WO2006038166A2

WO2006038166A2 - Device, method, computer-readable medium and program element for processing image data of a movable object, and rotational x-ray angiography apparatus

Info

Publication number: WO2006038166A2
Application number: PCT/IB2005/053217
Authority: WO
Inventors: Thomas BÜLOW; Cristian Lorenz; Babak Movassaghi; Volker Rasche
Original assignee: Philips Intellectual Property & Standards Gmbh; Koninklijke Philips Electronics N. V.
Priority date: 2004-10-05
Filing date: 2005-09-29
Publication date: 2006-04-13
Also published as: EP1800265A2; WO2006038166A3

Abstract

A device (104) for processing image data of a movable object (101), the device (104) comprising a processor adapted to carry out the method steps of receiving (410) an image data set reflecting a two-dimensional image according to a particular twodimensional projection of the movable object (101) at a particular point of time, determining (420) an initial model for the two-dimensional image, wherein the initial model is determined based on predetermined structural information about the movable object (101) and based on predetermined motion characteristics of the movable object (101), and modifying (430) the determined initial model to obtain a final model for the two-dimensional image.

Description

Device, method, computer-readable medium and program element for processing image data of a movable object, and rotational X-ray angiography apparatus

The invention relates to the field of processing image data. In particular, the invention relates to a device, to a method, to a computer-readable medium and to a program element for processing image data of a movable object, and to a rotational X-ray angiography apparatus for imaging a movable object.

Angiography is a medical imaging technique in which an X-ray picture is taken to visualize the inner opening of blood filled structures, including arteries, veins and the heart chambers. As blood has essentially the same radiodensity as the surrounding tissues, a radiocontrast agent (which absorbs X-rays) is added to the blood to make angiography visualization possible. The angiography X-ray image is actually a shadow picture of the openings within the cardiovascular structures carrying blood (actually the radiocontrast agent within). The blood vessels or heart chambers themselves remain largely invisible on the X-ray image.

The most common angiogram performed is to visualize the blood in the coronary arteries. A long, thin, flexible tube called a catheter is used so as to administer the radiocontrast agent at the desired area to be visualized. The catheter is threaded into an artery, and a tip is advanced through the arterial system into one of the major coronary arteries. X-ray images of the transient radiocontrast distribution within the blood flowing within the coronary arteries allows visualization of the size of the artery cross-section. Rotational X-ray angiography of the heart provides high resolution images of coronary arteries where the viewing direction is continuously changed between the acquisition of subsequent images. A variation between subsequent images is caused by the varying viewing direction, and additionally by the movement of the heart. In order to be able to compute a three-dimensional reconstruction from a plurality of two-dimensional projections (similar to CT reconstruction), three-dimensional rotational coronary angiography (3D-RCA) can be applied (see Rasche, V. et al. (2003) "Ecg-gated 3d rotational coronary angiography", RSNA, pages C 19-382, 83^rd Scientific Session). Hence, ECG-gated projections (belonging to the same cardiac phase) are selected and directly utilized for a filtered back projection procedure such as of Feldkamp reconstruction algorithm (see Grass, M. et al. (1999) "Three-dimensional reconstruction of high-contrast objects using c-arm image intensifier projection data", Comp Med Imag and Graphics 23, pages 311 to 321,).

A main obstacle is the limited reproducibility of the cardiac motion and the respiratory motion. To involve additional projections in the reconstruction procedure, acquired at different heart phase and to compensate for respiratory motion, 3D-RCA motion compensated reconstruction techniques can be applied (see Mouvassaghi, B. et al. (2003) "3d coronary reconstruction from calibrated motion-compensated 2d projections", Computer Assisted Radiology and Surgery, (CARS 2003), pages 1079 to 1084, Elsevier; Blondel, C. et al. (2003) "4d deformation field of coronary arteries from monoplane rotational x-ray angiography", Computer Assisted Radiology and Surgery, (CARS 2003), pages 1073 to 1078, Elsevier) . One way to compute the motion between two heart cycles is to semi-interactively determine a three-dimensional representation of the coronary artery centerlines in one heart phase utilizing coronary modeling techniques (see Mouvassaghi, B. et al. (2003) "Quantitative analysis of 3d coronary modeling in 3d rotational x-ray imaging", Proc. IEEE Nuclear Science Symposium (NSS) and Medical Imaging Conference (MC), Norfolk, Virginia; Mouvassaghi, B. et al. (2003) "An accurate coronary modeling procedure using 2d calibrated projections based on extracted 3d vessel centerlines", Computer Assisted Radiology and Surgery, (CARS 2003), page 1397, Elsevier). The three-dimensional centerlines can then be projected into the different images according to the imaging geometry. By optimizing a three-dimensional geometrical transformation of the centerlines, the projection is brought into alignment with the image of the coronary arteries at a different heart phase. Applying the inverse of the resulting 3D transformation (projected to a two-dimensional image plane) may yield the necessary information to approximately compensate the images for the cardiac motion. However, this static approach does not yield proper results in a critical scenario.

