CN104182965B - The method of chest muscle is split in a kind of galactophore image - Google Patents
The method of chest muscle is split in a kind of galactophore image Download PDFInfo
- Publication number
- CN104182965B CN104182965B CN201410028053.4A CN201410028053A CN104182965B CN 104182965 B CN104182965 B CN 104182965B CN 201410028053 A CN201410028053 A CN 201410028053A CN 104182965 B CN104182965 B CN 104182965B
- Authority
- CN
- China
- Prior art keywords
- chest muscle
- image
- galactophore image
- region
- pectoral region
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Abstract
The invention provides a kind of method splitting chest muscle in galactophore image, comprise the steps: that the region of galactophore image being carried out to morphology restriction increases, obtain the first pectoral region; First correction is carried out to described first pectoral region, obtains the galactophore image of segmentation pectoral region.Chest muscle dividing method provided by the invention has zero offset capability, can fast and effeciently split in the position breast X-ray figure of inside and outside oblique side and remove chest muscle part, there is higher degree of accuracy and robustness. for multi-stage grey scale level, edge fog, the irregular chest muscle in edge can both be split effectively.
Description
Technical field
The present invention relates to field of medical image processing, particularly relate to a kind of method splitting chest muscle in galactophore image.
Background technology
Human body pectoral region in inside and outside skew back (MLO) position of removing is the basic steps in the pre-service of mammography X (Mammograms) computer-aided diagnosis (Computer-aidedDiagnosis, CAD) system.In the galactophore image of MLO position, pectoral region is usually located at the upper left corner or the upper right corner of image, and general edge is comparatively mild, presents triangle, has brightness high or the highest in entire image.Remove the chest muscle step of MLO bit image, can avoid bringing harmful effect to subsequent treatment, such as, image regularization in lesion segmentation, breast density estimation etc., and can effectively suppress false positive for mammary gland CAD, thus improve focus verification and measurement ratio.
In prior art, the method splitting chest muscle in the galactophore image of MLO position mainly contains following several: (1) is based on the method for Threshold segmentation.First former figure is cut out to the area-of-interest comprising chest muscle part, then carry out Threshold segmentation, finally threshold process is carried out to image.Because grey scale pixel value difference in dynamic range of pectoral region in different galactophore images is very large, therefore cannot estimate threshold value, can only get empirical value, thus cause the method applicability wideless, segmentation result degree of accuracy is low; (2) based on the method that region increases.First, index mapping carried out to former figure and quantize to reduce gray level; Then Morphological scale-space is carried out; Be positioned at the prerequisite in the upper left corner at default chest muscle under, Seed Points is located at the pixel in the upper left corner; Then carry out region growing by the method for iteration, in the process of growth, add yardstick restriction (SizeRestriction), once outgrowth, then improve threshold value and regrow.The method can Fast Convergent, and can effectively control undue, but degree of accuracy is low, easily produces deviation; (3) based on the dividing method of Hough transform.The method is a kind of widely used method splitting chest muscle in prior art, is also the patented technology of each large manufacturer.These class methods can only be suitable for the situation of chest muscle real border near linear, the chest muscle border larger for curvature and the chest muscle having cortex to superpose inapplicable; (4) Hough transform and Snake mixed method.First, obtain the straight line close to real border by aforesaid Hough transform method, then with Snake alternative manner, straight line is corrected.The method is easily subject to the interference of noise and surrounding structure in Snake step and causes accuracy to reduce; (4) straight line is estimated and accurate adjustment (Refinement) method.First Threshold segmentation is carried out to former figure, obtain bianry image; The border of bianry image is utilized to estimate a straight line close to chest muscle real border; For each pixel on the straight line obtained, get the gradation of image on its traverse, obtain an one dimension waveform, this waveform of matching is removed with Sigmoid function, and find the most precipitous point (CliffDetection) with parametric technique, image coordinate corresponding to this point is the renewal to former frontier point.The shortcoming of the method is similar to the method for Hough transform, once real border is more bending, straight line differs comparatively far away with it, Refinement can not go back edge correction.Because Refinement is only applicable to accurate adjustment among a small circle, border will be caused not restrain once expand fine adjustment range.In addition, Sigmoid Function Fitting is inapplicable with waveform complicated situation.
