CN107154047A - Multi-mode brain tumor image blend dividing method and device - Google Patents
Multi-mode brain tumor image blend dividing method and device Download PDFInfo
- Publication number
- CN107154047A CN107154047A CN201710270990.4A CN201710270990A CN107154047A CN 107154047 A CN107154047 A CN 107154047A CN 201710270990 A CN201710270990 A CN 201710270990A CN 107154047 A CN107154047 A CN 107154047A
- Authority
- CN
- China
- Prior art keywords
- mrow
- msub
- image
- brain tumor
- msup
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/11—Region-based segmentation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/23—Clustering techniques
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/10—Image acquisition modality
- G06T2207/10072—Tomographic images
- G06T2207/10088—Magnetic resonance imaging [MRI]
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/30—Subject of image; Context of image processing
- G06T2207/30004—Biomedical image processing
- G06T2207/30016—Brain
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/30—Subject of image; Context of image processing
- G06T2207/30004—Biomedical image processing
- G06T2207/30096—Tumor; Lesion
Landscapes
- Engineering & Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Computer Vision & Pattern Recognition (AREA)
- General Physics & Mathematics (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Evolutionary Biology (AREA)
- Evolutionary Computation (AREA)
- Artificial Intelligence (AREA)
- General Engineering & Computer Science (AREA)
- Life Sciences & Earth Sciences (AREA)
- Magnetic Resonance Imaging Apparatus (AREA)
Abstract
The present invention relates to medicine equipment, medical image, to propose a kind of improved multi-mode brain tumor image blend partitioning algorithm, brain tumor region is extracted using FFCM, the border issue that tumor region is present is modified using mixed-level set algorithm.So that FFCM algorithms and level set algorithm can be more efficiently applied in MRI brain tumor images.The technical solution adopted by the present invention is, multi-mode brain tumor image blend dividing method, input Three models first include T1C, T2 and FLAIR MRI image, and processing is filtered to image using medium filtering and initial segmentation obtains pretreatment image, afterwards using linear fusion；FFCM cluster segmentations are carried out to fused images again, the larger region of wherein gray value is automatically extracted, obtained tumour less divided region carries out mixed-level collection segmentation.Present invention is mainly applied to the acquisition of medical image and processing.
Description
Technical field
It is an importance in Medical Imaging the present invention relates to medicine equipment.It swells in brain tumor cutting, brain
The fields such as knurl classification, brain tumor identification play an important roll.Concretely relate to improved multi-mode brain tumor image blend segmentation
Method and apparatus.
Background technology
In recent years, the brain tumor incidence of disease is in rising trend, accounts for the 5% of general tumour, accounts for the 70% of pediatric tumor.2015
Year, the new hair brain tumor case about 23000 only made a definite diagnosis in the U.S..The uncontrollable and indeterminate growth of cell causes brain tumor
Occur.If not carrying out early diagnosis and therapy to brain tumor, permanent brain damage is may result in, in addition it is dead.Nuclear magnetic resonance
The anomalous variation that (Magnetic Resonance Imaging, MRI) can be used for detection bodily tissue is imaged, brain tumor is to determine
The necessary means of therapeutic scheme, in all treatment methods, any information about knub position and size is all extremely important
, but it is due to that brain tumor is complex-shaped, size and location has randomness, the factor such as type difference is big causes not having also at present
Have the need for a kind of partitioning algorithm disclosure satisfy that clinic, real-time is also unable to reach requirement, different expert's manual segmentation brain tumors
The result of image also has very big difference, and cost of labor is higher.Therefore, studying accurately full-automatic brain tumor partitioning algorithm is
It is very important.
The automatic cutting techniques of brain tumor are all research hot topic direction all the time, and the dividing method of brain tumor image is divided into hand
It is divided into thresholding algorithm, clustering algorithm and deformation model again in dynamic segmentation, semi-automatic segmentation and full-automatic dividing, specific partitioning algorithm
Algorithm etc..Thresholding algorithm is used for image and split earliest, the problem of for brain tumor image, and OTSU algorithms are a kind of automatic to adapt to threshold
Value-based algorithm, can be prevented effectively from the error that fixed threshold is brought；A kind of On Local Fuzzy threshold value split for multi-region area image
(Fuzzy threshold, FTH) algorithm, which is directed to this complicated image of brain tumor, also certain effect, due to brain tumor
The complexity and thresholding algorithm of image consider not enough to pixel spatial information (si), cause the segmentation of threshold value class algorithm not can effectively solve the problem that
Brain tumor segmentation problem.