A disadvantage in the prior art approach mentioned above is the process of optimizing the three-dimensional transformation in order to bring the centerline projection in alignment with the image of the vessels. In particular, if the projection of the initial three-dimensional centerline tree is not close enough to the actual positions of the vessels, the optimization process frequently gets trapped at false structures like different vessels, images of the catheter or ECG (electrocardiogram) cables.

US2004/0066958 Al discloses a system for reconstruction of a three- dimensional representation of a moving arterial tree structure from a pair of sequences of time varying two-dimensional images. After having identified two-dimensional coronary arterial trees, moving three-dimensional coronary arterial trees are calculated. However, the system of US2004/0066958 Al requires to acquire and to analyze a pair of sequences of two-dimensional images and thus introduces a huge amount of data requiring powerful and expensive computing resources.

It is an object of the invention to provide a system for processing image data of a moving object allowing to determine the structure of the moving object with reasonable effort and high accuracy.

In order to achieve the object defined above, a device, a method, a computer-readable medium and a program element for processing image data of movable object, and a rotational X-ray angiography apparatus, with the features according to the independent claims are provided.

The device for processing image data of a movable object comprises a processor adapted to carry out the method steps of receiving an image data set reflecting a two-dimensional image according to a particular two-dimensional projection of the movable object at a particular point of time, determining an initial model for the two- dimensional image, wherein the initial model is determined based on predetermined structural information about the movable object and based on predetermined motion characteristics of the movable object, and modifying the determined initial model to obtain a final model for the two-dimensional image.

Further, a rotational X-ray angiography apparatus for imaging a movable object is provided, comprising an X-ray source, an X-ray detector and a device having the above-mentioned features for processing image data of a movable object. The X-ray source is adapted to irradiate X-rays to a movable object, the X-ray detector is adapted to capture X-rays scattered from the movable object to detect an image data set reflecting a two- dimensional image according to a particular two-dimensional projection of the movable object at a particular point of time. The device is coupled to the X-ray detector to receive the image data set. Moreover, a method for processing image data of a movable object is provided, the method comprising the above-mentioned method steps.

Beyond this, a computer-readable medium is provided, in which a computer program for processing image data of a movable object is stored which, when being executed by a processor, is adapted to carry out the above-mentioned method steps. Moreover, a program element for processing image data for a movable object is provided, which, when being executed by a processor, is adapted to carry out the above-mentioned method steps.

Thus, the processing of image data of a movable object, according to the invention, can be realized by a computer program, i.e. by software, or by using one or more special electronic optimization circuits, i.e. in hardware, or in hybrid form, i.e. by means of software components and hardware components.

The characteristic features according to the invention have particularly the advantage that image data of a movable object are processed, to calculate an image from the image data, under consideration of not only predetermined structural information about the movable object, but also by taking into account predetermined motion characteristics of the movable object. Static/geometric (e.g. anatomical) information and dynamic information (e.g. a repeated, periodic motion cycle) of a moving object are combined in an advantageous manner to determine a start model (e.g. a set of initial values for parameters of a geometrical model roughly approaching the actual image of the object) as a realistic and reasonable basis for a subsequent computer fit by which a final model (e.g. a set of final values for the parameters of the geometrical model properly representing the actual image of the object) are adjusted.

Thus, a two-step process for obtaining a proper fit between captured image data and the final model is provided. First, a proper initial model is determined, based on structural and dynamic frame conditions. Second, the actual fit procedure is carried out, in which the initial model is modified to refine the consistence between the experimentally measured image data set and the computationally derived final model for the image. The pre-analysis of the image data taught by the invention ensures that a reasonable set of start parameters for the subsequent fit is determined to avoid that a subsequent computer fit is trapped in an artificial (merely mathematical, non-physical) minimum to yield a wrong result. By ensuring that the fit starts with reasonable start parameters, it can be securely prevented that the fit yields meaningless results as a consequence of an initial model being too far away from reality.

