A Method for Identifying and Processing Artifact Regions in Nuclear Medicine Imaging Images

Abstract

This invention discloses a method for identifying and processing artifact regions in nuclear medicine imaging images, relating to the field of medical image registration technology. It includes: S1: acquiring PET dynamic image sequences, CT dual-energy spectral images, and respiratory motion monitoring signals to generate a registration quality map; S2: combining the PET dynamic image sequences and CT images to obtain the statistical dependence between them, and locating or reconstructing lesions in artifact regions of the CT images; S3: reconstructing artifacts from the optimized CT images and processing the reconstructed artifact regions using a dual-branch network model. This invention extracts the instantaneous phase of the respiratory signal using Hilbert transform and groups the PET dynamic image sequences according to respiratory phase, achieving accurate matching of in-phase frames. Simultaneously, the B-spline Demons algorithm generates a registration quality map, significantly reducing organ deformation artifacts caused by respiratory motion.

Description

Technical Field

[0001] This invention relates to the field of medical image registration technology, specifically to a method for identifying and processing artifact regions in nuclear medicine imaging images. Background Technology

[0002] Nuclear medicine imaging (such as PET / CT and SPECT) is an important medical imaging technology that observes the metabolic and functional state of the human body through the dynamic distribution of tracers. It has wide applications in fields such as tumor diagnosis, cardiovascular disease assessment, and neuroscience research. However, due to interference from various factors during the imaging process, artifacts often appear in the images, which can significantly affect image quality and diagnostic accuracy. The sources of artifacts are complex and mainly include: device-related artifacts (such as uneven detector sensitivity, system noise, photon statistical fluctuations, etc.), motion artifacts (image blurring or misalignment caused by patient breathing, heartbeat, or voluntary activities), attenuation artifacts (signal attenuation distortion caused by differences in tissue density), scattering artifacts (inelastic scattering of gamma rays or positron annihilation photons), and metal implant artifacts (the wire hardening effect caused by high-density materials such as titanium alloys).

[0003] Chinese invention patent CN110163897A discloses a multimodal image registration method based on synthetic ultrasound images. This method generates a realistic synthetic ultrasound image by constructing a generative adversarial network (GAN) including a generator and a discriminator, based on magnetic resonance imaging (MRI) images and real ultrasound images. The simulated synthetic ultrasound image is then registered with the real ultrasound image to obtain registration parameters. These parameters are applied to the MRI image to complete the final registration and fusion of the MRI and ultrasound images. This method can synthesize ultrasound images from MRI images in real time, meeting the requirements of real-time image-guided surgery. The synthesized realistic ultrasound image is closer to the real ultrasound image, with higher image quality and better preservation of important details. Even when the MRI image contains a tumor, the simulated ultrasound image can still be accurately synthesized. The final registration technique does not require a complex registration algorithm; a simple, traditional registration algorithm can achieve good registration results.

[0004] However, in nuclear medicine multimodal imaging, both high-uptake lesions (such as inflammation or tumors) and metal implant artifacts appear as localized abnormal radioactive uptake in PET / SPECT images, and they share similar grayscale distributions and texture features. Single-modal imaging cannot provide sufficient anatomical information to distinguish between the two, thus relying heavily on multimodal data such as CT / MRI for auxiliary diagnosis. However, due to differences in resolution between modalities and organ deformation caused by respiratory motion, new spatial misalignments and noise interference are introduced during cross-modal registration. This can not only mask the true lesion but also generate false signals during artifact correction, thereby affecting the accurate localization of the lesion and artifact suppression. Summary of the Invention

[0005] The purpose of this invention is to provide a method for identifying and processing artifact regions in nuclear medicine imaging images, so as to solve the problems mentioned in the background art.

[0006] To achieve the above objectives, the present invention provides the following technical solution: a method for identifying and processing artifact regions in nuclear medicine imaging images, comprising:

[0007] S1: Data Processing: Acquire PET dynamic image sequences, CT dual-energy spectral images, and respiratory motion monitoring signals, and generate registration quality maps using the B-spline Demons algorithm, including:

[0008] S1.1: Acquiring image data: Acquire PET dynamic image sequences, CT dual-energy spectral images, and respiratory waveform signals through a PET detector, CT scanner, and respiratory monitoring system, respectively;

[0009] S1.2: Perform phase grouping: Obtain the instantaneous phase of the respiratory waveform signal through Hilbert transform, and group the PET dynamic image sequence according to the instantaneous phase;

[0010] S1.3: CT parameter optimization: Based on the CT dual-energy spectral image, obtain the linear attenuation coefficient. Simultaneously, based on the linear attenuation coefficient and the mass attenuation coefficient data of the titanium alloy, construct the titanium alloy energy spectrum fitting formula, determine the photoelectric effect weighting coefficient and Compton scattering weighting coefficient of titanium, and adjust the CT scanner to obtain the optimized CT image.

[0011] S2: Obtain the generated quality map: Using the B-spline Demons algorithm, the grouped PET dynamic image sequence and the optimized CT image are combined to obtain the statistical dependency between the PET dynamic image sequence and the CT image, and based on the statistical dependency, the lesions in the artifact region of the CT image are located or reconstructed and compensated.

[0012] S3: Image Reconstruction: The optimized CT image is subjected to artifact reconstruction, and the reconstructed artifact region is processed through a dual-branch network model to obtain the reconstructed CT image.

[0013] Furthermore, the PET motion image sequence is grouped, including:

[0014] S1.2.1: Obtaining the analytical signal: Perform a Hilbert transform on the respiratory waveform signal to obtain the analytical signal, specifically:

[0015]

[0016] in: To analyze the signal, This is the original respiratory signal. The imaginary unit, The Hilbert transform of the respiratory signal. For Hilbert transform operators;

[0017] S1.2.2: Phase Extraction: Based on the analyzed signal, the corresponding instantaneous phase is obtained, and the instantaneous phase is normalized to obtain the normalized instantaneous phase, specifically as follows:

[0018]

[0019] in: For instantaneous phase, This is the original respiratory signal. The Hilbert transform of the respiratory signal. Pi Normalized instantaneous phase;

[0020] S1.2.3: Phase grouping: According to the preset separation interval, the normalized instantaneous phase is grouped to obtain a normalized instantaneous phase grouping table;

[0021] S1.2.4: PET image matching: Based on the acquisition time of each PET image frame, determine the displacement magnitude monitored by the respiratory monitoring system corresponding to the acquisition time, and repeat steps S1.2.1-S1.2.3 to obtain the normalized instantaneous phase corresponding to the displacement magnitude. Then, match the PET image corresponding to the normalized instantaneous phase with the normalized instantaneous phase grouping table to obtain the grouped PET images.

