Material weighting for projection-based spectral x-ray imaging

Abstract

There is provided a method (100) for projection-based spectral X-ray imaging, the method comprising: performing (110) projection-based material decomposition based on spectral X-ray data to generate a set of material-based sinograms, and performing (120) a weighted combination of at least a portion of at least two of the set of material-based sinograms based on the material weighting in the projection domain to obtain a reconstructed image.

Description

Background Art

[0001] The proposed technology relates to X-ray technology and X-ray imaging, and more particularly to reconstructing projection images. In particular, the proposed technology relates to an X-ray imaging system such as a computed tomography (CT) imaging system configured to reconstruct projection images using projection-based spectral X-ray imaging, and a corresponding computer program product to improve image quality.

[0002] Radiographic imaging, such as CT imaging systems and other more general X-ray imaging systems have been used for many years in medical applications such as medical diagnosis and treatment.

[0003] A typical X-ray imaging system, such as a CT imaging system, includes an X-ray source, an X-ray detector, and an associated image processing system. The X-ray detector includes a plurality of detector modules, which include one or more detector elements for independently measuring X-ray intensity. The X-ray source emits X-rays, which pass through the imaged subject or object and are received by the X-ray detector. The X-ray source and the X-ray detector are typically arranged to rotate around the subject or object on a rotating member of a gantry. The emitted X-rays are attenuated by the subject or object as they pass through, and the resulting transmitted X-rays are measured by the X-ray detector. The X-ray detector is connected to a digital acquisition system (DAS), and the measured X-ray data is transmitted to an image processing system to reconstruct an image of the subject or object.

[0004] refer to Figure 1A It may be useful to briefly summarize an illustrative, general X-ray imaging system according to the prior art. In this illustrative example, the X-ray imaging system 100 includes an X-ray source 10, an X-ray detector 20, and an associated image processing system 30. Generally speaking, the X-ray detector 20 is configured to record radiation from the X-ray source 10, which radiation has optionally been focused by an optional X-ray optics or collimator and has passed through an object, subject, or portion thereof. The X-ray detector 20 may be connected to the image processing system 30 via suitable readout electronics that are at least partially integrated in the X-ray detector 20 to enable the image processing system 30 to perform image processing and / or image reconstruction.

[0005] By way of example, a conventional CT imaging system includes an X-ray source and an X-ray detector, which is arranged such that projection images of an object or body can be acquired at different perspectives covering at least 180 degrees. This is most commonly achieved by mounting the source and detector on a support (e.g., a rotating member of a gantry) capable of rotating around the object or body. An image containing projections recorded in different detector elements for different perspectives is called a sinogram. Hereinafter, a set of projections recorded in different detector elements for different perspectives will be called a sinogram, even if the detector is two-dimensional (2D), thereby making the sinogram a three-dimensional (3D) image.

[0006] Figure 1B is a schematic diagram showing an example of an X-ray imaging system setup according to the prior art, showing the projection lines from the X-ray source through the object to the X-ray detector.

[0007] A further development of X-ray imaging is energy-resolved X-ray imaging, also known as spectral X-ray imaging, in which X-ray transmission is measured for several different energy levels. This can be achieved by rapidly switching the source between two different emission spectra, by using two or more X-ray sources emitting different X-ray spectra, or by using an energy-resolved detector that measures the incident radiation at two or more energy levels. An example of such a detector is a multi-bin photon-counting detector, in which each recorded photon generates a current pulse that is compared with a set of thresholds to count the number of photons incident in each of a plurality of energy bins.

[0008] Spectral X-ray projection measurements produce projection images for each energy level. Weighted sums of these projection images can be performed to optimize the contrast-to-noise ratio (CNR) for a specified imaging task, as described in "SNR and DQE analysis of broad-spectrum X-ray imaging", Tapiovaara and Wagner, Phys. Med. Biol. 30, 519.

[0009] Another technique enabled by energy-resolved X-ray imaging is basis material decomposition. This technique utilizes the fact that all materials composed of low atomic number elements (such as human tissues) have linear attenuation coefficients whose energy dependence can be well approximated as a linear combination of two (or more) basis functions:

[0010] μ(E) = a 1 f 1 (E) + a 2 f 2 (E)

[0011] where, f1 and f 2 is a basis function, and a 1 and a 2 is the corresponding basis coefficient. More generally, f i is a basis function, and a i are the corresponding basis coefficients, where i = 1, …, N, where N is the total number of basis functions. If one or more elements with high atomic numbers are present in the imaging volume, high enough to have a K absorption edge in the energy range used for imaging, one basis function must be added for each such element. In the field of medical imaging, such K-edge elements can typically be iodine or gadolinium, which are substances used as contrast agents.

[0012] Basis material decomposition has been described in "Energy-selective reconstructions in X-ray computerized tomography", Alvarez, Macovski, Phys Med Biol. 1976; 21(5): 733-744. In basis material decomposition, the integral of each basis coefficient in the basis coefficients (i=1, ..., N where N is the number of basis functions) is inferred from the measured data in each projection ray l from the source to the detector element. In one embodiment, this is done by first expressing the expected number of recorded counts in each energy bin as A i The function is implemented.

[0013]

[0014] where λ i is the expected number of counts in energy bin i, E is the energy, S i is a response function that depends on the shape of the spectrum incident on the imaged object, the quantum efficiency of the detector, and the sensitivity of energy bin i to X-rays of energy E. Although the term "energy bin" is most commonly used for photon counting detectors, this formula can also describe other energy-resolving X-ray imaging systems, such as multi-slice detectors, kVp switching sources, or multi-source systems.

[0015] Then, A can be estimated using maximum likelihood, assuming that the number of counts in each bin is a Poisson-distributed random variable. iThis is achieved by minimizing the negative log-likelihood function, see, for example, “K-edge imaging in X-ray computed tomography using multi-bin photon counting detectors”, Roessland Proksa, Phys. Med. Biol. 52 (2007), 4679-4696:

[0016]

[0017] Where m i is the number of counts measured in energy bin i, and M b is the number of energy bins.

[0018] When the resulting estimated basis coefficients for each projection line are integrated When arranged into an image matrix, the result is a material-specific projection image of each basis i, also called a basis image. This basis image can be viewed directly (e.g. in projection X-ray imaging) or can be used as a basis coefficient a for forming the interior of an object i The input to a reconstruction algorithm of the map (e.g. in CT imaging). In either case, the result of the basis decomposition can be viewed as one or more basis image representations, such as basis coefficient line integrals or the basis coefficients themselves.

[0019] Image quality improvement of X-ray imaging systems is undoubtedly a key factor in ensuring the quality and safety of patient care and is relevant to various methods, especially in the field of spectral X-ray imaging.

[0020] Therefore, there is still a general demand for image quality improvements, which are reflected in the following aspects: noise reduction, improvement of contrast-to-noise ratio (CNR), improvement of patient dose efficiency, etc. Summary of the invention

[0021] This summary introduces concepts that are described in more detail in the detailed description. It should not be used to identify essential features of the claimed subject matter, nor should it be used to limit the scope of the claimed subject matter.

[0022] According to one aspect, a method for projection-based spectral X-ray imaging is provided. The method includes: performing a projection-based material decomposition based on spectral X-ray data to generate a set of material basis sinusoidal graphs, and performing a weighted combination of at least a portion of at least two material basis sinusoidal graphs in the set of material basis sinusoidal graphs based on material weighting in the projection domain to obtain a reconstructed image.

[0023] According to another aspect, a CT imaging system is provided, the CT imaging system comprising an X-ray source configured to emit X-rays, an X-ray detector configured to generate spectral X-ray data, and a processor. The processor is configured to perform projection-based material decomposition on the spectral X-ray data to generate a set of material basis sinusoidal graphs. The processor is further configured to perform a weighted combination of at least a portion of at least two material basis sinusoidal graphs in the set of material basis sinusoidal graphs based on material weighting in the projection domain to obtain a reconstructed image.

[0024] The proposed technique enables a method of reconstructing an image based on projection-based spectral X-ray imaging. Reconstructing the image is based on adaptive / dynamic material weighting in the projection domain, i.e., reconstructing from the set of material-based sinusoidal graphs. The proposed technique allows for improved mapping of material-based sinusoidal graphs, thereby improving the quality of the reconstructed image. It should be understood that the X-ray imaging may be CT imaging, and the reconstructed image may be a reconstructed CT image. The quality improvement may, for example, be associated with an improved CNR, a reduced noise, and / or an enhancement of certain features in the image (such as anatomical features). In addition, the proposed technique enables improved patient dose efficiency without causing an intolerable loss in image quality. Furthermore, the proposed technique does not require additional calibration during use, thereby further facilitating the implementation of the technique. It should also be noted that the proposed technique is applicable to all photon counting X-ray detectors and other dual-source systems, such as X-ray and mammography systems. BRIEF DESCRIPTION OF THE DRAWINGS

[0025] Aspects of the present disclosure may be better understood after referring to the drawings and reading the detailed description.

