CN106204624A - A kind of method detecting picture structure change - Google Patents
A kind of method detecting picture structure change Download PDFInfo
- Publication number
- CN106204624A CN106204624A CN201610594607.6A CN201610594607A CN106204624A CN 106204624 A CN106204624 A CN 106204624A CN 201610594607 A CN201610594607 A CN 201610594607A CN 106204624 A CN106204624 A CN 106204624A
- Authority
- CN
- China
- Prior art keywords
- maximum
- thickness
- pixel
- fill
- spheroid
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Withdrawn
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/0002—Inspection of images, e.g. flaw detection
- G06T7/0012—Biomedical image inspection
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/60—Analysis of geometric attributes
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/10—Image acquisition modality
- G06T2207/10072—Tomographic images
- G06T2207/10081—Computed x-ray tomography [CT]
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/30—Subject of image; Context of image processing
- G06T2207/30004—Biomedical image processing
- G06T2207/30008—Bone
-
- 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
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Computer Vision & Pattern Recognition (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Geometry (AREA)
- Health & Medical Sciences (AREA)
- General Health & Medical Sciences (AREA)
- Medical Informatics (AREA)
- Nuclear Medicine, Radiotherapy & Molecular Imaging (AREA)
- Radiology & Medical Imaging (AREA)
- Quality & Reliability (AREA)
- Apparatus For Radiation Diagnosis (AREA)
Abstract
The invention discloses a kind of method detecting picture structure change, include following aspect successively: 1) or the improvement of maximum spheroid fill method circular to the maximum measuring two dimension or three dimensional structure thickness；2) maximum circular or largest square and maximum spheroid or maximum cube fill method is utilized to carry out the measurement of 2 and 3 dimensional organization thickness；3) linear processes is utilized to return the slight change of detection picture structure；4) two dimension significance plane picture is utilized to be analyzed the selection in region.The method increase the susceptiveness of detection picture structure slight change, there is good development and application prospect.
Description
Technical field
The invention belongs to technical field of image processing, relate to a kind of method detecting picture structure slight change.
Technical background
Osteoporosis is the senile metabolic disease of a kind of serious harm human health, and it mainly shows as bone amount and subtracts
Less, micro-architectural deterioration and bone fragility increase, thus cause patients with fractures's risk to rise.The diagnosis of osteoporosis relies primarily on survey
Amount bone amount and the change of two aspects of bone micro-structure.The instrument being currently used for measuring bone density mainly has Dual-energy X-rays absorptionmetry
(DEXA), bone quantitive CT (QCT) or Peripheral quantitative CT (pQCT).Dual-energy X-rays absorptionmetry (DEXA) measures bone density (BMD)
Change be the goldstandard of current diagnosis osteoporosis.Owing to bone quantitive CT is the highest to the accuracy of osseous tissue, therefore can not
It is used for carrying out the minor alteration of follow-up assessment bone density or micro structure.Although DEXA can measure the bone density of whole body any part,
But cannot distinguish between cortical bone density and cancellous bone density, therefore can not the change of effective reflected measurement position bone structure.PQCT is then
Can not only measure total BMD, and can distinguish cortical bone and spongy bone, therefore DEXA is compared in the change to detection bone density and structure
Sensitiveer.But can not effectively assess risk of bone fracture only according to bone density at present, and the change of bone density can not be comprehensive
The change of reflection bone micro-structure.Therefore the detection method of comprehensive bone density and trabecular bone microstructure change may be to osteoporosis
The assessment of diagnosis and risk of bone fracture significant.
Micro-CT scanning (MicroCT) is a kind of high accuracy three-dimensional CT imaging technique, and its spatial resolution can reach several micro-
Rice.Owing in the toy osseous tissue such as rat and mice, the thickness of trabecular bone is at tens micrometer ranges, conventional CT scan imaging is not
Can effectively show the fine structure of skeleton, therefore MicroCT has wide in terms of assessment toy bone amount and bone micro-structure change
General application.Due to proximal tibia and the trabecular bone structure-rich of the following certain area of distal femur epiphyseal growth plate, and sclerotin
The loose loss that can cause this region trabecular bone structure is destroyed, and therefore under proximal tibia and distal femur epiphyseal line, certain area is
MicroCT analyzes the preferred region that toy trabecular bone micro structure changes.The data of trabecular bone micro structure are divided by MicroCT at present
Analysis program is as follows: chooses the region of below the reference plane certain distance length-specific containing trabecular bone, successively chooses trabecular bone, so
Rear measurement is at the bone amount of this intra-zone, bone volume and the Domain Volume chosen, and then calculates bone density (BMD) and diaphysis
Long-pending density (BV/TV).Different model is finally utilized to calculate the morphometry information such as width of trabecular bone.Little at MicroCT
In beam bone micro-structure data analysis process, interpretation of result is had a major impact by the selection of analyzed area, but there is presently no system
The standard in the selection analysis region of, therefore in the document delivered, different research worker select different regions to be analyzed.
Due to trabecular bone skewness below epiphyseal line growth plate, approximate and increase with the distance of distance epiphyseal line and reduce, therefore with
Trabecular bone quantity will assist in raising detection bone micro-structure change along the selection of the analyzed area based on long bone radial distribution
Sensitivity.
Trabecular bone thickness is important Testing index based on trabecular bone structure.But at present for two dimension or 3-D view
Method for measuring thickness all have potential defect, cause measure trabecular bone thickness can not reflect the tight of osteoporosis symptoms comprehensively
Weight degree.Therefore, more accurate trabecular bone thickness measuring method is significant to the diagnosis of osteoporosis.
Summary of the invention
In order to solve above-mentioned technical problem, the present invention provides can a kind of method detecting picture structure slight change.
The technical solution adopted in the present invention is: a kind of detect picture structure change method, it is characterised in that include with
Lower step:
Step 1: for structure to be determined, measures two-dimensional structure or the thickness of three dimensional structure；
Step 2: calculate two-dimensional structure or the average thickness of three dimensional structure；
Step 3: successively calculate the parameters of structure to be determined；
Step 4: the parameters of structure to be determined is analyzed, the slight change of detection picture structure；
Step 5: structure to be determined utilizes two dimension Saliency maps picture select that different grouping is had significant difference and divides
Analysis region；
Step 6: to selected analyzed area, according to the distribution characteristics of parameters, utilize the parametric test of standard or non-
Parametric test method carries out the significance analysis of difference to different grouping.
As preferably, step 1 is the thickness utilizing maximum circular fill method to measure two-dimensional structure, described maximum circular
Fill method is filled maximum circular in being included in structure to be determined, the thickness corresponding to point in structure is for being completely contained in structure
Internal and comprise the maximum circular diameter of this point；Any one pixel be considered as the length of side be the square of 1 length in pixels,
Maximum circle can be with the center of any pixel as the center of circle, it is also possible to is with four angles of any pixel as the center of circle；Should
Maximum circular radius can be integer, it is also possible to be non-integer.
