(19)
(11)EP 2 410 491 B1

(12)EUROPEAN PATENT SPECIFICATION

(45)Mention of the grant of the patent:
10.09.2014 Bulletin 2014/37

(21)Application number: 11173551.0

(22)Date of filing:  12.07.2011
(51)International Patent Classification (IPC): 
G06T 11/00(2006.01)

(54)

System and method for reconstruction of X-ray images

System und Verfahren zur Rekonstruktion von Röntgenbildern

Système et procédé de reconstruction d'images à rayons X


(84)Designated Contracting States:
AL AT BE BG CH CY CZ DE DK EE ES FI FR GB GR HR HU IE IS IT LI LT LU LV MC MK MT NL NO PL PT RO RS SE SI SK SM TR

(30)Priority: 22.07.2010 US 841815

(43)Date of publication of application:
25.01.2012 Bulletin 2012/04

(73)Proprietor: GENERAL ELECTRIC COMPANY
Schenectady, NY 12345 (US)

(72)Inventors:
  • KHARE, Kedar Bhalchandra
    Niskayuna, New York 12309 (US)
  • EBERHARD, Jeffrey Wayne
    Niskayuna, New York 12309 (US)
  • CLAUS, Bernhard Erich Hermann
    Niskayuna, New York 12309 (US)

(74)Representative: Bedford, Grant Richard 
GPO Europe GE International Inc. The Ark 201 Talgarth Road Hammersmith
London W6 8BJ
London W6 8BJ (GB)


(56)References cited: : 
US-A1- 2007 297 656
  
      
    Note: Within nine months from the publication of the mention of the grant of the European patent, any person may give notice to the European Patent Office of opposition to the European patent granted. Notice of opposition shall be filed in a written reasoned statement. It shall not be deemed to have been filed until the opposition fee has been paid. (Art. 99(1) European Patent Convention).


    Description


    [0001] The subject matter disclosed herein relates generally to x-ray imaging systems, and more particularly to x-ray imaging systems that generate three-dimensional (3D) images.

    [0002] X-ray imaging systems have many different configurations and are used in many different applications. For example, known x-ray absorption based imaging systems cover a wide range of imaging geometries, such as helical or step-and-shoot type computed tomography (CT) scanners, mammography or radiology-tomography (rad-tomo) systems with single arc/line/two-dimensional (2D) source trajectories, multi-view security systems, laminography type inspection systems, among others. The helical/step-and-shoot type full CT scanners typically acquire about 1000 or more views circularly around an object (e.g., a patient or luggage) for imaging each slice of a 3D volume of the object. These scanners are suitable when accurate 3D representation of imaging volume is needed. On the other hand, mammography or rad-tomo type systems or multi-view baggage scanning systems typically involve a much smaller number of views (e.g., about 5 - 60 views) and are suitable when the application requires viewing the 3D image volume only along certain specified orientations.

    [0003] US 2007/0297656 describes a system and method for iterative image reconstruction in which sub iterations of iterations are performed on pixel subsets that may, for example, comprise neighbouring pixels.

    [0004] Accordingly, there is a wide gap in the application space between full CT scanners and other X-ray systems, due to the 3D image quality that is conventionally achievable with these different types of systems. For example, in x-ray tomosynthesis systems, planar two-dimensional (2D) or arc type source trajectories can acquire 3D depth information. However, when conventional reconstruction methods are employed for image formation, the resultant images are highly degraded. In particular, filtered backprojection type approaches often result in severe streaking in the images, and iterative approaches (even those based on using image domain priors or constraints, which can mitigate artifact levels to some extent) suffer from very slow convergence properties due to the highly incomplete nature of the data, thus resulting in images that lack high frequency details.

    [0005] Various aspects and embodiments of the present invention are thus defined in accordance with the appended claims.

    [0006] Various aspects and embodiments of the present invention will now be described in connection with the accompanying drawings, in which:

    Figure 1 is a system workflow diagram illustrating an image reconstruction process in accordance with various embodiments.

    Figure 2 is a flowchart of method for image reconstruction in accordance with various embodiments.

    Figure 3 is a flowchart of a method for automatically selecting weighting values in accordance with various embodiments.

    Figure 4 is a diagram illustrating images formed from different reconstruction processes.

    Figure 5 is a diagram showing one of the images of Figure 4 and an image formed by an image reconstruction process of one embodiment.

    Figure 6 is a diagram illustrating a source trajectory displacement method in accordance with various embodiments.

    Figure 7 is a block diagram of an imaging system formed in accordance with various embodiments.



    [0007] The foregoing summary, as well as the following detailed description of certain embodiments will be better understood when read in conjunction with the appended drawings. To the extent that the figures illustrate diagrams of the functional blocks of various embodiments, the functional blocks are not necessarily indicative of the division between hardware circuitry. Thus, for example, one or more of the functional blocks (e.g., processors or memories) may be implemented in a single piece of hardware (e.g., a general purpose signal processor or random access memory, hard disk, or the like) or multiple pieces of hardware. Similarly, the programs may be stand alone programs, may be incorporated as subroutines in an operating system, may be functions in an installed software package, and the like. It should be understood that the various embodiments are not limited to the arrangements and instrumentality shown in the drawings.