In sum, in prior art, mainly there is following problem for the method splitting chest muscle in position (MLO is) galactophore image of inside and outside oblique side: applicability is low, for multi-grey level, edge irregularity or fuzzy chest muscle can not be split effectively; Threshold value can not adjust in real time, can only adopt empirical value, and the selection of parameter is comparatively large by noise effect, causes segmentation result degree of accuracy to reduce.
Summary of the invention
The problem that the present invention solves is to provide a kind of method splitting chest muscle in galactophore image, in order to solve in the galactophore image of MLO position split chest muscle method can not effectively segmenting edge scrambling or fuzzy chest muscle, threshold value can not adjust in real time, empirical value can only be adopted, the selection of parameter is larger by noise effect, segmentation result is accurately low, poor robustness.
In order to solve the problem, the invention provides a kind of method splitting chest muscle in galactophore image, comprising the steps:
The region of galactophore image being carried out to morphology restriction increases, and obtains the first pectoral region;
First correction is carried out to described first pectoral region, obtains the galactophore image of segmentation pectoral region.
Further, the region growth of described morphology restriction comprises the steps:
(1) morphology restriction is defined;
(2) adaptive region carrying out morphology restriction increases, and the bottom threshold that described region increases is:
th
L=(1-t)*(T
max+T
min)／2
Wherein T
maxand T
minbe respectively maximal value and the minimum value of current growth district;
(3) judge: if the area S ' of growth district and the difference △ S≤a*S of the area S of morphology restricted area, then obtain described first pectoral region;
Otherwise, adjustment parametric t=t*b, and repeating said steps (2) and (3);
Wherein a, b, t are the constant between (0,1).
Further, the method for described definition morphology restriction comprises the steps: on described two-dimentional galactophore image
Defining point p (1, y): get the separation of chest muscle and breast in galactophore image the first row for a p;
Defining point q (x, 1): the line number m getting the some place of the col width maximal value of breast in galactophore image, and it is capable to move down c, order point x=m+c that q is expert at;
Get the line pq of described some p and some q, Iamge Segmentation is Part I and Part II by described line pq, and described Part I is the region of described morphology restriction.
Further, described first correction comprises the steps:
(1) lofty structure operation is removed to described first pectoral region, obtain the second pectoral region;
(2) on the galactophore image of n kind yardstick, accurate adjustment is carried out to described second pectoral region, obtain the galactophore image of segmentation pectoral region;
Wherein, described n be more than or equal to 1 positive integer.
Further, the lofty structure operation of described removal comprises the steps:
Travel through points all on described first edge, pectoral region, get any two terminal A and B, connect A, B two-end-point and form straight line AB, if straight line AB meets simultaneously:
Straight line AB distance
It is inner that straight line AB is positioned at described first pectoral region;
Straight line AB distance d
aB<d*C;
Then remove lofty structure;
If not, then do not deal with.
Wherein, described L is the overall length of chest muscle and mammary gland adjacent boundary in described first pectoral region; Described C is the chest muscle edge overall length connecting AB 2 in described first pectoral region; Described d be greater than 1 constant.
Further, described accurate adjustment comprises the steps:
Get the frontier point of chest muscle in described second pectoral region, get any one frontier point, make the normal of described frontier point, get at least 2 pixels that frontier point is adjacent on normal;
The gray-scale value of at least 3 pixels described in calculating, the difference defining the gray-scale value of adjacent 2 is gradient;
Determine gradient maximum point, define the renewal coordinate of image coordinate for got frontier point of this some correspondence;
Adopt the limit of Snake algorithm second pectoral region accurate adjustment chest muscle in the galactophore image of n kind yardstick, obtain the galactophore image of described segmentation pectoral region;
Wherein, the state equation of described Snake algorithm is:
State equation E
snake=E
int+ E
ext;
Described renewal coordinate is E in described Snake state equation
extfor external energy; The flatness on described chest muscle border is the internal energy E in described Snake state equation
int.
Further, described accurate adjustment carries out iteration accurate adjustment from low to high successively according to the resolution of described galactophore image.
Further, before described accurate adjustment, edge-smoothing operation is carried out to described second pectoral region.