[0026] Figure 1A and Figure 1B is a schematic diagram illustrating an example of an X-ray imaging system.

[0027] Figure 2 is a schematic diagram illustrating another example of an X-ray imaging system such as a CT imaging system.

[0028] Figure 3 is a schematic block diagram of a CT imaging system as an illustrative example of an X-ray imaging system.

[0029] Figure 4 is a schematic diagram illustrating another example of relevant components of an X-ray imaging system, such as a CT imaging system.

[0030] Figure 5 is a schematic diagram of a photon counting circuit and / or device according to an exemplary embodiment.

[0031] Figure 6is a schematic diagram illustrating an example of a semiconductor detector submodule according to an exemplary embodiment.

[0032] Figure 7 is a schematic diagram illustrating an example of a semiconductor detector submodule according to another exemplary embodiment.

[0033] Fig. 8A is a schematic diagram illustrating an example of a semiconductor detector sub-module according to yet another exemplary embodiment.

[0034] Figure 8B is a schematic diagram showing an example of a set of tiled detector submodules, wherein each detector submodule is a depth segmented detector submodule and an application specific integrated circuit (ASIC) or corresponding circuit system is arranged below the detector submodule as can be seen from the direction of the incident X-ray.

[0035] Fig. 9 is a schematic diagram illustrating an example of a CT imaging system.

[0036] Fig.10 is a schematic diagram showing an example of the design of an X-ray source and an X-ray detector system.

[0037] Fig.11 is a schematic flow chart illustrating a method for projection-based spectral X-ray imaging.

[0038] Fig.12 is a schematic flow chart illustrating a method for projection-based spectral X-ray imaging.

[0039] Fig.13A and Fig. 13B is a graph illustrating the relationship between variance and single energy.

[0040] Fig.14 is a schematic diagram illustrating certain relevant parts of an X-ray imaging system.

[0041] Fig.15 is a schematic diagram illustrating an example of a computer implementation according to an embodiment. DETAILED DESCRIPTION

[0042] Embodiments of the present disclosure will now be described by way of example with reference to the accompanying drawings.

[0043] For a better understanding, it may be useful to continue with the introductory description of a non-limiting example of an overall X-ray imaging system in which data processing and transmission according to the inventive concept may be implemented.

[0044] Figure 21 is a schematic diagram showing an example of an X-ray imaging system 100, such as a CT imaging system, which includes: an X-ray source 10 that emits X-rays; an X-ray detector 20 that has an X-ray detector that detects the X-rays after they have passed through an object; an analog processing circuit system 25 that processes and digitizes the raw electrical signals from the X-ray detector; a digital processing circuit system 40 that can perform further processing operations on the measurement data, such as applying corrections, temporary storage, or filtering; and a computer 50 that stores the processed data and can perform further post-processing and / or image reconstruction. The digital processing circuit system 40 may include a digital processor. According to an exemplary embodiment, all or part of the analog processing circuit system 25 may be implemented in the X-ray detector 20. The X-ray source and the X-ray detector may be coupled to a rotating member of a gantry 11 of the CT imaging system 100.

[0045] The entire X-ray detector may be considered an X-ray detector system 20 , or an X-ray detector 20 in combination with associated analog processing circuitry 25 .

[0046] The image processing system 30 communicates with and is electrically coupled to the analog processing circuit system 25, and may include a digital processing circuit system 40 and / or a computer 50, which may be configured to perform image reconstruction based on image data from the X-ray detector. Thus, the image processing system 30 may be viewed as the computer 50, or alternatively as a combination of the digital processing circuit system 40 and the computer 50, or, if the digital processing circuit system 40 is further specialized for image processing and / or reconstruction, it may be viewed as the digital processing circuit system itself.

[0047] An example of a commonly used X-ray imaging system is a CT imaging system, which may include an X-ray source or X-ray tube that generates a fan beam or cone beam of X-rays and an array of opposing X-ray detectors that measure the fraction of X-rays that pass through a patient or object. The X-ray source or X-ray tube and the X-ray detectors are mounted in a gantry 11 that can rotate around the imaging object.

[0048] Figure 3A CT imaging system 100 is schematically shown as an illustrative example of an X-ray imaging system. The CT imaging system includes a computer 50 that receives commands and scanning parameters from an operator via an operator console 60, which may have a display 62 and some form of operator interface, such as a keyboard, mouse, joystick, touch screen, or other input device. The commands and parameters provided by the operator are then used by the computer 50 to provide control signals to an X-ray controller 41, a gantry controller 42, and a table controller 43. Specifically, the X-ray controller 41 provides power and timing signals to the X-ray source 10 to control the emission of X-rays to an object or patient located on the table 12. The gantry controller 42 controls the rotation speed and position of the gantry 11 including the X-ray source 10 and the X-ray detector 20. For example, the X-ray detector 20 may be a photon counting X-ray detector. The table controller 43 controls and determines the position of the patient table 12 and the scan coverage of the patient. There is also a detector controller 44, which is configured to control and / or receive data from the X-ray detector 20.

[0049] In one embodiment, the computer 50 also performs post-processing and image reconstruction on the image data output from the X-ray detector 20. Thus, the computer 50 corresponds to FIG. 1 and FIG. Figure 2 The image processing system 30 is shown. An associated display 62 allows an operator to observe the reconstructed image and other data from the computer 50.

[0050] An X-ray source 10 arranged in a gantry 11 emits X-rays. An X-ray detector 20 (which may be in the form of a photon counting X-ray detector) detects the X-rays after they have passed through an object or patient. The X-ray detector 20 may be formed, for example, of a plurality of pixels (also referred to as sensors or detector elements) and associated image processing circuitry (such as an application specific integrated circuit (ASIC)) arranged in a detector module. A portion of the analog processing may be implemented in the pixels, while any remaining processing is implemented, for example, in the ASIC. In one embodiment, the image processing circuitry (ASIC) digitizes the analog signals from the pixels. The image processing circuitry (ASIC) may also include digital processing that may perform further processing operations on the measurement data, such as applying corrections, temporary storage and / or filtering. During scanning to acquire X-ray projection data, the gantry and components mounted thereon rotate around the isocenter 13.

[0051] Modern X-ray detectors generally require the conversion of incident X-rays into electrons, which often occurs via the photoelectric effect or Compton interaction, and the resulting electrons generally produce secondary visible light until their energy is lost and this light is in turn detected by a photosensitive material. Semiconductor-based detectors also exist, and in this case the electrons generated by the X-rays generate a charge from electron-hole pairs that are collected by an applied electric field.

[0052] There are detectors that work in energy integrating mode, in the sense that they provide an integrated signal from a large number of X-rays. The output signal is proportional to the total energy deposited by the detected X-rays.

[0053] X-ray detectors with photon counting and energy resolution capabilities are increasingly being used in medical X-ray applications. Photon counting detectors have the advantage that in principle the energy of each X-ray can be measured, which yields additional information about the composition of the object. This information can be used to improve image quality and / or reduce radiation dose.

[0054] Generally speaking, photon counting X-ray detectors determine the energy of a photon by comparing the height of an electrical pulse generated by the interaction of a photon in the detector material with a set of comparator voltages. These comparator voltages are also called energy thresholds. Generally speaking, the analog voltage in the comparator is set by a digital-to-analog converter (DAC). The DAC converts the digital setting sent by the controller into an analog voltage to which the height of the photon pulse can be compared.

[0055] Photon counting detectors count the number of photons that have interacted in the detector during the measurement time. New photons are generally identified by the fact that the height of the electrical pulse exceeds the comparator voltage of at least one comparator. When a photon is identified, the event is stored by incrementing a digital counter associated with the channel.

[0056] When several different thresholds are used, an energy-resolving photon counting detector is obtained, in which the detected photons can be classified into energy bins corresponding to various thresholds. Sometimes, this type of photon counting detector is also called a multi-bin detector. In general, energy information allows the creation of new kinds of images, in which new information is available and image artifacts inherent to conventional techniques can be removed. In other words, for an energy-resolving photon counting detector, the pulse height is compared with multiple programmable thresholds (T1-TN) in a comparator and classified according to the pulse height, which is in turn proportional to the energy. In other words, a photon counting detector comprising more than one comparator is referred to herein as a multi-bin photon counting detector. In the case of a multi-bin photon counting detector, the photon counts are stored in a set of counters, typically each counter corresponding to an energy threshold. For example, a count can be assigned to the highest energy threshold that a photon pulse has exceeded. In another example, a counter tracks the number of times a photon pulse crosses each energy threshold.