    [0008] As used herein, an element or step recited in the singular and proceeded with the word "a" or "an" should be understood as not excluding plural of said elements or steps, unless such exclusion is explicitly stated. Furthermore, references to "one embodiment" of the present invention are not intended to be interpreted as excluding the existence of additional embodiments that also incorporate the recited features. Moreover, unless explicitly stated to the contrary, embodiments "comprising" or "having" an element or a plurality of elements having a particular property may include additional such elements not having that property.

    [0009] Various embodiments provide methods and systems for image reconstruction, particularly for limited or reduced view tomographic data acquisition, which may be acquired by different x-ray imaging systems. For example, by practicing at least some embodiments, x-ray imaging systems with 10-100 times less views compared to full CT scanners may be used with reconstruction methods for enabling three-dimensional (3D) imaging applications. A technical effect of at least one embodiment is providing fast and high quality image reconstruction from limited tomographic data that enables 3D imaging with limited view tomographic systems. Additionally, increased or improved image quality for limited view x-ray imaging systems results from practicing at least some of the embodiments described herein.

    [0010] It should be noted that although the various embodiments may be described in connection with a particular imaging system, such as an x-ray imaging system for diagnostic imaging, the various embodiments are not limited to a particular imaging system or one having a particular configuration. For example, the various embodiments may be implemented in connection with non-medical, security and inspection systems, among others. Additionally, the various embodiments may include other projection based imaging modalities such as Positron Emission Tomography (PET), Single Photon Emission Computed Tomography (SPECT), among others.

    [0011] In accordance with various embodiments, optimal gradient type methods for iterative image reconstruction may be combined with prior information (e.g., prior iteration reconstruction solutions) that represents additional constraints on the image (e.g. smoothness, positivity, etc.). For example, two or more prior solutions or iterations may be used for a subsequent iterative update step in the reconstruction process.

    [0012] Various embodiments are used to generate 3D images in a CT imaging system wherein a limited or reduced dataset is acquired by the x-ray imaging system. For example, a reconstruction process 30 is illustrated by the system workflow diagram shown in Figure 1. The process 30 includes acquiring limited or reduced tomographic data 32, which is acquired by an x-ray imaging system 34. In some embodiments, the limited or reduced tomographic data 32 includes a dataset having a limited or reduced number of image or projection views 36, such as acquired by employing various source and/or detector trajectories, for example, source and detector motion in a circle, a fixed detector with a source moving in a straight line, an arc or other more complex curves, among others and combinations thereof. For example, in some embodiments, the dataset includes between about 50 and about 100 image views. However, other numbers of views may be acquired and used, such as above about 30 views to less than about 1000 views. Accordingly, in some embodiments, and merely for illustration, the x-ray imaging system 34 acquires more views than about 1-30, but less than 1000. Thus, the x-ray imaging system 34 may be configured as a low dose x-ray imaging system for 3D imaging that acquires, in some embodiments, more image views than, for example, an x-ray radiography system, a mammography system, a digital tomosynthesis system, a low cost small footprint system, among others. The source and detector trajectories may be selected as appropriate (e.g., as desired or needed) for obtaining or capturing object information based on a particular system application and constraints. For example, source and detector trajectories may be selected to optimize the object information being obtained based on the system application or constraints to capture the best (e.g., most clinically relevant) information. In these embodiments, the x-ray imaging system 34 also acquires less image views than a full CT scanner providing a complete 3D information dataset, which acquires 1000 or more views.

    [0013] Conventional image compression standards (e.g., Joint Photographic Experts Group (JPEG) compression) illustrate that natural images are highly redundant in the usual pixel/voxel basis representations and have sparse representations in transform domains (e.g., in wavelet/gradient/other transform domains). In accordance with various embodiments, less than the conventional amount or number of data / views is used by imposing transform-domain sparsity constraints on the 3D image during the reconstruction process.

    [0014] The process 30 includes using the limited or reduced tomographic data 32 acquired by the x-ray imaging system 34 in an iterative reconstruction process 38. The iterative reconstruction process 38 operates to reconstruct images wherein N pixels are reconstructed based on less than N measurements, such as less than the number of views used in a conventional CT scanner (e.g., full CT scanner), while providing high quality images that include diagnostically relevant information. The iterative reconstruction process 38 in some embodiments is performed to minimize a cost function that represents the data consistency and an additional set of priors (e.g., prior solutions) representing constraints on the solution or convergence of the iterative reconstruction. The iterative reconstruction process 38 is performed using a reduced number of iterations and uses update information 40 in an iterative update step that includes two or more prior solutions or iteration information, as well as automatic weighting for a cost function of the reconstruction process. For example, in some embodiments, an optimal gradient method may be used as described in more detail herein, wherein two prior iterations are used for updating the starting point for the next iteration.