Fuzzy clustering is the class algorithm for being adapted to the segmentation of brain tumor image, especially fuzzy C-mean algorithm (Fuzzy C-mean,
FCM) algorithm, simple advantage is realized with method, but because medical image is complicated, blur margin is clear, therefore, seed point
Choose influences very big to cluster result, and FCM algorithmic methods are difficult by the spatial information (si) of image, itself calculates complicated.Then
The problem of proposing quick FCM (Fast FCM, FFCM) algorithm improvement calculating speed；The problem of for spatial information deficiency, use
Correlation between space FCM (Spatial FCM, SFCM) algorithm segmentation figure picture, effective utilization space information, but SFCM is calculated
Method calculating speed can not meet the real-time of medical image requirement；Level set algorithm can effectively handle various contouring problems,
Fuzzy clustering is combined into (Fuzzy Clustering with Level Set Methods, FCLSM) with Level Set Method to calculate
Method efficiently solves level set edge problem, but FCLSM algorithms are the problem of have real-time and be easily trapped into local optimum.
Level set algorithm is the class for belonging to deformation model algorithm, and being also widely used in brain based on level-set segmentation algorithm swells
Knurl is split, but is due to that brain tumor tissue gray scale is uneven, and is used between brain tumor tissue often without obvious border
The problem of easily there is edge leakage in this kind of algorithm.Apart from regularization level set algorithm (Distance Regularized
Level Set Evolution, DRLSE) it is to eliminate counterweight apart from regularization effect in an effective algorithm, the algorithm
The need for new initialization, so as to avoid its caused local error；Other method also has mixed-level set algorithm.This method is used
Object bounds and area information realize robust and accurately segmentation.Boundary information can help the accurate position of detected target object
Put, and area information can prevent boundary leaking, but level set algorithm can not solve to be easily trapped into local optimum and to initial value
The problem of strong depend-ence.
The content of the invention
To overcome the deficiencies in the prior art, the present invention is directed to propose a kind of improved multi-mode brain tumor image blend segmentation
Algorithm, brain tumor region is extracted using FFCM, and the border issue that tumor region is present is modified using mixed-level set algorithm.
So that FFCM algorithms and level set algorithm can be more efficiently applied in MRI brain tumor images.The skill that the present invention is used
Art scheme is that multi-mode brain tumor image blend dividing method, first input Three models include T1C, T2 and FLAIR MRI
Image, is filtered processing to image using medium filtering and initial segmentation obtains pretreatment image, afterwards using linear fusion；
FFCM cluster segmentations are carried out to fused images again, the larger region of wherein gray value are automatically extracted, obtained tumour less divided area
Domain carries out mixed-level collection segmentation.
FFCM cluster segmentations are comprised the concrete steps that, data are divided into c classes by fuzzy C-mean algorithm theory, are schemed for a width M × N
Picture, if { hi, i=1,2 ..., n }, n=M × N, hiIt is the set that the pixel intensity value in image histogram is constituted, wherein M and N
It is the length and width of image, { vj, j=1,2 ..., c } and it is the set that cluster centre is constituted, and μj(hi) it is hiIt is under the jurisdiction of the person in servitude of j classes
Membership fuction, | | | | 2 norms are represented, b is one and is more than 1 constant, then：
Iterative (3) (4) if meeting stopping criterion for iteration, t ＞ T orThen stop, wherein t tables
Show iterations, ε is off condition, and T represents maximum iteration, after algorithm terminates, and pixel is divided by maximum membership degree
Class, if μj(hi) ＞ μj(hk), then by hiIt is classified as jth class region, k=1,2 ..., c；i≠k.
The segmentation of mixed-level collection comprises the concrete steps that, imbedding function φ null set be used to representing active contour C=X | φ (X)
=0 }, the point of profile inside/outside has positive/negative φ values, the function definition minimized the need for proposing：
I is image to be split in formula (5),It is the boundary characteristic figure related to image gradient,It is gradient
Operator, H (φ) is Heaviside function, and Ω is image area, and α and β are to predefine weight to balance two, and μ is to indicate target pair
The predefined parameter of the lower limit of the gray level of elephant；
Wherein,To point to the normal vector of curved exterior, therefore the dominant curve evolution partial differential equation of active contour are
In formulaAnd curvature<·,·>For inner product；Due to there was only the several of curve
What change is interested in segmentation, so from formula (6) it may be noted that the institute on curve a little all moves in the normal direction
Dynamic, formula (6) Section 1 is the biography of the contractile motion for the extension movement and exterior section for describing the curved portion inside destination object
Item is broadcast, Section 2 is advective term, curve is attracted to target by the curve movement in description vector field as caused by g gradient
The border of object, Section 3 description is mapped the curvature flow of g weightings by Gradient Features, and effect is that smooth boundary supports weak part
Curve；
In level set,WithIdentical curvilinear motion is described, if φ is that have symbolic measurement,
I.e.The derivative that level set imbedding function is changed over time is
G is decreasing function,C is control slope in formula.