[0057] For example, "side-on facing" is a special non-limiting design of a photon counting detector in which the X-ray sensors (such as X-ray detector elements or pixels) are oriented sideways toward incoming X-rays.

[0058] For example, such a photon counting detector may have pixels in at least two directions, wherein one of the two directions facing the sideways photon counting detector has a component in the direction of the X-ray. Such a sideways photon counting detector is sometimes referred to as a depth segmented photon counting detector having two or more pixel depth segments in the direction of the incoming X-ray. It should be noted that one detector element may correspond to one pixel, and / or multiple detector elements may correspond to one pixel, and / or data signals from multiple detector elements may be used for one pixel.

[0059] Alternatively, the pixels may be arranged in an array (non-depth segmented) in a direction substantially orthogonal to the incident X-rays, and each pixel may be oriented sideways toward the incident X-rays. In other words, the photon counting detector may be non-depth segmented while still being arranged sideways toward the incoming X-rays.

[0060] The absorption efficiency can be improved by arranging the side-facing photon counting detector to face sideways, in which case the absorption depth can be chosen to be any length and the side-facing photon counting detector can still be fully depleted without reaching very high voltages.

[0061] The conventional mechanism for detecting X-ray photons by direct semiconductor detectors works essentially as follows. The energy of the X-ray interaction in the detector material is converted into electron-hole pairs inside the semiconductor detector, where the number of electron-hole pairs is generally proportional to the photon energy. The electrons and holes drift toward the detector electrodes and back (or vice versa). During this drift, the electrons and holes induce a current in the electrodes, which can be measured.

[0062] like Figure 4 As shown in , the signals are routed via wiring paths 26 of the detector elements 22 of the X-ray detector to the input of an analog processing circuit system (e.g., an ASIC) 25. It should be understood that the term "application specific integrated circuit (ASIC)" should be broadly interpreted as any general purpose circuit used and configured for a particular application. The ASIC processes the charge generated from each X-ray and converts it into digital data, which can be used to obtain measurement data, such as photon counts and / or estimated energy. The ASIC is configured to be connected to the digital processing circuit system so that the digital data can be sent to the digital processing circuit system 40 and / or one or more memory circuits or components 45, and ultimately the data will be provided for Figure 2 The image processing circuitry 30 or computer 50 in generates input for reconstructing the image.

[0063] Since the number of electrons and holes from one X-ray event is proportional to the energy of the X-ray photon, the total charge in one induced current pulse is proportional to this energy. After the filtering step in the ASIC, the pulse amplitude is proportional to the total charge in the current pulse and therefore proportional to the X-ray energy. The pulse amplitude can then be measured by comparing the value of the pulse amplitude with one or more thresholds (THR) in one or more comparators (COMP) and introducing a counter by which the number of cases where the pulse is greater than the threshold can be recorded. In this way, the number of X-ray photons whose energy exceeds the energy corresponding to the corresponding threshold (THR) that has been detected within a certain time frame can be counted and / or recorded.

[0064] The ASIC typically samples the analog photon pulse once per clock cycle and records the output of the comparator. Depending on whether the analog signal is above or below the comparator voltage, the comparator (threshold) outputs a one or zero. The available information at each sample is, for example, a one or zero for each comparator, which indicates whether the comparator has been triggered (the photon pulse is above the threshold) or not triggered.

[0065] In a photon counting detector, there is typically a photon counting logic component that determines whether a new photon has been recorded and records the photon in a counter. In the case of a multi-bin photon counting detector, there are typically several counters, such as one counter for each comparator, and the photon counts are recorded in these counters based on an estimate of the photon energy. This logic component can be implemented in several different ways. Two of the most common categories of photon counting logic components are non-paralyzable counting mode and paralyzable counting mode. Other photon counting logic components include, for example, local maximum detection, which counts local maxima detected in a voltage pulse and may also record its pulse height.

[0066] Photon counting detectors have many benefits, including but not limited to: high spatial resolution; low sensitivity to electronic noise, good energy resolution; and material separation capabilities (spectral imaging capabilities). However, energy integrating detectors have the advantage of high count rate tolerance. The count rate tolerance comes from the fact / recognition that since the total energy of the photons is measured, adding one additional photon will always increase the output signal (within reasonable limits), regardless of the amount of photons currently being recorded by the detector. This advantage is one of the main reasons why energy integrating detectors are the standard for medical CT today.

[0067] Figure 5 A schematic diagram of a photon counting circuit and / or device according to an exemplary embodiment is shown.

[0068] When photons interact in a semiconductor material, a cloud of electron-hole pairs is generated. The charge carriers are collected by electrodes attached to the detector material by applying an electric field across the detector material. The signal is routed from the detector element to the input of a parallel processing circuit (e.g., an ASIC). In one example, the ASIC can process the charge so that a voltage pulse is generated whose maximum height is proportional to the amount of energy deposited by the photon in the detector material.

[0069] The ASIC may include a set of comparators 302, each of which compares the magnitude of the voltage pulse to a reference voltage. The comparator output is typically zero or one (0 / 1), depending on which of the two voltages being compared is larger. Here, we assume that if the voltage pulse is higher than the reference voltage, the comparator output is one (1); if the reference voltage is higher than the voltage pulse, the comparator output is zero (0). A digital-to-analog converter (DAC) 301 can be used to convert a digital setting that may be provided by a user or a control program into a reference voltage that may be used by the comparators 302. If the height of the voltage pulse exceeds the reference voltage of a particular comparator, we will refer to that comparator as being "triggered." Each comparator is typically associated with a digital counter 303, which increments based on the comparator output according to the photon counting logic component.

[0070] As mentioned earlier, when the resulting estimated basis coefficients for each projection ray are integrated When arranged into an image matrix, the result is a material-specific projection image of each basis i, also called a basis image. This basis image can be viewed directly (e.g. in projection X-ray imaging) or can be used as a basis coefficient a for forming the interior of an object i Input to a reconstruction algorithm of a map (eg in CT). In any case, the result of the basis decomposition can be viewed as one or more basis image representations, such as basis coefficient line integrals or the basis coefficients themselves.

[0071] It will be appreciated that the mechanisms and arrangements described herein can be implemented, combined and rearranged in various ways.

[0072] For example, embodiments may be implemented in hardware, or at least partially in software executed by appropriate image processing circuitry, or a combination thereof.

[0073] The steps, functions, processes and / or blocks described herein may be implemented in hardware (including both general purpose electronic circuitry and application specific circuitry) using any conventional technology such as discrete circuitry or integrated circuit technology.

[0074] Alternatively, or in addition, at least some of the steps, functions, processes and / or blocks described herein can be implemented in software, such as a computer program executed by appropriate image processing circuitry (such as one or more processors or processing units).

[0075] In the following, non-limiting examples of specific detector module implementations will be discussed. More specifically, these examples refer to side-facing oriented detector modules and depth-segmented detector modules. Other types of detectors and detector modules are also possible.

[0076] Figure 6 is a schematic diagram illustrating an example of a semiconductor detector submodule according to an exemplary embodiment. This is an example of a detector module 21 in which the semiconductor sensor has a plurality of detector elements or pixels 22, wherein each detector element (or pixel) is typically based on a diode having a charge collection electrode as a key component. X-rays enter through the edge of the detector module.

[0077] Figure 7 is a schematic diagram illustrating an example of a semiconductor detector submodule according to another exemplary embodiment. In this example, again assuming that X-rays enter through the edge of the detector module, the detector module 21 with semiconductor sensors is also divided into a plurality of depth segments or detector elements 22 in the depth direction.

[0078] Typically, a detector element is a single X-ray sensitive sub-element of a detector. Generally speaking, photon interactions occur in a detector element and the charge generated thereby is collected by corresponding electrodes of the detector element.

[0079] Each detector element typically measures the incident X-ray flux as a sequence of frames. A frame is the data measured during a specified time interval, called a frame time.

[0080] Depending on the detector topology, a detector element may correspond to a pixel, especially when the detector is a flat panel detector. A depth segmented detector may be viewed as having multiple detector strips, each strip having multiple depth segments. For such a depth segmented detector, each depth segment may be viewed as a separate detector element, especially if each of the depth segments is associated with its own separate charge collection electrode.

[0081] The detector strips of a depth segmented detector usually correspond to pixels of a common flat panel detector and are therefore sometimes also called pixel strips. However, a depth segmented detector can also be viewed as a three-dimensional pixel array, where each pixel corresponds to a separate depth segmenter / detector element.

