Scan image reconstruction method and scan imaging system

Abstract

The application provides a scanning image reconstruction method and a scanning imaging system, and relates to the fields of image processing, artificial intelligence and radiation imaging technology. The method comprises the following steps: acquiring dynamic projection data of an object to be imaged in response to ray scanning of the object to be imaged; dividing the dynamic projection data into a plurality of phases according to a predetermined first time window to obtain a plurality of sub-projection data located in the plurality of phases; processing the plurality of sub-projection data by using a pre-trained transient projection image reconstruction model to obtain a plurality of sub-phase reconstruction images; extracting prior information of a feature space according to the plurality of sub-phase reconstruction images; embedding the prior information of the feature space into a space-time continuous representation model; and obtaining a reconstruction image of the object to be imaged by using the dynamic projection data and the space-time continuous representation model with the embedded prior information. In this way, the dynamic projection data and the data-driven prior information can be fully combined to realize high space-time resolution image reconstruction.

Description

Technical Field

[0001] This application relates to the fields of image processing technology, artificial intelligence technology, and radiation imaging technology, and more specifically, to a scanning image reconstruction method and a scanning imaging system. Background Technology

[0002] With the increasing demand for high-precision imaging of high-speed motion processes, X-ray scanning imaging systems urgently need to improve the spatiotemporal resolution of imaging. For scanning image reconstruction, the improvement of spatiotemporal resolution will bring two major challenges: (1) The data dimension will be greatly increased, and the existing methods of modeling time and space using discretized grids will face challenges such as a surge in the number of parameters and insufficient characterization accuracy; (2) Insufficient projection sampling and highly ill-conditioned solution process: Since the scanning speed of existing scanning imaging equipment is already close to its limit, it is difficult to further improve it. The solution parameters required for high spatiotemporal resolution scanning image reconstruction are greatly increased, and the projection data sampling is seriously insufficient, resulting in a highly ill-conditioned solution process. For high spatiotemporal resolution scanning image reconstruction, how to effectively perform image reconstruction to obtain high-quality reconstructed images has always been one of the important topics of concern for researchers in related fields.

[0003] It should be noted that the information disclosed in this section is only used to understand the background of the inventive concept of this application. Therefore, the above information may include information that does not constitute prior art. Summary of the Invention

[0004] In view of at least one of the above-mentioned technical problems, embodiments of this application provide a scanning image reconstruction method, the method comprising: acquiring dynamic projection data of the object to be imaged in response to ray scanning; dividing the dynamic projection data into multiple time phases according to a predetermined first time window to obtain multiple sub-projection data located in the multiple time phases; processing the multiple sub-projection data using a pre-trained transient projection image reconstruction model to obtain multiple time-phase reconstructed images; extracting prior information of a feature space based on the multiple time-phase reconstructed images; embedding the prior information of the feature space into a spatiotemporal continuous representation model; and obtaining a reconstructed image of the object to be imaged using the dynamic projection data and the spatiotemporal continuous representation model embedded with the prior information.

[0005] For example, acquiring the dynamic projection data of the object to be imaged includes: during the process of the ray scanning of the object to be imaged, at least one attribute of the object to be imaged is dynamically changing; acquiring rays transmitted through the object to be imaged with the at least one dynamically changing attribute at a second time window period to acquire the dynamic projection data of the object to be imaged.

[0006] For example, at least one attribute of the object to be imaged includes at least one of the structure, shape and material form of the object to be imaged.

[0007] For example, the spatiotemporal continuous representation model includes: a model established in four dimensions, including three spatial dimensions and a time dimension, using a spatiotemporal coupling modeling approach.

[0008] For example, the spatiotemporal continuous representation model includes: a model established in four dimensions, including three spatial dimensions and a time dimension, using a spatiotemporal decoupling modeling approach.

[0009] For example, the spatiotemporal continuous representation model is used to directly describe the dynamic changes of the object to be imaged.

[0010] For example, the spatiotemporal continuous representation model includes an object anatomical structure model and a motion feature model.

[0011] For example, the spatiotemporal continuous representation model includes a spatiotemporal continuous representation model constructed with implicit neural representation or three-dimensional Gaussian sputtering.

[0012] For example, the prior information of the feature space includes location-encoded feature vectors or point cloud distributions.

[0013] For example, the step of extracting prior information of the feature space based on the plurality of time-separated reconstructed images includes: processing the plurality of time-separated reconstructed images using a pre-trained feature space information extraction model to extract the prior information of the feature space.

[0014] For example, the method further includes: determining the first time window based on the scanning speed and spatial resolution of the scanning imaging system.

[0015] For example, the step of processing the multiple sub-projection data using a pre-trained transient projection image reconstruction model to obtain multiple time-phase reconstructed images includes: inputting the multiple sub-projection data into the pre-trained transient projection image reconstruction model; the transient projection image reconstruction model performing image reconstruction on the multiple sub-projection data respectively to obtain the multiple time-phase reconstructed images; obtaining an average template image based on the multiple time-phase reconstructed images; and obtaining multiple motion displacement vector fields corresponding to the multiple time-phase reconstructed images based on the multiple time-phase reconstructed images and the average template image.

[0016] For example, the step of extracting prior information of the feature space based on the multiple time-phase reconstructed images includes: extracting prior information of the feature space using the average template image and the multiple motion displacement vector fields.

[0017] For example, the step of extracting prior information of the feature space using the average template image and the plurality of motion displacement vector fields includes: downsampling the average template image and the plurality of motion displacement vector fields to obtain a downsampled average template image and a plurality of motion displacement vector fields; and extracting prior information of the feature space using the downsampled average template image and the plurality of motion displacement vector fields.

[0018] For example, the prior information of the feature space includes first prior information and second prior information; the step of extracting the prior information of the feature space using the average template image and the plurality of motion displacement vector fields includes: extracting the first prior information using the plurality of motion displacement vector fields; and extracting the second prior information using the average template image.

[0019] For example, the feature space information extraction model includes a first sub-model and a second sub-model. The first sub-model is used to extract a first feature based on the first prior information, and the second sub-model is used to extract a second feature based on the second prior information.

[0020] For example, the spatiotemporal continuous representation model includes a motion displacement vector field sub-model and an average template image sub-model; embedding the prior information of the feature space into the spatiotemporal continuous representation model includes: using the output of the first sub-model as the input of the motion displacement vector field sub-model; and using the output of the second sub-model as the input of the average template image sub-model.

[0021] For example, obtaining a reconstructed image of the object to be imaged using the dynamic projection data and a spatiotemporally continuous representation model embedded with the prior information includes: using the plurality of motion displacement vector fields as input to a first sub-model, and the first sub-model extracting a first feature based on the first prior information; using the first feature as input to the motion displacement vector field sub-model, and the motion displacement vector field sub-model processing the first feature to output template image spatial coordinates; using the template image spatial coordinates as input to a second sub-model, and the second sub-model extracting a second feature based on the second prior information; using the second feature as input to the average template image sub-model, and the average template image sub-model processing the second feature to output an intermediate reconstructed image.

[0022] For example, obtaining a reconstructed image of the object to be imaged using the dynamic projection data and a spatiotemporal continuous representation model embedding the prior information further includes: calculating intermediate projection data of the intermediate reconstructed image based on a forward projection model of multiple temporal phases; and performing optimization based on the intermediate projection data and the dynamic projection data to obtain a reconstructed image of the object to be imaged.

[0023] For example, the optimization solution based on the intermediate projection data and the dynamic projection data includes: freezing the parameters of the first sub-model and the second sub-model during the optimization solution process, and optimizing the parameters of the motion displacement vector field sub-model and the average template image sub-model; or, freezing the parameters of the first sub-model and the second sub-model within a first dimension range during the optimization solution process, and optimizing the parameters of the first sub-model and the second sub-model within a second dimension range, as well as the parameters of the motion displacement vector field sub-model and the average template image sub-model, wherein the dimensions in the first dimension range are lower than a preset dimension value, and the dimensions in the second dimension range are higher than the preset dimension value.

[0024] For example, both the motion displacement vector field sub-model and the average template image sub-model include a multilayer perceptron.

[0025] For example, the first time window is larger than the second time window.

[0026] For example, the ratio of the first time window to the second time window is greater than 100.

[0027] For example, the object to be imaged includes biological tissue or industrial parts.

[0028] In another aspect, a scanning imaging system is provided, the system comprising: an imaging channel, at least a portion of which extends along a first direction, the imaging channel being used to place an object to be imaged during imaging; a radiation source for emitting radiation; a detector for detecting radiation emitted from the radiation source and passing through the object to be imaged, and generating projection data corresponding to the detected radiation; and a data processing device for generating a reconstructed image based on the projection data using the method described above.

[0029] For example, the scanning imaging system includes a CT scanning imaging system.

[0030] In the described method and system, a coarse-to-fine strategy is adopted. For the object to be imaged, coarse reconstruction is first performed on both spatial and temporal scales. Prior information is fully extracted through data-driven learning and efficiently integrated into the spatiotemporal high-resolution fine modeling and reconstruction process. Specifically, prior knowledge of the object to be imaged and its motion process is extracted through data-driven learning, providing effective prior information in the image space and the feature space of the novel representation method. This information is further embedded into the optimization process of the novel representation method, thereby achieving fine modeling. Subsequently, the novel representation method is used to perform high-precision modeling of the object and its motion process, fully combining dynamic projection data and data-driven prior information to achieve high spatiotemporal resolution image reconstruction. Attached Figure Description

[0031] The above and other objects, features and advantages of this application will become clearer from the following description of embodiments with reference to the accompanying drawings, in which:

[0032] Figure 1 A schematic diagram illustrating the projection relationship between the X-ray source, the object to be imaged, and the detector is shown.

[0033] Figure 2A This is a schematic diagram of the structure of a scanning imaging system according to some exemplary embodiments of this application;

[0034] Figure 2B This is a schematic diagram of the structure of a scanning imaging system according to some other exemplary embodiments of this application;

[0035] Figure 3A This is a schematic diagram of the structure of a slip ring CT scanning imaging system according to some exemplary embodiments of this application;

[0036] Figure 3B A planar schematic diagram of a linear trajectory scan performed in a scanning imaging system according to an embodiment of this application is shown;

[0037] Figure 3C This is a schematic diagram of the structure of a CT scanning imaging system based on a three-segment coplanar scanning structure according to some exemplary embodiments of this application;

[0038] Figure 4 This is a basic flowchart of the data-driven learning stage and the high spatiotemporal resolution reconstruction stage used in the scanned image reconstruction method according to the embodiments of this application.