Multi-mode brain tumor image blend segmenting device, is provided with computer, the MRI for handling T1C, T2 and FLAIR
Image, computer includes following module：Processing and initial segmentation module are filtered to image using medium filtering, pre- place is obtained
Manage image；Pretreatment image input linear Fusion Module afterwards；Fused image inputs FFCM cluster segmentation modules, automatically extracts
The wherein larger region of gray value, obtained tumour less divided region；Progress mixed-level collection segmentation resume module is inputted again, is obtained
To final result.
The features of the present invention and beneficial effect are：
The present invention splits the MRI figures with brain tumor by improved multi-mode brain tumor image blend partitioning algorithm
Picture, compared with the method that some are classical, its advantage is mainly reflected in：
1) novelty：Split the MRI image with brain tumor, root using FFCM algorithms and mixed-level set algorithm first
According to the characteristic of MRI brain tumor images, with reference to the advantage of FFCM algorithms and mixed-level collection, reach to brain tumor image Fast Segmentation
Purpose.
2) validity：The region of less divided can be fast and effectively obtained using FFCM, less divided region is input to mixing
It can accelerate to restrain border in level set, so that efficiently against the defect of algorithm, while improving accuracy.
3) practicality：Present partitioning algorithm is combined due to the requirement for being all difficult to reach practicality and real-time, the present invention
Reasonable fraction between mixed-level set algorithm and FFCM algorithms, so as to overcome the defect of some algorithms, increases to a certain extent
The practicality of quantity algorithm.And do further discussion for automatic segmentation brain tumor technology.
Brief description of the drawings：
Fig. 1 is the flow chart that improved multi-mode brain tumor image blend partitioning algorithm of the invention splits MRI brain tumors.
Fig. 2 is similarity factor (Dice) of the inventive algorithm in 10 brain tumor images.
Embodiment
1 is theoretical based on histogrammic quick FCM
FFCM core concept is that pixel intensity value seeks suitable degree of membership and cluster centre so that letter is expended in cluster
Several variances and iteration error are minimum, the weighted accumulation that the value for expending function, which is pixel, to be estimated to the norm of cluster centre 2 with.FFCM
Cluster segmentation algorithm is that by fuzzy C-mean algorithm theory data are divided into c classes, for a width M × N images, it is assumed that { hi, i=1,
2 ..., n }, n=M × N, hiIt is the set that the pixel intensity value in image histogram is constituted.{vj, j=1,2 ..., c } it is cluster
The set that center is constituted, and μj(hi) it is hiIt is under the jurisdiction of the membership function of j classes, so FFCM object function is
And
In formula, | | | | 2 norms are represented, b is one and is more than 1 constant, controls the fuzziness of cluster result.To calculate Jf's
Minimum value so that
It can be derived from formula (1) (2)
Iterative (3) (4) if meeting stopping criterion for iteration, t ＞ T orThen stop, wherein t tables
Show iterations, ε is off condition, and T represents maximum iteration.After algorithm terminates, pixel is divided by maximum membership degree
Class, if μj(hi) ＞ μj(hk), then by hiIt is classified as jth class region, k=1,2 ..., c；i≠k.
2 mixed model level set principles
Osher and Sethian proposes Level Set Method, and low-dimensional curve is expressed as to the zero level collection of higher-dimension curved surface.Arbitrarily
At the moment, only it is to be understood that φ can obtain its zero level collection curve, wherein φ (X) is level set function.Change to handle curved surface
Change, as long as being then that can be achieved for the zero level collection of the higher one-dimensional space by the representation of a surface.Compared with particle model and parameter model,
Level Set Models have significant advantage, and concept and Numerical Implementation, which are adapted to, solves the problems, such as any size；It can readily determine that
The inside and outside region of active contour.
Because the MRI image of brain tumor is extremely complex, the present invention is integrated using a kind of dividing method based on level set
Border issue and area information, while making up the border issue that FFCM algorithms are left., it is necessary to illustrate several before descriptive model
Parameter, imbedding function φ null set is used to representing active contour C={ X | φ (X)=0 }, and the point of profile inside/outside has positive/negative φ
Value.The function definition minimized the need for proposing
I is image to be split in formula (5),It is the boundary characteristic figure related to image gradient, wherein g is
Decreasing function,A is control slope in formula, and H (φ) is Heaviside function, and Ω is image area, and α and β are
Weight is predefined to balance two.μ is the predefined parameter of the lower limit for the gray level for indicating destination object.It is gradient operator.
