An automatic configuration system for X-ray machine examination parameters for pet examination

Abstract

This invention discloses an automatic parameter configuration system for X-ray examinations of pets, relating to the field of medical imaging equipment technology. The invention uses a displacement prediction model to predict the pet's convulsive state before exposure, and a dynamic exposure trigger activates only within a stable range, avoiding motion blur at its source. Compared to existing technologies that rely on post-exposure image feedback adjustment, this system avoids ineffective radiation and repeated exposure operations. A partitioned parameter mapping module, combined with a species characteristic database, maps the grayscale of the preview image to an equivalent thickness and independently generates region parameters. For extreme body size differences, such as the abdominal fat layer of obese dogs and the rib area of ​​emaciated cats, the system automatically assigns differentiated kV / mA parameters, eliminating overexposure or underexposure problems caused by traditional fixed penetration rate standards. A radiation fuse unit monitors the cumulative dose in real time and dynamically adjusts the safety threshold according to the pet's weight, terminating exposure before the dose exceeds the limit.

Description

Technical Field

[0001] This invention relates to the field of medical imaging equipment technology, and in particular to an automatic configuration system for X-ray machine examination parameters for pet examination. Background Technology

[0002] In veterinary clinical diagnosis, X-ray machines frequently need to deal with sudden convulsions or tremors in animals. These involuntary movements are common in scenarios involving pain stress, neurological diseases, or fear reactions, and are especially prevalent in examinations of pets that have not been adequately sedated. Newer automatic parameter configuration systems, such as adjustment schemes based on real-time image feedback, rely on post-exposure image quality assessment and dynamically adjust kV / mA parameters by comparing them with preset indicators.

[0003] The aforementioned systems have inherent limitations in dealing with sudden movements. Parameter adjustments require waiting for the entire imaging process to finish, with a delay of hundreds of milliseconds from exposure to feedback adjustment. Pet convulsions often occur within milliseconds, and existing solutions cannot respond to instantaneous motion changes within a single exposure cycle. Some solutions introduce external motion sensors, such as millimeter-wave radar to trigger emergency stops, but these can only interrupt the exposure and fail to achieve real-time parameter adaptation under motion conditions.

[0004] Motion artifacts cause image blurring, and the system can only compensate by repeated exposures, but cannot eliminate the ineffective radiation of the first exposure. At the same time, external sensors require modification of the equipment hardware structure, increasing the deployment cost in clinics. The existing parameter library also does not include the dose optimization model in the convulsive state, resulting in a break in the automated configuration chain for motion scenarios. Summary of the Invention

[0005] In view of the aforementioned existing problems, the present invention is proposed.

[0006] This invention provides an automatic configuration system for X-ray machine examination parameters for pet examinations, which solves the problems of existing systems being unable to suppress motion artifacts caused by pet twitching due to response delays and lacking cross-species parameter adaptation capabilities.

[0007] To solve the above-mentioned technical problems, the present invention provides the following technical solution:

[0008] This invention provides an automatic configuration system for X-ray machine examination parameters for pet examinations, including an X-ray generator, a detector, and a control unit, wherein:

[0009] The control unit integrates a displacement prediction model, which is used to analyze the continuous frame data stream collected by the detector in real time and output the predicted displacement.

[0010] The system is also equipped with a dynamic exposure trigger and a partition parameter mapping module. When the predicted displacement is lower than a preset threshold, the X-ray generator is activated to perform exposure according to the optimized parameters. The partition parameter mapping module is used to generate the optimized parameters.

[0011] As a preferred embodiment of the automatic configuration system for X-ray examination parameters for pet examination described in this invention, the displacement prediction model is a processing module based on a time series prediction algorithm, and its input includes grayscale matrix data of the current frame and at least 3 historical frames.

[0012] As a preferred embodiment of the automatic configuration system for X-ray examination parameters for pet examination described in this invention, wherein: in the displacement prediction model, after receiving the grayscale matrix of consecutive frames, the optical flow field is extracted using the Lucas-Conrad operator:

[0013] F t =LK(I t-1 ,I t ),

[0014] Among them, F t Let LK(·) represent the optical flow field of frame t, and let I represent the Lucas-Conrad optical flow operator. t-1 I t These are the detector grayscale matrices for frames t-1 and t, respectively.