[0039] Figure 5 This is a flowchart of a scanned image reconstruction method according to an embodiment of this application;

[0040] Figure 6 This illustration schematically depicts application scenarios of a scanned image reconstruction method according to some exemplary embodiments of this application, wherein, Figure 6 (a) schematically illustrates the motion cycle (e.g., cardiac cycle) of the object to be imaged (e.g., the heart). Figure 6 (b) schematically illustrates the acquisition cycle of a CT scan imaging system. Figure 6 (c) schematically illustrates the first time window used for coarse reconstruction. Figure 6 (d) schematically illustrates the second time window used for fine reconstruction;

[0041] Figure 7 The spatiotemporal continuous representation model used in the data-driven learning phase of a scanned image reconstruction method according to some exemplary embodiments of this application is illustrated schematically.

[0042] Figure 8 The illustration schematically shows the spatiotemporal continuous representation model used in the high spatiotemporal resolution reconstruction stage of a scanned image reconstruction method according to some exemplary embodiments of this application;

[0043] Figure 9 A block diagram of an electronic device suitable for implementing the method according to an embodiment of this application is illustrated schematically. Detailed Implementation

[0044] The embodiments of this application will now be described with reference to the accompanying drawings. However, it should be understood that these descriptions are exemplary only and are not intended to limit the scope of this application. In the following detailed description, numerous specific details are set forth to provide a thorough understanding of the embodiments of this application for ease of explanation. However, it will be apparent that one or more embodiments may be implemented without these specific details. Furthermore, descriptions of well-known structures and technologies are omitted in the following description to avoid unnecessarily obscuring the concepts of this application.

[0045] The terminology used herein is for the purpose of describing particular embodiments only and is not intended to limit the scope of this application. The terms “comprising,” “including,” etc., as used herein indicate the presence of the stated features, steps, operations, and / or components, but do not exclude the presence or addition of one or more other features, steps, operations, or components.

[0046] All terms used herein (including technical and scientific terms) have the meanings commonly understood by those skilled in the art, unless otherwise defined. It should be noted that the terms used herein are to be interpreted in a manner consistent with the context of this specification, and not in an idealized or overly rigid way.

[0047] When using expressions such as "at least one of A, B and C", they should generally be interpreted in accordance with the meaning that is commonly understood by those skilled in the art (e.g., "a system having at least one of A, B and C" should include, but is not limited to, a system having A alone, a system having B alone, a system having C alone, a system having A and B, a system having A and C, a system having B and C, and / or a system having A, B and C, etc.).

[0048] It should be noted that, in this application, computed tomography (CT) imaging refers to the process of using X-rays to perform a tomographic scan of the object to be imaged, then converting the analog signals received by the detector into digital signals, calculating the attenuation coefficient of each pixel by an electronic computer, and then reconstructing the image, thereby displaying the tomographic structure of each part of the object to be imaged.

[0049] The embodiments of this application will be described in detail below using a scanning imaging system for a specific tissue of a living organism or an industrial part as an example. It should be understood that the embodiments of this application are not limited to imaging scenarios for a specific tissue of a living organism or an industrial part; they can be applied to various scanning imaging scenarios. For example, they can be applied to scanning imaging scenarios involving various different inspection objects, including but not limited to vehicle scanning imaging, luggage / parcel scanning imaging, human or animal scanning imaging, organ / tissue scanning imaging, small object scanning imaging, and large object scanning imaging such as containers. It should be noted that the description of scanning imaging scenarios here is not exhaustive, and the exemplary descriptions below should not be construed as limiting the scope of protection of this application.

[0050] Figure 1 A schematic diagram illustrating the projection relationship between the X-ray source, the object to be imaged, and the detector. (Refer to...) Figure 1 In the embodiments of this application, rays emitted from the ray source 1 (e.g., X-rays, gamma rays, etc.) are incident on the object to be imaged 120, and the rays transmitted through the object to be imaged 120 are detected by the detector 2. A spatial point P1 on the object to be imaged is projected onto an image point P2 on the detector 2 via the ray source 1. In forward projection (also known as front projection), the pixel values ​​of the spatial points on the object to be imaged 120 are known, and the projection values ​​of the image points on the detector 2 are obtained. In backward projection (also known as back projection), the projection values ​​of the image points on the detector 2 are known, and the pixel values ​​of the spatial points on the object to be imaged 120 are obtained.

[0051] In some exemplary embodiments, the scanning imaging system may include only one scanning stage, i.e., only one source detection system. Figure 2A This is a schematic diagram of the structure of a scanning imaging system according to some exemplary embodiments of this application. (Refer to...) Figure 2AThe scanning imaging system according to embodiments of this application may include a scanning stage, a transport mechanism 110, a control device 140, and a data processing device 130. For example, the scanning stage may include an X-ray source and a detector.

[0052] Figure 2B This is a schematic diagram of the structure of a scanning imaging system according to some other exemplary embodiments of this application. (Refer to...) Figure 2B The scanning imaging system according to embodiments of this application may include multiple scanning stages (e.g., first scanning stage A, second scanning stage B, ...), a transport mechanism 110, a control device 140, and a data processing device 130. For example, the scanning stage may include an X-ray source and a detector.

[0053] Figure 3A This is a schematic diagram of the structure of a slip-ring CT scanning imaging system according to some exemplary embodiments of this application. Figure 3A As shown, a scanning imaging system according to some exemplary embodiments of this application may include: an X-ray source 1, a slip ring 150, a detector 2, a transmission mechanism 110, a control device 140, and a data processing device 130, etc. It should be understood that the X-ray source 1, the detector 2, and the slip ring 150 can be configured... Figure 2A or Figure 2B One of the scan levels.

[0054] For example, the slip ring 150 may include a rotating end and a fixed end. The rotating end of the slip ring 150 may include a conductive brush and a conductive ring, with a cantilever spring maintaining contact pressure to ensure reliable electrical contact during rotation. The fixed end of the slip ring 150 may include a data receiving electrode plate, forming capacitive coupling with the rotating end for non-contact transmission, avoiding signal attenuation caused by mechanical wear. The X-ray source 1 and detector 2 may be mounted on the slip ring 150. Specifically, the X-ray source 1 and detector 2 may be located at the rotating end of the slip ring 150; for example, the X-ray source 1 and detector 2 may be symmetrically fixed to the rotating end of the slip ring 150. Driven by the slip ring 150, the X-ray source 1 and detector 2 can rotate around a predetermined rotation axis. In the embodiments of this application, the scanning imaging system can perform CT scans on the object 120 to be imaged within a field of view centered on a predetermined rotation axis. Exemplarily, the object 120 to be imaged may include luggage, packages, biological tissue, industrial parts, etc.; however, the embodiments of this application do not impose particular limitations on the specific form of the object to be imaged.

[0055] For example, the radiation source 1 can be an X-ray machine. The detector 2 can be used to receive at least a portion of the radiation beam emitted by the X-ray machine. The detector 2 can include multiple detector modules for detecting the radiation beam passing through the object to be imaged 120, obtaining analog signals, and converting the analog signals into digital signals to output projection data of the object to be imaged 120 relative to the X-rays. The control device 140 is used to control the synchronous operation of various parts of the entire system. The data processing device 130 is used to process the projection data acquired by the detector 2, process and reconstruct the projection data, and output the results (e.g., a reconstructed image).

[0056] like Figure 3A As shown, X-ray source 1 is placed on one side of the object 120 to be imaged, and detector 2 is placed on the other side of the object 120. During the operation of the scanning imaging system, the focal point of X-ray source 1 emits a X-ray beam, and detector 2 acquires the transmission data and / or multi-angle projection data of the object 120. For example, the data acquisition unit in detector 2 may include a data amplification and shaping circuit, which can operate in (current) integration mode or pulse (counting) mode. The data output cable of detector 2 is connected to control device 140 and data processing device 130, and the acquired data is stored in data processing device 130 according to trigger commands.

[0057] The transmission mechanism 110 carries the object to be imaged 120 and passes through the scanning area between the X-ray source 1 and the detector 2 via the slip ring 150. Simultaneously, the slip ring 150 rotates around the translational direction of the object to be imaged 120, allowing the X-ray beam emitted by the X-ray source 1 to pass through the object to be imaged 120 and perform a CT scan. In other words, the transmission mechanism 110 can cause relative motion between the X-ray source 1 and the object to be imaged 120, causing the X-ray source 1 to move relative to the object to be imaged 120 along a predetermined scanning trajectory.

[0058] In some embodiments, during the scanning process of the X-ray source 1, the slip ring 150 drives the X-ray source 1 and detector 2 to rotate, while the object to be imaged 120 remains stationary; that is, the X-ray source 1 moves along a circular trajectory relative to the object to be imaged 120. In other embodiments, during the scanning process of the X-ray source 1, the slip ring 150 drives the X-ray source 1 and detector 2 to rotate, and the object to be imaged 120 translates under the drive of the conveying mechanism 110; that is, the X-ray source 1 moves along a helical trajectory relative to the object to be imaged 120.

[0059] Figure 3B A planar schematic diagram of linear trajectory scanning performed in a scanning imaging system according to an embodiment of this application is shown. Figure 2A , Figure 2B and Figure 3BAs shown, the object to be imaged 120 moves along a straight trajectory between the X-ray source 1 and the detector 2. During this movement, the X-ray source 1 emits X-rays according to the commands of the control system, penetrating the object to be imaged 120. The detector 2 receives the transmitted signal and acquires projection data under the control of the control device 140. It should be noted that... Figure 3B The scanning structure shown, including the radiation source and detector, can correspond to... Figure 2A and Figure 2B One scan level is shown in the image.

