ by three-dimensional visualization, and stereoscopic vision can be complete
Orientation helps hernia and stomach wall surgeon, can be used for the precise positioning of abdominal-wall defect and its surrounding abdominal wall muscle extent of disease.
Detailed description of the invention
Fig. 1 is the system composition schematic diagram of the abdominal wall muscle segmentation provided in an embodiment of the present invention based on image data；
Fig. 2 is the system composition schematic diagram of the abdominal wall muscle segmentation provided in an embodiment of the present invention based on image data；
Fig. 3 is the method flow diagram of the abdominal wall muscle segmentation provided in an embodiment of the present invention based on image data.
Specific embodiment
In order to make the objectives, technical solutions, and advantages of the present invention clearer, with reference to the accompanying drawings and embodiments, right
The present invention is further elaborated.It should be appreciated that the specific embodiments described herein are merely illustrative of the present invention, and
It is not used in the restriction present invention.
Referring to Fig. 1, in one embodiment of the invention, provide it is a kind of based on image data abdominal wall muscle segmentation
System 100, comprising:
The imaged image of preprocessing module 10, the abdominal wall muscle for that will acquire carries out adaptive bilateral filtering pretreatment；
First extraction module 20 is used for according to abdominal hernia muscle image feature, to the pretreated abdominal wall muscle
Imaged image sets pre-segmentation abdominal hernia muscle imagery zone, extracts initial abdominal hernia muscle cut zone；
Second extraction module 30, for setting the image feature function of the abdominal hernia muscle, in the initial stomach wall
Abdominal hernia muscle edge contour is extracted in hernia muscle cut zone；
Optimization module 40 obtains the final abdomen for optimizing processing to the abdominal hernia muscle edge contour
Wall hernia muscle cut zone.
In this embodiment, first of all for obtaining good image boundary, preprocessing module 10 is by collected stomach wall flesh
The imaged image of meat carries out adaptive bilateral filtering pretreatment by preprocessing module 10；And the imaged image of abdominal wall muscle is preparatory
Imaged image acquisition can be carried out by B ultrasound, CT and MRI.It is based on pretreated image later, the first extraction module 20 is according to abdomen
Wall hernia muscle image feature sets pre-segmentation abdominal hernia muscle image area to the imaged image of the pretreated abdominal wall muscle
Initial abdominal hernia muscle cut zone is extracted in domain；This is to interact segmentation according to abdominal hernia muscle image feature and manually set
Pre-segmentation abdominal hernia muscle imagery zone is determined, to extract interested target area.After extracting interested target area, the
Two extraction modules 30 set the image feature function of the abdominal hernia muscle, in the initial abdominal hernia muscle cut zone
Abdominal hernia muscle edge contour is extracted, to ultimately form segmentation object boundary.The abdominal hernia of the optimization module 40 according to acquisition
Muscle edge contour, application target regional morphology processing, optimizes processing to the abdominal hernia muscle edge contour, obtains
The final abdominal hernia muscle cut zone.It present embodiments provides as a result, and abdominal hernia disease is realized based on image data
The system of abdominal muscles divided automatically, should be based on the system 100 that the abdominal wall muscle of image data is divided especially in image number
It is partitioned into the Imaging Features of celiocele muscle automatically in, automatically extracts the imagery zone for being partitioned into celiocele muscle, realize
The visible of the 3 D stereo multi-angle of hernia muscle region.
Referring to fig. 2, in one embodiment of the invention, the preprocessing module 10 includes:
First computational submodule 11, for calculating the noise level spatial domain standard of the imaged image using interative computation
Poor Sigma S；
Second computational submodule 12, for calculating the codomain Gaussian function standard deviation Sigma R of the imaged image；
Submodule 13 is pre-processed, for according to the spatial domain standard deviation Sigma S and the codomain Gaussian function standard
Poor Sigma R carries out bilateral filtering pretreatment to the imaged image, obtains the striograph of the pretreated abdominal wall muscle
Picture.
In this embodiment, the noise level that the first computational submodule 11 calculates the imaged image using interative computation is empty
Between domain standard deviation Sigma S；The specific calculating process of first computational submodule 11 includes:
1. several 7 × 7 region can be marked off with each pixel of the imaged image for left vertex；
2. calculating noise level spatial domain Gaussian function using imaged image described in initial input as initial weak texture image
Standard deviation Sigma S (K)；
3. the mean value for carrying out gradient covariance matrix compares by initial weak texture image, new weak texture image is obtained
Region, and calculate noise level spatial domain standard deviation Sigma S (K+1)；
4. judging Sigma S (K) and Sigma S (K+1), if Sigma S (K) ≈ Sigma S (K+1), stop calculating
And step 5 is gone to, otherwise go to step 6.
5. exporting Sigma S=Sigma S (K+1).
6. enabling Sigma S (K)=Sigma S (K+1), step 2 is gone to, continues to calculate, in total iteration 3 times.