[0015] The average displacement is obtained by taking the modulus of each pixel and then averaging the values.

[0016]

[0017] Among them, s t Let u be the average displacement in frame t, and N be the total number of valid pixels included in the statistics. i (t) is the displacement vector of the i-th pixel in frame t;

[0018] Construct the first-order difference:

[0019] Δs t =s t -s t-1 ,

[0020] Where, Δs t s is the difference between the average displacement of frame t and frame (t-1). t-1 This represents the average displacement in frame t-1.

[0021] Modeling with ARIMA-(2,1,1):

[0022] Δs t =φ1Δs t-1 +φ2Δs t-2 +θ1ε t-1+ε t ,

[0023] Where φ1 and φ2 are autoregressive coefficients, θ1 is the moving average coefficient, and ε t-1 ,ε t Zero-mean random residuals;

[0024] Adaptive coefficient update using recursive least squares:

[0025]

[0026] Among them, K t Let P be the gain vector at step t. t-1 ,P t Let be the covariance matrix, and λ be the forgetting factor, 0 < λ << 1. Let y be the regression vector. t =Δs t For observation purposes, This is the updated coefficient vector;

[0027] Predict the displacement in the next frame:

[0028]

[0029] in, This is the predicted value of the average displacement for the (t+1)th frame.

[0030] As a preferred embodiment of the automatic configuration system for X-ray examination parameters for pets described in this invention, the optimized parameters are generated through a partitioned parameter mapping module, which performs the following operations:

[0031] Acquire a preview image output from the X-ray generator in low-dose mode;

[0032] Divide the preview image into gridded areas;

[0033] Calculate the equivalent thickness based on the grayscale values ​​of each region;

[0034] Each grid cell is independently generated with kV and mA parameter sets based on its equivalent thickness.

[0035] As a preferred embodiment of the automatic configuration system for X-ray examination parameters for pets described in this invention, the equivalent thickness calculation is associated with a species feature database, which stores the mapping relationship of anatomical features of different pet species, wherein the anatomical features include bone density and the equivalent aluminum thickness of soft tissue.

[0036] As a preferred embodiment of the automatic configuration system for X-ray examination parameters for pet examination described in this invention, the species feature database includes a parameter migration interface for receiving new species anatomical feature data input by the user and associating it with the partition parameter mapping module through an interpolation adaptation algorithm.

[0037] As a preferred embodiment of the automatic configuration system for X-ray examination parameters for pet examination described in this invention, the step of generating optimized parameters by the partition parameter mapping module includes:

[0038] After acquiring the preview image in low-dose mode, the image is divided into a grid of equal size:

[0039] G p,q ={(x,y)||x∈[x p ,x p+1 ),y∈[y q ,y q+1 )},

[0040] Among them, G p,q Let x, y be the pixel coordinates of the p-th row and q-th column. p ,x p+1 ,y q ,y q+1 For grid boundaries;

[0041] Calculate the average gray level of the grid:

[0042]

[0043] in, Let |G| be the average gray level of the p-th, q-th grid. p,q | represents the number of pixels within the grid, I prev (x,y) represents the grayscale values ​​of the preview image;

[0044] Calculate the equivalent thickness using the calibrated logarithmic mapping:

[0045]

[0046] Among them, T p,q α is the equivalent thickness of the p-th and q-th grids in mm, α is the thickness scaling factor in mm, and β is the gray-level normalization constant.

[0047] Preliminary exposure parameters are generated using a quadratic polynomial:

[0048]

[0049] Among them, kV p,q This refers to the local tube voltage, in kV and mA. p,qThe local tube current is given in mA, and k1, k2, k3, m1, m2, m3 are polynomial coefficients obtained through offline calibration.

[0050] If a balance needs to be struck between clarity and dosage, multi-objective optimization can be performed in a localized solution:

[0051]

[0052] Among them, Q p,q Let D be the image quality index for the p-th and q-th grids. p,q For dosage indicators, w Q ,w D It is a positive weight used to adjust the priority of image quality and dosage.