For example, the periodic motion of a beating heart (as a moving object which is investigated by rotational X-ray angiography) involves a sequence of events called the cardiac cycle. This consist of three major stages, the aterial systole, the ventricular systole and the complete cardiac diastole. During these different "operation modes" of the human or animal heart, the motion of the heart also influences the structure of the centerlines of the coronary arteries. Thus, when a set of images of the heart are captured by rotational X-ray angiography (which may take some seconds), the different images suffer from motion artefacts resulting from the fact that the heart beats during detecting the signals. According to the invention, such a pre-determinable motion is taken into account when initial fit parameters for a fit (e.g. a least-squares fit) for determining a structure of the heart are selected.

Both structural and dynamic information about the object under consideration are taken into account to estimate reasonable initial fit parameters for determining two-dimensional projections of the object. By particularly considering the motion characteristics of the movable object, motion artefacts can be compensated securely.

According to an exemplary embodiment of the invention, the following sequence of method steps may be carried out: measure image data sets representing two- dimensional projections of the moving object; estimate a starting model for a two- dimensional image for a particular image data set, assuming preknown information about structure (e.g. anatomy) and dynamics (e.g. a cardiac cycle) of the imaged object (e.g. a heart); calculate, starting with the initial model, a final model of the two-dimensional image for a particular image data set; reconstruct, on the basis of a plurality of two- dimensional images, a three-dimensional image of the moving object using known methods.

The invention discloses a method of motion modeling of the coronary arteries for three-dimensional reconstruction. Having such a model at hand, the three- dimensional motion of the coronary arteries from one cardiac phase to another cardiac phase can be predicted. The projection of the predicted three-dimensional centerline to the three-dimensional image plane will in general not coincide with the true vessel position. However, it can be expected that it will be close enough in order to provide a reasonable initialization of a fitting or optimization process. As a result, the optimization procedure is much less prone to be trapped at false structures, since motion artefacts are securely prevented from negatively influencing the evaluation procedure.

Particularly, the invention teaches a component of a fully automated system to motion compensated three-dimensional reconstruction, particularly of the coronaries of a heart, from rotational X-ray angiography. In other words, a motion model is used for motion compensated three-dimensional reconstruction of cardiac rotational X-ray projection. Thus, the invention can be advantageously applied in the field of three- dimensional X-ray reconstruction for three-dimensional coronary angiography. Particularly, the invention yields a contribution in the field of image processing for cardiac MSCT ("multi-slice computer tomography").

Motion compensation for three-dimensional reconstruction of cardiac rotational X-ray projections can be tackled by fitting the projection of a three-dimensional vessel centerline model to the different heart phases using a geometry three-dimensional transform. Further, a motion model of the coronary arteries is used for the initialization of such a fitting procedure.

According to one aspect of the invention, a geometric model of the centerlines of the coronary arteries for one cardiac phase plus additional information encoding the motion between different cardiac phases can be used. Such a geometric model can be extracted from ground-truth data, e.g. segmented cardiac MSCT data. The motion between different cardiac phases can be encoded as a geometric three-dimensional transformation. Such a three-dimensional transformation can be based, for instance, on a rigid model (taking into account translation and rotation), or by an affine model (additionally considering shearing), or based on a thin plate spline model (which further enhances the elasticity of the model). Such a generic model can be adapted to the actual data by an optimization procedure.

According to another aspect of the invention, the motion characteristics used for determining start parameters for the fit, may be estimated from an actual data set at hand. Low resolution three-dimensional reconstructions of the heart in various phases can be obtained and are computationally rather inexpensive. Motion fields between these three- dimensional reconstructions can be estimated, e.g. by optical flow methods. The estimated motion fields may then be used in order to update the coronary centerline tree from one cardiac phase to the next.

The motion model of the invention which may be used to improve the three- dimensional image of coronary artery centerlines may be stored in a file or may be provided by a database. It falls also under the scope of the invention that a three- dimensional coronary angiography equipment includes an update service in order to provide improved or additional models.

An idea of the invention is to carry out a motion compensation for a three- dimensional reconstruction of cardiac rotational X-ray by fitting a three-dimensional vessel centerline model to different heart phases, by using a motion model of the arteries, thus obtaining an initial model for an initialization of a fitting procedure.