As preferably, step 1 is utilize the maximum circular or thickness of largest square fill method measurement two-dimensional structure,
It implements and includes following sub-step:
Step is A.1: utilizes maximum circular fill method to measure the thickness of two-dimensional structure, i.e. fills in structure to be determined
Big circular, the thickness corresponding to point in structure is to be completely contained in inside configuration and the maximum circular diameter comprising this point；
Any one pixel be considered as the length of side be the square of 1 length in pixels, maximum circle can be with in any pixel
The heart is the center of circle, it is also possible to be with four angles of any pixel as the center of circle；The radius of this maximum circle can be integer, it is also possible to
It it is non-integer；
Step is A.2: fill largest square in structure to be determined, utilizes largest square to fill and obtain in computation structure
The maximum gauge of every bit；
Step is A.3: measure the thickness of two-dimensional structure；In structure to be determined, the thickness of every bit is that corresponding point utilizes maximum
The circular maximum filled in the maximum gauge obtained and two values utilizing largest square to fill the maximum gauge obtained.
As preferably, step 1 is the thickness utilizing maximum spheroid fill method to measure three dimensional structure, described maximum spheroid
Fill method fills maximum spheroid in being included in structure to be determined, the thickness corresponding to point in structure is for being completely contained in structure
Internal and comprise the diameter of maximum spheroid of this point；Any one pixel be considered as the length of side be the cube of 1 length in pixels,
Maximum spheroid can be with the center of any pixel as the centre of sphere, it is also possible to is with eight angles of any pixel as the centre of sphere；Should
The radius of maximum spheroid can be integer, it is also possible to be non-integer.
As preferably, step 1 is the thickness utilizing maximum spheroid or maximum cube fill method to measure three dimensional structure,
It implements and includes following sub-step:
Step is B.1: utilizes maximum spheroid fill method to measure the thickness of three dimensional structure, i.e. fills in structure to be determined
Big spheroid, the thickness corresponding to point in structure is the diameter being completely contained in inside configuration and the maximum spheroid that comprises this point；
Any one pixel be considered as the length of side be the cube of 1 length in pixels, maximum spheroid can be with in any pixel
The heart is the centre of sphere, it is also possible to be with eight angles of any pixel as the centre of sphere；The radius of this maximum spheroid can be integer, it is also possible to
It it is non-integer；
Step is B.2: then fill maximum cube in structure to be determined, utilizes maximum cube to fill in computation structure
The maximum gauge of the every bit obtained；
Step is B.3: measure the thickness of three dimensional structure；In structure to be determined, the thickness of every bit is that corresponding point utilizes maximum
Spheroid is filled the maximum gauge obtained and utilizes maximum cube to fill the maximum in two values of the maximum gauge obtained.
As preferably, calculate two-dimensional structure or the average thickness of three dimensional structure described in step 2, first calculate any thickness
The two-dimensional structure corresponding to part or three dimensional structure length, then calculate two-dimensional structure or the total length of three dimensional structure, finally
Calculate two-dimensional structure or the average thickness of three dimensional structure.
As preferably, calculate two-dimensional structure or the average thickness of three dimensional structure described in step 2, utilize the side reducing noise
Method calculates；First calculate the structure length corresponding to the part of any thickness, then ignore the thickness of maximum affected by noise
Degree and the length of correspondence thereof, finally calculate the total length of structure of the point without maximum affected by noise, so be calculated through
The average thickness of the structure division of correction.
As preferably, parameter described in step 3 includes thickness, density, bulk density etc..
As preferably, the process that implements of step 3 is, utilizes identification or manual identification methods analyst automatically interested
Structural region, then carries out the calculating of various structural parameters to this region.
As preferably, implementing of step 4 includes following sub-step:
Step 4.1: with the distance apart from reference plane as X-axis, maps with the parameters value of layered method for Y-axis；
Step 4.2: utilize linear regression or nonlinear regression model (NLRM) to carry out the matching of parameters data according to its distribution；
Step 4.3: the parameters of matched curve is utilized linear regression or nonlinear regression model (NLRM) to different grouping it
Between difference carry out the significance analysis added up.
As preferably, implementing of step 5 includes following sub-step:
Step 5.1: set up a two-dimensional array, this array includes initiateing all of any planar junction bundle from arbitrary plane
Possible analyzed area；
Step 5.2: be analyzed these regions respectively, obtains each structural parameters；
Step 5.3: utilize the statistical method of standard to calculate each structural parameters significance p of difference between different grouping
Value；
Step 5.4: according to the significance threshold value set, produce a corresponding two dimensional image, according to this two dimensional image, from
Image is selected the region with significance statistical discrepancy.
Thing in the present invention circular fill method maximum to utilization and maximum spheroid fill method measurement two and three dimensions image
The method of body structural thickness is improved, it is contemplated that institute's likely distributing position of the center of circle and the centre of sphere and they are a diameter of whole
All possible cases of number, improve the precision of thickness measure；The present invention propose that brand-new utilization maximum is circular or maximum just
Two and three dimensions objects in images thickness is analyzed by square fill method and maximum spheroid or maximum cube fill method
Method, these methods reduce the error of thickness measure；The present invention proposes a kind of distribution situation according to parameter and selects to divide
The method in analysis region.The above is improved, and can significantly improve the precision being analyzed target object, it will help improves and relies on
In the susceptiveness of the detection method of image, as being applied to bone loss and the assessment of bone structure destruction degree in osteoporosis, god
Through aspects such as the assessments that degeneration mesencephalic tissue changes.
Accompanying drawing explanation
Fig. 1 is the largest square filling of the embodiment of the present invention, maximum circular filling and maximum circle or largest square
Fill method schematic diagram；
Fig. 2 is the computational methods schematic diagram of the object thickness of the embodiment of the present invention；
Fig. 3 is the Comparative result schematic diagram of the two dimensional image method for measuring thickness of the embodiment of the present invention；
Fig. 4 is the Comparative result schematic diagram of the 3-D view method for measuring thickness of the embodiment of the present invention；
Fig. 5 is the interpretation of result that each analytical parameters of distal femur trabecular bone of the embodiment of the present invention is distributed along the femur longitudinal axis
Figure；
Fig. 6 is that the little molecule of nonlinear regression analysis osteogenic of the embodiment of the present invention is to trabecular bone BV, TV, BMC parameter
Impact analysis figure；
Fig. 7 is automatic trabecular bone subregion and the manual subregion shadow to trabecular bone BV, TV, BMC parameter of the embodiment of the present invention
Ring analysis chart；
Fig. 8 be the embodiment of the present invention trabecular bone micro structure different grouping between there is the choosing of analyzed area of significant difference
Select analysis chart；
Fig. 9 is the little molecule of osteogenic of the embodiment of the present invention impact analysis figure to trabecular bone thickness；
Figure 10 is the impact analysis figure that trabecular bone thickness is analyzed by the noise suppressed of the embodiment of the present invention；
Figure 11 is the selection of the trabecular bone analyzed area of the embodiment of the present invention impact analysis to trabecular bone microstructure analysis
Figure.
Detailed description of the invention
Understand and implement the present invention for the ease of those of ordinary skill in the art, below in conjunction with the accompanying drawings and embodiment is to this
Bright it is described in further detail, it will be appreciated that enforcement example described herein is merely to illustrate and explains the present invention, not
For limiting the present invention.