Wherein,To point to the normal vector of curved exterior.Therefore the dominant curve evolution partial differential equation of active contour are
In formulaAnd curvature<·,·>For inner product.Due to there was only the several of curve
What change is interested in segmentation, so from formula (6) it may be noted that the institute on curve a little all moves in the normal direction
It is dynamic.Formula (6) Section 1 is the biography of the contractile motion for the extension movement and exterior section for describing the curved portion inside destination object
Item is broadcast, Section 2 is advective term, curve is attracted to target by the curve movement in description vector field as caused by g gradient
The border of object, Section 3 description is mapped the curvature flow of g weightings by Gradient Features, and effect is that smooth boundary supports weak part
Curve.
In level set,WithIdentical curvilinear motion is described, if φ is that have symbolic measurement,
I.e.The derivative that level set imbedding function is changed over time is
Table 1
Table 1 is result of the inventive algorithm to 47 width brain tumor image procossings, wherein Jaccard coefficients, similarity factor
And recall is the most frequently used evaluation criterion (Dice).
Because MRI brain tumors image is of low quality in itself, it is impossible to be used in directly split, so the present invention uses Fig. 1 first
Shown mixing partitioning algorithm framework, because three kinds of modality images can provide part uncorrelated and complementary letter for lesion segmentation
Breath, the MRI image of present invention input Three models first includes T1C, T2 and FLAIR.Because there is certain make an uproar in itself in image
Sound, is filtered processing to image using medium filtering and initial segmentation obtains pretreatment image, afterwards using linear fusion；Melt
Close image and carry out FFCM clustering algorithm segmentations, automatically extract the larger region of wherein gray value, obtained tumour less divided region
Mixed-level collection segmentation is carried out, segmentation result is finally evaluated.FFCM clustering algorithms are combined with mixed model level set algorithm, a side
Face accelerates the speed of level set algorithm itself, and initial value is relied on not while also improving mixed model level set algorithm
Foot.To test out suitable integration percentage, the present invention finally show that appropriate ratio is by a variety of ratio test comparisons
FLAIR:T2:T1C=5:4:1.
Claims (4)
1. a kind of multi-mode brain tumor image blend dividing method, it is characterized in that, input first Three models include T1C, T2 and
FLAIR MRI image, is filtered processing to image using medium filtering and initial segmentation obtains pretreatment image, adopt afterwards
Use linear fusion；FFCM cluster segmentations are carried out to fused images again, the larger region of wherein gray value is automatically extracted, what is obtained is swollen
Knurl less divided region carries out mixed-level collection segmentation.