[0060] like Figure 3B As illustrated, the radiation source 1 may, by way of example, include an X-ray accelerator, an X-ray machine, or a radioactive isotope, as well as corresponding auxiliary equipment. Optionally, to make the horizontal beam angle (i.e., the beam fan angle) greater than 90 degrees, for example, between 90 and 180 degrees, two or more radiation sources may be used, selected according to the size of the object 120 to be imaged and the application scenario.

[0061] The conveying mechanism 110, such as a conveyor belt, can carry and smoothly transport the object to be imaged 120, and is used to move the carried object 120 along a straight trajectory during the scanning imaging process. Optionally, the conveying mechanism 110 moves the X-ray source and detector along a straight trajectory during the scanning imaging process, or moves the object to be imaged towards the X-ray source and detector. That is, the movement of the object to be imaged and the movement of the X-ray source and detector are relative movements.

[0062] Detector 2 may include a detector array for acquiring transmission projection data of rays by receiving rays that pass through the object to be imaged. Detector 2 may also include readout circuitry and a logic control unit for reading out the projection data on the detector array. The detector array may include multiple solid-state detector units, multiple gas detector units, or multiple semiconductor detector units.

[0063] In some exemplary embodiments of this application, the scanning imaging system may further include a multi-segment coplanar scanning structure. Below, a CT scanning imaging system with a three-segment scanning structure is used as an example to describe the embodiments of this application in detail. It should be understood that the following CT scanning imaging system based on a three-segment coplanar scanning structure is merely an exemplary structure for illustrative purposes. Those skilled in the art should understand that the number and position of the X-ray source and detector can be adjusted according to specific needs to accommodate objects of different sizes and shapes to be imaged.

[0064] Figure 3C This is a schematic diagram of a CT scanning imaging system based on a three-segment coplanar scanning structure according to some exemplary embodiments of this application. It should be noted that... Figure 3CThe scanning structure shown, including the radiation source and detector, can correspond to... Figure 2A and Figure 2B One scan level is shown in the image.

[0065] Combined with reference Figure 2A , Figure 2B and Figure 3C A scanning imaging system according to some exemplary embodiments of this application may include: a radiation source 1, a detector 2, and an imaging channel 160, etc. The imaging channel 160 extends along a first direction z and is used to place the object to be imaged 120 during the imaging process. For example, at least a portion of a support device may be disposed in the imaging channel 160, and the object to be imaged 120 may be placed on the support device. During scanning, the object to be imaged 120 may remain stationary in the imaging channel 160. As another example, at least a portion of a transport mechanism 110 is disposed in the imaging channel 160. The transport mechanism 110 carries the object to be imaged 120 through the scanning area between the radiation source 1 and the detector 2. The transport mechanism 110 can move the object to be imaged 120 along the first direction z, while the radiation emitted by the radiation source 1 can pass through the object to be imaged 120 to perform a CT scan.

[0066] Combined with reference Figure 2A , Figure 2B and Figure 3C In some exemplary embodiments, the X-ray source 1 of the CT scanning imaging system may include multiple distributed X-ray sources, for example, M distributed X-ray sources, where M is a positive integer greater than or equal to 2. Figure 3C In the exemplary embodiment shown, M=5, that is, five distributed radiation sources are provided. In this application, for ease of description, the five distributed radiation sources are respectively numbered as the first distributed radiation source 11, the second distributed radiation source 12, the third distributed radiation source 13, the fourth distributed radiation source 14, and the fifth distributed radiation source 15.

[0067] In the embodiments of this application, the distributed X-ray source includes multiple target points (also known as focal points). For example, multiple X-ray target points can be arranged at a high density within a single X-ray tube vacuum chamber. Compared to a traditional single X-ray source, a distributed X-ray source has multiple independent X-ray emission points, which can operate independently and be controlled individually. For example, the distributed X-ray source developed by the applicant of this application can integrate hundreds of X-ray target points within a single X-ray tube, each of which can be independently controlled and quickly switched as needed. For example, the emission of each target point can be precisely controlled, including parameters such as X-ray intensity and emission time, allowing for flexible adjustment of X-ray output based on different scanning locations, object density, and other factors during CT scans, thereby obtaining higher-quality images. For example, the intensity distribution of the X-ray beam can be modulated by controlling the X-ray intensity of different target points to adapt to the imaging requirements of different objects. Furthermore, by employing a multi-target distributed X-ray source, rapid scanning imaging capabilities can be achieved. Due to the presence of multiple X-ray target points, the distributed X-ray source can simultaneously or rapidly scan target objects sequentially from multiple different locations, greatly improving scanning speed. Compared to traditional CT scans, this method can acquire complete image information of an object in a shorter time, which is significant for scenarios requiring rapid imaging, such as medical and security checks. Furthermore, it enables high-quality imaging results; the collaborative work of multiple targets and precise control allow distributed X-ray sources to provide higher resolution and lower noise images. When imaging complex objects or minute structures, it can more clearly reveal the object's internal structure and details, providing more accurate information for doctors' diagnoses and other applications.

[0068] In some exemplary embodiments, the CT scanning imaging system may include multiple detectors 2, for example, N detectors, where N is a positive integer greater than or equal to 2. Figure 3C In the exemplary embodiment shown, N=3, that is, three detectors are provided. In this application, for ease of description, the three detectors are respectively numbered as the first detector 21, the second detector 22, and the third detector 23.

[0069] Reference Figure 3C In some exemplary embodiments, among the M distributed radiation sources, at least one distributed radiation source includes q target points 101, where q is a positive integer greater than or equal to 2. For example, each of the M distributed radiation sources includes more than two target points. Figure 3C In the illustrated embodiment, q=6, meaning a distributed radiation source comprises 6 target points. Exemplarily, the q target points are configured to be activated in a predetermined order to emit radiation.

[0070] It should be understood that the five distributed X-ray sources and three detectors mentioned here are merely exemplary embodiments and should not be construed as limiting the number of distributed X-ray sources and detectors in the embodiments of this application. In other embodiments, the CT scanning imaging system may include fewer distributed X-ray sources and detectors, such as two, or more, such as six, eight, ten, or more. When more than two distributed X-ray sources and detectors are set, X-rays can be emitted and received simultaneously or sequentially from different angles and positions, thereby improving the speed and resolution of imaging. Alternatively, three, four, or five distributed X-ray sources and detectors can be set. The combination of more than three distributed X-ray sources and detectors can provide a more comprehensive and detailed scan for imaging complex objects. For example, setting six distributed X-ray sources and detectors can achieve X-ray coverage over a larger area, which has significant advantages for imaging large objects or specific application scenarios requiring high resolution. In some cases, more than eight distributed X-ray sources and detectors can be set. The configuration of more than eight distributed X-ray sources and detectors can realize more complex scanning modes and multi-angle imaging, providing richer image information for security inspection scenarios, doctors' diagnoses, and researchers' analyses. It should be understood that as the number of distributed X-ray sources increases, CT scanning imaging systems can achieve more refined imaging, making them more adept at observing minute structures and performing high-precision detection tasks. In the embodiments of this application, the number of distributed X-ray sources and detectors can be flexibly adjusted according to specific application needs to meet the CT imaging requirements of different fields and scenarios.

[0071] It should also be understood that Figure 3C The distributed X-ray source in the figures, including six target points, and the number of target points shown in other figures are merely exemplary embodiments and should not be construed as a limitation on the number of target points in the embodiments of this application. In other embodiments, the distributed X-ray source of the CT scanning imaging system may include fewer target points, such as fewer than five, or more target points, such as 10, 20, 50, 100, hundreds, or more. In the embodiments of this application, the number of target points in the distributed X-ray source can be flexibly adjusted according to specific application requirements to meet the CT imaging requirements in different fields and scenarios.

[0072] It should be noted that, unless otherwise specified, the terms "first," "second," "third," etc., used in this application are merely for the convenience of referring to different components, such as the distributed radiation source and detector mentioned here, and should not be construed as imposing any limitation on the structure of the distributed radiation source and detector. In the embodiments of this application, the structures of the first distributed radiation source 11, the second distributed radiation source 12, the third distributed radiation source 13, the fourth distributed radiation source 14, and the fifth distributed radiation source 15 may be identical, or some of the distributed radiation sources may have the same structure while others may have different structures. For example, the first distributed radiation source 11, the second distributed radiation source 12, the third distributed radiation source 13, the fourth distributed radiation source 14, and the fifth distributed radiation source 15 may include the same number of target points, or the number of target points included in some of the distributed radiation sources may differ from the number of target points included in others, or any two distributed radiation sources may include different numbers of target points. In the embodiments of this application, the first detector 21, the second detector 22 and the third detector 23 may have the same structure, or some of the detectors may have the same structure while the other part of the detectors may have different structures.

[0073] Combined with reference Figure 2A , Figure 2B and Figure 3C In this application, for ease of description, the first direction z, the second direction x, the third direction y, the circumferential direction c of the imaging channel, and the radial direction r are described respectively. For example, the first direction z corresponds to the extension direction of the imaging channel 160, or, when the object to be imaged 120 moves, its direction of movement. The second direction x and the third direction y intersect, and the first direction z is perpendicular to both the second direction x and the third direction y. For example, the second direction x corresponds to the width direction of the scanning imaging system, and the third direction y corresponds to the height direction of the scanning imaging system. The circumferential direction c of the imaging channel intersects with the radial direction r, and the first direction z is perpendicular to both the circumferential direction c and the radial direction r of the imaging channel. For example, the radial direction r of the imaging channel 160 corresponds to the direction from the center of the imaging channel 160 to the outer periphery of the imaging channel 160, and the circumferential direction c of the imaging channel 160 corresponds to the direction surrounding the imaging channel 160. It should be noted that, in the embodiments of this application, the definitions and descriptions of various coordinate systems and directions are merely exemplary descriptions for the convenience of describing the embodiments of this application, and they are not intended to limit the embodiments of this application.

[0074] Reference Figure 2A , Figure 2B and Figure 3CFive distributed X-ray sources, namely the first distributed X-ray source 11 to the fifth distributed X-ray source 15, are arranged along the circumferential direction c of the imaging channel 160. Exemplarily, the five distributed X-ray sources are arranged at intervals along the circumferential direction c of the imaging channel 160. Three detectors, namely the first detector 21 to the third detector 23, are arranged along the circumferential direction c of the imaging channel 160. Exemplarily, the first detector 21 to the third detector 23 are arranged continuously or at intervals along the circumferential direction c of the imaging channel 160.

