Method and apparatus for scattered radiation estimation in computed tomography imaging systems, and information processing apparatus

Abstract

To provide scattered radiation estimation and correction having higher accuracy and shorter processing time than the prior art.SOLUTION: A method for scattered radiation estimation in a CT including a detector having multiple detector pixels includes: obtaining projection data by scanning an imaging object; reconstructing image data from the projection data; estimating, on the basis of the projection data, a first scattered radiation distribution; selecting, on the basis of the first scattered radiation distribution, a first subset of the pixels; calculating, on the basis of the projection data and the image data, a second scattered radiation distribution with respect to the selected first subset, the second scattered radiation distribution having higher accuracy than the first scattered radiation distribution; and acquiring, on the basis of the second scattered radiation distribution, a third scattered radiation distribution with respect to a second subset of the pixels, the third scattered radiation distribution having higher spatial resolution than the second scattered radiation distribution.SELECTED DRAWING: Figure 3

Description

[Technical field]

[0001] The present disclosure relates to scatter estimation and correction in computed tomography (CT) imaging systems, where estimation is performed using a cascaded neural network framework to increase accuracy and reduce processing time. [Background technology]

[0002] The background discussion herein is intended to generally indicate the context of the present disclosure. The work of the presently named inventors is not admitted expressly or impliedly as prior art to the present disclosure, to the extent that the work is described in the background section, as well as aspects of the description that do not qualify as prior art at the time of filing.

[0003] Generally, X-ray projection images contain a lot of scatter. Insufficient scatter correction tends to degrade and bias the quality of the projections as well as the reconstructed CT images. This effect leads to inaccurate CT numbers, low image contrast, and artifacts. In cone-beam CT imaging systems with wide beam geometry, scatter correction may be more important to maintain the diagnostic value of the CT examination.

[0004] Basically, there are two different approaches to dealing with X-ray scatter in CT imaging: scatter suppression and scatter estimation. For example, 2D detectors such as flat panel detectors deployed in CT machines can use hardware-based anti-scatter measures (e.g., anti-scatter grids, air gaps, etc.) to suppress scatter. Another method to correct for scatter is to estimate the scatter signal in the measurement data and remove it using software-based post-processing.

[0005] For example, deep learning-based methods have shown unique performance in CT scatter estimation and correction, providing higher accuracy in scatter estimation compared to traditional kernel methods and faster processing speed compared to physical model-based methods such as Monte Carlo and radiative transfer equation (RTE) methods. Summary of the Invention [Problem to be solved by the invention]

[0006] However, typical deep learning methods directly use the measured projection data as input to a trained convolutional neural network to estimate the scatter distribution. Although the projection data can be used to reconstruct images that reveal the structure of the imaging object, it is not suitable for obtaining information that represents the distribution of scatter. As a result, the accuracy of the scatter estimation and correction is insufficient.

[0007] The problem to be solved by the present disclosure is to improve accuracy and reduce processing time in scattered radiation estimation and correction, as compared to the conventional techniques.

[0008] The present disclosure relates to a method for performing scatter estimation in a computed tomography (CT) system, the CT system including a detector having a plurality of detector pixels. The method includes acquiring projection data by scanning an imaging object with the CT system, and reconstructing image data from the acquired projection data. The method also includes estimating a first scatter distribution based on the acquired projection data. The method further includes selecting a first subset of the plurality of detector pixels based on the estimated first scatter distribution. The method further includes calculating a second scatter distribution for the selected first subset of the plurality of detector pixels based on the acquired projection data and the reconstructed image data. The calculated second scatter distribution is more accurate than the estimated first scatter distribution. The method further includes acquiring a third scatter distribution for a second subset of the plurality of detector pixels based on the calculated second scatter distribution, the third scatter distribution having a higher spatial resolution than the calculated second scatter distribution. The second subset includes the first subset.

[0009] The present disclosure further relates to an apparatus for performing scatter estimation in a computed tomography (CT) system. The CT system includes a detector having a plurality of detector pixels. The apparatus includes a processing circuit for acquiring projection data by scanning an imaging object with the CT system and for reconstructing image data from the acquired projection data. The processing circuit is also for estimating a first scatter distribution based on the acquired projection data. The processing circuit is further for selecting a first subset of the plurality of detector pixels based on the estimated first scatter distribution. The processing circuit is further for calculating a second scatter distribution for the selected first subset of the plurality of detector pixels based on the acquired projection data and the reconstructed image data. The calculated second scatter distribution is more accurate than the estimated first scatter distribution. The processing circuit is further for acquiring a third scatter distribution for the second subset of the plurality of detector pixels based on the calculated second scatter distribution, the third scatter distribution having a higher spatial resolution than the calculated second scatter distribution. The second subset includes the first subset.