[0053] As a preferred embodiment of the automatic configuration system for X-ray machine examination parameters for pet examination described in this invention, it further includes a micro-motion phase analysis module, which extracts frequency domain feature components from the detector's raw data stream and sends an enable signal to the dynamic exposure trigger when the breathing phase is within a preset phase range.

[0054] As a preferred embodiment of the automatic configuration system for X-ray machine examination parameters for pet examination described in this invention, the step of extracting frequency domain feature components from the detector's raw data stream in the micro-motion phase analysis module includes:

[0055] In continuous acquisition mode, the original grayscale data D of the detector is obtained in frame t. t (x,y), first perform spatial integration over the preset region of interest Ω to obtain a one-dimensional breathing sequence:

[0056]

[0057] Among them, R t Let |Ω| represent the average grayscale of the ROI in frame t, and |Ω| represent the total number of pixels within the ROI. t (x,y) represents the grayscale value of pixel (x,y) in frame t, where x and y are pixel coordinates;

[0058] Zero-mean processing is applied within a sliding window of length L to eliminate slow drift. The processing formula is as follows:

[0059]

[0060] Where, r t For the baseline removal instantaneous signal, L is the number of sliding window frames, and l is the relative frame number within the window;

[0061] Perform short-time Fourier transform within the same window:

[0062]

[0063] Among them, S t (f k ) is the complex spectrum at the center of the t-th frame and frequency index k, and f k = kF s / L is the discrete frequency, F s is the frame sampling frequency, w n is the window function weight, n is the time index, and j is the imaginary unit;

[0064] Calculate the power spectrum:

[0065] P t (f k ) = |S t (f k )| 2 ,

[0066] Among them, P t (f k ) is the power spectral density, and |·| is the modulus operation;

[0067] At the pet's breathing zone Main frequency:

[0068]

[0069] Among them, f peak is the current main respiratory frequency, F min , f max are the lower and upper limits of the breathing zone;

[0070] According to f peak Adaptive design of an IIR band-pass filter, and the digital difference equation is written as:

[0071]

[0072] Among them, r bp (t) is the band-pass output, b m are the order-M numerator coefficients, M is the <IIR< filter numerator order, a n are the order-N f denominator coefficients, N f is the denominator order, and m and n are the numerator and denominator term indices respectively;

[0073] Use the Hilbert transform to form an analytic signal:

[0074]

[0075] Among them, z t is the analytic signal, is the discrete Hilbert operator;

[0076] Obtain the instantaneous phase from the analytic signal:

[0077]

[0078] in, For the instantaneous breathing phase, Im{·} and Re{·} take the imaginary and real parts respectively, and atan2(·,·) is the arctangent function in the four quadrants;

[0079] when At that time, an enable signal is sent to the dynamic exposure trigger, wherein, The boundary of the phase interval that allows for exposure.

[0080] As a preferred embodiment of the automatic configuration system for X-ray machine examination parameters for pet examination described in this invention, the system further includes a radiation fuse unit, which performs the following:

[0081] Receive the cumulative dose value fed back by the detector in real time;

[0082] When the cumulative dose exceeds the dynamic threshold generated based on the pet's weight, an exposure termination command is output to the X-ray generator.

[0083] The beneficial effects of this invention are as follows: This invention uses a displacement prediction model to predict the pet's twitching state before exposure, and the dynamic exposure trigger is activated only in the stable range, avoiding motion blur at its source. Compared to existing technologies that rely on post-exposure image feedback adjustment, this system avoids ineffective radiation and repeated exposure operations; the partition parameter mapping module, combined with a species characteristic database, maps the grayscale of the preview image to an equivalent thickness and independently generates region parameters. For extreme body size differences, such as the abdominal fat layer of obese dogs and the rib area of ​​emaciated cats, the system automatically assigns differentiated kV / mA parameters, eliminating overexposure or underexposure problems caused by traditional fixed transmittance standards.