The invention obviates the problem which usually occurs during capturing of image data of a beating heart. In case that such a beating heart is imaged by acquiring a plurality of data sets, each representing a particular two-dimensional projection of the beating heart at different times, is that the motion of the heart changes the structure of the heart at the different capture times and thus yields a geometric distortion of parts of the heart, i.e. centerlines of coronary arteries. For instance, when using a rotational X-ray angiography apparatus, a half circle may be scanned for a time of approximately twenty seconds, and a plurality of two-dimensional image data sets are acquired. The invention is based on the fact that it is desirable that all two-dimensional projections of the heart should be correspond to one particular heart phase to avoid motion artefacts. Advantageously, all two-dimensional pictures should be referred to a phase of the cardiac motion in which the actual motion is small. It is thus advantageous to transform all two-dimensional pictures to the chosen phase to suppress artefacts resulting from heartbeat. Thus, the invention preferably uses a model tree for coronary arteries as a structural model for the centerlines of the coronaries of a heart. Further, the motion of the heart is taken into account for determining a set of start parameters from which a subsequent fit to achieve a two- dimensional projection is provided. The dynamic model of the invention takes into account structural frame conditions and dynamic characteristics of the heart to provide sufficiently reasonable start parameters for a subsequent optimization calculation.

Referring to the dependent claims, further preferred embodiment of the invention will be described in the following. Next, preferred embodiments of the device for processing image data of a movable object will be described. These embodiments may also be applied for the method, for the rotational X-ray angiography apparatus, for the computer-readable medium and for the program element.

The device of the invention may be adapted to receive a plurality of image data sets, each reflecting a two-dimensional image according to a particular two- dimensional projection of the movable object at a particular point of time, wherein the steps of determining and modifying are carried out individually for each of the plurality of image data sets to obtain a final model for each of the two-dimensional images. In other words, different two-dimensional projection imaging data are acquired by the device and may be later used to reconstruct the three-dimensional structure of the object. However, the estimation of a final model for each of the sets of image data is performed individually, and as a start for a fit for estimating such a final model, a set of initial fit parameters to be varied during the fit are selected depending on structural and dynamic information about the movable object. The processor of the device may be adapted to reconstruct a three- dimensional projection of the movable object based on the final models for each of the two-dimensional images. After having performed the fits, starting with the previously chosen fit parameters, the plurality of two-dimensional projections are calculated. On the basis of this plurality of two-dimensional projections, a three-dimensional view of the object is reconstructed, using conventional methods as known by those skilled in the art. By improving the calculation for determining the two-dimensional projections according to the invention, the quality of the three-dimensional reconstruction is improved, the calculation time is reduced and the danger of motion artefacts is reduced as well. The processor of the device may be adapted to reconstruct a three- dimensional image of a heart based on the final models for each of the two-dimensional images. A human or animal heart investigated beats, according to a characteristic cardiac cycle, during acquiring several image data sets. This cardiac cycle, which is well known as a result of previous scientific investigations, is considered when evaluating the measured data.

Particularly, the processor of the device may be adapted to reconstruct a three-dimensional projection of centerlines of coronary arteries of a heart based on the final models for each of the two-dimensional images. The coronary arteries of a heart are regions of particular interests for medical investigations because they provide important information about the heart. Particularly, by injecting a radiocontrast agent in the coronary arteries, the coronary arteries become clearly visible by angiography.

The processor of the device may be adapted to, based on the predetermined motion characteristics of the movable object, compensate for inconsistencies of information concerning the three-dimensional structure of the movable object provided by different image data sets, which inconsistencies result from the motion of the movable object.

Thus, artefacts in the determined image of the heart are avoided, since these differences are at least partially compensated by the method of the invention.

The predetermined motion characteristics of the movable object may take into account a periodic motion of the movable object. By taken this measure, i.e. by assuming that the motion of the moving object follows a periodic cycle, the motion of the object in the future can be predicted so that a motion state and a correlated geometry can be determined for each point of time. For each image data set, the motion state of the moving object at the point of time at which the respective image data set has been measured can be estimated. In particular, the predetermined motion characteristics may take into account a cardiac cycle of a heart as the movable object. Consequently, the influence of the cardiac cycle to the structure of the heart can be considered, and the disturbing influence of the cardiac cycle to the reconstruction of the three-dimensional structure of the heart can thus be eliminated by the motion compensation of the invention.