2. multi-mode brain tumor image blend dividing method as claimed in claim 1, it is characterized in that, FFCM cluster segmentations are specific
Step is that data are divided into c classes by fuzzy C-mean algorithm theory, for a width M × N images, if { hi, i=1,2 ..., n }, n=
M × N, hiIt is the set that the pixel intensity value in image histogram is constituted, wherein M and N are the length and width of image, { vj, j=1,
2 ..., c } it is the set that cluster centre is constituted, and μj(hi) it is hiIt is under the jurisdiction of the membership function of j classes, | | | | represent 2 norms, b
It is one and is more than 1 constant, then：
<mrow>
<msub>
<mi>v</mi>
<mi>j</mi>
</msub>
<mo>=</mo>
<mfrac>
<mrow>
<munderover>
<mo>&Sigma;</mo>
<mrow>
<mi>i</mi>
<mo>=</mo>
<mn>1</mn>
</mrow>
<mi>n</mi>
</munderover>
<msup>
<mrow>
<mo>&lsqb;</mo>
<msub>
<mi>&mu;</mi>
<mi>j</mi>
</msub>
<mrow>
<mo>(</mo>
<msub>
<mi>h</mi>
<mi>i</mi>
</msub>
<mo>)</mo>
</mrow>
<mo>&rsqb;</mo>
</mrow>
<mi>b</mi>
</msup>
<msub>
<mi>h</mi>
<mi>i</mi>
</msub>
</mrow>
<mrow>
<munderover>
<mo>&Sigma;</mo>
<mrow>
<mi>i</mi>
<mo>=</mo>
<mn>1</mn>
</mrow>
<mi>n</mi>
</munderover>
<msup>
<mrow>
<mo>&lsqb;</mo>
<msub>
<mi>&mu;</mi>
<mi>j</mi>
</msub>
<mrow>
<mo>(</mo>
<msub>
<mi>h</mi>
<mi>i</mi>
</msub>
<mo>)</mo>
</mrow>
<mo>&rsqb;</mo>
</mrow>
<mi>b</mi>
</msup>
</mrow>
</mfrac>
<mo>,</mo>
<mi>j</mi>
<mo>=</mo>
<mn>1</mn>
<mo>,</mo>
<mn>2</mn>
<mo>,</mo>
<mo>...</mo>
<mo>,</mo>
<mi>c</mi>
<mo>-</mo>
<mo>-</mo>
<mo>-</mo>
<mrow>
<mo>(</mo>
<mn>3</mn>
<mo>)</mo>
</mrow>
</mrow>
<mrow>
<msub>
<mi>&mu;</mi>
<mi>j</mi>
</msub>
<mrow>
<mo>(</mo>
<msub>
<mi>h</mi>
<mi>i</mi>
</msub>
<mo>)</mo>
</mrow>
<mo>=</mo>
<mfrac>
<msup>
<mrow>
<mo>(</mo>
<mn>1</mn>
<mo>/</mo>
<mo>|</mo>
<mo>|</mo>
<msub>
<mi>h</mi>
<mi>i</mi>
</msub>
<mo>-</mo>
<msub>
<mi>v</mi>
<mi>j</mi>
</msub>
<mo>|</mo>
<msup>
<mo>|</mo>
<mn>2</mn>
</msup>
<mo>)</mo>
</mrow>
<mrow>
<mn>1</mn>
<mo>/</mo>
<mrow>
<mo>(</mo>
<mi>b</mi>
<mo>-</mo>
<mn>1</mn>
<mo>)</mo>
</mrow>
</mrow>
</msup>
<mrow>
<munderover>
<mo>&Sigma;</mo>
<mrow>
<mi>k</mi>
<mo>=</mo>
<mn>1</mn>
</mrow>
<mi>c</mi>
</munderover>
<msup>
<mrow>
<mo>(</mo>
<mn>1</mn>
<mo>/</mo>
<mo>|</mo>
<mo>|</mo>
<msub>
<mi>h</mi>
<mi>i</mi>
</msub>
<mo>-</mo>
<msub>
<mi>v</mi>
<mi>k</mi>
</msub>
<mo>|</mo>
<msup>
<mo>|</mo>
<mn>2</mn>
</msup>
<mo>)</mo>
</mrow>
<mrow>
<mn>1</mn>
<mo>/</mo>
<mrow>
<mo>(</mo>
<mi>b</mi>
<mo>-</mo>
<mn>1</mn>
<mo>)</mo>
</mrow>
</mrow>
</msup>
</mrow>
</mfrac>
<mo>,</mo>
<mi>i</mi>
<mo>=</mo>
<mn>1</mn>
<mo>,</mo>
<mn>2</mn>
<mo>,</mo>
<mo>...</mo>
<mo>,</mo>
<mi>n</mi>
<mo>;</mo>
<mi>j</mi>
<mo>=</mo>
<mn>1</mn>
<mo>,</mo>
<mn>2</mn>
<mo>,</mo>
<mo>...</mo>
<mo>,</mo>
<mi>c</mi>
<mo>-</mo>
<mo>-</mo>
<mo>-</mo>
<mrow>
<mo>(</mo>
<mn>4</mn>
<mo>)</mo>
</mrow>
</mrow>
Iterative (3) (4) if meeting stopping criterion for iteration, t>T orThen stop, wherein t represents to change
Generation number, ε is off condition, and T represents maximum iteration, after algorithm terminates, and pixel is classified by maximum membership degree, if
μj(hi)>μj(hk), then by hiIt is classified as jth class region, k=1,2 ..., c；i≠k.
