Micro-dose CT (Computed Tomography) system and data acquisition method

Abstract

The invention discloses a micro-dose CT system and a data acquisition method. The data acquisition method comprises the following steps: predicting a photon starvation state of a target in tomography; and based on the prediction result, whether an exposure enabling signal and an acquisition enabling signal are generated when the acquisition triggering signal is generated at the current tomography angle is judged, and synchronous exposure and acquisition are triggered circularly according to the exposure enabling signal and the acquisition enabling signal. According to the data acquisition method of the micro-dose CT system, photon hunger prediction is carried out, the sampling angle is dynamically adjusted according to the prediction result, data acquisition in a photon hunger area is avoided, synchronous exposure and acquisition are realized in combination with a pulse type exposure technology, the image quality is improved on the premise of keeping low radiation dose, and the data acquisition efficiency is improved. And the diagnosis accuracy is ensured.

Description

Technical Field

[0001] The present invention relates to the technical field of medical equipment, and in particular to a micro-dose CT system and a data acquisition method. Background Art

[0002] Computed tomography (CT) is an important medical diagnostic tool, but its radiation dose has always been a concern. Traditional CT scanning methods usually use higher radiation doses to ensure image quality, but this increases the radiation risk to patients. Microdose CT imaging technology aims to reduce the radiation dose to the lowest possible level to minimize the radiation risk to patients. However, in microdose CT imaging, the photon starvation phenomenon is more serious, resulting in increased image noise, more artifacts and reduced contrast, which seriously affects the accuracy of diagnosis. Summary of the invention

[0003] The purpose of the present invention is to provide a microdose CT system and a data acquisition method to solve the problem in the prior art that when the radiation dose is reduced in microdose CT imaging, the photon starvation phenomenon is aggravated, thereby causing increased image noise, increased artifacts and reduced contrast, which seriously affects the accuracy of diagnosis.

[0004] In order to solve the above technical problems, the technical solution adopted by the present invention is: A data acquisition method for a microdose CT system comprises the following steps: Predicting the photon starvation state of a target in tomography; Based on the prediction result, it is determined whether an exposure enable signal and an acquisition enable signal are generated when an acquisition trigger signal is generated at the current tomography angle, and synchronous exposure and acquisition are triggered cyclically according to the exposure enable signal and the acquisition enable signal.

[0005] Further, the photon starvation state is predicted based on a photon starvation prediction algorithm, and the photon starvation prediction algorithm includes: Through the analysis and prediction of the attenuation characteristics of the locator image, the target is scanned with a locator image, and the attenuation characteristics of the target scanning area obtained by the locator image scanning are combined to predict the photon-starved area of ​​the target in the tomographic scanning; Or / and, through real-time data analysis and prediction, the total number of pixels m that are photon-starved in a single acquisition of the target in the exposure scan is counted, and the change trend of the total number of pixels m that are photon-starved in multiple consecutive acquisitions of the target in the tomography scan is predicted; Or / and, the AI ​​model is used to predict the probability of a target being photon starved during scanning.

[0006] Furthermore, the photon starvation state is defined as K, the photon starvation index predicted by the positioning image attenuation characteristic analysis is k1, the photon starvation index predicted by the real-time data analysis is k2, and the photon starvation probability predicted by the AI ​​model is k3; When the photon starvation state K is predicted independently by analyzing the attenuation characteristics of the localization image, K=k1, and the threshold of K is 0; When the photon starvation state K is predicted independently through real-time data analysis, K = k2, and the threshold of K is 0; When the photon starvation state K is predicted independently by the AI ​​model, K=k3, and the threshold of K is k0; When the photon starvation state K is predicted by the analysis of the attenuation characteristics of the positioning image and the real-time data analysis, K=k1+k2, and the threshold of K is 0; When the photon starvation state K is predicted by the positioning image attenuation characteristic analysis and the AI ​​model, K=k1+k3, and the threshold of K is k0; When the photon starvation state K is predicted by real-time data analysis and AI model, K=k2+k3, and the threshold of K is k0; When the photon starvation state K is predicted through positioning image attenuation characteristic analysis, real-time data analysis and AI model, K=k1+k2+k3, and the threshold of K is k0; Among them, k0 is the photon starvation judgment threshold in the AI ​​model prediction.

[0007] Furthermore, based on the angle selection algorithm, it is determined whether to generate an exposure enable signal and an acquisition enable signal when an acquisition trigger signal is generated at the current tomography angle, and synchronous exposure and acquisition are triggered cyclically according to the exposure enable signal and the acquisition enable signal.

[0008] Further, based on the angle selection algorithm, it is determined whether to generate an exposure enable signal and an acquisition enable signal when an acquisition trigger signal is generated at the current tomography angle, and synchronous exposure and acquisition are triggered cyclically according to the exposure enable signal and the acquisition enable signal, specifically including: Step 1: Define the acquisition angle as A and the sparse acquisition angle period as B, where B=2π / D, and D is the total number of sparse acquisition angles; Step 2: Start the rack rotation; Step 3: Detect the gantry rotation angle in real time and generate scanning control signals, including scanning start and end signals. According to the acquisition angle cycle B, insert the acquisition trigger signal between the scanning start and end signals. The angle interval between adjacent acquisition trigger signals is B. When the first acquisition trigger signal starts, trigger a synchronous exposure and acquisition. Step 4: When each acquisition is completed, predict the K value for the next acquisition; If the K value exceeds the threshold, the photon starvation state is activated, and the sum of the unsampled angles W1 is calculated. When W1 ≥ B, the calculated K value is updated after each A angle. When W1 = xB, a synchronous exposure and acquisition is triggered, where x is a positive integer. If the K value is lower than the threshold, the photon starvation state is turned off and a synchronous exposure and acquisition is triggered. After the total angle of the acquired angle W2=yB angle, a synchronous exposure and acquisition is triggered, where y is a positive integer; Repeat the above steps.