    [0015] The iterative reconstruction process 38 is performed until a convergence is reached for reconstructing a 3D image 42, which may be displayed to a user or operator of the x-ray imaging system 34. In various embodiments, the determination of when convergence is reached is based on when an accuracy level criterion or threshold is satisfied, for example, a 1%, 2%, etc. level of accuracy (e.g., residual error) based on the acquired data, thereby defining a cost function minimization.

    [0016] The iterative reconstruction process 38 is performed in various embodiments to converge to a solution that is within an acceptable residual error level. The iterative reconstruction process 38 provides convergence such that images with high frequency details are obtained using a same number of iterations that would otherwise result in blurred images. The iterative reconstruction process 38 essentially solves a reconstruction problem defined by a linear system of equations as follows:



    [0017] In Equation 1, Y is the Sub-Nyquist sampled data, A is a system operator (forward projection from image to data space) and f is the image to be reconstructed. The iterative reconstruction process 38 may perform an iterative convergence method to find the least squared solutions for Equation 1, wherein Y represents a dataset having less data (e.g., a limited or reduced number of image or projection views) than would be needed to perform image reconstruction. In particular, a minimization problem may be defined as cost function C(f) as follows:



    [0018] In Equation 2, the || y - A f ||2 term represents data consistency, which is to be reduced or minimized and the λ1||Wf||1 + λ2 ||∇f ||1 terms represent sparsity transformations, which are additional constraints on the image (e.g. attenuation coefficients, smoothness, positivity, etc.). In this Equation 2, λ1, λ2 are weighting or penalty factors, Wf represents a wavelet transform of the image and ∇f represents the gradient of the image so that the corresponding term in Equation 2 represents the total variation of the image f. The terminology ||...||1 refers to the L1-norm (sum of absolute values of components) of the quantity inside and || ... ||2 refers to the L2-norm (sum of absolute value squares of components). As described in more detail herein, the λ1, λ2 terms are selected automatically and determines the significance or value of the terms to data consistency.

    [0019] Referring now to Figure 2, a method 50 for image reconstruction is shown, which may be performed, for example, during the iterative reconstruction process 38. The method 50 generally includes obtaining limited view data at 52, which may be, for example, the limited or reduced tomographic data 32 that includes the dataset having a limited or reduced number of image or projection views 36 (both show in Figure 1). For example, in some embodiments, a dataset having between about 50 and 100 views or projections is used by the method 50. Thus, the acquired data includes a plurality of views or projections generated by the x-ray imaging system. Additionally, information is received from the x-ray imaging system or user, such as information regarding the parameters used for the acquisition. Such information may include, for example, patient information, calibration data, and image acquisition timing data, among other information.

    [0020] Thereafter, an iterative process is initialized at 54 using two of more starting points, for example, two or more initial values or approximations (which may be arbitrarily selected using, for example, backprojection of the limited data to image space and by applying further processing, such as smoothing and/or filtering) and then at 56 a linear combination is formed based on the last two solutions. As the iterative method 50 continues, the last two solutions are defined by the prior two iterations of the method 50 (or by the initial values if the process is in the first iteration). The initial values may be initial voxel or pixel values for a volume of interest. Thus, two prior iterations are used for updating the next approximation or reconstruction iteration. The linear combination provides a solution for the data consistency term. For example, in some embodiments, a Nesterov optimal gradient algorithm is used for the data consistency term and is defined as follows:





    [0021] In Equations 3 and 4, fn is the prior approximation at the nth iteration, gn is the linear combination of fn and fn-1 , and other like terms have similar definitions to the terms in Equations 1 and 2. A and AT represent the forward projection (3D image space to projection space) and the backprojection (projection to the 3D image space), respectively, and τ is a scaling constant (e.g., as appropriate according to system configuration).

    [0022] Thereafter, sparsity transformations and constraints are applied at 58. For example, sparsity transformations and constraints as described in more detail above in connection with Equation 2 are added to the cost function minimization equation. The result is then forward projected and subtracted from the data (as defmed by the dataset at 52) to obtain a residual error at 60. The forward projection is represented in Equation 3 by A gn and the subtraction is represented by the (y - A gn) term. The term AT (y-Agn) represents the backprojection of the residual (y - A gn) from the projection to the 3D image space.

    [0023] A determination is then made at 62 as to whether a residual error is within an acceptable level, for example, smaller or less than a predetermined threshold accuracy percentage. The threshold value may be, for example, a 1%, 2%, 3%, etc. accuracy level. If the residual error is within the acceptable level, namely lower than the threshold level, then the method 50 stops at 64 as the solution, namely that the voxel values for the reconstructed image have converged to an acceptable solution. The reconstructed image then may be displayed.