The segmentation of mixed-level collection comprises the concrete steps that, imbedding function φ null set is used to representing active contour C={ X | φ (X)=0 },
The point of profile inside/outside has positive/negative φ values, the function definition minimized the need for proposing：
<mrow>
<mi>M</mi>
<mi>i</mi>
<mi>n</mi>
<mi>&epsiv;</mi>
<mrow>
<mo>(</mo>
<mi>&phi;</mi>
<mo>)</mo>
</mrow>
<mo>=</mo>
<mo>-</mo>
<mi>&alpha;</mi>
<munder>
<mo>&Integral;</mo>
<mi>&Omega;</mi>
</munder>
<mrow>
<mo>(</mo>
<mi>I</mi>
<mo>-</mo>
<mi>&mu;</mi>
<mo>)</mo>
</mrow>
<mi>H</mi>
<mrow>
<mo>(</mo>
<mi>&phi;</mi>
<mo>)</mo>
</mrow>
<mi>d</mi>
<mi>&Omega;</mi>
<mo>+</mo>
<mi>&beta;</mi>
<munder>
<mo>&Integral;</mo>
<mi>&Omega;</mi>
</munder>
<mi>g</mi>
<mo>|</mo>
<mo>&dtri;</mo>
<mi>H</mi>
<mrow>
<mo>(</mo>
<mi>&phi;</mi>
<mo>)</mo>
</mrow>
<mo>|</mo>
<mi>d</mi>
<mi>&Omega;</mi>
<mo>-</mo>
<mo>-</mo>
<mo>-</mo>
<mrow>
<mo>(</mo>
<mn>5</mn>
<mo>)</mo>
</mrow>
</mrow>
I is image to be split in formula (5),It is the boundary characteristic figure related to image gradient,It is that gradient is calculated
Son, H (φ) is Heaviside function, and Ω is image area, and α and β are to predefine weight to balance two, and μ is to indicate destination object
Gray level lower limit predefined parameter；
Wherein,To point to the normal vector of curved exterior, therefore the dominant curve evolution partial differential equation of active contour are
<mrow>
<msub>
<mi>C</mi>
<mi>t</mi>
</msub>
<mo>=</mo>
<mi>&alpha;</mi>
<mrow>
<mo>(</mo>
<mi>I</mi>
<mo>-</mo>
<mi>&mu;</mi>
<mo>)</mo>
</mrow>
<mover>
<mi>N</mi>
<mo>&RightArrow;</mo>
</mover>
<mo>-</mo>
<mi>&beta;</mi>
<mo><</mo>
<mo>&dtri;</mo>
<mi>g</mi>
<mo>&CenterDot;</mo>
<mover>
<mi>N</mi>
<mo>&RightArrow;</mo>
</mover>
<mo>></mo>
<mo>+</mo>
<mi>&beta;</mi>
<mi>g</mi>
<mi>&kappa;</mi>
<mover>
<mi>N</mi>
<mo>&RightArrow;</mo>
</mover>
<mo>-</mo>
<mo>-</mo>
<mo>-</mo>
<mrow>
<mo>(</mo>
<mn>6</mn>
<mo>)</mo>
</mrow>
</mrow>
In formulaAnd curvature<·,·>For inner product；Due to there was only the change of the geometry of curve
It is interested to change in segmentation, so may be noted that the institute on curve a little all moves in the normal direction from formula (6), formula
(6) Section 1 is the propagation of the contractile motion for the extension movement and exterior section for describing the curved portion inside destination object,
Section 2 is advective term, the curve movement in description vector field as caused by g gradient, and curve is attracted into destination object
Border, Section 3 description is mapped the curvature flow of g weightings by Gradient Features, and effect is the curve that smooth boundary supports weak part；
In level set,WithIdentical curvilinear motion is described, if φ is that have symbolic measurement, i.e.,The derivative that level set imbedding function is changed over time is
<mrow>
<msub>
<mi>&phi;</mi>
<mi>t</mi>
</msub>
<mo>=</mo>
<mi>&alpha;</mi>
<mrow>
<mo>(</mo>
<mi>I</mi>
<mo>-</mo>
<mi>&mu;</mi>
<mo>)</mo>
</mrow>
<mo>+</mo>
<mi>&beta;</mi>
<mi>d</mi>
<mi>i</mi>
<mi>v</mi>
<mrow>
<mo>(</mo>
<mi>g</mi>
<mo>&dtri;</mo>
<mi>&phi;</mi>
<mo>)</mo>
</mrow>
<mo>-</mo>
<mo>-</mo>
<mo>-</mo>
<mo>(</mo>
<mn>7</mn>
<mo>)</mo>
<mo>.</mo>
</mrow>
1
3. multi-mode brain tumor image blend dividing method as claimed in claim 1, it is characterized in that, g is decreasing function,C is control slope in formula.