[0009] Furthermore, the synchronous exposure and acquisition are triggered according to the exposure enable signal and the acquisition enable signal, specifically: The exposure enable signal and the acquisition enable signal are generated simultaneously and have the same duration; Alternatively, the exposure enable signal and the acquisition enable signal are generated successively, and the exposure enable signal at least satisfies the requirement of generating a rising edge of the exposure enable start slightly earlier than the start time of the acquisition enable signal, and generating a falling edge of the exposure enable end slightly later than the end time of the acquisition enable signal.

[0010] Furthermore, the target is scanned with a scouting image, and the attenuation characteristics of the target scanning area obtained by the scouting image are combined to predict the photon-starved area of ​​the target in the tomographic scan, specifically including: Collecting a forward positioning image or / and a lateral positioning image orthogonal to the target, wherein the forward positioning image is defined as a 0-degree positioning image, and the lateral positioning image is defined as a 90-degree positioning image; Obtain the attenuation characteristics of the target scanning area at 0 degrees and 90 degrees, and calculate the orthogonal attenuation a0 and a90 of each Z coordinate corresponding to the fault plane; Establish the attenuation distribution curve of each Z coordinate corresponding to the fault plane, and obtain the estimated attenuation a (θ) of the target scanning area at each scanning angle based on the fault plane shape of the target scanning area; The attenuation threshold of photon starvation is set to bs, and the real-time attenuation of the target scanning area at each scanning angle during scanning is predicted to be b(θ) based on the estimated attenuation a(θ). When b(θ) < bs, the photon starvation index k1 is set to 0, otherwise k1 is set to 1.

[0011] Furthermore, the calculation formulas for attenuation a0 and a90 are as follows: ; ; Among them, M and N are the pixel matrix sizes of the detector, For each pixel size, is the offset of each pixel corresponding to the 0-degree positioning image, The offset of each pixel corresponding to the 90-degree positioning image; The calculation formula of real-time attenuation b(θ) is as follows: ; Wherein, kVt is the tube voltage for scout image scanning, mAt is the tube current for scout image scanning, kV(θ) is the tube voltage for tomography scanning, mA(θ) is the tube current for tomography scanning, and c is the tube voltage proportionality coefficient.

[0012] Furthermore, the total number of pixels m that are photon-starved in a single acquisition of the target in the exposure scan is counted, and the change trend of the total number of pixels m that are photon-starved in multiple consecutive acquisitions of the target in the tomography scan is predicted, specifically including: Collect unexposed data at V angles by rotating within 360 degrees, V ≥ 100, calculate the expected value μ and standard deviation σ of each pixel under the background noise level, and set the photon starvation judgment threshold THR according to the expected value μ and standard deviation σ; Pulse scanning obtains real-time scanning data and performs pixel-level difference calculation, compares all pixels with the photon starvation judgment threshold THR, and obtains P comparison results, P≤M×N, and sets the total number of pixels below the photon starvation judgment threshold THR to m, m≤P, where P is the total number of detector pixels involved in photon starvation prediction, and M and N are the pixel matrix sizes of the detector; Store the three most recent m values: mn, mn-1, mn-2, calculate the acceleration a of the rate of change of the m value, a=(mn-mn-1)-(mn-1-mn-2), predict the next periodic change ∆m=(mn-mn-1)+a, set the total number of pixels where photon starvation occurs, a baseline value mo, and the periodic change baseline value ∆mo, when m>mo and ∆m>∆mo, k2=1, otherwise k2=0.

[0013] Furthermore, the AI ​​model is used to predict the probability of a target being photon starved during scanning. Specifically: The AI ​​model network uses a multimodal convolutional neural network; Input data: Target positioning image and scanning parameter information; Or, real-time scanned data and scan parameter information; Or, target positioning image, scanning parameter information and real-time scanning data; Output data: photon starvation probability k3, the range of k3 is 0~1.

[0014] Furthermore, the AI ​​model network structure includes: An input layer, for receiving multimodal data, wherein the multimodal data includes images, attenuation curves and scanning parameters; The CNN branch is used to process anatomical structure images and extract spatial features; The fully connected branch is used to process the attenuation curve and scanning parameters and extract physical features; The feature fusion layer is used to concatenate multimodal features and input them into the fully connected layer; The output layer generates the photon starvation probability k3 based on the Sigmoid activation function.

[0015] Furthermore, the method further includes reconstructing the collected data to obtain a CT image, wherein the reconstruction of the CT image adopts an iterative reconstruction algorithm, an AI reconstruction algorithm, or a noise reduction algorithm.

[0016] The present invention also provides a micro-dose CT system, based on the above data acquisition method, comprising: frame; An X-ray tube is arranged on the frame, and the X-ray tube comprises a cathode, an anode and a grid-controlled electrode arranged between the cathode and the anode; A high voltage generator is arranged on the frame and connected to the X-ray tube, and is used to provide the X-ray tube with a working voltage and a pulse grid voltage; A detector, arranged on the frame and connected to the X-ray tube, for receiving X-rays and converting them into electrical signals; A data acquisition unit, connected to the detector signal, for acquiring, processing and storing the electrical signal output by the detector; An image reconstruction unit, connected to the data acquisition unit by signal, and used to reconstruct the data acquired by the data acquisition unit to generate a CT image; The control unit is arranged on the frame and is respectively connected to the X-ray tube, the high voltage generator, the detector, the data acquisition unit and the image reconstruction unit by signals.