Later, the second computational submodule 12 calculates the codomain Gaussian function standard deviation Sigma R of the imaged image；Second
The specific calculating process of computational submodule 12 includes:
1. each pixel value of imaged image obtains edge pixel I (s) after Boundary algorithm detects；
2. calculating visible edge pixel summation Is；
3. according to formula: e=Is/N, (N is pixel sum) calculate the imaged image edge strength e；
4. according to formula: Sigma R=be, (b is linear coefficient) calculate codomain Gaussian function standard deviation Sigma R.
Finally, pretreatment submodule 13 is according to the spatial domain standard deviation Sigma S and the codomain Gaussian function standard
Poor Sigma R carries out bilateral filtering pretreatment to the imaged image, obtains the striograph of the pretreated abdominal wall muscle
Picture.As a result, according to the spatial domain standard deviation Sigma S and codomain Gaussian function standard deviation Sigma R adaptively measured, carry out double
Side filter preprocessing.
Referring to fig. 2, in one embodiment of the invention, the first extraction module 20 includes:
First delineates submodule 21, is used for according to the abdominal hernia muscle image feature, the stomach wall after the pre-treatment
Serial cross-sectional image first in the imaged image of muscle, which is opened, delineates the first abdominal hernia muscle imagery zone wheel with last
It is wide；
Second delineates submodule 22, is used for according to abdominal hernia muscle image feature, the abdominal wall muscle after the pre-treatment
Imaged image in the coronal bit image first of series open and delineate the second abdominal hernia muscle imagery zone profile with last；
Submodule 23 is generated, for generating institute according to the first, second abdominal hernia muscle imagery zone profile delineated
State the first segmentation result of abdominal muscles imaged image.
In this embodiment, it is based on the pretreated imaged image, according to the abdominal hernia muscle image feature, interaction
Divide automatic or manual setting pre-segmentation abdominal hernia muscle imagery zone, extracts interested target area.It can be by above-mentioned
First delineate submodule 21, second delineate submodule 22 and generate submodule 23 and predetermined search setting or manually set
It is fixed.
Referring to fig. 2, in one embodiment of the invention, the second extraction module 30 includes:
Initial submodule 31, for determining initial profile line according to first segmentation result；The initial profile line packet
Contain or not comprising all initial abdominal hernia muscle cut zone；
Submodule 32 is measured, for being divided into the initial abdominal hernia muscle cut zone according to the initial profile line
Inside and outside two parts measure the average gray value of inside and outside two parts；
Third computational submodule 33, for calculate inside and outside two parts region image grayscale apart from advantage；
4th computational submodule 34, for being iterated and calculating inside and outside two parts wheel according to C-V Level Set Models
Exterior feature forms the boundary for dividing the initial abdominal hernia muscle cut zone, obtains the abdominal hernia muscle edge contour.
In this embodiment, image of the initial submodule 31 based on first segmentation result, according to the first segmentation result
Contour line be defined as initial profile line, measurement submodule 32 is divided into inside and outside two regions according to described image initial profile line, according to
Image is drawn C1, C2 in two sub-sections according to inside and outside area image average gray；Third computational submodule 33 is interior according to profile length
Exterior domain area and inside and outside area grayscale distance capabilities, the 4th computational submodule 34 carry out level set according to C-V Level Set Models
Develop, moveable contour gradually tends to the boundary of inside and outside under the action of internal force and external force, ultimately forms segmentation object boundary.
Wherein, C-V Level Set Models are a kind of new active contour models that Chan and Vese were proposed in 2001, based on simplification
Mumford-Shah model, with level set thought, by the minimum of energy function come the curve that develops.
Referring to fig. 2, in one embodiment of the invention, optimization module 40 includes:
First optimization submodule 41, for swollen to the abdominal hernia muscle edge contour progress free growth, burn into is obtained
The Morphological scale-space of swollen and empty filling keeps the integrality and flatness of the abdominal hernia muscle edge contour；
Second optimization submodule 42, for being obscured to described to the smeared out boundary in the abdominal hernia muscle edge contour
The profile point on boundary sorts along the smeared out boundary, obtains the final abdominal hernia muscle cut zone.
In this embodiment, according to the abdominal hernia muscle edge contour obtained, application target regional morphology is handled,
Initial profile optimization processing finally finally divides target, obtains the abdominal hernia muscle region.Optimization module 40 is specific
Optimization process it is as follows:
First step: to the Morphological scale-space of target area；Target area Morphological scale-space includes free growth, corrosion,
Expansion, cavity filling, keeps the integrality and flatness of profile.Specific process content is realized by the first optimization submodule 41:
1. by closed operation come filling cavity, extended contour makes up the boundary of missing；
2. eliminating noise at the boundary by opening operation, profile is shunk to eliminate bur noise.
Second step, the optimization processing of initial profile；
In order to handle the smeared out boundary of target area in current layer medical image, need to sort to profile point along boundary.