[0084] The radiation-breaking unit of this invention monitors the cumulative dose in real time and dynamically adjusts the safety threshold according to the pet's weight, terminating exposure before the dose exceeds the limit. Compared to industrial equipment that increases redundant dose for safety, this invention can ensure image quality while strictly adhering to ethical limits for live animals. The micro-motion phase analysis module directly extracts respiratory frequency domain features from the detector data stream to lock the optimal exposure phase for pets with shallow breathing. It eliminates the need for external sensors, solving the problem of existing hardware solutions failing to detect micro-motion, and improving the success rate of single exposures. It introduces a species characteristic database, supporting user input of new species parameters and rapid adaptation for exotic pet examinations through interpolation. Veterinarians only need to select the species / weight, and the system automatically completes parameter optimization, eliminating the time-consuming manual adjustments. Attached Figure Description

[0085] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the following description of the embodiments will be briefly introduced. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0086] Figure 1 This is a schematic diagram of the framework of the automatic configuration system for X-ray machine examination parameters for pet examination in Example 1. Detailed Implementation

[0087] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings.

[0088] Many specific details are set forth in the following description in order to provide a full understanding of the invention. However, the invention may also be practiced in other ways different from those described herein, and those skilled in the art can make similar extensions without departing from the spirit of the invention. Therefore, the invention is not limited to the specific embodiments disclosed below.

[0089] Secondly, the term "one embodiment" or "embodiment" as used herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The phrase "in one embodiment" appearing in different places in this specification does not necessarily refer to the same embodiment, nor is it a single or selective embodiment that is mutually exclusive with other embodiments.

[0090] Example 1, referring to Figure 1 This embodiment provides an automatic configuration system for X-ray machine examination parameters for pet examinations, including an X-ray generator, a detector, and a control unit, wherein:

[0091] The control unit integrates a displacement prediction model to analyze the continuous frame data stream collected by the detector in real time and output the predicted displacement.

[0092] The system is also equipped with a dynamic exposure trigger and a partition parameter mapping module. When the predicted displacement is lower than the preset threshold, the X-ray generator is activated to perform exposure according to the optimized parameters; the partition parameter mapping module is used to generate the optimized parameters.

[0093] The displacement prediction model is a processing module based on a time series prediction algorithm. Its input includes grayscale matrix data of the current frame and at least three historical frames.

[0094] In the displacement prediction model, after receiving the grayscale matrix of consecutive frames, the optical flow field is extracted using the Lucas-Conrad operator:

[0095] F t =LK(It-1 ,I t ),

[0096] Among them, F t Let LK(·) represent the optical flow field of frame t, and let I represent the Lucas-Conrad optical flow operator. t-1 I t These are the detector grayscale matrices for frames t-1 and t, respectively.

[0097] The average displacement is obtained by taking the modulus of each pixel and then averaging the values.

[0098]

[0099] Among them, s t Let u be the average displacement in frame t, and N be the total number of valid pixels included in the statistics. i (t) is the displacement vector of the i-th pixel in frame t;

[0100] Construct the first-order difference:

[0101] Δs t =s t -s t-1 ,

[0102] Where, Δs t s is the difference between the average displacement of frame t and frame (t-1). t-1 This represents the average displacement in frame t-1.

[0103] Modeling with ARIMA-(2,1,1):

[0104] Δs t =φ1Δs t-1 +φ2Δs t-2 +θ1ε t-1 +ε t ,

[0105] Where φ1 and φ2 are autoregressive coefficients, θ1 is the moving average coefficient, and ε t-1 ,ε t Zero-mean random residuals;

[0106] Adaptive coefficient update using recursive least squares:

[0107]

[0108] Among them, K t Let P be the gain vector at step t. t-1 ,P t Let be the covariance matrix, and λ be the forgetting factor, 0 < λ << 1. Let y be the regression vector. t =Δs t For observation purposes, This is the updated coefficient vector;

[0109] Predict the displacement in the next frame:

[0110]

[0111] in, This is the predicted value of the average displacement for the (t+1)th frame;

[0112] Specifically, this model converts the optical flow average into a quantifiable displacement sequence, then uses differential ARIMA to suppress the trend term and capture respiratory amplitude and jitter; recursive least squares real-time correction coefficients enable the prediction to maintain stable accuracy under different body sizes and positions; the dynamic exposure trigger therefore only works within a low displacement window, shortening the exposure waiting time and improving image consistency.