[0017] Due to the application of the above technical solution, the beneficial effects of the present application compared with the prior art are: The data acquisition method of the microdose CT system of the present application performs photon starvation prediction and dynamically adjusts the sampling angle according to the prediction results to avoid data acquisition in photon starvation areas. At the same time, it combines pulse exposure technology to achieve synchronous exposure and acquisition, thereby improving image quality while maintaining a low radiation dose, thereby ensuring the accuracy of diagnosis. BRIEF DESCRIPTION OF THE DRAWINGS

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

[0019] Figure 1 is a flowchart of the angle selection algorithm in Embodiment 1 of the present invention; Figure 2 Schematic diagram of the timing of the photon starvation state and the acquisition enable signal in Example 1 of the present invention; Figure 3 Schematic diagram of the structure of the X-ray tube in Example 2 of the present invention.

[0020] Description of reference numerals: 1- cathode; 2- anode; 3- gate-controlled electrode. DETAILED DESCRIPTION

[0021] In order to enable those skilled in the art to better understand the solution of the present application, the technical solution in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work should fall within the scope of protection of the present application.

[0022] It should be noted that the terms "first", "second", etc. in the specification and claims of the present application and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequential order. It should be understood that the data used in this way can be interchanged where appropriate, so that the embodiments of the present application described here. In addition, the terms "including" and "having" and any of their variations are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device comprising a series of steps or units is not necessarily limited to those steps or units clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0023] In the present application, the terms "upper", "lower", "left", "right", "front", "back", "top", "bottom", "inner", "outer", "middle", "vertical", "horizontal", "lateral", "longitudinal" and the like indicate positions or positional relationships based on the positions or positional relationships shown in the drawings. These terms are mainly used to better describe the present invention and its embodiments, and are not used to limit the indicated devices, elements or components to have a specific orientation, or to be constructed and operated in a specific orientation.

[0024] In addition, some of the above terms may be used to express other meanings in addition to indicating orientation or positional relationship. For example, the term "on" may also be used to express a certain dependency or connection relationship in some cases. For those skilled in the art, the specific meanings of these terms in the present invention can be understood according to specific circumstances.

[0025] In addition, the terms "installed", "set", "provided with", "connected", "connected", and "socketed" should be understood in a broad sense. For example, it can be a fixed connection, a detachable connection, or an integral structure; it can be a mechanical connection or an electrical connection; it can be a direct connection, or an indirect connection through an intermediate medium, or it can be an internal connection between two devices, elements, or components. For those of ordinary skill in the art, the specific meanings of the above terms in the present invention can be understood according to specific circumstances.

[0026] It should be noted that, in the absence of conflict, the embodiments and features in the embodiments of the present application can be combined with each other. The present application will be described in detail below with reference to the accompanying drawings and in combination with the embodiments.

[0027] Example 1 This embodiment provides a data acquisition method for a microdose CT system, which can be applied to various CT scanning scenarios, such as chest CT screening, pediatric CT examination, etc., and is particularly suitable for occasions where radiation dose needs to be controlled. This application does not limit this. The data acquisition method includes the following steps: Predicting the photon starvation state of a target in tomography; Based on the prediction result, it is determined whether an exposure enable signal and an acquisition enable signal are generated when an acquisition trigger signal is generated at the current tomography angle, and synchronous exposure and acquisition are triggered cyclically according to the exposure enable signal and the acquisition enable signal.

[0028] In this embodiment, the photon starvation state is predicted based on the photon starvation prediction algorithm. The core purpose of the photon starvation prediction algorithm is to optimize the exposure and data acquisition strategy during the CT scanning process by predicting the occurrence of photon starvation during image acquisition, thereby reducing radiation dose and improving image quality.

[0029] In this embodiment, the photon starvation prediction algorithm includes performing a positioning image scan on the target through positioning image attenuation characteristic analysis and prediction, combining the attenuation characteristics of the target scanning area obtained by the positioning image scan to predict the area where the target is photon starved in the tomographic scan; or / and, through real-time data analysis and prediction, counting the total number of pixels m that are photon starved in a single acquisition of the target in the exposure scan, and predicting the changing trend of the total number of pixels m that are photon starved in multiple consecutive acquisitions of the target in the tomographic scan; or / and predicting the probability of the target being photon starved during scanning through an AI model.

[0030] It is worth noting that the above three photon starvation prediction algorithms can be used to perform photon starvation prediction independently, or can be combined in pairs or all three together for photon prediction, and this application does not make specific limitations on this. Of course, in other implementation schemes, other photon starvation prediction algorithms can also be used, and this application does not elaborate on them here.

[0031] Specifically, the photon starvation state is defined as K, the photon starvation index predicted by the positioning image attenuation characteristic analysis is k1, the photon starvation index predicted by real-time data analysis is k2, and the photon starvation probability predicted by the AI ​​model is k3. The photon starvation state prediction is as follows: When the photon starvation state K is predicted independently through the positioning image attenuation characteristic analysis, K=k1, and the threshold of K is 0; when the photon starvation state K is predicted independently through real-time data analysis, K=k2, and the threshold of K is 0; when the photon starvation state K is predicted independently through the AI ​​model, K=k3, and the threshold of K is k0; when the photon starvation state K is predicted jointly through the positioning image attenuation characteristic analysis and real-time data analysis, K=k1+k2, and the threshold of K is 0; when the photon starvation state K is predicted jointly through the positioning image attenuation characteristic analysis and the AI ​​model, K=k1+k3, and the threshold of K is k0; when the photon starvation state K is predicted jointly through the real-time data analysis and the AI ​​model, K=k2+k3, and the threshold of K is k0; when the photon starvation state K is predicted jointly through the positioning image attenuation characteristic analysis, the real-time data analysis and the AI ​​model, K=k1+k2+k3, and the threshold of K is k0.

