Low-dose SPECT image reconstruction method and system

Abstract

The invention relates to a low-dose SPECT image reconstruction method and system. The method comprises the following steps: acquiring a plurality of projection images corresponding to a plurality of scanning angles when a target part of a patient is subjected to SPECT imaging scanning; converting the plurality of projection drawings into a plurality of first chordal diagrams, and inputting the plurality of first chordal diagrams into an image reconstruction model to obtain a plurality of second chordal diagrams; wherein the image reconstruction model is obtained by performing model training on a mask auto-encoder model MAE based on a plurality of sample chordal diagrams, a three-dimensional encoder and a three-dimensional decoder are adopted as an encoder and a decoder of the mask auto-encoder model MAE, and the plurality of sample chordal diagrams are obtained by processing a plurality of sample projection diagrams; and reconstructing based on the plurality of second chordal diagrams to obtain a fault diagram of the target part.

Description

Technical Field

[0001] The disclosed embodiments relate to the technical field of medical image processing, and in particular to a low-dose SPECT image reconstruction method and system. Background Art

[0002] Currently, Single Photon Emission Computed Tomography (SPECT) is a nuclear medicine imaging technology that can display the distribution of radioactive drugs in the body, thereby reflecting the functional status of organs. SPECT imaging technology acquires data by detecting the gamma rays emitted by radioactive tracers in the body and reconstructs images using these data. SPECT imaging is widely used in many fields such as heart disease, brain function assessment, and tumor diagnosis.

[0003] Tomographic reconstruction is a key step in SPECT imaging. The tomogram can show the structural information inside the tissue, which is crucial for the diagnosis of diseases. The reconstruction method is to recover the spatial distribution of the source distribution from the projection data. The quality of the tomogram is closely related to the projection image directly obtained by scanning, and the quality of the projection image is the basis for reconstructing the image.

[0004] However, conventional SPECT imaging usually requires the injection of a certain amount of radioactive tracers into the patient's body, which not only increases the patient's radiation dose but may also cause adverse reactions. In order to reduce the radiation impact of SPECT imaging on patients, the input dose or activity of the tracer needs to be reduced, but this will bring a series of image quality problems, such as Poisson noise, radiation artifacts, and decreased spatial resolution, which will lead to a decrease in the quality of the reconstructed image and affect the accuracy of clinical diagnosis. Summary of the invention

[0005] In order to solve the above technical problems or at least partially solve the above technical problems, the embodiments of the present disclosure provide a low-dose SPECT image reconstruction method and system.

[0006] In a first aspect, an embodiment of the present disclosure provides a low-dose SPECT image reconstruction method, comprising: Acquire multiple projection images corresponding to multiple scanning angles when performing SPECT imaging scanning on a target part of a patient; The multiple projection images are converted into multiple first chord images, and the multiple first chord images are input into an image reconstruction model to obtain multiple second chord images; wherein the image reconstruction model is obtained by training a masked autoencoder model MAE based on multiple sample chord images, the encoder and decoder of the masked autoencoder model MAE use a three-dimensional encoder and a three-dimensional decoder, and the multiple sample chord images are obtained by processing multiple sample projection images; A tomographic image of the target site is reconstructed based on the multiple second chord images.

[0007] In one embodiment, the low-dose scanning strategy of the multiple projection images during SPECT imaging scanning includes any one of the following strategies under a preset total scanning time: Sparsely sample the original multiple scanning angles to reduce the number of scanning angles that need to capture images during the scanning process, thereby allocating more scanning time to the scanning angle corresponding to each captured image; ii) sample the scanning time to maintain a complete sampling image of the multiple scanning angles and reduce the scanning time of each scanning angle.

[0008] In one embodiment, the 3D encoder and the 3D decoder both use 3D Vision Transformer.

[0009] In one embodiment, the plurality of first chord diagrams and the plurality of second chord diagrams are both three-dimensional chord diagrams.

[0010] In one embodiment, reconstructing a tomogram of the target site based on the multiple second chord diagrams includes: The target image reconstruction algorithm is used to reconstruct the multiple second chord images to obtain a tomographic image of the target part; wherein the target image reconstruction algorithm at least includes an image iterative reconstruction algorithm.

[0011] In one embodiment, the loss function adopted by the masked autoencoder model MAE includes a mean square error loss function.

[0012] In one embodiment, the target site includes any one of the liver, heart and brain.