Image analysis processing includes the analyzing and processing of two dimensional image and the analyzing and processing of 3-D view.The most no matter to two dimension
The detection method of image or 3-D view structure slight change is the sensitiveest.The problem to be solved in the present invention is to provide one
The method of kind, it is possible to accurate quantification two dimensional image or the slight change of 3-D view structure.The present invention is possible not only to based on CT imaging
Be analyzed, and can based on nuclear magnetic resonance image, ultrasonoscopy etc. toy, big animal and human body carried out osseous tissue,
The structure mutation analysis of cerebral tissue, cardiovascular, pulmonary, kidney etc..The present invention is not limited only to medical image, but can apply to appoint
What image.Due to the fact that the precision that improve graphical analysis, be therefore beneficial to carry out osteoporosis, the heart based on structure change
The diagnosis of the diseases such as cerebrovascular disease, neurodegenerative diseases, kidney, pulmonary.
For achieving the above object, the invention provides the two kinds of methods measuring two-dimensional structure thickness, fill out including maximum circle
Fill method and maximum is circular or largest square fill method.Present invention below assumes that two-dimensional digital image is made up of pixel, often
One pixel is a square, You Yige center and four angles.
Each pixel in structure is calculated in being completely contained in this structure and comprises by maximum circular fill method
The maximum circle of this pixel.First the theory of the method was reported by Garrahan etc. in 1987.Hildebrand is equal to 1997
The method is expanded to do not rely on the situation of any model and carried out the realization of two and three dimensions THICKNESS CALCULATION by year.But this
The implementation method of a little reports only considered the center of circle situation at each pixel center point, and only considered round radius is integer
Situation, therefore its result of calculation has the biggest error.The method of the present invention not only allows for the situation of the center of circle point of the heart within the pixel,
And consider the center of circle situation at four angles of pixel.Meanwhile, the method for the present invention considers the feelings that diameter of a circle is odd number
Condition, i.e. round radius is non-integral situation.These improvement substantially increase the precision measuring structure.It mainly walks
Suddenly it is:
1) each to two dimensional image belongs to the pixel of structure to be determined, it is judged that whether it is positioned at the border of image.As
Fruit has the pixel of structure to be determined to be positioned at image boundary, then outer image border increases a row or column background pixel so that institute
Structure to be determined is had the most not to be positioned at the border of image.Initializing the maximum gauge that in structure to be determined, every bit is corresponding is 0.Right
Each pixel in structure to be determined, performs following steps 2) to step 7).
2) for pixel to be analyzed, set this pixel center as the center of circle, set the radius of circle as 0.5 length in pixels, if
The maximum gauge of the circle surely comprised is 0.
3) circle is drawn with the center of circle set and radius.
4) for each pixel comprised in circle, distance d in its central point distance center of circle is calculated.For all of in circle
D, less than the pixel of the radius of circle, checks whether that pixel is not belonging to structure to be determined, if it has, then enter step 6)；
If all d belong to structure to be determined less than the point of the radius of circle in circle, then enter step 5).
5) first recording current diameter of a circle, then set the center of circle constant, the radius length of circle increases by 0.5 pixel, enters
Rapid 3).
6) diameter of a circle recorded is with the point set as the center of circle, the maximum circle being included in structure to be determined
Diameter.For in circle, each distance away from the center of circle is less than or equal to the pixel of the radius of circle, if its current thickness is less than
Diameter of a circle, then updating its thickness is diameter of a circle.
7) respectively with four angles of pixel to be analyzed as the center of circle, initially sets the radius of circle as 0.5 length in pixels, set and wrap
The maximum gauge of the circle contained is 0, then repeats step 3) to step 6), respectively obtain and as the center of circle and be completely contained in four angles
The maximum circle of inside configuration to be determined, and update the maximum gauge of respective point in structure to be determined.
8) each pixel in structure to be determined is all by step 2) to step 7) calculate the center with this point and four angles thereof
For the maximum circular of the center of circle the thickness that updates reference point, therefore first the present invention has obtained treating that each point of geodesic structure is corresponding
Maximum gauge, then adds up all pixel numbers Si that each thickness (i) comprises, the then knot that each thickness (i) is corresponding
Structure length (Li) is Si/i.The average thickness of structure the most to be determined isWherein ∑ Si is to all in structure
Si sues for peace, and ∑ Li is to sue for peace all Li in structure.
First each pixel in structure is calculated and is completely contained in by maximum circular or largest square fill method
This structure is interior and comprises the maximum circular of this pixel, obtains each point in structure to be determined and utilizes circular filling of maximum to obtain
Thickness, then calculate the largest square being completely contained in this structure and comprising this pixel, obtain in structure to be determined
Each point utilizes largest square to fill the thickness obtained, and the thickness of each point in last structure to be determined is for utilizing
Big circular filling obtains thickness and largest square fills the maximum in the thickness obtained, and is the maximum thick of this pixel
Degree.Same structure utilize maximum circular and largest square filling calculating maximum gauge method yet there are no report at present simultaneously
Road.Maximum circular fill method can not accurately measure rectangle and containing change and the two-dimensional structure of sharp edges suddenly, maximum
Square fill method can not accurately measure the structure of circle, and maximum circle or largest square fill method are by choosing circle
The maximum thickness that shape is filled and calculated in square filling, can significantly reduce the error of irregularly shaped object thickness measure, increase
Precision that structure is measured.It mainly comprises the following steps:
1) each to two dimensional image belongs to the pixel of structure to be determined, it is judged that whether it is positioned at the border of image.As
Fruit has the pixel of structure to be determined to be positioned at image boundary, then outer image border increases a row or column so that all to be determined
Structure be not the most positioned at the border of image.
2) for structure to be determined, the maximum gauge of the every bit in this structure is calculated according to maximum circular fill method.
3) for structure to be determined, the maximum thick of every bit in this structure is calculated according to largest square fill method
Degree.It mainly comprises the following steps: for the every bit in structure to be determined, centered by four angles of the center of this point and this point respectively
In structure to be determined, fill largest square by the step that maximum circular filling is same, just then update these maximums filled
The thickness that the point of square inner is corresponding, makes any affected point the most only record all fillings maximum square including this point
Maximal side in shape.
4) maximum circular filling thickness is all chosen for each point in structure to be determined and largest square fills thickness
Maximum in degree, is the maximum gauge of this pixel.
5) add up all pixel numbers Si that each thickness (i) comprises, then the structure that each thickness (i) is corresponding is long
Degree (Li) is Si/i.The average thickness of structure the most to be determined isWherein ∑ Si is for ask Si all in structure
With, ∑ Li is to sue for peace all Li in structure.
For realizing the purpose of above-mentioned accurate measurement three dimensional structure thickness, the invention provides and measure the two of three dimensional structure thickness
The method of kind, including maximum spheroid fill method and maximum spheroid or maximum cube fill method.Present invention below assumes three-dimensional
Digital picture is made up of pixel, and each pixel is a cube, You Yige center and eight angles.