[0010] It is noted that this Summary section does not identify all of the embodiments and / or additional novel aspects of the disclosed or claimed invention. The Summary of the Invention merely provides a preliminary discussion of various embodiments and corresponding novel points. For additional details and / or possible aspects of the disclosure and embodiments, read the Detailed Description section and corresponding drawings of the present disclosure, which are further discussed below. [Brief description of the drawings]

[0011] Various embodiments of the present disclosure, proposed by way of example, will now be described in detail with reference to the following drawings, in which like reference numerals refer to like elements, and in which: [Figure 1] FIG. 1 shows an exemplary process for directing x-rays that are attenuated and scattered through an imaged object to a detector where both the primary flux and the scattered flux are detected. [Diagram 2]FIG. 2 illustrates a non-limiting example block diagram of a scatter correction circuit for estimating scatter distribution with improved accuracy in accordance with one embodiment of the present disclosure. [Diagram 3] FIG. 3 illustrates a non-limiting example flow chart of a scatter correction process for estimating scatter distribution with improved accuracy in accordance with one embodiment of the present disclosure. [Figure 4] FIG. 4 illustrates a non-limiting example block diagram of a scatter correction circuit for estimating scatter distribution with improved accuracy and reduced processing time in accordance with one embodiment of the present disclosure. [Diagram 5] FIG. 5 illustrates a non-limiting example flow chart of a scatter correction process for estimating scatter distribution with improved accuracy and reduced processing time in accordance with one embodiment of the present disclosure. [Figure 6A] FIG. 6A illustrates a non-limiting example of selecting a subset of detector pixels included in a detector of a CT device, according to one embodiment of the present disclosure. [Figure 6B] FIG. 6B illustrates a non-limiting example of selecting a subset of detector pixels included in a detector of a CT device, according to one embodiment of the present disclosure. [Figure 7] FIG. 7 illustrates a schematic block diagram of a CT device, according to one embodiment of the present disclosure. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0012] In the following disclosure, many different embodiments or examples are provided for implementing different features of the given subject matter. Specific examples of components and arrangements that simplify the disclosure are described below. These are, of course, merely examples and are not intended to be limiting.

[0013] For example, the order in which different steps are discussed herein is presented for clarity. In general, these steps can be performed in any suitable order. In addition, although different features, techniques, configurations, etc. may each be discussed in different locations in this disclosure, it is intended that each concept can be practiced independently of one another or in combination with one another. Thus, the disclosure can be embodied and viewed in many different ways.

[0014] Also, as used herein, the words "a," "an," and the like generally mean "one or more," unless otherwise noted.

[0015] FIG. 1 shows an example of how X-rays travel through an imaging object while being attenuated and scattered to a detector, where both primary and scattered fluxes are detected by the detector. As shown in FIG. 1, the imaging object may be, for example, a phantom, or in a clinical application, the imaging object may be, for example, a patient. The primary flux includes unscattered X-rays. Along with the primary flux, the detector also detects a first-order scattered flux, which includes X-rays that have undergone a single scattering event. The detector also detects a multiple-scattered flux, which includes X-rays that have undergone multiple scattering events. For example, the multiple-scattered flux may include a second-order scattered flux (corresponding to X-rays that have been scattered twice), a third-order scattered flux (corresponding to X-rays that have been scattered three times), and a higher-order scattered flux (corresponding to X-rays that have been scattered four or more times).

[0016] As mentioned above, in deep learning-based scatter correction schemes, when a neural network is trained end-to-end to infer the scatter distribution from the measured projection data, the accuracy of the scatter estimation is limited. To obtain a higher estimation accuracy, the first-order scatter distribution is accurately calculated and used as an input to the trained neural network to obtain the total scatter distribution.

[0017] FIG. 2 illustrates a non-limiting example of a block diagram of a scattered radiation correction circuit 200 according to an embodiment of the present disclosure. The scattered radiation correction circuit 200 includes an image reconstruction circuit 210, a first scattered radiation distribution calculation circuit 220, a total scattered radiation distribution determination circuit 230, and a scattered component removal circuit 240. The image reconstruction circuit 210 reconstructs image data from measured projection data and sends the image data to the first scattered radiation distribution calculation circuit 220. Based on the projection data and the reconstructed image data, the first scattered radiation distribution calculation circuit 220 calculates a first scattered radiation distribution using a model based on physical processing. The total scattered radiation distribution determination circuit 230 receives the first scattered radiation distribution from the first scattered radiation distribution calculation circuit 220 and determines a total scattered radiation distribution. Based on the total scattered radiation distribution determined by the total scattered radiation distribution determination circuit 230, the scattered radiation correction circuit 240 performs scattered radiation correction by removing the scattered components from the projection data. The scattered radiation correction circuit 200 is an example of an information processing device, the image reconstruction circuit 210 is an example of a reconstruction section, the first scattered radiation distribution calculation circuit 220 is an example of an estimation section, and the total scattered radiation distribution determination circuit 230 is an example of a calculation section. The first scattered radiation distribution calculation circuit 220 calculates, for example, a primary scattered radiation distribution as the scattered radiation distribution. Furthermore, the first scattered radiation distribution calculation circuit 220 can also calculate a secondary scattered radiation distribution and a tertiary scattered radiation distribution as the scattered radiation distribution.

[0018] 3 illustrates a non-limiting example flow chart of a scatter correction process 300 in accordance with one embodiment of the present disclosure. The process 300 begins at step S310, where raw projection data obtained by scanning an imaging object with a CT device is received. For example, the raw projection data may be sinogram data generated from a CT scan.