[0013] In a second aspect, the present disclosure provides a low-dose SPECT image reconstruction system, comprising: An image acquisition module, used to acquire multiple projection images corresponding to multiple scanning angles when a target part of a patient is subjected to SPECT imaging scanning; An image reconstruction module is used to convert the multiple projection images into multiple first chord images, and input the multiple first chord images into an image reconstruction model to obtain multiple second chord images; wherein the image reconstruction model is obtained by training a masked autoencoder model MAE based on multiple sample chord images, the encoder and decoder of the masked autoencoder model MAE use a three-dimensional encoder and a three-dimensional decoder, and the multiple sample chord images are obtained by processing multiple sample projection images; An image processing module is used to reconstruct a tomographic image of the target part based on the multiple second chord images.

[0014] In a third aspect, an embodiment of the present disclosure provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the low-dose SPECT image reconstruction method described in any of the above embodiments.

[0015] In a fourth aspect, an embodiment of the present disclosure provides an electronic device, including: Processor; and Memory for storing computer programs; Wherein, the processor is configured to perform the low-dose SPECT image reconstruction method described in any of the above embodiments by executing the computer program.

[0016] Compared with the prior art, the technical solution provided by the embodiments of the present disclosure has the following advantages: The low-dose SPECT image reconstruction method and system provided by the disclosed embodiment obtain multiple projection images corresponding to multiple scanning angles when performing SPECT imaging scanning on the target part of the patient; convert the multiple projection images into multiple first chord images, and input the multiple first chord images into the image reconstruction model to obtain multiple second chord images; wherein the image reconstruction model is obtained by training the masked autoencoder model MAE based on multiple sample chord images, the encoder and decoder of the masked autoencoder model MAE use a three-dimensional encoder and a three-dimensional decoder, and the multiple sample chord images are obtained by processing multiple sample projection images; based on the multiple second chord images, a tomogram of the target part is reconstructed. The scheme of this embodiment can improve the quality of reconstructed images during SPECT imaging, such as low-dose imaging, thereby improving the accuracy of auxiliary clinical diagnosis, while reducing the radiation impact of SPECT imaging examinations on patients. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the present disclosure and, together with the description, serve to explain the principles of the present disclosure.

[0018] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, for ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative labor.

[0019] Figure 1 This is a flow chart of a low-dose SPECT image reconstruction method according to an embodiment of the present disclosure; Figure 2 is an exemplary chord diagram of an embodiment of the present disclosure; Figure 3 Schematic diagram of the architecture of the masked autoencoder model MAE according to an embodiment of the present disclosure; Figure 4 is a schematic diagram of a low-dose SPECT image reconstruction system according to an embodiment of the present disclosure; Figure 5 Schematic diagram of an electronic device according to an embodiment of the present disclosure. DETAILED DESCRIPTION

[0020] In order to more clearly understand the above-mentioned objectives, features and advantages of the present disclosure, the scheme of the present disclosure will be further described below. It should be noted that the embodiments of the present disclosure and the features in the embodiments can be combined with each other without conflict.

[0021] In the following description, many specific details are set forth to facilitate a full understanding of the present disclosure, but the present disclosure may also be implemented in other ways different from those described herein; it is obvious that the embodiments in the specification are only part of the embodiments of the present disclosure, rather than all of the embodiments.

[0022] It should be understood that, in the following, "at least one (item)" means one or more, and "plurality" means two or more. "And / or" is used to describe the association relationship of associated objects, indicating that three relationships may exist. For example, "A and / or B" can mean: only A exists, only B exists, and A and B exist at the same time, where A and B can be singular or plural. The character " / " generally indicates that the associated objects before and after are in an "or" relationship. "At least one of the following" or similar expressions refers to any combination of these items, including any combination of single or plural items. For example, at least one of a, b or c can mean: a, b, c, "a and b", "a and c", "b and c", or "a and b and c", where a, b, c can be single or multiple.

[0023] As described in the background technology section, when reconstructing low-dose SPECT scanning tomographic images, if the traditional image reconstruction method is still used directly, the image quality after imaging will have high noise, low resolution, etc. This low-quality image does not meet the clinical diagnostic standards and affects the accuracy of diagnosis.

[0024] In low-dose SPECT scanning, in order to ensure the basic quality of the scanned image, the overall scanning time should not be significantly increased, because the longer the total imaging time, the more disadvantages it will bring, including worse image quality and lower scanning efficiency due to more patient motion, more patient discomfort and less patient throughput. Therefore, how to reduce the dose of radioactive tracers in SPECT imaging while maintaining the quality of reconstructed images to meet clinical diagnostic needs has become a hot research issue.