    [0024] However, if at 62, the residual error is not within the acceptable level, such that the value is above the threshold value, then the process continues through another iteration, and in particular, at 66 a backprojection of the residual data is performed, for example, backprojecting to the voxels in the volume that were mapped to a pixel in the forward projection. For example, the residual data is backprojected to the image space (voxels) and added to the current iteration solution. In particular, this data is combined with two previous iterations or solutions, which may include an initial value if the method 50 is not past the second iteration. Thus, a linear combination is again formed at 56 based on at least the last two iteration solutions. It should be noted that more than two previous solutions may be used.

    [0025] It also should be noted that the process includes reducing or limiting artifacts, noise, etc. using weighting, which may include applying a penalty weighting when forming the linear combination and as described in more detail herein. The determination of the weighting value is performed automatically. For example, the artifact or noise may be estimated at any point in the method 50 and the weighting value determined based on the size of the artifact or noise. The weighting value is then used in the next step of the method 50 or in the next iteration.

    [0026] In various embodiments, the penalty weights are automatically selected. In particular, the values for λ1 , λ2 etc. are automatically selected, for example, in a data driven manner as illustrated in the method 70 shown in Figure 3. The cost function or equation described herein is parameterized using the values for λ1 , λ2 etc., with each value generating a different solution.

    [0027] In particular, an intermediate solution is determined at 72. The intermediate solution is formed by the linear combination of the two previous solutions, for example, which is the result of step 56 in Figure 2. Thus, the intermediate solution is the linear combination of the previous two solutions at each iteration of the method 50 (shown in Figure 2). Accordingly, an intermediate solution is generated at each iteration of the method 50. It should be noted that the method 70 in various embodiments is performed on full image data or data for local regions of an image. Additionally, the method may be performed at different resolution scales.

    [0028] After obtaining the intermediate solution, a gradient, wavelet or other sparsifying transform is performed at 74. For example, the wavelets or sparsity transforms may be added to the cost function illustrated in Equation 2. Thereafter, a statistical spread, such as a determination of a standard deviation (sigma value) is determined at 76 using any suitable method. For example, the statistical spread of the transform coefficients are determined using the standard deviation or median-absolute deviation. Thereafter, the weighting values or parameters are set at 78 based on the statistical spread. It should be noted that the variance (sigma) in various embodiments is about constant for each of the weighting parameters. Also, it should be noted that in various embodiments, when sigma is constant, the weighting value is selected as a smaller number, such as in the order of 1-2.

    [0029] Thus, in accordance with various embodiments, a converging iterative reconstruction process is provided that uses two prior solutions or iterations for calculating the value(s) for the next iteration. For example, each iteration of the image reconstruction process includes a linear combination of two previous solutions or results from two previous iterations. Accordingly, in various embodiments, one iteration of a reconstruction process may be defined as follows:





    [0030] It should be noted that like variables are defined similarly unless otherwise noted. In Equation 5, P represents the sparsification operation, which may include, for example, thresholding in the image or other transform domain (e.g. wavelets), total variation minimization, other constraints based on prior information (e.g., positivity, smoothness requirements, etc.), filtering operations in the data and/or image domain, other thresholding operations, etc. Thus, P is a term that includes the sparsity transforms, such as image transforms, etc.

    [0031] Thus, various embodiments employ optimal gradient type methods for iterative image reconstruction combined with at least two prior solutions or values that represent additional constraints on the image (e.g. smoothness, positivity, etc.). The iterative reconstruction methods can provide 3D imaging with limited view x-ray systems. The image reconstruction for limited view tomographic data systems, thus, involves solving an incomplete system of linear equations as set forth in Equation 1 (y = Af), where again, γ is the projection data, A is the system operator, and f is the image to be determined.

    [0032] The iteration performed in various embodiments includes the use of two or more prior solutions for the update step. In various embodiments, an iteration is thus performed as follows: fn+1 = P[ gn + tAT (y - Agn ) ] , where gn = a fn + b fn-1 for constants a and b that typically depend on the iteration number. In particular, a and b may be selected according to Nesterov's optimal gradient method. The L2-norm convergence for such methods is known to be about O(1/n2) . Thus, a typical number of iterations in order to obtain L2-norm error of E is about 1/ sqrt(E), representing a reduction in the number of iterations for convergence for limited view X-ray imaging applications. Various embodiments use more than one prior image solution (instead of only one) for the iterative update, which provides a convergence rate such that higher quality images are generated with the same number of iterations that would otherwise result in lower quality images (e.g., blurred images).

    [0033] Thus, as illustrated in Figure 4, a dataset, illustrated as a sinogram dataset (having 1000 views) represented by the image 80 may be used to reconstruct an image 82, which in this embodiment is an imaged pelvic region. The reconstruction method used is a filtered backprojection reconstruction method. As can be seen in the reconstructed image 84, which is reconstructed using the iterative process described herein, an image 84 of similar quality to the image 82 may be generated with less data, for example, a sinogram dataset having 250 views as illustrated by the image 86.