[0019] In step S320, image data is reconstructed from the received projection data. The reconstructed image data can provide various information about the imaging object, which is used in calculating the primary scatter distribution. As an example, a filtered back-projection (FBP) method can be used for the image reconstruction. However, this is not limiting, and other reconstruction methods are possible.

[0020] In step S330, a first-order scattered radiation distribution is calculated based on the projection data and the reconstructed image data. For example, the first-order scattered radiation distribution, which is a distribution of the first-order scattered radiation flux, can be accurately simulated using a physics-based model. The physics-based model can be expressed as follows using the radiative transfer equation (RTE):

[0021]

number

[0022] This is subject to the following boundary conditions:

[0023]

number

[0024]

number

[0025] The RTE can be solved to calculate an accurate scatter solution based on the acquired raw CT projection data and various other parameters. These parameters may include the x-ray source spectrum, CT geometry, voxel-based phantom / tissue geometry information, and the scattering cross section and total attenuation of different materials. For example, the voxel-based phantom / tissue geometry information can be obtained from the CT image data reconstructed using the uncorrected projection data. Various other input parameters such as the scattering cross section and total attenuation of different materials can be obtained or extrapolated from a combination of the reconstructed image and prior knowledge about the imaged object. Furthermore, the given parameters such as the x-ray source spectrum and CT geometry can be measured in advance during calibration or obtained from the known specifications of the CT device.

[0026]

number

[0027] The primary scattered radiation distribution is calculated as described above using the RTE method, but is not limited to this. As long as it can be calculated with high accuracy, other methods can be used, including but not limited to physical model-based methods such as the Monte Carlo method.

[0028] In step S340, a total scatter distribution is determined based on the calculated first scatter distribution. This determination can be performed by a trained neural network. The calculated first scatter distribution is input to the neural network, and the output of the neural network is obtained as a total scatter distribution. The training data set for the neural network can be collected from simulations, research experiments on phantoms or volunteers, and / or clinical diagnosis of patients. For example, a physical model-based simulation, such as a simulation based on the RTE method or the Monte Carlo method, can be performed on the CT device to generate different first scatter distributions and their corresponding total scatter distributions. The neural network can be trained with the collected data to learn the mapping between different first scatter distributions and the corresponding total scatter distributions.

[0029] In step S350, the projection data can be corrected based on the total scatter distribution determined in step S340.

[0030] As mentioned above, the primary scatter distribution is the distribution of the primary component of scatter detected by the CT scanner detector. This component arises from a single scattering event in the imaged object. The primary scatter distribution is a close approximation of the total scatter distribution. The total scatter distribution can be inferred with improved accuracy from an accurately calculated primary scatter distribution.

[0031] 2 and 3, the first scatter distribution is calculated to obtain the total scatter distribution, however, it is feasible to determine the total scatter distribution with greater accuracy using the first component and one or more higher order components, which may include a distribution of first and second scatter components or a distribution of first, second and third order scatter components, etc.

[0032] In one embodiment of the present disclosure, a cascaded neural network framework is devised to further accelerate the scatter correction process. Figure 4 shows a non-limiting example block diagram of a scatter correction circuit 400 for estimating scatter distribution with improved accuracy and reduced processing time in accordance with one embodiment of the present disclosure.

[0033] The scattered radiation distribution estimation circuit 400 includes a primary scattered radiation distribution estimation circuit 410, an image reconstruction circuit 420, a detector pixel selection circuit 430, a second scattered radiation distribution calculation circuit 440, a primary scattered radiation distribution decomposition circuit 450, a total scattered radiation distribution determination circuit 460, and a scattered component removal circuit 470. The scattered radiation distribution estimation circuit 400 is an example of an information processing device, the image reconstruction circuit 410 is an example of a reconstruction unit, the detector pixel selection circuit 430 is an example of a selection unit, the second scattered radiation distribution calculation circuit 440 or the primary scattered radiation distribution decomposition circuit 450 is an example of an estimation unit, and the total scattered radiation distribution determination circuit 460 is an example of a calculation unit. The second scattered radiation distribution calculation circuit 440 calculates, for example, a primary scattered radiation distribution as a scattered radiation distribution. Furthermore, the second scattered radiation distribution calculation circuit 440 can also calculate a secondary scattered radiation distribution and a tertiary scattered radiation distribution as a scattered radiation distribution.

[0034] The first-scatter ray distribution estimation circuit 410 receives raw projection data, estimates the first-scatter ray distribution, and sends it to a detector pixel selection circuit 430. The image reconstruction circuit 420 reconstructs image data from the projection data and sends the reconstructed image data to a second scatter ray distribution calculation circuit 440. Based on the estimated first-scatter ray distribution, the detector pixel selection circuit 430 selects a subset of a plurality of detector pixels included in the detector of the CT apparatus. Using a physical model-based method, the second scatter ray distribution calculation circuit 440 calculates the first-scatter ray distribution for the detector pixels in the selected subset.