Each pixel in structure is calculated in being completely contained in this structure and comprises by maximum spheroid fill method
The maximum spheroid of this pixel.First the theory of the method is reported equal to 1997 by Hildebrand.But its method only considers
The situation of the centre of sphere point of the heart within the pixel, and only considered the situation that radius is integer of spheroid, therefore its result of calculation has
The biggest error.The method of the present invention not only allows for the situation of the centre of sphere point of the heart within the pixel, and considers the centre of sphere in pixel
The situation at eight angles.Meanwhile, it is non-whole that the method for the present invention considers the situation of a diameter of odd number of spheroid, the i.e. radius of spheroid
The situation of number.These improvement substantially increase the precision measuring three dimensional structure.It mainly comprises the following steps:
1) each to 3-D view belongs to the pixel of structure to be determined, it is judged that whether it is positioned at the border of image.As
Fruit has the pixel of structure to be determined to be positioned at image boundary, then outer image border increases a row or column or one layer of background pixel,
All structures to be determined are made the most not to be positioned at the border of image.Initialize the maximum gauge that in structure to be determined, every bit is corresponding
It is 0.For each pixel in structure to be determined, perform following steps 2) to step 7).
2) for pixel to be analyzed, set this pixel center as the centre of sphere, set the radius of ball as 0.5 length in pixels, if
The maximum gauge of the ball surely comprised is 0.
3) spheroid is drawn with the centre of sphere set and radius.
4) for each pixel comprised in spheroid, distance d of its central point distance centre of sphere is calculated.For institute in spheroid
Some d, less than the pixel of the radius of ball, check whether that pixel is not belonging to structure to be determined, if it has, then enter step
Rapid 6)；If all d belong to structure to be determined less than the point of the radius of ball in spheroid, then enter step 5).
5) first recording the diameter of current spheroid, then set the centre of sphere constant, the radius length of spheroid increases by 0.5 pixel,
Enter rapid 3).
6) diameter of the spheroid recorded is with the point set as the centre of sphere, the maximum spheroid being included in structure to be determined
Diameter.Each distance away from the centre of sphere in spheroid is less than or equal to the pixel of the radius of spheroid, if its current thickness
Degree less than the diameter of spheroid, then updates the diameter that its thickness is spheroid.
7) respectively with eight angles of pixel to be analyzed as the centre of sphere, initially sets the radius of spheroid as 0.5 length in pixels, setting
The maximum gauge of the spheroid comprised is 0, then repeats step 3) to step 6), respectively obtain and as the centre of sphere and wrap completely with eight angles
It is contained in the maximum spheroid of inside configuration to be determined, and updates the maximum gauge of respective point in structure to be determined.
8) each pixel in structure to be determined is all by step 2) to step 7) calculate the central point and eight with this pixel
Individual angle is the maximum spheroid of the centre of sphere and updates the thickness of reference point, and therefore first the present invention has obtained treating each point on geodesic structure
Corresponding maximum gauge, then adds up all pixel numbers Si that each thickness (i) comprises, then each thickness (i) is right
The structure length (Li) answered is Si/i2.The average thickness of structure the most to be determined isWherein ∑ Si is to knot
All Si summation in structure, ∑ Li is to sue for peace all Li in structure.
First each pixel in structure is calculated and is completely contained in by maximum spheroid or maximum cube fill method
In this structure and comprise the maximum spheroid of this pixel, obtain each point in structure to be determined and utilize maximum spheroid to fill to obtain
Thickness, then calculate the maximum cube being completely contained in this structure and comprising this pixel, obtain in structure to be determined
Each point utilizes maximum cube to fill the thickness obtained, and the thickness of each point in last structure to be determined is for utilizing
Big spheroid fills the thickness obtained and maximum cube fills the maximum in the thickness obtained, and is the maximum thick of this pixel
Degree.The method that same structure utilizes maximum spheroid and maximum cube filling calculate maximum gauge at present simultaneously yet there are no report
Road.Maximum spheroid fill method can not accurately measure cuboid or containing change and the three dimensional structure of sharp edges suddenly, maximum
Cube fill method can not accurately measure the structure of spheroid, and maximum spheroid or maximum cube fill method are by choosing
The maximum thickness that big spheroid is filled and calculated in maximum cube filling, can significantly reduce the mistake of irregularly shaped object thickness measure
Difference, adds the precision measuring structure.It mainly comprises the following steps:
1) each to 3-D view belongs to the pixel of structure to be determined, it is judged that whether it is positioned at the border of image.As
Fruit has the pixel of structure to be determined to be positioned at image boundary, then outer image border increases a row or column or one layer so that all
Structure to be determined is not the most positioned at the border of image.
2) for structure to be determined, the maximum gauge of the every bit in this structure is calculated according to maximum spheroid fill method.
3) for structure to be determined, the maximum thick of every bit in this structure is calculated according to maximum cube fill method
Degree.It mainly comprises the following steps: for the every bit in structure to be determined, centered by eight angles of the center of this point and this point respectively
In structure to be determined, fill same step by maximum spheroid fill maximum cube, then update these maximums filled and stand
The thickness that point within cube is corresponding, makes any affected point the most only record the maximum cube of all fillings including this point
Maximal side in body.
4) maximum spheroid filling thickness is all chosen for each point in structure to be determined and maximum cube fills thickness
Maximum in degree, is the maximum gauge of this pixel.
5) add up all pixel numbers Si that each thickness (i) comprises, then the structure that each thickness (i) is corresponding is long
Degree (Li) is Si/i2.The average thickness of structure the most to be determined isWherein ∑ Si is to Si all in structure
Summation, ∑ Li is to sue for peace all Li in structure.
Owing to carrying out object in two dimension or 3-D view sampling process, noise is inevitable, and noise is typically at figure
In random distribution in Xiang, therefore its to the structure influence with single pixel thick bigger and to the knot with many pixel thick
Structure impact is smaller.Therefore to reduce the noise impact on result, when the present invention calculates object thickness, only ignore maximum gauge
There is the point of or a few pixels thickness.Such process greatly improves detection small thickness between different disposal sample
The susceptiveness of change.
Ask for an interview Fig. 1, be that the largest square of the embodiment of the present invention is filled, maximum circular fill and maximum is circular or maximum just
Square fill method schematic diagram.A is the structure of thickness to be determined；B is for filling by maximum circular method；C is for using largest square
Fill；D is for filling by maximum circular or largest square.
Ask for an interview Fig. 2, the computational methods schematic diagram of the object thickness of the embodiment of the present invention.A is the X-Y scheme of object to be determined
Sheet schematic diagram, each pixel is represented by a little square, and object to be determined is made up of hypographous pixel；B is for maximum
Square method is filled, the maximum gauge that digitized representation this pixel calculated in pixel is corresponding；C is two-dimension picture thickness
Computational methods, wherein i be maximum gauge be the pixel of i, Si is that all maximum gauges are total number of the pixel of i in object,
Li be thickness be object length corresponding to the pixel of i, ∑ Si is all pixels that object comprises, and ∑ Li is the length of object,
Average thickness for two-dimensional bodies；D is the computational methods of three-dimensional body thickness.Assume that thickness is the pixel count of i in three-dimensional body
For Si, then Li be thickness be object length corresponding to the pixel of i, ∑ Si is all pixels that object comprises, and ∑ Li is object
Length,Average thickness for three-dimensional body.