Further, before obtaining described first pectoral region, the direction of mammary gland in first described judgement image;
Further, the method in described judgement mammary gland direction comprises the steps:
Get any a line 1 containing mammary gland part in described galactophore image and middle column m, it is Part III and Part IV that described 1st row and m arrange the first half symmetry division of described first image;
Calculate the average gray T of described third and fourth two parts image
rightand T
left;
If T
right>T
left, then mammary gland is positioned at the right side of described first image;
If T
right<T
left, then mammary gland is positioned at the left side of described first image.
Compared with prior art, the present invention has the following advantages: (1) the invention provides a kind of method splitting chest muscle in galactophore image, regulate threshold value in real time by morphologic limiting adaptive ground in the process that can increase in region, thus effectively suppress to overflow on a large scale; (2) the present invention effectively can remove spilling among a small circle by the method (method of RollingBall) removing lofty structure, greatly improves the robustness of algorithm; (3) the present invention can carry out accurate adjustment to chest muscle edge on multiple dimensioned galactophore image, effectively ensures the flatness at chest muscle edge in accurate adjustment process, makes segmentation result more accurate; (4) the method applicability is comparatively wide, accuracy is high, effectively can split multi-grey level, edge fog, the irregular chest muscle in edge.
Accompanying drawing explanation
Figure 1 shows that the method flow schematic diagram splitting chest muscle in one embodiment of the invention in galactophore image;
Figure 2 shows that in one embodiment of the invention the method flow schematic diagram determining mammary gland direction;
Figure 3 shows that the schematic diagram of Part III and Part IV in one embodiment of the invention;
Figure 4 shows that the area schematic of morphology restriction in one embodiment of the invention;
Figure 5 shows that the method flow schematic diagram that the region of morphology restriction in one embodiment of the invention increases;
Figure 6 shows that in one embodiment of the invention the method flow schematic diagram defining morphology restriction;
Figure 7 shows that the schematic diagram of Part I in one embodiment of the invention and second;
Figure 8 shows that the first method flow schematic diagram corrected in one embodiment of the invention;
Figure 9 shows that in one embodiment of the invention the schematic flow sheet removing lofty structure operation;
Figure 10 shows that the schematic diagram of lofty structure in one embodiment of the invention;
Figure 11 shows that the close-up schematic view of R-portion in Figure 10;
Figure 12 shows that the schematic flow sheet of accurate adjustment in one embodiment of the invention;
Figure 13 a ~ 13e is each phase results schematic diagram splitting chest muscle in one embodiment of the invention.
Embodiment
Set forth a lot of detail in the following description so that fully understand the present invention.But the present invention can be much different from alternate manner described here to implement, those skilled in the art can when without prejudice to doing similar popularization when intension of the present invention, therefore the present invention is by the restriction of following public concrete enforcement.
Secondly, the present invention utilizes schematic diagram to be described in detail, and when describing the embodiment of the present invention in detail, for ease of illustrating, described schematic diagram is example, and it should not limit the scope of protection of the invention at this.
In order to solve the technical matters in background technology, the invention provides a kind of method splitting chest muscle in galactophore image, first the region growing methods of morphology restriction is adopted, in the morphology restricted area of definition, automatically adjust threshold parameter, regulate threshold value in real time, the effective suppression overflows on a large scale, avoid the shortcoming adopting empirical value method in prior art, improve applicability of the present invention and accuracy; Then adopt the removal method (RollingBall) of lofty structure and the method on Snake algorithm optimization chest muscle border to carry out the first correction to region growth structure, effectively split multi-grey level, edge fog, the irregular chest muscle in edge.
Describe the present invention below in conjunction with accompanying drawing.Shown in the method flow schematic diagram splitting chest muscle in galactophore image as Fig. 1 one embodiment of the invention, described method comprises the steps:
First, step S10 is performed: determine mammary gland direction in described galactophore image, be convenient to follow-up unified operation; Describedly determine that the method in mammary gland direction is as shown in the method flow schematic diagram determining mammary gland direction in Fig. 2 one embodiment of the invention, described method comprises the steps:
First, perform step S11: get containing any a line 1 of mammary gland part and the middle column m of described galactophore image in described galactophore image, described 1st row and m arrange the Part III 3 and Part IV 4 that to be divided into by described galactophore image as shown in Figure 3;
Then, step S12 ~ S13: the average gray T calculating image in described Part III 3 and Part IV 4 is respectively performed
rightand T
left, compare the size of two values:
If T
right>T
left, then mammary gland is positioned at the right side of described first image;
If T
right< T
left, then mammary gland is positioned at the left side of described first image.