[0025] In order to solve this problem, the solution disclosed in the present invention aims to use artificial intelligence technology represented by deep learning to develop a new image reconstruction model, and based on a certain overall scanning time and low-dose tracer input, perform a series of processing on the scanned SPECT projection images to reconstruct high-quality SPECT tomograms.

[0026] Figure 1 The flowchart of a low-dose SPECT image reconstruction method according to an embodiment of the present disclosure is shown below. The low-dose SPECT image reconstruction method can be executed by a computing device such as a computer or a mobile terminal, and includes the following steps: Step S101: Acquire multiple projection images corresponding to multiple scanning angles when a target part of a patient is subjected to a SPECT imaging scan.

[0027] Exemplarily, the target part may include but is not limited to any one of the liver, heart and brain. In this embodiment, the liver is taken as an example. The specific content of the SPECT imaging scan can be understood by referring to the prior art, and the multiple projection images of multiple scanning angles obtained by scanning can be stored for backup. When necessary, the stored multiple projection images can be read for subsequent analysis.

[0028] Step S102: convert the multiple projection images into multiple first chord images, and input the multiple first chord images into an image reconstruction model to obtain multiple second chord images; wherein the image reconstruction model is obtained by training a masked autoencoder model MAE based on multiple sample chord images, the encoder and decoder of the masked autoencoder model MAE use a three-dimensional encoder and a three-dimensional decoder, and the multiple sample chord images are obtained by processing multiple sample projection images.

[0029] It should be noted that under the condition of low-dose SPECT imaging scanning, artifacts, loss of details, reduced spatial resolution and other problems may be introduced. In short, the image quality is reduced, and it is necessary to accurately reconstruct a high-quality image. Therefore, the original projection image in this embodiment needs to be pre-processed and converted into a chord diagram (Sinogram). Take a spect scan of the liver as an example, Figure 2 Shown is a chord diagram of the projection of one scan angle obtained from the original acquisition (ie, angular sparse sampling).

[0030] A chord diagram is a form of image representation obtained through a random (Radon) transform, in which each pixel represents the intensity of rays passing through an object such as a target part at a specific angle and position. For a full-angle scan, a chord diagram can be generated for each layer, thus forming a three-dimensional data set composed of multiple layers of chord diagrams. The purpose of this step is to provide a data format that is more suitable for mathematical operations for subsequent processing, because the original projection images come from different scanning angles. From the perspective of image features, the continuity and correlation between them are relatively small. After conversion to a chord diagram, the continuity and correlation between the data are greatly increased, which is more helpful for the model algorithm to combine neighborhood information for denoising or interpolation, thereby improving the quality of subsequent reconstructed images.

[0031] In view of the challenge of image quality degradation caused by low-dose SPECT imaging scanning, it is necessary to design a special algorithm to deal with these defects. Specifically, the method proposed in this embodiment is to reconstruct the converted three-dimensional chord map with high quality, and then reconstruct a high-quality tomogram through the reconstructed high-quality chord map.

[0032] This embodiment uses masked autoencoders (MAE) as the basic model for high-quality image reconstruction. MAE is a deep learning model that was originally used in the field of natural language processing to process sequence data and train the model by predicting masked tokens. It can be regarded as a generative model, but its generation method is different from that of traditional generative models. Generative models in the traditional sense, such as GAN (Generative Adversarial Networks) or VAE (Variational Autoencoders), usually directly generate new sample data. GAN generates new samples by adversarial learning, while VAE generates new samples by sampling the latent space. In contrast, the main task of MAE is to restore or "generate" the masked or missing parts, which is more suitable for the image reconstruction task in the scenario of this embodiment. It should be noted that this embodiment is different from the traditional MAE. The traditional MAE processes text data with a 2d encoder and decoder. In this embodiment, because three-dimensional data is to be processed, it is innovatively replaced with a 3d encoder and decoder. In one embodiment, the plurality of first chord diagrams and the plurality of second chord diagrams are all three-dimensional chord diagrams, which contain information of more dimensions, so that the quality of the subsequently reconstructed image is higher.