Ask for an interview Fig. 3, the Comparative result of the two dimensional image method for measuring thickness of the embodiment of the present invention.Icon Square,
Circle, CirSquare and BoneJ represent respectively utilize largest square, maximum circular, maximum circular or largest square,
The thickness of BoneJ software measurement standard two-dimensional image.A be various method to the length of side from 1 pixel to the standard square of 20 pixels
The result that measures of thickness.B be various method be 30 pixels to regular length, width from 1 pixel to the standard of 20 pixels
The result that rectangular thickness measures.C is various method to be entered the diameter thickness from 1 pixel to the standard circular of 20 pixels
The result that row is measured.Wherein, maximum circle or largest square fill method are the most accurate to the measurement of these three body, and do not have
There is any error.
Ask for an interview Fig. 4, the Comparative result of the 3-D view method for measuring thickness of the embodiment of the present invention.Icon Cube, Sphere,
CubSphere, BoneJ and Scanco represent respectively the maximum cube of utilization, maximum spheroid, maximum spheroid or maximum cube,
The thickness of BoneJ and Scanco software measurement standard three-dimensional image.A be various method to the length of side from 1 pixel to the mark of 20 pixels
The result that the thickness of pseudo-cubic measures.B be various method be 30 pixels to regular length, fixed width is 30 pixels,
The result that height measures from the thickness of the standard cuboid of 1 pixel to 20 pixels.C be various method to diameter from 1 pixel
The result measured to the thickness of standard ball of 20 pixels.D and E be various method be 30 pixels to level altitude, bottom surface
The result that circular diameter measures to the thickness of the standard cylinder of 20 pixels from 1 pixel, wherein D is all of flat to cylinder
The result that face is analyzed, E is for after various methods analysts, choosing the aspect from cylinder distance upper surface bottom surface radius length
Start the interval terminated to the aspect of distance lower surface bottom surface radius length, this interval is carried out the knot that thickness recalculates
Really.Result surface, maximum spheroid or maximum cube fill method are the most accurate to the measurement of these several bodies.
The invention provides a kind of method structure being analyzed and measuring, comprise the steps (below with analysis of control
Between group and process group as a example by long bone trabecular bone microstructure change, but the structure that the present invention is not limited only to analyze long bone trabecular bone becomes
Change):
1) to each sample, from the beginning of specific reference plane, the most full-automatic or semi-automatic or manually choose sense emerging
The analyzed area of interest；
2) differentiation object under test and the threshold value of background noise are set.According to the threshold value set, successively to selected region
The zoning gross area (TA), wherein more than threshold value object area (BA), more than the object girth (PM) of threshold value, more than threshold value
Object pixel value (or gray value or intensity level or density value) (BC), calculate object thickness with maximum circular fill method
And length (Len2D.Cr), utilize maximum circular or largest square fill method to calculate the thickness of object (Th2D.Cr)
And the parameter such as length (Len2D.CrSq) (Th2D.CrSq)；
3) for structure to be analyzed in each sample, calculate the surface area (BS) of this structure to be analyzed, use biggest ball
Body fill method calculates object thickness (Th3D.Sp) and length (Len3D.Sp) and utilizes maximum spheroid or maximum cube to fill out
Fill method and calculate thickness (Th3D.SpCb) and length (Len3D.SpCb), each pixel in structure the most to be analyzed of object
Put equal labelling Th3D.Sp, Th3D.SpCb value and whether be positioned at body structure surface.From the beginning of reference plane, successively calculate each layer
Middle Th3D.Sp, Th3D.SpCb, Len3D.Sp, Len3D.SpCb and BS value；
4) according to each two and three dimensions parameter of layered method, with the distance (or number of plies) apart from reference plane as X-axis,
The parameters value of layered method is Y-axis, and mapping also determines to use linear regression or nonlinear regression march according to its distribution
Line Fitting Analysis；Utilize linearly or nonlinearly regression analysis to each parameter calculated difference between matched group and process group
Carry out significance statistical analysis；
5) from the beginning of reference plane, the accumulative gross area, object area, girth, surface area, gray value, thickness etc. are calculated
Parameter.Its calculation procedure is as follows: set up a two-dimensional table, and this form includes initiateing any planar junction bundle from arbitrary plane
All possible analyzed area.It is characterized by: in first row, each layer record by ground floor to the analyzed area of current layer
Interior calculated parameters value；In a second column, each layer record is counted by the second layer in the analyzed area of current layer
The parameters value calculated；The rest may be inferred, calculates last string and last layer always.Wherein, the calculating of each layer cumulative thickness
Formula is that accumulative object area is divided by accumulative object length.Accumulative parameters result is carried out t check analysis, calculates
Between different grouping, the significance p value of difference, obtains the two-dimensional diagram of p value.The p value two-dimensional diagram obtained is analyzed, root
According to the significant difference threshold value (such as α=0.05) set, calculate the p value region less than this threshold value.This region i.e. p value is in difference
Packet has the region of significant difference.The present invention chooses any point in this region, the row at its place i.e. corresponding to start into
The aspect of row cumulative analysis, and the plane that the plane at its row place i.e. terminates corresponding to cumulative analysis, say, that if this
The bright structure to terminating in plane from the plane proceeding by cumulative analysis to cumulative analysis carries out t check analysis, then this
Bright will be able to detect that analytical parameters significant difference between different grouping.
6) region with significant difference that previous step selects is utilized, to all parameters in matched group and the difference of process group
Different carry out t check analysis.
Below in conjunction with concrete osteogenic little molecule, the impact of rat femur far-end trabecular bone micro structure is expanded on further this
Invention；
1. the process of laboratory animal；
Six-month-old female rats, divides three groups, often group 10, respectively with PBS (matched group), (experiment of the little molecule of osteogenic
Group) or PTH drug administration by injection, after three months, separate femur.Four monthly age female rats, divide four groups, often group 12, and three groups carry out double
Side ovary excision (OVX group), one group carries out sham-operation operation (Sham group), and after performing the operation eight weeks, OVX group is respectively with PBS (comparison
Group), the little molecule of osteogenic (experimental group) or PTH drug administration by injection, Sham group PBS drug administration by injection.After being administered three months, separate
Femur.The microscopic CT scanning instrument of all distal femur Scanco, with the spatial resolution scan of 15 microns, is rebuild.
2. each analytical parameters of distal femur trabecular bone is along the distribution of the femur longitudinal axis；
Automatically identify femur external boundary first with software, then choose external boundary inner distance external boundary 0.6 millimeter with
On interior zone and be set as that trabecular bone region to be analyzed is (due to JIUYUE rat femur in age distal bone cortical thickness about 0.5
Millimeter, the step for purpose be remove rat femur far-end cortical bone region, only retain trabecular bone region to be analyzed).