Preferably, in the present embodiment, adopt mammary gland direction to be positioned on the left of image, when mammary gland is positioned at the right side of described galactophore image, made by Image Reversal mammary gland be positioned at the left side of described galactophore image.
Then, step S20 is performed: carry out region growth in the morphology restricted area shown in the area schematic of galactophore image morphology restriction in such as Fig. 4 one embodiment of the invention.
The method of the region growth of described morphology restriction is as shown in the method flow schematic diagram of the region growth of morphology restriction in Fig. 5 one embodiment of the invention, and described method comprises the steps:
First, step S21 is performed: the restriction of definition morphology; Be different from prior art, in the present embodiment, it is carry out in the region of morphology restriction that region increases, can effective control area large-scale overflow problem in increasing, and improves the accuracy that chest muscle is split.The method of described definition morphology restriction is as shown in the method flow schematic diagram defining morphology restriction in Fig. 6 one embodiment of the invention:
Perform step S211: defining point p and some q on described galactophore image;
Defining point p (1, y): get the separation of chest muscle and breast in galactophore image the first row for a p; In the galactophore image of usual MLO position, chest muscle is triangular in shape, in entire image, there is high or the highest brightness, the boundary of general the first row chest muscle part and background or breast portion is comparatively clear, and due to the existence of body of gland, focus or other tissue below image, the boundary of chest muscle and other parts thickens, therefore be chosen at the separation of described galactophore image the first row chest muscle and breast in the present embodiment for a p, be convenient to anchor point p quickly and accurately.
Defining point q (x, 1): the line number m getting the some place of the col width maximal value of breast in galactophore image, and it is capable to move down c, order point x=m+c that q is expert at; Because breast shape is different, the separation of chest muscle lower end not necessarily at the col width maximal value place of breast, so, the line number m at the some place of col width maximal value is moved down c capable, is expert at as a q, made to comprise pectoral region as far as possible, improve segmentation chest muscle accuracy.
After determining complete some p and putting q, perform step S212: the line pq getting described some p and some q, described line pq by Iamge Segmentation for shown in the schematic diagram of Part I 1 in such as Fig. 7 one embodiment of the invention and Part II 2, wherein said Part I 1 is the mammary gland parts of images near side, thoracic cavity, chooses the region that described Part I 1 is the restriction of described morphology.
Continue to perform step S22: the adaptive region described galactophore image being carried out to morphology restriction increases, and the bottom threshold that described region increases is: th
l=(1-t) * (T
max+ T
min)/2, wherein, T
maxand T
minbe respectively maximal value and the minimum value of current growth district.
Then step S23 is performed: the area S ' of growth district and the size of the difference △ S of the area S of described the Part I 1 and area S of described Part I more, if △ is S≤a*S, obtains described first pectoral region;
Otherwise, adjustment parametric t=t*b, and repeating said steps S22 and S33; Wherein a, b, t are the constant between (0,1).
It should be noted that, in the present embodiment, the bottom threshold th that described region increases
lbe regulated in real time by threshold parameter t, effectively prevent the spilling in threshold value propagation process, be applicable to the chest muscle segmentation of the galactophore image of different gray scale.In the present embodiment, the size of preferred a is 0.1, namely when difference in areas △ S exceedes the region area 1/10 of described morphology restriction, adjustment parametric t=t*b, in the present embodiment, the size of t is preferably 0.4, and the size of described b is preferably 0.9, brings the parameter after adjustment into th
l=(1-t) * (T
max+ T
minre-start region in)/2 to increase, until meet morphology restriction.
The region adopting morphology to limit in the present embodiment increases, and in the process that can increase in region, utilizes morphology to limit and automatically adjusts adaptive threshold parametric t, thus regulate threshold value th in real time
l, effectively suppress to overflow on a large scale in propagation process, improve the accuracy of segmentation chest muscle.
