Medical rotary CBCT (cone beam computed tomography) scanning imaging system

Abstract

The invention discloses a medical rotary type CBCT scanning imaging system, and relates to the technical field of CBCT scanning diagnos.The medical rotary type CBCT scanning imaging system is characterized in that breathing modes and CBCT scanning areas of a patient are preset, the set CBCT scanning areas are collected through a CBCT scanner, CBCT signal data collected in the breathing modes are processed in a signal processing mode after collection is completed, and the CBCT signal data are obtained; meanwhile, on the basis of the CBCT signal data processed in each breathing mode, the CBCT signal data processed in each breathing mode is reconstructed through a reconstruction algorithm, after reconstruction is completed, the CBCT three-dimensional model reconstructed in each breathing mode is optimized through a deep learning mode, correction parameters of the CBCT three-dimensional model in each breathing mode are calculated, and the correction parameters of the CBCT three-dimensional model in each breathing mode are calculated. And finally, the CBCT signal data acquired in real time is corrected based on the correction parameters of the CBCT three-dimensional model in each breathing mode, so that the accuracy of CBCT scanning imaging is improved.

Description

Technical Field

[0001] This invention relates to the field of CBCT scanning diagnostic technology, specifically to a medical rotating CBCT scanning imaging system. Background Technology

[0002] While CT technology brings convenience, it also poses threats. Because CT scans use X-rays, patients inevitably receive radiation doses. Although the radiation dose from a single scan is within safe limits, multiple scans can lead to cumulative doses, increasing radiation risks, especially for vulnerable groups such as pregnant women and infants. Furthermore, the public health risks caused by radiation exposure are receiving increasing attention. Therefore, improving image quality while reducing the radiation dose received by the human body is currently a major challenge for CT technology.

[0003] Existing technology, such as the invention patent application with publication number CN116849689A, discloses an X-ray beam baffle and a low-dose CBCT imaging method. The method includes: First, extracting a portion of data from the training set to ensure the remaining CBCT images are equal in number to the registered CT images, thus avoiding imbalanced training data. Second, pre-training a Transformer using the extracted data in a self-supervised manner to avoid training difficulties caused by adding a randomly initialized Transformer to a CycleGAN. Finally, adding the pre-trained Transformer to a CycleGAN to obtain a hybrid network, and performing unsupervised training using the remaining data, combining the advantages of both Transformer and convolution to improve the network's ability to enhance CBCT image quality.

[0004] As can be seen from the above solutions, most current CBCT scanning imaging focuses on improving CBCT image quality, neglecting intelligent correction of CBCT images, which has certain limitations. Summary of the Invention

[0005] The purpose of this invention is to provide a medical rotating CBCT scanning imaging system that solves the problems existing in the prior art.

[0006] To solve the above-mentioned technical problems, the present invention adopts the following technical solution: The present invention provides a medical rotating CBCT scanning imaging system. S1. Preset the patient's breathing mode and CBCT scanning area, and collect the set CBCT scanning area through the CBCT scanner to obtain CBCT signal data after collection under each breathing mode; S2. Process the CBCT signal data acquired under each respiratory mode using signal processing methods to obtain the processed CBCT signal data under each respiratory mode. S3. Based on the CBCT signal data processed under each breathing mode, the CBCT signal data processed under each breathing mode is reconstructed using a reconstruction algorithm to obtain the reconstructed CBCT three-dimensional model under each breathing mode. S4. Optimize the reconstructed CBCT 3D model under each breathing mode using deep learning, and calculate the correction parameters of the CBCT 3D model under each breathing mode. S5. Real-time acquisition of the patient's breathing mode and the corresponding CBCT 3D model, and correction of the real-time acquired CBCT 3D model based on the correction parameters of the CBCT 3D model in each breathing mode.

[0007] Preferably, the preset patient breathing mode and CBCT scanning area, and the acquisition of CBCT signal data in each breathing mode by using a CBCT scanner to obtain the preset CBCT scanning area, include the following steps: Setting a breathing pattern includes: respiratory rate and amplitude; The respiratory rate and amplitude of various patients are dynamically collected, and the respiratory rate and amplitude are divided into four levels on average based on the respiratory rate and amplitude. The amplitude is set to four levels: 5mm, 10mm, 20mm, and 30mm; the respiratory rate is set to four levels: 10 breaths / min, 12 breaths / min, 15 breaths / min, and 20 breaths / min. The CBCT scanner parameters are set based on the set respiratory rate and amplitude, and the set CBCT scanning area is acquired based on the set CBCT scanner parameters. The acquired CBCT signal data for each respiratory mode are then summarized.