[0082] The semiconductor sensor can be implemented as a so-called multi-chip module (MCM), in the sense that it serves as a base substrate for electrical wiring and multiple ASICs, which are preferably attached by so-called flip-chip technology. The wiring will include signal connections from each pixel or detector element to the ASIC input and connections from the ASIC to external memory and / or digital data processing. The power to the ASIC can be supplied by similar wiring, but also by separate connections, taking into account the increase in the cross-section required for the high currents in these connections. The ASIC can be positioned to the side of the active sensor, which means that it can be protected from incident X-rays if an absorber cover is placed on top, and it can also be protected from scattered X-rays coming from the side by positioning the absorber in this direction as well.

[0083] Fig. 8A8,183,535 is a schematic diagram showing a detector module implemented as an MCM similar to the embodiment in U.S. Patent No. 8,183,535. In this example, it is shown how the semiconductor sensor 21 can also have the function of the substrate in the MCM. The signal is routed from the detector element 22 to the input of the parallel processing circuit 24 (e.g., ASIC) positioned next to the active sensor area via the wiring path 23. The ASIC processes the charge generated from each X-ray and converts it into digital data, which can be used to detect photons and / or estimate the energy of the photons. The ASIC can have their own digital processing circuit system and memory for small tasks. In addition, the ASIC can be configured to connect to the digital processing circuit system and / or memory circuit or component located outside the MCM, and ultimately use the data as input for reconstructing the image.

[0084] However, the adoption of deep segmentation also brings two notable challenges to silicon-based photon counting detectors. First, a large number of ASIC channels must be employed to process the data fed from the associated detector segments. In addition to the increased number of channels due to both smaller pixel size and deep segmentation, multiple energy bins also increase the data size. Second, since a given X-ray input count is divided into smaller pixels, segments, and energy bins, each with a much lower signal, detector calibration / correction requires several orders of magnitude more calibration data to minimize statistical uncertainty.

[0085] Naturally, in addition to requiring larger computing resources, hard drives, memory, and central processing units (CPUs) or graphics processing units (GPUs), the data size that is several orders of magnitude larger also slows down both data processing and preprocessing. For example, when the data size is 10GB instead of 10MB, the processing time for reading and writing data may be as long as 1000 times.

[0086] A problem in any detector that counts X-ray photons is the pile-up problem. When the flux rate of X-ray photons is high, problems may arise in distinguishing two subsequent charge pulses. As mentioned above, the pulse length after the filter depends on the shaping time. If this pulse length is greater than the time between two X-ray photon induced charge pulses, the pulses will grow together and the two photons will be indistinguishable and can be counted as one pulse. This is called pile-up. Therefore, one way to avoid pile-up at high photon fluxes is to use a small shaping time, or to use deep segmentation.

[0087] For pile-up calibration vector generation, the pile-up calibration data needs to be pre-processed for jet correction. For material decomposition vector generation, the material decomposition data should preferably be pre-processed for both jet correction and pile-up correction. For patient scan data, the data needs to be pre-processed for jet, pile-up, and material decomposition before image reconstruction ensue. These are simplified examples to explain pre-processing, as actual pre-processing steps may include several other calibration steps such as reference normalization and air calibration as needed. The term processing may only indicate the last step of each calibration vector generation or patient scan, but may be used interchangeably in some cases.

[0088] Figure 8B is a schematic diagram showing an example of a set of tiled detector submodules, wherein each detector submodule is a deeply segmented detector submodule and an ASIC or corresponding circuit system 24 is arranged below the detector elements 22 as can be seen from the direction of the incoming X-rays, thereby allowing wiring paths 23 from the detector elements 22 to parallel processing circuits 24 (e.g., ASICs) in the space between the detector elements.

[0089] Fig. 9 1 is a schematic diagram illustrating an outline example of a CT imaging system. In this schematic example, the overall CT imaging system 100 includes a gantry 111, a patient table 112, which can be inserted into an opening 114 of the gantry 111 during a patient scan and / or a calibration scan. The direction of the rotation axis of the rotating member of the gantry around the imaged object or patient is denoted as the z-direction. The angular direction of the CT imaging system is denoted as the x-direction, and the direction of the incident X-ray is referred to as the y-direction.

[0090] It should be understood, however, that the rotating and fixed components of the gantry need not be part of the CT imaging system, but may be arranged and / or configured in other ways, such as for linear and / or translational relative movement without rotation. As an example, the X-ray source and detector combination may move in a linear and / or translational manner relative to the fixed components of the entire gantry. For example, the X-ray source and detector may move together as a combined assembly unit along a table axis (commonly referred to as the z-axis). Alternatively, the patient table moves while the X-ray source and detector combination is stationary; relative motion is key. This also includes geometric system configurations in which the patient can stand, such as in a so-called phone booth scanner.

[0091] Fig.10is a schematic diagram showing an example of an overall design of an X-ray source-detector system. In this example, a schematic diagram of an X-ray detector including a plurality of detector modules and an X-ray source emitting X-rays is shown. Each detector module may have a set of detector elements defining corresponding pixels. For example, the detector modules may be arranged in parallel and oriented with the side facing the detector module pointing to the X-ray source, and they may be arranged in a slightly curved overall configuration. As described above, the direction of the incident X-rays is referred to as the y-direction. Multiple detector pixels in the direction of the rotation axis of the gantry (referred to as the z-direction) enable multi-plane image acquisition. Multiple detector pixels in the angular direction (referred to as the x-direction) enable simultaneous measurement of multiple projections in the same plane, and this is applied to fan / cone beam CT. The x-direction is sometimes also referred to as the channel direction. Most detectors have detector pixels in both the plane (z) direction and the angular (x) direction.

[0092] Fig.11 is a schematic flow chart illustrating an example of a method for projection-based spectral X-ray imaging. According to an embodiment, a method S0 for projection-based spectral X-ray imaging. The method includes a first step S1 of performing a projection-based material decomposition based on spectral X-ray data to generate a set of material basis sinusoids. The next step S2 includes performing a weighted combination of at least a portion of at least two material basis sinusoids in the set of material basis sinusoids based on material weighting in the projection domain to obtain a reconstructed image.

[0093] A projection-based material decomposition is performed for each projection measurement from the spectral X-ray data. In other words, the material decomposition is performed in the projection domain. According to an example, a pile-up correction technique may be applied to the spectral X-ray data before performing the projection-based material decomposition. This example may improve the linearization of the X-ray detector response.

[0094] The set of material basis sinusoidal graphs is obtained by material decomposition based on projection. The generated set of material basis sinusoidal graphs may include two material basis sinusoidal graphs or more material basis sinusoidal graphs. According to an example, the material decomposition based on projection generates two material basis sinusoidal graph estimates given as follows:

[0095]

[0096] where μ(E) is the monoenergetic attenuation measured by projection, A 1 and A 2 are the material basis sinusoidal maps for bin i and pixel j for a given material, also called the material basis estimate, and w 1 and w 2 is the corresponding energy-dependent linear decay, which is also called the material-based estimation coefficient or simply the material weight. Therefore, the choice of the single energy E determines the material-based sinusoidal diagram A1 and A 2 The relative weight of . The single energy E may be referred to as monochromatic energy. According to an example, the base material image is a base sinusoidal image.

[0097] The weighted combination of at least a portion of the material basis sinusoids is based on adaptive material weighting. In other words, (at least a portion of) the material basis sinusoids are individually weighted. Therefore, weights are determined for a portion and / or all of the material basis sinusoids. Therefore, a single energy can be selected for the material basis sinusoids, respectively, to obtain a reconstructed image that is optimized in, for example, the following aspects: maximizing CNR and / or minimizing noise. Since the noise and / or CNR vary based on the path length of the material used for each projection, there is no single single energy that can be selected in the image domain that minimizes noise and / or maximizes CNR. An advantage of the present invention is that a weighted combination of material basis sinusoids based on (adaptive) material weighting in the projection domain is performed. This requires a more versatile, more flexible and / or more accurate reconstruction of the projections into a reconstructed image. Therefore, the proposed technique allows for an improved image quality, which is reflected in, for example, the following aspects: noise reduction, improved CNR and improved signal-to-noise ratio (SNR). In addition, the present invention enables an improved patient dose efficiency. According to an example, when the present invention is combined with mA modulation (also known as tube current modulation), patient dose efficiency can be further improved.

[0098] Fig.12 is a schematic flow chart illustrating a method for projection-based spectral X-ray imaging. Fig.12 The schematic flow chart and Fig.11 The schematic flow charts in have several common features and are hereby referenced Fig.11 and related text to increase understanding of at least some of the features and / or functions in the flowchart. It should also be noted that for simplicity, Fig.12 The schematic flow chart in may not disclose every possible combination, arrangement, etc. of the steps of method S0 of the exemplary embodiment. Therefore, Fig.12 It should be construed as being non-limiting with respect to possible modifications and / or combinations within and between the steps of method S0 of the exemplary embodiment.