[0035] The first scattered ray distribution decomposition circuit 450 receives the first scattered ray distribution calculated by the first scattered ray distribution calculation circuit 440, and decomposes the first scattered ray distribution with respect to a plurality of detector pixels of the detector.

[0036] The total scattered ray distribution determination circuit 460 determines the total scattered ray distribution based on the primary scattered ray distribution decomposed by the primary scattered ray distribution decomposition circuit 450. The scattered component removal circuit 470 performs scattered ray correction based on the total scattered ray distribution determined by the total scattered ray distribution determination circuit 460. Note that the total scattered ray distribution determination circuit 460 is optional and can be omitted. In this case, the scattered component removal circuit 470 can perform scattered ray correction directly using the primary scattered ray distribution from the primary scattered ray distribution decomposition circuit 450.

[0037] 5 illustrates a non-limiting example flow chart of a scatter correction process 500 for estimating scatter distribution with improved accuracy and reduced computation time in accordance with an embodiment of the present disclosure. Process 500 begins at step S510, where raw projection data acquired by scanning an imaging object using a CT scanner is received. At step S520, image data is reconstructed from the acquired projection data. As with the embodiment illustrated in FIGS. 2 and 3, image reconstruction can be performed using a variety of methods, including, but not limited to, the FBP method.

[0038] In step S530, the first-scatter distribution is estimated based on the projection data. For example, a neural network can be used to obtain a rough estimate of the first-scatter distribution from the projection data. The resolution of the first-scatter distribution can be matched to the resolution of the projection data, or can be reduced to allow for downsampling. For example, the resolution of the projection data can be 600×400 and the resolution of the first-scatter distribution can be lower, 200×100.

[0039] Training data sets for the neural network can be collected from simulations, research experiments on phantoms or volunteers, and / or clinical diagnostic procedures on patients. For example, physical model-based simulations, such as RTE or Monte Carlo-based simulations, can be performed with a CT scanner to generate projection data and associated primary scatter distributions. These databases can then be used to train a neural network to learn the mapping from projection data to primary scatter distributions.

[0040] In step S540, a subset of detector pixels of the detector is selected based on the estimated first-scatter distribution. The selection process can be achieved by adaptive downsampling using a certain criterion. For example, detector pixels with local maxima or minima in first-scatter values ​​can be selected. Other criteria are also possible, such as criteria based on the physical part of the imaged object. In addition to these approaches, any other method that helps to reduce the number of detector pixels and thus improve computational efficiency can also be applied. These methods include, but are not limited to, uniform sampling to obtain the subset of detector pixels.

[0041] In step S550, the first-order scatter distribution is calculated for the selected detector pixels based on the projection data and the reconstructed image data. Any method capable of obtaining an accurate scatter distribution can be used, including but not limited to physical model-based methods. Thanks to the adaptive sparse sampling performed in step S540, the first-order scatter distribution is calculated only for the selected detector pixels, which results in a substantial reduction in processing time.

[0042] In step S560, the first-order scatter distribution is resolved for all detector pixels of the detector based on the first-order scatter distribution calculated for the selected detector pixel. Despite its accuracy, the scatter distribution obtained in step S550 may not accurately represent the scatter distribution across the entire detector. The resolution in step S560 can be achieved by an interpolation algorithm or via a neural network. For example, a super-resolution network can be used to recover the values ​​of the non-selected pixels. The input of the network is the first-order scatter distribution of the selected detector pixel, and the target is the scatter distribution of all detector pixels. There is no particular limitation on the network configuration or loss function.

[0043] In step S570, a total scattered radiation distribution is determined based on the decomposed first scattered radiation distribution. This determination can be performed in the same manner as described in the embodiment shown in Figures 2 and 3. In step S580, the projection data can be corrected based on the total scattered radiation distribution determined in step S570. As described above, the processing of step S570 is optional and can be omitted, so in step S580, the projection data can be directly corrected using the first scattered radiation distribution decomposed in step S560.

[0044] Again, the distribution of the first order component and one or more higher order components can be used to more accurately determine the total scatter distribution, which may include a distribution of first and second order scatter components or a distribution of first, second and third order scatter components, etc.

[0045] 6A and 6B illustrate a non-limiting example of selecting a subset of a plurality of detector pixels included in a detector of a CT device, according to one embodiment of the present disclosure.

[0046] As mentioned above, calculating first-order scatter values ​​for only a limited number of selected detector pixels can be an efficient way to speed up the estimation process, for example, adaptive sparse sampling can be applied based on an approximate estimated first-order scatter distribution to determine the locations of the selected detector pixels.

[0047] A non-limiting example of adaptive sparse sampling is shown in FIG. 6A, where detector pixels with local maxima or minima are identified. The scatter distribution has low frequency characteristics and therefore changes very slowly. A peak / valley detection algorithm can be used to capture the location of the detector pixels with the maxima or minima.

[0048] An example of a non-limiting scatter distribution along a row of detector pixels is shown in Figure 6B. If the selected peak / valley pixels (represented by squares in the figure) are too sparse to depict the scatter distribution, more pixels (represented by dots in the figure) between the peak and valley regions can be sparsely selected until the number of selected detector pixels reaches a predefined threshold. For example, the threshold can be determined a priori based on the performance of the super-resolution network.