[0008] Preferably, the step of processing the CBCT signal data acquired under each respiratory mode using image processing methods to obtain the processed CBCT signal data under each respiratory mode includes the following steps: S21, Through Filters and The filter filters the CBCT signal data acquired under each breathing mode to obtain filtered CBCT signal data. right Filters and The filtered CBCT signal data is fitted to obtain the filtered CBCT signal data. S22. Calculate the noise of the filtered CBCT signal data, and determine whether the filtered CBCT signal data is qualified based on the calculation results; S23. Summarize the qualified filtered CBCT signal data to obtain the processed CBCT signal data for each breathing mode.

[0009] Preferably, the step of calculating the noise of the filtered CBCT signal data and determining whether the filtered CBCT signal data is qualified based on the calculation results includes the following steps: Based on the filtered CBCT signal data for each band, excluding the DC band, and calculating the total power of the entire peak band based on the band with the maximum value (max),... ; Based on the filtered CBCT signal data for each band, the DC band, peak band, and harmonic band are removed. Then, the median of the remaining portion is calculated, the removed portions are assigned values, and finally, the sum of the entire power spectrum is calculated. ; The signal-to-noise ratio is set as a standard for noise evaluation. The signal-to-noise ratio formula is as follows: ; in, This indicates the signal-to-noise ratio (SNR). The higher the SNR, the stronger the noise suppression effect. Set a signal-to-noise ratio (SNR) threshold. If the SNR calculated from the filtered CBCT signal data is less than the set SNR threshold, the filtering result is unqualified. Filter again. If the SNR calculated from the filtered CBCT signal data is greater than or equal to the set SNR threshold, the filtering result is qualified.

[0010] Preferably, the step of reconstructing the CBCT signal data processed under each respiratory mode using a reconstruction algorithm to obtain the reconstructed CBCT three-dimensional model under each respiratory mode includes the following steps: S31. Initialize the processed CBCT signal data for each respiratory mode to obtain the initialized CBCT signal data for each respiratory mode. S32. Set the projection plane and project the initialized CBCT signal data corresponding to each breathing mode onto the corresponding plane to obtain the projected CBCT signal data. At the same time, record the time frame corresponding to the projected CBCT signal data. S33. Measure and compare the projected CBCT signal data of adjacent time frames, and calculate the residuals of all projected data; The projected CBCT signal data contours were obtained by measuring the projected CBCT signal data from adjacent time frames. Projected data is calculated using a similarity comparison function. The degree of overlap of the residuals; ; in, This represents the residual of the CBCT signal data after projection of adjacent frames. These represent the CBCT signal data after projection of the i-th frame and the (i-1)-th frame, respectively. For two contours The area of ​​the overlapping portion. and These are the areas of the corresponding contours; S34. Weight the residuals of all projection data and iteratively update the projection image; S35. After the update is completed, when the obtained projection images converge, the iteration stops, and the reconstructed CBCT three-dimensional models under each breathing mode are obtained. The formula for updating the projected image is as follows: ; in, This indicates the projected image to be updated. Indicates the number of iterations. This represents the relaxation factor, used to control the convergence rate. This represents the parameter matrix of the reconstruction algorithm. For projection image, This indicates the matrix transpose.

[0011] Preferably, the optimization of the reconstructed CBCT 3D model under each breathing mode using deep learning, and the calculation of the correction parameters of the CBCT 3D model under each breathing mode, includes the following steps: S41. Acquire a 3D model of the CBCT scan area in a static state, and set multiple reference points in the 3D model of the CBCT in a static state; S42. Set up a three-dimensional coordinate system and determine the coordinates of all reference points in the CBCT three-dimensional model in the static state and the reconstructed CBCT three-dimensional model. S43. Compare the differences in the coordinates of the corresponding reference points in the CBCT 3D model under static conditions and the reconstructed CBCT 3D model, and construct a corrected model based on the comparison results; S44. Based on the constructed correction model, the reconstructed CBCT 3D model under each breathing mode is optimized using deep learning to determine the correction parameters of the CBCT 3D model under each breathing mode.