The processor of the device may be adapted to provide, as the initial model, a set of parameter values as initial parameters for a subsequent fit. Thus, the processor computes a set of reasonable fit parameters which have a proper probability to be sufficiently close to the final model so that a danger that the fit is trapped in a "wrong" minimum, i.e. a minimum which does not reflect the real structure of the object under consideration, is significantly reduced. Particularly, the processor can be adapted, to obtain the final model from the initial model, to carry out a fit to adjust the parameters of the image data set, using the set of parameter values as initial parameters. Such a fit adjusts one or more parameters of a theoretical image model to achieve a sufficient or optimized consistence with measured data.

The predetermined motion characteristics of the movable object may include a theoretical model predicting the time dependence of the motion of the movable object. Such a theoretical model allows the processor to reduce or even to eliminate the influence of motion according to the chosen theoretical model to improve the quality of the set of initial fit parameters.

The processor may be adapted to determine motion characteristics of the movable object by reconstructing a plurality of three-dimensional projections of the movable object based on a part of the data of the image data set by determining motion characteristics by comparing different three-dimensional projections. According to the described embodiment, the measured data are taken as a basis for a fast determination of a rough approximation of a three-dimensional structure of the object. In order to keep the computational burden low, thus reducing the calculation time, only a (small) part of the captured data (e.g. a half or a quarter of all data, e.g. every second or every forth data point) can be used for this determination. By analyzing one such three-dimensional structure, or by comparing different such three-dimensional structures derived from data captured at different points of time and thus related to different motion states of the investigated object, centers or regions of strong or low motion can be estimated, and based on this motion information, the initial model can be determined.

Referring to the previously described embodiment, the processor may be adapted to determine the motion characteristics using an optical flow method.

It is noted that predetermined structural information about the movable object may be, for the example of a heart comprising coronary artery trees, be modeled as described by Lorenz, C. et al. "Modeling the coronary artery tree", Shape Modeling International (SMI) 2004, Genova, June 6 to 9, 2004. In the following, a preferred embodiment of the rotational X-ray angiography apparatus for imaging a movable object will be described. This embodiment also applies for the device, for the method, for the computer-readable medium and for the 1

program element for processing image data of a movable object.

The rotational X-ray angiography apparatus may be adapted to receive a plurality of image data sets, each reflecting a two-dimensional image according to a particular two-dimensional projection of a movable object at a particular point of time, wherein the steps of determining and modifying are carried out individually for each of the plurality of image data sets to obtain a final model for each of the two-dimensional images.

The aspects defined above and further aspects of the invention are apparent from the examples of the embodiments to be described hereinafter and are explained with reference to these examples of embodiment. The invention will be described in more detail hereinafter with reference to examples of embodiment but to which the invention is not limited.

Fig.l shows a rotational X-ray angiography apparatus according to a preferred embodiment of the invention.

Fig.2 and Fig.3 illustrate the improvement of initialization according to the method of the invention.

Fig.4 shows a flow-chart illustrating method steps according to a method for processing image data of a movable object according to a first embodiment of the invention.

Fig.5 shows a flow-chart illustrating method steps according to a method for processing image data of a movable object according to a second embodiment of the invention.

Fig.6 shows a flow-chart illustrating method steps according to a method for processing image data of a movable object according to a third embodiment of the invention.

The illustration in the drawings is schematically. In different drawings, similar or identical elements are provided with the same reference signs. In the following, referring to Fig.1 , a rotational X-ray angiography apparatus 100 for imaging a beating heart 101 according to a preferred embodiment of the invention will be described.

Fig.l shows a rotational X-ray angiography apparatus 100 for imaging a beating heart 101. The rotational X-ray angiography apparatus 100 comprises an X-ray tube 102 as an X-ray source, an X-ray detector 103 for detecting X-rays, and a computer system 104 for processing image data of the beating heart 101.

The beating heart 101 is located within the body of a human being (not shown) under investigation. The beating heart 101 of the human being (a patient) shall be examined.

The X-ray detector 103 is controlled by a first control unit 106 which, in term, is controlled by the computer system 104, and is capable of detecting X-rays. Based on detected X-rays, the X-ray detector 103 estimates an image data set reflecting a two- dimensional projection of the beating heart 101 with respect to a current position of the X- ray detector 103 and the X-ray tube 102. Each actual image data set is transmitted from the X-ray detector 103 to the computer system 104 for further processing. In other words, the X-ray detector 103 is connected to the first control unit 106 which is connected to the computer system 104 and which receives control information for controlling the X-ray detector 103. The X-ray detector 103 functions in correspondence with the control information received from the computer system 104 and outputs measured data concerning the two-dimensional projection of the beating heart 101 to the computer system 104. For calculating back the three-dimensional image of the heart 101, the data measured by the X- ray detector 103 are transmitted to the computer system 104.

