CN103729845A - Breast X-ray image registration method and system based on barycenter - Google Patents

Abstract

The invention discloses a breast X-ray image registration method based on barycenter. The method includes the steps that left-breast X-ray image DICOM data to be registered and right-breast X-ray image DICOM data to be registered are loaded; window width and window level transformation is performed on original left-breast X-ray image DICOM data and original right-breast X-ray image DICOM data respectively to obtain a transformed left-breast image and a transformed right-breast image; a barycentric coordinate R1 of the transformed left-breast image and a barycentric coordinate Rr of the transformed right-breast image are calculated; the transformed left-breast image or the transformed right-breast image is selected as a moving image Imgs, and the other image is used as a template image Imgd; the horizontal moving amount is calculated, the moving image Imgs is made to move horizontally so as to enable the left-breast image and the right-breast image to be aligned; a registered breast image is displayed. According to the method and system, registering alignment of a left breast and a right breast is achieved by moving the breasts horizontally according to the distance difference of barycenter positions of the left-breast image and the right-breast image, and the registered and aligned breast image is obtained, so that doctors can perform contrastive analysis conveniently.

Description

Method for registering and the registration arrangement of the Mammography image based on barycenter

Technical field

The invention belongs to digital image processing techniques field, be specifically related to a kind of method for registering and registration arrangement of the Mammography image based on barycenter.

Background technology

Mammary gland disease is a kind of disease of serious harm women's health, 30% left and right that at present breast cancer has accounted for women's de novo malignancy, first of rank women's Cancer Mortality.Mastography (Mammography) is by special molybdenum palladium X-ray machine to the breast formation Mammography of taking pictures, and the research of using at present Mammography to carry out mammary gland disease as basic data becomes the most general mode of use in developed country.In the process of the research of mammary gland disease, conventionally need the Mammography image (abbreviation breast image) of comparative analysis left and right sides, the breast image that also needs if desired the same patient of comparative analysis to take at different times.But, uncertain due to camera site, between the breast image of left and right sides, always there are some position difference, meeting impact analysis result, therefore,, in order to navigate to lump rapidly and calcified regions is analyzed, doctor wishes breast image to be registrated to symmetrical position as far as possible, be analyzed easily, thereby promote work efficiency and improve the accuracy rate of analysis result.

Current, the digitizing of medical image has become the inexorable trend of development, 90 mid-terms in year of 20th century, (the American Colleage of Radiology of ACR, ACR) and American National electrical equipment manufacture (the National Electrical Manufactures Association of association, NEMA) combining the medical digital image of having released files and transmission standard DICOM(Digital Imaging and Communication in Medicine) 3.0, the various medical imaging equipments of producing for different manufacturers are (as CT, nuclear magnetic resonance, endoscope, ultrasound wave etc.) interface standard and interaction protocol are provided, for utilizing computer technology directly to process and showing that various medical images provide condition.Data Dynamic scope large (pixel depth is not less than 4096 gray levels conventionally) due to medical image.Therefore, general display is difficult to provide high dynamic range like this once to show the full detail of entire image.In order to utilize the full detail that demonstrates more realistically medical digital images, conventionally adopt Window Display DICOM data, i.e. window width and window level adjustment.

Traditional breast image matching technique mainly contains two kinds: a kind of is feature matching method based on key message, and another kind is the minimum entropy method based on gradation of image value.Characteristic matching method based on key message is first calculated teat placement and the profile information on breast image, and calculates textural characteristics and the contour feature of relevant position, thinks that left and right mammary gland reaches registration position when these features of left and right mammary gland match; The method needs precomputation teat placement and profile information, if these information miscounts can directly cause last registration results failure, its poor robustness; Method based on gradation of image value is got a width breast image as with reference to image, and another width breast image or its part travel through reference picture as floating image, find the maximum position of reference picture and floating image simple crosscorrelation information (Mutual Information).The shortcoming of the method is the breast image that needs searching loop all, and calculated amount is very large, and processing speed is very slow, causes being difficult to requirement of real time.

Summary of the invention

For the defect existing in above-mentioned prior art or deficiency, the object of the invention is to, a kind of method for registering and registration arrangement of the Mammography image based on barycenter are proposed, the method and the system range difference translation mammary gland by the centroid position of left and right breast image reaches left and right mammary gland registration and aligns, and the galactophore image after the registration alignment obtaining facilitates doctor to be analyzed.

In order to realize the demand, of the present invention being achieved as follows:

A method for registering for Mammography image based on barycenter, specifically comprises the steps:

Step 1, load left and right side Mammography image DICOM data subject to registration;

Step 2, respectively original left side Mammography image and right side Mammography image DICOM data are carried out to window width and window bit map, the left side breast image after being converted and right side breast image;

The center-of-mass coordinate R of the left side breast image after step 3, calculation procedure 2 obtains respectively conversion _lcenter-of-mass coordinate R with right side breast image _r;

Step 4, utilize formula 7 from conversion left side breast image and right side breast image in select a pair as moving images Img _s, using another width as template image Img _d;

$\mathrm{Im}{g}_s=\left\{\begin{array}{cc}\mathrm{Im}{g}_r& |R_l-R|\≤|R_r-R|\\ \mathrm{Im}{g}_l& |R_l-R|>|R_r-R|\end{array}\right.---\left(7\right)$

In formula, Img _land Img _rrepresent respectively left side breast image and the right side breast image after conversion; R _l, R _rrepresent with R that respectively the center-of-mass coordinate of left side breast image after conversion is, the centre coordinate of arbitrary width breast image in the center-of-mass coordinate of right side breast image and this two width breast image;

Step 5, calculate translational movement, according to translational movement by moving images Img _scarry out translation and make the alignment of left and right side breast image.

Step 6: show the breast image after registration.

Further, described step 3 specifically comprises the steps:

Utilize formula 4 to calculate the coordinate of the barycenter of breast image:

$R_l=\frac{1}{M_l}{\∫\∫}_{\mathrm{Im}{g}_l}{\mathrm{\ρ}}_l\left[r(x,y)\right]r(x,y)\mathrm{dxdy},R_r=\frac{1}{M_r}{\∫\∫}_{\mathrm{Im}{g}_r}{\mathrm{\ρ}}_r\left[r(x,y)\right]r(x,y)\mathrm{dxdy}---\left(4\right)$

In formula, R _lrepresent the left side breast image Img after conversion _lcenter-of-mass coordinate; R _rrepresent the right breast image Img after conversion _rcenter-of-mass coordinate; R (x, y) represents the coordinate of any point on breast image; ρ _l[r (x, y)] represents left side breast image Img _lthe density of locating at coordinate r (x, y); ρ _r[r (x, y)] represents right side breast image Img _rthe density of locating at coordinate r (x, y), their computing formula is as follows:

${\mathrm{\ρ}}_l\left[r(x,y)\right]=g_{\mathrm{Im}{g}_l}\left(d\right),{\mathrm{\ρ}}_r\left[r(x,y)\right]=g_{\mathrm{Im}{g}_r}\left(d\right)---\left(5\right)$

In formula,

with

be respectively through the left side breast image after window width and window level conversion and the pixel value of right side breast image;

M _lthe total pixel value that represents left side breast image; M _rthe total pixel value that represents right side breast image; Their computing formula is:

$M_l={{\∫\∫}_{\mathrm{Im}{g}_l}\mathrm{\ρ}}_l\left[r(x,y)\right]\mathrm{dxdy},M_r={\∫\∫}_{\mathrm{Im}{g}_r}{\mathrm{\ρ}}_r\left[r(x,y)\right]\mathrm{dxdy}---\left(6\right).$

Further, the computing formula of the translational movement Δ of described step 5 is as follows:

Δ＝R _l-R _r （8）

In formula, R _l, R _rthe center-of-mass coordinate that represents respectively left and right side breast image.

Further, in described step 5, moving images being carried out to translation refers to: as moving images Img _sbe confirmed as left side breast image Img _ltime, to left side breast image Img _lcarry out translation, translational movement is-Δ; As moving images Img _sbe confirmed as right side breast image Img _rtime, to right side breast image Img _rcarry out translation, translational movement is Δ.

Further, the breast image after the demonstration registration of described step 6 refers to the breast image in left side is presented to the left side, and the breast image on right side is presented to the right.

Apply a registration arrangement for the method for registering of the above-mentioned Mammography image based on barycenter, specifically comprise the module being connected successively as follows:

Breast image load module subject to registration: for loading left and right side Mammography image DICOM data subject to registration;

Image window width subject to registration and window bit map module: for original left side Mammography image and right side Mammography image DICOM data are carried out to window width and window bit map, the left side breast image after being converted and right side breast image;

Breast image centroid calculation module: for the center-of-mass coordinate R of the left side breast image after computational transformation _lcenter-of-mass coordinate R with right side breast image _r;

Template image and moving images are selected module: for utilizing left side breast image and the right side breast image of formula 7 from conversion to select a pair as moving images Img _s, using another width as template image Img _d;

$\mathrm{Im}{g}_s=\left\{\begin{array}{cc}\mathrm{Im}{g}_r& |R_l-R|\≤|R_r-R|\\ \mathrm{Im}{g}_l& |R_l-R|>|R_r-R|\end{array}\right.---\left(7\right)$

In formula, Img _land Img _rrepresent respectively left side breast image and the right side breast image after conversion; R _l, R _rrepresent with R that respectively the center-of-mass coordinate of left side breast image after conversion is, the centre coordinate of arbitrary width breast image in the center-of-mass coordinate of right side breast image and this two width breast image;

Breast image centroid calculation module: for calculating translational movement, according to translational movement by moving images Im _gscarry out translation and make the alignment of left and right side breast image;

Breast image display module after registration: for showing the breast image after registration.

Further, in described breast image centroid calculation module, the center-of-mass coordinate of breast image utilizes formula 4 to calculate:

$R_l=\frac{1}{M_l}{\∫\∫}_{\mathrm{Im}{g}_l}{\mathrm{\ρ}}_l\left[r(x,y)\right]r(x,y)\mathrm{dxdy},R_r=\frac{1}{M_r}{\∫\∫}_{\mathrm{Im}{g}_r}{\mathrm{\ρ}}_r\left[r(x,y)\right]r(x,y)\mathrm{dxdy}---\left(4\right)$

In formula, R _lrepresent the left side breast image Img after conversion _lcenter-of-mass coordinate; R _rrepresent the right breast image Img after conversion _rcenter-of-mass coordinate; R (x, y) represents the coordinate of any point on breast image; ρ _l[r (x, y)] represents left side breast image Img _lthe density of locating at coordinate r (x, y); ρ _r[r (x, y)] represents right side breast image Img _rthe density of locating at coordinate r (x, y), their computing formula is as follows:

${\mathrm{\ρ}}_l\left[r(x,y)\right]=g_{\mathrm{Im}{g}_l}\left(d\right),{\mathrm{\ρ}}_r\left[r(x,y)\right]=g_{\mathrm{Im}{g}_r}\left(d\right)---\left(5\right)$

In formula,

with

be respectively through the left side breast image after window width and window level conversion and the pixel value of right side breast image;

M _lthe total pixel value that represents left side breast image; M _rthe total pixel value that represents right side breast image; Their computing formula is:

$M_l={{\∫\∫}_{\mathrm{Im}{g}_l}\mathrm{\ρ}}_l\left[r(x,y)\right]\mathrm{dxdy},M_r={\∫\∫}_{\mathrm{Im}{g}_r}{\mathrm{\ρ}}_r\left[r(x,y)\right]\mathrm{dxdy}---\left(6\right).$

Further, in described breast image centroid calculation module, the computing formula of described translational movement Δ is as follows:

Δ＝R _l-R _r （8）

In formula, R _l, R _rthe center-of-mass coordinate that represents respectively left and right side breast image.

Further, moving images carried out to translation refer to described in described breast image centroid calculation module: as moving images Img _sbe confirmed as left side breast image Img _ltime, to left side breast image Img _lcarry out translation, translational movement is-Δ; As moving images Img _sbe confirmed as right side breast image Img _rtime, to right side breast image Img _rcarry out translation, translational movement is Δ.

Further, after described registration, described in breast image display module, show that breast image after registration refers to the breast image in left side is presented to the left side, the breast image on right side is presented to the right.

Compared with prior art, advantage of the present invention is as follows:

The range difference translation mammary gland of the present invention by the centroid position of left and right breast image reaches left and right mammary gland registration and aligns, can automatically tell in the breast image barycenter of left and right from the nearest width breast image of image center, then another width galactophore image of translation, reaches the object of rationally utilizing screen space to show galactophore image.Through check, result of calculation of the present invention is accurate, and processing speed is fast, and robustness is good, has good adaptability, can effectively meet the registration demand that breast image is read sheet work.

Accompanying drawing explanation

Fig. 1 is the process flow diagram of the method for registering of the Mammography image based on barycenter of the present invention.

Fig. 2 is the process flow diagram of the registration arrangement of the Mammography image based on barycenter of the present invention.

Fig. 3 is that the breast image of two left and right sides subject to registration in embodiments of the invention is at the image of registration front and back.Wherein, (a) be the breast image of two left and right sides subject to registration; (b) be the breast image of the left and right side after above-mentioned two width Image registrations.

Fig. 4 is that the breast image of two left and right sides subject to registration of embodiments of the invention is at the barycenter sign picture of registration front and back.Wherein, (a) be the center-of-mass coordinate of left and right side breast image subject to registration; (b) for adopting the center-of-mass coordinate of the left and right side breast image after method registration of the present invention.

Fig. 5 is the breast image comparing result before and after the registration of randomly drawing in embodiments of the invention.Wherein, (a) and (b), (c), (d), (e) are the contrast situation of 5 couple of choosing at random breast image subject to registration before and after registration.

Embodiment

The term relating in the present invention:

1, the barycenter of object: the barycenter of object is the unique point of in space.Around center of mass point, the quality of object is uniformly distributed.In other words, in space the position vector of the weight of object for center of mass point and equal zero.According to the understanding to barycenter, the object V of quality continuous distribution should be 0 to the Density Weighted integration of the relative position of centroid position R, meets following formula:

∫ _Vρ(r)(r-R)dV＝0 （1）

In formula, V represents to treat the object of integration; R represents to treat the coordinate position of any point on the object of integration; ρ (r) represents that V is in the density at coordinate position r place; R represents the center-of-mass coordinate of object.Solution formula 1 obtains the computing formula of the center-of-mass coordinate of object:

$R=\frac{1}{M}{\∫}_V\mathrm{\ρ}\left(r\right)\mathrm{rdV},$ Wherein: $M={\∫}_V\mathrm{\ρ}\left(r\right)\mathrm{dV}---\left(2\right)$

In formula, M represents the gross mass of object V, other parameter-definition cotypes 1.

2, the window width of Mammography image and window bit map: due to the non-constant width of data codomain scope of Mammography image, in order to show Mammography image on display, need be mapped in given relatively narrow scope, now need according to the following formula it to be carried out to window width and window bit map.

$g\left(d\right)=\left\{\begin{array}{cc}0& d<c-w/2\\ \frac{g_{\max }}{w}(d+w/2-c)& c-w/2\&le;d\&le;c+w/2\\ g_{\max }& d>c+w/2\end{array}\right.---\left(3\right)$

In formula, d represents original DICOM data; G (d) represents the image pixel value of original DICOM data d after window width and window bit map; g _maxthe maximal value of the image pixel value of expression after window width and window bit map, it generally gets the maximum gradation value that display can show; C refers to window position, and w refers to window width, and they are known numeric value.

The method for registering of the Mammography image based on barycenter of the present invention, specifically comprises the steps:

Step 1, load left and right side Mammography image DICOM data subject to registration;

Step 2, respectively original left side Mammography image and right side Mammography image DICOM data are carried out to window width and window bit map according to formula 3, the left side breast image after being converted and right side breast image;

Left side breast image after step 3, calculation procedure 2 obtains respectively conversion and the center-of-mass coordinate of right side breast image; Specifically comprise the steps:

Utilize formula 4 to calculate the center-of-mass coordinate of breast image:

$R_l=\frac{1}{M_l}{\&Integral;\&Integral;}_{\mathrm{Im}{g}_l}{\mathrm{\&rho;}}_l\left[r(x,y)\right]r(x,y)\mathrm{dxdy},R_r=\frac{1}{M_r}{\&Integral;\&Integral;}_{\mathrm{Im}{g}_r}{\mathrm{\&rho;}}_r\left[r(x,y)\right]r(x,y)\mathrm{dxdy}---\left(4\right)$

In formula, R _lrepresent the left side breast image Img after conversion _lcenter-of-mass coordinate; R _rrepresent the right breast image Img after conversion _rcenter-of-mass coordinate; R (x, y) represents the coordinate of any point on breast image; ρ _l[r (x, y)] represents left side breast image Img _lthe density of locating at coordinate r (x, y); ρ _r[r (x, y)] represents right side breast image Img _rthe density of locating at coordinate r (x, y), their computing formula is as follows:

${\mathrm{\&rho;}}_l\left[r(x,y)\right]=g_{\mathrm{Im}{g}_l}\left(d\right),{\mathrm{\&rho;}}_r\left[r(x,y)\right]=g_{\mathrm{Im}{g}_r}\left(d\right)---\left(5\right)$

In formula, with be respectively through the left side breast image after window width and window level conversion and the pixel value of right side breast image;

M _lthe total pixel value that represents left side breast image; M _rthe total pixel value that represents right side breast image; Their computing formula is:

$M_l={{\&Integral;\&Integral;}_{\mathrm{Im}{g}_l}\mathrm{\&rho;}}_l\left[r(x,y)\right]\mathrm{dxdy},M_r={\&Integral;\&Integral;}_{\mathrm{Im}{g}_r}{\mathrm{\&rho;}}_r\left[r(x,y)\right]\mathrm{dxdy}---\left(6\right)$

Step 4, utilize formula 7 from conversion left side breast image and right side breast image in select a pair as moving images Img _s, using another width as template image Img _d;

$\mathrm{Im}{g}_s=\left\{\begin{array}{cc}\mathrm{Im}{g}_r& |R_l-R|\&le;|R_r-R|\\ \mathrm{Im}{g}_l& |R_l-R|>|R_r-R|\end{array}\right.---\left(7\right)$

In formula, Img _land Img _rrepresent respectively left side breast image and the right side breast image after conversion; R _l, R _rrepresent with R that respectively the center-of-mass coordinate of left side breast image after conversion is, the centre coordinate of arbitrary width breast image (the present invention gives tacit consent to two width breast image equal and opposite in directions) in the center-of-mass coordinate of right side breast image and this two width breast image;

Step 5, calculating translational movement, by moving images Img _scarry out translation and make the alignment of left and right side breast image.As moving images Img _sbe confirmed as left side breast image Img _ltime, to its left side breast image Img _lcarry out translation, translational movement is-Δ; Otherwise, as moving images Img _sbe confirmed as right side breast image Img _rtime, just need to be to right side breast image Img _rcarry out translation, translational movement is Δ.

Δ calculates by following formula:

Δ＝R _l-R _r （8）

In formula, R _l, R _rthe center-of-mass coordinate that represents respectively left and right side breast image;

Step 6: show the breast image after registration: watch in order to facilitate doctor to contrast, by the symmetrical demonstration of the breast image of the left and right side obtaining after step 5 translation, the breast image in left side is presented to the left side, the breast image on right side is presented to the right.

For validity of the present invention is described.Inventor adopts system of the present invention to carry out registration test 1000 pairs of breast images.Operating system in test is that Win764, CPU are Intel (R) Core (TM) i5-2300CPU2.80GHz.As shown in Fig. 3 (a), between the breast image of the left and right side that a pair is subject to registration, there is certain dislocation; Adopt method of the present invention to carry out after registration, obtain the breast image of the left and right side after the registration as shown in figure (b).Before and after registration, the centroid position of this pair of breast image is as shown in Fig. 4 (a) and Fig. 4 (b).The result demonstration of 1000 pairs of images of this test, be 23.11ms the averaging time of the registration process of every pair of breast image, and its result is reliable and stable, and Fig. 5 (a)-Fig. 5 (e) has shown the registration effect of 5 pairs of breast images randomly drawing.

Claims (10)

1. a method for registering for the Mammography image based on barycenter, is characterized in that, specifically comprises the steps:

Step 1, load left and right side Mammography image DICOM data subject to registration;

Step 2, respectively original left side Mammography image and right side Mammography image DICOM data are carried out to window width and window bit map, the left side breast image after being converted and right side breast image;

The center-of-mass coordinate R of the left side breast image after step 3, calculation procedure 2 obtains respectively conversion _lcenter-of-mass coordinate R with right side breast image _r;

Step 4, utilize formula 7 from conversion left side breast image and right side breast image in select a pair as moving images Img _s, using another width as template image Img _d;

$\mathrm{Im}{g}_s=\left\{\begin{array}{cc}\mathrm{Im}{g}_r& |R_l-R|\&le;|R_r-R|\\ \mathrm{Im}{g}_l& |R_l-R|>|R_r-R|\end{array}\right.---\left(7\right)$

In formula, Img _land Img _rrepresent respectively left side breast image and the right side breast image after conversion; R _l, R _rrepresent with R that respectively the center-of-mass coordinate of left side breast image after conversion is, the centre coordinate of arbitrary width breast image in the center-of-mass coordinate of right side breast image and this two width breast image;

Step 5, calculate translational movement, according to translational movement by moving images Img _scarry out translation and make the alignment of left and right side breast image;

Step 6: show the breast image after registration.

2. the method for registering of the Mammography image based on barycenter as claimed in claim 1, is characterized in that, described step 3 specifically comprises the steps:

Utilize formula 4 to calculate the coordinate of the barycenter of breast image:

$R_l=\frac{1}{M_l}{\&Integral;\&Integral;}_{\mathrm{Im}{g}_l}{\mathrm{\&rho;}}_l\left[r(x,y)\right]r(x,y)\mathrm{dxdy},R_r=\frac{1}{M_r}{\&Integral;\&Integral;}_{\mathrm{Im}{g}_r}{\mathrm{\&rho;}}_r\left[r(x,y)\right]r(x,y)\mathrm{dxdy}---\left(4\right)$

In formula, R _lrepresent the left side breast image Img after conversion _lcenter-of-mass coordinate; R _rrepresent the right breast image Img after conversion _rcenter-of-mass coordinate; R (x, y) represents the coordinate of any point on breast image; ρ _l[r (x, y)] represents left side breast image Img _lthe density of locating at coordinate r (x, y); ρ _r[r (x, y)] represents right side breast image Img _rthe density of locating at coordinate r (x, y), their computing formula is as follows:

${\mathrm{\&rho;}}_l\left[r(x,y)\right]=g_{\mathrm{Im}{g}_l}\left(d\right),{\mathrm{\&rho;}}_r\left[r(x,y)\right]=g_{\mathrm{Im}{g}_r}\left(d\right)---\left(5\right)$

In formula,

with

be respectively through the left side breast image after window width and window level conversion and the pixel value of right side breast image;

M _lthe total pixel value that represents left side breast image; M _rthe total pixel value that represents right side breast image; Their computing formula is:

$M_l={{\&Integral;\&Integral;}_{\mathrm{Im}{g}_l}\mathrm{\&rho;}}_l\left[r(x,y)\right]\mathrm{dxdy},M_r={\&Integral;\&Integral;}_{\mathrm{Im}{g}_r}{\mathrm{\&rho;}}_r\left[r(x,y)\right]\mathrm{dxdy}---\left(6\right).$

3. the method for registering of the Mammography image based on barycenter as claimed in claim 1, is characterized in that, the computing formula of the translational movement Δ of described step 5 is as follows:

Δ＝R _l-R _r （8）

In formula, R _l, R _rthe center-of-mass coordinate that represents respectively left and right side breast image.

4. the method for registering of the Mammography image based on barycenter as claimed in claim 1, is characterized in that, in described step 5, moving images is carried out to translation and refers to: as moving images Img _sbe confirmed as left side breast image Img _ltime, to left side breast image Img _lcarry out translation, translational movement is-Δ; As moving images Img _sbe confirmed as right side breast image Img _rtime, to right side breast image Img _rcarry out translation, translational movement is Δ.

5. the method for registering of the Mammography image based on barycenter as claimed in claim 1, is characterized in that, the breast image after the demonstration registration of described step 6 refers to the breast image in left side is presented to the left side, and the breast image on right side is presented to the right.

6. application rights requires a registration arrangement for the method for registering of the Mammography image based on barycenter described in 1, it is characterized in that, specifically comprises the module being connected successively as follows:

Breast image load module subject to registration: for loading left and right side Mammography image DICOM data subject to registration;

Image window width subject to registration and window bit map module: for original left side Mammography image and right side Mammography image DICOM data are carried out to window width and window bit map, the left side breast image after being converted and right side breast image;

Breast image centroid calculation module: for the center-of-mass coordinate R of the left side breast image after computational transformation _lcenter-of-mass coordinate R with right side breast image _r;

Template image and moving images are selected module: for utilizing left side breast image and the right side breast image of formula 7 from conversion to select a pair as moving images Img _s, using another width as template image Img _d;

$\mathrm{Im}{g}_s=\left\{\begin{array}{cc}\mathrm{Im}{g}_r& |R_l-R|\&le;|R_r-R|\\ \mathrm{Im}{g}_l& |R_l-R|>|R_r-R|\end{array}\right.---\left(7\right)$

In formula, Img _land Img _rrepresent respectively left side breast image and the right side breast image after conversion; R _l, R _rrepresent with R that respectively the center-of-mass coordinate of left side breast image after conversion is, the centre coordinate of arbitrary width breast image in the center-of-mass coordinate of right side breast image and this two width breast image;

Breast image centroid calculation module: for calculating translational movement, according to translational movement by moving images Img _scarry out translation and make the alignment of left and right side breast image;

Breast image display module after registration: for showing the breast image after registration.

7. the method for registering of the Mammography image based on barycenter as claimed in claim 6, is characterized in that, in described breast image centroid calculation module, the center-of-mass coordinate of breast image utilizes formula 4 to calculate:

$R_l=\frac{1}{M_l}{\&Integral;\&Integral;}_{\mathrm{Im}{g}_l}{\mathrm{\&rho;}}_l\left[r(x,y)\right]r(x,y)\mathrm{dxdy},R_r=\frac{1}{M_r}{\&Integral;\&Integral;}_{\mathrm{Im}{g}_r}{\mathrm{\&rho;}}_r\left[r(x,y)\right]r(x,y)\mathrm{dxdy}---\left(4\right)$

In formula, R _lrepresent the left side breast image Img after conversion _lcenter-of-mass coordinate; R _rrepresent the right breast image Img after conversion _rcenter-of-mass coordinate; R (x, y) represents the coordinate of any point on breast image; ρ _l[r (x, y)] represents left side breast image Img _lthe density of locating at coordinate r (x, y); ρ _r[r (x, y)] represents right side breast image Img _rthe density of locating at coordinate r (x, y), their computing formula is as follows:

${\mathrm{\&rho;}}_l\left[r(x,y)\right]=g_{\mathrm{Im}{g}_l}\left(d\right),{\mathrm{\&rho;}}_r\left[r(x,y)\right]=g_{\mathrm{Im}{g}_r}\left(d\right)---\left(5\right)$

In formula, with

be respectively through the left side breast image after window width and window level conversion and the pixel value of right side breast image;

M _lthe total pixel value that represents left side breast image; M _rthe total pixel value that represents right side breast image; Their computing formula is:

$M_l={{\&Integral;\&Integral;}_{\mathrm{Im}{g}_l}\mathrm{\&rho;}}_l\left[r(x,y)\right]\mathrm{dxdy},M_r={\&Integral;\&Integral;}_{\mathrm{Im}{g}_r}{\mathrm{\&rho;}}_r\left[r(x,y)\right]\mathrm{dxdy}---\left(6\right).$

8. the method for registering of the Mammography image based on barycenter as claimed in claim 6, is characterized in that, in described breast image centroid calculation module, the computing formula of described translational movement Δ is as follows:

Δ＝R _l-R _r （8）

In formula, R _l, R _rthe center-of-mass coordinate that represents respectively left and right side breast image.

9. the method for registering of the Mammography image based on barycenter as claimed in claim 6, is characterized in that, moving images is carried out to translation refer to described in described breast image centroid calculation module: as moving images Img _sbe confirmed as left side breast image Img _ltime, to left side breast image Img _lcarry out translation, translational movement is-Δ; As moving images Img _sbe confirmed as right side breast image Img _rtime, to right side breast image Img _rcarry out translation, translational movement is Δ.

10. the method for registering of the Mammography image based on barycenter as claimed in claim 6, it is characterized in that, after described registration, described in breast image display module, showing that breast image after registration refers to is presented at the left side by the breast image in left side, and the breast image on right side is presented to the right.

Images