[0099] By way of example, the step of performing S2 the weighted combination of at least a portion of at least two basis sinusoids may include determining S3 at least one of: a viewing angle-dependent material weight for each projection in a set of a plurality of projections individually, and a pixel-dependent material weight for each pixel in a set of a plurality of pixels individually. In other words, a corresponding weight is determined for each projection. The corresponding weight may be viewing angle-dependent and / or pixel-dependent. Thus, a combination of viewing angle-dependent and pixel-dependent material weights for each projection is feasible, and also only one of the viewing angle-dependent and pixel-dependent material weights is feasible. The viewing angle-dependent and / or pixel-dependent material weights (which are also collectively referred to as "material weights") may be determined such that they minimize the noise in the monoenergetic attenuation μ(E) of the generated material basis sinusoids for the projection. Since the material weights are energy-dependent, this may be achieved by determining a monoenergetic attenuation μ(E) that minimizes the noise in the monoenergetic attenuation μ(E) of the projection. Thus, the material weights may be determined using projection-specific noise characteristics.

[0100] The step of performing S2 the weighted combination of at least a portion of the at least two basis sinusoidal graphs may further comprise: combining S4 the at least a portion of the at least two material basis sinusoidal graphs in the projection domain based on the at least one of the view-dependent material weight and the pixel-dependent material weight. Thus, the determined view-dependent and / or pixel-dependent material weights are applied to the at least two material basis sinusoidal graphs in the projection domain, and a combined image may be generated. When the determined weights are applied to the material basis sinusoidal graphs, respectively, the material basis sinusoidal graphs may be referred to as weighted basis material sinusoidal graphs or weighted sinusoidal graphs. The weighted sinusoidal graphs may then be reconstructed into an image, i.e. a reconstructed image.

[0101] In a specific example, the step of determining S3 at least one of the view-dependent material weight and the pixel-dependent material weight may be performed based on at least one of a variance and a covariance between the at least a portion of the at least two material basis sinusoidals.

[0102] The covariance between the at least one portion of the at least two material basis sinusoids may be measured in a calibration step and / or modeled, estimated, calculated, etc. According to an example, the covariance between the at least one portion of the at least two material basis sinusoids is estimated by a forward model given by: ij (A N jk ), where i is the bin, and A N jk is the material basis estimate for pixel j, view angle k, and material N of sinogram A. The forward model can be determined by computing the Cramér-Rao lower bound (CRLB). The required Fisher information in the case of uncorrelated bin counts can be given by computing the Fisher matrix given as follows:

[0103]

[0104] Then, the Fisher information F j CRLB is determined by the inverse of . In addition, the forward model may be specific to each detector element in the X-ray imaging system. Furthermore, for the additional noise model, the same method of estimating the variance between the at least one portion of at least two material basis sinusoids may be applied, but the definition of the Fisher matrix is ​​different. According to an example, before determining the covariance between (at least a portion of) the material basis sinusoids, (at least a portion of) the material basis sinusoids may be filtered in at least one of the view direction and the pixel direction.

[0105] The variance of a given single energy is given by: w(E) T F -1 w(E). According to the example, for a given F j , the variable w is determined by minimizing the variance of a given single energy min The variable w min Represents the material basis sinusoidal graph A jk According to another example, w can be determined by maximizing the CNR given by the following formula: min :

[0106]

[0107] According to an example, before determining the variance of (at least a portion of) the material basis sinusoidal graph, (at least a portion of) the material basis sinusoidal graph may be filtered in at least one of a viewing angle direction and a pixel direction, respectively.

[0108] In a specific example, the step of determining S3 at least one of the view-dependent material weight and the pixel-dependent material weight may include maximizing or minimizing S5 the first objective function with respect to at least one of the noise, artifacts, depiction of anatomical features, pile-up, ratio, signal-to-noise ratio (SNR) and contrast-to-noise ratio (CNR) of the reconstructed image. Thus, by maximizing or minimizing S5 the first objective function, i.e. optimizing the first objective function, all or only some of the image characteristics may be taken into account. The first objective function may also be optimized in a manner that achieves a compromise between the image characteristics.

[0109] In a non-limiting example, determining S3 at least one of a view-dependent material weight and a pixel-dependent material weight may include selecting S6 a region of interest (ROI) and determining at least one of a view-dependent material weight and a pixel-dependent material weight in the ROI.

[0110] It should be noted that when determining at least one of the S3 view-dependent and pixel-dependent material weights, the selected ROI may be a common, shared, identical ROI for all of the at least one portion of the at least two material-based sinusoids. The selected ROI may be a plurality of ROIs in the corresponding material-based sinusoids, wherein the plurality of ROIs overlap, thus forming a (common) selected ROI. The view-dependent and / or pixel-dependent material weights may be calculated based on measurements in the selected ROI, but then the view-dependent and / or pixel-dependent material weights may be applied to a second region of the material-based sinusoids. In other words, a weighted combination of the at least one portion of the at least two material-based sinusoids is performed in the second region. The second region may include the selected ROI, or only a portion of the selected ROI. In some examples, the selected ROI is excluded from the second region. The second region is preferably larger than the selected ROI. According to an example, determining the S3 view-dependent and / or pixel-dependent material weights may be performed by forward projection based on measurements in the selected ROI. According to an example, the material to be decomposed may be selected, for example, based on a selected scanning protocol. Accordingly, A can be selected based on these materials. 1 and A 2 This can be achieved, for example, by segmenting the material and applying a forward projection model to obtain the segmented material basis sinusoidal graphs.

[0111] According to an example, a classifier algorithm may identify at least two materials in a selected ROI. The classifier algorithm may be part of a machine learning system. View-dependent and / or pixel-dependent material weights may be determined to maximize the detectability of the at least two materials in the selected ROI. According to another example, multiple ROIs may be selected for the at least a portion of the at least two material basis sinusoidal graphs, wherein the multiple ROIs do not need to completely overlap. Therefore, several selected ROIs may be the same, or at least different to some extent. In addition, according to this example, a classifier algorithm may identify multiple at least two materials in the multiple selected ROIs. The materials in different ROIs may be the same or different. A combination of view-dependent and / or pixel-dependent material weights may be determined to maximize the detectability of the at least two materials in the selected ROI.

[0112] In another non-limiting example, at least one of the viewing angle-dependent material weight and the pixel-dependent material weight may be at least one of: constrained within a predetermined reference interval based on a predetermined material weight, and parameterized according to a material or signal level. The viewing angle-dependent and / or pixel-dependent material weight may be parameterized in a manner that achieves a trade-off between, for example, CNR and artifact suppression. In addition, the material weight referred to as w refThe predetermined reference interval of A can be parameterized according to the material or signal level. For example, the material function can depend on A N jk .

[0113] In a specific example, the step of determining at least one of the viewing angle dependent material weight and the pixel dependent material weight may include maximizing or minimizing S7 a second objective function with respect to material separation of the plurality of materials. The second objective function may be the same or different than the first objective function.

[0114] In a specific example, at least one of the view-dependent material weight and the pixel-dependent material weight and the predetermined material weight may be parameterized as a third objective function of at least one of: i) an estimated gradient in at least one of the view direction and the pixel direction of the projection; and ii) an estimated gradient of the reconstructed image to compensate for the deviation. The third objective function may be the same or different than the first objective function and / or the second objective function. For example, the third objective function may employ additional parts and / or aspects of the material basis sinusoidal graph compared to the first and / or second objective functions.

[0115] In a non-limiting example, the step of determining at least one of the view-dependent material weight and the pixel-dependent material weight may be performed by a first machine learning system. The first machine learning system may include one or several machine learning models, such as a decision tree, a support vector machine, a neural network, etc. According to an example, the first machine learning system may include a machine learning architecture and / or a machine learning algorithm that may be based at least in part on a convolutional neural network model. According to another example, the first machine learning system may include a machine learning architecture and / or a machine learning algorithm that may be based at least in part on at least one of a decision tree model and a support vector machine model.

[0116] In another non-limiting example, the first machine learning system may include a trained first convolutional neural network (CNN). The step of determining at least one of the view-dependent material weight and the pixel-dependent material weight is performed by the first CNN. The first CNN is trained for at least one of mapping the at least a portion of the at least two material basis sinusoids to at least one of the view-dependent material weight and the pixel-dependent material weight, and estimating at least one of the variance and the covariance based on the mapping between the at least a portion of the at least two material basis sinusoids and the fourth objective function.