The X-ray tube 102 is controlled by a second control unit 107 which, in term, is controlled by the computer system 104, and is capable of emitting X-rays to irradiate the beating heart 101 inside the human being's body in a controllable manner. In other words, the second control unit 107 is capable of receiving control information from the computer system 104 and is coupled with the X-ray tube 102 to control emission of X- rays. The computer system 104 comprises a processor (not shown) adapted to carry out the method steps of receiving, from the X-ray detector 103, an image data set reflecting a two-dimensional image according to a particular two-dimensional projection of the beating heart 101 at a particular point of time. The X-ray detector 103 rotates in correspondence with the rotating X-ray tube 102 on a rotation circle 105 to detect X-rays which are produced by the X-ray tube 102 and scattered by the beating heart 101. These scattered X-rays are captured by the X-ray detector 103 as image data sets. During a rotation cycle, a plurality of data sets are acquired by the X-ray detector 103, each data set corresponding to a two-dimensional projection of the beating heart 101 with respect to the actual position of the X-ray tube 102 and of the X-ray detector 103.

Further, the processor of the computer system 104 is adapted to determine an initial model for the two-dimensional image, wherein the initial model is determined based on predetermined structural information about the beating heart 101 and based on predetermined motion characteristics of the beating heart 101. The computer system 104 determines an initial model for the two-dimensional image according to an image data set measured by the X-ray detector 103, wherein the initial model is determined based on structural information about the beating heart 101 and based on predetermined motion characteristics of the beating heart 101. Thus, for each set of data representing a two- dimensional projection of the beating heart 101, the computer system 104 calculates back the respective two-dimensional image. This is performed in a two-step process. First, an initial model is determined, including determining a set of starting parameters for a subsequent fit to determine a final model. Since the determination of the initial model includes structural and dynamic information about the beating heart 101, motion artefacts can be eliminated and the subsequent fit has a better chance to be tapped into a physical minimum, so that the optimization procedure is much less prone to be trapped at a false structure. To modify the determined initial model, a fit is carried out to obtain a final model for the two-dimensional image. Based on the plurality of two-dimensional proj ections of the beating heart 101 , the processor 104 then reconstructs a three- dimensional image of the movable object based on the final models for each of the two- dimensional images.

According to the described embodiment, the structural information about the movable object used to derive the initial model includes a geometrical model of the structure of a human heart which is known from basic anatomical investigations (e.g. typical ratios between dimensions of different parts of a "standard" human heart). Particularly, a geometric model of the centerlines of the coronary arteries for particular cardiac phases is used.

Additionally, information encoding the motion between different cardiac phases is used according to the embodiment described referring to Fig.l for selecting the initial model. For this purpose, a dynamic model of the cardiac cycle of a human heart which is known from basic anatomical investigations is included in the image data processing of the invention.

In the following, referring to Fig.2 and Fig.3, the principle of the invention of improving a start model for a subsequent fit to retrieve a two-dimensional projection from image data will be illustrated.

Fig.2 shows a first projection 200 illustrating, on a projection area 201, a plurality of centerlines of the coronary artery of a heart. Fig.2 shows a projection of a centerline tree 202 from a previous phase of the heart beat. As can be seen on the projection area 201, a poor initialization is obtained so that there exists an imminent danger that the subsequent optimization procedure will be trapped at a false structure.

Fig.3 shows a second projection 300 according to which a significantly improved initialization on a projection area 301 is obtained, as a result of applying a motion model, i.e. a motion of the centerline tree 302 is allowed when searching for an initial model. By increasing the degree of freedom of the pre-fit of the invention, a better initialization is achieved so that a subsequent main fit is much less prone to be trapped in a false structure.

In the following, referring to Fig.4, a method for processing image data of a beating heart according to a first embodiment of the invention will be described.

Starting with an acquire step 410, a plurality of two-dimensional data sets of coronary arteries of the heart 101 is acquired. Each of the two-dimensional data sets represents a particular projection of the three-dimensional heart 101, according to an actual position of the X-ray tube 102 and of an X-ray detector 103 with respect to the heart 101. Thus, each of the sets of data is acquired at a different time and at a different angle of rotation along rotation circle 105, so that different pictures may be taken at different phases of the cardiac cycle.