Good ranking results can be obtained based on the sequence of polar profile point.It is realized by the second optimization submodule 42, it is specific to walk
It is rapid as follows:
1. the acquisition of profile central point.Two-dimensional coordinate system is established according to the plane of delineation, the top left corner apex of image is origin,
It is horizontally to the right X-axis positive direction, is vertically downward Y-axis positive direction.Central point is calculated according to the profile point that coarse segmentation obtains.
2. the foundation of polar coordinate system.Contouring central point is sat as pole 0, while along the plane of delineation in the plane of delineation
It is polar axis that the positive direction of the x-axis of mark system, which draws a ray, will counterclockwise be set to the positive direction of angle.
3. the sequence of profile point.The institutional framework of most medical image is all comparison rule, therefore for profile
The sequence of point, there are mainly two types of situation: first is that when the profile central point of the target area of medical image is located within profile, i.e. mesh
Mark shape be convex form, then the polar angle of each profile point be it is different, profile point can be carried out as long as according to polar angle
Sequence, profile point is small into big deposit profile queue with people according to polar angle；Second is that when profile central point is located at except profile, i.e.,
Target shape is concave shape, then needs first to sort to profile point from small to large according to polar angle, then the equal situation of polar angle is discussed,
After having traversed all the points, the profile point in stack is successively popped up, is stored in profile queue.Final extract is partitioned into abdomen as a result,
The imagery zone of portion's hernia muscle realizes the visible of the 3 D stereo multi-angle of hernia muscle region.
Belong to medical image in the system 100 for the abdominal wall muscle segmentation based on image data that above-mentioned multiple embodiments provide
Three-dimensional visualization processing and computer medicine assisting in diagnosis and treatment system regions, in particular to using Artificial intellectual technology to doctor
Learn segmentation, identification and the three-dimensional visualization imaging technique of the abdominal wall muscle of the abdominal hernia disease in image.
In order to realize another goal of the invention of the invention, the present invention also provides use above-mentioned multiple abdomens based on image data
The method that the system 100 of wall muscle segmentation realizes the abdominal wall muscle segmentation based on image data.
Referring to Fig. 3, in one embodiment of the invention, provide it is a kind of based on image data abdominal wall muscle segmentation
Method, comprising:
The imaged image of the abdominal wall muscle of acquisition is carried out adaptive bilateral filtering and located in advance by step S301, preprocessing module 10
Reason；To obtain the imaged image boundary of the good abdominal wall muscle；
Step S302, the first extraction module 20 is according to abdominal hernia muscle image feature, to the pretreated stomach wall flesh
The imaged image of meat sets pre-segmentation abdominal hernia muscle imagery zone, extracts initial abdominal hernia muscle cut zone；With basis
Abdominal hernia muscle image feature interacts segmentation and makes a reservation for or manually set pre-segmentation abdominal hernia muscle imagery zone, to extract
Interested target area；
Step S303, the second extraction module 30 set the image feature function of the abdominal hernia muscle, described initial
Abdominal hernia muscle edge contour is extracted in abdominal hernia muscle cut zone；By the image feature letter for setting the abdominal hernia muscle
Number ultimately forms segmentation object to extract abdominal hernia muscle edge contour in the initial abdominal hernia muscle cut zone
Boundary；
Step S304, optimization module 40 optimize processing to the abdominal hernia muscle edge contour, obtain final institute
State abdominal hernia muscle cut zone.According to the abdominal hernia muscle edge contour obtained, application target regional morphology is handled,
Initial profile optimization processing finally finally divides target, obtains the abdominal hernia muscle region.In addition, working as to handle
The smeared out boundary of target area in front layer medical image sorts to profile point along boundary, is sorted based on polar profile point
Good ranking results can be obtained, to obtain the final abdominal hernia muscle cut zone.
It may be implemented using the method for the above-mentioned abdominal wall muscle segmentation based on image data by physical, the physiology of human body
Property parameter digitized, be intelligent, Mobile state Medical Image Processing of going forward side by side, accurate three-dimensional is rebuild and Quantitative Study, is
Diagnosis, treatment and the education of doctor and training of disease provide a kind of brand new technical means more true to nature.Therefore, celiocele
Muscle automatic segmentation result to it is objective, quantitatively assess abdominal hernia (hernia ring transverse diameter, hernia anchor ring product, hernical sac volume) the case where, it is preceding
It the case where abdominal wall muscle (rectus aabdominis, musculus obliquus externus abdominis, oblique, musculus trasversus abdomins), is of great significance.Accurate segmentation result warp
It crosses three-dimensional visualization and rebuilds the anatomical structure that can intuitively show human internal organ, stereoscopic vision can comprehensive help hernia and abdomen
Wall surgeon can be used for the precise positioning of abdominal-wall defect and its surrounding abdominal wall muscle extent of disease.