[0075] It should be understood that Figure 3C A schematic diagram of a plane perpendicular to the first direction z, i.e. Figure 3C The plane of the paper is perpendicular to the first direction z. For example... Figure 3C As shown, in a plane perpendicular to the first direction z, five distributed ray sources 11 to 15 and three detectors 21 to 23 surround and form a closed or nearly closed polygon.

[0076] In the above embodiments, circular trajectory, spiral trajectory, single-segment straight line trajectory and multi-segment straight line trajectory are used as examples to describe the scanning imaging system. It should be noted that the scanning imaging system according to the embodiments of this application is not limited to this. For example, in other embodiments, the scanning imaging system according to the embodiments of this application may include a tomographic imaging system.

[0077] It should be understood that, in the embodiments of this application, during the scanning imaging process, the X-ray source 1 emits a X-ray beam to scan the object to be imaged 120; in response to the X-ray beam scanning the object to be imaged 120, the detector 2 can acquire the projection data of the object to be imaged 120. The data processing device 130 can process the projection data acquired by the detector 2, process and reconstruct the projection data, and generate a reconstructed image.

[0078] It should be noted that, in the embodiments of this application, during the X-ray scanning of the object to be imaged 120, at least one attribute of the object to be imaged 120 is dynamically changing. The at least one attribute of the object to be imaged 120 may include at least one of the following: structure, shape, and physical form. For example, the object to be imaged 120 may include at least one tissue of a living organism, such as a heart, whose structure and shape change periodically with its beating. As another example, the object to be imaged 120 may include an industrial part, such as an industrial part undergoing heat treatment, during which the physical form of at least a portion of the industrial part changes. For such an object to be imaged 120, when performing image reconstruction, it is necessary not only to consider the voxels in its spatial location but also to consider the characteristics of the voxels at each spatial location changing over time; therefore, high spatiotemporal resolution image reconstruction is required.

[0079] The inventors discovered through research that, with the increasing demand for high-precision imaging of high-speed motion processes, X-ray scanning imaging systems urgently need to improve the spatiotemporal resolution of imaging and image reconstruction; in other words, they need high spatiotemporal resolution image reconstruction methods. High spatiotemporal resolution image reconstruction requires both efficient representation methods for high-dimensional spaces to effectively address the massive parameter optimization challenges brought about by the dramatic increase in data dimensionality, and a full integration of data-driven prior knowledge to effectively address the challenges posed by incomplete data sampling.

[0080] In related technologies, image reconstruction methods typically employ discretized voxels to model the image object and solve for the linear attenuation coefficients corresponding to each discretized voxel, thereby achieving image reconstruction. When facing ill-conditioned problems such as finite-angle CT scans and sparse-view CT scans, on the one hand, artificially designed sparsity constraints are introduced in conjunction with iterative image reconstruction methods to reduce the ill-conditioned nature of the reconstruction problem; on the other hand, deep learning reconstruction methods can effectively extract valuable prior information from large-scale training data, improving the quality of reconstructed images through projection domain data completion, image domain post-processing, dual-domain joint optimization, and deep unfolding models. Recently, generative artificial intelligence methods, represented by diffusion models, have shown excellent results in missing data recovery and image inpainting. However, the above-mentioned image reconstruction methods based on discretized voxel representation have two main limitations: (1) As the spatiotemporal resolution increases, the parameter scale required for discretized voxel representation will increase significantly, resulting in limitations in computational efficiency, solution complexity, data storage, and representation accuracy; (2) Although existing deep learning methods can effectively mine data-driven prior information, most of them perform data recovery phase by phase, making it difficult to efficiently capture the correlation between time series; at the same time, as the data dimension increases, the computational consumption will increase significantly. In other words, in the existing methods for image reconstruction of the object to be imaged using discretized voxels, although the traditional discretized grid representation method can effectively combine data-driven prior knowledge, its representation accuracy for high spatiotemporal resolution scenes is insufficient, and the parameter utilization efficiency is low.

[0081] To address at least one aspect of the aforementioned problems, embodiments of this application provide a scanning image reconstruction method, the method comprising: acquiring dynamic projection data of the object to be imaged in response to a ray scanning of an object to be imaged; dividing the dynamic projection data into multiple temporal phases according to a predetermined first time window to obtain multiple sub-projection data located in the multiple temporal phases; processing the multiple sub-projection data using a pre-trained transient projection image reconstruction model to obtain multiple temporally phased reconstructed images; extracting prior information of a feature space based on the multiple temporally phased reconstructed images; embedding the prior information of the feature space into a spatiotemporally continuous representation model; and obtaining a reconstructed image of the object to be imaged using the dynamic projection data and the spatiotemporally continuous representation model embedded with the prior information.

[0082] In the embodiments of this application, a coarse-to-fine strategy is adopted. For the object to be imaged, coarse reconstruction is first performed on both spatial and temporal scales. Prior information is fully extracted through data-driven learning and efficiently integrated into the spatiotemporal high-resolution fine modeling and reconstruction process. Specifically, prior knowledge of the object to be imaged and its motion process is extracted through data-driven learning, providing effective prior information in the image space and the feature space of the novel representation method. This information is further embedded into the optimization process of the novel representation method, thereby achieving fine modeling. Subsequently, the novel representation method is used to perform high-precision modeling of the object and its motion process, fully combining dynamic projection data and data-driven prior information to achieve high spatiotemporal resolution image reconstruction.

[0083] In some exemplary embodiments of this application, high spatiotemporal resolution image reconstruction can be achieved using continuous spatiotemporal representation models such as Implicit Neural Representation (INR) and 3D Gaussian Splatting (3DGS). The inventors have found that continuous spatiotemporal representation models, such as INR and 3D Gaussian Splatting, overcome the limitations of traditional discretized grids and have significant potential in the field of high-dimensional scene reconstruction. For example, INR can represent continuous functions using a multilayer perceptron, constructing a mapping relationship from position coordinates to linear decay coefficients, which has significant advantages in representing the continuity and multi-order smoothness of high-dimensional space. As another example, 3DGS can represent 3D scenes using a series of Gaussian functions, effectively reducing the parameter scale of the representation and improving representation efficiency. Compared to traditional discretized representation methods, these continuous spatiotemporal representation models show greater potential in terms of parameter utilization efficiency and high-dimensional space representation accuracy.

[0084] In the embodiments of this application, the method for achieving high spatiotemporal resolution image reconstruction using a continuous spatiotemporal representation model can be divided into a data-driven learning stage and a high spatiotemporal resolution reconstruction stage.

[0085] It should be understood that the data-driven learning stage is the foundation for the spatiotemporal continuous representation model to achieve image reconstruction. Its role is to extract the continuous mapping rules of object structure and cross-object statistical commonalities from training projection data or scanning training data of similar objects. During execution, the model encodes the features of three-dimensional spatial coordinates, learns the correspondence between coordinates and object intensity values, and uses multi-object joint training to explore the structural commonalities between different scanned objects. It also optimizes model parameters through loss functions to achieve robust adaptation to measurement noise and data sparsity, ultimately outputting a set of model parameters and regularity representations with generalization capabilities, rather than directly generating reconstructed images.

[0086] The high spatiotemporal resolution reconstruction stage is the step that transforms the learned patterns into actual images. Its function is to generate high spatiotemporal resolution images of the object to be imaged based on the trained model. During execution, the dynamic projection data of the object to be reconstructed is input as constraints. The model can densely sample the full spatial coordinates of the object, inputting each coordinate point into the trained model and directly outputting the intensity value at the corresponding location. For dynamic CT scenes, it can also combine the learned spatiotemporal patterns to generate continuous images of different phases. The entire process does not rely on a fixed-resolution voxel grid, allowing for flexible output of reconstruction results at any resolution. Simultaneously, it relies on previously learned rules to suppress artifacts, ensuring the accuracy of image details.

[0087] Figure 4 This is a basic flowchart of the data-driven learning stage and the high spatiotemporal resolution reconstruction stage used in the scanned image reconstruction method according to embodiments of this application. (Refer to...) Figure 4 The data-driven learning stage may include steps S410 to S440, and the high spatiotemporal resolution reconstruction stage may include steps S450 to S480.

[0088] In the data-driven learning phase, in step S410, a training dataset can be constructed or obtained. For example, projection simulation or other methods can be used. Figures 2A to 3C The scanning imaging system shown acquires projection data of the entire motion process of the object 120 to be imaged, ground truth images at each time phase, and corresponding feature space information through actual experiments. For example, the feature space information can vary depending on the representation method used in the high spatiotemporal resolution reconstruction stage. For instance, for the INR model, the feature space information can be a position-encoded feature vector; for the 3DGS model, the feature space information can be a point cloud distribution.

[0089] In step S420, a transient projection image reconstruction model can be trained. For example, the scale of the reconstruction time window can be determined according to the temporal resolution requirements, and the projection data of the entire motion process can be divided into each time window. For instance, in a CT scanning imaging system, due to the limitations of CT scanning speed, there is a significant problem of missing projection data within each time window. The image reconstruction resolution is determined according to the spatial resolution requirements. By constructing and training the transient projection image reconstruction model, artifacts in the reconstructed images of each temporal phase can be suppressed, obtaining temporally segmented reconstructed images.

[0090] In step S430, a feature space information extraction model can be trained. For example, the learning objective for feature space information can be determined based on the representation method of the high spatiotemporal resolution reconstruction stage. The feature space information extraction model can learn prior information of the feature space from the reconstructed images of each time phase. For instance, if a spatiotemporal decoupling model is used, an anatomical structure model and a motion feature model need to be constructed separately. The prior information of the feature space for the anatomical structure model can be obtained by learning the average images of each time phase. The prior information of the feature space for the motion feature model can be obtained by estimating motion features such as the displacement vector field of the images of each time phase. As another example, if a spatiotemporal coupling model is used, the prior information of the feature space corresponding to different time phases can be directly obtained from the images of each time phase.