[0113] Optimization parameters are generated through the partition parameter mapping module, which performs the following operations:

[0114] Acquire a preview image output from the X-ray generator in low-dose mode;

[0115] Divide the preview image into gridded areas;

[0116] Calculate the equivalent thickness based on the grayscale values ​​of each region;

[0117] Each grid cell is generated independently based on its equivalent thickness, using kV and mA parameter sets.

[0118] Equivalent thickness calculation is associated with a species feature database, which stores the mapping relationship of anatomical features of different pet species, including the equivalent aluminum thickness of bone density and soft tissue.

[0119] The species feature database includes a parameter migration interface, which receives new species anatomical feature data input by the user and associates it with the partition parameter mapping module through an interpolation adaptation algorithm;

[0120] The steps in generating optimization parameters by the partition parameter mapping module include:

[0121] After acquiring the preview image in low-dose mode, the image is divided into a grid of equal size:

[0122] G p,q ={(x,y)||x∈[x p ,x p+1 ),y∈[y q ,y q+1 )},

[0123] Among them, G p,q Let x, y be the pixel coordinates of the p-th row and q-th column. p ,xp+1 ,y q ,y q+1 For grid boundaries;

[0124] Calculate the average gray level of the grid:

[0125]

[0126] in, Let |G| be the average gray level of the p-th, q-th grid. p,q | represents the number of pixels within the grid, I prev (x,y) represents the grayscale values ​​of the preview image;

[0127] Calculate the equivalent thickness using the calibrated logarithmic mapping:

[0128]

[0129] Among them, T p,q α is the equivalent thickness of the p-th and q-th grids in mm, α is the thickness scaling factor in mm, and β is the gray-level normalization constant.

[0130] Preliminary exposure parameters are generated using a quadratic polynomial:

[0131]

[0132] Among them, kV p,q This refers to the local tube voltage, in kV and mA. p,q The local tube current is given in mA, and k1, k2, k3, m1, m2, m3 are polynomial coefficients obtained through offline calibration.

[0133] If a balance needs to be struck between clarity and dosage, multi-objective optimization can be performed in a localized solution:

[0134]

[0135] Among them, Q p,q Let D be the image quality index for the p-th and q-th grids. p,q For dosage indicators, w Q ,w D A positive weight is used to adjust the priority between image quality and dosage;

[0136] Specifically, this module can infer the local equivalent thickness with a single low-dose preview, and then obtain coarse exposure parameters through polynomial fast mapping; multi-target optimization automatically balances image quality and dose in dense bone and soft tissue areas, avoiding overexposure or underexposure caused by overall settings, and reducing repeated adjustments by the operator; real-time computation is low, making it suitable for online execution by embedded control units;

[0137] It also includes a micro-motion phase analysis module, which extracts frequency domain feature components from the detector's raw data stream and sends an enable signal to the dynamic exposure trigger when the breathing phase is in a preset phase range.

[0138] In the micro-motion phase analysis module, the steps for extracting frequency domain feature components from the detector's raw data stream include:

[0139] In continuous acquisition mode, the original grayscale data D of the detector is obtained in frame t. t (x,y), first perform spatial integration over the preset region of interest Ω to obtain a one-dimensional breathing sequence:

[0140]

[0141] Among them, R t Let |Ω| represent the average grayscale of the ROI in frame t, and |Ω| represent the total number of pixels within the ROI. t (x,y) represents the grayscale value of pixel (x,y) in frame t, where x and y are pixel coordinates;

[0142] Zero-mean processing is applied within a sliding window of length L to eliminate slow drift. The processing formula is as follows:

[0143]

[0144] Where, r t For the baseline removal instantaneous signal, L is the number of sliding window frames, and l is the relative frame number within the window;

[0145] Perform short-time Fourier transform within the same window:

[0146]

[0147] Among them, S t (f k f is the complex spectrum at the center of the t-th frame and at frequency index k. k =kF s / L is the discrete frequency, F s w is the frame sampling frequency. n Here, n is the time index and j is the imaginary unit;