In one embodiment of the invention, the step S301 includes:
First computational submodule 11 calculates the noise level spatial domain standard deviation of the imaged image using interative computation
Sigma S；
Second computational submodule 12 calculates the codomain Gaussian function standard deviation Sigma R of the imaged image；
In this embodiment, pretreatment submodule 13 is according to the spatial domain standard deviation Sigma S and the codomain Gauss
Functional standard difference Sigma R carries out bilateral filtering pretreatment to the imaged image, obtains the pretreated abdominal wall muscle
Imaged image.
First computational submodule 11 calculates the noise level spatial domain standard deviation of the imaged image using interative computation
Sigma S；It is specific:
1. several 7 × 7 region can be marked off with each pixel of the imaged image for left vertex；
2. calculating noise level spatial domain Gaussian function using imaged image described in initial input as initial weak texture image
Standard deviation Sigma S (K)；
3. the mean value for carrying out gradient covariance matrix compares by initial weak texture image, new weak texture image is obtained
Region, and calculate noise level spatial domain standard deviation Sigma S (K+1)；
4. judging Sigma S (K) and Sigma S (K+1), if Sigma S (K) ≈ Sigma S (K+1), stop calculating
And step 5 is gone to, otherwise go to step 6.
5. exporting Sigma S=Sigma S (K+1).
6. enabling Sigma S (K)=Sigma S (K+1), step 2 is gone to, continues to calculate, in total iteration 3 times.
Later, the second computational submodule 12 calculates the codomain Gaussian function standard deviation Sigma R of the imaged image；Second
The specific calculating process of computational submodule 12 includes:
1. each pixel value of imaged image obtains edge pixel I (s) after Boundary algorithm detects；
2. calculating visible edge pixel summation Is；
3. according to formula: e=Is/N, (N is pixel sum) calculate the imaged image edge strength e；
4. according to formula: Sigma R=be, (b is linear coefficient) calculate codomain Gaussian function standard deviation Sigma R.
Finally, pretreatment submodule 13 is according to the spatial domain standard deviation Sigma S and the codomain Gaussian function standard
Poor Sigma R carries out bilateral filtering pretreatment to the imaged image, obtains the striograph of the pretreated abdominal wall muscle
Picture.As a result, according to the spatial domain standard deviation Sigma S and codomain Gaussian function standard deviation Sigma R adaptively measured, carry out
Bilateral filtering pretreatment.
In one embodiment of the invention, the step S302 includes:
First delineates submodule 21 according to the abdominal hernia muscle image feature, the abdominal wall muscle after the pre-treatment
Serial cross-sectional image first in imaged image, which is opened, delineates the first abdominal hernia muscle imagery zone profile with last；
Second delineates submodule 22 according to abdominal hernia muscle image feature, the image of the abdominal wall muscle after the pre-treatment
The coronal bit image first of series in image, which is opened, delineates the second abdominal hernia muscle imagery zone profile with last；
Submodule 23 is generated according to the first, second abdominal hernia muscle imagery zone profile delineated, generates the abdomen
First segmentation result of portion's muscle imaged image.Using above-mentioned steps, it can be based on pretreated image, and according to the abdomen
Wall hernia muscle image feature, the automatic or manual setting pre-segmentation abdominal hernia muscle imagery zone of interactive segmentation extract interested
Target area.
In one embodiment of the invention, the step S303 includes:
Initial submodule 31 determines initial profile line according to first segmentation result；The initial profile line include or
Person does not include all initial abdominal hernia muscle cut zone；
The initial abdominal hernia muscle cut zone is divided into inside and outside by measurement submodule 32 according to the initial profile line
Two parts measure the average gray value of inside and outside two parts；
Third computational submodule 33 calculates the image grayscale in the region of inside and outside two parts apart from advantage；
4th computational submodule 34 is iterated according to C-V Level Set Models and calculates inside and outside two parts profile, shape
At the boundary for dividing the initial abdominal hernia muscle cut zone, the abdominal hernia muscle edge contour is obtained.
In this embodiment, the image based on the first segmentation result is defined as initially taking turns according to the contour line of the first segmentation
Profile, is divided into inside and outside two regions according to image initial contour line, and image is divided two according to inside and outside area image average gray
Part C1, C2；According to profile length, inside and outside region area and inside and outside area grayscale distance capabilities, according to C-V Level Set Models into
Row level set movements, moveable contour gradually tends to the boundary of inside and outside under the action of internal force and external force, ultimately forms segmentation
Object boundary.Wherein, steps are as follows for the iterative calculation of the 4th computational submodule 34:
1. exterior domain average gray value in calculating；
2. calculating inside and outside area grayscale value distance to keep；
3. progressive propagate the moveable contour for calculating two regions；
4. active contour is inwardly searched in outside, active contour is searched for outward in inside；
Inhibit moveable contour 5. calculating；
6. judging whether moveable contour is stable；Contour line is calculated if stablized, otherwise goes to step 2 and continues to count
It calculates.
7. obtaining stable contour line.