[0022] Furthermore, the grouped PET images were validated, including:

[0023] W1: Perform intra-group consistency verification: Obtain the SUVmax value of all frame PET images corresponding to the normalized instantaneous phase group, determine the standard deviation of the SUVmax corresponding to the normalized instantaneous phase group, and determine the consistency quality result corresponding to the normalized instantaneous phase group based on the comparison result of the standard deviation of the SUVmax and the preset standard deviation threshold.

[0024] W2: Verify respiratory cycle stability: Based on the number of consecutive complete respiratory cycles corresponding to the normalized instantaneous phase group, obtain the standard deviation of the respiratory cycle duration, and determine the cycle stability quality result corresponding to the normalized instantaneous phase group based on the comparison result of the standard deviation of the respiratory cycle duration and the preset duration standard deviation threshold.

[0025] Furthermore, based on the comparison between the standard deviation of SUVmax and a preset standard deviation threshold, the consistency quality result corresponding to the normalized instantaneous phase group is determined, specifically as follows:

[0026] When the standard deviation of SUVmax is less than the preset standard deviation threshold, the normalized instantaneous phase group corresponding to the standard deviation of SUVmax passes the consistency quality test; otherwise, the normalized instantaneous phase group corresponding to the standard deviation of SUVmax fails the consistency quality test. At the same time, steps S1.1-S1.2.4 are repeated to reacquire the respiratory waveform signal and perform group processing. Step W1 is repeated until the standard deviation of SUVmax is less than the preset standard deviation threshold.

[0027] The formula for obtaining the standard deviation of SUVmax is as follows:

[0028]

[0029] in: The standard deviation of SUVmax This represents the total number of PET image frames within the current normalized instantaneous phase group. Let be the SUVmax value of the i-th PET image. This is the average SUVmax of all frames of PET images within the current normalized instantaneous phase group. This is the frame index of the PET image within the current normalized instantaneous phase group.

[0030] Furthermore, based on the comparison between the standard deviation of the respiratory cycle duration and a preset standard deviation threshold, the periodic stability quality result corresponding to the normalized instantaneous phase group is determined, specifically as follows:

[0031] When the standard deviation of the respiratory cycle duration is less than the preset duration standard deviation threshold, the normalized instantaneous phase group corresponding to the standard deviation of the respiratory cycle duration passes the cycle stability quality test; otherwise, the normalized instantaneous phase group corresponding to the standard deviation of the respiratory cycle duration fails the cycle stability quality test. At the same time, steps S1.1-S1.2.4 are repeated to adjust the normalized instantaneous phase grouping table, and step W2 is repeated until the standard deviation of the respiratory cycle duration is less than the preset duration standard deviation threshold.

[0032] The formula for obtaining the standard deviation of the respiratory cycle duration is as follows:

[0033]

[0034] in: The standard deviation of the respiratory cycle duration. This represents the total number of respiratory cycles. Let j be the duration of the j-th respiratory cycle. The average respiratory cycle duration. This is an index for the respiratory cycle.

[0035] Furthermore, the photoelectric effect weighting coefficients and Compton scattering weighting coefficients of titanium are corrected, including:

[0036] M1: Periodic verification: According to the set detection cycle, the photoelectric effect weight coefficient and Compton scattering weight coefficient are periodically compared with the preset photoelectric effect weight threshold and the preset Compton scattering weight threshold, respectively, and the final weight coefficient is determined based on the comparison results.

[0037] M2: Coefficient Adjustment: By obtaining the actual photon energy of the X-rays through a high-voltage sensor, the photoelectric effect weighting coefficient is corrected, specifically as follows:

[0038]

[0039] in: The corrected photoelectric effect weighting coefficients for titanium. The photoelectric effect weighting coefficient for titanium is... This refers to the previously measured X-ray photon energy. This refers to the actual measured X-ray photon energy;

[0040] The real-time operating temperature of the detector is obtained using a temperature sensor, and the Compton scattering weighting coefficients are corrected accordingly, specifically as follows:

[0041]

[0042] in: The corrected Compton scattering weighting coefficients for titanium. The Compton scattering weighting coefficient for titanium is... This is the temperature compensation coefficient. This refers to the actual operating temperature measured by the detector. This refers to the detector's previous operating temperature.

[0043] Furthermore, the error ratios between the photoelectric effect weighting coefficient and the preset photoelectric effect weighting threshold, and the error ratios between the Compton scattering weighting coefficient and the preset Compton scattering weighting threshold are obtained. These error ratios are then compared with preset threshold ratios, and the final weighting coefficients are determined based on the comparison results. Specifically:

[0044] When the error ratio is less than the preset ratio threshold, the photoelectric effect weighting coefficient and Compton scattering weighting coefficient corresponding to the error ratio are the final photoelectric effect weighting coefficient and Compton scattering weighting coefficient. Otherwise, the next step M2 is executed to correct the photoelectric effect weighting coefficient and Compton scattering weighting coefficient. At the same time, step M1 is repeated according to the corrected weighting coefficient until the error ratio is less than the preset ratio threshold.

[0045] Furthermore, the lesions in the artifact regions of the CT images are located or reconstructed and compensated, including:

[0046] S2.1: Obtaining Statistical Dependencies Between Images: Based on the deformation field constructed in the B-spline Demons algorithm, the gray values ​​of the PET image and the CT image, the statistical dependencies between the PET image and the CT image are obtained, specifically:

[0047]

[0048] in: To determine the statistical dependence between PET and CT images, The total grayscale value of the CT image. The total grayscale value of the PET image. The grayscale value of the CT image. The grayscale value of the PET image. This represents the joint grayscale probability of CT and PET images at the same spatial location. grayscale values ​​in CT images The probability of occurrence grayscale values ​​of PET images The probability of occurrence;

[0049] S2.2: Determine the final optimized CT image: Compare the statistical dependency with a preset dependency threshold, and determine the final optimized CT image based on the comparison result, specifically:

[0050] When the statistical dependency is greater than the preset dependency threshold, the optimized CT image corresponding to the statistical dependency is the final optimized CT image. Otherwise, repeat steps S1.3-S2.2 to re-acquire the optimized CT image until the statistical dependency is greater than the preset dependency threshold.