[0091] In step S440, a data distribution-aware image reconstruction model can be constructed and trained. For example, the data distribution-aware image reconstruction model can be constructed and trained based on the transient projection image reconstruction model and the feature space information extraction model described above.

[0092] In the high spatiotemporal resolution reconstruction stage, in step S450, dynamic projection data of the object to be imaged can be acquired. For example, this can be achieved using... Figures 2A to 3C The scanning imaging system shown acquires dynamic projection data of the object 120 to be imaged. Based on the scanning speed and spatial resolution of the scanning imaging system, the image reconstruction resolution and time window used for coarse reconstruction can be determined, dividing the dynamic projection data into different temporal phases, thus providing a foundation for subsequent refined modeling.

[0093] It should be noted that, in this application, for the sake of brevity, the term "temporal phase" is used to describe the time phase, which can represent the scanning phase divided into time windows during the scanning imaging of an object that changes dynamically over time.

[0094] In step S460, a spatiotemporally continuous representation model can be constructed. For example, this model can be constructed using representation methods such as INR or 3DGS. For instance, with "spatial three-dimensional + temporal" four-dimensional modeling as the core, a spatiotemporally coupled modeling approach can be used to directly describe the dynamic changes of an object; alternatively, a spatiotemporally decoupled modeling approach can be used to construct an anatomical structure model and a motion feature model of the object separately, and the motion feature model can be reduced in dimensionality using methods such as orthogonal decomposition to improve computational efficiency.

[0095] In step S470, pre-trained transient projection image reconstruction models and feature space information extraction models can be used to obtain time-phase reconstructed images and feature space prior information, respectively, and then embedded into the spatiotemporal continuous representation model. Specifically, the pre-trained transient projection image reconstruction model can process the CT projection data of each time phase to obtain time-phase reconstructed images; the pre-trained feature space information extraction model can extract feature space prior information. The extracted feature space prior information is embedded into the spatiotemporal continuous representation model to provide data-driven prior support.

[0096] In step S480, high spatiotemporal resolution image reconstruction can be achieved using dynamic projection data. For example, by integrating prior information from the feature space with dynamic projection data, the spatiotemporal continuous representation model is optimized to obtain a high spatiotemporal resolution reconstruction result. In some exemplary embodiments, to effectively preserve data-driven prior information, some feature parameters can be frozen, or they can be explicitly introduced as constraints into the objective function to further enhance the model's reconstruction accuracy and stability.

[0097] Figure 5 This is a flowchart of a scanned image reconstruction method according to an embodiment of this application. (Refer to...) Figure 5 The scanned image reconstruction method may include steps S510 to S560.

[0098] In step S510, in response to the ray scanning of the object to be imaged, dynamic projection data of the object to be imaged is acquired.

[0099] In some exemplary embodiments, it is possible to utilize Figures 2A to 3CThe scanning imaging system shown performs scanning imaging on the object 120 to be imaged. During the X-ray scanning of the object 120, at least one property of the object 120 is dynamically changing. The at least one property of the object 120 may include at least one of the following: structure, shape, and physical form. For example, the object 120 may include at least one tissue of a living organism, such as a heart, whose structure and shape change periodically with its beating. As another example, the object 120 may include an industrial part, such as an industrial part undergoing heat treatment, during which at least a portion of the industrial part undergoes a change in physical form.

[0100] It should be noted that the "dynamic" in "dynamic projection data" here can mean that at least one attribute of the object to be imaged 120 is dynamically changing during the process of ray scanning of the object to be imaged 120.

[0101] In some exemplary embodiments, in step S510, obtaining dynamic projection data of the object to be imaged may include: during the process of ray scanning of the object to be imaged, at least one attribute of the object to be imaged is dynamically changing; acquiring rays that transmit through the object to be imaged with at least one dynamically changing attribute at a second time window period to obtain dynamic projection data of the object to be imaged.

[0102] In step S520, the dynamic projection data is divided into multiple time phases according to a predetermined first time window to obtain multiple sub-projection data located in the multiple time phases.

[0103] Exemplarily, prior to step S520, the method may further include a step of determining a first time window. That is, determining the duration of the first time window. The method may further include: determining the first time window based on the scanning speed and spatial resolution of the scanning imaging system. For example, in the scenario of a slip-ring CT scanning imaging system, the first time window can be determined based on factors such as the movement speed of the slip ring, the acquisition speed of the detector, and the spatial resolution of the coarse reconstructed image. In some exemplary embodiments, the first time window may be larger than the second time window. For example, the ratio of the first time window to the second time window is greater than 100. Thus, the reconstructed image obtained by dividing the first time window into multiple sub-projection data and then performing image reconstruction based on these multiple sub-projection data can be understood as a coarse reconstructed image; the reconstructed image obtained by reconstructing the actual dynamic projection data obtained through the second time window (e.g., the actual acquisition cycle) can be understood as a fine reconstructed image.

[0104] In some exemplary embodiments, the first time window may be substantially equal to the second time window. For example, in a static CT scanning imaging system, projection data from one or more viewpoints can be obtained at each moment. In this way, the discretized time window can be compressed to a single projection exposure, that is, the first time window can be determined as the duration of a single exposure. In this way, the temporal resolution accuracy of coarse reconstruction can be improved.

[0105] In step S530, the multiple sub-projection data are processed using a pre-trained transient projection image reconstruction model to obtain multiple time-phase reconstructed images.

[0106] For example, in step S530, the plurality of sub-projection data can be input into a pre-trained transient projection image reconstruction model; the transient projection image reconstruction model performs image reconstruction on the plurality of sub-projection data respectively to obtain the plurality of time-phase reconstructed images; an average template image is obtained based on the plurality of time-phase reconstructed images; and a plurality of motion displacement vector fields corresponding to the plurality of time-phase reconstructed images are obtained based on the plurality of time-phase reconstructed images and the average template image.

[0107] In the embodiments of this application, the method adopts a "coarse-to-fine" approach. First, it uses "coarse reconstruction" with discretized grid representation, combined with data-driven learning, to extract key features from large-scale data, thereby providing priors for spatiotemporally continuous "fine reconstruction". This effectively reduces the large-scale computing resource bottleneck of ultra-high resolution data-driven learning.

[0108] In step S540, prior information of the feature space is extracted based on the multiple time-phase reconstructed images.

[0109] For example, in step S540, the multiple time-series reconstructed images can be processed using a pre-trained feature space information extraction model to extract prior information of the feature space.

[0110] For example, the prior information of the feature space may include location-encoded feature vectors or point cloud distributions.

[0111] For example, the step of extracting prior information of the feature space based on the multiple time-phase reconstructed images may include: extracting prior information of the feature space using the average template image and the multiple motion displacement vector fields.

[0112] In some exemplary embodiments, the step of extracting prior information of the feature space using the average template image and the plurality of motion displacement vector fields may include: downsampling the average template image and the plurality of motion displacement vector fields to obtain a downsampled average template image and a plurality of motion displacement vector fields; and using the downsampled average template image and the plurality of motion displacement vector fields to extract prior information of the feature space.

[0113] In some exemplary embodiments, the prior information of the feature space includes first prior information and second prior information. Extracting the prior information of the feature space using the average template image and the plurality of motion displacement vector fields may include: extracting the first prior information using the plurality of motion displacement vector fields; and extracting the second prior information using the average template image.

[0114] In some exemplary embodiments, the feature space information extraction model may include a first sub-model and a second sub-model, wherein the first sub-model is used to extract a first feature based on the first prior information, and the second sub-model is used to extract a second feature based on the second prior information.

[0115] For example, the spatiotemporal continuous representation model may include a motion displacement vector field sub-model and an average template image sub-model.

[0116] In some exemplary embodiments, embedding the prior information of the feature space into the spatiotemporal continuous representation model includes: using the output of the first sub-model as the input of the motion displacement vector field sub-model; and using the output of the second sub-model as the input of the average template image sub-model.

[0117] In step S550, the prior information of the feature space is embedded into the spatiotemporal continuous representation model.

[0118] In some exemplary embodiments, the spatiotemporal continuous representation model may include: a model established in four dimensions, including three spatial dimensions and a time dimension, using a spatiotemporal coupling modeling approach.

[0119] In some exemplary embodiments, the spatiotemporal continuous representation model may include: a model established in four dimensions, including three spatial dimensions and a time dimension, using a spatiotemporal decoupling modeling approach.

[0120] In some exemplary embodiments, the spatiotemporal continuous representation model is used to directly describe the dynamic changes of the object to be imaged.

[0121] In some exemplary embodiments, the spatiotemporal continuous representation model includes an object anatomical structure model and a motion feature model.

[0122] In some exemplary embodiments, the spatiotemporal continuous representation model includes a spatiotemporal continuous representation model constructed using implicit neural representation or three-dimensional Gaussian sputtering.

[0123] In step S560, the reconstructed image of the object to be imaged is obtained by using the dynamic projection data and the spatiotemporal continuous representation model embedded with the prior information.

[0124] In some exemplary embodiments, step S560 may include: using the plurality of motion displacement vector fields as input to the first sub-model, and the first sub-model extracting a first feature based on the first prior information; using the first feature as input to the motion displacement vector field sub-model, and the motion displacement vector field sub-model processing the first feature to output template image spatial coordinates; using the template image spatial coordinates as input to the second sub-model, and the second sub-model extracting a second feature based on the second prior information; using the second feature as input to the average template image sub-model, and the average template image sub-model processing the second feature to output an intermediate reconstructed image.

[0125] In some exemplary embodiments, obtaining a reconstructed image of the object to be imaged using the dynamic projection data and a spatiotemporal continuous representation model embedding the prior information may further include: calculating intermediate projection data of the intermediate reconstructed image based on a forward projection model of multiple temporal phases; and performing optimization based on the intermediate projection data and the dynamic projection data to obtain a reconstructed image of the object to be imaged.

[0126] In some exemplary embodiments, the optimization solution based on the intermediate projection data and the dynamic projection data may include: freezing the parameters of the first sub-model and the second sub-model during the optimization solution process, and optimizing the parameters of the motion displacement vector field sub-model and the average template image sub-model. Optionally or additionally, the optimization solution based on the intermediate projection data and the dynamic projection data may include: freezing the parameters of the first sub-model and the second sub-model within a first dimension range during the optimization solution process, optimizing the parameters of the first sub-model and the second sub-model within a second dimension range, as well as the parameters of the motion displacement vector field sub-model and the average template image sub-model, wherein the dimensions in the first dimension range are lower than a preset dimension value, and the dimensions in the second dimension range are higher than the preset dimension value.