In one embodiment of the invention, the step S304 includes:
First 41 pairs of submodule of optimization obtain the abdominal hernia muscle edge contours carry out free growth, burn into expansion with
And the Morphological scale-space of cavity filling, keep the integrality and flatness of the abdominal hernia muscle edge contour；
First optimization submodule 41 is to the smeared out boundary in the abdominal hernia muscle edge contour, to the smeared out boundary
Profile point sorts along the smeared out boundary, obtains the final abdominal hernia muscle cut zone.
In this embodiment, according to the abdominal hernia muscle edge contour obtained, application target regional morphology is handled,
Initial profile optimization processing finally finally divides target, obtains the abdominal hernia muscle region.Specifically, the first optimization
Submodule 41 utilizes the Morphological scale-space of target area；Target area Morphological scale-space includes free growth, is corroded, and expansion is empty
Hole filling, keeps the integrality and flatness of profile.Specific process content: by closed operation come filling cavity, extended contour comes
Make up the boundary of missing；Noise at the boundary is eliminated by opening operation, shrinks profile to eliminate bur noise.First optimization submodule
41 carry out the optimization processing of initial profile.This is the smeared out boundary in order to handle target area in current layer medical image, is needed
It sorts to profile point along boundary.Good ranking results can be obtained based on the sequence of polar profile point.Specific steps packet
It includes:
1. the acquisition of profile central point.Two-dimensional coordinate system is established according to the plane of delineation, the top left corner apex of image is origin,
It is horizontally to the right X-axis positive direction, is vertically downward Y-axis positive direction.Central point is calculated according to the profile point that coarse segmentation obtains.
2. the foundation of polar coordinate system.Contouring central point is sat as pole 0, while along the plane of delineation in the plane of delineation
It is polar axis that the positive direction of the x-axis of mark system, which draws a ray, will counterclockwise be set to the positive direction of angle.
3. the sequence of profile point.The institutional framework of most medical image is all comparison rule, therefore for profile
The sequence of point, there are mainly two types of situation: first is that when the profile central point of the target area of medical image is located within profile, i.e. mesh
Mark shape be convex form, then the polar angle of each profile point be it is different, profile point can be carried out as long as according to polar angle
Sequence, profile point is small into big deposit profile queue with people according to polar angle；Second is that when profile central point is located at except profile, i.e.,
Target shape is concave shape, then needs first to sort to profile point from small to large according to polar angle, then the equal situation of polar angle is discussed,
After having traversed all the points, the profile point in stack is successively popped up, is stored in profile queue.
The method of the above-mentioned abdominal wall muscle segmentation based on image data provided by the invention as a result, will be to celiocele muscle
Realize automatic segmentation, and to it is objective, quantitatively assess abdominal hernia (hernia ring transverse diameter, hernia anchor ring product, hernical sac volume) the case where, it is preceding
It the case where abdominal wall muscle (rectus aabdominis, musculus obliquus externus abdominis, oblique, musculus trasversus abdomins), is of great significance.Accurate segmentation result warp
It crosses three-dimensional visualization and rebuilds the anatomical structure that can intuitively show human internal organ, stereoscopic vision can comprehensive help hernia and abdomen
Wall surgeon can be used for the precise positioning of abdominal-wall defect and its surrounding abdominal wall muscle extent of disease.
In conclusion the present invention is located in advance by the way that the imaged image of collected abdominal wall muscle is carried out adaptive bilateral filtering
Good image boundary is obtained after reason, pretreated image is based on later, according to abdominal hernia muscle image feature, after pretreatment
The imaged image of the abdominal wall muscle set pre-segmentation abdominal hernia muscle imagery zone, extract initial abdominal hernia muscle segmentation
Region obtains interested target area；After extracting interested target area, the image for setting the abdominal hernia muscle is special
Function is levied, abdominal hernia muscle edge contour is extracted in the initial abdominal hernia muscle cut zone, to ultimately form segmentation
Object boundary；Last application target regional morphology processing, optimizes processing to the abdominal hernia muscle edge contour, obtains
The final abdominal hernia muscle cut zone.The iconography for being partitioned into celiocele muscle automatically in image data as a result, is special
Point automatically extracts the imagery zone for being partitioned into celiocele muscle, realizes the visualization of the 3 D stereo multi-angle of hernia muscle region
Imaging.Accurate segmentation result rebuilds the anatomical structure that can intuitively show human internal organ, stereopsis by three-dimensional visualization
Feel can comprehensive help hernia and stomach wall surgeon, can be used for the accurate of abdominal-wall defect and its surrounding abdominal wall muscle extent of disease
Positioning.
Certainly, the present invention can also have other various embodiments, without deviating from the spirit and substance of the present invention, ripe
It knows those skilled in the art and makes various corresponding changes and modifications, but these corresponding changes and change in accordance with the present invention
Shape all should fall within the scope of protection of the appended claims of the present invention.