[0051] S2.3: J-value comparison: The final optimized CT image and the grouped PET dynamic image sequence are used as inputs to the B-spline Demons algorithm model, and the output is the J-value of each location point between the CT image and the PET image, specifically:

[0052]

[0053] in: For Jacobi determinant, The spatial coordinates in the image are The displacement vector corresponding to the control point Let Jacobian matrix be the displacement vector. The weight coefficient for the k-th control point is... Let B-spline basis functions be used for the k-th control point. To control the number of points, These are the spatial coordinates in the image. Let be the determinant of the matrix;

[0054] The J value is compared with a preset J threshold range, and image fusion / reconstruction / labeling is performed based on the comparison result, specifically as follows:

[0055] When the J value is not less than the upper limit of the preset J threshold range, step S2.4 is executed to fuse the image region corresponding to the J value. When the J value is within the preset J threshold range, step S3 is executed to reconstruct the image region corresponding to the J value. When the J value is less than the lower limit of the preset J threshold range, the image region corresponding to the J value is manually marked, deleted, and re-acquired.

[0056] S2.4: Image Fusion: The gray values ​​of the corresponding regions in the PET image and the corresponding gray values ​​in the CT image are fused to obtain the gray values ​​of the regions in the fused image. Specifically:

[0057]

[0058] in: The spatial coordinates in the fused image are The grayscale value corresponding to the control point. These are the weighting coefficients for the PET images. The spatial coordinates in the PET image are The grayscale value corresponding to the control point. The spatial coordinates in the CT image are The grayscale value corresponding to the control point. These are the spatial coordinates in the image.

[0059] Furthermore, the reconstructed CT images are acquired, including:

[0060] S3.1: Artifact Reconstruction: Based on the point spread function of the titanium alloy, artifact quantization correction is performed, specifically as follows:

[0061]

[0062] in: Let be the point spread function. Let be the standard deviation of the Gaussian distribution. is the base of the natural logarithm. The distance at which the metal artifact is affected. This is the inverse operation of convolution. This is the corrected standardized intake value. SUV value measured from the original PET;

[0063] S3.2: Acquiring Feature Data: The artifact-quantified and corrected CT image is used as the input to the anatomical branch of the two-branch network model, and the time-activity curve corresponding to the PET image is used as the input to the functional branch of the two-branch network model. The output is to obtain the proliferation and inflammation features, specifically:

[0064]

[0065] in: This is the time-activity curve. The initial activity of the first component. The initial activity of the second component. Let be the decay rate constant of the first component. The decay rate constant of the second component is... For time, is the base of the natural logarithm;

[0066] S3.3: Obtaining Attention Weights: Based on the feature vectors of the CT image and the PET image, determine the magnitude of the cross-modal attention weights, specifically as follows:

[0067]

[0068] in: The correlation between PET and CT features at locations m and n. Let m be the query vector for the m-th PET feature. This is the inversion of the key vector of the nth CT feature. The dimension of the feature vector. , For a trainable linear transformation weight matrix, The feature vector corresponding to the PET image. The feature vector corresponding to the CT image;

[0069] S3.4: Image Reconstruction: Based on the proliferation features, inflammation features, and cross-modal attention weights, the optimized CT image is used as the input to the dual-branch network model, and the reconstructed CT image is output.

[0070] Compared with the prior art, the present invention has the following beneficial effects:

[0071] Firstly, this invention extracts the instantaneous phase of the respiratory signal through Hilbert transform and groups the PET dynamic image sequence according to the respiratory phase, which can achieve accurate matching of in-phase frames. At the same time, it performs non-rigid registration through the B-spline Demons algorithm to generate a registration quality map, which can significantly reduce organ deformation artifacts caused by respiratory motion. That is, through dynamic group registration, the spatial consistency between PET images and CT images is improved, avoiding lesions being masked or misaligned due to respiratory motion.

[0072] Secondly, this invention constructs a model of the photoelectric effect and Compton scattering weighting coefficients of titanium alloy using dual-energy spectral CT data, and dynamically adjusts CT scanning parameters. Combined with real-time tube voltage feedback and temperature drift compensation, it can improve attenuation correction accuracy, effectively suppress metal artifacts, avoid confusion between high-density artifacts and high-uptake lesions, and improve the accuracy of tumor or inflammation localization.

[0073] Thirdly, this invention quantifies the statistical dependence between PET and CT images through mutual information and selects the optimal registration result. At the same time, it combines the Jacobian determinant to analyze the local characteristics of the deformation field, thereby adaptively distinguishing artifacts from lesions, avoiding erroneous fusion or reconstruction, and improving the reliability of multimodal images. Attached Figure Description

[0074] Figure 1 This is a flowchart illustrating the artifact region identification and processing method in this invention.

[0075] Figure 2 This is a schematic diagram of the process for obtaining grouped PET images in this invention;

[0076] Figure 3 This is a flowchart illustrating the process of determining the image processing method in this invention;

[0077] Figure 4 This is a diagram illustrating the stability analysis of the respiratory cycle in this invention;

[0078] Figure 5 This is an iterative calibration diagram of the photoelectric effect weighting coefficients in this invention. Detailed Implementation

[0079] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0080] In nuclear medicine multimodal imaging, high-uptake lesions (such as inflammation or tumors) and metal implant artifacts both appear as localized abnormal radioactive uptake in PET / SPECT images, and they share similar grayscale distributions and texture features. Single-modal imaging cannot provide sufficient anatomical information to distinguish between the two, thus relying heavily on multimodal data such as CT / MRI for auxiliary diagnosis. However, due to differences in resolution between modalities and organ deformation caused by respiratory motion, new spatial misalignments and noise interference are introduced during cross-modal registration. This can not only mask the true lesion but also generate false signals during artifact correction, thereby affecting the accurate localization of the lesion and artifact suppression. The technical solution of this application achieves in-phase non-rigid registration by processing the correlation of PET dynamic image sequences, CT dual-energy spectral images and respiratory motion monitoring signals, thus eliminating motion artifacts. At the same time, a titanium alloy attenuation coefficient model is constructed based on CT dual-energy spectral data, the weight coefficients are optimized, and optimized CT images are obtained. The statistical dependence between PET images and CT images is evaluated through mutual information, and the image regions are divided into three categories: fusion, reconstruction and marker deletion, and corresponding processing is performed. This can effectively suppress metal artifacts, avoid confusion between high-density artifacts and high-uptake lesions, and improve the accuracy of tumor or inflammation localization.