[0127] The embodiments of this application will be illustrated below using the heart of a living organism as an example of the object to be imaged 120.

[0128] In an exemplary embodiment, it is possible to utilize Figures 2A to 3CThe scanning imaging system shown scans the heart region of an organism. For example, it can be used... Figure 3A The slip-ring CT scanning imaging system shown scans the heart region of an organism.

[0129] It should be understood that during a scan, the heart undergoes periodic movements according to a specific cardiac cycle. The cardiac cycle represents the time required for the heart to complete one full contraction and relaxation cycle; that is, the interval between the start of one heartbeat and the start of the next. A cardiac cycle consists of two main phases: diastole and systole. During diastole, the atria and ventricles of the heart relax sequentially, and blood flows back from the veins to the atria and then into the ventricles. At this time, the heart is in a state of full blood flow, and this phase accounts for a large proportion of the cardiac cycle. During systole, the atria contract first, squeezing blood into the ventricles, followed by the ventricles contracting, pumping blood into the arteries, completing one ejection cycle.

[0130] It should also be understood that in CT scans of the heart, the heart is always beating. The morphology of the heart varies greatly at different cardiac phases (such as end-diastole and end-systole). It is necessary to accurately match the specific phase of the cardiac cycle for scanning and reconstruction in order to obtain a clear heart image without motion artifacts.

[0131] Figure 6 This illustration schematically depicts application scenarios of a scanned image reconstruction method according to some exemplary embodiments of this application, wherein, Figure 6 (a) schematically illustrates the motion cycle (e.g., cardiac cycle) of the object to be imaged (e.g., the heart). Figure 6 (b) schematically illustrates the acquisition cycle of a CT scan imaging system. Figure 6 (c) schematically illustrates the first time window used for coarse reconstruction. Figure 6 (d) in the diagram schematically illustrates the second time window used for fine reconstruction.

[0132] In some exemplary embodiments, reference is made to Figure 6 One cardiac cycle is 1 second. When using a slip-ring CT scanning imaging system to scan the heart region of a living organism, the slip ring rotates at a speed of 0.25 seconds per revolution. Each revolution scans 1440 angles, and 1440 ÷ 360 = 4, meaning there are 4 subdivisions within each degree. In 1 second, the gantry rotates 4 times, scanning a total of 5760 angles, covering a range of 1440 degrees (calculated using 360*4). In a slip-ring CT scanning imaging system, the detector acquires projection data at a certain acquisition cycle. For example, the acquisition cycle can be 0.17 ms, meaning the detector acquires one projection data point every 0.17 ms. Within one cardiac cycle, approximately 5882 projection data points can be acquired, i.e., dynamic projection data.

[0133] Reference Figure 6 In (c), the dynamic projection data can be divided into partially overlapping or non-overlapping temporal phases. It should be noted that in... Figure 6 In (c) of the diagram, two time phases are schematically shown, and the two time phases partially overlap; however, the embodiments of this application are not limited thereto. In other embodiments, at least two time phases may not overlap.

[0134] In some exemplary embodiments, a first time window can be determined based on the scanning speed and spatial resolution of the scanning imaging system. Then, the dynamic projection data is divided into multiple temporal phases according to the predetermined first time window. For example, in Figure 6 In the illustrated embodiment, based on the scanning speed and spatial resolution of the exemplary scanning imaging system, the first time window can be determined to be 22.22 ms. The projection data within 22.22 ms contains a total of 128 angles, covering a range of 32 degrees. That is, each phase lasts 22.22 ms.

[0135] In cardiac CT imaging, phases represent different time stages within the cardiac cycle, with each phase corresponding to a specific state during cardiac contraction or relaxation. A complete cardiac cycle can be divided into multiple consecutive phases, and at least one of the following—cardiac morphology, chamber size, or myocardial thickness—may differ significantly across different phases. For example, at end-diastole, the heart chambers are full of blood and at their largest volume, making it suitable for observing myocardial thickness and cardiac chamber structure; at end-systole, the ventricles shrink to their smallest volume, making it suitable for assessing cardiac systolic function.

[0136] In some exemplary embodiments, the second time window can be substantially equal to the acquisition period; for example, the second time window can be 0.17 ms. That is, the first time window used for coarse reconstruction is larger than the second time window used for fine reconstruction; in this example, the ratio of the first time window to the second time window is approximately 131.

[0137] In some exemplary embodiments, a spatiotemporally decoupled implicit neural representation (INR) method can be used to construct a spatiotemporally continuous representation model.

[0138] Figure 7 The illustration schematically shows the spatiotemporal continuous representation model used in the data-driven learning phase of a scanned image reconstruction method according to some exemplary embodiments of this application.

[0139] Reference Figure 7 The spatiotemporal continuous representation model can include two INR sub-models—a motion displacement vector field sub-model and an average template image sub-model. It should be noted that, in an exemplary embodiment, the motion displacement vector field sub-model is represented as... ,in, These are the learnable parameters of the model; the average template image submodel is represented as... ,in, These are the learnable parameters of the model.

[0140] For example, the motion displacement vector field sub-model and average template image sub-model Each of these can include a multilayer perceptron.

[0141] In some exemplary embodiments, the feature space information extraction model may include a first sub-model and a second sub-model, wherein the first sub-model is used to extract a first feature based on first prior information, and the second sub-model is used to extract a second feature based on second prior information. For example, in... Figure 7 In the example shown, the first sub-model is represented as ,in, These are the learnable parameters of the model, containing the first prior information. This first sub-model is used to determine the spatial location at time t. Converted into the first feature. The second sub-model is represented as... ,in, These are the learnable parameters of the model, containing second prior information. This second sub-model is used to convert the template image spatial coordinates... Converted to the second feature.

[0142] Continue to refer to Figure 7 Motion displacement vector field sub-model Used to characterize the position in space at time t. Location and template image spatial coordinates Mapping relationship: Average template image submodel Used to characterize the second feature With linear attenuation coefficient The mapping relationship between them. That is, the linear attenuation coefficient of the object to be imaged at any given time. .

[0143] It should be understood that the template image space can be interpreted as a baseline reference space in a dynamic scene. It can represent a fixed, static reference space, with the average template image corresponding to the spatial structure of the dynamic scene in its "average state" or "typical state," such as the average morphology of the heart over multiple cardiac cycles or the intermediate stable state of an object's motion. The template image space provides a unified reference benchmark for dynamic changes at all times, avoiding spatial chaos during motion modeling, and ensuring that the scene position at all times can be associated with this reference space through displacement mapping. For example, in cardiac CT scan imaging, the template image space can be the three-dimensional space of the heart at end-diastole (the stable state with the smallest amplitude of motion).

[0144] In some exemplary embodiments of this application, the first sub-model can be provided in a data-driven manner. Second sub-model It provides a wealth of prior information.

[0145] In the data-driven learning phase, under the aforementioned scanning and imaging conditions, historical dynamic CT reconstructed images or moving phantoms can be used to simulate and generate dynamic CT projection data. According to a set first time window, the dynamic CT projection data of the entire process is divided into various time phases. Optionally, for the projection data of each time phase, analytical reconstruction or iterative reconstruction methods can be used to obtain preliminary reconstructed images with discretized mesh modeling. Due to the very limited coverage of the projection angle, the preliminary reconstructed images contain severe artifacts. In some exemplary embodiments, convolutional neural networks or diffusion models can be used as the basic framework of the image reconstruction network, and a transient projection image reconstruction model can be trained on paired training datasets to obtain multiple time-phase reconstructed images with artifact suppression. Based on the obtained multiple time-phase reconstructed images, the dynamic reconstruction methods in the field can be referenced to refine the model for each training sample. Calculate the average template image Furthermore, based on the multiple time-series reconstructed images and the average template image, multiple motion displacement vector fields corresponding to the multiple time-series reconstructed images are obtained. Next, for each training sample... Using the average template image and motion displacement vector field By using the following formulas (1) and (2) for optimization, the optimized values ​​of the learnable parameters of each of the above models are obtained:

[0146] (1),

[0147] (2),

[0148] in, To utilize the average template image The obtained spatial position The pixel or voxel value, To utilize the average template image sub-model Second sub-model The obtained spatial position The pixel or voxel value, formula (1) represents the two learnable parameters obtained by optimization when the second norm of the above two values ​​is minimized. The value of ; To utilize the motion displacement vector field The obtained time t and spatial position eigenvalues ​​at that location To utilize the motion displacement vector field sub-model and the first sub-model The obtained time t and spatial position The eigenvalues, Equation (2) represents the two learnable parameters obtained by optimization when the second norm of the above two values ​​is minimized. The value of .

[0149] Using the above formulas (1) and (2), the optimized values ​​of the parameters of the first sub-model can be obtained. The optimized values ​​of the parameters of the second sub-model are For example, the first sub-model and the second sub-model can be feature encoding functions, and correspondingly, the feature encoding reference value corresponding to the first sub-model can be... The feature encoding reference value corresponding to the second sub-model is .

[0150] In some exemplary embodiments, in order to efficiently obtain reference values ​​of the feature space directly from the time-series reconstructed image, a feature space information extraction model can be trained based on a multilayer perceptron or a lightweight Transformer model. For example, the first sub-model can be represented as follows: The second sub-model can be represented as It can extract prior information (e.g., prior encoded feature information) from the motion displacement vector field sub-model and the average template image sub-model, respectively. This is similar to the position encoding function in the traditional INR model. Different, feature space information extraction models It is necessary to consider each training sample This involves constructing a mapping from spatial location to encoded features. Therefore, it requires the motion displacement vector field... and average template image As additional input, to perceive differences in anatomical structure and motion characteristics among different samples.