[0012] Preferably, the step of comparing the differences in the coordinates of corresponding reference points between the CBCT 3D model in a static state and the reconstructed CBCT 3D model, and constructing a corrected model based on the comparison results, includes the following steps: The coordinates of a reference point in a static CBCT 3D model are set as the standard coordinates, and the coordinates of the corresponding reference point in the reconstructed CBCT 3D model are set as the real-time coordinates. The difference between the standard coordinates and the real-time coordinates is calculated, and a set of data is obtained. ,in, These are the real-time coordinates of the corresponding reference point. These are the standard coordinates of the corresponding reference point. To determine the difference between standard coordinates and real-time coordinates, we collect comparison results from all reference points and create a dataset for all reference points. A modified model is constructed using polynomial fitting. The corrected model formula is shown below: ; ; in, This represents the horizontal component offset value between the standard coordinates of the corresponding reference point and the real-time coordinates of the corresponding reference point. This represents the vertical component offset value from the standard coordinates of the corresponding reference point to the real-time coordinates of the corresponding reference point. This represents the horizontal component of the real-time coordinates of the corresponding reference point. This represents the vertical component of the real-time coordinates of the corresponding reference point. This represents the difference between standard coordinates and real-time coordinates; , Indicates two groups Secondary Lagrange basis functions.

[0013] Preferably, the modified model based on the constructed model optimizes the reconstructed CBCT 3D model under each breathing mode using deep learning, and the determination of the modification parameters of the CBCT 3D model under each breathing mode includes the following steps: S441. Summarize the population of CBCT 3D models reconstructed under each respiratory mode; Initialize the constructed population, define each particle as a set of reconstructed CBCT 3D models of the detection point, set the population size, and set the maximum number of iterations. Random position of particles Particle velocity and inertia factor ; S442. Set the constructed modified model as a fitness function, and calculate the fitness of each particle based on the fitness function; The formula for calculating the fitness of each particle is as follows: ; in, Indicates particle fitness; S443, updating the optimal position of a single particle; The formula for updating the velocity of a single particle is as follows: ; in, Represents particles In the Speed ​​during each iteration Represents particles In the Speed ​​during each iteration Represents particles In the Position in the next iteration. , Represents the acceleration constant. , Represents a random number within the interval [0,1]. Represents particles Individual extreme values, Represents the global extremum of all particles; The formula for updating the position of a single particle is as follows: ; in, Represents particles In the Position in the next iteration; For each particle being computed, its fitness at its current position is compared with the best position it has passed. The fitness of the current position is compared with that of the best position it has visited. Based on the fitness level, the current position is taken as the current optimal position. If the fitness of the current position is less than or equal to the best position it has visited, The fitness of the position does not change the current optimal position. ; S444, Update the group's optimal position; For each particle being computed, its fitness at its current position is compared to the best position passed by any particle in the population. The fitness of the current position is compared with that of the best position traversed by particles in the population. Based on the fitness level, the current position is taken as the current optimal position. If the fitness of the current position is less than or equal to the best position passed by a particle in its population, The fitness of the position does not change the current optimal position. ; S45. Update the inertia factor. Based on the updated inertia factor, update the position and velocity of all particles. The formula for updating the inertia factor is as follows: ; in, The inertia factor represents the initial inertia factor at the start of the iteration. This represents the inertia factor at the final iteration. Indicates the current iteration number. Indicates the maximum number of iterations; S446. Repeat steps S442-S445 until the maximum number of iterations is reached, and output the CBCT 3D model correction parameters corresponding to the optimal position under each breathing mode.

[0014] Preferably, the real-time acquisition of the patient's breathing pattern and the corresponding CBCT three-dimensional model under each breathing pattern, and the correction of the real-time acquired CBCT three-dimensional model based on the correction parameters of the CBCT three-dimensional model under each breathing pattern, includes the following steps: The system acquires the patient's breathing pattern and the corresponding CBCT 3D model in real time, and matches the patient's breathing pattern with the set gear. After the matching is completed, the CBCT scanner acquires the CBCT 3D model of the corresponding area of ​​the patient in real time. After the acquisition is completed, the CBCT 3D model of the corresponding area of ​​the patient is corrected based on the correction parameters of the CBCT 3D model in each breathing mode.