    [0034] With reference now to Figure 5, various embodiments of reconstruction method described herein provide improved image quality as illustrated by the image 88, where eight iterations were performed and is compared to the image 84 reconstructed with an iterative process that uses only one prior solution (as opposed to a linear combination of multiple prior solutions) for update. Thus, as can be seen, images with more clinically relevant high-frequency detail information may be generated not only using less views, but with fewer iterations. In some embodiments, a 3D image formed from the limited tomographic dataset has an image quality less than and within an image margin of a full CT scanner image quality. For example, the image margin may be a variance or difference, such as 1%, 2%, 3%, etc. of the image quality of a full CT scanner image quality.

    [0035] It should be noted that variations and modifications are contemplated. For example, as illustrated in Figure 6, a modification in source trajectory, for example, the trajectory of an x-ray source may be provided. In particular, a one-dimensional (1D) trajectory 90 and a 2D trajectory 100 are illustrated as a plurality of trajectory points 92 and 102, respectively. The trajectories 90 and 100 correspond to a location 110 on an area 112 of a detector. In various embodiments, at least some, for example, every third trajectory point 104 of the 2D trajectory 100 is shifted laterally, which may include an alternating shifting pattern as shown. For example, the source is displaced in the lateral direction to obtain more image information, thereby allowing new trajectories. As can be seen, the images 106 and 108 (especially image 108) generated using the 2D trajectory 100 have greater depth resolution than the images 96 and 98 generated using the 1D trajectory 90. The images 96 and 106 show a horizontal slice through the object that is parallel to the detector and the images 98 and 108 show coronal slices (slices perpendicular to the detector) through the object.

    [0036] The various embodiments, including the reconstruction methods may be implemented in connection with any type of x-ray imaging system. For example, as shown in Figure 7, an imaging system 120, which is illustrated as an x-ray system can be used to acquire image views or projections for use in image reconstruction in accordance with various embodiments. The imaging system 120 generally includes an x-ray detector 122 having an array of detector cells 124 defining a scan area, and an x-ray source 126. An object 128, such as a patient is positioned between the x-ray source 126 and the x-ray detector 122, which may be one or more detectors or detector modules. The imaging system 120 also includes a data acquisition system 130 with readout electronics 132.

    [0037] In one embodiment, the x-ray detector(s) 122 may be flat-panel detector systems such as an amorphous silicon flat panel detector or other type of digital x-ray image detector. In another embodiment, the x-ray detector(s) 122 may include a scintillator having a screen that is positioned in front of the x-ray detector(s) 122.

    [0038] It should be noted that the imaging system 120 may be implemented as a non-mobile or mobile imaging system. Moreover, the imaging system 120 may be provided in different configurations. For example, the image data may be generated with the x-ray source 126 at discrete foci along an arc that completely surrounds the object 128 to generate the image information using CT procedures and processes. In another embodiment, the image data is generated with the x-ray source 126 at discrete foci along a smaller arc above the object to generate the image information using computed tomosynthesis procedures and processes. In other embodiments, the x-ray source 126 and the x-ray detector 122 are both mounted at opposite ends of a gantry 134, which may be a C-arm that rotates about the object 128. The rotatable C-arm is a support structure that allows rotating the x-ray source 126 and the x-ray detector 122 around the object 128 along a substantially circular arc, to acquire a plurality of projection images of the object 128 at different angles (e.g., different views or projections). It should be appreciated that various embodiments of image reconstruction may be provided in connection with a plurality of measured projections or image views regardless of the implementation of the apparatus or system that generates the projections or image data. It also should be appreciated that the various embodiments of image reconstruction may be provided in connection with a combination of two or more of the above described source/detector trajectories.

    [0039] In operation, the object 128 is positioned in the imaging system 120 for performing an imaging scan. For example, the x-ray source 126 may be positioned above, below or around the object 128. For example, the x-ray source 126 (and the x-ray detector(s) 122) may be moved between different positions around the object 128 using the gantry 134. X-rays are transmitted from the x-ray source 126 through the object 128 to the x-ray detector(s) 122, which detect x-rays impinging thereon.

    [0040] The readout electronics 132 may include a reference and regulation board (RRB) or other data collection unit. The RRB may accommodate and connect data modules to transfer data (e.g., a plurality of views or projections) from the x-ray detector(s) 122 to the data acquisition system 130. Thus, the readout electronics 132 transmit the data from the x-ray detector(s) 122 to the data acquisition system 130. The data acquisition system 130 forms an image from the data and may store, display on a display 133, and/or transmit the image. For example, the various embodiments may include an image reconstruction module 136, which may be implemented in hardware, software, or a combination thereof, that allows the data acquisition system to reconstruct images using limited tomographic data acquired from the x-ray detector(s) 122 and as described in more detail herein.