After described first pectoral region is determined, continue to perform step S30: carry out the first correction to described first pectoral region, for optimizing, the adaptive region of accurate adjustment above-mentioned morphology restriction increases the edge, pectoral region obtained, and improves the accuracy of chest muscle segmentation.Described first method corrected is as shown in the first method flow schematic diagram corrected in Fig. 8 one embodiment of the invention:
First step S31 is performed: lofty structure operation is removed to described first pectoral region, obtains the second pectoral region; Described lofty structure refers in the region propagation process that morphology limits, at the irregular structure at described first edge, pectoral region, may be some mammary gland, or lesion tissue, or some spillings among a small circle, causing the over-segmentation of chest muscle, in order to improve the accuracy of chest muscle dividing method in the present embodiment galactophore image, needing to remove lofty structure.
Particularly, the lofty structure operation of described removal, as shown in the schematic flow sheet removing lofty structure operation in Fig. 9 one embodiment of the invention, comprises the steps:
Perform step S311: first travel through points all on described first edge, pectoral region, get any two terminal A and B, connect A, B two-end-point and form straight line AB, particularly choose two end points of any one lofty structure on described first edge, pectoral region, as shown in Figure 10 and Figure 11; Then step S312 ~ S313 is performed: if straight line AB meets following three conditions simultaneously: a) straight line AB distance
b) straight line AB is positioned at described first inside, pectoral region; C) straight line AB distance d
aB<d*C; Then remove lofty structure, be about to the border of straight line AB as described chest muscle of connection A, B two-end-point; If not, then do not deal with, retain this lofty structure.
Wherein, described L is the overall length of chest muscle and mammary gland adjacent boundary in described first pectoral region, and described condition is a) for controlling the aperture width of lofty structure; For the lofty structure meeting aperture width, conjugation condition c) judged result determine whether to remove lofty structure; Described condition b) operation of removing lofty structure is defined in described first inside, pectoral region; Described C is the chest muscle edge overall length connecting AB 2 in described first pectoral region, described condition c) for controlling the lofty degree of lofty structure, remove lofty structure in this case by conditional operation; Described d be greater than 1 constant, preferred d=2 in the present embodiment.
In the present embodiment, RollingBall is adopted to remove lofty structure, the segmentation of chest muscle effectively can be optimized to the condition of the three groups of AND operations arranged in RollingBall method, the part of spilling among a small circle particularly in above-mentioned zone propagation process, such as elongated body of gland part, by the intersection operation of condition setting, such as, control the lofty degree of lofty structure to distinguish chest muscle and body of gland part, thus obtain effective segmentation of chest muscle part.
Finally, perform step S40: on the galactophore image of n kind yardstick, accurate adjustment is carried out to described second pectoral region, obtain the galactophore image of segmentation pectoral region;
Particularly, described accurate adjustment operation, as shown in the schematic flow sheet of accurate adjustment operation in Figure 12 one embodiment of the invention, comprises the steps:
First, perform step S41, edge-smoothing operation is carried out to described second pectoral region obtained; Then step S42 ~ S43 is performed: points all on the border of described second pectoral region after traversal is level and smooth, get any one frontier point, make the normal of described frontier point; get at least 2 pixels that frontier point is adjacent in the normal direction with described frontier point; the gray-scale value of at least 3 pixels described in calculating, the difference defining the gray-scale value of adjacent 2 is gradient; Determine gradient maximum point, described gradient maximum point is relevant with the mode that gradient defines, if forward gradient before being exactly a bit, backward gradient after being exactly a bit, define the renewal coordinate that image coordinate corresponding to described gradient maximum point is described frontier point.Then, perform step S44: bring in the state equation of Snake algorithm by the renewal coordinate that above-mentioned steps obtains, adopt Snake algorithm accurate adjustment chest muscle border in the second pectoral region of the galactophore image of n kind yardstick, the galactophore image of the described segmentation pectoral region of final acquisition, wherein, described n be more than or equal to 1 positive integer.
Wherein, the state equation of described Snake algorithm is: E
snake=E
int+ E
ext, described renewal coordinate is E in described Snake state equation
extexternal energy; The flatness on described chest muscle border is the internal energy E in described Snake state equation
int, for maintaining the flatness at chest muscle edge, restriction external energy E
extexcessive variation, namely when the renewal coordinate chosen is abnormal, such as, there is the tissue of the high brightness such as body of gland in galactophore image because of frontier point, the greatest gradient value point calculated is not the boundary of chest muscle and breast, causes the renewal coordinate exception obtained.