[0081] Example 1

[0082] refer to Figures 1-3 This embodiment provides a method for identifying and processing artifact regions in nuclear medicine imaging images. The method specifically includes the following steps:

[0083] Step S1: Data Processing. This involves acquiring PET dynamic image sequences, CT dual-spectrum images, and respiratory motion monitoring signals via timestamp synchronization. Simultaneously, the PET images are grouped based on respiratory phase. Using the B-spline Demons algorithm, frame-by-frame non-rigid registration is performed on the PET and CT images to obtain a registration quality map. Details are as follows:

[0084] Step S1.1: Acquire image data. This involves acquiring dynamic PET image sequences using a PET detector, acquiring dual-energy spectral CT images using a CT scanner, and acquiring the displacement of patient surface markers (such as the sternal notch and xiphoid process) using a respiratory monitoring system (such as a C-RAD Catalyst optical surface monitoring system), and converting these displacements into corresponding respiratory waveform signals.

[0085] Furthermore, during the data acquisition process, nanosecond-level time synchronization is established between the PET detector, CT scanner, and respiratory monitoring system through the IEEE 1588 Precision Time Protocol, thereby enabling the acquisition of PET dynamic image sequences, CT dual-energy spectral images, and respiratory motion monitoring signals through timestamp synchronization.

[0086] Step S1.2: Perform phase grouping. Based on the respiratory waveform signal obtained in step S1.1, perform a Hilbert transform on the signal and determine the corresponding instantaneous phase based on the transform result. Simultaneously, divide the instantaneous phases evenly according to a preset grouping size, obtain the phase interval for each group, and group the PET images according to this phase interval. Specifically:

[0087] Step S1.2.1: Obtain the analytical signal. This involves performing a Hilbert transform on the respiratory waveform signal obtained in step S1.1 to obtain the corresponding analytical signal, specifically:

[0088]

[0089] in: To analyze the signal, This is the original respiratory signal. The imaginary unit, The Hilbert transform of the respiratory signal. This is the Hilbert transform operator.

[0090] Step S1.2.2: Phase Extraction. This involves determining the corresponding instantaneous phase based on the analytical signal obtained in step S1.2.1, specifically as follows:

[0091]

[0092] in: For instantaneous phase, This is the original respiratory signal. The Hilbert transform of the respiratory signal. Pi is the mathematical constant of a circle.

[0093] Furthermore, the acquired instantaneous phase is normalized so that its magnitude is mapped to the range of 0%-100%, specifically:

[0094]

[0095] in: To normalize the instantaneous phase, For instantaneous phase, Pi is the mathematical constant of a circle.

[0096] Step S1.2.3: Phase grouping. That is, according to the preset separation interval (which can be specifically set according to actual needs, so it will not be specifically described in this embodiment), the normalized instantaneous phases obtained in step S1.2.2 are grouped to obtain a normalized instantaneous phase grouping table.

[0097] In the specific implementation process, the preset separation interval is set to 12.5%. Therefore, the normalized instantaneous phase within the 0%-100% range can be divided into 8 groups, as shown in Table 1 below:

[0098] Table 1: Normalized Instantaneous Phase Grouping Table

[0099]

[0100] Step S1.2.4: PET image matching. That is, based on the acquisition time corresponding to each PET image frame, determine the displacement magnitude recorded by the respiratory monitoring system at that acquisition time, and repeat steps S1.2.1-S1.2.3 to obtain the normalized instantaneous phase corresponding to the displacement magnitude. That is, the normalized instantaneous phase is the normalized instantaneous phase corresponding to that PET image frame.

[0101] Furthermore, the normalized instantaneous phase corresponding to each frame of the obtained PET image is matched with the phase interval after grouping in step S1.2.3, thereby matching the PET image with the normalized instantaneous phase grouping table in step S1.2.3 to obtain the grouped PET image.

[0102] During the implementation process, the acquisition time for the 15th PET image was 70 seconds, and its corresponding normalized instantaneous phase was 43.3%. According to the phase interval size divided in Table 1, the 15th PET image matches the 4th group in the normalized instantaneous phase grouping table.

[0103] Step S1.3: CT parameter optimization. This involves decomposing the CT dual-energy spectral image obtained in step S1.1 into a linear combination of the photoelectric effect and Compton scattering, and obtaining the corresponding linear attenuation coefficient for the CT dual-energy spectral image. Specifically:

[0104]

[0105] in: For matter at X-ray photon energy The linear decay coefficient under the condition, These are the weighting coefficients for the photoelectric effect term. For matter at X-ray photon energy The photoelectric effect attenuation coefficient under the following conditions These are the weighting coefficients for the Compton scattering term. For matter at X-ray photon energy The Compton scattering attenuation coefficient under these conditions, This refers to the energy of X-ray photons.

[0106] Furthermore, based on the formula for obtaining the linear attenuation coefficient and the mass attenuation coefficient data of the titanium alloy, a titanium alloy energy spectrum fitting formula is constructed. Simultaneously, based on the titanium alloy energy spectrum fitting formula, the photoelectric effect weighting coefficient and the Compton scattering weighting coefficient of titanium are determined. Based on these determined photoelectric effect weighting coefficients and Compton scattering weighting coefficients, adjustments are made to the X-ray tube energy spectrum filtering, detector response curve, and high-voltage generator stability obtained by the CT scanner. Optimized CT images are then obtained using the adjusted CT scanner.

[0107] In this embodiment, the energy spectrum fitting formula for the titanium alloy is as follows:

[0108]

[0109] in: The linear decay coefficient of titanium at 80 kVp energy. The photoelectric effect weighting coefficient for titanium is... The linear decay coefficient of titanium at 140 kVp energy is given. The Compton scattering weighting coefficient for titanium is... This represents the photoelectric effect attenuation coefficient of titanium at 80 kVp energy. This represents the photoelectric effect attenuation coefficient of titanium at 140 kVp energy. The Compton scattering attenuation coefficient of titanium at 80 kVp energy is given. The Compton scattering attenuation coefficient of titanium at 140 kVp energy is given.

[0110] Step S2: Obtain the generated quality map. This involves processing the grouped PET images obtained in step S1.2.4 using the B-spline Demons algorithm to obtain in-phase frames within the PET images. Simultaneously, the optimized CT images from step S1.3 are processed to suppress metal artifacts. The in-phase frames from the PET images and the CT images with suppressed metal artifacts are then combined to obtain the statistical dependency between them. Based on this statistical dependency, lesions in the artifact regions are located or reconstructed and compensated. Details are as follows:

[0111] Step S2.1: Obtain the statistical dependency between images. Specifically, based on the deformation field constructed in the B-spline Demons algorithm, and using the grayscale values ​​of the PET and CT images within the deformation field, the statistical dependency between the PET and CT images is determined.