4. a kind of multi-mode brain tumor image blend segmenting device, it is characterized in that, be provided with computer, for handle T1C, T2 and
FLAIR MRI image, computer includes following module：Processing and initial segmentation mould are filtered to image using medium filtering
Block, obtains pretreatment image；Pretreatment image input linear Fusion Module afterwards；Fused image inputs FFCM cluster segmentation moulds
Block, automatically extracts the larger region of wherein gray value, obtained tumour less divided region；Progress mixed-level collection segmentation is inputted again
Resume module, obtains final result.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710270990.4A CN107154047A (en) | 2017-04-24 | 2017-04-24 | Multi-mode brain tumor image blend dividing method and device |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710270990.4A CN107154047A (en) | 2017-04-24 | 2017-04-24 | Multi-mode brain tumor image blend dividing method and device |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN107154047A true CN107154047A (en) | 2017-09-12 |
Family
ID=59793903
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201710270990.4A Pending CN107154047A (en) | 2017-04-24 | 2017-04-24 | Multi-mode brain tumor image blend dividing method and device |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN107154047A (en) |
Cited By (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107845098A (en) * | 2017-11-14 | 2018-03-27 | 南京理工大学 | Liver cancer image full-automatic partition method based on random forest and fuzzy clustering |
| CN107909577A (en) * | 2017-10-18 | 2018-04-13 | 天津大学 | Fuzzy C-mean algorithm continuous type max-flow min-cut brain tumor image partition method |
| CN108416792A (en) * | 2018-01-16 | 2018-08-17 | 辽宁师范大学 | Medical computer tomoscan image dividing method based on movable contour model |
| CN108961214A (en) * | 2018-05-31 | 2018-12-07 | 天津大学 | Brain tumor MRI three-dimensional dividing method based on improved continuous type maximum-flow algorithm |
| CN109671054A (en) * | 2018-11-26 | 2019-04-23 | 西北工业大学 | The non-formaldehyde finishing method of multi-modal brain tumor MRI |
| CN109685767A (en) * | 2018-11-26 | 2019-04-26 | 西北工业大学 | A kind of bimodal brain tumor MRI dividing method based on Cluster-Fusion algorithm |
| CN110309827A (en) * | 2019-05-06 | 2019-10-08 | 上海海洋大学 | A kind of area of edema parted pattern based on OCT image |
| CN112634280A (en) * | 2020-12-08 | 2021-04-09 | 辽宁师范大学 | MRI image brain tumor segmentation method based on energy functional |
| CN112686916A (en) * | 2020-12-28 | 2021-04-20 | 淮阴工学院 | Curved surface reconstruction system based on heterogeneous multi-region CT scanning data processing |
| CN116363160A (en) * | 2023-05-30 | 2023-06-30 | 杭州脉流科技有限公司 | CT perfusion image brain tissue segmentation method and computer equipment based on level set |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106127794A (en) * | 2016-07-29 | 2016-11-16 | 天津大学 | Based on probability FCM algorithm MRI tumor image dividing method and system |
| CN106296699A (en) * | 2016-08-16 | 2017-01-04 | 电子科技大学 | Cerebral tumor dividing method based on deep neural network and multi-modal MRI image |
-
2017
- 2017-04-24 CN CN201710270990.4A patent/CN107154047A/en active Pending
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106127794A (en) * | 2016-07-29 | 2016-11-16 | 天津大学 | Based on probability FCM algorithm MRI tumor image dividing method and system |
| CN106296699A (en) * | 2016-08-16 | 2017-01-04 | 电子科技大学 | Cerebral tumor dividing method based on deep neural network and multi-modal MRI image |
Non-Patent Citations (4)
| Title |
|---|
| SHAHEEN AHMED,ET.AL: "Efficacy of Texture, Shape, and Intensity Feature Fusion for Posterior-Fossa Tumor Segmentation in MRI", 《IEEE TRANSACTIONS ON INFORMATION TECHNOLOGY IN BIOMEDICINE》 * |
| YAN ZHANG，ET.AL: "Medical Image Segmentation Using New Hybrid Level-Set Method", 《FIFTH INTERNATIONAL CONFERENCE BIOMEDICAL VISUALIZATION: INFORMATION VISUALIZATION IN MEDICAL AND BIOMEDICAL INFORMATICS》 * |
| 张腾达: "基于模糊水平集的脑肿瘤MR图像分割方法", 《现代电子技术》 * |
| 王黎明: "基于模糊C均值算法的医学图像分割研究", 《硕士学位论文全文数据库》 * |
Cited By (15)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107909577A (en) * | 2017-10-18 | 2018-04-13 | 天津大学 | Fuzzy C-mean algorithm continuous type max-flow min-cut brain tumor image partition method |
| CN107845098A (en) * | 2017-11-14 | 2018-03-27 | 南京理工大学 | Liver cancer image full-automatic partition method based on random forest and fuzzy clustering |