[0151] Optionally, in some exemplary embodiments, to avoid the limitations of high-dimensional input on computational complexity and video memory, the following can be implemented separately: and Downsampling was performed to obtain and Only the main key features are retained. During the training of the feature space information extraction model, the training objective function of the feature space information extraction model can be expressed by the following formulas (3) and (4):

[0152] (3),

[0153] (4),

[0154] in, The reference values ​​for the feature encoding corresponding to the second sub-model are as follows. To utilize the downsampled average template image Second sub-model The obtained spatial position The eigenvalues, formula (3) represent the learnable parameters obtained by optimization when the second norm of the above two values ​​is minimized. The value of ; The reference values ​​for the feature encodings corresponding to the first sub-model are as follows. To utilize the downsampled motion displacement vector field and the first sub-model The obtained time t and spatial position The eigenvalues, formula (4) represent the learnable parameters obtained by optimization when the second norm of the above two values ​​is minimized. The value of .

[0155] It should be noted that in some application scenarios, if it is necessary to obtain the coding features of the spatiotemporal continuous representation model and the prior knowledge of the multilayer perceptron at the same time, the feature space information extraction model can be skipped. Instead, for each test sample, the initial reference value of the spatiotemporal continuous representation model can be obtained directly from the average template image and motion displacement vector field obtained by the transient projection image reconstruction model in the test phase, referring to formulas (1) and (2). Then, the high spatiotemporal resolution image reconstruction can be completed directly based on the dynamic projection data.

[0156] It should be noted that the inventors' research revealed that while unsupervised optimization of neural model parameters or Gaussian function parameters in related technologies does not rely on pre-labeled training data and can achieve superior reconstruction results compared to traditional methods by leveraging the inherent advantages of representation methods, image reconstruction quality remains limited when dealing with highly ill-conditioned problems. Furthermore, these methods require re-optimization of parameters and scene reconstruction for different objects, resulting in lengthy reconstruction processes and low computational efficiency, limiting their widespread application in real-time reconstruction scenarios. Additionally, while novel spatiotemporally continuous representation methods such as INR or 3DGS can effectively reduce dimensionality in high-dimensional spaces, achieving sparse parameterized representation and continuous modeling, their scene-by-scene and object-by-object optimization reconstruction methods have limitations in computational efficiency. Moreover, due to differences in representation methods, existing technical approaches are difficult to explicitly combine with data-driven approaches, resulting in limited effectiveness in highly ill-conditioned CT reconstruction problems.

[0157] Figure 8 The illustration schematically shows the spatiotemporal continuous representation model used in the high spatiotemporal resolution reconstruction stage of a scanned image reconstruction method according to some exemplary embodiments of this application.

[0158] In the high spatiotemporal resolution reconstruction stage, the temporal and spatial resolution of the reconstruction results obtained in the data-driven stage are relatively coarse due to limitations in the discretized reconstruction time window and resolution. In practical applications, for each test individual, dynamic projection data acquired from actual scanning at each time step can be utilized. By combining data-driven prior information with INR's refined modeling capabilities, high spatiotemporal resolution reconstructed images can be obtained.

[0159] In some exemplary embodiments, the dynamic projection data acquired by actual scanning is used. Based on the data-driven learning phase setup, it is divided into various time phases. That is, according to... Figure 6 The first time window shown will dynamically project data. The data is divided into different time phases to obtain multiple sub-projection data located in multiple time phases.

[0160] The multiple sub-projection data are processed using a pre-trained transient projection image reconstruction model to obtain multiple time-phase reconstructed images. It should be noted that a low-resolution average template image can also be obtained. and motion displacement vector field To avoid the limitations of high-dimensional input on computation and memory, the average template image can be processed separately. and motion displacement vector field Perform downsampling to obtain the downsampled average template image. and motion displacement vector field .

[0161] Based on this, the first sub-model and the second sub-model mentioned above are obtained by combining the pre-trained feature space information extraction model, for example, as shown in the following formulas (5) and (6):

[0162] (5),

[0163] (6).

[0164] Subsequently, dynamic projection data was used. and spatiotemporal continuous representation models (such as) Figure 8 As shown), the forward projection model is defined by combining the positions of each temporal X-ray source and the detector. High spatiotemporal resolution reconstructed images can be obtained by solving the following optimization problems (e.g., as shown in equations (7), (8), or (9) below):

[0165] (7).

[0166]

[0167] In some exemplary embodiments, in conjunction with reference to Figure 8 And formula (7), multiple motion displacement vector fields As the first sub-model The input, and the first sub-model Based on the first prior information (e.g., at least the parameters in formula (7) Extracting the first feature (which reflects the prior information) yields... The first feature As a field model of motion displacement vector The input, and the motion displacement vector field sub-model Process the first feature to output the template image spatial coordinates The template image spatial coordinates As the second sub-model The input, and the second sub-model Based on the second prior information (e.g., at least the parameters in formula (7) Extracting the second feature (which reflects the prior information) yields the result. The second feature As an average template image sub-model The input, and the average template image sub-model The second feature is processed to output the intermediate reconstructed image, thus obtaining the intermediate reconstructed image. Then, it can be based on the forward projection model of multiple time phases. Calculate the intermediate projection data of the intermediate reconstructed image. For example, in formula (7), by Calculate intermediate projection data. Further, based on the intermediate projection data and the dynamic projection data... The reconstructed image of the object to be imaged is obtained by performing optimization and solving.

[0168] For example, in formula (7), the optimization solution can be understood as: obtaining the intermediate projection data and the dynamic projection data through optimization solution. Multiple learnable parameters when the second norm value is minimized The value of . Refer to Figure 8 By obtaining the multiple learnable parameters The optimized value can be obtained through the optimized spatiotemporal continuous representation model. Through this It is possible to obtain a high spatiotemporal resolution reconstructed image of the object to be imaged.

[0169] Optionally, in some exemplary embodiments, during the optimization process based on the intermediate projection data and the dynamic projection data, the parameters of the first sub-model and the second sub-model can be frozen, and the parameters of the motion displacement vector field sub-model and the average template image sub-model can be optimized.

[0170] For example, referring to formula (8) below, the parameters of the first sub-model and the second sub-model (e.g., encoding function feature parameters) can be completely frozen. and Only optimize the network parameters of the multilayer perceptron. and This preserves, to the greatest extent possible, the prior information driven by data.

[0171] (8).

[0172]

[0173] Optionally, in some exemplary embodiments, during the optimization solution process based on the intermediate projection data and the dynamic projection data, the parameters in the first dimension range of the first sub-model and the second sub-model can be frozen, and the parameters in the second dimension range of the first sub-model and the second sub-model, as well as the parameters of the motion displacement vector field sub-model and the average template image sub-model, can be optimized. The dimensions in the first dimension range are lower than a preset dimension value, and the dimensions in the second dimension range are higher than the preset dimension value.

[0174] For example, see formula (9) below, which can freeze only. and Encoding parameters related to low-dimensional features and Fine-tuning the encoding parameters related to high-dimensional features and Meanwhile, the parameters of the multilayer perceptron are optimized. and This improves the ability to recover high-frequency details.

[0175]

[0176] In the optimization process represented by formulas (8) and (9) above, data-driven prior knowledge and fidelity constraints of projected data can be effectively utilized.

[0177] In the embodiments of this application, by constructing a method framework that efficiently combines data-driven priors and high spatiotemporal resolution reconstruction, the challenges of insufficient modeling accuracy and the dramatic increase in data dimensionality with resolution in traditional discretization modeling methods are effectively overcome, and the challenge of spatiotemporally continuous modeling methods being unable to be effectively and deeply integrated with data-driven learning is solved.

[0178] In the embodiments of this application, the proposed "coarse-to-fine" approach first uses "coarse reconstruction" with discretized grid representation, combined with data-driven learning, to extract key features from large-scale data, thereby providing priors for spatiotemporally continuous "fine reconstruction" and effectively reducing the large-scale computing resource bottleneck of ultra-high resolution data-driven learning.

[0179] In the embodiments of this application, a novel method for efficiently embedding data-driven prior knowledge into a spatiotemporally continuous representation is proposed. By encoding feature embedding, the data-driven prior knowledge can be fully inherited, effectively preserving the ability of spatiotemporally continuous modeling to recover high spatiotemporal resolution detailed structures. Compared with the traditional framework of using data-driven reconstructed images as prior images, the method proposed in this application organically integrates the data-driven approach into a novel spatiotemporally continuous representation method, effectively avoiding the limitations on computational efficiency and high-frequency structure representation capabilities caused by alternating between continuous and discretized representations for optimization.

[0180] It should be noted that the numbering of each step in the above method is not a restriction on the order of the method. In the absence of conflict, the steps in the method can be executed in parallel or in a different order than that described in this application.

[0181] Figure 9 A block diagram of an electronic device suitable for implementing the methods described above, according to an embodiment of this application, is illustrated schematically. Figure 9The illustrated electronic device is merely an example and should not be construed as limiting the functionality or scope of use of the embodiments of this application. For example, either the data processing device 130 or the control device 140 may be implemented as the described electronic device. In embodiments of this application, the data processing device 130 and the control device 140 may be implemented separately; for example, the data processing device 130 may be implemented as... Figure 9 The illustrated electronic device, control unit 140, can be implemented as a display device with a display screen. In other embodiments, data processing unit 130 and control unit 140 can be implemented as an integrated device, that is, they can be unified into a single electronic device. For example, data processing unit 130 and control unit 140 can be implemented as... Figure 9 The illustrated electronic device is in the form of a display screen. The embodiments of this application do not impose particular limitations on the specific implementation of the data processing device 130 and the control device 140.

[0182] like Figure 9 As shown, an electronic device 900 according to an embodiment of this application includes a processor 901, which can perform various appropriate actions and processes according to a program stored in a read-only memory 902 or a program loaded from a storage portion 908 into a random access memory 903. The processor 901 may include, for example, a general-purpose microprocessor, an instruction set processor and / or an associated chipset and / or a dedicated microprocessor. The processor 901 may also include onboard memory for caching purposes. The processor 901 may include a single processing unit or multiple processing units for executing different steps of the method flow according to an embodiment of this application.