Wherein, described accurate adjustment process carries out accurate adjustment from low to high successively according to the resolution of described galactophore image, and such as former figure size is the galactophore image of 100 × 100, by its down-sampled to resolution be 50 × 50, further down-sampled to resolution be 25 × 25.First in the low resolution of 25 × 25, adjust in galactophore image that in the second pectoral region accurate adjustment chest muscle border, then in the resolution of 50 × 50, adjust chest muscle border, finally on former Figure 100 × 100, adjust chest muscle border.By carrying out the accurate adjustment chest muscle border operation of Snake algorithm to resolution galactophore image from low to high, reduction noise and surrounding structure, to the interference of Snake algorithm, improve the accuracy optimizing border result.If Figure 13 a ~ 13e is each phase results schematic diagram splitting chest muscle in one embodiment of the invention.
By carrying out accurate adjustment to the edge of described second pectoral region on the galactophore image of different scale in the present embodiment, thus multi-grey level, edge fog or irregular chest muscle effectively being optimized, being improved accuracy and the robustness of chest muscle segmentation.
Although the present invention with preferred embodiment openly as above; but it is not for limiting the present invention; any those skilled in the art without departing from the spirit and scope of the present invention; the Method and Technology content of above-mentioned announcement can be utilized to make possible variation and amendment to technical solution of the present invention; therefore; every content not departing from technical solution of the present invention; the any simple modification done above embodiment according to technical spirit of the present invention, equivalent variations and modification, all belong to the protection domain of technical solution of the present invention.
Claims (8)
1. split a method for chest muscle in galactophore image, it is characterized in that, comprise the steps:
The region of galactophore image being carried out to morphology restriction increases, and obtains the first pectoral region;
First correction is carried out to described first pectoral region, obtains the galactophore image of segmentation pectoral region;
The region growth of described morphology restriction comprises the steps:
(1) morphology restriction is defined;
(2) adaptive region carrying out morphology restriction increases, and the bottom threshold that described region increases is:
th
L＝(1-t)*(T
max+T
min)/2
Wherein T
maxand T
minbe respectively maximal value and the minimum value of current growth district;
(3) judge: if the difference Δ S≤a*S of the area S ＇ of growth district and the area S of morphology restricted area, then obtain described first pectoral region;
Otherwise, adjustment parametric t=t*b, and repeating said steps (2) and (3);
Wherein a, b, t are the constant between (0,1);
Described first correction comprises the steps:
(1) lofty structure operation is removed to described first pectoral region, obtain the second pectoral region;
(2) on the galactophore image of n kind yardstick, accurate adjustment is carried out to described second pectoral region, obtain the galactophore image of segmentation pectoral region;
Wherein, described n be more than or equal to 1 positive integer.
2. split the method for chest muscle in galactophore image as claimed in claim 1, it is characterized in that, described definition morphology restriction comprises the steps: on described galactophore image
(1) defining point p (1, y): the separation getting chest muscle and breast in galactophore image the first row is a some p;
(2) defining point q (x, 1): the line number m getting the some place of the col width maximal value of breast in galactophore image, and it is capable to move down c, order point x=m+c that q is expert at;
(3) get the line pq of described some p and some q, Iamge Segmentation is Part I and Part II by described line pq, and described Part I is the region of described morphology restriction.
3. split the method for chest muscle in galactophore image as claimed in claim 1, it is characterized in that, the lofty structure operation of described removal comprises the steps:
Travel through points all on described first edge, pectoral region, get any two terminal A and B, connect A, B two-end-point and form straight line AB, if straight line AB meets simultaneously:
A) straight line AB distance
B) straight line AB is positioned at described first inside, pectoral region;
C) straight line AB distance d
aB< d*C;
Then remove lofty structure;
If not, then do not deal with;
Wherein, described L is the overall length of chest muscle and mammary gland adjacent boundary in described first pectoral region; Described C is the chest muscle edge overall length connecting A, B 2 in described first pectoral region; Described d be greater than 1 constant.