[0117] Therefore, the first CNN is trained to: map (at least a portion of) the material basis sinusoidal graph to the view-dependent and / or pixel-dependent material weights, and / or estimate the variance and / or covariance based on the mapping between (at least a portion of) the material basis sinusoidal graph and the fourth objective function. The fourth objective function may be the same as or different from the first objective function, the second objective function, and the third objective function. The fourth objective function may be CRLB. According to the example of obtaining the trained first CNN, the view-dependent and / or pixel-dependent material weights for the common imaging task may be determined for several individual phantom images. Then, the determined view-dependent and / or pixel-dependent material weights may be used to train the first CNN. The output from the first CNN may be a map of view-dependent and / or pixel-dependent material weights. According to another example, the CNN may be trained to output a low-noise variance map. The low-noise variance map may be used to determine at least one of the view-dependent and / or pixel-dependent material weights. The advantage of using the first CNN is that further optimized material weights may be obtained, and thus a reconstructed image with relatively low noise and high resolution may be obtained. Furthermore, providing a first CNN for determining material weights may reduce the need for human interaction, decision making, etc., or may eliminate such need entirely. Furthermore, using a first CNN to determine material weights may result in reduced computation time.

[0118] In a specific example, the step of performing S2 a weighted combination of at least a portion of at least two material basis sinusoids may be performed by a second machine learning system by mapping the at least a portion of at least two material basis sinusoids into a reconstructed image. The second machine learning system may be the same or different than the first machine learning system. The second machine learning system may include one or several machine learning models, such as a decision tree, a support vector machine, a neural network, etc. According to an example, the second machine learning system may include a machine learning architecture and / or a machine learning algorithm that may be based at least in part on a convolutional neural network. According to another example, the second machine learning system may include a machine learning architecture and / or a machine learning algorithm that may be based at least in part on at least one of a decision tree model and a support vector machine model. According to an example, the (same) machine learning system is configured to: determine S3 view-dependent and / or pixel-dependent material weights, and perform a weighted combination of (at least a portion of) the S2 material basis sinusoids to obtain a reconstructed image. It should be noted that it is also feasible that the machine learning system only performs the step of performing a weighted combination of (at least a portion of) the S2 material basis sinusoids, i.e., without determining the S3 material weights.

[0119] In a specific example, the second machine learning system may include a trained second convolutional neural network (CNN), wherein the step of performing the weighted combination of the at least two material basis sinusoids is performed by the second CNN, the second CNN being trained based on at least one of a sinusoid obtained from a simulation image, a phantom image, and a patient image. The second CNN may be the same or different than the first CNN.

[0120] According to another specific example, the final reconstructed image may be generated by a fourth machine learning system including a third trained CNN. The fourth machine learning system may be the same or different than the first and / or second machine learning system. The third CNN may be the same or different than the first and / or second CNN. The CNN may be trained based at least on sinograms obtained from simulated images, phantom images, and / or patient images. The CNN may be optimized so that the final reconstructed image achieves optimal CNR for different path lengths. The CNN may be further optimized so that the final reconstructed image achieves minimum deviation and / or minimum retained edge. The input to the CNN may be various material-based sinograms A, A S , A f and material weight w min 、w ref Alternatively, the CNN may be configured to generate a deviation map. The generated deviation map may then be compared with the determined w min (A S jk )A jk are combined to generate the final reconstructed image.

[0121] For example, a training set of material basis sinusoids and basis material weights may be generated, and an optimal combination of these sinusoids may be selected manually or by numerical optimization in order to select a weighting scheme that gives the optimal image quality of the reconstructed image as measured by an objective or subjective image quality metric. For example, optimal weights may be calculated for several different imaging tasks, and a combination of these optimal weights may be selected manually, at least in part, by forming angle-dependent and pixel-dependent weighted averages of different weight maps. After applying this method to find the optimal weighted sinusoids for the sinusoids in the training set, a CNN may be trained to map a set of at least two of the parameters of the various material basis sinusoids and material weights to an optimal combination of basis sinusoids. In this way, a CNN may be used to obtain the same image quality as the training set without having to go through a manual or optimization-based search to find the optimal way to combine the sinusoids.

[0122] In a non-limiting example, the step of performing S2 the weighted combination of at least a portion of the at least two basis sinusoids may comprise iterative forward projection. The iterative forward projection comprises calculating S8 an estimated forward projection of the at least a portion of the at least two material basis sinusoids, wherein the weighted combination may be based at least in part on the estimated forward projection.

[0123] In another non-limiting example, the method S0 may include at least one of the following operations: filtering S9 and downsampling S10 at least a portion of at least one of the spectral X-ray data and the at least a portion of the at least two material basis sinusoids to compensate for at least one of low signal, photon starvation, and noise. Thus, the filtering S9 and / or downsampling S10 may be performed before and / or after performing S1 projection-based material decomposition. For example, the low-pass filtered material basis sinusoidal portion A S jk The material basis sinusoidal graph A can be obtained by jk The low-pass filtered material base sinusoidal graph A is obtained by low-pass filtering and decimation of at least a portion of the material base sinusoidal graph A. S jk can be varied to a relatively small extent so that the minimum noise estimate itself does not introduce additional noise. According to this example, the risk of suppressing high frequency content is reduced. For each bin i, the minimum noise is selected for each extracted pixel j and viewing angle k. Then, the adaptively weighted material basis sinusoidal map A for pixel j and viewing angle k can be obtained f jk Come as w min (A S jk )A jk According to the example, the extraction is omitted, but based on A S jk To select the minimum noise i for each pixel and viewing angle.

[0124] According to an example, different levels of filtered material basis sinusoidal graph A S jk Can be used to reduce, for example, A S jk The variance is calculated to compensate for the base sinusoidal pattern including relatively low signal and / or photon-starved materials.

[0125] In another non-limiting example, the method further includes bias-adjusting at least one of the first weighted material basis sinusoidal graph and the reconstructed image to generate a bias-adjusted reconstructed image. The bias-adjusting is accomplished by at least one of the following operations: performing Fourier domain segmentation on the first weighted sinusoidal graph and the weighted, filtered sinusoidal graph in the projection domain, and combining at least a portion of the first weighted sinusoidal graph after the Fourier domain segmentation and at least a portion of the weighted, filtered sinusoidal graph after the Fourier domain segmentation to generate a bias-adjusted material sinusoidal graph, wherein the bias-adjusted sinusoidal graph is reconstructed into a bias-adjusted reconstructed image, and performing Fourier domain segmentation on a first reconstructed image reconstructed from the weighted sinusoidal graph and a second reconstructed image reconstructed from the weighted, filtered sinusoidal graph in the image domain, and combining at least a portion of the first reconstructed image after the Fourier domain segmentation and at least a portion of the second reconstructed image after the Fourier domain segmentation to generate a bias-adjusted reconstructed image. It should be understood that a weighted sinusoid or a weighted material-based sinusoid may be defined as a combination of the at least two material-based sinusoids based on the determined material weights in the projection domain. A weighted, filtered sinusoid may be a material-based sinusoid that is filtered or downsampled, for example, to compensate for low signal, photon starvation, and / or noise and to which material weights are applied. Material weights may be determined in different ways, as disclosed throughout the application. A reference sinusoid may be a different weighted sinusoid and / or a filtered, weighted sinusoid. A reconstructed image may be reconstructed from a weighted sinusoid, a filtered, weighted sinusoid, and / or a bias-adjusted material-based sinusoid. A bias-adjusted material-based sinusoid may be referred to as a weighted, bias-adjusted material-based sinusoid. It should also be understood that a Fourier domain segmented sinusoid or a Fourier domain segmented reconstructed image is a sinusoid or image that has undergone Fourier domain segmentation.

[0126] Thus, different portions of the material basis sinusoidal graph may be processed in different ways, such as by filtering and / or bias adjustment. For example, a weighted sinusoidal graph may be combined with a weighted, filtered sinusoidal graph by performing Fourier domain segmentation on two weighted, filtered, or unfiltered material basis sinusoidal graphs to generate a Fourier domain segmented material basis sinusoidal graph, which in turn may be combined to generate a weighted, bias-adjusted material basis sinusoidal graph. It should be noted that at least a portion of the material basis sinusoidal graph may be omitted in the filtering and / or bias adjustment, thereby producing an initial portion, an unprocessed portion, a native portion, etc. of the material basis sinusoidal graph. At least two of the various portions of the material basis sinusoidal graph, such as an initial portion A of the material basis sinusoidal graph, a filtered portion A of the material basis sinusoidal graph, and the like, may be mixed in the Fourier domain. S and the deviation-adjusted portion A of the material basis sinusoidal diagram fcTherefore, when Fourier domain segmentation is performed, the initial part A of the material basis sinusoidal graph, the filtered part A S and the bias-adjusted portion A fc Any combination of presence is possible.