In a calculating step 420, a start model for a subsequent fit of each of the two-dimensional data sets is calculated taking into account a static model of centerlines of coronary arteries and taking into account a motion model concerning the motion of the heart. In other words, a pre-evaluation is performed to find a plurality of reasonable start parameters for a subsequent fit allowing the subsequent main fit to find, with a higher probability, a correct numerical minimum reflecting the real image of the heart. Each set of two-dimensional projection data is evaluated separately using known motion information and known geometry information about the heart 101.

Based on this start model, in a fit step 430, each of the two-dimensional data sets are fit to obtain a respective two-dimensional projection image of the centerlines of the coronary arteries, starting with the previously calculated start model. Since the start model provides a plurality of reasonable fit parameters compatible with structural and dynamic frame conditions, the fit is fast and yields reliable results, i.e. the two-dimensional projection images according to the plurality of image data sets.

In a reconstructing step 440, a three-dimensional image of the coronary arteries of the heart 101 is reconstructed from each of the two-dimensional projection images obtained in step 430.

In step 450, the three-dimensional image is displayed on a screen wherein differences between the two-dimensional projections related to motion of the heart 101 are at least partially eliminated or compensated.

In the following, referring to Fig.5, a flowchart 500 will be described illustrating a method for processing image data of a beating heart 101 according to a second embodiment of the invention.

As can be seen in Fig.5, in the steps 410, 430, 440 and 450 equal to the first embodiment shown in Fig.4. However, calculating 420 a start model for a subsequent fit is based on a geometrical model of centerlines of the coronary arteries 510 and is based on a theoretical model of motion between different cardiac phases 520. In other words, a geometrical model reflecting a typical array of centerlines of the coronary arteries of a human heart are included in the calculating step for determining a start model. Further, typical motion of the heart 101, namely a known cardiac cycle, is introduced as a theoretical model 520 in the fit. In the following, referring to Fig.6, a flowchart 600 is shown illustrating a method for processing image data of a beating heart 101 according to a third embodiment of the invention. Steps 410, 430, 440, 450 equal to the method described according to Fig.4. The calculation of the start model for a subsequent fit, namely the calculating step 420, is carried out on the basis of a geometrical model 510 like in the case of Fig.5, and additionally motion information is included in the calculating step 420 by an experimental model of motion 610, which estimates motion by a low resolution 3D reconstruction using an optical flow method. For this purpose, the data according to the acquire step 410 are provided to the experimental model of motion step 610, wherein the experimental model of motion step 610 uses only a small part of the data provided by step 410 to calculate a rough approximation of the 3D reconstruction from which motion data is derived. It should be noted that the term "comprising" does not exclude other elements or steps and the "a" or "an" does not exclude a plurality. Also elements described in association with different embodiments may be combined.

It should also be noted that reference signs in the claims shall not be construed as limiting the scope of the claims.

LIST OF REFERENCE SIGNS:

100 rotational X-ray angiography apparatus 101 heart

102 X-ray tube

103 X-ray detector

104 computer system

105 rotation circle 106 first control unit

107 second control unit

200 first projection

201 projection area

202 centerline tree 300 second projection

301 projection area

302 centerline tree 400 flowchart

410 Acquire 420 Calculate

430 Fit

440 Reconstruct

450 Display 500 flowchart 510 Geometrical model

520 Theoretical model of motion 600 flowchart

610 Experimental model of motion

Claims

CLAIMS:

1. A device (104) for processing image data of a movable object (101), the device (104) comprising a processor adapted to carry out the following method steps: receiving (410) an image data set reflecting a two-dimensional image according to a particular two-dimensional projection of the movable object (101) at a particular point of time; determining (420) an initial model for the two-dimensional image, wherein the initial model is determined based on predetermined structural information about the movable object (101) and based on predetermined motion characteristics of the movable object (101); modifying (430) the determined initial model to obtain a final model for the two-dimensional image.

2. The device (104) according to claim 1, being adapted to receive (410) a plurality of image data sets, each reflecting a two-dimensional image according to a particular two-dimensional projection of the movable object (101) at a particular point of time, wherein the steps of determining (420) and modifying (430) are carried out individually for each of the plurality of image data sets to obtain a final model for each of the two-dimensional images.