[0183] Random access memory 903 stores various programs and data required for the operation of electronic device 900. Processor 901, read-only memory 902, and random access memory 903 are interconnected via bus 904. Processor 901 executes various steps of the method flow according to embodiments of this application by executing programs stored in read-only memory 902 and / or random access memory 903. It should be noted that the programs may also be stored in one or more memories other than read-only memory 902 and random access memory 903. Processor 901 may also execute various steps of the method flow according to embodiments of this application by executing programs stored in said one or more memories.

[0184] According to embodiments of this application, the electronic device 900 may further include an input / output interface 905, which is also connected to a bus 904. The electronic device 900 may also include one or more of the following components connected to the input / output interface 905: an input section 906 including a keyboard, mouse, etc.; an output section 907 including a cathode ray tube, liquid crystal display, etc., and a speaker, etc.; a storage section 908 including a hard disk, etc.; and a communication section 909 including a model interface card such as a local area network card, modem, etc. The communication section 909 performs communication processing via a model such as the Internet. A driver 910 is also connected to the input / output interface 905 as needed. A removable medium 911, such as a disk, optical disk, magneto-optical disk, semiconductor memory, etc., is installed on the driver 910 as needed so that computer programs read from it can be installed into the storage section 908 as needed.

[0185] Embodiments of this application also provide a computer-readable storage medium, which may be included in the device / apparatus / system described in the above embodiments; or it may exist independently and not assembled into the device / apparatus / system. The computer-readable storage medium carries one or more programs, which, when executed, implement the method according to the embodiments of this application.

[0186] According to embodiments of this application, the computer-readable storage medium can be a non-volatile computer-readable storage medium, such as including but not limited to: portable computer disks, hard disks, random access memory, read-only memory, erasable programmable read-only memory, portable compact disk read-only memory, optical storage devices, magnetic storage devices, or any suitable combination thereof. In embodiments of this application, the computer-readable storage medium can be any tangible medium containing or storing a program that can be used by or in conjunction with an instruction execution system, apparatus, or device. For example, according to embodiments of this application, the computer-readable storage medium may include the read-only memory 902 described above, and / or random access memory 903, and / or one or more memories other than read-only memory 902 and random access memory 903.

[0187] Embodiments of this application also include a computer program product comprising a computer program containing program code for performing the methods shown in the flowchart. When the computer program product is run on a computer system, the program code is used to cause the computer system to implement the methods provided in the embodiments of this application.

[0188] In one embodiment, the computer program may rely on a tangible storage medium such as an optical storage device or a magnetic storage device. In another embodiment, the computer program may also be transmitted and distributed in the form of signals on a model medium, and downloaded and installed via the communication section 909, and / or installed from a removable medium 911. The program code contained in the computer program can be transmitted using any suitable model medium, including but not limited to: wireless, wired, etc., or any suitable combination thereof.

[0189] In embodiments of this application, the computer program can be downloaded and installed from the model via communication section 909, and / or installed from removable medium 911. When the computer program is executed by processor 901, it performs the functions defined in the system of embodiments of this application. According to embodiments of this application, the systems, devices, apparatuses, modules, units, etc., described above can be implemented by computer program modules.

[0190] According to embodiments of this application, program code for executing the computer programs provided in the embodiments of this application can be written in any combination of one or more programming languages. Specifically, these computational programs can be implemented using high-level procedural and / or object-oriented programming languages, and / or assembly / machine languages. The program code can be executed entirely on the user's computing device, partially on the user's device, partially on a remote computing device, or entirely on a remote computing device or server. In cases involving remote computing devices, the remote computing device can be connected to the user's computing device via any type of model, including a local area network (LAN) or a wide area network (WAN), or it can be connected to an external computing device (e.g., via the Internet using an Internet service provider).

[0191] The embodiments of this application have been described above. However, these embodiments are merely illustrative and not intended to limit the scope of this application. Although various embodiments have been described above, this does not mean that the measures in the various embodiments cannot be used advantageously in combination. Without departing from the scope of this application, those skilled in the art can make various substitutions and modifications, all of which should fall within the scope of this application.

Claims

1. A method for reconstructing a scanned image, characterized in that, The method includes: In response to a ray scan of the object to be imaged, dynamic projection data of the object to be imaged is acquired; According to a predetermined first time window, the dynamic projection data is divided into multiple time phases to obtain multiple sub-projection data located in the multiple time phases; The multiple sub-projection data are processed using a pre-trained transient projection image reconstruction model to obtain multiple time-phase reconstructed images, wherein the transient projection image reconstruction model is based on a convolutional neural network or a diffusion model. Based on the multiple time-phase reconstructed images, prior information of the feature space is extracted, wherein the step of extracting prior information of the feature space based on the multiple time-phase reconstructed images includes: processing the multiple time-phase reconstructed images using a pre-trained feature space information extraction model to extract the prior information of the feature space; The prior information of the feature space is embedded into the spatiotemporal continuous representation model, wherein the spatiotemporal continuous representation model includes a spatiotemporal continuous representation model constructed with implicit neural representation or three-dimensional Gaussian sputtering. Using the dynamic projection data and the spatiotemporal continuous representation model embedded with the prior information, a reconstructed image of the object to be imaged is obtained. The process of using a pre-trained transient projection image reconstruction model to process the multiple sub-projection data to obtain multiple time-phase reconstructed images includes: The multiple sub-projection data are input into a pre-trained transient projection image reconstruction model; The transient projection image reconstruction model performs image reconstruction on the multiple sub-projection data respectively to obtain the multiple time-phase reconstructed images; Based on the multiple time-phase reconstructed images, an average template image is obtained; Based on the multiple time-series reconstructed images and the average template image, obtain multiple motion displacement vector fields corresponding to the multiple time-series reconstructed images; The prior information of the feature space includes first prior information and second prior information; The step of extracting prior information in the feature space based on the multiple time-phase reconstructed images includes: extracting the first prior information using the multiple motion displacement vector fields; and extracting the second prior information using the average template image. The feature space information extraction model includes a first sub-model and a second sub-model. The first sub-model is used to extract a first feature based on the first prior information, and the second sub-model is used to extract a second feature based on the second prior information. The spatiotemporal continuous representation model includes a motion displacement vector field sub-model and an average template image sub-model; embedding the prior information of the feature space into the spatiotemporal continuous representation model includes: using the output of the first sub-model as the input of the motion displacement vector field sub-model; and using the output of the second sub-model as the input of the average template image sub-model; Using the dynamic projection data and a spatiotemporal continuous representation model embedding the prior information, a reconstructed image of the object to be imaged is obtained, including: The plurality of motion displacement vector fields are used as input to the first sub-model, and the first sub-model extracts the first feature based on the first prior information; The first feature is used as the input of the motion displacement vector field sub-model, and the motion displacement vector field sub-model processes the first feature to output the template image spatial coordinates; The template image spatial coordinates are used as input to the second sub-model, and the second sub-model extracts the second feature based on the second prior information; The second feature is used as input to the average template image sub-model, and the average template image sub-model processes the second feature to output an intermediate reconstructed image; Based on the forward projection model of multiple time phases, the intermediate projection data of the intermediate reconstructed image is calculated; The reconstructed image of the object to be imaged is obtained by optimizing the solution based on the intermediate projection data and the dynamic projection data.

2. The method according to claim 1, characterized in that, The process of acquiring the dynamic projection data of the object to be imaged includes: During the process of the ray scanning of the object to be imaged, at least one attribute of the object to be imaged changes dynamically; The system acquires rays that transmit through the object to be imaged as at least one attribute changes dynamically, using a second time window as the period, to obtain dynamic projection data of the object to be imaged.

3. The method according to claim 2, characterized in that, At least one attribute of the object to be imaged includes at least one of the structure, shape, and material form of the object to be imaged.

4. The method according to any one of claims 1-3, characterized in that, The spatiotemporal continuous representation model includes a model established in four dimensions, including three spatial dimensions and a time dimension, using a spatiotemporal coupling modeling approach.

5. The method according to any one of claims 1-3, characterized in that, The spatiotemporal continuous representation model includes a model established in four dimensions, including three spatial dimensions and a time dimension, using a spatiotemporal decoupling modeling approach.

6. The method according to claim 4, characterized in that, The spatiotemporal continuous representation model is used to directly describe the dynamic changes of the object to be imaged.

7. The method according to claim 5, characterized in that, The spatiotemporal continuous representation model includes an object anatomical structure model and a motion feature model.

8. The method according to any one of claims 1-3, characterized in that, The prior information of the feature space includes location-encoded feature vectors or point cloud distributions.

9. The method according to any one of claims 1-3, 6 and 7, characterized in that, The method further includes: determining the first time window based on the scanning speed and spatial resolution of the scanning imaging system.

10. The method according to claim 1, characterized in that, The optimization solution based on the intermediate projection data and the dynamic projection data includes: During the optimization process, the parameters of the first sub-model and the second sub-model are frozen, and the parameters of the motion displacement vector field sub-model and the average template image sub-model are optimized; or, During the optimization process, the parameters in the first dimension range of the first sub-model and the second sub-model are frozen, and the parameters in the second dimension range of the first sub-model and the second sub-model, as well as the parameters of the motion displacement vector field sub-model and the average template image sub-model, are optimized. The dimensions in the first dimension range are lower than the preset dimension value, and the dimensions in the second dimension range are higher than the preset dimension value.

11. The method according to claim 1 or 10, characterized in that, Both the motion displacement vector field sub-model and the average template image sub-model include a multilayer perceptron.

12. The method according to claim 2 or 3, characterized in that, The first time window is larger than the second time window.

13. The method according to claim 12, characterized in that, The ratio of the first time window to the second time window is greater than 100.

14. The method according to any one of claims 1-3, 6, 7 and 10, characterized in that, The object to be imaged includes biological tissue or industrial parts.

15. A scanning imaging system, characterized in that, The system includes: An imaging channel, at least a portion of which extends along a first direction, the imaging channel being used to place an object to be imaged during the imaging process; A radiation source, wherein the radiation source is used to emit radiation; A detector is used to detect rays emitted from the ray source and passing through the object to be imaged, and to generate projection data corresponding to the detected rays; A data processing apparatus for generating a reconstructed image based on the projection data using the method described in any one of claims 1-14.

16. The system according to claim 15, characterized in that, The scanning imaging system includes a CT scanning imaging system.

Images