[0033] Specifically, if Figure 3The diagram shows the architecture of the 3D MAE proposed in this embodiment. The input data is a low-quality three-dimensional chord diagram, and the trained MAE can reconstruct it into a high-quality three-dimensional chord diagram. The model training process may include: first, the input image such as the sample chord diagram is divided into multiple fixed-size image patches (patches), and then some image patch patches are randomly extracted according to a certain mask rate, and only the visible image patch patches are sent to the encoder (the order can be randomly shuffled), and then the output of the encoder (latent representations) and the mask token are arranged in order as the input of the decoder, and finally the decoder reconstructs the entire image. The sample chord diagram is obtained by processing the sample projection image of the historical patient.

[0034] Step S103: reconstructing a tomographic image of the target part based on the multiple second chord images.

[0035] Exemplarily, after reconstructing a plurality of high-quality second chord images, a tomographic image of the target part can be reconstructed based on the plurality of second chord images, and the reconstruction method can adopt a traditional image reconstruction algorithm.

[0036] This embodiment proposes a low-dose SPECT image reconstruction method based on a 3D mask autoencoder model, which can maintain the scanning time unchanged, combine with a low-dose SPECT scanning scheme, use the trained mask autoencoder model, i.e., the image reconstruction model, to reconstruct and restore high-quality chord images, and then use an image reconstruction algorithm or other appropriate methods to reconstruct tomographic images. This can improve the quality of reconstructed images during SPECT imaging, such as low-dose imaging, and thus improve the accuracy of auxiliary clinical diagnosis, while reducing the radiation impact of SPECT imaging examinations on patients.

[0037] Based on the above embodiments, in one embodiment, the low-dose scanning strategy for the multiple projection images during SPECT imaging scanning includes any one of the following strategies under a preset total scanning time: i) sparse sampling of the original multiple scanning angles to reduce the number of scanning angles at which images need to be acquired during the scanning process, thereby allocating more scanning time to the scanning angle corresponding to each acquired image; ii) sampling the scanning time to maintain a complete sampling image of the multiple scanning angles, thereby reducing the scanning time of each scanning angle.

[0038] That is to say, under the premise of maintaining a certain overall scanning time, there are two exemplary schemes for low-dose SPECT scanning: Solution 1: Sparsely sample the original multiple scanning angles (the number is greater than the multiple scanning angles in step 101), and reduce the number of scanning angles for acquiring images during the scanning process, thereby allowing more scanning time to be allocated to each scanning angle for acquiring images. This means that higher signal strength and better signal-to-noise ratio (SNR) can be obtained at fewer scanning angles, because the counting time for each scanning angle is longer and the statistical fluctuation is smaller. Fewer scanning angles means that the redundancy of the original image data is reduced, which may make the reconstruction process more complicated and require more powerful algorithms to ensure image quality. If the algorithm cannot accurately fill in the missing angle data, artifacts may be introduced, affecting the accuracy and reliability of the image.

[0039] Solution 2: Scanning time sampling, maintaining the complete number of scanning angle samples (i.e., the total number of multiple scanning angles in step 101), reducing the scanning time of each scanning angle, but because the count is reduced, the signal-to-noise ratio is reduced, which will lead to a reduction in the quality of a single projection image. This method requires the algorithm to denoise the image and super-resolution it, that is, to convert a low-resolution image into a high-resolution image through an algorithm, thereby improving the clarity and details of the image. If the algorithm cannot effectively remove these noises, the image quality may further deteriorate, such as loss of details, reduced spatial resolution, and other problems.

[0040] For the two different low-dose SPECT image scanning schemes mentioned above: angle sparse sampling and scanning time sampling. The former will cause discontinuities in the chord diagram, while the latter will produce blur and high noise effects. These characteristics pose challenges to subsequent image processing, and special algorithms need to be designed to deal with these defects. The method proposed in this embodiment is to first convert it into a chord diagram, and then reconstruct the chord diagram data with high quality, and then reconstruct a high-quality tomogram through the reconstructed high-quality chord diagram, thereby improving the quality of the reconstructed image during SPECT imaging such as low-dose imaging, thereby improving the accuracy of auxiliary clinical diagnosis, and at the same time reducing the radiation impact of SPECT imaging examinations on patients.

[0041] In one embodiment, the three-dimensional encoder and the three-dimensional decoder both use 3D Vision Transformer. Different from the traditional MAE, the traditional MAE uses 2d ViT (Vision Transformer) as the encoder and decoder. In the present invention, because three-dimensional data needs to be processed, 2d ViT is replaced by 3d ViT, thereby further improving the quality of the final reconstructed tomogram, thereby further improving the quality of the reconstructed image during SPECT low-dose imaging.