[0032] It should be noted that the above k0 is the photon starvation judgment threshold in the AI ​​model prediction, and the K threshold of 0 means that when K=0, the CT system is in a normal state. In some optional implementations, k0 is usually greater than a certain value, such as 0.6, or a number of thresholds between 0.6 and 1 are set as judgment conditions of different levels, such as 0.6, 0.7, 0.8, and 0.9. The larger the value, the more relaxed the probability of allowing the photon starvation state to occur.

[0033] In this embodiment, a locator image scan is performed on the target, and the attenuation characteristics of the target scanning area obtained by the locator image scan are combined to predict the photon-starved area of ​​the target in the tomographic scan, specifically including: Collecting a forward positioning image or / and a lateral positioning image orthogonal to the target, wherein the forward positioning image is defined as a 0-degree positioning image and the lateral positioning image is defined as a 90-degree positioning image; Obtain the attenuation characteristics of the target scanning area at 0 degrees and 90 degrees, and calculate the orthogonal attenuation a0 and a90 of each Z coordinate corresponding to the fault plane; Establish the attenuation distribution curve of each Z coordinate corresponding to the fault plane, and obtain the estimated attenuation a(θ) of the target scanning area at each scanning angle in combination with the fault plane shape of the target scanning area. For example, if an elliptical shape is used to simulate the fault plane shape of the human body, the calculation formula of a(θ) is: ; The attenuation threshold of photon starvation is set to bs, and the real-time attenuation of the target scanning area at each scanning angle during scanning is predicted to be b(θ) based on the estimated attenuation a(θ). When b(θ) < bs, the photon starvation index k1 is set to 0, otherwise k1 is set to 1.

[0034] In an optional implementation, a forward positioning image and a lateral positioning image orthogonal to the target are collected, and the calculation formulas of the attenuation a0 and a90 are as follows: ; ; Among them, M and N are the pixel matrix sizes of the detector, For each pixel size, is the offset of each pixel corresponding to the 0-degree positioning image, The offset of each pixel corresponding to the 90-degree positioning image; The calculation formula of real-time attenuation b(θ) is as follows: ; Among them, kVt is the tube voltage of the scout image scan, mAt is the tube current of the scout image scan, kV(θ) is the tube voltage of the tomography scan, mA(θ) is the tube current of the tomography scan, and c is the tube voltage proportionality coefficient. The empirical value of c is generally 1.7.

[0035] In another optional implementation, a forward scout image or a lateral scout image orthogonal to the target is collected, and the attenuation characteristics of another uncollected scout image are obtained by accumulating and normalizing all pixels in the same Z direction. This is a conventional method and will not be described in detail here.

[0036] In some optional implementation schemes, the total number of pixels m that are photon-starved in a single acquisition of the target in an exposure scan is counted, and the change trend of the total number of pixels m that are photon-starved in multiple consecutive acquisitions of the target in a tomographic scan is predicted, specifically including: Collect unexposed data at V angles by rotating within 360 degrees, V ≥ 100, calculate the expected value μ and standard deviation σ of each pixel under the background noise level, and set the photon starvation judgment threshold THR according to the expected value μ and standard deviation σ; Pulse scanning obtains real-time scanning data and performs pixel-level difference calculation, compares all pixels with the photon starvation judgment threshold THR, and obtains P comparison results, P≤M×N, and sets the total number of pixels below the photon starvation judgment threshold THR to m, m≤P, where P is the total number of detector pixels involved in photon starvation prediction, and M and N are the pixel matrix sizes of the detector; Store the three most recent m values: mn, mn-1, mn-2, calculate the acceleration a of the rate of change of the m value, a=(mn-mn-1)-(mn-1-mn-2), predict the next periodic change ∆m=(mn-mn-1)+a, set the total number of pixels where photon starvation occurs, a baseline value mo, and the periodic change baseline value ∆mo, when m>mo and ∆m>∆mo, k2=1, otherwise k2=0.

[0037] It is worth noting that the above-mentioned non-exposure data acquisition step can be selected to be executed at any time when the CT system is not exposed, and is preferably automatically executed once before each exposure acquisition.

[0038] In some optional implementation schemes, an AI model is used to predict the probability of photon starvation of a target during scanning. Specifically, the AI ​​model network adopts a multimodal convolutional neural network, and the input data is a target positioning image and scanning parameter information; or, the input data is real-time scanning data and scanning parameter information; or, the input data is a target positioning image, scanning parameter information, and real-time scanning data, and the output data is a photon starvation probability k3, where k3 ranges from 0 to 1.

[0039] It is worth noting that the attenuation characteristics of the target scanning area are actually derived from the target positioning image. Usually, the positioning image or attenuation characteristics can be directly input. For AI models, the processing is usually simplified and the end-to-end positioning image input is directly used. The scanning parameter information mainly includes the tube current, tube voltage, rotation time, etc. of the tomography. The real-time scanning data is mainly the exposure data of the detector at each angle during the tomography process. The total number of photon-starved pixels m in the real-time scan is also a derivative of this parameter.

[0040] In detail, the AI ​​model network structure includes an input layer, a CNN branch, a fully connected branch, a feature fusion layer, and an output layer. The input layer is used to receive multimodal data, which includes images, attenuation curves, and scanning parameters. The CNN branch is used to process anatomical structure images and extract spatial features. The fully connected branch is used to process attenuation curves and scanning parameters and extract physical features. The feature fusion layer is used to splice multimodal features and input them into the fully connected layer. The output layer generates the photon starvation probability k3 based on the Sigmoid activation function. This network structure achieves accurate prediction of the photon starvation state by fusing multimodal data and combining the spatial features of the image, the attenuation curve, and the physical features of the scanning parameters.