| CN108416792B (en) * | 2018-01-16 | 2021-07-06 | 辽宁师范大学 | Medical computed tomography image segmentation method based on active contour model |
| CN108416792A (en) * | 2018-01-16 | 2018-08-17 | 辽宁师范大学 | Medical computer tomoscan image dividing method based on movable contour model |
| CN108961214A (en) * | 2018-05-31 | 2018-12-07 | 天津大学 | Brain tumor MRI three-dimensional dividing method based on improved continuous type maximum-flow algorithm |
| CN109685767A (en) * | 2018-11-26 | 2019-04-26 | 西北工业大学 | A kind of bimodal brain tumor MRI dividing method based on Cluster-Fusion algorithm |
| CN109671054A (en) * | 2018-11-26 | 2019-04-23 | 西北工业大学 | The non-formaldehyde finishing method of multi-modal brain tumor MRI |
| CN110309827A (en) * | 2019-05-06 | 2019-10-08 | 上海海洋大学 | A kind of area of edema parted pattern based on OCT image |
| CN110309827B (en) * | 2019-05-06 | 2022-10-14 | 上海海洋大学 | Edema region segmentation model based on OCT image |
| CN112634280A (en) * | 2020-12-08 | 2021-04-09 | 辽宁师范大学 | MRI image brain tumor segmentation method based on energy functional |
| CN112634280B (en) * | 2020-12-08 | 2023-06-16 | 辽宁师范大学 | MRI image brain tumor segmentation method based on energy functional |
| CN112686916A (en) * | 2020-12-28 | 2021-04-20 | 淮阴工学院 | Curved surface reconstruction system based on heterogeneous multi-region CT scanning data processing |
| CN112686916B (en) * | 2020-12-28 | 2024-04-05 | 淮阴工学院 | Curved surface reconstruction system based on heterogeneous multi-region CT scanning data processing |
| CN116363160A (en) * | 2023-05-30 | 2023-06-30 | 杭州脉流科技有限公司 | CT perfusion image brain tissue segmentation method and computer equipment based on level set |
| CN116363160B (en) * | 2023-05-30 | 2023-08-29 | 杭州脉流科技有限公司 | CT perfusion image brain tissue segmentation method and computer equipment based on level set |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN107154047A (en) | Multi-mode brain tumor image blend dividing method and device | |
| CN107644420B (en) | Blood vessel image segmentation method based on centerline extraction and nuclear magnetic resonance imaging system | |
| Li et al. | Integrating spatial fuzzy clustering with level set methods for automated medical image segmentation | |
| Zhang et al. | 3D skeletonization feature based computer-aided detection system for pulmonary nodules in CT datasets | |
| WO2017092182A1 (en) | Method for automatically recognizing liver tumor type in ultrasonic image | |
| CN104268873B (en) | Breast tumor partition method based on nuclear magnetic resonance images | |
| Liu et al. | A CADe system for nodule detection in thoracic CT images based on artificial neural network | |
| McClure et al. | A novel NMF guided level-set for DWI prostate segmentation | |
| Rampun et al. | Automated 2d fetal brain segmentation of mr images using a deep u-net | |
| Dubey et al. | Semi-automatic segmentation of MRI brain tumor | |
| Ma et al. | A novel deep learning framework for automatic recognition of thyroid gland and tissues of neck in ultrasound image | |
| US20220301224A1 (en) | Systems and methods for image segmentation | |
| Guo et al. | A novel myocardium segmentation approach based on neutrosophic active contour model | |
| Yuan et al. | ResD-Unet research and application for pulmonary artery segmentation | |
| Jayadevappa et al. | A hybrid segmentation model based on watershed and gradient vector flow for the detection of brain tumor | |
| Farag et al. | Variational approach for segmentation of lung nodules | |
| Badura | Virtual bacterium colony in 3D image segmentation | |
| Jia et al. | A novel lung nodules detection scheme based on vessel segmentation on CT images | |
| Bahreini et al. | Gradient vector flow snake segmentation of breast lesions in dynamic contrast-enhanced MR images | |
| CN106340022A (en) | Image segmentation method based on region correlation | |
| Liming et al. | A new algorithm of automatic lung parenchyma segmentation based on CT images | |
| Dai et al. | Lung segmentation with improved graph cuts on chest CT images | |
| Nomura et al. | Can the spherical gold standards be used as an alternative to painted gold standards for the computerized detection of lesions using voxel-based classification? | |
| Walczak et al. | Segmenting lungs from whole-body CT scans | |
| Wang et al. | Breast Lesion Segmentation in Ultrasound Images by CDeep3M |
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 | ||
| WD01 | Invention patent application deemed withdrawn after publication | ||
| WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20170912 |