    [0041] Thus, various embodiments provide iterative image reconstruction methods that form images, particularly 3D images, based on data having a lower number of image views or projections, for example, compared to a conventional CT scanner. Various embodiments also provide a fast converging method and apply to multiple system geometry configurations. Additionally, various embodiments include automated data driven reconstruction methods.

    [0042] Some embodiments provide a tangible and non-transitory machine-readable medium or media having instructions recorded thereon for a processor or computer to operate an imaging apparatus to perform one or more embodiments of the methods described herein, for example, to implement the processes of the image reconstruction module. The medium or media may be any type of CD-ROM, DVD, floppy disk, hard disk, optical disk, flash RAM drive, or other type of computer-readable medium or a combination thereof.

    [0043] The various embodiments and/or components, for example, the processors, or modules, components and controllers therein, also may be implemented as part of one or more computers or processors. The computer or processor may include a computing device, an input device, a display unit and an interface, for example, for accessing the Internet. The computer or processor may include a microprocessor. The microprocessor may be connected to a communication bus. The computer or processor may also include a memory. The memory may include Random Access Memory (RAM) and Read Only Memory (ROM). The computer or processor further may include a storage device, which may be a hard disk drive or a removable storage drive such as a floppy disk drive, optical disk drive, and the like. The storage device may also be other similar means for loading computer programs or other instructions into the computer or processor.

    [0044] As used herein, the term "computer" may include any processor-based or microprocessor-based system including systems using microcontrollers, reduced instruction set computers (RISC), application specific integrated circuits (ASICs), logic circuits, and any other circuit or processor capable of executing the functions described herein. The above examples are exemplary only, and are thus not intended to limit in any way the definition and/or meaning of the term "computer".

    [0045] The computer or processor executes a set of instructions that are stored in one or more storage elements, in order to process input data. The storage elements may also store data or other information as desired or needed. The storage element may be in the form of an information source or a physical memory element within a processing machine.

    [0046] The set of instructions may include various commands that instruct the computer or processor as a processing machine to perform specific operations such as the methods and processes of the various embodiments of the invention. The set of instructions may be in the form of a software program. The software may be in various forms such as system software or application software and which may be embodied as a tangible and non-transitory computer readable medium. Further, the software may be in the form of a collection of separate programs, a program module within a larger program or a portion of a program module. The software also may include modular programming in the form of object-oriented programming. The processing of input data by the processing machine may be in response to user commands, or in response to results of previous processing, or in response to a request made by another processing machine.

    [0047] As used herein, the terms "software" and "firmware" are interchangeable, and include any computer program stored in memory for execution by a computer, including RAM memory, ROM memory, EPROM memory, EEPROM memory, and nonvolatile RAM (NVRAM) memory. The above memory types are exemplary only, and are thus not limiting as to the types of memory usable for storage of a computer program.

    [0048] It is to be understood that the above description is intended to be illustrative, and not restrictive. For example, the above-described embodiments (and/or aspects thereof) may be used in combination with each other. In addition, many modifications may be made to adapt a particular situation or material to the teachings of the invention without departing from its scope. While the dimensions and types of materials described herein are intended to define the parameters of the invention, they are by no means limiting and are exemplary embodiments. Many other embodiments will be apparent to those of skill in the art upon reviewing the above description. The scope of the invention should, therefore, be determined with reference to the appended claims, along with the full scope of equivalents to which such claims are entitled. In the appended claims, the terms "including" and "in which" are used as the plain-English equivalents of the respective terms "comprising" and "wherein." Moreover, in the following claims, the terms "first," "second," and "third," etc. are used merely as labels, and are not intended to impose numerical requirements on their objects. This written description uses examples to disclose the invention, including the preferred mode, and also to enable any person skilled in the art to practice the invention, including making and using any devices or systems and performing any incorporated methods. The patentable scope of the invention is defined by the claims, and may include other examples that occur to those skilled in the art. Such other examples are intended to be within the scope of the claims if they have structural elements that do not differ from the literal language of the claims, or if they include equivalent structural elements with insubstantial differences from the literal languages of the claims.


    Claims

    1. A method (50) for reconstructing images, the method comprising:

    acquiring (52) a plurality of image views using an x-ray imaging system, the plurality of image views defining a limited tomographic dataset;

    initializing (54) an iterative process by selecting two or more image views;

    performing (56, 58, 60, 62) three-dimensional (3D) image reconstruction using the selected image views in an iterative reconstruction (38), wherein the iterative reconstruction (38) includes forming (56) a linear combination based on a plurality of previous iteration results per iteration, in order to reconstruct images wherein a number N pixels are reconstructed based on less than N measurements; the image reconstruction comprising

    applying (58) sparsity transformations and constraints to the current iteration result;

    forward projecting (60) the result and subtracting the result from the acquired image view data to obtain a residual error;

    determining (62) whether the residual error is within an acceptable level, and
    if the residual error is within the acceptable level

    displaying (64) an image based on the image reconstruction, the image including clinically relevant high-frequency detail information; and if the residual error is not within the acceptable level

    backprojecting (66) the residual data to the image space and adding it to the current iteration solution; the method further comprising

    using at least two prior iteration reconstruction results to update a convergence for the image reconstruction; and

    using an optimal gradient process for the iterative reconstruction to minimize a data consistency term for a cost function used in the iterative reconstruction, wherein
    the iterative reconstruction includes the cost function minimization to a threshold level and wherein a value for a penalty weight for the cost function is automatically determined based on acquired image data including the plurality of image views.