[0041] The training method of the above AI model mainly includes two key steps: data set construction and data enhancement. First of all, the construction of the data set requires the collection of fully annotated CT images. These images should cover the scanning data of different parts of the body, such as the chest, pelvis and other areas, and each image should be equipped with corresponding annotation information. In addition, the scanning technical parameters of the image, such as tube voltage, tube current, attenuation value, scanning intensity and other information, also need to be recorded. These technical parameters can help the model better understand the characteristics and structure of the image. In order to further enhance the model's sensitivity to image quality, it is also necessary to annotate the image with quality issues such as photon starvation to clarify whether there is any degradation in image quality during the scanning process.

[0042] Data augmentation is an effective technique to improve the generalization ability of the model. By rotating, translating, scaling, mirror flipping, and injecting noise into the original image, different scanning conditions and possible image variations can be simulated. This augmentation method helps to expand the diversity of training data and ensure that the model can maintain high robustness and accuracy in the face of various variations and challenges in practical applications, thereby improving its adaptability and performance in different clinical scenarios.

[0043] Please combine Figure 1 and Figure 2 In this embodiment, based on the angle selection algorithm, it is determined whether to generate an exposure enable signal and an acquisition enable signal when an acquisition trigger signal is generated at the current tomography angle, and synchronous exposure and acquisition are triggered cyclically according to the exposure enable signal and the acquisition enable signal. The specific steps are as follows: Step 1: Define the acquisition angle as A and the sparse acquisition angle period as B, where B=2π / D, and D is the total number of sparse acquisition angles; Step 2: Start the rack rotation; Step 3: Detect the gantry rotation angle in real time and generate scanning control signals, including scanning start and end signals. According to the acquisition angle cycle B, insert the acquisition trigger signal between the scanning start and end signals. The angle interval between adjacent acquisition trigger signals is B. When the first acquisition trigger signal starts, trigger a synchronous exposure and acquisition. Step 4: When each acquisition is completed, predict the K value for the next acquisition; If the K value exceeds the threshold, the photon starvation state is activated, and the sum of the unsampled angles W1 is calculated. When W1 ≥ B, the calculated K value is updated after each A angle. When W1 = xB, a synchronous exposure and acquisition is triggered, where x is a positive integer. If the K value is lower than the threshold, the photon starvation state is turned off and a synchronous exposure and acquisition is triggered. After the total angle of the acquired angle W2=yB angle, a synchronous exposure and acquisition is triggered, where y is a positive integer; Repeat the above steps.

[0044] When photon starvation does not occur, the system performs sparse pulse exposure and data acquisition normally, generally with uniform angle exposure and acquisition to achieve micro-dose scanning; when photon starvation occurs, the system dynamically determines the angle interval of pulse exposure and acquisition according to the angle selection algorithm, and the angle interval is no longer uniform, affected by the photon starvation distribution angle. Photon starvation prediction is calculated at least once at each originally evenly distributed angle, and the calculation result is used to update the photon starvation state K. At the same time, by forcing a synchronous exposure and acquisition at W1=xB and W2=yB, the rationality of the acquisition angle distribution is ensured.

[0045] In this embodiment, triggering synchronous exposure and acquisition according to the exposure enable signal and the acquisition enable signal is specifically as follows: The exposure enable signal and the acquisition enable signal are generated simultaneously and have the same duration; Alternatively, the exposure enable signal and the acquisition enable signal are generated sequentially, and the exposure enable signal at least satisfies the requirement of generating a rising edge of the exposure enable start slightly earlier than the acquisition enable signal start time, and generating a falling edge of the exposure enable end slightly later than the acquisition enable signal end time. The advantage of sequential generation is that the X-ray exposure is sufficiently stable during the acquisition enable duration.

[0046] It is worth noting that in some optional implementation schemes, the exposure enable signal generates a rising edge of the exposure enable start several microseconds to several hundred microseconds earlier than the acquisition enable signal starts, and generates a falling edge of the exposure enable end several microseconds to several hundred microseconds later than the acquisition enable signal ends, such as 1 microsecond, 10 microseconds, 20 microseconds, 100 microseconds, 500 microseconds, and 800 microseconds, which are not limited in this application.

[0047] In some optional embodiments, the scan start feature and the scan end feature may be characterized by features of the same period but different duty cycles.

[0048] In some optional implementation schemes, the scan start signal is the first acquisition trigger signal, and the scan end signal is the last acquisition trigger signal.

[0049] The data collection method of the present application also includes reconstructing the collected data to obtain a CT image. The reconstruction of the CT image adopts an iterative reconstruction algorithm, an AI reconstruction algorithm, or a noise reduction algorithm.

[0050] In this embodiment, an AI reconstruction algorithm is used for reconstruction. Specifically, a method of FBP-based reconstruction and neural network high-quality restoration and reconstruction is used. This AI model belongs to the image domain AI model. Combining the computational efficiency of FBP and the image enhancement capability of the neural network, it can effectively remove noise and artifacts, retain anatomical structure details, improve the quality of low-dose and low-angle CT images, meet clinical real-time processing needs, and thus improve the accuracy and value of diagnosis.

[0051] Detailed, FBP basic reconstruction part: 1. Data preprocessing: Obtain projection data, including offset correction, air correction, ray hardening correction, and nonlinear correction.

[0052] 2. FBP filter back projection: After obtaining the projection data through the above steps, the projection data is filtered based on the FBP algorithm, and then the image is reconstructed by back projection. The mathematical model of this process is based on Fourier transform, and the internal structure of the object is inferred from the information in the projection data. This is a conventional method and will not be described in detail here.