[0042] In one embodiment, in step S103, reconstructing the tomogram of the target part based on the multiple second chord diagrams may specifically include: using a target image reconstruction algorithm to reconstruct the multiple second chord diagrams to obtain the tomogram of the target part; wherein the target image reconstruction algorithm at least includes an image iterative reconstruction algorithm.

[0043] Exemplary image iterative reconstruction algorithms include Maximum Likelihood Expectation Maximization (MLEM), Filtered Back Projection (FBP), etc., but are not limited thereto. By using the image iterative reconstruction algorithm and the reconstructed high-quality three-dimensional chord map, a higher-quality tomogram can be reconstructed, further improving the quality of reconstructed tomograms during SPECT imaging, such as low-dose imaging, thereby improving the accuracy of auxiliary clinical diagnosis and reducing the radiation impact of SPECT imaging on patients.

[0044] In one embodiment, the loss function used by the masked autoencoder model MAE includes a mean square error loss function. In this embodiment, the loss function uses MSE (Mean Squared Error) to constrain the consistency of the reconstructed image and the gold standard image. MSE measures the average of the absolute values ​​of the difference between the predicted value and the true value, and is particularly suitable for image reconstruction tasks because it can give greater penalties for larger errors, which helps the model to adjust the predicted value more finely. MSE tends to produce smoother results and helps to restore details in the image. Therefore, for the image reconstruction task of this embodiment, MSE can help the model better capture subtle features, and the combination with the MAE model, through a large number of masked patches, the model is forced to learn more powerful feature representations, thereby improving generalization capabilities, and ultimately further improving the quality of the model-reconstructed image.

[0045] MSE is shown in the following formula: ; in, yi It is i The true value of the samples, y ^ i The model is i The predicted value of samples, N is the total number of samples.

[0046] For the test process, the input image data such as the chord diagram is also divided into a series of fixed-size patches, all patches are arranged in order, and input into the encoder network to obtain the corresponding embedding vectors, and the embedding vectors are sent to the decoder network to obtain the reconstructed high-quality image. This can improve the quality of reconstructed tomographic images during SPECT imaging such as low-dose imaging, thereby improving the accuracy of auxiliary clinical diagnosis, while reducing the radiation impact of SPECT imaging examinations on patients.

[0047] It should be noted that, although the steps of the method in the present disclosure are described in a specific order in the accompanying drawings, this does not require or imply that the steps must be performed in this specific order, or that all the steps shown must be performed to achieve the desired results. Additionally or alternatively, some steps may be omitted, multiple steps may be combined into one step for execution, and / or one step may be decomposed into multiple steps for execution, etc. In addition, it is also easy to understand that these steps may be, for example, executed synchronously or asynchronously in multiple modules / processes / threads.

[0048] like Figure 4 As shown, the embodiment of the present disclosure provides a low-dose SPECT image reconstruction system, comprising: An image acquisition module 401 is used to acquire multiple projection images corresponding to multiple scanning angles when a target part of a patient is subjected to a SPECT imaging scan; An image reconstruction module 402 is used to convert the multiple projection images into multiple first chord images, and input the multiple first chord images into an image reconstruction model to obtain multiple second chord images; wherein the image reconstruction model is obtained by training a masked autoencoder model MAE based on multiple sample chord images, the encoder and decoder of the masked autoencoder model MAE use a three-dimensional encoder and a three-dimensional decoder, and the multiple sample chord images are obtained by processing multiple sample projection images; The image processing module 403 is used to reconstruct a tomographic image of the target part based on the multiple second chord images.

[0049] In one embodiment, the low-dose scanning strategy of the multiple projection images during SPECT imaging scanning includes any one of the following strategies under a preset total scanning time: Sparsely sample the original multiple scanning angles to reduce the number of scanning angles that need to capture images during the scanning process, thereby allocating more scanning time to the scanning angle corresponding to each captured image; ii) sample the scanning time to maintain a complete sampling image of the multiple scanning angles and reduce the scanning time of each scanning angle.

[0050] In one embodiment, the 3D encoder and the 3D decoder both use 3D Vision Transformer.

[0051] In one embodiment, the plurality of first chord diagrams and the plurality of second chord diagrams are both three-dimensional chord diagrams.