4. split the method for chest muscle in galactophore image as claimed in claim 1, it is characterized in that, described accurate adjustment comprises the steps:
(1) get the frontier point of chest muscle in described second pectoral region, get any one frontier point, make the normal of described frontier point, get at least 2 pixels that frontier point is adjacent in the normal direction with frontier point;
(2) gray-scale value of at least 3 pixels described in calculating, the difference defining the gray-scale value of adjacent 2 is gradient;
(3) determine gradient maximum point, the image coordinate defining this some correspondence is the renewal coordinate of described frontier point;
(4) in the second pectoral region the border of accurate adjustment chest muscle in the galactophore image of n kind yardstick to adopt Snake algorithm, obtains the galactophore image of described segmentation pectoral region;
Wherein, the state equation of described Snake algorithm is:
State equation E
snake=E
int+ E
ext;
Described renewal coordinate is the external energy E in described Snake state equation
ext; The flatness on described chest muscle border is the internal energy E in described Snake state equation
int.
5. split the method for chest muscle in the galactophore image as described in claim 1 or 4 any one, it is characterized in that, described accurate adjustment carries out iteration accurate adjustment from low to high successively according to the resolution of described galactophore image.
6. split the method for chest muscle in galactophore image as claimed in claim 1, it is characterized in that, before described accurate adjustment, edge-smoothing operation is carried out to described second pectoral region.
7. split the method for chest muscle in galactophore image as claimed in claim 1, it is characterized in that, before obtaining described first pectoral region, judge the direction of mammary gland in described galactophore image, obtain the first image.
8. split the method for chest muscle in galactophore image as claimed in claim 7, it is characterized in that, the method in described judgement mammary gland direction comprises the steps:
Get any a line l containing mammary gland part in described galactophore image and middle column m, described l are capable and m to arrange the first half symmetry division of described first image be Part III and Part IV;
Calculate the average gray T of described third and fourth two parts image
rightand T
left;
If T
right>T
left, then mammary gland is positioned at the right side of described first image;
If T
right<T
left, then mammary gland is positioned at the left side of described first image.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201410028053.4A CN104182965B (en) | 2014-01-22 | 2014-01-22 | The method of chest muscle is split in a kind of galactophore image |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201410028053.4A CN104182965B (en) | 2014-01-22 | 2014-01-22 | The method of chest muscle is split in a kind of galactophore image |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN104182965A CN104182965A (en) | 2014-12-03 |
| CN104182965B true CN104182965B (en) | 2016-03-30 |
Family
ID=51963983
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201410028053.4A Active CN104182965B (en) | 2014-01-22 | 2014-01-22 | The method of chest muscle is split in a kind of galactophore image |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN104182965B (en) |
Families Citing this family (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105447879B (en) * | 2015-12-15 | 2018-04-03 | 上海联影医疗科技有限公司 | The method and device of chest muscle is detected in breast image |
| CN108471995B (en) | 2015-09-30 | 2022-03-29 | 上海联影医疗科技股份有限公司 | System and method for determining breast regions in medical images |
| US9959617B2 (en) | 2016-01-28 | 2018-05-01 | Taihao Medical Inc. | Medical image processing apparatus and breast image processing method thereof |
| TWI587844B (en) * | 2016-01-28 | 2017-06-21 | 太豪生醫股份有限公司 | Medical image processing apparatus and breast image processing method thereof |
| CN113393475B (en) * | 2021-06-30 | 2024-02-20 | 神州医疗科技股份有限公司 | Mammary gland molybdenum target image segmentation device, electronic equipment, medical equipment and medium |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101334895B (en) * | 2008-08-07 | 2011-09-14 | 清华大学 | Image division method aiming at dynamically intensified mammary gland magnetic resonance image sequence |
| CN101612047B (en) * | 2004-08-03 | 2013-03-06 | 株式会社东芝 | Image displaying apparatus |