[0053] Neural network high-quality restoration and reconstruction part: Network architecture: It uses an improved U-Net network structure, including an encoder-decoder architecture, and retains image detail information through skip connections. Skip connections can help the network make full use of the details in the input image when reconstructing in the decoding stage.

[0054] Improved U-Net network features: Multi-scale feature extraction capability: U-Net extracts multi-scale features in images through multi-level convolution and pooling operations, which helps to improve sensitivity to details of different sizes, especially the recognition of important features such as edges and textures.

[0055] Attention mechanism: enables the network to better focus on important feature areas and ignore unimportant areas, thereby improving the quality of reconstructed images.

[0056] Processing flow: Input: The network input is low-quality CT images, which are obtained by FBP reconstruction and usually have problems such as noise and artifacts.

[0057] Network processing: The improved U-Net network processes low-quality CT images, using its deep feature extraction and restoration capabilities to gradually denoise, restore lost details, and enhance image quality.

[0058] Output: After network processing, the output is a high-quality CT image, in which the noise and artifacts are effectively suppressed and the details are enhanced.

[0059] Training strategy: Use paired datasets for training, i.e. low-quality images and corresponding high-quality reference images. Paired datasets help the network learn how to transform low-quality images into high-quality images.

[0060] Acquisition of high-quality reference images: By performing FBP reconstruction of the original data of the complete scan at all angles, generally about 1000 angles, a high-quality reference image can be obtained.

[0061] Sparse projection data generates low-quality images: The projection data of the full-angle data is downsampled to simulate the sparse projection data in the real scene, and then the FBP algorithm is used to reconstruct it to obtain a low-quality image.

[0062] Data enhancement: In order to improve the generalization ability of the network, data enhancement techniques such as rotation, scaling, and flipping are used to enable the model to handle various image changes.

[0063] Dataset features: This dataset is generated based on completely real CT scan scenes. These data are preprocessed and paired with the FBP algorithm. Since actual CT scan data is used, network training can better simulate real reconstruction scenes, thereby improving the quality of reconstruction results.

[0064] In other embodiments, a projection domain AI model is used to reconstruct CT images. Specifically, sparse projection data is used to generate full-angle projection data through an AI model. Commonly used models include Unet, CNN, or their variants, such as adding residual connections, introducing wavelet transforms, etc. The following are the steps of the method: 1. Data preprocessing: The preprocessing in this stage is the same as the FBP data preprocessing method mentioned above, and the purpose is to obtain sparse projection data. Usually, the sparse projection data will come from 100 angles or less, which will be used as the input of the AI ​​model.

[0065] 2. Model processing: Input the sparse projection data into the trained AI model, and the model will output full-angle projection data. The number of full-angle projection data will generally reach 1,000 angles or more. The purpose is to supplement the missing angle information for subsequent image reconstruction.

[0066] 3. Image reconstruction: Using the generated full-angle projection data, the traditional FBP algorithm is applied to reconstruct the CT image. In this way, a relatively complete CT image can be obtained, although the initial input data is sparse.

[0067] 4. Model training: The training method is similar to the training process of other AI models. First, use the real complete scan data, and then downsample it to obtain sparse projection data by reducing the number of angles to form a paired data set. These paired data sets are used to train the AI ​​model so that it can accurately predict the full-angle projection data from the sparse projection data.

[0068] The advantage of this method is that the AI ​​model can recover complete projection data from limited projection data by learning the rules and structure of image reconstruction, thereby improving scanning efficiency and reducing dependence on large amounts of projection data.

[0069] In some other embodiments, the image domain AI model is used in combination with the projection domain AI model, which is not specifically limited in the present application.

[0070] Example 2 This embodiment provides a micro-dose CT system. Based on the data acquisition method of Embodiment 1, the micro-dose CT system includes a rack and an X-ray tube, a control unit, a high-voltage generator and a detector, a data acquisition unit, and an image reconstruction unit arranged on the rack. The control unit is respectively connected to the X-ray tube, the high-voltage generator, the detector, the data acquisition unit, and the image reconstruction unit by signals.

[0071] For details, see Figure 3 The X-ray tube includes a cathode 1, an anode 2, and a grid-controlled electrode 3 disposed between the cathode 1 and the anode 2. Only when a high voltage is applied between the cathode 1 and the anode 2, electrons are emitted from the cathode 1 to the anode 2 to generate X-rays. When a certain voltage is applied to the grid-controlled electrode 3, electrons will no longer be emitted to the anode 2, and the generation of X-rays will be interrupted. By intermittently applying a voltage to the grid-controlled electrode 3, intermittent pulse exposure can be achieved.

[0072] The high voltage generator is connected to the X-ray tube to provide the X-ray tube with an operating voltage and a pulsed grid voltage. It is worth noting that during the duration of the scanning control signal, the tube voltage of the X-ray tube is continuously loaded, while the voltage of the grid-controlled electrode 3 is kept loaded by default.

[0073] The detector is set on the frame and connected to the X-ray tube to receive X-rays and convert them into electrical signals.

[0074] The data acquisition unit is connected to the detector signal and is used to collect, process and store the electrical signal output by the detector. Furthermore, it is preferred that the data acquisition unit has an FPGA to realize real-time data analysis, perform real-time parameter calculation on the collected data, calculate the photon hunger index k2, calculate the photon hunger statistical value m in real time, and provide k2 to the control unit.