[0112]

[0113] in: To determine the statistical dependence between PET and CT images, The total grayscale value of the CT image. The total grayscale value of the PET image. The grayscale value of the CT image. The grayscale value of the PET image. This represents the joint grayscale probability of CT and PET images at the same spatial location. grayscale values ​​in CT images The probability of occurrence grayscale values ​​of PET images The probability of occurrence.

[0114] Step S2.2: Determine the final optimized CT image. This involves comparing the statistical dependency obtained in step S2.1 with a preset dependency threshold, and determining the final optimized CT image based on the comparison result. Specifically:

[0115] When the obtained statistical dependency is greater than the preset dependency threshold, the optimized CT image corresponding to the obtained statistical dependency is the final optimized CT image. Otherwise, repeat steps S1.3-S2.2 to re-obtain the optimized CT image until the obtained statistical dependency is greater than the preset dependency threshold.

[0116] Step S2.3: J-value comparison. The final optimized CT image determined in step S2.2 and the grouped PET images obtained in step S1.2.4 are used as inputs to the B-spline Demons algorithm model. The output is the J-value of each point between the CT and PET images. Specifically:

[0117]

[0118] in: For Jacobi determinant, The spatial coordinates in the image are The displacement vector corresponding to the control point Let Jacobian matrix be the displacement vector. The weight coefficient for the k-th control point is... Let B-spline basis functions be used for the k-th control point. To control the number of points, These are the spatial coordinates in the image. Let be the determinant of the matrix.

[0119] Furthermore, the obtained J value is compared with a preset J threshold range, and image fusion / reconstruction / labeling is performed based on the comparison result. Specifically:

[0120] If the obtained J value is not less than the upper limit of the preset J threshold range, then step S2.4 is executed to fuse the image region corresponding to the J value. If the obtained J value is within the preset J threshold range, then step S3 is executed to reconstruct the image region corresponding to the J value. If the obtained J value is less than the lower limit of the preset J threshold range, then the image region corresponding to the J value is manually marked, deleted, and re-obtained.

[0121] Step S2.4: Image Fusion. Based on the image regions identified in Step S2.3 that can be fused, the grayscale values ​​of these regions in the PET image and their corresponding grayscale values ​​in the CT image are fused to obtain the grayscale values ​​of the fused region in the final image. Specifically:

[0122]

[0123] in: The spatial coordinates in the fused image are The grayscale value corresponding to the control point. These are the weighting coefficients for the PET images. The spatial coordinates in the PET image are The grayscale value corresponding to the control point. The spatial coordinates in the CT image are The grayscale value corresponding to the control point. These are the spatial coordinates in the image.

[0124] Step S3: Image Reconstruction. This involves reconstructing artifacts in the CT image region corresponding to J values ​​within a preset J threshold range, and processing the reconstructed artifact regions using a two-branch network model to obtain the reconstructed CT image. Details are as follows:

[0125] Step S3.1: Artifact Reconstruction. Based on the physical parameters of the titanium alloy, namely its density and photon energy, the corresponding point spread function is obtained. Then, based on this point spread function, the marked artifact regions requiring reconstruction are quantized and corrected using point spread function deconvolution. Specifically:

[0126]

[0127] in: Let be the point spread function. Let be the standard deviation of the Gaussian distribution. is the base of the natural logarithm. The distance at which the metal artifact is affected. This is the inverse operation of convolution. This is the corrected standardized intake value. SUV value measured by the original PET.

[0128] Step S3.2: Obtain feature data. This involves acquiring 64-dimensional anatomical feature vectors and proliferation and inflammation features from the anatomical and functional branches of the constructed dual-branch network model. In other words, the CT image after artifact quantification correction in step S3.1 is used as input to the anatomical branch, outputting the corresponding bone texture gradient and metal-tissue interface sharpness, i.e., the 64-dimensional anatomical feature vector, to determine the location of metal in the CT image.

[0129] Furthermore, based on the identified metal locations in the CT images, the time-activity curve of the corresponding PET image, i.e., the rate of change of the SUVmax value over time in the PET image, is used as the input to the functional branch, outputting the corresponding proliferative and inflammatory features, specifically:

[0130]

[0131] in: This is the time-activity curve. The initial activity of the first component. The initial activity of the second component. Let be the decay rate constant of the first component. The decay rate constant of the second component is... For time, is the base of the natural logarithm.

[0132] Step S3.3: Obtain attention weights. This involves determining the cross-modal attention weights based on the feature vectors corresponding to the CT images (e.g., bone density, metal artifact contours) and the feature vectors corresponding to the PET images (e.g., metabolic activity, proliferation features, and inflammation features obtained in step S3.2). Specifically:

[0133]

[0134] in: The correlation between PET and CT features at locations m and n. Let m be the query vector for the m-th PET feature. This is the inversion of the key vector of the nth CT feature. The dimension of the feature vector. , For a trainable linear transformation weight matrix, The feature vector corresponding to the PET image. This is the feature vector corresponding to the CT image.

[0135] Step S3.4: Image Reconstruction. Based on the proliferation and inflammation features determined in Step S3.2 and the cross-modal attention weights determined in Step S3.3, the CT image region that needs to be reconstructed, as determined in Step S2.3, is used as the input to the dual-branch network model in this embodiment, and the corresponding reconstructed CT image is obtained as the output.

[0136] Example 2

[0137] This embodiment provides a method for identifying and processing artifact regions in nuclear medicine imaging images. The specific implementation method is the same as in Embodiment 1, except that the grouping effect after matching is verified during the PET image matching process in step S1.2.4. The invention will be illustrated below with specific examples of this embodiment.

[0138] In this embodiment, the grouping effect of PET image matching in step S1.2.4 is verified through intragroup consistency verification and respiratory cycle stability verification, as follows:

[0139] Step W1: Perform intra-group consistency verification. This involves obtaining the SUVmax value for each PET frame within the same normalized instantaneous phase group, and then calculating the standard deviation of all PET frames within the same group based on this SUVmax value.

[0140]

[0141] in: The standard deviation of SUVmax This represents the total number of PET image frames within the current normalized instantaneous phase group. Let be the SUVmax value of the i-th PET image. This is the average SUVmax of all frames of PET images within the current normalized instantaneous phase group. This is the frame index of the PET image within the current normalized instantaneous phase group.