[0049] Compared to deep learning methods that use end-to-end training between projection data and scatter distributions, the embodiment shown in Figures 4 and 5 cannot improve the total scatter estimation by using an accurately calculated first-order scatter distribution, but it can reduce the processing time by introducing appropriate detector pixel downsampling and a super-resolution network.

[0050] 7 shows a schematic block diagram of a CT device or scanner according to one embodiment of the present disclosure. As shown in FIG. 7, a radiography gantry 750 is shown from the side, and includes an X-ray tube 751, an annular frame 752, and a multi-row or two-dimensional array type X-ray detector 753. The X-ray tube 751 and the X-ray detector 753 are mounted on the annular frame 752 diametrically across an object OBJ, and the annular frame 752 is rotatably supported about a rotation axis RA. A rotating unit 757 rotates the annular frame 752 at a high speed, for example 0.4 seconds / revolution, while the object OBJ is moved along the axis RA into or out of the illustrated page.

[0051] An embodiment of an X-ray CT device according to the present disclosure will be described below with reference to the views of the attached drawings. Note that the X-ray CT device includes various devices, such as a rotating / rotating type device in which an X-ray tube and an X-ray detector rotate together around an object to be inspected, and a fixed / rotating type device in which many detection elements are arranged in a ring or plane and only the X-ray tube rotates around an object to be inspected. The present disclosure can be applied to both types. In this case, the rotating / rotating type, which is currently mainstream, is exemplified.

[0052] The multi-slice X-ray CT apparatus further includes a high voltage device 759 that generates a tube voltage to be applied to the X-ray tube 751 via a slip ring 758 so that the X-ray tube 751 generates X-rays. X-rays are emitted toward an object OBJ whose cross-sectional area is represented by a circle. For example, in the X-ray tube 751, the average X-ray energy during a first scan is smaller than the average X-ray energy during a second scan. In this manner, two or more scans can be obtained corresponding to different X-ray energies. The X-ray detector 753 is located on the opposite side of the X-ray tube 751 through the object OBJ to detect the X-rays that have been emitted and passed through the object OBJ. The X-ray detector 753 further includes individual detector elements or units.

[0053] The CT apparatus further includes other devices for processing signals detected from the X-ray detector 753. A data acquisition circuit or data acquisition system (DAS) 754 converts signals output from the X-ray detector 753 into voltage signals for each channel, amplifies the signals, and further converts the signals into digital signals. The X-ray detector 753 and the DAS 754 are configured to handle a predetermined total number of projections per rotation (TPPR).

[0054] The above mentioned data is sent via a non-contact data transmitter 755 to a pre-processing unit 756 housed in a console outside the radiography gantry 750. The pre-processing unit 756 applies given corrections to the raw data, such as sensitivity corrections. A memory 762 stores the resulting data, also called projection data at the stage immediately before the reconstruction process. The memory 762 is connected to a system controller 760 via a data / control bus 761, together with a reconstructor 764, an input device 765 and a display 766. The system controller 760 controls a current regulator 763, which limits the current to a level sufficient to drive the CT system.

[0055] The detectors are rotated and / or fixed relative to the patient in various generations of CT scanner systems. In one implementation, the CT system described above may be an example of a combination of third generation geometry and fourth generation geometry systems. In a third generation system, the x-ray tube 751 and x-ray detector 753 are mounted diametrically on an annular frame 752 and rotate around the object OBJ as the annular frame 752 rotates about the rotation axis RA. In a fourth generation geometry system, the detectors are fixedly placed around the patient and the x-ray tube rotates around the patient. In an alternative embodiment, the radiography gantry 750 has multiple detectors disposed on an annular frame 752 supported by a C-arm and a stand.

[0056] The memory 762 can store measurements representative of the exposure of X-rays at the X-ray detector 753. The memory 762 can also store specialized programs for performing CT image reconstruction, material decomposition, and motion estimation and motion correction methods, including those described herein.

[0057] The reconstruction unit 764 can perform the above-mentioned methods described herein, and can also perform pre-reconstruction image processing, such as volume rendering and image difference processing, if necessary.

[0058] Reconstruction pre-processing of the projection data performed by pre-processor 756 can include, for example, detector calibration, detector non-linearity, and polar effect correction.

[0059] Post-reconstruction processing performed by the reconstructor 764 may include image filtering and smoothing, volume rendering, and image differencing, if necessary. Image reconstruction may be performed using filtered backprojection, iterative image reconstruction, or stochastic image reconstruction. The reconstructor 764 may use memory to store, for example, projection data, reconstructed images, calibration data and parameters, and computer programs.