Claims (10)
1. a kind of method of the abdominal wall muscle segmentation based on image data characterized by comprising
A, the imaged image of the abdominal wall muscle of acquisition is subjected to adaptive bilateral filtering pretreatment；
B, according to abdominal hernia muscle image feature, pre-segmentation stomach wall is set to the imaged image of the pretreated abdominal wall muscle
Hernia muscle imagery zone extracts initial abdominal hernia muscle cut zone；
C, the image feature function for setting the abdominal hernia muscle, extracts abdomen in the initial abdominal hernia muscle cut zone
Wall hernia muscle edge contour；
D, processing is optimized to the abdominal hernia muscle edge contour, obtains the final abdominal hernia muscle cut zone.
2. the method according to claim 1, wherein the step A includes:
A1, the noise level spatial domain standard deviation Sigma S that the imaged image is calculated using interative computation；
A2, the codomain Gaussian function standard deviation Sigma R for calculating the imaged image；
A3, according to the spatial domain standard deviation Sigma S and the codomain Gaussian function standard deviation Sigma R, to the image
Image carries out bilateral filtering pretreatment, obtains the imaged image of the pretreated abdominal wall muscle.
3. according to the method described in claim 2, it is characterized in that, the step B includes:
B1, the series according to the abdominal hernia muscle image feature, in the imaged image of the abdominal wall muscle after the pre-treatment
Cross-sectional image first, which is opened, delineates the first abdominal hernia muscle imagery zone profile with last；
B2, according to abdominal hernia muscle image feature, in the imaged image of the abdominal wall muscle after the pre-treatment series it is coronal
Bit image first, which is opened, delineates the second abdominal hernia muscle imagery zone profile with last；
The first, second abdominal hernia muscle imagery zone profile that B3, basis are delineated, generates the abdominal muscles imaged image
The first segmentation result.
4. according to the method described in claim 3, it is characterized in that, the step C includes:
C1, according to first segmentation result, determine initial profile line；The initial profile line includes or not comprising whole institute
State initial abdominal hernia muscle cut zone；
C2, the initial abdominal hernia muscle cut zone is divided by inside and outside two parts according to the initial profile line, measured
The average gray value of inside and outside two parts；
C3, calculate inside and outside two parts region image grayscale apart from advantage；
C4, according to C-V Level Set Models, be iterated and calculate inside and outside two parts profile, formed and divide the initial abdomen
The boundary of wall hernia muscle cut zone, obtains the abdominal hernia muscle edge contour.
5. according to the method described in claim 4, it is characterized in that, the step D includes:
D1, to obtaining, the abdominal hernia muscle edge contour carries out free growth, burn into expands and the morphology of cavity filling
Processing, keeps the integrality and flatness of the abdominal hernia muscle edge contour；
D2, to the smeared out boundary in the abdominal hernia muscle edge contour, to the profile point of the smeared out boundary along the mould
Boundary sequence is pasted, the final abdominal hernia muscle cut zone is obtained.
6. a kind of system of the abdominal wall muscle segmentation based on image data characterized by comprising
The imaged image of preprocessing module, the abdominal wall muscle for that will acquire carries out adaptive bilateral filtering pretreatment；
First extraction module is used for according to abdominal hernia muscle image feature, to the striograph of the pretreated abdominal wall muscle
As setting pre-segmentation abdominal hernia muscle imagery zone, initial abdominal hernia muscle cut zone is extracted；
Second extraction module, for setting the image feature function of the abdominal hernia muscle, in the initial abdominal hernia muscle
Abdominal hernia muscle edge contour is extracted in cut zone；
Optimization module obtains the final abdominal hernia flesh for optimizing processing to the abdominal hernia muscle edge contour
Meat cut zone.
7. system according to claim 6, which is characterized in that the preprocessing module includes:
First computational submodule, for calculating the noise level spatial domain standard deviation of the imaged image using interative computation
Sigma S；
Second computational submodule, for calculating the codomain Gaussian function standard deviation Sigma R of the imaged image；
Submodule is pre-processed, for according to the spatial domain standard deviation Sigma S and the codomain Gaussian function standard deviation Sigma
R carries out bilateral filtering pretreatment to the imaged image, obtains the imaged image of the pretreated abdominal wall muscle.
8. system according to claim 7, which is characterized in that first extraction module includes:
First delineates submodule, for according to the abdominal hernia muscle image feature, the abdominal wall muscle after the pre-treatment
Serial cross-sectional image first in imaged image, which is opened, delineates the first abdominal hernia muscle imagery zone profile with last；
Second delineates submodule, is used for according to abdominal hernia muscle image feature, the image of the abdominal wall muscle after the pre-treatment
The coronal bit image first of series in image, which is opened, delineates the second abdominal hernia muscle imagery zone profile with last；
Submodule is generated, for generating the abdomen according to the first, second abdominal hernia muscle imagery zone profile delineated
First segmentation result of muscle imaged image.