[0142] Furthermore, the standard deviation of the SUVmax corresponding to the current normalized instantaneous phase group is compared with a preset standard deviation threshold, and the consistency quality result corresponding to the current normalized instantaneous phase group is determined based on the comparison result. Specifically:

[0143] If the standard deviation of the SUVmax corresponding to the current normalized instantaneous phase group is less than the preset standard deviation threshold, then the current normalized instantaneous phase group passes the consistency quality test. Otherwise, the current normalized instantaneous phase group fails the consistency quality test. In this case, steps S1.1-S1.2.4 need to be repeated to reacquire the respiratory waveform signal, perform grouping processing, and repeat step W1 until the standard deviation of the SUVmax corresponding to the current normalized instantaneous phase group is less than the preset standard deviation threshold, and the current normalized instantaneous phase group passes the consistency quality test.

[0144] During the implementation process, a consistency quality inspection data table as shown in Table 2 is set up, specifically as follows:

[0145] Table 2: Consistency Quality Inspection Data Table

[0146]

[0147] According to the data in Table 2 above, the average SUVmax value of the PET images within this group of frames is 2.45, and the standard deviation of the corresponding SUVmax is 0.13. Meanwhile, the preset standard deviation threshold in this embodiment is set to 0.5, meaning that this normalized instantaneous phase group passes the consistency quality test.

[0148] Step W2: Verify respiratory cycle stability. This involves obtaining a consecutive number of complete respiratory cycles from each normalized instantaneous phase group, acquiring the time difference between adjacent inspiratory initiation points, and determining the standard deviation of the respiratory cycle duration using this time difference. Specifically:

[0149]

[0150] in: The standard deviation of the respiratory cycle duration. This represents the total number of respiratory cycles. Let j be the duration of the j-th respiratory cycle. The average respiratory cycle duration. This is an index for the respiratory cycle.

[0151] It is worth noting that the number of complete respiratory cycles can be set according to actual needs, so it will not be described in this embodiment.

[0152] Furthermore, the standard deviation of the obtained respiratory cycle duration is compared with a preset duration standard deviation threshold, and based on the comparison result, the cycle stability quality result corresponding to the current normalized instantaneous phase group is determined. Specifically:

[0153] If the standard deviation of the obtained respiratory cycle duration is less than the preset duration standard deviation threshold, then the current normalized instantaneous phase group passes the cycle stability quality test. Otherwise, the current normalized instantaneous phase group fails the cycle stability quality test. In this case, steps S1.2.1-S1.2.4 need to be repeated to adjust the current normalized instantaneous phase grouping table, and step W2 needs to be repeated until the standard deviation of the obtained respiratory cycle duration is less than the preset duration standard deviation threshold, and the current normalized instantaneous phase group passes the cycle stability quality test.

[0154] During the specific implementation process, the periodic stability quality test data table shown in Table 3 below is set up as follows:

[0155] Table 3: Periodic Stability Quality Test Data Table

[0156]

[0157] According to the data in Table 3 above, the average respiratory cycle duration within this normalized instantaneous phase group is 5 seconds, and the corresponding standard deviation of the respiratory cycle duration is approximately 0.15 seconds. Meanwhile, the preset duration standard deviation threshold in this embodiment is set to 0.3 seconds, meaning that this normalized instantaneous phase group passes the cycle stability quality test.

[0158] refer to Figure 4 , Figure 4 This is a respiratory cycle stability analysis diagram in this embodiment, by Figure 4 It can be seen that when the standard deviation of the respiratory cycle duration is 0.13 seconds, it is significantly lower than the preset threshold of 0.3 seconds. At this time, the respiratory signal has high stability, with a fluctuation range of 4.8-5.2 seconds. This ensures that the dynamic PET image registration is not affected by the fluctuation of the respiratory cycle, providing a stable basis for motion artifact correction.

[0159] Example 3

[0160] This embodiment provides a method for identifying and processing artifact regions in nuclear medicine imaging images. The specific implementation method is the same as in Embodiment 1, except that in step S1.3, the determined photoelectric effect weighting coefficient and Compton scattering weighting coefficient are corrected based on real-time tube voltage feedback and temperature drift compensation. The invention will be illustrated below with specific examples of this embodiment.

[0161] In this embodiment, the photoelectric effect weighting coefficient is corrected through real-time tube voltage feedback, and the Compton scattering weighting coefficient is corrected through temperature drift compensation. Specifically:

[0162] Step M1: Periodic Verification. This involves verifying the photoelectric effect weighting coefficients and Compton scattering weighting coefficients determined in step S1.3 according to the set detection cycle. Specifically, the determined photoelectric effect weighting coefficients are compared with preset photoelectric effect weighting thresholds, and the determined Compton scattering weighting coefficients are compared with preset Compton scattering weighting thresholds. Based on the comparison results, the final weighting coefficients are determined.

[0163] Furthermore, the error ratios between the photoelectric effect weighting coefficient and the preset photoelectric effect weighting threshold, and the error ratios between the Compton scattering weighting coefficient and the preset Compton scattering weighting threshold are obtained. These error ratios are then compared with preset threshold values, and based on the comparison results, the next step M2 is executed. Specifically:

[0164] When the obtained error ratio is less than the preset ratio threshold, the determined photoelectric effect weighting coefficient and Compton scattering weighting coefficient are the final photoelectric effect weighting coefficient and Compton scattering weighting coefficient. Otherwise, the next step M2 is executed to correct the photoelectric effect weighting coefficient and Compton scattering weighting coefficient, and step M1 is repeated according to the corrected weighting coefficient until the obtained error ratio is less than the preset ratio threshold.

[0165] In this embodiment, the formula for obtaining the error ratio is as follows:

[0166]

[0167] in: This is the error ratio. The actual weighting coefficients obtained. The preset weight threshold.

[0168] Step M2: Coefficient Adjustment. This involves obtaining the photon energy of the actual X-rays using a high-voltage sensor, and then correcting the photoelectric effect weighting coefficients based on the obtained photon energy. The corrected photoelectric effect weighting coefficients are obtained as follows:

[0169]

[0170] in: The corrected photoelectric effect weighting coefficients for titanium. The photoelectric effect weighting coefficient for titanium is... This refers to the previously measured X-ray photon energy. This represents the actual measured X-ray photon energy.

[0171] Furthermore, the real-time operating temperature of the detector is obtained through a temperature sensor, and the Compton scattering weighting coefficients are corrected based on the real-time operating temperature of the detector to obtain the corrected Compton scattering weighting coefficients, specifically:

[0172]

[0173] in: The corrected Compton scattering weighting coefficients for titanium. The Compton scattering weighting coefficient for titanium is... This is the temperature compensation coefficient. This refers to the actual operating temperature measured by the detector. This refers to the detector's previous operating temperature.