| CN103473767A (en) * | 2013-09-05 | 2013-12-25 | 中国科学院深圳先进技术研究院 | Segmentation method and system for abdomen soft tissue nuclear magnetism image |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US9111174B2 (en) * | 2012-02-24 | 2015-08-18 | Riverain Technologies, LLC | Machine learnng techniques for pectoral muscle equalization and segmentation in digital mammograms |
-
2014
- 2014-01-22 CN CN201410028053.4A patent/CN104182965B/en active Active
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101612047B (en) * | 2004-08-03 | 2013-03-06 | 株式会社东芝 | Image displaying apparatus |
| CN101334895B (en) * | 2008-08-07 | 2011-09-14 | 清华大学 | Image division method aiming at dynamically intensified mammary gland magnetic resonance image sequence |
| CN103473767A (en) * | 2013-09-05 | 2013-12-25 | 中国科学院深圳先进技术研究院 | Segmentation method and system for abdomen soft tissue nuclear magnetism image |
Non-Patent Citations (1)
| Title |
|---|
| 基于模型的乳腺X线图像胸肌分割算法研究;徐伟栋等;《浙江大学学报（工学版）》;20050331;第39卷(第3期);第427页-第432页 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN104182965A (en) | 2014-12-03 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN104182965B (en) | The method of chest muscle is split in a kind of galactophore image | |
| CN101576997B (en) | Abdominal organ segmentation method based on secondary three-dimensional region growth | |
| Ji et al. | Cell image segmentation based on an improved watershed algorithm | |
| CN103164855B (en) | A kind of Bayesian decision foreground extracting method in conjunction with reflected light photograph | |
| CN101667137B (en) | Method for extracting finger vein grain by using directional filtering technique | |
| CN110689545A (en) | Fuzzy boundary image automatic segmentation method based on active contour and deep learning | |
| CN103903254B (en) | A kind of x-ray image processing method, system and x-ray image processing equipment | |
| CN104376551A (en) | Color image segmentation method integrating region growth and edge detection | |
| CN105389811A (en) | Multi-modality medical image processing method based on multilevel threshold segmentation | |
| CN109753997B (en) | Automatic accurate robust segmentation method for liver tumor in CT image | |
| CN104346799B (en) | The extracting method of spinal cord in a kind of CT images | |
| CN109671092A (en) | A kind of improved Canny image partition method and system | |
| CN113296095B (en) | Target hyperbola edge extraction method for pulse ground penetrating radar | |
| CN104574327A (en) | Mammary gland image processing method and device | |
| Li et al. | Watershed algorithm based on morphology for dental X-ray images segmentation | |
| CN115147710A (en) | Sonar image target processing method based on heterogeneous filtering detection and level set segmentation | |
| CN109410147A (en) | A kind of supercavity image enchancing method | |
| CN116468641A (en) | Infrared blood vessel image enhancement processing method | |
| CN109636822B (en) | Improved Canny self-adaptive edge extraction method based on newly-constructed membership function | |
| CN109886973B (en) | Blood vessel extraction method and device and computer readable storage medium | |
| CN106447659B (en) | A kind of region growing detection method based on multiple decision | |
| CN104794689A (en) | Preprocessing method for enhancing sonar image contract | |
| Behrad et al. | Automatic spleen segmentation in MRI images using a combined neural network and recursive watershed transform | |
| CN110930330B (en) | Image segmentation and region growth based salt and pepper noise reduction algorithm | |
| CN109741265A (en) | A kind of spine CT image bilateral filtering denoising method based on feature selecting |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| C14 | Grant of patent or utility model | ||
| GR01 | Patent grant | ||
| CP01 | Change in the name or title of a patent holder |
Address after: 201815 No. 1180 Xingxian Road, Jiading Industrial Zone, Jiading District, Shanghai Patentee after: Shanghai Lianying Medical Technology Co., Ltd Address before: 201815 No. 1180 Xingxian Road, Jiading Industrial Zone, Jiading District, Shanghai Patentee before: SHANGHAI UNITED IMAGING HEALTHCARE Co.,Ltd. |
|
| CP01 | Change in the name or title of a patent holder | ||
| CP02 | Change in the address of a patent holder |
Address after: 201807 2258 Chengbei Road, Jiading District, Shanghai Patentee after: Shanghai Lianying Medical Technology Co.,Ltd. Address before: 201815 No. 1180 Xingxian Road, Jiading Industrial Zone, Jiading District, Shanghai Patentee before: Shanghai Lianying Medical Technology Co.,Ltd. |
|
| CP02 | Change in the address of a patent holder |