[0075] The image reconstruction unit is connected to the data acquisition unit signal, and is used to reconstruct the data acquired by the data acquisition unit to generate a CT image. Furthermore, it participates in the calculation of the photon starvation prediction and angle selection algorithm, calculates the photon starvation index k1, undertakes the photon starvation state prediction of the AI ​​model, obtains the m value of the data acquisition unit in real time, obtains the exposure parameters of the control unit in real time, uses the AI ​​model to infer and predict the photon starvation probability k3, and provides the photon starvation state k1 or / and k3 to the control unit.

[0076] The control unit is used to control the scanning frame, X-ray tube, detector and data acquisition unit, provide exposure related parameters, such as tube voltage, tube current, angle information, etc., obtain k1, k2, k3 and fuse them, make the final prediction of the photon starvation state, implement the process control of the angle selection algorithm based on the prediction results, generate exposure and data acquisition synchronization signals according to the angle selection algorithm, and realize accurate pulse exposure and data acquisition synchronization. It is worth noting that the control unit can use a single-chip microcomputer or an embedded processor as the core controller.

[0077] Finally, it should be noted that the above are only preferred embodiments of the present invention and are not intended to limit the present invention. Although the present invention has been described in detail with reference to the aforementioned embodiments, it is still possible for those skilled in the art to modify the technical solutions described in the aforementioned embodiments, or to make equivalent substitutions for some of the technical features therein. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the protection scope of the present invention.

Claims

1. A data acquisition method for a microdose CT system, characterized in that: The following steps are involved: Predicting the photon starvation state of a target in tomography; Based on the prediction result, it is determined whether an exposure enable signal and an acquisition enable signal are generated when an acquisition trigger signal is generated at the current tomography angle, and synchronous exposure and acquisition are triggered cyclically according to the exposure enable signal and the acquisition enable signal.

2. A data acquisition method for a microdose CT system as claimed in claim 1, characterized in that: The photon starvation state is predicted based on a photon starvation prediction algorithm, wherein the photon starvation prediction algorithm comprises: Through the analysis and prediction of the attenuation characteristics of the locator image, the target is scanned with a locator image, and the photon-starved area of ​​the target in the tomographic scan is predicted by combining the attenuation characteristics of the target scanning area obtained by the locator image scan; Or / and, through real-time data analysis and prediction, the total number of pixels m that are photon-starved in a single acquisition of the target in the exposure scan is counted, and the change trend of the total number of pixels m that are photon-starved in multiple consecutive acquisitions of the target in the tomography scan is predicted; Or / and, the AI ​​model is used to predict the probability of a target being photon starved during scanning.

3. The data acquisition method of a micro-dose CT system according to claim 2, characterized in that: The photon starvation state is defined as K, the photon starvation index predicted by the positioning image attenuation characteristic analysis is k1, the photon starvation index predicted by the real-time data analysis is k2, and the photon starvation probability predicted by the AI ​​model is k3; When the photon starvation state K is predicted independently by analyzing the attenuation characteristics of the localization image, K=k1, and the threshold of K is 0; When the photon starvation state K is predicted independently through real-time data analysis, K = k2, and the threshold of K is 0; When the photon starvation state K is predicted independently by the AI ​​model, K=k3, and the threshold of K is k0; When the photon starvation state K is predicted by the analysis of the attenuation characteristics of the positioning image and the real-time data analysis, K=k1+k2, and the threshold of K is 0; When the photon starvation state K is predicted by the positioning image attenuation characteristic analysis and the AI ​​model, K=k1+k3, and the threshold of K is k0; When the photon starvation state K is predicted by real-time data analysis and AI model, K=k2+k3, and the threshold of K is k0; When the photon starvation state K is predicted through positioning image attenuation characteristic analysis, real-time data analysis and AI model, K=k1+k2+k3, and the threshold of K is k0; Among them, k0 is the photon starvation judgment threshold in the AI ​​model prediction.

4. The data acquisition method of a microdose CT system according to any one of claims 1 to 3, characterized in that: Based on the angle selection algorithm, it is determined whether to generate an exposure enable signal and an acquisition enable signal when an acquisition trigger signal is generated at the current tomography angle, and synchronous exposure and acquisition are triggered cyclically according to the exposure enable signal and the acquisition enable signal.

5. The data acquisition method of a micro-dose CT system according to claim 4, characterized in that: Based on the angle selection algorithm, it is determined whether to generate an exposure enable signal and an acquisition enable signal when an acquisition trigger signal is generated at the current tomography angle, and synchronous exposure and acquisition are triggered cyclically according to the exposure enable signal and the acquisition enable signal, specifically including: Step 1: Define the acquisition angle as A and the sparse acquisition angle period as B, where B=2π / D, and D is the total number of sparse acquisition angles; Step 2: Start the rack rotation; Step 3: Detect the gantry rotation angle in real time and generate scanning control signals, including scanning start and end signals. According to the acquisition angle cycle B, insert the acquisition trigger signal between the scanning start and end signals. The angle interval between adjacent acquisition trigger signals is B. When the first acquisition trigger signal starts, trigger a synchronous exposure and acquisition. Step 4: When each acquisition is completed, predict the K value for the next acquisition; If the K value exceeds the threshold, the photon starvation state is activated, and the sum of the unsampled angles W1 is calculated. When W1 ≥ B, the calculated K value is updated after each A angle. When W1 = xB, a synchronous exposure and acquisition is triggered, where x is a positive integer. If the K value is lower than the threshold, the photon starvation state is turned off and a synchronous exposure and acquisition is triggered. After the total angle of the acquired angle W2=yB angle, a synchronous exposure and acquisition is triggered, where y is a positive integer; Repeat the above steps.