[0060] The reconfiguration device 764 may include a CPU (processing circuitry) that may be implemented as discrete logic gates, an Application Specific Integrated Circuit (ASIC), a Field Programmable Gate Array (FPGA), or other Complex Programmable Logic Device (CPLD). The FPGA or CPLD implementation may be coded in VDHL, Verilog, or any other hardware description language, and the code may be stored directly in electronic memory within the FPGA or CPLD or as a separate electronic memory. Additionally, the memory 762 may be non-volatile, such as a ROM, EPROM, EEPROM, or FLASH memory. The memory 762 may be volatile, such as static or dynamic RAM, and a processor, such as a microcontroller or microprocessor, may be provided to manage the electronic memory and the interaction of the FPGA or CPLD with the memory.

[0061] Alternatively, the CPU in the reconstruction device 764 may execute a computer program including a set of computer readable instructions to perform the functions described herein, the program being stored in the non-transitory electronic memory described above and / or a hard disk drive, a CD, a DVD, a FLASH drive, or any other known storage medium. The computer readable instructions may also be provided as a utility application, a background daemon, or a component of an operating system, or a combination thereof, and executed in conjunction with a processor, such as an Intel Xeon processor or an AMD Opteron processor, and an operating system, such as Microsoft 10, UNIX, Solaris, LINUX, Apple, MAC-OS, or other operating systems known to those skilled in the art. Additionally, the CPU may be implemented as multiple processors cooperating to execute instructions in parallel. In one implementation, the reconstructed image may be displayed on a display 766. The display 766 may be an LCD display, a CRT display, a plasma display, an OLED, an LED, or any other display known in the art.

[0062] Memory 762 may be a hard disk drive, a CD-ROM drive, a DVD drive, a FLASH drive, RAM, ROM, or any other electronic storage device known in the art.

[0063] Numerous modifications and variations are possible in light of the above teachings, and therefore, within the scope of the appended claims, the present disclosure may be practiced other than as specifically described herein.

[0064] Embodiments of the present disclosure may be as described in the brackets below.

[0065] (1) A method for performing scattered radiation estimation in a computed tomography (CT) system equipped with a detector having a plurality of detector pixels, the method including: acquiring projection data by scanning an object to be imaged using the CT system; reconstructing image data from the acquired projection data; estimating a first scattered radiation distribution based on the acquired projection data; selecting a first subset of the plurality of detector pixels based on the estimated first scattered radiation distribution; calculating a second scattered radiation distribution for the selected first subset of the plurality of detector pixels based on the acquired projection data and the reconstructed image data, the second scattered radiation distribution being more accurate than the estimated first scattered radiation distribution; and acquiring a third scattered radiation distribution for the second subset of the plurality of detector pixels based on the calculated second scattered radiation distribution, the third subset including the first subset.

[0066] (2) The method according to (1), further comprising determining a fourth scattered radiation distribution based on the acquired third scattered radiation distribution, the fourth scattered radiation distribution being more accurate than the estimated first scattered radiation distribution and including higher order scattered components.

[0067] (3) The method of (2), wherein the estimation further includes estimating a distribution for a first scatter component of the scattered radiation as the estimated first scatter distribution, the first scatter component resulting from a single scattering event occurring within the imaged object, and the determination further includes determining a total scatter distribution as the determined fourth scatter distribution, the total scatter distribution relating to the first scatter component and all higher order components of the scattered radiation, each of the higher order scatter components resulting from a corresponding multiple scattering event occurring within the imaged object.

[0068] (4) The method according to (3), wherein the estimation further includes estimating a distribution of the first scattering component and the second scattering component of the scattered radiation as the estimated first scattering radiation distribution, the second scattering component being caused by two scattering events occurring within the imaging object.

[0069] (5) The method according to (3), wherein the estimation further includes estimating a distribution of the first scattering component, the second scattering component, and the third scattering component of the scattered radiation as the estimated first scattering radiation distribution, the second scattering component being caused by two scattering events occurring within the imaged object, and the third scattering component being caused by three scattering events occurring within the imaged object.

[0070] (6) The method described in (3), wherein the determination includes inputting the acquired third scatter distribution into a trained neural network and obtaining an output of the trained neural network as the determined fourth scatter distribution.

[0071] (7) The method of (6), further comprising: performing a physics-based simulation on the CT system to generate different distributions for the first scatter component and corresponding total scatter distributions as a training data set for the neural network; and training the neural network with the generated training data set to learn a mapping between the different distributions for the first scatter component and the corresponding total scatter distribution.

[0072] (8) The method described in (1), wherein the estimated first scattered radiation distribution has a coarse granularity compared to the acquired projection data.

[0073] (9) The method described in (1), wherein the estimation further includes inputting the acquired projection data into a trained neural network and obtaining an output of the trained neural network as the estimated first scatter distribution.

[0074] (10) The method of (7), further comprising: performing a physics-based simulation on the CT system to generate different projection data and corresponding scatter distributions as a training data set for the neural network; and training the neural network with the generated training data set to learn a mapping between the different projection data and the corresponding scatter distributions.

[0075] (11) The method of (1), wherein the selecting further comprises downsampling the plurality of detector pixels based on a predetermined criterion to obtain the selected first subset of the plurality of detector pixels.

[0076] (12) The method of (11), wherein the downsampling further includes generating the selected first subset of the plurality of detector pixels by determining the number of detector pixels for which the estimated first scatter distribution has a local maximum or a local minimum.