9. system according to claim 8, which is characterized in that second extraction module includes:
Initial submodule, for determining initial profile line according to first segmentation result；The initial profile line include or
Not comprising all initial abdominal hernia muscle cut zone；
Submodule is measured, it is inside and outside for being divided into the initial abdominal hernia muscle cut zone according to the initial profile line
Two parts measure the average gray value of inside and outside two parts；
Third computational submodule, for calculate inside and outside two parts region image grayscale apart from advantage；
4th computational submodule is formed for being iterated and calculating inside and outside two parts profile according to C-V Level Set Models
The boundary for dividing the initial abdominal hernia muscle cut zone, obtains the abdominal hernia muscle edge contour.
10. system according to claim 9, which is characterized in that the optimization module includes:
First optimization submodule, for obtain the abdominal hernia muscle edge contour carry out free growth, burn into expansion and
The Morphological scale-space of cavity filling, keeps the integrality and flatness of the abdominal hernia muscle edge contour；
Second optimization submodule, for the smeared out boundary in the abdominal hernia muscle edge contour, to the smeared out boundary
Profile point sorts along the smeared out boundary, obtains the final abdominal hernia muscle cut zone.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810638182.3A CN108961278B (en) | 2018-06-20 | 2018-06-20 | Method and system for abdominal wall muscle segmentation based on image data |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810638182.3A CN108961278B (en) | 2018-06-20 | 2018-06-20 | Method and system for abdominal wall muscle segmentation based on image data |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN108961278A true CN108961278A (en) | 2018-12-07 |
| CN108961278B CN108961278B (en) | 2023-01-31 |
Family
ID=64490371
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810638182.3A Active CN108961278B (en) | 2018-06-20 | 2018-06-20 | Method and system for abdominal wall muscle segmentation based on image data |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN108961278B (en) |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110111386A (en) * | 2019-05-10 | 2019-08-09 | 深圳大学 | The method for automatic measurement and system of structural angle in a kind of musculature |
| CN111583286A (en) * | 2020-04-09 | 2020-08-25 | 天津大学 | Abdomen MRI (magnetic resonance imaging) image contour extraction method based on Flow-XDoG operator |
| CN113409309A (en) * | 2021-07-16 | 2021-09-17 | 北京积水潭医院 | Muscle CT image delineation method, system, electronic equipment and machine storage medium |
| CN116309385A (en) * | 2023-02-27 | 2023-06-23 | 之江实验室 | Abdominal fat and muscle tissue measurement method and system based on weak supervision learning |
Citations (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2002222410A (en) * | 2001-01-25 | 2002-08-09 | Hitachi Medical Corp | Image diagnostic device |
| US20090226060A1 (en) * | 2008-03-04 | 2009-09-10 | Gering David T | Method and system for improved image segmentation |
| CN101887583A (en) * | 2010-06-24 | 2010-11-17 | 东软集团股份有限公司 | Method and device for extracting brain tissue image |
| CN103886564A (en) * | 2013-11-29 | 2014-06-25 | 沈阳东软医疗***有限公司 | PET heart image myocardial contour line segmentation method and device |
| CN104809723A (en) * | 2015-04-13 | 2015-07-29 | 北京工业大学 | Three-dimensional liver CT (computed tomography) image automatically segmenting method based on hyper voxels and graph cut algorithm |
| CN106447682A (en) * | 2016-08-29 | 2017-02-22 | 天津大学 | Automatic segmentation method for breast MRI focus based on Inter-frame correlation |
| CN106570856A (en) * | 2016-08-31 | 2017-04-19 | 天津大学 | Common carotid artery intima-media thickness measuring device and method combining level set segmentation and dynamic programming |
| CN106890009A (en) * | 2017-03-17 | 2017-06-27 | 深圳大学 | A kind of method for automatic measurement and device of skeletal muscle Volume Changes |
| CN107481252A (en) * | 2017-08-24 | 2017-12-15 | 上海术理智能科技有限公司 | Dividing method, device, medium and the electronic equipment of medical image |
-
2018
- 2018-06-20 CN CN201810638182.3A patent/CN108961278B/en active Active
Patent Citations (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2002222410A (en) * | 2001-01-25 | 2002-08-09 | Hitachi Medical Corp | Image diagnostic device |
| US20090226060A1 (en) * | 2008-03-04 | 2009-09-10 | Gering David T | Method and system for improved image segmentation |
| CN101887583A (en) * | 2010-06-24 | 2010-11-17 | 东软集团股份有限公司 | Method and device for extracting brain tissue image |
| CN103886564A (en) * | 2013-11-29 | 2014-06-25 | 沈阳东软医疗***有限公司 | PET heart image myocardial contour line segmentation method and device |