[0174] refer to Figure 5 , Figure 5 This is the iterative calibration diagram of the photoelectric effect weighting coefficients in this embodiment, derived from... Figure 5 It can be seen that the photoelectric effect weighting coefficient converged from the initial value of 1.1 to 1.01 after iterative calibration, reaching the target threshold. In other words, the dual-energy spectral CT parameters were precisely optimized through a closed-loop feedback mechanism (such as tube voltage / temperature compensation), ensuring the accuracy of metal artifact correction and avoiding over-correction or under-correction.

[0175] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended embodiments and their equivalents.

Claims

1. A method for identifying and processing artifact regions in nuclear medicine imaging images, characterized in that, Including: S1: Data Processing: Acquire PET dynamic image sequences, CT dual-energy spectral images, and respiratory motion monitoring signals, and generate registration quality maps using the B-spline Demons algorithm, including: S1.1: Acquiring image data: Acquire PET dynamic image sequences, CT dual-energy spectral images, and respiratory waveform signals through a PET detector, CT scanner, and respiratory monitoring system, respectively; S1.2: Perform phase grouping: Obtain the instantaneous phase of the respiratory waveform signal through Hilbert transform, and group the PET dynamic image sequence according to the instantaneous phase; S1.3: CT parameter optimization: Based on the CT dual-energy spectral image, obtain the linear attenuation coefficient. Simultaneously, based on the linear attenuation coefficient and the mass attenuation coefficient data of the titanium alloy, construct the titanium alloy energy spectrum fitting formula, determine the photoelectric effect weighting coefficient and Compton scattering weighting coefficient of titanium, and adjust the CT scanner to obtain the optimized CT image. S2: Obtaining the generated quality map: Using the B-spline Demons algorithm, the grouped PET dynamic image sequence and the optimized CT image are combined to obtain the statistical dependency between the PET dynamic image sequence and the CT image. Based on the statistical dependency, lesions in the artifact regions of the CT image are located or reconstructed and compensated, including: S2.1: Obtain statistical dependencies between images; S2.2: Determine the final optimized CT image: Compare the statistical dependency with a preset dependency threshold, and determine the final optimized CT image based on the comparison result, specifically: When the statistical dependency is greater than the preset dependency threshold, the optimized CT image corresponding to the statistical dependency is the final optimized CT image. Otherwise, repeat steps S1.3-S2.2 to re-acquire the optimized CT image until the statistical dependency is greater than the preset dependency threshold. S2.3: J-value comparison: The final optimized CT image and the grouped PET dynamic image sequence are used as inputs to the B-spline Demons algorithm model. The output is the J-value of each location point between the CT image and the PET image. Simultaneously, the J-value is compared with a preset J-threshold range. Based on the comparison result, image fusion / reconstruction / labeling is performed, specifically as follows: When the J value is not less than the upper limit of the preset J threshold range, step S2.4 is executed to fuse the image region corresponding to the J value. When the J value is within the preset J threshold range, step S3 is executed to reconstruct the image region corresponding to the J value. When the J value is less than the lower limit of the preset J threshold range, the image region corresponding to the J value is manually marked, deleted, and re-acquired. S2.4: Image fusion: The gray values ​​of the fusion image region in the PET image are fused with the corresponding gray values ​​in the CT image to obtain the gray values ​​of the region in the fused image. S3: Image Reconstruction: The optimized CT image is subjected to artifact reconstruction, and the reconstructed artifact region is processed through a dual-branch network model to obtain the reconstructed CT image.

2. The method for identifying and processing artifact regions in nuclear medicine imaging images according to claim 1, characterized in that, Grouping the PET motion image sequence includes: S1.2.1: Obtaining the analytical signal: Perform a Hilbert transform on the respiratory waveform signal to obtain the analytical signal, specifically: ； in: To analyze the signal, This is the original respiratory signal. The imaginary unit, The Hilbert transform of the respiratory signal. For Hilbert transform operators; S1.2.2: Phase Extraction: Based on the analyzed signal, the corresponding instantaneous phase is obtained, and the instantaneous phase is normalized to obtain the normalized instantaneous phase, specifically as follows: ； in: For instantaneous phase, This is the original respiratory signal. The Hilbert transform of the respiratory signal. Pi Normalized instantaneous phase; S1.2.3: Phase grouping: According to the preset separation interval, the normalized instantaneous phase is grouped to obtain a normalized instantaneous phase grouping table; S1.2.4: PET image matching: Based on the acquisition time of each PET image frame, determine the displacement magnitude monitored by the respiratory monitoring system corresponding to the acquisition time, and repeat steps S1.2.1-S1.2.3 to obtain the normalized instantaneous phase corresponding to the displacement magnitude. Then, match the PET image corresponding to the normalized instantaneous phase with the normalized instantaneous phase grouping table to obtain the grouped PET images.

3. The method for identifying and processing artifact regions in nuclear medicine imaging images according to claim 2, characterized in that, The grouped PET images were validated, including: W1: Perform intra-group consistency verification: Obtain the SUVmax value of all frame PET images corresponding to the normalized instantaneous phase group, determine the standard deviation of the SUVmax corresponding to the normalized instantaneous phase group, and determine the consistency quality result corresponding to the normalized instantaneous phase group based on the comparison result of the standard deviation of the SUVmax and the preset standard deviation threshold. W2: Verify respiratory cycle stability: Based on the number of consecutive complete respiratory cycles corresponding to the normalized instantaneous phase group, obtain the standard deviation of the respiratory cycle duration, and determine the cycle stability quality result corresponding to the normalized instantaneous phase group based on the comparison result of the standard deviation of the respiratory cycle duration and the preset duration standard deviation threshold.

4. The method for identifying and processing artifact regions in nuclear medicine imaging images according to claim 3, characterized in that, Based on the comparison between the standard deviation of SUVmax and the preset standard deviation threshold, the consistency quality result corresponding to the normalized instantaneous phase group is determined, specifically as follows: When the standard deviation of SUVmax is less than the preset standard deviation threshold, the normalized instantaneous phase group corresponding to the standard deviation of SUVmax passes the consistency quality test; otherwise, the normalized instantaneous phase group corresponding to the standard deviation of SUVmax fails the consistency quality test. At the same time, steps S1.1-S1.2.4 are repeated to reacquire the respiratory waveform signal and perform group processing. Step W1 is repeated until the standard deviation of SUVmax is less than the preset standard deviation threshold. The formula for obtaining the standard deviation of SUVmax is as follows: ； in: The standard deviation of SUVmax This represents the total number of PET image frames within the current normalized instantaneous phase group. Let be the SUVmax value of the i-th PET image. This is the average SUVmax of all frames of PET images within the current normalized instantaneous phase group. This is the frame index of the PET image within the current normalized instantaneous phase group.