3. The device (104) according to claim 2, wherein the processor is adapted to reconstruct (440) a three-dimensional image of the movable object (101) based on the final models for each of the two-dimensional images.

4. The device (104) according to claim 2, wherein the processor is adapted to reconstruct (440) a three-dimensional image of a heart (101) based on the final models for each of the two-dimensional images.

5. The device (104) according to claim 2, wherein the processor is adapted to reconstruct a three-dimensional image of centerlines of coronary arteries of a heart (101) based on the final models for each of the two-dimensional images.

6. The device (104) according to claim 2, wherein the processor is adapted to, based on the predetermined motion characteristics of the movable object (101), compensate for inconsistencies of information concerning the three-dimensional structure of the movable object (101) provided by different image data sets, which inconsistencies result from the motion of the movable object (101).

7. The device (104) according to claim 1, wherein the predetermined motion characteristics of the movable object (101) take into account a periodic motion of the movable object (101).

8. The device (104) according to claim 1, wherein the predetermined motion characteristics take into account a cardiac cycle of a heart (101) as the movable object.

9. The device (104) according to claim 1, wherein the processor is adapted to provide, as the initial model, a set of parameter values as initial parameters for a subsequent fit.

10. The device (104) according to claim 9, wherein the processor is adapted, to obtain the final model from the initial model, to carry out a fit to adjust the parameters to the image data set, using the set of parameter values as initial parameters.

11. The device (104) according to claim 1 , wherein the predetermined motion characteristics of the movable object (101) include a theoretical model (520) predicting the time dependence of the motion of the movable object (101).

12. The device (104) according to claim 2, wherein the processor is adapted to determine motion characteristics'of the movable object (101) by reconstructing (610) a plurality of three-dimensional images of the movable object (101) based on a part of the of data of the image data set and by determining motion characteristics by comparing different three-dimensional images.

13. The device (104) according to claim 12, wherein the processor is adapted to determine the motion characteristics using an optical flow method.

14. A rotational X-ray angiography apparatus (100) for imaging a movable object (101), comprising an X-ray source (102); an X-ray detector (103); a device (104) according to claim 1 for processing image data of a movable object (101); wherein the X-ray source (102) is adapted to irradiate X-rays on a movable object (101); wherein the X-ray detector (103) is adapted to capture X-rays scattered from the movable object (101) to detect (410) an image data set reflecting a two-dimensional image according to a particular two-dimensional projection of the movable object (101) at a particular point of time; wherein the device (104) is coupled to the X-ray detector (103) to receive the image data set.

15. The rotational X-ray angiography apparatus (100) according to claim 14, being adapted to receive (410) a plurality of image data sets, each reflecting a two- dimensional image according to a particular two-dimensional projection of the movable object (101) at a particular point of time, wherein the steps of determining (420) and modifying (430) are carried out individually for each of the plurality of image data sets to obtain a final model for each of the two-dimensional images.

16. A method (400) for processing image data of a movable object (101), the method (400) comprising the steps of: receiving (410) an image data set reflecting a two-dimensional image according to a particular two-dimensional projection of the movable object (101) at a particular point of time; determining (420) an initial model for the two-dimensional image, wherein the initial model is determined based on predetermined structural information about the movable object (101) and based on predetermined motion characteristics of the movable object (101); modifying (430) the determined initial model to obtain a final model for the two-dimensional image.

17. A computer-readable medium, in which a computer program for processing image data of a movable object (101) is stored which, when being executed by a processor, is adapted to carry out the following method steps: receiving (410) an image data set reflecting a two-dimensional image according to a particular two-dimensional proj ection of the movable obj ect ( 101 ) at a particular point of time; determining (420) an initial model for the two-dimensional image, wherein the initial model is determined based on predetermined structural information about the movable object (101) and based on predetermined motion characteristics of the movable object (101); modifying (430) the determined initial model to obtain a final model for the two-dimensional image.

18. A program element for processing image data of a movable object (101), which, when being executed by a processor, is adapted to carry out the following method steps: receiving (410) an image data set reflecting a two-dimensional image according to a particular two-dimensional projection of the movable object (101) at a . particular point of time; determining (420) an initial model for the two-dimensional image, wherein the initial model is determined based on predetermined structural information about the movable object (101) and based on predetermined motion characteristics of the movable object (101); modifying (430) the determined initial model to obtain a final model for the two-dimensional image.