[0052] In one embodiment, the image processing module reconstructs a tomogram of the target part based on the multiple second chord diagrams, including: The target image reconstruction algorithm is used to reconstruct the multiple second chord images to obtain a tomographic image of the target part; wherein the target image reconstruction algorithm at least includes an image iterative reconstruction algorithm.

[0053] In one embodiment, the loss function adopted by the masked autoencoder model MAE includes a mean square error loss function.

[0054] In one embodiment, the target site includes any one of the liver, heart and brain.

[0055] Regarding the system in the above embodiment, the specific manner in which each module performs operations and the corresponding technical effects brought about have been described in detail in the embodiment of the method, and will not be elaborated here.

[0056] It should be noted that although several modules or units of the device for action execution are mentioned in the above detailed description, this division is not mandatory. In fact, according to the embodiment of the present disclosure, the features and functions of two or more modules or units described above can be concretized in one module or unit. Conversely, the features and functions of a module or unit described above can be further divided into multiple modules or units for concretization. The components displayed as modules or units may or may not be physical units, that is, they may be located in one place, or they may be distributed on multiple network units. Some or all of the modules can be selected according to actual needs to achieve the purpose of the disclosed solution. Those of ordinary skill in the art can understand and implement it without paying creative work.

[0057] The present disclosure also provides a computer-readable storage medium having a computer program stored thereon. When the computer program is executed by a processor, the low-dose SPECT image reconstruction method described in any one of the above embodiments is implemented.

[0058] Exemplarily, the readable storage medium may be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, device or device, or any combination thereof. More specific examples (non-exhaustive list) of readable storage media include: an electrical connection with one or more wires, a portable disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination thereof.

[0059] The computer readable storage medium may include a data signal propagated in a baseband or as part of a carrier wave, wherein a readable program code is carried. This propagated data signal may take a variety of forms, including but not limited to an electromagnetic signal, an optical signal, or any suitable combination of the above. The readable storage medium may also be any readable medium other than a readable storage medium, which may send, propagate, or transmit a program for use by an instruction execution system, an apparatus, or a device or used in combination with it. The program code contained on the readable storage medium may be transmitted with any appropriate medium, including but not limited to wireless, wired, optical cable, RF, etc., or any suitable combination of the above.

[0060] The present disclosure also provides an electronic device, including a processor and a memory, wherein the memory is used to store a computer program, wherein the processor is configured to execute the low-dose SPECT image reconstruction method in any of the above embodiments by executing the computer program.

[0061] Refer to the following Figure 5 The electronic device 600 according to this implementation of the present embodiment is described. Figure 5 The electronic device 600 shown is only an example and should not bring any limitation to the functions and scope of use of the present embodiment.

[0062] like Figure 5 As shown, the electronic device 600 is in the form of a general computing device. The components of the electronic device 600 may include but are not limited to: at least one processing unit 610, at least one storage unit 620, a bus 630 connecting different system components (including the storage unit 620 and the processing unit 610), a display unit 640, etc.

[0063] The storage unit stores a program code, which can be executed by the processing unit 610, so that the processing unit 610 performs the steps of various exemplary implementations of this embodiment described in the above method embodiment section of this specification. For example, the processing unit 610 can perform the following steps: Figure 1 The steps of the method shown in .

[0064] The storage unit 620 may include a readable medium in the form of a volatile storage unit, such as a random access memory unit (RAM) 6201 and / or a cache memory unit 6202 , and may further include a read-only memory unit (ROM) 6203 .

[0065] The storage unit 620 may also include a program / utility 6204 having a set (at least one) of program modules 6205, such program modules 6205 including but not limited to: an operating system, one or more application programs, other program modules and program data, each of which or some combination may include the implementation of a network environment.

[0066] Bus 630 may represent one or more of several types of bus structures, including a memory unit bus or memory unit controller, a peripheral bus, an accelerated graphics port, a processing unit, or a local bus using any of a variety of bus architectures.

[0067] The electronic device 600 may also communicate with one or more external devices 700 (e.g., keyboards, pointing devices, Bluetooth devices, etc.), may also communicate with one or more devices that enable a user to interact with the electronic device 600, and / or communicate with any device that enables the electronic device 600 to communicate with one or more other computing devices (e.g., routers, modems, etc.). Such communication may be performed through an input / output (I / O) interface 650. In addition, the electronic device 600 may also communicate with one or more networks (e.g., local area networks (LANs), wide area networks (WANs), and / or public networks, such as the Internet) through a network adapter 660. The network adapter 660 may communicate with other modules of the electronic device 600 through a bus 630. It should be understood that, although not shown in the figure, other hardware and / or software modules may be used in conjunction with the electronic device 600, including but not limited to: microcode, device drivers, redundant processing units, external disk drive arrays, RAID systems, tape drives, and data backup storage systems, etc.