5. The method for identifying and processing artifact regions in nuclear medicine imaging images according to claim 3, characterized in that, Based on the comparison between the standard deviation of the respiratory cycle duration and a preset standard deviation threshold, the cycle stability quality result corresponding to the normalized instantaneous phase group is determined, specifically as follows: When the standard deviation of the respiratory cycle duration is less than the preset duration standard deviation threshold, the normalized instantaneous phase group corresponding to the standard deviation of the respiratory cycle duration passes the cycle stability quality test; otherwise, the normalized instantaneous phase group corresponding to the standard deviation of the respiratory cycle duration fails the cycle stability quality test. At the same time, steps S1.1-S1.2.4 are repeated to adjust the normalized instantaneous phase grouping table, and step W2 is repeated until the standard deviation of the respiratory cycle duration is less than the preset duration standard deviation threshold. The formula for obtaining the standard deviation of the respiratory cycle duration is as follows: ； in: The standard deviation of the respiratory cycle duration. This represents the total number of respiratory cycles. Let j be the duration of the j-th respiratory cycle. The average respiratory cycle duration. This is an index for the respiratory cycle.

6. The method for identifying and processing artifact regions in nuclear medicine imaging images according to claim 1, characterized in that, The photoelectric effect weighting coefficients and Compton scattering weighting coefficients for titanium are corrected, including: M1: Periodic verification: According to the set detection cycle, the photoelectric effect weight coefficient and the Compton scattering weight coefficient are periodically compared with the preset photoelectric effect weight threshold and the preset Compton scattering weight threshold, respectively, and the final weight coefficient is determined based on the comparison results. M2: Coefficient Adjustment: By obtaining the actual photon energy of the X-rays through a high-voltage sensor, the photoelectric effect weighting coefficient is corrected, specifically as follows: ； in: The corrected photoelectric effect weighting coefficients for titanium. This is the weighting coefficient for the photoelectric effect of titanium. This refers to the previously measured X-ray photon energy. This refers to the actual measured X-ray photon energy; The real-time operating temperature of the detector is obtained using a temperature sensor, and the Compton scattering weighting coefficients are corrected accordingly, specifically as follows: ； in: The corrected Compton scattering weighting coefficients for titanium. The Compton scattering weighting coefficient for titanium is... This is the temperature compensation coefficient. This refers to the actual operating temperature measured by the detector. This refers to the detector's previous operating temperature.

7. The method for identifying and processing artifact regions in nuclear medicine imaging images according to claim 6, characterized in that, The error ratios between the photoelectric effect weighting coefficient and the preset photoelectric effect weighting threshold, and the error ratios between the Compton scattering weighting coefficient and the preset Compton scattering weighting threshold are obtained. These error ratios are then compared with preset threshold ratios, and the final weighting coefficients are determined based on the comparison results. Specifically: When the error ratio is less than the preset ratio threshold, the photoelectric effect weighting coefficient and Compton scattering weighting coefficient corresponding to the error ratio are the final photoelectric effect weighting coefficient and Compton scattering weighting coefficient. Otherwise, the next step M2 is executed to correct the photoelectric effect weighting coefficient and Compton scattering weighting coefficient. At the same time, step M1 is repeated according to the corrected weighting coefficient until the error ratio is less than the preset ratio threshold.

8. The method for identifying and processing artifact regions in nuclear medicine imaging images according to claim 1, characterized in that, To obtain the statistical dependencies between images, specifically: based on the deformation field constructed in the B-spline Demons algorithm, the gray values ​​of the PET image and the CT image, the statistical dependencies between the PET image and the CT image are obtained, specifically: ； in: To determine the statistical dependence between PET and CT images, The total grayscale value of the CT image. The total grayscale value of the PET image. The grayscale value of the CT image. The grayscale value of the PET image. This represents the joint grayscale probability of CT and PET images at the same spatial location. grayscale values ​​in CT images The probability of occurrence grayscale values ​​of PET images The probability of occurrence; The J value at each location point between the CT image and the PET image is specifically as follows: ； in: For Jacobi determinant, The spatial coordinates in the image are The displacement vector corresponding to the control point Let Jacobian matrix be the displacement vector. The weight coefficient for the k-th control point is... Let B-spline basis functions be used for the k-th control point. To control the number of points, These are the spatial coordinates in the image. Let be the determinant of the matrix; The grayscale values ​​of the regions in the fused image are specifically as follows: ； in: The spatial coordinates in the fused image are The grayscale value corresponding to the control point. These are the weighting coefficients for the PET images. The spatial coordinates in the PET image are The grayscale value corresponding to the control point. The spatial coordinates in the CT image are The grayscale value corresponding to the control point. These are the spatial coordinates in the image.

9. The method for identifying and processing artifact regions in nuclear medicine imaging images according to claim 1, characterized in that, Obtain the reconstructed CT images, including: S3.1: Artifact Reconstruction: Based on the point spread function of the titanium alloy, artifact quantization correction is performed, specifically as follows: ； in: Let be the point spread function. Let be the standard deviation of the Gaussian distribution. is the base of the natural logarithm. The distance at which the metal artifact is affected. This is the inverse operation of convolution. This is the corrected standardized intake value. SUV value measured from the original PET; S3.2: Acquiring Feature Data: The artifact-quantified and corrected CT image is used as the input to the anatomical branch of the two-branch network model, and the time-activity curve corresponding to the PET image is used as the input to the functional branch of the two-branch network model. The output obtains the proliferation and inflammation features, specifically: ； in: This is the time-activity curve. The initial activity of the first component. The initial activity of the second component. Let be the decay rate constant of the first component. The decay rate constant of the second component is... For time, is the base of the natural logarithm; S3.3: Obtaining Attention Weights: Based on the feature vectors of the CT image and the PET image, determine the magnitude of the cross-modal attention weights, specifically as follows: ； in: The correlation between PET and CT features at locations m and n. Let m be the query vector for the m-th PET feature. This is the inversion of the key vector of the nth CT feature. The dimension of the feature vector. , For a trainable linear transformation weight matrix, The feature vector corresponding to the PET image. The feature vector corresponding to the CT image; S3.4: Image Reconstruction: Based on the proliferation features, inflammation features, and cross-modal attention weights, the optimized CT image is used as the input to the dual-branch network model, and the reconstructed CT image is output.

Images