In selected trabecular bone region, set and distinguish trabecular bone and the threshold value of medullary cavity, the most successively calculate trabecular bone area (BV), choosing
The fixed region gross area (TV), (BMC, by the gray value of trabecular bone according to a series of hydroxyapatite marks for the bone content of trabecular bone
The gray value of quasi-product is calculated), volume bone density (BV/TV), bone density (BMD), maximum circular or largest square fill
The two-dimentional trabecular bone thickness (Thickness-2D) of method measurement and maximum spheroid or the three of maximum cube fill method measurement
Dimension trabecular bone thickness (Thickness-3D).With the number of plies apart from distal femur growth plate as X-axis, each structural parameters of trabecular bone
For Y-axis, map (Fig. 5).Result shows, BV, TV, BMC parameter value from distal femur growth plate plane along femur major axis along with
The distance of distance growth plate increases and is gradually lowered, and approximates exponentially attenuation distribution, and BV/TV, BMD parameter value is along femur major axis
Approximate linearly attenuation distribution.Two dimension trabecular bone thickness and three-dimensional trabecular bone thickness parameter are then approximately one and are parallel to Y-axis
Straight line, prompting the two parameter is the inherent key character of trabecular bone structure, and the distance change with distance growth plate is little.
3. the exponential damping nonlinear regression statistical analysis detection osteogenic little molecule impact on trabecular bone analytical parameters；
Ask for an interview Fig. 6, be that trabecular bone BV, TV, BMC are joined by the little molecule of nonlinear regression analysis osteogenic of the embodiment of the present invention
The impact analysis figure of number.The rat femur sample that PBS (CTRL) or the little molecule of osteogenic (TR) process, below selected growth plate from
Distance growth plate 1.5 millimeters starts the trabecular bone region terminated to distance growth plate 3.72 millimeters, utilizes nonlinear regression index
The change of BV, TV and BMC parameter analyzed by attenuation model.
Rat femur sample is divided into matched group (CTRL) and process group (TR), BV, TV, BMC parameter profit according to processing mode
Statistical analysis (Fig. 6) is carried out with nonlinear regression exponential decay model.Result shows, nonlinear regression exponential decay model is added up
Analyze the sensitiveest, small parameter change can be detected.
4. the osteogenic little molecule impact on trabecular bone structure；
Ask for an interview Fig. 7, be that the automatic trabecular bone subregion of the embodiment of the present invention and manual subregion are to trabecular bone BV, TV, BMC parameter
Impact.1.5 millimeters of 3.72 mm of thickness regions started below rat distal femoral sample growth plate, according to automatically identifying
Or manual identification (μ CT) mode of Micro-CT scanning standard operation (auto), selected bone territory, girder district to be analyzed, then according to micro-
The standard operating procedure of CT workbook analyzes the impact on trabecular bone of the osteogenic little molecule.
Rat distal femoral sample is according to automatically identifying (auto) or manual identification (μ CT) mode of Micro-CT scanning standard operation
Select bone territory, girder district to be analyzed, then according to the standard operating procedure of Micro-CT scanning workbook analyzes the little molecule of osteogenic
Impact on trabecular bone.Result shows, osteogenic molecule appreciable impact BV/TV and BMD, on the impact of BV, TV and BMC without significantly
Sex differernce (Fig. 7).
5. the osteogenic little molecule selection to the analyzed area that trabecular bone parameter has a significant impact；
Ask for an interview Fig. 8, be the analysis area between the trabecular bone micro structure different grouping of the embodiment of the present invention with significant difference
The selection analysis figure in territory.
First rat distal femoral sample starts successively to calculate accumulative BV, TV, BV/TV and BMD parameter value from growth plate,
Then in first row, each layer record by ground floor to the parameters value calculated in the analyzed area of current layer；?
In secondary series, each layer record by the second layer to the parameters value calculated in the analyzed area of current layer；The rest may be inferred,
Calculate last string and last layer always.Setting up a two dimensional surface according to the method, this plane includes from growth
Any one layer below plate starts to terminate the parameters value in all regions to any one layer plane.Utilize t inspection to experimental group and
Matched group carries out statistical analysis, obtains the two-dimentional significance p value plane of parameters.The p value that t is checked by the present invention is less than 0.05
Point be set to 1, its residual value is set to 0, then the present invention has obtained the two-dimentional significance black and white binary image of parameters.
Owing to the measured value of BV, TV, BMC is all exponentially decayed along femur major axis with the increase of distance growth plate plane,
T inspection is then that the aggregate-value in employing analyzed area is analyzed, the most only in specific analyzed area scope, this
The bright significant difference that experimental group and matched group just can be detected.In order to find such region, the present invention establishes one two
Dimensional plane, this plane includes all possible analyzed area initiateing any planar junction bundle from arbitrary plane.To this plane bag
The all regions contained calculate parameters value respectively, and add up these parameter values t inspection to experimental group and matched group
Analyze.If the p value of t inspection is less than 0.05, the pixel value of this point is set to 1, and otherwise, the pixel value of this point is set to 0.This sample is sent out
The bright significance plane bianry image just having obtained a p value, wherein pixel value be 1 Regional Representative's experimental group and matched group have
The region (Fig. 8) of significant difference.The present invention selects point, then this place at the intra-zone with significant difference
Row represent the aspect proceeding by analysis from growth plate direction, and the termination that the row at this place represents trabecular bone analyzed area is put down
Face.According to such selection mode, the present invention ensure that if be analyzed in this region, can be at experimental group and matched group
The difference of significance detected, also imply that such analysis method is more sensitive to trabecular bone microstructure change.Result shows,
BV, BMC, BV/TV and BMD parameter of trabecular bone is all able to detect that experimental group and matched group in sizable selection region
Significant difference, and outside these salient regions, the present invention can not detect the most significant change.
6. the osteogenic little molecule impact on trabecular bone thickness；
Ask for an interview Fig. 9, be the impact on trabecular bone thickness of the little molecule of osteogenic of the embodiment of the present invention.Rat distal femoral sample
This starts bed-by-bed analysis two and three dimensions trabecular bone thickness from growth plate place plane, and presses according to different stimulation process situations
The distance mapping of distance growth plate.Wherein packet is respectively PBS group (CTRL), osteogenic little molecule stimulation group (TREATMENT)
With PTH group (PTH).
The two and three dimensions trabecular bone thickness calculated is mapped by the present invention along the number of plies of distance femur major axis growth plate, result
Showing, two and three dimensions trabecular bone thickness increases with the distance of distance distal femur growth plate and changes the least, points out two peacekeepings
Three-dimensional trabecular bone thickness is a key character of trabecular bone micro structure.The little molecule of osteogenic can stimulate the little of trabecular bone thickness
Amount however, it will be apparent that increase, and strong osteogenic stimulant PTH can cause trabecular bone thickness significantly dramatically increase (figure
9)。
7. selectivity ignores the impact that trabecular bone thickness is analyzed by the part trabecular bone micro structure of a pixel thick；
Ask for an interview Figure 10, be the impact that trabecular bone thickness is analyzed of the noise suppressed of the embodiment of the present invention.Rat distal femoral
1.5 millimeters of 3.72 mm of thickness regions started below sample growth plate, calculate two and three dimensions trabecular bone by standard mode thick
Degree (Thick-2D and Thick-3D).In order to reduce effect of noise, two and three dimensions trabecular bone thickness (Thick-2D.2 and
Thick-3D.2) representing respectively maximum gauge is the structure of 2 and the above pixel result that carries out Thickness Analysis.