[0127] The deviation-adjusted material basis sinusoidal diagram A can be obtained by performing the following fc :

[0128]

[0129] This allows combinations from ref The low spatial frequencies of the weighted sinogram are combined with the min The high spatial frequencies of the weighted sinogram are combined and a bias-corrected, weighted sinogram is produced, which can then be reconstructed into an image. Mixing different frequencies can be achieved by segmenting the frequencies of various parts of the (weighted) material basis sinogram, thereby generating a Fourier domain segmented (weighted) material basis sinogram. Then, the Fourier domain segmented material basis sinograms are combined. For example, relatively high frequencies can be obtained by filtering the filtered portion A of the material basis sinogram. f This can result in a reduction in the noise present. In addition, relatively low frequencies can be dominated by the initial portion A of the material basis sinusoid. This can result in an increase in bias correction. According to an example, different levels of filtered material basis sinusoid A S jk can be combined to improve bias correction. The bias adjustment can be given by:

[0130]

[0131] Where m represents different filtering levels.

[0132] According to an example, Fourier domain segmentation can be performed in the image domain rather than in the projection domain. This can be done by first transforming the weighted material basis sinusoidal graph Reconstruct the weighted material basis image and the reference material basis sinusoidal image A ref jk This is achieved by reconstructing the reference weighted material basis image. ref jk It may refer to, for example, a basis sinusoidal graph w weighted with a reference weight ref A jk , refers to the original determined material basis sinusoidal diagram A jk , or refers to the filtered material base sinusoidal graph A SThe reconstructed weighted material basis image and the reference weighted basis material image are then blended in the Fourier domain. For example, both images may be low pass filtered, such as by Fourier transformation, weighting and inverse transformation, after which the bias corrected basis material image is generated as follows:

[0133]

[0134] Wherein R represents the reconstruction operator, and S represents the low-pass filter.

[0135] In a specific example, the step of performing S2 a weighted combination of at least two material basis sinusoids may include determining S15 a material weight matrix including at least a first material weight and a second material weight separately for each at least one of the projection and the pixel, wherein the second material weight is orthogonal to the first material weight. Thus, the generated material basis sinusoids are composed of vector-valued material basis sinusoids.

[0136] The matrix referred to as W includes material weights including at least a first material weight w min (A S jk ) and the second material weight w ⊥ min (A S jk ). Therefore, the second material weight is a unit vector orthogonal to the first material weight. By replacing w with W in the previously mentioned equation ref To obtain the vector valued material basis sinusoidal graph. The next step is to combine the equation with the constant orthogonal matrix W ref The resulting bias-adjusted material basis sinusoidal graph A is fc It can be finally obtained by the following formula:

[0137]

[0138] Since W(A S jk ) and W ref are orthogonal matrices, so their product is a rotation matrix. The given example comprising the matrix W, comprising at least the first material basis sinusoidal graph and the second material basis sinusoidal graph can therefore be interpreted as rotating around A S jk Rotate A jk transformation.

[0139] Fig.13A and Fig. 13B is a graph illustrating the relationship between variance and single energy. Fig.13A and Fig. 13B In FIG. 1 , CNR is shown on the vertical axis, and the single energy MonoE in keV is shown on the horizontal axis. Fig.13AIn the middle, the dotted line L 1 The first relationship between the difference and the single energy is shown at longer path lengths. The dashed line L 2 The second relationship between the variance and the single energy is shown at shorter path lengths. The CNR / noise varies based on the material thickness used for each projection, which means that the optimal single energy may vary with each viewing angle and / or pixel. It is for L 1 The optimized single energy of minimizing noise / maximizing CNR, and It is for L 2 The optimized single energy for minimizing noise / maximizing CNR. and Therefore, there is no single energy that can be chosen in the image domain that maximizes CNR and / or minimizes noise.

[0140] exist Fig. 13B In the middle, the dotted line L 3 Represents the relationship between variance and single energy. L 3 represents the CNR of an insert as a function of monoE. The dashed line L 4 represents the CNR of a reconstructed image after weighted combination of at least two material basis sinusoids in the projection domain according to an embodiment of the present invention.

[0141] Fig.14 4 is a schematic diagram illustrating certain relevant parts of an X-ray imaging system, such as a CT imaging system. The CT imaging system 400 includes an X-ray source 410 configured to emit X-rays, an X-ray detector 420 configured to generate spectral X-ray data, and a processor 430. The processor 430 is configured to perform a projection-based material decomposition on the spectral X-ray data to generate a set of material basis sinusoidal graphs, and to perform a weighted combination of at least a portion of at least two material basis sinusoidal graphs in the set of material basis sinusoidal graphs based on material weighting in the projection domain to obtain a reconstructed image.

[0142] Processor 430 may include digital and / or analog processing circuitry. Processor 430 may form part of an imaging processing system or the entire imaging processing system. For a better understanding of processors, please refer to Figure 2 To Figure 8 and associated text.

[0143] In a specific example, the processor 430 may be configured to determine at least one of: a separate view-dependent material weight for each projection in a set of a plurality of projections, and a separate pixel-dependent material weight for each pixel in a set of a plurality of pixels. The processor 430 is further configured to combine the at least a portion of the at least two material basis sinusoids in the projection domain based on the at least one of the view-dependent material weight and the pixel-dependent material weight.

[0144] In another specific example, the processor 430 may be configured to determine at least one of the view-dependent material weight and the pixel-dependent material weight based on at least one of the variance and the covariance between the at least a portion of the at least two material basis sinusoidals.

[0145] In a non-limiting example, the processor 430 may be configured to determine at least one of the view-dependent material weights and the pixel-dependent material weights based on maximizing or minimizing a fifth objective function based on at least one of noise, artifacts, delineation of anatomical features, pile-up, ratio, signal-to-noise ratio (SNR), and contrast-to-noise ratio (CNR) relative to the reconstructed image.

[0146] In another non-limiting example, the processor 430 may include a third machine learning system 440 configured to perform a weighted combination of at least a portion of the at least two material basis sinusoidal graphs by mapping the at least a portion of the at least two material basis sinusoidal graphs into a reconstructed image.

[0147] It should be understood that the CT imaging system can include any 3-D X-ray based medical imaging modality. For example, the CT imaging system can include a general tomography system that is adapted to take images from different angles and combine the information therein to create an image volume in which each slice can be viewed individually. The CT imaging system can include a CT system that is adapted to use an angular range of a full 180 degree range and / or a limited angular range. In other words, the CT system can include a general computed tomography system that uses a full 180 degree range, or a breast tomosynthesis system (also known as a mammography system) that uses a limited angular range.

[0148] In this example, the CT system includes an X-ray source 110 and an X-ray detector 120, which are arranged in the beam path of the X-rays so that projection images of the subject or object can be acquired from different viewing angles. This is most commonly achieved by mounting the X-ray source 110 and the X-ray detector 120 on a support (e.g., a rotating member of a gantry) that can rotate around the subject or object.

[0149] As mentioned, at least some of the steps, functions, processes and / or blocks described herein may be implemented in software, such as a computer program, for execution by appropriate image processing circuitry, such as one or more processors or processing units.

[0150] Fig.152 is a schematic diagram illustrating an example of a computer-specific implementation according to an embodiment. In this particular example, the system 200 includes a processor 210 and a memory 220, the memory including instructions executable by the processor, whereby the processor is operable to perform the steps and / or actions described herein. The instructions are typically organized as a computer program 225, 235, which may be pre-configured in the memory 220 or downloaded from an external memory device 230. Optionally, the system 200 includes an input / output interface 240, which may be interconnected to the processor 210 and / or the memory 220 to enable input and / or output of relevant data such as input parameters and / or resulting output parameters.

[0151] In a particular example, the memory 220 includes a set of instructions executable by the processor, whereby the processor is operable to perform the steps and / or actions described herein.

[0152] The term "processor" should be interpreted in a general sense as any system or device capable of executing program code or computer program instructions to perform specific processing, determination or computing tasks.

[0153] Thus, image processing circuitry including one or more processors is configured to perform well-defined processing tasks (such as those described herein) when executing a computer program.

[0154] The image processing circuit system does not have to be dedicated only to performing the above-mentioned steps, functions, processes and / or blocks, but may also perform other tasks.

[0155] The proposed technology also provides a computer program product comprising a computer readable medium 220, 230 having such a computer program stored thereon.

[0156] By way of example, the software or computer program 225, 235 may be implemented as a computer program product, which is typically carried or stored on a computer readable medium 220, 230, specifically a non-volatile medium. The computer readable medium may include one or more removable or non-removable memory devices, including but not limited to: read-only memory (ROM), random access memory (RAM), compact disc (CD), digital versatile disc (DVD), Blu-ray disc, universal serial bus (USB) memory, hard disk drive (HDD) storage device, flash memory, magnetic tape, or any other conventional memory device. Thus, the computer program may be loaded into the operating memory of a computer or equivalent processing device for execution by its image processing circuit system.