| CN104809723A (en) * | 2015-04-13 | 2015-07-29 | 北京工业大学 | Three-dimensional liver CT (computed tomography) image automatically segmenting method based on hyper voxels and graph cut algorithm |
| CN106447682A (en) * | 2016-08-29 | 2017-02-22 | 天津大学 | Automatic segmentation method for breast MRI focus based on Inter-frame correlation |
| CN106570856A (en) * | 2016-08-31 | 2017-04-19 | 天津大学 | Common carotid artery intima-media thickness measuring device and method combining level set segmentation and dynamic programming |
| CN106890009A (en) * | 2017-03-17 | 2017-06-27 | 深圳大学 | A kind of method for automatic measurement and device of skeletal muscle Volume Changes |
| CN107481252A (en) * | 2017-08-24 | 2017-12-15 | 上海术理智能科技有限公司 | Dividing method, device, medium and the electronic equipment of medical image |
Non-Patent Citations (1)
| Title |
|---|
| 王亚琼: ""医学图像分割算法研究及在腿部肌肉组织分割中的应用"", 《中国优秀硕士学位论文全文数据库信息科技辑》 * |
Cited By (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110111386A (en) * | 2019-05-10 | 2019-08-09 | 深圳大学 | The method for automatic measurement and system of structural angle in a kind of musculature |
| CN110111386B (en) * | 2019-05-10 | 2023-01-24 | 深圳大学 | Method and system for automatically measuring structural angle in muscle tissue |
| CN111583286A (en) * | 2020-04-09 | 2020-08-25 | 天津大学 | Abdomen MRI (magnetic resonance imaging) image contour extraction method based on Flow-XDoG operator |
| CN111583286B (en) * | 2020-04-09 | 2023-01-20 | 天津大学 | Abdomen MRI (magnetic resonance imaging) image contour extraction method based on Flow-XDoG operator |
| CN113409309A (en) * | 2021-07-16 | 2021-09-17 | 北京积水潭医院 | Muscle CT image delineation method, system, electronic equipment and machine storage medium |
| CN113409309B (en) * | 2021-07-16 | 2023-10-27 | 北京积水潭医院 | Muscle CT image sketching method, system, electronic equipment and machine storage medium |
| CN116309385A (en) * | 2023-02-27 | 2023-06-23 | 之江实验室 | Abdominal fat and muscle tissue measurement method and system based on weak supervision learning |
| CN116309385B (en) * | 2023-02-27 | 2023-10-10 | 之江实验室 | Abdominal fat and muscle tissue measurement method and system based on weak supervision learning |
Also Published As
| Publication number | Publication date |
|---|---|
| CN108961278B (en) | 2023-01-31 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US11455732B2 (en) | Knowledge-based automatic image segmentation | |
| CN108961278A (en) | The method and its system of abdominal wall muscle segmentation based on image data | |
| EP1264277A1 (en) | Three-dimensional reconstruction of a breast from two x-ray mammograms | |
| CN111008984A (en) | Method and system for automatically drawing contour line of normal organ in medical image | |
| US9727975B2 (en) | Knowledge-based automatic image segmentation | |
| CN104504737A (en) | Method for obtaining three-dimensional tracheal tree from lung CT (computed tomography) images | |
| Feng et al. | Convolutional neural network‐based pelvic floor structure segmentation using magnetic resonance imaging in pelvic organ prolapse | |
| Tsagaan et al. | Segmentation of kidney by using a deformable model | |
| CN106651874A (en) | Brain tumor post-operation spatial domain partitioning method based on multimode MRI data | |
| Qiu et al. | Lung Nodules Detection in CT Images Using Gestalt‐Based Algorithm | |
| Yunchao et al. | Application of computer vision and deep learning in breast cancer assisted diagnosis | |
| Gallego et al. | Automatic model-based 3D segmentation of the breast in MRI | |
| US20080285822A1 (en) | Automated Stool Removal Method For Medical Imaging | |
| Stough et al. | Regional appearance in deformable model segmentation | |
| Zhang et al. | Segmenting the thoracic, abdominal and pelvic musculature on CT scans combining atlas-based model and active contour model | |
| CN112419330A (en) | Temporal bone key anatomical structure automatic positioning method based on spatial relative position prior | |
| Zhou et al. | Preliminary study for automated recognition of anatomical structure from torso CT images | |
| CN109785296A (en) | A kind of spherical assessment of indices method of three-dimensional based on CTA image | |
| Shan et al. | Threshold algorithm for pancreas segmentation in Dixon water magnetic resonance images | |
| Joseph et al. | Three dimensional reconstruction models for medical modalities: a comprehensive investigation and analysis | |
| Butler et al. | The effectiveness of self-supervised pre-training for multi-modal endometriosis classification | |
| Athira et al. | Atlas based breast registration and segmentation in the Mediolateral Oblique and Craniocaudal Views | |
| Tan et al. | A pulmonary vascular segmentation algorithm of chest CT images based on fast marching method | |
| Ding et al. | Removal of abdominal wall for 3D visualization and segmentation of organs in CT volume | |
| Yang et al. | 3D tumor segmentation in breast MRIs using 3D modified active contour method |
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 | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |