CA2258730A1  Method and apparatus for threedimensional reconstruction of coronary vessels from angiographic images  Google Patents
Method and apparatus for threedimensional reconstruction of coronary vessels from angiographic imagesInfo
 Publication number
 CA2258730A1 CA2258730A1 CA 2258730 CA2258730A CA2258730A1 CA 2258730 A1 CA2258730 A1 CA 2258730A1 CA 2258730 CA2258730 CA 2258730 CA 2258730 A CA2258730 A CA 2258730A CA 2258730 A1 CA2258730 A1 CA 2258730A1
 Authority
 CA
 Grant status
 Application
 Patent type
 Prior art keywords
 vessel
 points
 biplane
 projection
 object
 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.)
 Abandoned
Links
Classifications

 G—PHYSICS
 G06—COMPUTING; CALCULATING; COUNTING
 G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
 G06T11/00—2D [Two Dimensional] image generation
 G06T11/003—Reconstruction from projections, e.g. tomography
 G06T11/006—Inverse problem, transformation from projectionspace into objectspace, e.g. transform methods, backprojection, algebraic methods

 G—PHYSICS
 G06—COMPUTING; CALCULATING; COUNTING
 G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
 G06T2211/00—Image generation
 G06T2211/40—Computed tomography
 G06T2211/404—Angiography
Abstract
Description
Wo 97/49065 PCT/US97/10194 METHOD AND APPARATUS FOR THRE~DIl\~ENSIONAL RECONSTRUCTION
2 OF CORONARY VESSELS EROM ANGIOGRAPHIC Il\~AGES
3 A portion of the dicclos~re of this patent document corlt~in~ m~t~ri~l which 4 is subject to copyright ~lote~ion. The copyright owner has no objection to the facsimilP
repro~ otion by anyone of the patent ~oc~ or the patent ~i~closure~ as it appears in the 6 Patent and Trademark Office patent file or records, but otherwise reserves all copyright 7 rights wLaL~Ge~
8 BACKGROI~ND OF THE INVENTION
9 The present invention relates generally to a method for ~ or~ cting images of coronary vessels and more ~e.,irically to a method for threedim~mion~l (3D) 11 reconstruction of colonal,y vessels from two two~iimPn~ional biplane projection images.
12 Several investi~tnrs have l~Ollcd col~ulcr assisted m~tho~s for estirn~tit~n of the 13 3D coronaly arteries from biplane projection data. These known m~thr~ are based on the 14 known or standard Xray g~ ft~ y of the projections, pl~rf .n~l~t of l~lu~,~.l.!;, known vessel ~5 shape, and on ilelaLiv~ Id~ ~lirc~;on of m~trhing structures in two or more views. Such 16 methods are described in a publication entitled "3D digital subtraction angiograph~," IEEE
17 Trans. Med. Irnag., vol. MIl, pp. 152158, 1982 by H.C. Kim, B.G. Min, T.S. Lee, et.
18 al. and in a publication entitled "Methods for evAll~tin~ cardiac wall motion in 3D using 19 bifurcation points of the co.o,a~y arterial tree," Invest. Radioiogy, Jan.Feb. pp. 4756, 1983 by M.J. Potel, J.M. Rubin, and S. A. Mackay, et al. Because the computation was designed WO 97149065 PCT/USg7/10194 for predefined views only, it is not suitab}e to solve the reconstruction problem on the basis 2 of two projection images acquired at url~llr~r~y and unknown relative orientations.
3 Another known method is based on motion and multiple views acquired in a 4 singleplane im~gin~ system. Such a method is described in a publication entitled S "Recoll~L,lcting the 3d medial axes of cGl( n~)~ arteries in singleview cin~n~iograms, "
6 IEEE Trans. MI, vol. 13, no. 1, pp. 4860, 1994 by T.V. Nguyen and J. Sklansky uses 7 motion llallsrolllations of the heart model. ~owever, the motion transformations of the 8 heart model consist only of rotation and scaling. By incorporation of the centerleft,.~llced 9 mlo.thnrl, initial depth coo1~inatcs, and center cool.lil~ates, a 3D skeleton of the colonaly arteries was obtained. However, the reaI heart motion during the contraction involves five 11 specific movements: translation, rotation, Wlil~,illg, accordionlike motion, and movement 12 toward the center of the vPntri~ r ch~mber. Therefore, the model employed is not general 13 enough to portray the tme motion of ~e heart, especi~lly toward the endsystole.
14 Knowledgebased or rulebased systems have been proposed for 3D reconstruction of coronary arteries by use of a vascular r~,twolh model. One such knowledgebased system 16 is d~sc,il)ed in a publication entitled "An expert system for the labeling and 3D
17 lccol~Lluction of the coronary arteries from two projections," Inte~nationa~ Jou~nal of 18 Imaging, Vol. 5, No. 23, pp. 145154, 1990 by Smets, Vandewerf, Suctens, and 19 Oosterlinck. Because the rules or knowledge base were ol~ani~ed for certain specific conditions, it does not generalize the 3D ~~co~ u~;lion process to arbitrary projection data.
21 In other knowledgebased systems, the 3D COlOl~al,y arteries were reconstructed from a set 22 of Xray perspective projections by use of an algorithrn from computed tomography. Due 23 to the motion of the heart and only a limited number of projections (four or six), accurate 24 reconstruction and qll~ntit~tive measul~llelll are not easily achieved.
., . _~
MISSING UPON TIME OF PUBLICATION
MISS~G UPON TIME OF PUBLICATION
CA 02258730 l998l2l8 Angiograms of fifteen patients were analyzed in which two cases are selected for 2 discussion hereinafter. The biplane imaging geolllcny was first deterrnined without a 3 calibration object, and the 3D coronary arterial trees were reconstructed, including both left 4 and right coron~ artery systems. Various t~vo~imPn~ioll~l (2D) projection images of the 5 l~co~laLIu._t~d 3D cor~naly arterial tree were gelL.aLed and compared to other viewing 6 angles obtained in the actual patient study. Similarity beL~.~,en the real and iecohsuucl~d 7 arterial ~ clul~ s was eYcel~nt The ac~ul~ of this method was evaluated by using a 8 col~ J~rlsim~ te~ colol~aly arterial tree. Rootmeansquare (RMS) errors in the 3D
9 pG~iliOll and the 3D confi~ ration of vessel c~ lt~ s and in the angles defining the R
matrix and t vector were 0.9  5.5 mm, 0.7  1.0 rnm, and less than 1.5 and 2.0 degrees, 11 l~s~eelively, when using 2D vessel c~ lines with RMS normally distributed errors varying 12 from 0.4  4.2 pixels (0.25  1.26 rnm).
13 More sl,ec;r~Ally, the method for three~im~ ion~l l.,cons~ ction of a target object 14 from twodim~n~ional images involves a target object having a plurality of branchlike 15 vessels. The method includes the steps of: a) placing the target object in a position to be 16 sc~nn~(l by an im~in~ system, the im~inE system having a plurality of im~EinE polLions;
17 b) ac4ui~ g a plurality of projection images of the target object, each im~ging portion 18 providing a projection irnage of the target object, each im~gin~ portion disposed at an 19 unknown olic~ ion relative to the other im~in~ portions; c) identifying twolimt~n~ n~l 20 vessel ce,lle,lines for a predetermined number of the vessels in each of the projection images;
21 d) creating a vessel hierarchy data ~Lr~lCLule for each projection image, the data structure 22 including the identified two(limen.~ional vessel centerlines defined by a plurality of data 23 points in the vessel hierarchy data structure; e) calculating a predetermined number of 24 bifurcation points for èach projection image by traversing the corresponding vessel hierarchy , . ~ .~ ........ . . .. .
wo 97/49065 PCT/IJS97/10194 data strUclure, the bifurcation points defined by intersections of the twodimensional vessel 2 centerlines; f) dele.llli,ing a transforrnation in the forrn of a rotation matrix and a translation 3 vector utili7ing the biru~;ation points corresponding to each of the projections images, the 4 rotation matrix, and the translation vector le~l~senlin~ im~ging pal~a,.,ctel~ corresponding to 5 the relative olie,l~lions of the im~ging portions of the im~ging system; g) lltili7.ing the data 6 points and the ll~r~r~ alion to establish a co,~ .olldence b~,.weell the two~ c;~r~al 7 vessel centerlines col~,s~onding to each of the projection images such that each data point 8 co,.~;",onding to one projection image is linked to a data point corresponding to the other 9 projection images, the linked data points le~leàe~ g an identir~l location in the vessel of 10 the target object after the projections; h) c~lMll~tirlg three~ .,P,~ional vessel cel,Lellilles 11 ~tili7.ing the twodimensional vessel centerlines and the c~,,c~llol~dence between the data 12 points of the twodimensional vessel centerlines; and i) ,ecoris~.u~ ,g a three~im~cional 13 visual ,~ ese~ .tinn of the target object ~ased on the three~imP.~ onal vessel centerlines and 14 di~snPters of each vessel esl;~ along the threeAimPncional centerline of each vessel; and j) determining the optimal view of the vessel segm~n1c with minim~l vessel ro~sho~lLI.i,~g.
16 BRIEF DESCRIPrION OF THE DRAWINGS
17 The features of the present invention which are believed to be novel are set forth with 18 particularity in the appended claims. The invention, together with further objects and 19 advantages thereof, may best be ul~de,~lood by ,ef~ ce to the following description in conju"t;lion with the accompanying drawings.
21 Fig. 1 is a high level flowchart illustrating the steps acco,di~,g to a specific 22 embodiment of the present inventive method;
23 Figs. 2A and 2B are schematic views of the imaging system parLicularly showing the 24 position of the isocemer;
.. . . . ...
W O 97/4906S PCT~US97/10194 Figs. 3A3B are orthogonal views of typical biplane projection images based on a2 cor,ll)ule~ sim~ t~d cor.)llaly arterial tree where each vessel centerline is lc~le~e~ d by a 3 sequence of data points;
4 Fig. 4 is a s~ view of a specific embodiment of a biplane imaging system model for ~fining a 3D ob3eet point in the 3D coordillale systems xyz and x'y'z' and their 6 pro3ections in the 2D irnage plane coordinate systems uv and u'v', l.,~e~.lively, according 7 to the present invention;
8 Fig. 5 is a sch~ ~a~;r view showing two initial solutions used for searching the optimal 9 solutions of im~gin~ pa,.~ t~"~ as well as the 3D object in the given biplane systems;
Fig. 6 is a s~hF~ view showing coneshape bounding regions ~oci~d with the 11 c~lcnl~t~d 3D points; and 12 Fig. 7 is a sçl~ ;c view showing two setups of a biplane imaging system resulting 13 from the employed two initial con~itior.C yielding two sets of lccon~L,ucL~d 3D objects A'14 D' and A  D (real 3D object points) as shown in gray and black circles, ~ ecLi~ly.
DETA~LED DESCRIPIION OF THE INVENTION
16 Referring now to Fig. 1, the present invention method inrlutles eight major steps: (1) 17 acquiring biplane p, je~,Lion images of the cO~ y ~LluClul~, as shown in step 20, (2) 18 ~etPcting, segmenting, and idc.lli~yii g vessel centerlines and coll~Llu~Lillg a vessel hie,~,~;hy 19 ,epl~s~"LaLion, as illu~LIat~d in step 22, (3) calc~ tin~ bi~,ca~ion points and measuring vessel ~ te,~ in colonal~ angiograms, as shown in step 24, but only if biplane im~ging 21 geometry data is not available, as shown by the "no" branch of step 26, (4) (leterrninin~
22 biplane im~gin~ parameters in terrns of a rotation matrix R and a unit translation vector t 23 based on the identifred bifurcation points, as illustrated in step 28, (5) retrieving im~ging 24 parameters, as shown in step 30, but only if known biplane imaging geometrv data is . . . ~, . .
available, as shown by the "yes' branch of step 26, (6) establishing the cen~erline 2 correspondences of the twodimensional arterial re~l~selltations, as shown in step 3~, (7) 3 c~lc~ tin~e and recovering the 3D spacial inrullllalioll of the coronary arterial tree based on 4 the c~lrul~t~1 biplane im~in~ parameters, the cor.e~l on~f n~.s of vessel centerlines, and the 5 vessel diallleL~ls, as illustrated in step 34, and (8) rendering the reconstructed 3D coronary 6 tree and estim~ting an optimal view of the va~cul~tllre to ~ e vessel overlap and vessel 7 foresho~ g, as shown in step 36. The abovedesc~ibed steps will be described in greater 8 detail he~vilh.
9 Acquirin~ Ima~es From The Biplane Im~in~ Svstem As shown in step 20 of Fig. 1, biplane projection images are acquired using an Xray 11 based im~in~ system. However, other nonXray based im~in~ systems may be used, as 12 will be described h~lclllGrl~l. Such Xray based images are preferably created using a 13 biplane im~ing system where two projection images are produced s~lbst~nti~lly 14 simultaneously. The patient is placed in a position so that the target object, in this illustrated 15 embodiment, the heart, is sc~nned by the im~ging system. The imaging system preferably 16 inr~ P~ a plurality of im~in~ portions where each ;..~ag;i,~ portion provides a projection 17 image of the COI~ ullelul~. Due to ill~lr~ .ence ~ ell the im~in~ portions during 18 Xray Pmi~siQn) one image portion must be turned off while the other image portion is 19 active. However, the duration of exposure is extremely short so that the two images are 20 sequentially taken such that the heart does not signific~ntly move during the imaging period.
21 Such an im~ging system may be, for example, a SeimPnc BICORE system. It is the time 22 between e~osul~s that must be short. Otherwise blurring of the vessels may result. The 23 motion of the heart is significant throughout most of the heart cycle, thus most investigators 24 use enddiastole.
Wo 97/49065 PCT/US97tlO194 Referring now to Fig. 1, and Figs. 2A2B, Figs. 2A2B schPm~tir~l]y illustrate a 2 typical im~ging system configuration where only one gantry arm is shown for clarity. An 3 Xray source is located at the focus of a cone shaped Xray beam which diverges outwardly 4 toward an Xray image ;"t. n~;rer (image plane or view). The Xray beam passes through S the target object and is received by the Xray image i,l. r;~irer In a biplane system. two 6 such Xray sources and image illlr,l.cirl~,s are present. The gantry arm is rotatably mounted 7 such that the Xray source and the image intensifier move relative to the target:object but 8 always remain fixed relative to each other. The gantrv arm permits the Xray source and the 9 image il.~r~~;rFr to be pos;~ n~d about the target object in a p~d~t~ lined orientation.
The present h.~elllioll is not limited to a biplane im~ging system having only two 11 im~ing portions. A multiplane im~in~ system may be used having a plurality of im~ing 12 portions such as, for example, three, four, or more im~in,~ portions, each providing a 13 projection image of the cool~ly structure. Such an im~vin~ system is ess~nti~lly limited by 14 the size of the system relative to the treatment facility. The scope of the present inventive 15 method includes use of a system providing two or more projection images.
16 The projection images may or may not include orientation information describing the 17 relative gcollleLly b~t~ the gantry arms. Even if the individual gantry position 18 i..rollllation is available, the derivation of relative oLi~lL~tion based on the known gantry 19 pG~iLiOllS beco...es a nontrivial process especi~lly when the two icocf.,l~ ,s are not aligned.
20 A signifit ~nt feature of the present inventive method permits reconstruclion of a 3D model 21 of the target object even when the relative o..c.llaLion of the gantry arms is unknown. Other ~ prior art methods for lcco~ cLion the 3D model re~uires either known relative orientation 23 between the two views, known individual gantry position (resulting in ten parameters for wo 97/49065 PCT/US97/10194 biplane imaging geometry which may be sufficiently defined by five parameters), or at least 2 more that eight pairs of accurate input corresponding points.
3 The projection images are digitized and are entered into a work station. The 4 projection images are ~ ed by a series of individual pixels limited only by the 5 resolution of the im~Pin~ system and the memory of the work station. The work station, for 6 example, may be an IBM RISC/6000 work station or a Silicon Graphics Indigo2/High 7 Impact work station or the like. An input device, such as a keyboard, a mouse, a light pen, 8 or other known input devices is included to permit the operator to enter data.
9 Se~mentation and Peature E~L,dclion of the TwoDimensional Coronary Arterial Tree False detection of arteries is inevitable using a fully ~ u,~tic vessel tracker, 11 especially when vessels overlap or the signaltonoise ratio of the angiogram is low. In the 12 present inventive meth~l, a semialtom~tiC system based on a t~rllnirluP using a deformation 13 model is employed for the idf~ntifir~tion of the 2D coronary arterial treein the angiograms, 14 as will be described in greater detail he~ ,afler~ The required user interaction involves only 15 the indication of several points inside the lumen, near the projection of vessel centerline in 16 the angiogram, as is illuctrat~cl in step 22 of Fig. 1.
17 Refening now to Figs. 3A and 3B, co~ uL, l sim~ t~d biplane l)lojecLion images are 18 shown. Typically, the operator or the physician inspects the ~ iti7Pd biplane projection 19 images. Using a mouse or other data entry device, the physician marks a series of points 20 (data points) within a major vessel to define the initial twodimPncional ce,ll. .lhle of the 21 vessel. After the major vessel has been marked, five additional b.~nclling vessels are marked 22 in the same manner such that a total of six twodimensional centerlines are identjfiPd. Once 23 the major vessel has been marked, the rem~inin~J five vessels may be identified and marked . . ~ . , . . ~
Wo 97/49065 PCTIUS97/10194 in any order. Note, for pulposes of clarity only, Figs. 3B and 3B show one major vessel and 2 only four branching vessels.
3 The above process is then pelrullllcd for the other biplane projection image(s).
4 Again, the operator or the phy~icidn id~";r~s and marks the major vessel and then identif1~s and marks the five a~t~ition~l blan~ vessels. The branching vessels may be ide~tifi~d and 6 marked in any order, which may be different from the order marked in the first biplane 7 projection image. The l~h~ n must mark the same six vessels in each biplane projection 8 image. The present inventive method renders excellent results with use of only six i~lentifi~d 9 vessel centerlines. Other known systems require many more id~ntifi~d data points. An example of such a system is desciL.cd in a publication entitled "~ lllF " of Diffuse 11 Coronary Artery Disease by Qu~ t;~e Analysis of Coronary Morphology Based Upon 3D
12 ReconsLIuction from Biplane Angiograms," IEEE Trans. on Med. Imag., Vol. 14, No. 2, 13 pp. 230241, 1995 by A. Wahle and E. Wellnhofer et. al. This system requires at least ten 14 or more identified corresponding points due to ten variables that need to be optimized for determination of biplane im~ging ~al~l,.,t~ls. Such a burdensome requirement ~ignifir~ntly 16 increases col.u~r ~.ocessing tirne. Since the derived objective f~ll LiOll employs five 17 variables to characterize the biplane im~gin~ ge~ chy, it only requires five col,;.l,ond.l.g 18 points.
19 Next, by use of the defolll.aLion model and ridgepoint o~e,dtor, described 20 he.th~afL~, the initially identified cc~L~linc is gradually deformed and made to finally reside 21 on the actual centerline of vessel.
... . .... . .
W O 97/4906~ PCTrUS97/10194 Deforrnation model 2 The behavior of the deformation contour is controlled by internal and ex~ernal forces.
3 The internal forces serve as a smoothness constraint and the external forces guide the active 4 contour toward image ~alul~s. A deformable contour can be defined as a llla~ g of a S material coodi,~te s ~ [0,17 into Rt.
6 v(s) = (x(s),y(s)) 7 Its associated energy function can be defined as the sum of an internal energy and an external 8 energy. The external energy ~ccoun~ for the ima~e force ~j",~" such as a line or edge 9 e~ ed from the image content, and other e~rtPm~l constraint forces ~o~rr~ such as a pulling force intentionally added by the user. T~e energy function is the sum of the internal energy 11 and the external energy and is written as:
12 E(v) = E,n,(v) + E,~,(v~
13 = E~n~(v) + Ejm~(V) + Eocr!(v) 14 Equ. (A) The internal energy, which is caused by s~et.;l~illg and bending, characterizes the deformable 16 material and is modeled as:
Eint(~r) = 2J~{a (s) Iv~(s) ¦Z+~ (S) ~ (S) l2}dS
18 Equ. (B) 19 where cx(s) and l~(s) control the tension and the rigidity at point v(s). The first order term, measuring the length of ds, resists ~k~lchillg~ while the second order term, measuring the 21 curvature at ds, resists bending.
Let ~~(v) denote the image force at point v(s), which is the directional maximum (or 2 minimum~ response of gray level in a region with m by m pixels These point sources of 3 force are referred to as ridge points which will act on the contour. Based on the image 4 force, the image energy is defined as Eimg(~) =~ 2 J ~AI(V(S) ) l2ds 6 Equ. (C) 7 Here, 02l1y the ridge points are considered as the ~rtçrn~l force. The shape of CoMOUr under 8 the forces becoll.es a problem of .l~ AtiQn of the energy funrtion E(~) = 2 J ~ (s) ¦V~(S) 12+~(s) ¦V11(5) 12 ~ I (V (S) ) ¦2~d Equ. (D) 11 A nt~cessh.y condition for a contour function to minimi~o Equ. (D) is that it must satisfy the 12 following EulerLagrange equ~tion13 (~v')' + G6'v")" + F,m"(v) = O
14 Equ. (E) 15 When all the forces that act on the contour are bal~nred7 the shape change on the contour is 16 negligible and results in an e~uilibrium state.
17 The deformation process starts from an initial contour. The contour modifies its 18 shape dy~ ir~lly according to the force fields described above until it reaches an 19 equilibrium state. The user m~ml~lly in~ic~t~s points near the vessel centerline and a 20 splinecurve is formed based on the selected points. This serves as the initial centerline of .
l4the vessel. Without loss of generality, the artery is imaged as darlc imensity aYainst the 2 background in the angiogram. According to the densitometric profile model, the 2D
3 crosssectional profile of the coronary arteries has a minimllm intensity at the vessel 4 centerline. An m by m operator is convolved with a given arterial image by which the ridge S pixel is idPntifieti if it is a directional Illil;llllllll on intensity. By use of the defo~lllalion 6 model, the set of ridge pixels seNes as the external forces which act on the initial model 7 curve such that it will gradually be deformed and finally reside on the real cente~ e of the 8 vessel.
9 A coll,~uultrbased editing tool is employed for the correction in case a falsenegaliv~
10 or falsepositive detection occurs. The editing tool provides the operator with the ability to 11 change the shape of the vessel ce,~ lille. This is done by the modification of control points, 12 such as addition, deletion, or dragging of the points on a splinebased curve that models the 13 vessel centerline.
14 The identified centerlines and the bldlchilg relationships are used to construct the 15 hierarchy in each image by their labeling according to the ap~lu~liate anatomy of the 16 primary and secol~d~.y colonal~ arteries. The l~elin~ process on the colollaly tree is 17 pe.r~ led ~ ic~lly by application of the breadthfirst search algorithm to Ll~
18 i(l~ntifiPcl vessel cellL~,lillcs, as is known in the art. From each vessel of the coronary tree 19 that is ~;ull~ ly visited, this ap~r~ach sealches as broadly as possible by next visiting all of 20 the vessel centerlines that are adjacent to it. This finally results in a vessel hierarchically 21 directed graph (digraph~ cont~ining a set of nodes corresponding to each individual artery and 22 two types of arcs (descen~nt and sibling arcs) defining the coronary an~Lol"y.
23 ln addition to the coordinates of the vessel centerlines, the rela~ive di~mPt~rs of the 24 vessels are deterrnined. The diameter of each vessel is esrim~ted based on the maximum .... , ..... ....... . . _~ .
W O 97/49065 PCT~US97/10194 vessel diameter at a beginning portion of the vessel and a minimllm vessel diame~er at an Z ending portion of the vessel. This step is perforrned within step 22 (Fig. 1) and is also 3 pclru~ ed by the operator or the physician. The physician measures the minimllm and 4 m~Aximllm vessel tliAmeter on the projection image at the begil~ling of a vessel and at the end of the vessel and enters the data into the work station. Only the minimllm and mAximllm 6 diA..~ t~ ~ are l~uh~d since typical values of vessel taper are known and are sll~stAnti~lly 7 Col~L~n~ from patient to patient. For e~a~ lc, a vessel may have a maximum (l;Am~ter of 8 0.35 mm at the proximal RCA and a ".i.. ;.. """ rli~meter of 0.02 mm at the distal RCA. The 9 remAinin~ riiAIllP.~ bc~ ,e,l the two points can be rAlolllAte~l by linear interpolation based 10 on the maxilllUlll and Illill;lll~llll r1;A~
11 Next, a ~et~ .,;.~AI;on is made whether biplane imA~in~ geolll~ly is available, as 12 ill~lsLldltd in step 26 of Fig. 1. As described above, the geometric orientation of the gantries 13 during e,~yo~ul~ may not be available or alteTT~t~ly, if it iS availab!e, may require a 14 calibration process. Based on the current imaging technology, the information of a single plane system includes the gantry orientation (LAO and CAUD angles), SID (focal spot to 16 image ;.,lrn~;r.~ tAnre), and m~.,;rra~ion. However, such information is defined based 17 on each individual l~Ç~,,c.lce system. In other words, the relative o~ alion that 18 cll~ac~l.,es the two views is unknown. Th ~er ~le, it is n~eeCc~ y to tletermin~ the biplane 19 geometry. If the two l~ ce points, which are the location of the isocenters, are made to coincide, the relative orientation can be calculated directly from the recorded inforrnation.
21 However, such coincidence of the lerel~llce points is difficult to achieve in a practical 22 environment. If the accurate relative orientation data is available from the mechAnir~l 23 hardware of the gantry, steps 3236 of ~ig 1 may be employed to calculate the 3D coronary 24 arterial structures. However, it is difficult to obtain biplane transformation data based on . .
W097/4gO65 PCT~Usg7/10194 current instrumental technolo~y. A si~nificant advantage of the present inventive method is 2 that 3D reconstruction is accurately rendered from 2D projection images when such 3 orientation information is unavailable.
4 If biplane ;.n~ gc~ Ctly is not available, the bifurcation points are calculated, as S shown in step 24 of Fig. 1. An important step in the present inventive method relies on the 6 accurate establ;~l~ ,l of co,l~,~pondence beL~ ell image fcalu.~s, such as points or curve 7 se~ ; between projecli~l~s. The bifurcation points on the vascular tree are prnmin,nr 8 features and are often Iccoglli~ed in both images to f~cilit~te the delf i"~tiQn of biplane 9 im~in~ geometry. Using the hie.al~llical digraph, bifurcation points are then c~lrul~t~d by use of each pair of ~Ccen~A~.l and descPntl~nt nodes (or vessels). C~iven two sets of 2D
11 poin~s l't~l~;Sell;l)~ the re~l,c~Li~/e centerlines of vessels Concliluli~E a birulca~ion, they can 12 be modeled as two curYes~ pfr) and q(t), where O S r, t S l are the pald,l.cL~l~ based on 13 the splinebased curvefltting algoritl~n, such as is described by R.H. Bartels, J.C. Beatty, 14 and B.A. 13arsky in a publication entitled "An introduction to splines for use in computer grap~ics and geometric modeling, " Morgan Ka~fm~nn Publishers Inc., Los Altos, California, 16 as is known in the art. Let curve q(t) denote the branch of the p~ ,aly vessel as modeled ~7 by a curYe~ p(r). The birulc~tion point can then be obtained by c~lr~ tion of the il,~claeclion 18 of the tangent line denoted as a vector qO at the point q(O) and the curve of the primary 19 vessel, p(r), by ,;";.~ the objective function ~bl,7(r) as follows:
CA 022~8730 19981218 W O 97/49065 PCT~US97/10194 r ~bifl (I ) =~p (r) ~ qo~
~oqo (p ~(r)p(r))(qTqO)(q ~ (r)) 2 ( qOg~ ) 2 Equ. (F) 3 subject to 4 OSrCl.
5 The results of p(~), where r 7 satisfies Equ. (F), are the c~lr~ tPd bifurcation points which are saved into the nodes 8 associated with the branching vessels in the hierarchical digraph. On the basis of the vessel 9 hierarchy digraph, the relationships of vessel correspondence among the multiple projections 10 are established by traversal of the ~C~oci~tPd hierarchical digraphs via the tlesce~nt and 11 sibling arcs.
12 DeL~ ation of Pdlalll~t~ of the Biplane Tm~ing Svstem 13 Referring now to Figs. 1, 2A2B, 4, and step 28 of Fig. 1, the biplane im~gin~
14 system inlutles a pair of singleplane imaging systems (Figs. ~A2B). Each Xray source 15 (or focal spot) functions as the origin of 3D coordinate space and the spatial rela~ionship 16 between each im~ging portion of the biplane system can be characterized by a transformation 17 in the form of a rotation matrix R and a translation vector t . In the first projection view, 1~ let (u;, ~;) denote the image coordinates of the ith object point, located at position (xi, y" zi) , . . . .
WO 97/490~5 PCTrUS97/10194 We have ui = Dxj/z" ~, = Dyj/z" where D is the perpendicular distance between the xray 2 focal spot and the image plane. Let (~ ,) denote scaled image coordinates, defined as 3 = u;/D = x,/z;, rli = Vj/D = Yi/Zi The second projection view of the biplane imaging system 4 can be described in terms of a second pair of image and object coordinate systems u'v' and 5 x'y 'z ' defined in an analogous manner. Scaled image coordinates t~ Ij, 71 ',.) in the second view 6 (second projection image) for the ith object point at position Ixl;, ylj, z';) are given by ~
7 u',/D' = x'j/z',, t1'; = V'j/D' = y~j/Z~;. The g~ L~;cal relationship between the two views 8 can be characterized by XI i' "Xi' ' rll rl2 rl3 'XitX' Y' i =R ' Yi t ~= r2l r22 r23' Yity ~Z'i ~ Zi , r3l r32 I33 i Z
Equ. (1) 11Fig. 4 illustrates the graphical ~ ellt~tion of the biplane system defined by a 12 m~thern~tjcal model. In the inventive method, the required prior inforrnation (i.e., the 13 intrinsic ~a~ et~,~ of each singleplane i...~ g system) for ~PtPrrnin~tion of biplane 14 im~ging geometry inr.!ll~Ps: (I) the ~1ict~nre between each focal spot and its image plane, SID
15 (focal_pot to im~gingplane (li~t~nre), (2) the pixel size, p5j~ (e.g., .3 mm/pixel), (3) the 16 ~i~t~nr.e ff 1~ between the two focal spots or the known 3D distance between two points in the projection 19 images. and (4) for each view, an apploxi.llation of the factor MF (e.g., 1.2), which is the _19 ratio of the SID and the approximate distance of the object to the focal spon Item (4), 2 immt~di~t~ly above, is optional but may provide a more accurate estim~te if it is available.
3 An e~nti~l step in featurebased 3D reconstruction from two views relies on the 4 accurate establi.chm~nt of correspondence in image features, such as points or curve segments between projections, as is illustrated in step 32 of Fig. 1. The bifurcation points on the 6 vascular tree are prominent features and can often be recognized in both images to facilitate 7 the determination of biplane im~ing y~eolllclly~ T~ec~ e the vessel correspondences are 8 m~inl~in~d based on the hierarchical digraphs, the correspondences of bifurcation points are 9 h~ ly established and can be ,eLlie~d by traversing the ~soc;~l~d hierarchical digraphs (data structures). The established pairs of bifurcation points are used for the calculation of 11 the biplane im~in~ geometry. Note that the "pin~ hion distortions" on birulcation points 12 and image points are cGll~;led first before the estim~tiQn of biplane imaging geometry 13 proceeds. The co,l~clion of pinrushion error can be implPmentPd based on known 14 algorithms. For example, a method described in a publication entitled "Correction of Image Deformation from Lens Distortion Using Bezier Patches", COll~l1Lel Viusion. Graphics 16 Ima~e Processing, Vol. 47, 1989, pp. 385394, may be used, as is Icnown in the art. In the 17 present inventive method, the pinrucllion distortion does not conci~erahly affect the accuracy 18 of the 3D recor~lu.;tion due to the small field of view (i.e., 100 cm SID and 17 cm x 17 19 cm II). The prior information (SID, P5e~ MF) and the 2D inputs are employed to serve as constraints such that the intermediate solutions resulting from each iterative calculation 21 remains in the vicinity of the true solution space and converges to an "optimal" solution 22 Initial Estimates of Biplane Tm~in~ Geometry 23 When the input data error of corresponding points is moderate (e.g., less than 1 pixel 24 ~ .3 mm RMS error in coronary angiography), the estimate of the 3D imaging geometry .
. .
provided by the linear algorithm is generally sufficient to ensure proper conver~ence for 2 further optimization. Such a linear algorithm is described in a publication by C. E. Metz and 3 L. E. Fencil entitled "De~ ion of threedimensional structure in biplane radiography 4 without prior knowledge of the rel~tioll~hir ~cLween the two views: Theory," Medical S Physics, 16 (1), pp. 4551, Jan/Feb 1989, as is known in the art.
6 However, when input data error is large, the initial estim~t~ provided by the linear 7 algolilhl,l may be grossly inac._u,dle, and the m;l;",i,;1lion procedure may become trapped 8 in a local minimnm. In the problem of biplane angiography, the centroid of a target object 9 or the region of interest to be imaged is usually aligned with the jCocentpr of the ima~inC
10 system as closely as possible such that the content of projection image inr~ es the desired 11 focus of attention at any viewing angle. The i~oc~ , is the location ~t~ the focal spot 12 and image ;"~ .iri~l with respect to the rotary motion of the gantry arm, as illustrated in 13 Figs. 2A2B. It is usually llledsui~d as the relative ~ t~nre frorn the focal spot. Hence, the 14 inforrnation with respect to the isocellL~r is employed and converted to the approximate MF
15 value if the distance between the object and focai spotis not available.
16 The required initial e3l;,.. ~s include a rotation matrix 17 a UIUt ~r~n~l~tion vector tut 18 and scaled 3D points P~i (x i,Y i~Z~ 2~ n.
With large amounts of noise on the input of the 2D corresponding points e~tracted from the 2 biplane images, the estim~t~d im~in~ geometry, as well as the 3D objects by use of the 3 linear algorithrn may considerably deviate from the real solution and, therefore are not 4 suitable to serve as the initial e~ t~ for the refinement process. Such a situation can be S il1~ntifi~d if (1) not all of the c~lr~ tPd 3D points are in front of both (or all) focal spots, 6 (2) the RMS image point errors are large (e.g., > 50 pixels) or (3) the projections of the 7 c~ t~d 3D points are not in the image plane. To remedy this problem, the estim~tPs of R, tu and p, ~ s 9 must be redefin~ so that their values are in compliance with the initial biplane ~eonleL,y 10 setup for the optimi7~ti~n. Without loss of generality, the initial estim~t~s of the z and z' 11 axes of the two imaging systems are taken to be otthogonal in this situation, and the unit 12 translation vector is set to be on the xz plane. ~et (~, cx') and (D, D'), denote the MF
13 factors and SID of the biplane im~in~ systems in xyz and x'y 'z ' coordinates"ei,~Je.,Li~,ly.
14 Two dirr~.e~lL initial solutions (R~, tu ) and ~2, t",) 15 are employed as follows:
Dl Dl 0 0 1 a~ t O O 1 altd Rl= 0 1 0 , tu~= o , and R~= O 1 0, tu = O
~ ~~t atd }7 Equ.(2) , .. . . . . .
W O 97/49065 PCTAUS97/lOlg4 where t~ replesents the magnitude of t. If the magnitude of t is not available, an 2 approximated measuremen~ is calculated as follows:
td=~¦ ( D ~ 2 + ( D ~ 2 4 Referring now to Fig. 5, Fig. 5 illustrates the ~laphical representation of the predefined initial solutions. The scaled 3D points ~, y'j, z',) defined in the x'y'z' 6 coordinate system are initialized as a/t ~ Y i a/ td a' td 8 where (u 'j, v';) denotes the 2D input points on image plane in the x 'y 'z ' singleplane system.
9 Final F..c~im~t~s Based on Constrained O~ i.l;o"
Although the linear algorithm di~cc~ed above is co",l"~ lionally fast, the solution is not Il optimal in the ~se,~ce of very noisy data (e.g., RMS error > 1 pixel). Hence, itiS
12 potentially advantageous to employ another method aiming at global optimization to improve 13 the accuracy of the solution in image locations of co.l~ onding points. In the approach 14 described herein, an objective function defined as the surn of squares of the Euclidean 15 rli~l;..,~es bf,l~.en the 2D input data and the projections of the c~lçul~ted 3D data points is 16 employed. Given the set of 2D points extracted from the biplane images, an "optimal"
17 estim~te of the biplane im~ging geometry and 3D object structures is ~e obtained by 1 8 minimi7.in~
wo 97/49065 PCT/USg7/l0194 plp,  1 (P~ P ) =~ z ) 2+(~i Yi ) 2+(~li X i)2+(~, _ Y'i~2~
Equ. (3) 2 where n denotes the number of pairs of corresponding points e~Lr~el~d from the biplane 3 images, and P and P' denote the sets of 3D object position vectors pj = (xj,yj,zJ and p, =
4 (x'j,y'j,z'J, where i=l,..., n, l~s~ec~ ely. The first two terms of the objective function 5 F,(P,P') denote the square of t~iC~nre b.,.~ n the input of image data and the pro~ection of 6 c~lcul~t~d 3D data at the ith point. The last two terms are similarly defined as the square 7 of 2D ~lict~nre error in the second image plane. Since the relationship b~L~een the two 8 im~in~ systems can be ch~lack,.i~ed by a rotation matrix R and a translation vector t =
9 ~t~r~ty~tz]t~ as shown in Eq. (1), Eq. (3) can be eA~ ed as mln F2~R, t,P~ Z ) +(~ i Zl ) + (~i c p' ~+t ) + ~rli , Y~ ~' Equ. (4) 11 where c,~ denotes the ,.,s~e~ e kth column vectors of matrix R. From a pair of projectiorls, 12 the 3D objects can only be recovered up to a scale factor, at best. This fact is reflected by 13 the in~pection of each quotient term involving the 3D points in Equ. (4) as follows:
min F(R t p,~ X'~ )+~ i_Y' ~/leI) u~ ~ (C~p' i'+tX~ /It l~l2 ~ (C2~1 j+ty) /~tl~2~) (Z3Pl i+tz) /lt 1) ~ (Z3p~i+tz) /ltl)S
~2 ~ ~2 * i ) Z+ ('Tl I ~ i ) 2+ ~i P i ~ + r~ i Y
i ~1 i Z i Z3 ~3 1 i tU~ ~ ~ Z3 P I i ur J
14 Equ (5) .
where P' 2 denotes the set of scaled 3D points ~i ( x i, Y ~
4 where i = 1,..., n, to within a scale factor of the m~ninl~e of the translation vector It S and where [tu~, tu, tu 3 t 7 denotes the unit translation vector corRsponding to t.
8 It is well known that any 3D rigid motion can be uniquely decomposed into a translation 9 and a rotation by an angle 6 around an axis vu passing through the origin of the coordinate 10 system. In the present inventive method, a quaternion denoted as q=(S,W)=(S,Wl,W2,W3) 12 of norrn equal to 1 is employed to ~ sel,t the rotation transformation as w=sin (~/2)~,1, s=cos (~/2) .
14 Equ. (6) T
WO 97/49065 PCT/US97tlO194 Similarly, for any quaternion q=~s~w)=(s~wl~w2~w3) 3 of norm 1, there exists a rotation R sali~ryillg Eq. (6) and is defined as follows:
s + (wl) 2 w w sw sW2+wlw3 2 W,W2+5W3 S2+ (w2) 2  2 W2W35Wl WlW3SW2 SWl+W2w3 52~ (W3) 2  2 4 Equ. (7) S With this qualclllon ~ lc~ iQn, Eq. (4) can be l~wl.L~ as:
n~in F,(q,tOf~ 1 X~/~2 ~ I Y/ 2 2(s~wl ~l12)x~ 2(wlw sw~)j,12(w~ sw )i,lt Z~ 2(sw wlw~ 2(w2w~sw~ 2(s2w~
2~wlw2sw~ 2(sl~w22l12)j~~2(swl~w2w~
2(SW2~W~W~ 2(W2W~SWI);~ ~2(S2~W~2_ ~ t"
Equ. (8) W 097149065 PCTrUS97/10194 subject to the cons~raints:
c, S2 ,(~ +(W2)2~(W3)2=
C2 (t )2~(t )2~(~ )2=~
C3: O<~j,i=l,...,n C,: 0<2(sw2 I W~w3)~l~2(w2w3~ 1 2(s2~(w )2_ 1 )i' ~1 I
2 where constraint C, ck ~ ~ iLes the ~ua~cll~ion norm, constraint C2 ensures a unit 3 translation vector, and cor,~ C3 and C4 force the scaled coordinates z'; and z; to be in 4 front of the lcs~ecLive focal spots.
If the isocenter .l;~Ai~re~ of employed biplane im~g~ng systems or MF factors are 6 available, the constraints C3 and C4 in Eq.(8) can be modified as:
d '~h d ' ~h C~ d ~h<zj=2[(sw2~wlw3)~l+(w2w3swl)g'j+(s2+(w3)2  2)~d+t" ' 1 _ i=l,...,n, 8 where d = D/o~ and d'=D'/o~' are the apyluxilllate distances between the object's centroid 9 and the ~s~ec~ive focal spots, ~ and ~' denote the MF factors and ~h ( Z 12.5 i 2.0 cm) denotes the maximal length of the heart along a longaxis view at enddiastole, as is known 11 and described by A.E. Weyman in a publication entitled "CrossSectional 12 Echocardiography. " Lea & Febiger, Philadelphia, 1982. For each 3D object point, the ray 13 connectinsg that point and the focal spot intersects the image plane near the associated 2D
14 image point, even when the input data is corrupted by noise. In addition to the constraints imposed on the z and z' coordinates, two other constrains are incorporated to confine the x, 2 x', y, and y' coordinates of each calculated 3D point as follows 3 For each 3D object point, the ray co~ that point and the focal spot i~,~e.~
4 the image plane near the associated 2D image point, even when the input data are corrupted by noise. In ~lition to the cor~ imposed on the z and z' coordiantes, two other 6 constraints are iluolyol~l~d to confine the x, x ', y, and y ' coordinates of each c~lrul~tPd 3D
7 point as follows:
C5 ( ~ )2+( ~ .)2S( D ' ~2, i=1,...,n, X~ )2 (Yi ~l,)2~( ~5~ )2, i=1,...,n, 9 where ~c defines the radius of a circular disk (e.g., 20 pixels) centered at (~ ;) or (~';, ~7';) and psize re~l.,sel~ls the pixel size.
11 Referring now to Fig. 6, Fig. 6 shows the bounding regions based on the employed 12 constraint C3 to C6 in x 'y 'z' system. If two initial solutions are employed (as described under 13 the s~lbhP~rling of Initial F~l;~l~s of Biplane Tm~gin~ Geometry), in general, two sets of 14 biplane im~in~ geolllehy and their ~Coci~tPd 3D scaled object points will be obtained:
IRl,t"~ .y ~ =1,...nJ
17 and .
~ ,tU,(~2,y~2~2i)~ i=l,...n].
2 Rer. ~ g now to Fig. 7, Fig. 7 illustrates a typical example by use of several object 3 point RMS errors on image points a~oc;~l~d with the true solutions defined by one imaging 4 g~O~ !  y (e.g.,RI and tu,) 6 is smaller than those defined by the other im~gin~ y (e.g.,~2 and t",).
8 Therefore, the c~lc~ t~d im~gin~ parameters, which have a smaller RMS error on the image 9 points, are selected as the optimal solution. To d~te.lllhlc the absolute size of the object, the 10 ~ illJ~o of the translation vector (i.e., the (lict~nr,e ~el~ the two focal spots ff',) 12 or the real 3D distance between any two object points projected onto the biplan~ images 13 needs to be known. In the forrner case, the actual 3D object points can be recovered easily 14 by multiplying the scaled object points by the m~nit~l~e. Othervvise, the scale factor Sf is 15 calculated and employed to obtain the absolute 3D object point as wo 97/49065 PCT/USg7/10194 Xi Xj y;=Sf ~ Y;, where Sl Ld Zi Zi ~(XPIxp2) +~PIyP2)~+(zP _Zp)2 3 and Ld denotes the known 3D (~ict~nre ~oci~ed with the tw~o scaled 3D object points (xp,,ypt~zpl) and ($~,Yp2,zp~).
Recove~y of 3D Spatial Il,~o,.ation 6 After the biplane im~ing geo",cll~ that defines the two views is obtained, the 7 orientation i,~llllaLion is used to establish the point co"~ ondences on vessel cel,le~ .es 8 in ~e pair of images and is further used to c~lrui~ 3D morphologic structures of co,u~
9 arterial tree, as is illll~trat~d in step 34 of Fig. 1. The c~lr~ t~d i."~ g geol"elly in conj~l"~;Lion with the epipolar constraints are employed as the framework for establishing t'ne 11 point coll~ ol,~t ~res on the vessel c~ s based on the two identifi~d 2D coronary 12 arterial trees.
13 According to the epipolar con~L~ ;, the co.. ~s.po.~en~e of a point in one image 14 must lie on the epipolar line in the other image. Two types of ambiguity may arise in the 15 twoview correspondence problem: (1) the epipolar line may hlL~I~.ect more that one vessel 16 in the coronary arterial tree, and (2) the epipolar line may intersect more than one point on 17 a single vessel. The first ambiguity is resolved by means of the co..~.~u~L~d hierarchical 18 digraph defining the anatomy of the 2D co,onaly arterial tree such that the epipolar 19 constraints are applied iteratively to every pair of corresponding vessels in the two coronary ..... .. ....... ..
Wo 97/49065 PCT/US97/10194 trees. For example, the corresponding centerline points of the left anterior desce~ing artery 2 in the angiogram acquired from the first view is uniquely deterrnine by finding the 3 hlLe.secLions of the epipolar line and the 2D centerline of the left anterior ~escen~;"g artery 4 in the angiogram acquired from the second view.
S When the i"t~.~e~ n point is c~lr~ tPcl~ each 2I) vessel centerline is modeled by 6 a splinebased curvefitting functionf(s) = (x;, yj), 0 ~ s ~ 1 (the same method used for 7 c~lrul~tion of birul~dLion points described above) where s is the paldllleLl ic argument defining 8 the location of points (xj, YiJ on the vessel centerline. If there are n ;~ ecliQn points 9 beL~en the epipolar line and the vessel centerline due to the tortuous vessel shape, the 10 locations of these points can be defined based on the pa~ eLlic a~ Luen~s (e.g., f(s,), 11 f(s2),.. f(sn), s, = 0.2, 52 = 0.35,, Sn = 0.5). The point with the ~a1;1111CtliC argument Sk, 12 1 5 k ~ n is sel~ct~d as the desired co~ onding point if s~ is the cm~llçst value larger 13 than the Pal~ll~ iC a~ t of the last ~etecf~d correspondi~g point. Based on such a 14 method, the second type of ambiguity is resolved.
With the point correspon~ienreS on 2D vessel c~ t,.lh~es (~ 1.) and (~ 1',) and the 16 im~ging geometry, the 3D CeJI~ points of coron~,y arteries (x;, y;, zj)'s can then be 17 e~lr~ trd based on the following eq~tionc rllr3l~'~rl2r32~i r.3r33~i  a t r2~r3~r22r3211 i r23r33~ i ~, bt ~ ~i ~
O 1 t1i  . O
18 Equ. (9) Wo 97/49065 PCT/USg7/10194 where a and b are two vectors defined as follows:
(r~lr3l~ i) (r2lr31rl'i) a (rl2r32~ i) . b= (r22(rl3r33~',) (r23r33~
2 Equ. (10) 3 and r,j denotes the colllpori~lll of the rotation matrix R.
4 Renderin~ of Recoll~LIlcted 3D Coronary Tree and Estimation of an Optimal View After the 3D vessel cGlll~.lilles are obtained which define the 3D location of the 6 arterial tree, as shown in step 34 of Fig. 1, the anatomical morphology of the arterial tree 7 is generated by a surface based lc,~loduclion techni~ue, as illustrated in step 36 of Fig. 1, 8 as is known in the art. Such a surface based reproduction technique is described by S.Y.
9 Chen, K.R. Hoffmann, C.T. Chen, and J.D. Carroll in a publication entitled "Modeling the Human Heart based on Cardiac Tomography," SPIE, vol. 1778, 1992, pp. 1418.
11 The 3D lumen surface is leylcsentEd by a se~uellce of crosssectional contours.
12 Each contour V; along the vessel is ~e~)~sellted by a djmm circular disk centered at and 13 perpe~dir~ r to the 3D vesselce~ lille. The surface be~ e.l each pair of cnn~ecltive 14 contours Vj and Vj+, is ~elle~aL~d based upon a number of polygonal patches. Utilizing the modeled lumen surfaces, the morphology of the r~coll~ cted COlullalr arterial tree is 16 reproduced by employing the technique of compulel graphics, as is known in the art.
17 When an arbitrary computergenerated image is produced, the gantry inforrnation 18 defining the current projection is c~lcul~ted in the form of LAO/RAO (on the yz plane) and 19 CAUD/CRAN (on the xz plane) angles by which the gantry arm moves along the LAO/RAO
. .
W O 97t49065 PCTAUS97/lOlg4 angle followed by the CAUD/CRAN angle. The focal spot of the gantry can be formulated 2 as 0 0 cos(O O sin(,B) Rx(y)Ry(~J= O cos(y) sin(y) O 1 0 O sin(y) cos(y) sin(O O cos(,B) ' cos(a) O Sin(a) ' = sin(y)sin(aJ cos(y) sin(y)cos(,B), cosf y)sinf a) sin(y) cos(y)cos( a).
3 Equ. (11) 4 where R,~ and Ry denote the rigid rolaliolls with respect to the xaxis and yaxis, ~ cc~ ely, S and where y and ~ denote the LAO and CAUD angles, ~~ec~ ly.
6 Let pj, i = O, 1,.. ,m denote the points on the c~ of a 3D vessel. Let l~=[ljS,13~,1jS]' ~ld Ij, J=1,2,...,m 8 denote the vector and length of the se~.~le~ be~weenpj~ andp~ cli~/ely. The Inin~m~l 9 foreshortening of the vessel seg..~ are o~l~ined in te~ns of the gantry oli~n~ io~ y and 10 ,B angles) by mini.~ the objective function as follows:
min F(pl,y"B)=~1/l,cos(~.)//2 Y.~ m ~ (l ,2 12 Equ. (12) W O 97/49065 PCTrUS97/10194 subject to the constraints 2 goo ~ ~y < goo, 40~ < ,B ~ 40~, 3 where ". " denotes the inner product and ~j is the angle between the directional vector I j and 4 ~e projection vector zp is defined as cos(y)sint~) zp= sin( y) cos(y)cos( ,B) 6 Egu. (13) 7 In prior art nl~thn(ls, due to the problem of vessel overlap and vessel foreshorterung, 8 mllltirl~ pluje.,lions are i~eC~c~y to adequately evaluate the coronary arterial tree using 9 arteriography. Hence, the patient may receive additional or unn~eded radiation and contrast material during ~ gnl)stir and interventional procedures. This known tra~lition~ trial and 11 error method may provide views in which overlapping and foreshortening are somewhat 12 minimi7Pd, but only in terrns of the ~llbje~Live experiencebased judgement of the 13 angio~,lapher. In the present inventive mPthnd, the r~col~,u.;led 3D colonaly arterial tree 14 can be rotated to any sel~ct~d viewing angle yielding multiple cull,~ulclg.,n~,at~,d ~fùj~
to (leL~ r for each patient which ~L~ da~l views are useful and which are of no clinical 16 value due to excessive overlap. The.cfo~c, the 3D CU111~ULel a.5ci~t~nre provides a means 17 to improve the quality and utility of the images subsequently acquired.
18 Expelill,c.lL~l Results 19 The accuracy of the present inventive method was evaluated by use of bifurcation points in a computersimt~ Pd coronary arterial tree. For a~secs,l.ell~ of the rotation matrix, R is 21 further decomposed into a rotation around an axis vu (a unit vector) passing through the .
origin of the coordinate system with the angle ~. The differences between the calculated and 2 real rotation axes ~ and rotation angles Ee are employed for error analysis. The error in 3 the translation vector E, is the angle between the real and e~lcul~tPd translation vectors. The 4 error in the 3D absolute position ~3~tiS defined as the RMS ~i~t~nre between the c~lrul~t~d 5 and the real 3D data sets; while the error in the 3D configuration 6 is defined as the RMS dist~nre between the r~lr~ te~ and the real 3D data sets after the 7 centroids of these two data sets have been made to coincide. In simlll~ted expe~ eL~, the 8 parameters of the biplane im~ing geolllell~ were varied to investigate the effects of the 9 system geometry on the ae~;ulaey with which the 3D point positions could be recovered.
Both D and D ' were equal to 100 cm.
11 To assess the reliability of the technique under realistic conditions, a set of 12 ~ lents was ~im~ tlod by adding indel,ende"~ errors to the 2D vessel centerlines 13 resulting from the projection of the ~imlll~ted 3D arterial tree. The effect of the relative 14 angle between the biplane im~ging views was ac~ec~ed by varying ~ from 30~ to 150~. By use of the colllyu~el sim~ t~d colonaly arterial tree, RMS errors in angles defir~ing the R
16 matrix and t vector were less than O.S (Ev), 1.2 (E~), and 0.7 (E,) degrees, lc~yecti~ly, 17 when ten corresponding points were used with RMS normally distributed errors varying from 18 0.7  4.2 pixels (0.21  1.32 mm) in fifty configurations; when only the linear based 19 Met~Fencil method was employed, the respective errors varied from 0.5  8.0 degrees, 6.0  40.0 degrees, and 3.7  34.1 degrees. The simulation shows subst~nti~l improvement in 21 the estimation of biplane imaging geometry based on the new technique, which facilitates accurate reconstruction of 3D coronary arterial structures. The RMS errors in 3D absolute 2 position (E3d) and configuration (E3d) 3 of the l~consLllcted arterial tree were 0.9  5.5 mm and 0.7  1.0 rnm, Itsl.c~Liv~ly. The 4 following table shows one of the simlllAtion results based on an orthogonal biplane setup:
Erro~ inim~ging~ ~ S e~or in3D
6 R~S error tn2D
7centerlines ~el) Ev E~ E, ~hill E~ E~
8 0.5 0.30~ 0.61~ 0.11~ 1.8n~n 1.08 ~ O.9Q~n 9 1.0 0.33~ 0.72~ 0.40~ 6.4 ~ 1.37 ~ 3.7 0 1.5 0.49~ 1.19~ 0.64~ }0.2nun1.56 ~ 5.5n~n 2.0 0.40~ 0.65~ 0.~~ 6.5 ~ 1.79mm 3.6 ~
2 2.5 0.42~ 0.95~ 0.40~ 6.4 ~ 0.74mm 2.8mm 3 3.0 0.42~ 0.86~ 0.40~ 6.5 ~ 0.96n~n 2.~nun 15where 16 denotes the deviation angle between the true v and the c~lcul~tP.d v ' rotational axes and E~
17 denotes the angle dirr~,c. ce be~,en the t~ue ~ and the cAlrulA~.ed ~' rotational angles.
18 Note that the 3D absolute position error is due primarily to displ~r.f.. "~ t error E5h,fl =
19 (Drr ~E,) that results from inaccurate e~l;...A~;un of the tran.cl~tion vector, where Dfr is the 20 (~ re be~weell the focal spots of two imaging systems and ~ denotes the deviation angle 21 between the real and c~lcul~tPd translation vectors. The RMS error in the 3D configuration (E3d) 22 decreases due to the reduction of the displAr~..r ~ error after the centroids of the real and 23 calculated data are made to coincide. In general, the results show a great similarity between , .. . .
W097/49065 PCT~US97/10194 the reconstructed and the real 3D vessel centerlines. The simulation shows highly accurate 2 results in the estimation of biplane imaging geu~etry, vessel correspondences (less than 2 mm 3 RMS error), and 3D coronary arterial sllucLul~5 (less than 2 mm ~MS error in configuration 4 and 0.5 cm RMS error in absolute position, ~ .u;Lively) when a col,,yulersim~ t~d coronary 5 arterial tree is used.
6 Angiograms of fifteen patients were analyzed where each patient had multiple biplan 7 irnage ac~lu;c;lloll~. The biplane im~in~ geoll,etLy was first tlele~ d without the need of 8 a calibration object, and the 3D coronary arterial trees inrl~ldin~ the left and the right coronary 9 artery sys~ems were lecollaLIu~Led. Similarity b~ e,l the real and reconsllucled arterial 10 ~Llu~;lul~s was excellent.
1 1 Conclusions 12 The present inventive method is novel in several ways: (1) the 3D coronary v~c.~ lre 13 is lecorsLlu~ d from a pair of projection angiograms based on a biplane im~ing system or 14 multiple pairs of angiograms acquired from a singleplane system in the sarne phase of the 15 cardiac cycle at different viewing angles without use of a calibration object to achieve 16 ac~ ies in ~ A~ion and Ima~in~ ~O~ of better than 2% and three degrees, 17 I~,~pec~ ly; (2) a beating 3D COIOllaly v~rul~hlre can be reproduced throughout the cardiac 18 cycles in the temporal se.luellces of images to f~cilit~t~ the study of heart movement; (3) the 19 choice of an optirnal view of the v~ccul~1llre of interest can be achieved on the basis of the 20 capability of rotating the ~consLI~lcted 3D coronary arterial tree; and (4) the inventive method 21 can be implemented on most digital singleplane or biplane systems. A calculated 3D coronary 22 tree for each patient predicts which projections are clinically useful thus providing an optimal 23 vi~ li7.ation strategy which leads to more efficient and successful dia~nostic and thel~p~ ic WO 97/49065 PCTtUS97/10194 procedures. The elimination of coronary artery views with excessive overlap may reduce 2 contrast and radiation.
3 Note that the present inventive method is not limited to Xray based im~ging systems 4 For example, suitable ;l"~;ng systems may include particlebeam im~ging systems, radar im~in~ systems, ultrasound im~ging systems, photographic im~ging systems, and laser im~ging 6 systems. Such im~ing systems are suitable when pe.~.~c~ eprojection images of the target 7 object are provided by the systems.
8 Please refer to Appendix A for a source code listing of the abovedescribed method.
9 The software is written in C Progr~mmin~ T ~n~ge inr~ in~ GL Graphics Library Functions 10 and Tk. Tcl Library filnrtionc compiled on a Unixbased C Compiler.
11 Specific embo(li~ of a method and apparatus for threedim~ncional reconstruction 12 of colonaly vessels from angiographic images according to the present invention have been 13 described for the purpose of illusl~atillg the maMer in which the invention may be made and 14 used. It should be understood that impl~"~~r~ t;on of otner variations and modifi~tions of the }S invention and its various aspects will be appale,~t to those skilled in the art, and that the 16 invention is not limited by the specific embo~im~nt~s] described. It is ~ eÇole contemplated 17 to cover by the present invention any and all mo~ifir~tinnc, variations, or equivalents that fall 18 within the true spirit and scope of the basic w~dellying principles tlicclosed and claimed herein.
Claims (20)
a) placing the target object in a position to be scanned by an imaging system, said imaging system having a plurality of imaging portions;
b) acquiring a plurality of projection images of the target object, each imaging portion providing a projection image of the target object, each imaging portion disposed at an unknown orientation relative to the other imaging portions;
c) identifying twodimensional vessel centerlines for a predetermined number of the vessels in each of the projection images;
d) creating a vessel hierarchy data structure for each projection image, said data structure including the identified twodimensional vessel centerlines defined by a plurality of data points in the vessel hierarchy data structure;
e) calculating a predetermined number of bifurcation points for each projection image by traversing the corresponding vessel hierarchy data structure, said bifurcation points defined by intersections of the twodimensional vessel centerlines;
f) determining a formulation in the form of a rotation matrix and a translation vector utilizing the bifurcation points corresponding to each of the projections images, said rotation matrix and said translation vector representing imaging parameters corresponding to the orientation of each imaging portion relative to the other imaging portions of the imaging system;
g) utilizing the data points and the transformation to establish a correspondence between the twodimensional vessel centerlines corresponding to each of the projection images such that each data point corresponding to one projection image is linked to a data point corresponding to the other projection images, said linked data points representing an identical location in the vessel of the target object;
h) calculating threedimensional vessel centerlines utilizing the twodimensional vessel centerlines and the correspondence between the data points of the twodimensional vessel centerlines; and i) reconstructing a threedimensional visual representation of the target object based on the threedimensional vessel centerlines and diameters of each vessel estimated along the threedimensional centerline of each vessel.
a) providing an optimal threedimensional visual representation of the target object such that vessel overlap and vessel foreshortening are minimized in the visual representation by rotating the threedimensional visual representation in at least one of three dimensions;
b) calculating image parameters corresponding to the rotated threedimensional visual representation; and c) providing said calculated parameters to the imaging system to permit the imaging system to further scan the target object such that optimal projection images of the target object are produced.
a) placing the target object in a position to be scanned by a biplane imaging system, said biplane imaging system having first and second imaging portions;
b) acquiring biplane projection images of the target object, each imaging portion providing a biplane projection image of the target object, each imaging portion disposed at an unknown orientation relative to the other imaging portion;
c) identifying twodimensional vessel centerlines for a predetermined number of the vessels in each of the biplane projection images;
d) creating a vessel hierarchy data structure for each biplane projection image, said data structure including the identified twodimensional vessel centerlines defined by a plurality of data points in the vessel hierarchy data structure;
e) calculating a predetermined number of bifurcation points for each biplane projection image by traversing the corresponding vessel hierarchy data structure, said bifurcation points defined by intersections of the twodimensional vessel centerlines;
f) determining a transformation in the form of a rotation matrix and a translation vector utilizing the bifurcation points corresponding to each of the biplane projections images, said rotation matrix and said translation vector representing biplane imaging parameters corresponding to the orientation of each imaging portion relative to the other imaging portion of the biplane imaging system;
g) utilizing the data points and the transformation to establish a correspondence between the twodimensional vessel centerlines corresponding to each of the biplane projection images such that each data point corresponding to one biplane projection image is linked to a data point corresponding to the other biplane projection image, said linked data points representing an identical location in the vessel of the target object;
h) calculating threedimensional vessel centerlines utilizing the twodimensional vessel centerlines and the correspondence between the data points of the twodimensional vessel centerlines; and i) reconstructing a threedimensional visual representation of the target object based on the threedimensional vessel centerlines and diameters of each vessel estimated along the threedimensional centerline of each vessel.
a) providing an optimal threedimensional visual representation of the target object such that vessel overlap and vessel foreshortening are minimized in the visual representation by rotating the threedimensional visual representation in at least one of three dimensions;
b) calculating image parameters corresponding to the rotated threedimensional visual representation; and c) providing said calculated parameters to the biplane imaging system to permit the imaging system to further scan the target object such that optimal projection images of the target object are produced.
a) placing the target object in a position to be scanned by a singleplane system, said imaging system having one imaging portion;
b) acquiring projection images of the target object, the imaging portion providing a plurality of projection images of the target object produced at different times, each projection image produced during an identical phase of a cardiac cycle of the target object, the imaging portion corresponding to one of the plurality of projection images disposed at an unknown orientation relative to the imaging portion corresponding to the others of the plurality of projection images;
c) identifying twodimensional vessel centerlines for a predetermined number of the vessels in each of the projection images;
d) creating a vessel hierarchy data structure for each projection image, said data structure including the identified twodimensional vessel centerlines defined by a plurality of data points in the vessel hierarchy data structure;
e) calculating a predetermined number of bifurcation points for each projection image by traversing the corresponding vessel hierarchy data structure, said bifurcation points defined by intersections of the twodimensional vessel centerlines;
f) determining a transformation in the form of a rotation matrix and a translation vector utilizing the bifurcation points corresponding to each of the projections images, said rotation matrix and said translation vector representing parameters corresponding to the relative orientations of the imaging portion corresponding to the plurality of projection images;
g) utilizing the data points and the information to establish a correspondence between the twodimensional vessel centerlines corresponding to each of the projection images such that each data point corresponding to one projection image is linked to a data point corresponding to the other projection images, said linked data points representing an identical location in the vessel of the target object;
h) calculating threedimensional vessel centerlines utilizing the twodimensional vessel centerlines and the correspondence between the data points of the twodimensional vessel centerlines; and i) reconstructing a threedimensional visual representation of the target object based on the threedimensional vessel centerlines and diameters of each vessel estimated along the threedimensional centerline of each vessel.
Priority Applications (3)
Application Number  Priority Date  Filing Date  Title 

US08/665,836  19960619  
US08665836 US6047080A (en)  19960619  19960619  Method and apparatus for threedimensional reconstruction of coronary vessels from angiographic images 
PCT/US1997/010194 WO1997049065A1 (en)  19960619  19970617  Method and apparatus for threedimensional reconstruction of coronary vessels from angiographic images 
Publications (1)
Publication Number  Publication Date 

CA2258730A1 true true CA2258730A1 (en)  19971224 
Family
ID=24671758
Family Applications (1)
Application Number  Title  Priority Date  Filing Date 

CA 2258730 Abandoned CA2258730A1 (en)  19960619  19970617  Method and apparatus for threedimensional reconstruction of coronary vessels from angiographic images 
Country Status (4)
Country  Link 

US (2)  US6047080A (en) 
EP (1)  EP0978101A1 (en) 
CA (1)  CA2258730A1 (en) 
WO (1)  WO1997049065A1 (en) 
Families Citing this family (206)
Publication number  Priority date  Publication date  Assignee  Title 

US6236743B1 (en) *  19950915  20010522  Greg Pratt  Threedimensional digitizing system and method 
DE19648935B4 (en) *  19961126  20080515  IMEDOS Intelligente Optische Systeme der Medizin und Messtechnik GmbH  Apparatus and method for examining vessels 
EP0880108A1 (en) *  19970523  19981125  Philips Electronics N.V.  Image processing method including a chaining step and medical imaging apparatus including means for carrying out this method 
JPH11155881A (en) *  19970926  19990615  Olympus Optical Co Ltd  Operative path retrieval device 
JP4698780B2 (en) *  19980915  20110608  シーメンス アクチエンゲゼルシヤフトＳｉｅｍｅｎｓ Ａｋｔｉｅｎｇｅｓｅｌｌｓｃｈａｆｔ  Image reconstruction method and measurement data acquisition method 
US6226636B1 (en) *  19981120  20010501  Philips Electronics North America Corp.  System for retrieving images using a database 
US6721759B1 (en) *  19981224  20040413  Sony Corporation  Techniques for spatial representation of data and browsing based on similarity 
US6385332B1 (en) *  19990219  20020507  The John P. Roberts Research Institute  Automated segmentation method for 3dimensional ultrasound 
US6793496B2 (en) *  19990415  20040921  General Electric Company  Mathematical model and a method and apparatus for utilizing the model 
US8442618B2 (en) *  19990518  20130514  Mediguide Ltd.  Method and system for delivering a medical device to a selected position within a lumen 
US7343195B2 (en) *  19990518  20080311  Mediguide Ltd.  Method and apparatus for real time quantitative threedimensional image reconstruction of a moving organ and intrabody navigation 
US7840252B2 (en)  19990518  20101123  MediGuide, Ltd.  Method and system for determining a three dimensional representation of a tubular organ 
US7386339B2 (en) *  19990518  20080610  Mediguide Ltd.  Medical imaging and navigation system 
US9833167B2 (en) *  19990518  20171205  Mediguide Ltd.  Method and system for superimposing virtual anatomical landmarks on an image 
US7778688B2 (en)  19990518  20100817  MediGuide, Ltd.  System and method for delivering a stent to a selected position within a lumen 
US9572519B2 (en) *  19990518  20170221  Mediguide Ltd.  Method and apparatus for invasive device tracking using organ timing signal generated from MPS sensors 
US6718054B1 (en) *  19990623  20040406  Massachusetts Institute Of Technology  MRA segmentation using active contour models 
US6389417B1 (en) *  19990629  20020514  Samsung Electronics Co., Ltd.  Method and apparatus for searching a digital image 
US6385476B1 (en)  19990921  20020507  Biosense, Inc.  Method and apparatus for intracardially surveying a condition of a chamber of a heart 
FR2799029B1 (en) *  19990924  20020524  Ge Medical Syst Sa  Method for reconstructing a threedimensional image of a moving object, in particular a threedimensional image of the human heart vessels 
FR2799031B1 (en) *  19990924  20020104  Ge Medical Syst Sa  Method for reconstructing a section by cross example, of an element of interest contained in an object, in particular a vessel of the human heart 
US6546271B1 (en)  19991001  20030408  Bioscience, Inc.  Vascular reconstruction 
US7151996B2 (en) *  20000414  20061219  Mobileye Technologies Limited  System and method for generating a model of the path of a roadway from an image recorded by a camera 
US7321677B2 (en) *  20000509  20080122  Paieon Inc.  System and method for threedimensional reconstruction of an artery 
US6381302B1 (en) *  20000705  20020430  Canon Kabushiki Kaisha  Computer assisted 2D adjustment of stereo Xray images 
US6980675B2 (en) *  20001018  20051227  Paieon, Inc.  Method for processing images of coronary arteries 
US7778685B2 (en) *  20001018  20100817  Paieon Inc.  Method and system for positioning a device in a tubular organ 
US6754376B1 (en) *  20001122  20040622  General Electric Company  Method for automatic segmentation of medical images 
US6829379B1 (en) *  20001127  20041207  Ge Medical Systems Global Technology Company, Llc  Methods and apparatus to assist and facilitate vessel analysis 
US6473488B2 (en)  20001220  20021029  Cedara Software Corp.  Three dimensional image reconstruction from single plane Xray fluorograms 
US6470070B2 (en)  20001220  20021022  Cedara Software Corp.  Image reconstruction using multiple Xray projections 
US20040091143A1 (en) *  20010122  20040513  Qingmao Hu  Two and three dimensional skeletonization 
US7697744B2 (en) *  20010306  20100413  Kabushiki Kaisha Toshiba  Xray diagnostic apparatus and image processor 
US6944322B2 (en)  20010328  20050913  Visiongate, Inc.  Optical tomography of small objects using parallel ray illumination and postspecimen optical magnification 
US6741730B2 (en)  20010810  20040525  Visiongate, Inc.  Method and apparatus for threedimensional imaging in the fourier domain 
US7907765B2 (en)  20010328  20110315  University Of Washington  Focal plane tracking for optical microtomography 
FR2823057B1 (en)  20010328  20030704  Ge Med Sys Global Tech Co Llc  Process for the magnification factor of determination of a vascular Xray image including 
US20060023219A1 (en) *  20010328  20060202  Meyer Michael G  Optical tomography of small objects using parallel ray illumination and postspecimen optical magnification 
JP4347571B2 (en) *  20010402  20091021  コーニンクレッカ フィリップス エレクトロニクス エヌ ヴィ  Model of the heart using a template 
DE10122875C1 (en) *  20010511  20030213  Siemens Ag  Combined 3D angiovolume reconstruction process 
US7217266B2 (en) *  20010530  20070515  Anderson R Rox  Apparatus and method for laser treatment with spectroscopic feedback 
WO2002097735A8 (en) *  20010531  20031218  Kent Ridge Digital Labs  System and method of anatomical modeling 
US7251352B2 (en) *  20010816  20070731  Siemens Corporate Research, Inc.  Marking 3D locations from ultrasound images 
US7079132B2 (en) *  20010816  20060718  Siemens Corporate Reseach Inc.  System and method for threedimensional (3D) reconstruction from ultrasound images 
US6754522B2 (en)  20010905  20040622  Medimag C.V.I., Inc.  Imaging methods and apparatus particularly useful for two and threedimensional angiography 
US7372983B2 (en) *  20011016  20080513  Koninklijke Philips Electronics N.V.  Method for automatic branch labeling 
DE60237719D1 (en) *  20011016  20101028  Koninkl Philips Electronics Nv  A process for the selection of branches for the alignment of probes 
FR2831698A1 (en) *  20011030  20030502  Koninkl Philips Electronics Nv  medical imaging station has a function of path extraction within a branched object 
US20030160786A1 (en) *  20020228  20030828  Johnson Richard K.  Automatic determination of borders of body structures 
JP2005518917A (en) *  20020307  20050630  メドトロニック・インコーポレーテッド  Computer implementation type of a software system for dynamic threedimensional reconstruction of a nonlinear implantable medical device 
US7260253B2 (en) *  20020419  20070821  Visiongate, Inc.  Method for correction of relative objectdetector motion between successive views 
US20050085708A1 (en) *  20020419  20050421  University Of Washington  System and method for preparation of cells for 3D image acquisition 
US7811825B2 (en) *  20020419  20101012  University Of Washington  System and method for processing specimens and images for optical tomography 
US7738945B2 (en) *  20020419  20100615  University Of Washington  Method and apparatus for pseudoprojection formation for optical tomography 
US6697508B2 (en) *  20020510  20040224  Visiongate, Inc.  Tomographic reconstruction of small objects using a priori knowledge 
US6770893B2 (en) *  20020513  20040803  Visiongate, Inc.  Method and apparatus for emission computed tomography using temporal signatures 
JP2005528157A (en) *  20020604  20050922  コーニンクレッカ フィリップス エレクトロニクス エヌ ヴィＫｏｎｉｎｋｌｉｊｋｅ Ｐｈｉｌｉｐｓ Ｅｌｅｃｔｒｏｎｉｃｓ Ｎ．Ｖ．  Hybrid 3d reconstruction of the coronary artery structure based on the rotation angiography 
JP4444100B2 (en) *  20020605  20100331  コーニンクレッカ フィリップス エレクトロニクス エヌ ヴィ  Analysis method of multidimensional structure 
US7020510B2 (en) *  20020725  20060328  Koninklijke Philips Electronics, N.V.  Optimal view map V.0.01 
US7113623B2 (en) *  20021008  20060926  The Regents Of The University Of Colorado  Methods and systems for display and analysis of moving arterial tree structures 
DE10247832A1 (en) *  20021014  20040422  Philips Intellectual Property & Standards Gmbh  Producing 4D image data records of moving tubular structure involves recording 3D central line points projected into second 2D images to obtain 2D central line points, iterating for more phases 
US7289841B2 (en) *  20021025  20071030  Koninklijke Philips Electronics N.V.  Method and apparatus for volumetric cardiac computed tomography imaging 
EP1588325B1 (en) *  20030115  20061011  Philips Electronics N.V.  Image processing method for automatic adaptation of 3d deformable model onto a subtantially tubular surface of a 3d object 
JP2006516440A (en) *  20030131  20060706  コーニンクレッカ フィリップス エレクトロニクス エヌ ヴィＫｏｎｉｎｋｌｉｊｋｅ Ｐｈｉｌｉｐｓ Ｅｌｅｃｔｒｏｎｉｃｓ Ｎ．Ｖ．  Reconstruction method of threedimensional object 
US7574026B2 (en) *  20030212  20090811  Koninklijke Philips Electronics N.V.  Method for the 3d modeling of a tubular structure 
US7616818B2 (en) *  20030219  20091110  Agfa Healthcare  Method of determining the orientation of an image 
US7426256B2 (en) *  20030314  20080916  Koninklijke Philips Electronics N.V.  Motioncorrected threedimensional volume imaging method 
US7639855B2 (en) *  20030402  20091229  Ziosoft, Inc.  Medical image processing apparatus, and medical image processing method 
US7978887B2 (en)  20030617  20110712  Brown University  Methods and apparatus for identifying subject matter in view data 
DE602004032540D1 (en) *  20030617  20110616  Univ Brown  Method and apparatus for detection of a modelbased structure in projection data 
CN100468458C (en) *  20030716  20090311  皇家飞利浦电子股份有限公司  Reconstruction of an image of a moving object from volumetric data 
US7687167B2 (en) *  20030718  20100330  Panasonic Corporation  Power supply unit 
US7587074B2 (en) *  20030721  20090908  Paieon Inc.  Method and system for identifying optimal image within a series of images that depict a moving organ 
WO2005011500A1 (en)  20030804  20050210  Koninklijke Philips Electronics N.V.  Mapping the coronary arteries on a sphere 
DE602004031656D1 (en) *  20030821  20110414  Philips Intellectual Property  Apparatus and method for generating a threedimensional vascular model 
US7742629B2 (en) *  20030925  20100622  Paieon Inc.  System and method for threedimensional reconstruction of a tubular organ 
US20050074150A1 (en) *  20031003  20050407  Andrew Bruss  Systems and methods for emulating an angiogram using threedimensional image data 
US20050093859A1 (en) *  20031104  20050505  Siemens Medical Solutions Usa, Inc.  Viewing direction dependent acquisition or processing for 3D ultrasound imaging 
US20050110791A1 (en) *  20031126  20050526  Prabhu Krishnamoorthy  Systems and methods for segmenting and displaying tubular vessels in volumetric imaging data 
US20070201737A1 (en) *  20031126  20070830  Wenli Cai  System And Method For Vascular Visualization Using Planar Reformation Of Vascular Central Axis Surface With Biconvex Slab 
WO2005055496A3 (en) *  20031126  20050804  Wenli Cai  System and method for optimization of vessel centerlines 
JP4613172B2 (en) *  20031128  20110112  コーニンクレッカ フィリップス エレクトロニクス エヌ ヴィ  3D reconstruction method and apparatus object from the projection photographing image 
US7813549B2 (en) *  20031208  20101012  Koninklijke Philips Electronics N.V.  Image segmentation in a volume data set 
DE102004011154B3 (en)  20040308  20051124  Siemens Ag  A method for registering a sequence of 2D image data of a cavity organ with 3D image data of the organ cavity 
KR100695142B1 (en) *  20040308  20070314  삼성전자주식회사  Adaptive nth power of 2 nary tree generating method, and method and apparatus for encoding and decoding 3D volume data using it 
DE102004012057A1 (en) *  20040311  20050929  Siemens Ag  Xray examination apparatus and method for calibrating 
US8010175B2 (en) *  20040505  20110830  Siemens Medical Solutions Usa, Inc.  Patientspecific coronary territory mapping 
WO2005114575A1 (en) *  20040518  20051201  Koninklijke Philips Electronics N.V.  Image processing system for automatic segmentation of a 3d treelike tubular surface of an object, using 3d deformable mesh models 
US7338449B2 (en) *  20040525  20080304  Siemens Medical Solutions Usa, Inc.  Three dimensional locator for diagnostic ultrasound or medical imaging 
US7489419B2 (en) *  20040615  20090210  Infoprint Solutions Company, Llc  Method and system for specifying halftone spot shapes 
US20060034513A1 (en) *  20040723  20060216  Siemens Medical Solutions Usa, Inc.  View assistance in threedimensional ultrasound imaging 
US7639847B2 (en) *  20040908  20091229  Ge Medical Systems Information Technologies, Inc.  Coronary artery tree imaging system and method 
US7583829B2 (en) *  20040913  20090901  Siemens Medical Solutions Usa, Inc.  Method and apparatus for embolism analysis 
US20060074285A1 (en) *  20040924  20060406  Paieon Inc.  Apparatus and method for fusion and inoperatingroom presentation of volumetric data and 3D angiographic data 
EP1800265A2 (en) *  20041005  20070627  Philips Intellectual Property & Standards GmbH  Device, method, computerreadable medium and program element for processing image data of a movable object, and rotational xray angiography apparatus 
US6991738B1 (en)  20041013  20060131  University Of Washington  Flowthrough drum centrifuge 
US7127095B2 (en)  20041015  20061024  The Brigham And Women's Hospital, Inc.  Factor analysis in medical imaging 
US20060096358A1 (en) *  20041028  20060511  University Of Washington  Optical projection tomography microscope 
US7494809B2 (en) *  20041109  20090224  Visiongate, Inc.  Automated cell sample enrichment preparation method 
US20060155188A1 (en) *  20041124  20060713  Walczak Alan M  Method and system for determination of vessel tortuousity 
US8744146B2 (en) *  20041206  20140603  Siemens Aktiengellschaft  Vascular reformatting using curved planar reformation 
WO2006061814A1 (en) *  20041208  20060615  Paieon Inc.  Method and apparatus for finding the tubular organs blood velocity and flow and related parameters 
EP1830701A1 (en) *  20041208  20070912  Paieon Inc.  Method and apparatus for blood vessel parameter determinations 
EP1835855B1 (en) *  20050111  20170405  Volcano Corporation  Vascular image coregistration 
DE102005007893B4 (en) *  20050221  20070510  Siemens Ag  A method for determining the position of an instrument having an Xray system 
US8335357B2 (en) *  20050304  20121218  Kabushiki Kaisha Toshiba  Image processing apparatus 
JP2006239253A (en)  20050304  20060914  Toshiba Corp  Image processing device and image processing method 
WO2006103644A1 (en) *  20050331  20061005  Paieon Inc.  Method and apparatus for positioning a device in a tubular organ 
US8295577B2 (en) *  20050331  20121023  Michael Zarkh  Method and apparatus for guiding a device in a totally occluded or partly occluded tubular organ 
DE102005023167B4 (en) *  20050519  20080103  Siemens Ag  Method and apparatus for registration of 2D projection images relative to a 3D image data set 
DE102005027678A1 (en) *  20050615  20061228  Siemens Ag  Threedimensionalstructure marking method of twodimensional projection image, used in medical interventions, by calculating theoretical position of position marker on further twodimensional projection image in threedimensional space 
DE102005027963B3 (en) *  20050616  20061207  Siemens Ag  Method of reconstructing three dimensional image data set of moved object from set of projected images taken from different directions involves dividing images into several partial sets taken during given phases of object movement 
EP1903943A1 (en) *  20050621  20080402  Philips Electronics N.V.  Method and device for imaging a blood vessel 
JP2008543511A (en) *  20050624  20081204  ヴォルケイノウ・コーポレーション  Image method for manufacturing a vascular 
US7970187B2 (en) *  20050906  20110628  Pie Medical Imaging B.V.  Method, apparatus and computer program for contour detection of vessels using xray densitometry 
JP2007089992A (en) *  20050930  20070412  Terumo Corp  Energy irradiation device, control device and control method 
US20080273777A1 (en) *  20051021  20081106  Vincent Luboz  Methods And Apparatus For Segmentation And Reconstruction For Endovascular And Endoluminal Anatomical Structures 
US8731367B2 (en) *  20060301  20140520  Kabushiki Kaisha Toshiba  Image processing apparatus 
JP5523656B2 (en) *  20060301  20140618  株式会社東芝  Image processing apparatus 
WO2007127157A3 (en) *  20060428  20080410  Nicholas A Accomando  System and method for biometric retinal identification 
US7786898B2 (en)  20060531  20100831  Mobileye Technologies Ltd.  Fusion of far infrared and visible images in enhanced obstacle detection in automotive applications 
EP2046202B1 (en) *  20060628  20140611  Koninklijke Philips N.V.  Optimal rotational trajectory determination for ra based on predetermined optimal view map 
CN101479769A (en) *  20060628  20090708  皇家飞利浦电子股份有限公司  Modelbased determination of the contraction status of a periodically contracting object 
JP5121173B2 (en) *  20060629  20130116  株式会社東芝  3dimensional image generating device 
US7961924B2 (en) *  20060821  20110614  Stereotaxis, Inc.  Method of threedimensional device localization using singleplane imaging 
DE102006046285A1 (en) *  20060929  20090423  Siemens Ag  Object vessellike structures segmented representing method for patient, involves using determined probabilities as criterion for presence of vessel in segmentation process for representation of vessel structures 
US7940977B2 (en)  20061025  20110510  Rcadia Medical Imaging Ltd.  Method and system for automatic analysis of blood vessel structures to identify calcium or soft plaque pathologies 
US7940970B2 (en) *  20061025  20110510  Rcadia Medical Imaging, Ltd  Method and system for automatic quality control used in computerized analysis of CT angiography 
US7860283B2 (en)  20061025  20101228  Rcadia Medical Imaging Ltd.  Method and system for the presentation of blood vessel structures and identified pathologies 
US7983459B2 (en) *  20061025  20110719  Rcadia Medical Imaging Ltd.  Creating a blood vessel tree from imaging data 
US7873194B2 (en) *  20061025  20110118  Rcadia Medical Imaging Ltd.  Method and system for automatic analysis of blood vessel structures and pathologies in support of a triple ruleout procedure 
US8503741B2 (en) *  20061130  20130806  Siemens Aktiengesellschaft  Workflow of a service provider based CFD business model for the risk assessment of aneurysm and respective clinical interface 
US8543338B2 (en) *  20070116  20130924  Simbionix Ltd.  System and method for performing computerized simulations for imageguided procedures using a patient specific model 
GB2459225B (en) *  20070116  20110720  Simbionix Ltd  Preoperative surgical simulation 
US20110282586A1 (en) *  20070123  20111117  Kassab Ghassan S  Systems and methods to determine optimal diameters of vessel segments in bifurcation 
WO2008091568A3 (en)  20070123  20081127  Dtherapeutics Llc  Applications of scaling laws of tree structures 
US9591994B2 (en)  20070123  20170314  Dtherapeutics, Llc  Systems and methods to obtain a myocardial mass index indicative of an atrisk myocardial region 
US7835561B2 (en)  20070518  20101116  Visiongate, Inc.  Method for image processing and reconstruction of images for optical tomography 
DE102007024154A1 (en) *  20070524  20081127  Siemens Ag  Method for automatic selection of a display mode for an image data set of an organ to be examined 
WO2008150945A3 (en)  20070530  20090402  Cleveland Clinic Foundation  Automated centerline extraction method and generation of corresponding analytical expression and use thereof 
JP4388104B2 (en) *  20070629  20091224  ザイオソフト株式会社  Image processing method, image processing program and an image processing device 
US20100189337A1 (en) *  20070711  20100729  Koninklijke Philips Electronics N.V.  Method for acquiring 3dimensional images of coronary vessels, particularly of coronary veins 
US7660383B2 (en) *  20070823  20100209  Kabushiki Kaisha Toshiba  Three dimensional image processing apparatus and Xray diagnosis apparatus 
WO2009027889A3 (en) *  20070831  20100325  Koninklijke Philips Electronics N.V.  Uncertainty maps for segmentation in the presence of metal artifacts 
US8934604B2 (en) *  20070928  20150113  Kabushiki Kaisha Toshiba  Image display apparatus and Xray diagnostic apparatus 
US8571277B2 (en) *  20071018  20131029  Eigen, Llc  Image interpolation for medical imaging 
US7787112B2 (en)  20071022  20100831  Visiongate, Inc.  Depth of field extension for optical tomography 
US8428316B2 (en) *  20071203  20130423  Siemens Aktiengesellschaft  Coronary reconstruction from rotational Xray projection sequence 
US8064675B2 (en) *  20080130  20111122  Carestream Health, Inc.  Computeraided tubing detection 
EP2255312A1 (en) *  20080317  20101201  Koninklijke Philips Electronics N.V.  Perfusion imaging 
DE102008026035A1 (en) *  20080530  20091210  Siemens Aktiengesellschaft  A method of operating a pivotable Polyplan imaging system for imaging a moving object to be examined 
US8200466B2 (en)  20080721  20120612  The Board Of Trustees Of The Leland Stanford Junior University  Method for tuning patientspecific cardiovascular simulations 
RU2508907C2 (en) *  20080811  20140310  Конинклейке Филипс Электроникс Н.В.  Detecting perspective optimal representation maps with provision for crosssectional shape of vessel in cardiovascular xray systems 
US8577108B2 (en)  20080813  20131105  Carestream Health, Inc.  Method for detecting anatomical structures 
CN102223843B (en)  20081126  20131023  皇家飞利浦电子股份有限公司  Visualization of coronary artery tree 
US20100185672A1 (en) *  20090121  20100722  Rising Iii Hawley K  Techniques for spatial representation of data and browsing based on similarity 
CN102365653B (en) *  20090327  20150225  皇家飞利浦电子股份有限公司  Improvements to medical imaging 
JP5491914B2 (en)  20090428  20140514  株式会社東芝  The image display apparatus and the xray diagnostic apparatus 
US8934686B2 (en) *  20091126  20150113  Algotec Systems Ltd.  User interface for selecting paths in an image 
US20120303289A1 (en) *  20091202  20121129  Anders Ohrn  Combined onlattice/offlattice optimization method for rigid body docking 
US8346007B2 (en)  20091222  20130101  Carestream Health, Inc.  Noise suppression in cone beam CT projection data 
CN102113900B (en)  20100105  20150715  深圳迈瑞生物医疗电子股份有限公司  Relevant method and device for color blood flow dynamic frame 
US8526699B2 (en) *  20100312  20130903  Siemens Aktiengesellschaft  Method and system for automatic detection and classification of coronary stenoses in cardiac CT volumes 
DE102010012621A1 (en) *  20100324  20110929  Siemens Aktiengesellschaft  Method and device for automatic adaptation of a reference image 
US20110306025A1 (en) *  20100513  20111215  Higher Education  Ultrasound Training and Testing System with MultiModality Transducer Tracking 
JP5839822B2 (en) *  20100517  20160106  株式会社東芝  Image processing apparatus and xray ct device 
GB201008281D0 (en)  20100519  20100630  Nikonovas Arkadijus  Indirect analysis and manipulation of objects 
US8157742B2 (en)  20100812  20120417  Heartflow, Inc.  Method and system for patientspecific modeling of blood flow 
US8315812B2 (en)  20100812  20121120  Heartflow, Inc.  Method and system for patientspecific modeling of blood flow 
US9536314B2 (en) *  20101020  20170103  Siemens Medical Solutions Usa, Inc.  Image reconstruction 
US8494237B2 (en) *  20101108  20130723  Cranial Technologies, Inc  Method and apparatus for processing digital image representations of a head shape 
US8442288B2 (en) *  20101108  20130514  Cranial Technologies, Inc.  Method and apparatus for processing threedimensional digital mesh image representative data of threedimensional subjects 
US9875339B2 (en)  20110127  20180123  Simbionix Ltd.  System and method for generating a patientspecific digital imagebased model of an anatomical structure 
US8358823B2 (en) *  20110330  20130122  Mitsubishi Electric Research Laboratories, Inc.  Method for tracking tumors in biplane images 
US9504588B2 (en)  20110605  20161129  The Research Foundation For The State University Of New York  System and method for simulating deployment configuration of an expandable device 
US9097739B2 (en) *  20110627  20150804  The Johns Hopkins University  System for lightweight image processing 
US8805038B2 (en) *  20110630  20140812  National Taiwan University  Longitudinal image registration algorithm for infrared images for chemotherapy response monitoring and early detection of breast cancers 
EP2546804A1 (en) *  20110710  20130116  Dürr Dental AG  Method and tomography apparatus for reconstruction of a 3D volume 
US9105200B2 (en)  20111004  20150811  Quantant Technology, Inc.  Semiautomated or fully automated, network and/or webbased, 3D and/or 4D imaging of anatomy for training, rehearsing and/or conducting medical procedures, using multiple standard Xray and/or other imaging projections, without a need for special hardware and/or systems and/or preprocessing/analysis of a captured image data 
US9129417B2 (en) *  20120221  20150908  Siemens Aktiengesellschaft  Method and system for coronary artery centerline extraction 
US20130223712A1 (en) *  20120228  20130829  Canon Kabushiki Kaisha  Information processing apparatus, information processing method and radiation imaging system 
US8548778B1 (en)  20120514  20131001  Heartflow, Inc.  Method and system for providing information from a patientspecific model of blood flow 
US9053551B2 (en)  20120523  20150609  International Business Machines Corporation  Vessel identification using shape and motion mapping for coronary angiogram sequences 
US9814433B2 (en) *  20121024  20171114  Cathworks Ltd.  Creating a vascular tree model 
US9858387B2 (en)  20130115  20180102  CathWorks, LTD.  Vascular flow assessment 
US8864763B2 (en)  20130313  20141021  DePuy Synthes Products, LLC  External bone fixation device 
CA2905865A1 (en)  20130313  20141002  DePuy Synthes Products, Inc.  External bone fixation device 
US9039706B2 (en)  20130313  20150526  DePuy Synthes Products, Inc.  External bone fixation device 
US9925009B2 (en)  20130315  20180327  Covidien Lp  Pathway planning system and method 
US9459770B2 (en)  20130315  20161004  Covidien Lp  Pathway planning system and method 
US9639666B2 (en)  20130315  20170502  Covidien Lp  Pathway planning system and method 
CN104346368A (en) *  20130730  20150211  腾讯科技（深圳）有限公司  Indoor scene switch displaying method and device and mobile terminal 
DE102013226858A1 (en) *  20131220  20150625  Siemens Aktiengesellschaft  A process for producing an at least threedimensional display data set, xray device and computer program 
US20150223902A1 (en) *  20140207  20150813  Hansen Medical, Inc.  Navigation with 3d localization using 2d images 
CN103985121B (en) *  20140513  20170201  北京航空航天大学  An underwater projector cursor given process configuration 
WO2016004030A1 (en)  20140702  20160107  Covidien Lp  System and method for segmentation of lung 
US9603668B2 (en)  20140702  20170328  Covidien Lp  Dynamic 3D lung map view for tool navigation inside the lung 
US9754367B2 (en)  20140702  20170905  Covidien Lp  Trachea marking 
CN106232010A (en)  20140702  20161214  柯惠有限合伙公司  System and method for detecting trachea 
US9770216B2 (en)  20140702  20170926  Covidien Lp  System and method for navigating within the lung 
DE102014213408A1 (en) *  20140710  20160114  Siemens Aktiengesellschaft  A method for determining a threedimensional model data set of at least one vessel segment comprising the vascular system 
US9642586B2 (en) *  20140818  20170509  Siemens Aktiengesellschaft  Computeraided analysis of medical images 
EP3188059A1 (en) *  20140829  20170705  KNUIndustry Cooperation Foundation  Method for determining patientspecific blood vessel information 
CN104318557B (en) *  20141017  20170329  重庆大学  Vascular Reconstruction skeleton 
JP2016097261A (en) *  20141126  20160530  株式会社東芝  Image processing device, image processing program, image processing method and therapy system 
WO2017199245A1 (en) *  20160516  20171123  Cathworks Ltd.  System for vascular assessment 
Family Cites Families (4)
Publication number  Priority date  Publication date  Assignee  Title 

US4839822A (en) *  19870813  19890613  501 Synthes (U.S.A.)  Computer system and method for suggesting treatments for physical trauma 
US4875165A (en) *  19871127  19891017  University Of Chicago  Method for determination of 3D structure in biplane angiography 
US5839440A (en) *  19940617  19981124  Siemens Corporate Research, Inc.  Threedimensional image registration method for spiral CT angiography 
US5709206A (en) *  19951127  19980120  Teboul; Michel  Imaging system for breast sonography 
Also Published As
Publication number  Publication date  Type 

US6047080A (en)  20000404  grant 
US6501848B1 (en)  20021231  grant 
EP0978101A1 (en)  20000209  application 
WO1997049065A1 (en)  19971224  application 
Similar Documents
Publication  Publication Date  Title 

Mikić et al.  Human body model acquisition and tracking using voxel data  
Wahle et al.  Assessment of diffuse coronary artery disease by quantitative analysis of coronary morphology based upon 3D reconstruction from biplane angiograms  
US5289373A (en)  Method and apparatus for realtime tracking of catheter guide wires in fluoroscopic images during interventional radiological procedures  
US5588033A (en)  Method and apparatus for three dimensional image reconstruction from multiple stereotactic or isocentric backprojections  
Li et al.  Establishing a normative atlas of the human lung: intersubject warping and registration of volumetric CT images  
US7231076B2 (en)  ROI selection in image registration  
US7327865B2 (en)  Fiducialless tracking with nonrigid image registration  
US20050015003A1 (en)  Method and device for determining a threedimensional form of a body from twodimensional projection images  
US7366278B2 (en)  DRR generation using a nonlinear attenuation model  
Tuceryan et al.  Calibration requirements and procedures for a monitorbased augmented reality system  
US20080039713A1 (en)  Dynamic tracking of moving targets  
US5602891A (en)  Imaging apparatus and method with compensation for object motion  
US20080262345A1 (en)  Image registration of multiple medical imaging modalities using a multiple degreeoffreedomencoded fiducial device  
US6563941B1 (en)  Modelbased registration of cardiac CTA and MR acquisitions  
US5287274A (en)  Method for acquisition of radiological data in multiple orthogonal orientations using a 2D detector relating to a body irradiated with xrays and for reconstruction of structures corresponding to said body using an algebraic algorithm  
Penney et al.  A comparison of similarity measures for use in 2D3D medical image registration  
US5175773A (en)  Method of threedimensional reconstruction of arborescence by labeling  
US5768405A (en)  Digital image processing method for local determination of the center and the width of objects in the form of contrasting bands on a background  
US7117026B2 (en)  Physiological model based nonrigid image registration  
Stein  Lens distortion calibration using point correspondences  
Vandemeulebroucke et al.  The POPImodel, a pointvalidated pixelbased breathing thorax model  
US5868675A (en)  Interactive system for local intervention inside a nonhumogeneous structure  
US7113623B2 (en)  Methods and systems for display and analysis of moving arterial tree structures  
Pellot et al.  A 3D reconstruction of vascular structures from two Xray angiograms using an adapted simulated annealing algorithm  
US20050008210A1 (en)  System and method for threedimensional reconstruction of an artery 
Legal Events
Date  Code  Title  Description 

FZDE  Dead 