Noise mainly affects and has the structure of single pixel thick and impact on having many dot structures object is smaller,
Therefore to reduce the noise impact on result, when the present invention calculates object thickness, only consider the pixel maximum gauge calculated
The system point (Figure 10) at the pixel thick place more than 1.Result shows, through so processing, greatly improves little point of osteogenic
Son promotes the significance level that trabecular bone thickness increases.
8. the analysis result selecting appreciable impact trabecular bone microstructure change of trabecular bone analyzed area；
Ask for an interview Figure 11, be the embodiment of the present invention trabecular bone analyzed area select shadow to trabecular bone microstructure analysis
Ring.The Saliency maps picture utilizing two dimension p value selects the 1.5 mm of thickness regions that below distal femur growth plate 1.5 millimeters starts
(Guided) and standard Micro-CT scanning handbook select distal femur growth plate below 1.5 millimeters start 3.72 millimeters of region (μ
CT) on trabecular bone BV/TV and the impact of BMD parameter.
The present invention, by analysis method above, have selected specific trabecular bone district by the accumulative Parameter analysis of two dimensional surface
Territory is analyzed, and result shows, by the analyzed area selected with upper type than the district recommended by Micro-CT scanning standard operating procedure
Territory sensitiveer to the change of trabecular bone micro structure (Figure 11).
It should be appreciated that the part that this specification does not elaborates belongs to prior art.
It should be appreciated that the above-mentioned description for preferred embodiment is more detailed, can not therefore be considered this
The restriction of invention patent protection scope, those of ordinary skill in the art, under the enlightenment of the present invention, is weighing without departing from the present invention
Profit requires under the ambit protected, it is also possible to make replacement or deformation, within each falling within protection scope of the present invention, this
The bright scope that is claimed should be as the criterion with claims.
Claims (11)
1. the method detecting picture structure change, it is characterised in that comprise the following steps:
Step 1: for structure to be determined, measures two-dimensional structure or the thickness of three dimensional structure；
Step 2: calculate two-dimensional structure or the average thickness of three dimensional structure；
Step 3: successively calculate the parameters of structure to be determined；
Step 4: the parameters of structure to be determined is analyzed, the slight change of detection picture structure；
Step 5: structure to be determined utilizes two dimension Saliency maps picture select the analysis area to different grouping with significant difference
Territory；
Step 6: to selected analyzed area, according to the distribution characteristics of parameters, utilize parametric test or the nonparametric of standard
The method of inspection carries out the significance analysis of difference to different grouping.
The method of detection picture structure the most according to claim 1 change, it is characterised in that: step 1 is utilize maximum
Circular fill method measures the thickness of two-dimensional structure, and described maximum circular fill method fills maximum in being included in structure to be determined
Circle, the thickness corresponding to point in structure is to be completely contained in inside configuration and the maximum circular diameter comprising this point；Appoint
The pixel of anticipating be considered as the length of side be the square of 1 length in pixels, maximum circle can be with the center of any pixel
For the center of circle, it is also possible to be with four angles of any pixel as the center of circle；The radius of this maximum circle can be integer, it is also possible to be
Non-integer.
The method of detection picture structure the most according to claim 1 change, it is characterised in that: step 1 is utilize maximum
Circular or largest square fill method measures the thickness of two-dimensional structure, and it implements and includes following sub-step:
Step is A.1: utilizes maximum circular fill method to measure the thickness of two-dimensional structure, i.e. fills greatest circle in structure to be determined
Shape, the thickness corresponding to point in structure is to be completely contained in inside configuration and the maximum circular diameter comprising this point；Arbitrarily
One pixel be considered as the length of side be the square of 1 length in pixels, maximum circle can be to be with the center of any pixel
The center of circle, it is also possible to be with four angles of any pixel as the center of circle；The radius of this maximum circle can be integer, it is also possible to right and wrong
Integer；
Step is A.2: fill largest square in structure to be determined, utilizes largest square to fill obtain every in computation structure
The maximum gauge of a bit；
Step is A.3: measure the thickness of two-dimensional structure；In structure to be determined, the thickness of every bit is that corresponding point utilizes maximum circle
Fill the maximum gauge obtained and utilize largest square to fill the maximum in two values of the maximum gauge obtained.
The method of detection picture structure the most according to claim 1 change, it is characterised in that: step 1 is utilize maximum
Spheroid fill method measures the thickness of three dimensional structure, and described maximum spheroid fill method fills maximum in being included in structure to be determined
Spheroid, the thickness corresponding to point in structure is the diameter being completely contained in inside configuration and the maximum spheroid that comprises this point；Appoint
The pixel of anticipating be considered as the length of side be the cube of 1 length in pixels, maximum spheroid can be with the center of any pixel
For the centre of sphere, it is also possible to be with eight angles of any pixel as the centre of sphere；The radius of this maximum spheroid can be integer, it is also possible to be
Non-integer.
The method of detection picture structure the most according to claim 1 change, it is characterised in that: step 1 is utilize maximum
Spheroid or maximum cube fill method measure the thickness of three dimensional structure, and it implements and includes following sub-step:
Step is B.1: utilizes maximum spheroid fill method to measure the thickness of three dimensional structure, i.e. fills biggest ball in structure to be determined
Body, the thickness corresponding to point in structure is the diameter being completely contained in inside configuration and the maximum spheroid that comprises this point；Arbitrarily
One pixel be considered as the length of side be the cube of 1 length in pixels, maximum spheroid can be to be with the center of any pixel
The centre of sphere, it is also possible to be with eight angles of any pixel as the centre of sphere；The radius of this maximum spheroid can be integer, it is also possible to right and wrong
Integer；
Step is B.2: then fill maximum cube in structure to be determined, utilizes maximum cube to fill and obtain in computation structure
The maximum gauge of every bit；
Step is B.3: measure the thickness of three dimensional structure；In structure to be determined, the thickness of every bit is that corresponding point utilizes maximum spheroid
Fill the maximum gauge obtained and utilize maximum cube to fill the maximum in two values of the maximum gauge obtained.
The method of detection picture structure the most according to claim 1 change, it is characterised in that: calculate two described in step 2
Dimension structure or the average thickness of three dimensional structure, first calculate the two-dimensional structure corresponding to the part of any thickness or three dimensional structure be long
Degree, then calculates two-dimensional structure or the total length of three dimensional structure, finally calculates two-dimensional structure or the average thickness of three-dimensional.
The method of detection picture structure the most according to claim 1 change, it is characterised in that: calculate two described in step 2
Dimension structure or the average thickness of three dimensional structure, utilize the method reducing noise to calculate；First the part of any thickness is calculated
Corresponding structure length, then ignores the thickness of maximum affected by noise and the length of correspondence thereof, finally calculates without being made an uproar
Sound shadow rings the total length of the structure of maximum point, and then is calculated the average thickness of calibrated structure division.
The method of detection picture structure the most according to claim 1 change, it is characterised in that: parameter bag described in step 3
Include thickness, density, bulk density.
9. according to the method for the detection picture structure change described in claim 1 or 8, it is characterised in that: implementing of step 3
Process is, the structural region that utilization identifies automatically or manual identification methods analyst is interested, then this region is carried out various knot
The calculating of structure parameter.