[0157] The method flow, when executed by one or more processors, can be regarded as a computer action flow. The corresponding device, system and / or apparatus can be defined as a set of functional modules, wherein each step executed by the processor corresponds to a functional module. In this case, the functional module is implemented as a computer program running on the processor. Therefore, the device, system and / or apparatus can alternatively be defined as a set of functional modules, wherein these functional modules are implemented as computer programs running on at least one processor.

[0158] The computer program resident in the memory may therefore be organized into appropriate functional modules that are configured to perform at least part of the steps and / or tasks described herein when the computer program is executed by a processor.

[0159] Alternatively, these modules may be implemented primarily through hardware modules, or alternatively through hardware. The degree of software versus hardware is purely a choice of the particular implementation.

[0160] The embodiments of the present disclosure shown in the drawings and described above are exemplary embodiments only and are not intended to limit the scope of the appended claims, including any equivalents included within the scope of the claims. Those skilled in the art will appreciate that various modifications, combinations and changes may be made to these embodiments without departing from the scope of the invention as defined by the appended claims. It is intended that any combination of non-mutually exclusive features described herein is within the scope of the invention. That is, the features of the embodiments may be combined with any appropriate aspect described above, and the optional features of any one aspect may be combined with any other appropriate aspect. Similarly, the features listed in the dependent claims may be combined with non-mutually exclusive features of other dependent claims, particularly where the dependent claims are dependent on the same independent claim. In some jurisdictions that require single claim dependencies, these dependencies may have been used as practice, but this should not be taken to mean that the features in the dependent claims are mutually exclusive.

[0161] It should also be noted that, unless explicitly stated otherwise, the concept of the present invention relates to all possible feature combinations. In particular, where technically possible, different partial solutions in different embodiments can be combined in other configurations.

[0162] Definition of terms

[0163] In the following description of non-limiting examples, the following symbols and definitions are mainly used herein:

[0164] Index i = bin, j = pixel, k = view

[0165] A NFor material N, the original MD sinusoidal diagram, where N = 1, 2, ... A N Often referred to simply as A for a given material; the sinusoidal graph A may be appropriately selected and optionally pre-processed for a particular region of interest.

[0166] MD Material Decomposition.

[0167] Material based estimate for pixel j and view angle k of sinogram A. Material based estimate is often interchangeably referred to as material thickness estimate. Sometimes referred to as the

[0168] A S Low-pass filtered MD sinusoidal graph.

[0169] Low-pass filtered material basis estimate for pixel j and view angle k.

[0170] A reference material basis estimate for pixel j and view angle k, which represents the previously determined material basis estimate.

[0171] A f Adaptively weighted MD sinusoidal map.

[0172] Adaptively weighted material basis estimate for pixel j and view angle k.

[0173] Adaptively weighted, bias-corrected MD sinusoidal graph.

[0174] Adaptively weighted, bias-corrected basis estimate for pixel j and view angle k.

[0175] f ij (A jk ) The forward model for pixel j, bin i.

[0176] F ij Fisher matrix for pixel j, bin i.

[0177] w(E) is the material weight for the single energy E.

[0178] w opt (A) A function that returns the optimal weights for a given projection A and a task function.

[0179] w ref Represents the reference weighting for a given reference single energy.

[0180] Used for The optimal single energy material weight.

[0181] CRLB Cramero lower bound.

Claims

1. A method (100) for projection-based spectral X-ray imaging, the method comprising: performing (110) a projection-based material decomposition based on the spectral X-ray data to generate a set of material basis sinusoidal diagrams, A weighted combination of at least a portion of at least two material basis sinusoidals in the set of material basis sinusoidals is performed (120) based on a material weighting in the projection domain to obtain a reconstructed image.

2. The method of claim 1 , wherein said step of performing said weighted combination of said at least a portion of at least two material basis sinusoidal graphs comprises: Determine (130) at least one of the following: a view-dependent material weight applied separately to each projection in a set of multiple projections, and a pixel-dependent material weight, individually for each pixel in a group of multiple pixels, and The at least a portion of at least two material basis sinusoidal graphs are combined (140) in the projection domain based on the at least one of the viewing angle dependent material weight and the pixel dependent material weight.

3. The method according to claim 2, wherein the step of determining at least one of the view-dependent material weight and the pixel-dependent material weight is performed based on at least one of the variance and the covariance between the at least a portion of at least two material basis sinusoidal graphs.

4. The method according to claim 2 or 3, wherein the step of determining at least one of the viewing angle-dependent material weight and the pixel-dependent material weight comprises: A first objective function is maximized or minimized (150) with respect to at least one of noise, artifacts, delineation of anatomical features, pile-up, ratio, signal-to-noise ratio (SNR), and contrast-to-noise ratio (CNR) of the reconstructed image.

5. The method according to any of the preceding claims, wherein the step of determining at least one of the viewing angle dependent material weight and the pixel dependent material weight comprises: A region of interest ROI is selected (160) and at least one of the view-dependent material weight and the pixel-dependent material weight in the ROI is determined.

6. The method according to any one of claims 2 to 5, wherein at least one of the viewing angle-dependent material weight and the pixel-dependent material weight is subjected to at least one of the following operations: Constraints are within predetermined reference intervals based on predetermined material weights and parameterized according to material or signal level.

7. The method according to any one of claims 2 to 6, wherein the step of determining at least one of the viewing angle-dependent material weight and the pixel-dependent material weight comprises: A second objective function is maximized or minimized (180) with respect to material separation of the plurality of materials.

8. The method of any one of the preceding claims, wherein the method further comprises bias-adjusting (220) at least one of the first weighted material basis sinusoidal map and the reconstructed image to generate a bias-adjusted reconstructed image by at least one of the following operations: performing Fourier domain segmentation on the first weighted sinusoid and the weighted, filtered sinusoid in the projection domain, and combining at least a portion of the Fourier domain segmented first weighted sinusoid and at least a portion of the Fourier domain segmented weighted, filtered sinusoid to generate a bias-adjusted material sinusoid, wherein the bias-adjusted sinusoid is reconstructed into the bias-adjusted reconstructed image, and Fourier domain segmentation is performed in the image domain on a first reconstructed image reconstructed from a weighted sinusoidal map and a second reconstructed image reconstructed from a weighted, filtered sinusoidal map, and at least a portion of the first reconstructed image after the Fourier domain segmentation and at least a portion of the second reconstructed image after the Fourier domain segmentation are combined to generate the deviation-adjusted reconstructed image.

9. A method according to any one of the preceding claims, wherein at least one of the viewing angle dependent material weight and the pixel dependent material weight, and Predetermine the material weights, an objective function parameterized as at least one of: i) an estimated gradient in at least one of the view direction and the pixel direction of the projection to compensate for the deviation; and ii) Estimated gradient of the reconstructed image.

10. The method according to any one of claims 2 to 9, wherein the step of determining at least one of the viewing angle dependent material weight and the pixel dependent material weight is performed by a machine learning system.

11. The method of claim 10, wherein the machine learning system comprises a trained convolutional neural network (CNN), wherein the step of determining at least one of the view-dependent material weight and the pixel-dependent material weight is performed by the CNN, wherein the CNN is trained for at least one of the following operations: Mapping the at least a portion of the at least two material basis sinusoidal graphs to the at least one of the view-dependent material weights and the pixel-dependent material weights, and estimating at least one of the variance and the covariance based on the mapping between the at least a portion of the at least two material basis sinusoidal graphs and the objective function.

12. The method according to any one of the preceding claims, wherein the step of performing the weighted combination of at least two material basis sinusoidal graphs comprises: A material weight matrix is ​​determined (250) including at least a first material weight and a second material weight, separately for each at least one of a projection and a pixel, wherein the second material weight is orthogonal to the first material weight.

13. A CT imaging system (400), comprising: an X-ray source (410) configured to emit X-rays, an X-ray detector (420) configured to generate spectral X-ray data, A processor (430), the processor being configured to: performing a projection-based material decomposition on the spectral X-ray data to generate a set of material basis sinusoidal diagrams, and A weighted combination of at least a portion of at least two material basis sinusoidals in the set of material basis sinusoidals is performed based on a material weighting in the projection domain to obtain a reconstructed image.

14. The CT imaging system of claim 13, wherein the processor is configured to: Determine at least one of the following: separate view-dependent material weights for each projection in a set of multiple projections, and a separate pixel-dependent material weight for each pixel in a set of multiple pixels, and The at least a portion of at least two material basis sinusoidals are combined in the projection domain based on the at least one of the viewing angle dependent material weight and the pixel dependent material weight.

15. A computer program product (500), comprising a non-volatile computer readable storage medium having stored thereon a computer program, the computer program comprising instructions which, when executed by a processor, cause the processor to perform the steps of the method according to any one of claims 1 to 12.

Images