     
    2. A method (50) in accordance with claim 1, wherein the automatic determination comprises using a statistical spread to set the weighting value.
     
    3. A method (50) in accordance with any preceding claim, wherein each iteration of the iterative reconstruction is defined by:




    wherein P is a constraint operation, y is projection data corresponding to the projection views, A is a system operator, AT is the backprojection operator and f is the image to be displayed.
     
    4. A method (50) in accordance with any preceding claim, wherein the limited tomographic dataset comprises between about 50 image views and about 100 image views.
     
    5. A method (50) in accordance with any preceding claim, wherein the 3D image formed from the limited tomographic dataset has an image quality less than and within an image margin of a full computed tomography (CT) scanner image quality.
     
    6. A method (50) in accordance with any preceding claim, wherein the imaging system acquiring the plurality of image views operates an x-ray source along one of a two-dimensional trajectory having laterally displaced trajectory points a circle, spiral, multiple line and sideways to an object trajectory.
     


    Ansprüche

    1. Verfahren (50) zum Rekonstruieren von Bildern, wobei das Verfahren Folgendes umfasst:

    Erfassen (52) mehrerer Bildansichten unter Verwendung eines Röntgensystems, wobei die mehreren Bildansichten einen begrenzten tomografischen Datensatz definieren;

    Initialisieren (54) eines iterativen Prozesses durch Auswählen von zwei oder mehr Bildansichten;

    Ausführen (56, 58, 60, 62) der dreidimensionalen (3D)-Bildrekonstruktion unter Verwendung der ausgewählten Bildansichten in einer iterativen Rekonstruktion (38),

    wobei die iterative Rekonstruktion (38) das Bilden (56) einer linearen Kombination auf der Basis mehrerer vorheriger Iterationsergebnisse per Iteration umfasst,

    um Bilder zu rekonstruieren, wobei eine Zahl N Pixel auf der Basis von weniger als N Messungen rekonstruiert wird; wobei die Bildrekonstruktion Folgendes umfasst

    Anwenden (58) von Seltenheitstransformationen und

    Nebenbedingungen auf das aktuelle Iterationsergebnis;

    Fortschreiben (60) des Ergebnisses und Subtrahieren des Ergebnisses von den erfassten Bildansichtsdaten,

    um einen Restfehler zu erhalten;

    Bestimmen (62), ob der Restfehler innerhalb eines akzeptablen Niveaus liegt und ob der Restfehler innerhalb dieses akzeptablen Niveaus liegt;

    Anzeigen (64) eines Bildes auf der Basis der Bildrekonstruktion, wobei das Bild klinisch relevante hochfrequente Detailinformationen enthält; und ob der Restfehler nicht innerhalb des akzeptablen Niveaus liegt;

    Rückprojizieren (66) der Restdaten in den Bildraum und

    Hinzufügen zur aktuellen Iterationslösung; wobei das Verfahren ferner umfasst

    Verwenden von mindestens zwei vorherigen Iterationsrekonstruktionsergebnissen, um eine Konvergenz für die Bildrekonstruktion zu aktualisieren; und

    Verwenden eines optimalen Gradientenverfahrens für die iterative Rekonstruktion, um einen Datenkonsistenzterm für eine Kostenfunktion, die in der iterativen Rekonstruktion verwendet wird, zu minimieren, wobei

    die iterative Rekonstruktion die Minimierung der Kostenfunktion bis zu einem Schwellwertniveau umfasst und wobei ein Wert für eine Strafgewichtung für die Kostenfunktion automatisch auf der Basis von erfassten Bilddaten, die die mehreren Bildansichten enthalten, bestimmt wird.


     
    2. Verfahren (50) nach Anspruch 1, wobei die automatische Bestimmung das Verwenden einer statistischen Streuung umfasst, um den Gewichtungswert festzulegen.
     
    3. Verfahren (50) nach einem der vorherigen Ansprüche, wobei jede Iteration der iterativen Rekonstruktion durch Folgendes definiert ist:




    wobei P eine Restriktionsoperation ist, y Projektionsdaten sind, die den Projektionsansichten entsprechen, A ein Systemoperator ist, AT der Rückprojektionsoperator ist und f das Bild ist, das angezeigt werden soll.
     