Detection picture structure the most according to claim 1 change method, it is characterised in that step 4 implement bag
Include following sub-step:
Step 4.1: with the distance apart from reference plane as X-axis, maps with the parameters value of layered method for Y-axis；
Step 4.2: utilize linear regression or nonlinear regression model (NLRM) to carry out the matching of parameters data according to its distribution；
Step 4.3: between utilizing linear regression or nonlinear regression model (NLRM) to different grouping the parameters of matched curve
Difference carries out the significance analysis added up.
11. detection picture structures according to claim 1 change method, it is characterised in that step 5 implement bag
Include following sub-step:
Step 5.1: set up a two-dimensional array, this array includes initiateing the institute of any planar junction bundle likely from arbitrary plane
Analyzed area；
Step 5.2: be analyzed these regions respectively, obtains each structural parameters；
Step 5.3: utilize the statistical method of standard to calculate each structural parameters significance p value of difference between different grouping；
Step 5.4: according to the significance threshold value set, produce a corresponding two dimensional image, according to this two dimensional image, from image
In select the region with significance statistical discrepancy.
Priority Applications (3)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610594607.6A CN106204624A (en) | 2016-07-25 | 2016-07-25 | A kind of method detecting picture structure change |
| PCT/CN2017/094076 WO2018019202A1 (en) | 2016-07-25 | 2017-07-24 | Method and device for detecting change of structure of image |
| CN201710606324.3A CN107464235A (en) | 2016-07-25 | 2017-07-24 | A kind of method and device of detection image structure change |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610594607.6A CN106204624A (en) | 2016-07-25 | 2016-07-25 | A kind of method detecting picture structure change |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN106204624A true CN106204624A (en) | 2016-12-07 |
Family
ID=57495345
Family Applications (2)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201610594607.6A Withdrawn CN106204624A (en) | 2016-07-25 | 2016-07-25 | A kind of method detecting picture structure change |
| CN201710606324.3A Pending CN107464235A (en) | 2016-07-25 | 2017-07-24 | A kind of method and device of detection image structure change |
Family Applications After (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201710606324.3A Pending CN107464235A (en) | 2016-07-25 | 2017-07-24 | A kind of method and device of detection image structure change |
Country Status (2)
| Country | Link |
|---|---|
| CN (2) | CN106204624A (en) |
| WO (1) | WO2018019202A1 (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2018019202A1 (en) * | 2016-07-25 | 2018-02-01 | 武汉大学 | Method and device for detecting change of structure of image |
Families Citing this family (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110942095A (en) * | 2019-11-27 | 2020-03-31 | 中国科学院自动化研究所 | Method and system for detecting salient object area |
| CN116109570B (en) * | 2022-12-22 | 2023-08-08 | 北京医准智能科技有限公司 | Bone mineral density measuring method, device, equipment and storage medium |
Family Cites Families (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US6975894B2 (en) * | 2001-04-12 | 2005-12-13 | Trustees Of The University Of Pennsylvania | Digital topological analysis of trabecular bone MR images and prediction of osteoporosis fractures |
| CN1689020A (en) * | 2002-09-16 | 2005-10-26 | 成像治疗仪股份有限公司 | Imaging markers in musculoskeletal disease |
| US20100068211A1 (en) * | 2008-02-08 | 2010-03-18 | Bateman Ted A | Use of antiresorptive compounds to prevent ionizing radiation-induced activation of osteoclasts and resulting bone loss |
| CN101332127B (en) * | 2008-07-25 | 2011-02-16 | 中国人民解放军第四军医大学 | Construction method of rat hepatocirrhosis and osteoporosis animal model |
| IN2012DN03065A (en) * | 2009-09-11 | 2015-07-31 | Straxcorp Pty Ltd | |
| CN104240241B (en) * | 2014-09-05 | 2017-03-15 | 哈尔滨医科大学 | A kind of method and device of Micro CT graphical analyses quantitative evaluation molar socket bone |
| CN106204624A (en) * | 2016-07-25 | 2016-12-07 | 武汉大学 | A kind of method detecting picture structure change |
-
2016
- 2016-07-25 CN CN201610594607.6A patent/CN106204624A/en not_active Withdrawn
-
2017
- 2017-07-24 WO PCT/CN2017/094076 patent/WO2018019202A1/en active Application Filing
- 2017-07-24 CN CN201710606324.3A patent/CN107464235A/en active Pending
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2018019202A1 (en) * | 2016-07-25 | 2018-02-01 | 武汉大学 | Method and device for detecting change of structure of image |
Also Published As
| Publication number | Publication date |
|---|---|
| CN107464235A (en) | 2017-12-12 |
| WO2018019202A1 (en) | 2018-02-01 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Irie et al. | Use of micro-computed tomography for bone evaluation in dentistry | |
| CN100550004C (en) | A kind of method that the three-dimensional medical image that comprises region of interest is cut apart | |
| US8092387B2 (en) | Ultrasonic imaging technique for differentiating the distribution of scatterers within a tissue | |
| NL1024869C2 (en) | Method and system for measuring tissue changes relevant to disease. | |
| CN100528078C (en) | Method and system for airway measurement | |
| JPH10509075A (en) | Bone densitometer with improved indicating characteristics | |
| US8649843B2 (en) | Automated calcium scoring of the aorta | |
| CN106204624A (en) | A kind of method detecting picture structure change | |
| JPS6282942A (en) | Phantom for inspecting quality of bone and ct scanner and scanning method utilizing the same | |
| CN110363765A (en) | Osteoporosis parameter automatic measurement method based on CT image | |
| CN101091656A (en) | Methods and apparatus for relative perfusion and/or viability | |
| CN113034389B (en) | Image processing method, device, computer equipment and storage medium | |
| CN106880900A (en) | Method for automatically determining contrast agent infusion protocol | |
| CN110503652A (en) | Mandibular kinesiography and adjacent teeth and mandibular canal relationship determine method, apparatus, storage medium and terminal | |
| KR102342575B1 (en) | Optical fiber bundle image processing method and apparatus | |
| US20110148861A1 (en) | Pet data processing system, an arrangement, a method and a computer program product for determining a distribution of a tracer uptake | |
| CN108324244A (en) | The construction method and system of automatic augmentation training sample for the diagnosis of AI+MRI Image-aideds | |
| US20150086094A1 (en) | Echogenicity quantification method and calibration method for ultrasonic device using echogenicity index | |
| US20090127451A1 (en) | Devices and Methods for Calibrating Nuclear Medical and Radiological Images | |
| CN113515875B (en) | Bone biomechanics modeling method, system and device based on multi-modal image | |
| CN115115813B (en) | Intelligent construction method for human skeleton standard body position | |
| JP4652935B2 (en) | X-ray CT apparatus and image processing method | |
| CN110151211A (en) | Exact method is estimated in the scattering for improving scan vision boudary portion in imaging system | |
| US20020181755A1 (en) | Method for measuring bone mineral density by using X-ray image | |
| US11561273B2 (en) | Medical image processing apparatus and method |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| WW01 | Invention patent application withdrawn after publication | ||
| WW01 | Invention patent application withdrawn after publication |
Application publication date: 20161207 |