[0077] (13) The method of (12), wherein the generating further includes, in response to a sparsity of detector pixels in the selected first subset of the plurality of detector pixels exceeding a predetermined threshold, adding another number of detector pixels to the selected first subset of the plurality of detector pixels until a total number of detector pixels in the selected first subset of the plurality of detector pixels reaches a predetermined number.

[0078] (14) The method according to (1), wherein the calculation further includes performing a physics-based simulation to generate the calculated second scattered radiation distribution.

[0079] (15) The method according to (14), wherein the performing further includes performing the simulation based on a radiative transfer equation (RTE) method or a Monte Carlo method.

[0080] (16) A method for estimating scattered radiation in a computed tomography (CT) system, the method including: acquiring projection data by scanning an object to be imaged using the CT system; reconstructing image data from the acquired projection data; calculating a first scattered radiation distribution using a physics-based model based on the acquired projection data and the reconstructed image data; and determining a second scattered radiation distribution based on the calculated first scattered radiation distribution, the second scattered radiation distribution including higher-order scattered components compared to the calculated first scattered radiation distribution.

[0081] (17) The method of (16), wherein the calculating further includes calculating a distribution for a first scatter component of the scattered radiation as the calculated first scatter distribution, the first scatter component resulting from a single scattering event occurring within the imaging object, and the determining further includes determining a total scatter distribution as the determined second scatter distribution, the total scatter distribution relating to the first scatter component and all higher order components of the scattered radiation, each of the higher order scatter components resulting from a corresponding multiple scattering event occurring within the imaging object.

[0082] (18) The method described in (17), wherein the calculation further includes calculating a distribution of the first scattered component and the second scattered component of the scattered radiation as the calculated first scattered radiation distribution, the second scattered component being caused by two scattering events occurring within the imaging object.

[0083] (19) The method according to (17), wherein the calculation further includes calculating a distribution of the first scattering component, the second scattering component, and the tertiary scattering component of the scattered radiation as the calculated first scattering radiation distribution, wherein the second scattering component is caused by two scattering events occurring within the imaged object, and the tertiary scattering component is caused by three scattering events occurring within the imaged object.

[0084] (20) An apparatus for performing scattered radiation estimation in a computed tomography (CT) system having a detector having a plurality of detector pixels, the apparatus comprising a processing circuit for performing the following operations: acquiring projection data by scanning an imaging object using the CT system; reconstructing image data from the acquired projection data; estimating a first scattered radiation distribution based on the acquired projection data; selecting a first subset of the plurality of detector pixels based on the estimated first scattered radiation distribution; calculating a second scattered radiation distribution for the selected first subset of the plurality of detector pixels based on the acquired projection data and the reconstructed image data, the second scattered radiation distribution being higher in accuracy than the estimated first scattered radiation distribution; and acquiring a third scattered radiation distribution for the second subset of the plurality of detector pixels based on the calculated second scattered radiation distribution, the second subset including the first subset.

[0085] Those skilled in the art will appreciate that many variations are possible in the operation of the techniques described above while still achieving the same objectives of the disclosure. Such variations are intended to be covered by the scope of the disclosure. As such, the foregoing description of the embodiments of the disclosure are not intended to be limiting. Rather, any limitations to the embodiments of the disclosure are set forth in the following claims. [Explanation of symbols]

[0086] 210 Image reconstruction circuit 220 First scattered radiation distribution calculation circuit 230 Total scattered radiation distribution determination circuit 240 Scattered component removal circuit 400 Scattered radiation correction circuit 410 Primary scattered radiation distribution estimation circuit 420 Image Reconstruction Circuit 430 Detector pixel selection circuit 440 Second scattered radiation distribution calculation circuit 450 Primary scattered ray distribution decomposition circuit 460 Fully scattered line distribution determines the circuit 470 Scattered component removal circuit

Claims

1. 1. A method for performing scatter estimation in a computed tomography (CT) system having a detector with a plurality of detector pixels, the method comprising: acquiring projection data by scanning an imaging object with the CT system; reconstructing image data from the acquired projection data; estimating a first scattered radiation distribution based on the acquired projection data; selecting a first subset of the plurality of detector pixels based on the estimated first scatter distribution; calculating a second scatter distribution for the selected first subset of the plurality of detector pixels based on the acquired projection data and the reconstructed image data, the second scatter distribution being more accurate than the estimated first scatter distribution; and acquiring a third scattered radiation distribution for a second subset of the plurality of detector pixels based on the calculated second scattered radiation distribution, the third scattered radiation distribution having a higher spatial resolution than the calculated second scattered radiation distribution; the second subset includes the first subset; method.

2. The method according to claim 1 , further comprising determining a fourth scattered radiation distribution based on the acquired third scattered radiation distribution, the fourth scattered radiation distribution being more accurate and including higher order scattered components compared to the estimated first scattered radiation distribution.