    4. Verfahren (50) nach einem der vorherigen Ansprüche, wobei der begrenzte tomografische Datensatz zwischen etwa 50 Bildansichten und etwa 100 Bildansichten umfasst.
     
    5. Verfahren (50) nach einem der vorherigen Ansprüche, wobei das 3D-Bild, das aus dem begrenzten tomografischen Datensatz gebildet wird, eine Bildqualität hat, die kleiner als die Scannerbildqualität ist und innerhalb eines Bildfehlerraums einer vollständig berechneten Tomografie (CT) liegt.
     
    6. Verfahren (50) nach einem der vorherigen Ansprüche, wobei das Abbildungssystem, das die mehreren Bildansichten erfasst, eine Röntgenstrahlenquelle entlang einer aus einer zweidimensionalen Bahn, die seitlich versetzte Bahnpunkte hat, eines Kreises, einer Spirale, mehrerer Linien und seitwärts von einer Objektbahn steuert.
     


    Revendications

    1. Une méthode (50) pour reconstruire des images, la méthode comprenant :

    l'acquisition (52) d'une pluralité de vues d'image en utilisant un système d'imagerie à rayons X, la pluralité de vues d'image définissant un ensemble limité de données tomographiques ;

    l'initialisation (54) d'un processus itératif en sélectionnant deux ou plusieurs vues d'image ;

    l'exécution (56, 58, 60, 62) d'une reconstruction d'image tridimensionnelle (3D) en utilisant les vues d'image sélectionnées dans une reconstruction itérative (38), dans laquelle la reconstruction itérative (38) inclut la formation (56) d'une combinaison linéaire basée sur une pluralité de résultats itératifs précédents par itération, afin de reconstruire des images dans lesquelles un nombre N de pixels sont reconstruits en se basant sur moins de N mesures ; la reconstruction d'image comprenant
    l'application (58) de transformations et de contraintes de rareté au résultat itératif courant ;

    la projection directe (60) du résultat et la soustraction du résultat depuis les données de vue d'image acquises pour obtenir une erreur résiduelle ;

    la détermination (62) si l'erreur résiduelle est à l'intérieur d'un niveau acceptable, et si l'erreur résiduelle est à l'intérieur du niveau acceptable

    l'affichage (64) d'une image sur la base de la reconstruction d'image, l'image incluant des informations de détail haute fréquence cliniquement appropriées ; et si l'erreur résiduelle n'est pas à l'intérieur du niveau acceptable

    la rétroprojection (66) des données résiduelles vers l'espace d'image et l'addition de celles-ci à la solution itérative courante ; la méthode comprenant en outre

    l'utilisation d'au moins deux résultats antérieurs de reconstruction par itération pour mettre à jour une convergence pour la reconstruction d'image ; et

    l'utilisation d'un processus de gradient optimal pour la reconstruction itérative pour minimiser un terme de cohérence de données pour une fonction de coût utilisée dans la reconstruction itérative, dans laquelle la reconstruction itérative inclut la minimisation de fonction de coût à un niveau de seuil et dans laquelle une valeur pour un poids de pénalité pour la fonction de coût est automatiquement déterminée sur la base des données d'image acquises incluant la pluralité de vues d'image.


     
    2. Une méthode (50) selon la revendication 1, dans laquelle la détermination automatique comprend l'utilisation d'une diffusion statistique pour fixer la valeur de pondération.
     
    3. Une méthode (50) selon une quelconque revendication précédente, dans laquelle chaque itération de la reconstruction itérative est définie par :




    dans lesquelles P est une opération de contrainte, y est une donnée de projection correspondant aux vues de projection, A est un opérateur de système, AT est l'opérateur de rétroprojection et f est l'image à afficher.
     
    4. Une méthode (50) selon une quelconque revendication précédente, dans laquelle l'ensemble limité de données tomographiques comprend entre environ 50 vues d'image et environ 100 vues d'image.
     
    5. Une méthode (50) selon une quelconque revendication précédente, dans laquelle l'image 3D formée à partir de l'ensemble limité de données tomographiques a une qualité d'image inférieure et à l'intérieur d'une marge d'image d'une pleine qualité d'image de tomodensitomètre (CT).
     
    6. Une méthode (50) selon une quelconque revendication précédente, dans laquelle le système d'imagerie acquérant la pluralité de vues d'image met en oeuvre une source de rayons X le long de l'un d'une trajectoire bidimensionnelle ayant des points de trajectoire latéralement déplacés, d'un cercle, d'une spirale, d'une ligne multiple et de chemins latéraux à une trajectoire d'objet.
     




    Drawing


























    Cited references

    REFERENCES CITED IN THE DESCRIPTION



    This list of references cited by the applicant is for the reader's convenience only. It does not form part of the European patent document. Even though great care has been taken in compiling the references, errors or omissions cannot be excluded and the EPO disclaims all liability in this regard.

    Patent documents cited in the description