[0068] Through the description of the above implementation methods, it is easy for those skilled in the art to understand that the example implementation methods described here can be implemented by software, or by combining software with necessary hardware. Therefore, the technical solution according to the implementation methods of the present disclosure can be embodied in the form of a software product, which can be stored in a non-volatile storage medium (which can be a CD-ROM, a USB flash drive, a mobile hard disk, etc.) or on a network, and includes a number of instructions to enable a computing device (which can be a personal computer, a server, or a network device, etc.) to execute the method steps of the above-mentioned embodiments according to the implementation methods of the present disclosure.

[0069] It should be noted that, in this article, relational terms such as "first" and "second" are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of further restrictions, the elements defined by the sentence "comprise a ..." do not exclude the existence of other identical elements in the process, method, article or device including the elements.

[0070] The above description is only a specific embodiment of the present disclosure, so that those skilled in the art can understand or implement the present disclosure. Various modifications to these embodiments will be apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present disclosure. Therefore, the present disclosure will not be limited to the embodiments described herein, but will conform to the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A low-dose SPECT image reconstruction method, characterized in that: include: Acquire multiple projection images corresponding to multiple scanning angles when performing SPECT imaging scanning on a target part of a patient; The multiple projection images are converted into multiple first chord images, and the multiple first chord images are input into an image reconstruction model to obtain multiple second chord images; wherein the image reconstruction model is obtained by training a masked autoencoder model MAE based on multiple sample chord images, the encoder and decoder of the masked autoencoder model MAE use a three-dimensional encoder and a three-dimensional decoder, and the multiple sample chord images are obtained by processing multiple sample projection images; A tomographic image of the target site is reconstructed based on the multiple second chord images.

2. The method according to claim 1, characterized in that The low-dose scanning strategy of the multiple projection images during SPECT imaging scanning includes any one of the following strategies under a preset total scanning time: Sparsely sample the original multiple scanning angles to reduce the number of scanning angles that need to capture images during the scanning process, thereby allocating more scanning time to the scanning angle corresponding to each captured image; ii) sample the scanning time to maintain a complete sampling image of the multiple scanning angles and reduce the scanning time of each scanning angle.

3. The method according to claim 1 or 2, characterized in that: The 3D encoder and the 3D decoder both use 3D Vision Transformer.

4. The method according to claim 1 or 2, characterized in that: The plurality of first chord diagrams and the plurality of second chord diagrams are both three-dimensional chord diagrams.

5. The method according to claim 1 or 2, characterized in that: The reconstructing the tomographic image of the target part based on the multiple second chord images comprises: The target image reconstruction algorithm is used to reconstruct the multiple second chord images to obtain a tomographic image of the target part; wherein the target image reconstruction algorithm at least includes an image iterative reconstruction algorithm.

6. The method according to claim 1 or 2, characterized in that: The loss function adopted by the masked autoencoder model MAE includes a mean square error loss function.

7. The method according to claim 1 or 2, characterized in that: The target site includes any one of the liver, heart and brain.

8. A low-dose SPECT image reconstruction system, characterized in that: include: An image acquisition module, used to acquire multiple projection images corresponding to multiple scanning angles when a target part of a patient is subjected to SPECT imaging scanning; An image reconstruction module is used to convert the multiple projection images into multiple first chord images, and input the multiple first chord images into an image reconstruction model to obtain multiple second chord images; wherein the image reconstruction model is obtained by training a masked autoencoder model MAE based on multiple sample chord images, the encoder and decoder of the masked autoencoder model MAE use a three-dimensional encoder and a three-dimensional decoder, and the multiple sample chord images are obtained by processing multiple sample projection images; An image processing module is used to reconstruct a tomographic image of the target part based on the multiple second chord images.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the low-dose SPECT image reconstruction method according to any one of claims 1 to 7 is implemented.

10. An electronic device, characterized in that: include: processor; as well as Memory for storing computer programs; Wherein, the processor is configured to perform the low-dose SPECT image reconstruction method according to any one of claims 1 to 7 by executing the computer program.

Images