3. The estimation further includes estimating a distribution of a first scattering component of scattered radiation as the estimated first scattering radiation distribution, the first scattering component being caused by a single scattering event occurring within the imaging object; determining a total scatter distribution as the determined fourth scatter distribution, the total scatter distribution being related to the first scatter component and all higher order components of the scatter, each of the higher order scatter components resulting from a corresponding multiple scattering event occurring within the imaged object; The method of claim 2.

4. 4. The method of claim 3, wherein the estimation further comprises estimating a distribution of the first scattered radiation component and a second scattered radiation component as the estimated first scattered radiation distribution, the second scattered radiation component resulting from two scattering events occurring within the imaging object.

5. 4. The method of claim 3, wherein the estimation further includes estimating a distribution of the first scattering component, the second scattering component, and the third scattering component of the scattered radiation as the estimated first scattering radiation distribution, wherein the second scattering component is caused by two scattering events occurring within the imaging object, and the third scattering component is caused by three scattering events occurring within the imaging object.

6. The said decision: inputting the acquired third scattered radiation distribution into a trained neural network; obtaining an output of the trained neural network as the determined fourth scatter distribution; The method of claim 3 , comprising:

7. performing a physics-based simulation on the CT system to generate different distributions for the first scatter component and corresponding total scatter distributions as a training data set for the neural network; training the neural network with the generated training data set to learn a mapping between the different distributions for the first scatter components and the corresponding total scatter distribution; The method of claim 6 further comprising:

8. The method of claim 1 , wherein the estimated first scattered radiation distribution is coarse-grained compared to the acquired projection data.

9. The estimation is inputting the acquired projection data into a trained neural network; obtaining an output of the trained neural network as the estimated first scatter distribution; The method of claim 1 further comprising:

10. performing a physics-based simulation on the CT system to generate different projection data and corresponding scatter distributions as a training data set for the neural network; training the neural network with the generated training data set to learn a mapping between the different projection data and the corresponding scatter distributions; The method of claim 7 further comprising:

11. The method of claim 1 , wherein the selecting further comprises downsampling the plurality of detector pixels based on a predetermined criterion to obtain the selected first subset of the plurality of detector pixels.

12. 12. The method of claim 11, wherein the downsampling further comprises determining a number of detector pixels for which the estimated first scatter distribution has a local maximum or a local minimum to generate the selected first subset of the plurality of detector pixels.

13. 13. The method of claim 12, wherein the generating further comprises: in response to a sparsity of detector pixels in the selected first subset of the plurality of detector pixels exceeding a predetermined threshold, adding another number of detector pixels to the selected first subset of the plurality of detector pixels until a total number of detector pixels in the selected first subset of the plurality of detector pixels reaches a predetermined number.

14. The method of claim 1 , wherein the calculating further comprises performing a physics-based simulation to generate the calculated second scattered radiation distribution.

15. The method of claim 14 , wherein the performing further comprises performing the simulation based on a radiative transfer equation (RTE) method or a Monte Carlo method.

16. 1. A method for performing scatter estimation in a computed tomography (CT) system, the method comprising: acquiring projection data by scanning an imaging object with the CT system; reconstructing image data from the acquired projection data; calculating a first scattered radiation distribution using a physics-based model based on the acquired projection data and the reconstructed image data; determining a second scattered ray distribution including a higher order scattered component compared to the calculated first scattered ray distribution based on the calculated first scattered ray distribution; method.

17. The calculation further includes calculating a distribution of a first scattering component of the scattered radiation as the calculated first scattering radiation distribution, the first scattering component being caused by a single scattering event occurring within the imaging object; The determining further includes determining a total scatter distribution as the determined second scatter distribution, the total scatter distribution being related to the first scatter component and all higher order components of the scatter, each of the higher order scatter components resulting from a corresponding multiple scattering event occurring within the imaged object.

17. The method of claim 16.

18. 20. The method of claim 17, wherein the calculation further comprises calculating a distribution of the first scattered radiation component and a second scattered radiation component as the calculated first scattered radiation distribution, the second scattered radiation component resulting from two scattering events occurring within the imaging object.

19. The method of claim 17, wherein the calculation further includes calculating a distribution of the first scattered component, the second scattered component, and the third scattered component of the scattered radiation as the calculated first scattered radiation distribution, wherein the second scattered component is caused by two scattering events occurring within the imaging object, and the third scattered component is caused by three scattering events occurring within the imaging object.

20. An information processing device for performing scattered radiation estimation in a computed tomography (CT) system equipped with a detector having a plurality of detector pixels, comprising: a reconstruction unit that reconstructs image data from projection data acquired by scanning an imaging object using the CT system; an estimation unit that estimates a first scattered radiation distribution based on the acquired projection data; a selection unit that selects a first subset of the plurality of detector pixels based on the estimated first scattered radiation distribution; a calculation unit that calculates a second scattered ray distribution, the second scattered ray distribution being more accurate than the estimated first scattered ray distribution, for the selected first subset of the plurality of detector pixels based on the acquired projection data and the reconstructed image data, and acquires a third scattered ray distribution, the second subset of the plurality of detector pixels being more accurate than the estimated first scattered ray distribution, based on the calculated second scattered ray distribution; a processing circuit for performing the second subset includes the first subset; Information processing device.

Images