6. The data acquisition method of a micro-dose CT system according to claim 5, characterized in that: Trigger synchronous exposure and acquisition according to the exposure enable signal and acquisition enable signal. Specifically: The exposure enable signal and the acquisition enable signal are generated simultaneously and have the same duration; Alternatively, the exposure enable signal and the acquisition enable signal are generated successively, and the exposure enable signal at least satisfies the requirement of generating a rising edge of the exposure enable start slightly earlier than the start time of the acquisition enable signal, and generating a falling edge of the exposure enable end slightly later than the end time of the acquisition enable signal.

7. The data acquisition method of a micro-dose CT system according to claim 3, characterized in that: Perform a scouting image scan on the target, and combine the attenuation characteristics of the target scanning area obtained by the scouting image scan to predict the photon starved area of ​​the target in the tomography scan, including: Collecting a forward positioning image or / and a lateral positioning image orthogonal to the target, wherein the forward positioning image is defined as a 0-degree positioning image, and the lateral positioning image is defined as a 90-degree positioning image; Obtain the attenuation characteristics of the target scanning area at 0 degrees and 90 degrees, and calculate the orthogonal attenuation a0 and a90 of each Z coordinate corresponding to the fault plane; Establish the attenuation distribution curve of each Z coordinate corresponding to the fault plane, and obtain the estimated attenuation a (θ) of the target scanning area at each scanning angle based on the fault plane shape of the target scanning area; The attenuation threshold of photon starvation is set to bs, and the real-time attenuation of the target scanning area at each scanning angle during scanning is predicted to be b(θ) based on the estimated attenuation a(θ). When b(θ) < bs, the photon starvation index k1 is set to 0, otherwise k1 is set to 1.

8. The data acquisition method of a micro-dose CT system according to claim 7, characterized in that: The calculation formulas for attenuation a0 and a90 are as follows: ； ； Among them, M and N are the pixel matrix sizes of the detector, For each pixel size, is the offset of each pixel corresponding to the 0-degree positioning image, The offset of each pixel corresponding to the 90-degree positioning image; The calculation formula of real-time attenuation b(θ) is as follows: ； Wherein, kVt is the tube voltage for scout image scanning, mAt is the tube current for scout image scanning, kV(θ) is the tube voltage for tomography scanning, mA(θ) is the tube current for tomography scanning, and c is the tube voltage proportionality coefficient.

9. The data acquisition method of a micro-dose CT system according to claim 3, characterized in that: The total number of pixels m that are photon-starved in a single acquisition of the target in the exposure scan is counted, and the change trend of the total number of pixels m that are photon-starved in multiple consecutive acquisitions of the target in the tomography scan is predicted, including: Collect unexposed data at V angles by rotating within 360 degrees, V ≥ 100, calculate the expected value μ and standard deviation σ of each pixel under the background noise level, and set the photon starvation judgment threshold THR according to the expected value μ and standard deviation σ; Pulse scanning obtains real-time scanning data and performs pixel-level difference calculation, compares all pixels with the photon starvation judgment threshold THR, and obtains P comparison results, P≤M×N, and sets the total number of pixels below the photon starvation judgment threshold THR to m, m≤P, where P is the total number of detector pixels involved in photon starvation prediction, and M and N are the pixel matrix sizes of the detector; Store the three most recent m values: mn, mn-1, mn-2, calculate the acceleration a of the rate of change of the m value, a=(mn-mn-1)-(mn-1-mn-2), predict the next periodic change ∆m=(mn-mn-1)+a, set the total number of pixels where photon starvation occurs, a baseline value mo, and the periodic change baseline value ∆mo, when m>mo and ∆m>∆mo, k2=1, otherwise k2=0.

10. The data acquisition method of a micro-dose CT system according to claim 3, characterized in that: The AI ​​model is used to predict the probability of photon starvation of the target during scanning. Specifically: The AI ​​model network uses a multimodal convolutional neural network; Input data: Target positioning image and scanning parameter information; Or, real-time scanned data and scan parameter information; Or, target positioning image, scanning parameter information and real-time scanning data; Output data: photon starvation probability k3, the range of k3 is 0~1.

11. The data acquisition method of a micro-dose CT system according to claim 10, characterized in that: The AI ​​model network structure includes: An input layer, for receiving multimodal data, wherein the multimodal data includes images, attenuation curves and scanning parameters; The CNN branch is used to process anatomical structure images and extract spatial features; The fully connected branch is used to process the attenuation curve and scanning parameters and extract physical features; The feature fusion layer is used to concatenate multimodal features and input them into the fully connected layer; The output layer generates the photon starvation probability k3 based on the Sigmoid activation function.

12. The data acquisition method of a micro-dose CT system according to claim 1, characterized in that: The method also includes reconstructing the collected data to obtain a CT image, wherein the reconstruction of the CT image adopts an iterative reconstruction algorithm, an AI reconstruction algorithm, or a noise reduction algorithm.

13. A microdose CT system, based on the data acquisition method according to any one of claims 1 to 12, characterized in that: include: frame; An X-ray tube is arranged on the frame, and the X-ray tube comprises a cathode, an anode and a grid-controlled electrode arranged between the cathode and the anode; A high voltage generator is arranged on the frame and connected to the X-ray tube, and is used to provide the X-ray tube with a working voltage and a pulse grid voltage; A detector, arranged on the frame and connected to the X-ray tube, for receiving X-rays and converting them into electrical signals; A data acquisition unit, connected to the detector signal, for acquiring, processing and storing the electrical signal output by the detector; An image reconstruction unit, connected to the data acquisition unit by signal, and used to reconstruct the data acquired by the data acquisition unit to generate a CT image; The control unit is arranged on the frame and is respectively connected to the X-ray tube, the high voltage generator, the detector, the data acquisition unit and the image reconstruction unit by signals.

Images