[0148] Calculate the power spectrum:

[0149] P t (f k )=|S t (f k )| 2 ,

[0150] Among them, P t (f k ) represents the power spectral density, and |·| represents the modulus operation;

[0151] In the pet breathing zone Main frequency:

[0152]

[0153] Among them, f peak is the current main breathing frequency, f min , f max lower and upper limits of the breathing zone;

[0154] According to f peak Adaptive design of an IIR band-pass filter, the digital difference equation is written as:

[0155]

[0156] Among them, r bp (t) is the band-pass output, b m is the order M numerator coefficient, M is the <IIR< filter numerator order, a n is the order N f denominator coefficient, N f is the denominator order, m and n are the numerator and denominator term indices respectively;

[0157] Use the Hilbert transform to form an analytic signal:

[0158]

[0159] Among them, z t is the analytic signal, is the discrete Hilbert operator;

[0160] Find the instantaneous phase from the analytic signal:

[0161]

[0162] Among them, is the instantaneous breathing phase, Im{·}, Re{·} respectively take the imaginary and real parts, and atan2(·,·) is the four-quadrant arctangent function;

[0163] When , send an enable signal to the dynamic exposure trigger, where is the phase interval boundary allowing exposure;

[0164] Specifically, this algorithm compresses the two-dimensional original grayscale stream into a one-dimensional breathing sequence, uses STFT to capture spectral peaks in real time and adaptively filters according to the main frequency, achieving fast locking of the breathing rhythms of different species and body positions; instantaneous phase discrimination makes the exposure fall in the interval with the minimum breathing movement speed, significantly reducing motion blur and maintaining image sharpness in the low-dose mode;

[0165] The system also includes a radiation fuse unit, which performs the following:

[0166] Receive the cumulative dose value fed back by the detector in real time;

[0167] When the cumulative dose exceeds the dynamic threshold generated based on the pet's weight, an exposure termination command is output to the X-ray generator.

[0168] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention, and all such modifications or substitutions should be covered within the scope of the claims of the present invention.

Claims

1. An automatic configuration system for X-ray machine examination parameters for pet examination, comprising an X-ray generator, a detector, and a control unit, characterized in that: The control unit integrates a displacement prediction model, which is used to analyze the continuous frame data stream collected by the detector in real time and output the predicted displacement. The system is also equipped with a dynamic exposure trigger and a partition parameter mapping module. When the predicted displacement is lower than a preset threshold, the X-ray generator is activated to perform exposure according to the optimized parameters. The partition parameter mapping module is used to generate the optimized parameters. The displacement prediction model is a processing module based on a time series prediction algorithm, and its input includes grayscale matrix data of the current frame and at least three historical frames. In the displacement prediction model, after receiving the grayscale matrix of consecutive frames, the optical flow field is extracted using the Lucas-Conrad operator: ， in, For the first Frame light flow field, This represents the Lucas-Conrad optical flow operator. , The first , Frame detector grayscale matrix; The average displacement is obtained by taking the modulus of each pixel and then averaging the values. ， in, For the first Average frame displacement The total number of valid pixels included in the statistics. For the first The pixel in the first The frame's displacement vector; Construct the first-order difference: ， in, For the first Frame and the The difference in average frame displacement. For the first Average frame displacement; Modeling with ARIMA-(2,1,1): ， in, These are the autoregressive coefficients. The moving average coefficient is... Zero-mean random residuals; Adaptive coefficient update using recursive least squares: ， ， ， wherein, is the step gain vector, is the covariance matrix, is the forgetting factor, 0 < < 1, is the regression vector, is the observed quantity, is the updated coefficient vector; Predict the displacement in the next frame: ， in, For the first Predicted value of frame average displacement.

2. The automatic configuration system for X-ray machine examination parameters for pet examination as described in claim 1, characterized in that, The optimization parameters are generated through a partition parameter mapping module, which performs the following operations: Acquire a preview image output from the X-ray generator in low-dose mode; Divide the preview image into grid areas; Calculate the equivalent thickness based on the grayscale values ​​of each region; Each grid cell is independently generated with kV and mA parameter sets based on its equivalent thickness.

3. The automatic configuration system for X-ray machine examination parameters for pet examination as described in claim 2, characterized in that, The equivalent thickness calculation is associated with a species feature database, which stores the mapping relationship of anatomical features of different pet species, including bone density and equivalent aluminum thickness of soft tissue.

4. The automatic configuration system for X-ray machine examination parameters for pet examination as described in claim 3, characterized in that, The species feature database includes a parameter migration interface, which receives new species anatomical feature data input by the user and associates it with the partition parameter mapping module through an interpolation adaptation algorithm.

5. The automatic configuration system for X-ray machine examination parameters for pet examination as described in claim 4, characterized in that, The steps for generating optimization parameters by the partition parameter mapping module include: After acquiring the preview image in low-dose mode, the image is divided into a grid of equal size: ， in, For the first Line number Column grid, For pixel coordinates, For grid boundaries; Calculate the average gray level of the grid: ， in, For the first Average gray level of the grid The number of pixels within the grid. To preview the grayscale values ​​of the image; Calculate the equivalent thickness using the calibrated logarithmic mapping: ， in, For the first Mesh equivalent thickness, in mm. This is a thickness scaling factor, in mm. This is the grayscale normalization constant; Preliminary exposure parameters are generated using a quadratic polynomial: ， ， in, This refers to the local tube voltage, in kV. This refers to the local tube current, measured in mA. The coefficients are polynomials, obtained through offline calibration; If a balance needs to be struck between clarity and dosage, multi-objective optimization can be performed in a localized solution: ， in, For the first Grid image quality index For dosage indicators, It is a positive weight used to adjust the priority of image quality and dosage.

6. The automatic configuration system for X-ray machine examination parameters for pet examination as described in claim 1, characterized in that, It also includes a micro-motion phase analysis module, which extracts frequency domain feature components from the detector's raw data stream and sends an enable signal to the dynamic exposure trigger when the breathing phase is within a preset phase range.

7. The automatic configuration system for X-ray machine examination parameters for pet examination as described in claim 6, characterized in that, The step of extracting frequency domain feature components from the detector's raw data stream in the micro-motion phase analysis module includes: In continuous acquisition mode, at the first Frames obtain the detector's raw grayscale data First, select the preset region of interest. Spatial integration yields a one-dimensional respiratory sequence: ， in, Indicates the first Average grayscale of frame ROI Indicates the total number of pixels within the ROI. Indicates the first Frame pixels grayscale value, These are pixel coordinates; In length of Zero-mean processing is applied within the sliding window to eliminate slow drift. The processing formula is as follows: ， in, For baseline removal instantaneous signal, The number of frames for the sliding window. The relative frame number within the window; Perform short-time Fourier transform within the same window: ， in, For the first Frame center, frequency index The complex spectrum at the location, For discrete frequencies, The frame sampling frequency, For window function weights, For time indexing, The imaginary unit; Calculate the power spectrum: ， in, For power spectral density, For modulo length calculation; In the pet breathing bag Clock speed: ， in, The current dominant respiratory frequency, Lower and upper limits of the breathing zone; in accordance with The digital difference equation for adaptively designing an IIR bandpass filter is written as: ， in, For bandpass output, For the order Molecular coefficient For the molecular order of the IIR filter For the order Denominator coefficient, The order of the denominator These are the indexes for the numerator and denominator terms, respectively. Using the Hilbert transform to generate analytic signals: ， in, To analyze the signal, For discrete Hilbert operators; Determining the instantaneous phase from an analytic signal: ， in, This refers to the instantaneous respiratory phase. Take the imaginary part and the real part respectively. It is the arctangent function in the four quadrants; when At that time, an enable signal is sent to the dynamic exposure trigger, wherein, The boundary of the phase interval that allows for exposure.

8. The automatic configuration system for X-ray machine examination parameters for pet examination as described in claim 1, characterized in that, The system also includes a radiation fuse unit, which performs the following: Receive the cumulative dose value fed back by the detector in real time; When the cumulative dose exceeds the dynamic threshold generated based on the pet's weight, an exposure termination command is output to the X-ray generator.

Images