[0015] The present invention also provides a medical rotating CBCT scanning imaging system, comprising: a data acquisition module, a data processing module, a data reconstruction module, a model correction module, and a correction and adjustment module; The data acquisition module is used to preset the CBCT scanner and acquire CBCT signal data after acquisition under each breathing mode; The data processing module is used to process the acquired CBCT signal data; The data reconstruction module is used to reconstruct the CBCT signal data processed under each breathing mode using a reconstruction algorithm, so as to obtain the reconstructed CBCT three-dimensional model under each breathing mode. The model correction module is used to optimize the reconstructed CBCT 3D model under each breathing mode through deep learning, and to calculate the correction parameters of the CBCT 3D model under each breathing mode. The correction and adjustment module is used to correct the CBCT 3D model acquired in real time.

[0016] The beneficial effects of this invention are as follows: (1) This invention presets the patient's breathing mode and CBCT scanning area, and collects the set CBCT scanning area through a CBCT scanner. After the collection is completed, the CBCT signal data collected under each breathing mode is processed by signal processing. At the same time, based on the processed CBCT signal data under each breathing mode, the CBCT signal data processed under each breathing mode is reconstructed by a reconstruction algorithm. After the reconstruction is completed, the reconstructed CBCT three-dimensional model under each breathing mode is optimized by deep learning. The correction parameters of the CBCT three-dimensional model under each breathing mode are calculated. Finally, the real-time collected CBCT signal data is corrected based on the correction parameters of the CBCT three-dimensional model under each breathing mode, thereby improving the accuracy of CBCT scanning imaging.

[0017] (2) The present invention is achieved through Filters and The filter filters the CBCT signal data acquired under each breathing mode. At the same time, it calculates the noise of the filtered CBCT signal data in real time during the filtering process, and judges whether the filtered CBCT signal data is qualified based on the calculated noise, thereby improving the reliability of CBCT signal data processing.

[0018] (3) This invention initializes the processed CBCT signal data under each breathing mode, sets the projection plane, and measures and compares the projected CBCT signal data of adjacent time frames to calculate the residuals of all projection data. At the same time, the residuals of all projection data are weighted and the projection image is iteratively updated. After the update is completed, when the obtained projection image converges, the iteration stops and the reconstructed CBCT three-dimensional model under each breathing mode is obtained, which improves the accuracy and intelligence of CBCT three-dimensional model reconstruction. Attached Figure Description

[0019] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. 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.

[0020] Figure 1 This is a schematic diagram of the process of a medical rotating CBCT scanning imaging system according to the present invention. Detailed Implementation

[0021] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0022] In a specific embodiment of the present invention, Reference Figure 1 As shown, the present invention provides a medical rotating CBCT scanning imaging system, comprising the following steps: S1. Preset the patient's breathing mode and CBCT scanning area, and collect the set CBCT scanning area through the CBCT scanner to obtain CBCT signal data after collection under each breathing mode; S2. Process the CBCT signal data acquired under each respiratory mode using signal processing methods to obtain the processed CBCT signal data under each respiratory mode. S3. Based on the CBCT signal data processed under each breathing mode, the CBCT signal data processed under each breathing mode is reconstructed using a reconstruction algorithm to obtain the reconstructed CBCT three-dimensional model under each breathing mode. S4. Optimize the reconstructed CBCT 3D model under each breathing mode using deep learning, and calculate the correction parameters of the CBCT 3D model under each breathing mode. S5. Real-time acquisition of the patient's breathing mode and the corresponding CBCT 3D model under the breathing mode, and correction of the real-time acquired CBCT 3D model based on the correction parameters of the CBCT 3D model under each breathing mode. Furthermore, referring to Figure 1 As shown, the patient's breathing mode and CBCT scanning area are preset, and the CBCT scanner is used to acquire CBCT signal data in the preset CBCT scanning area. The process includes the following steps: Setting a breathing pattern includes: respiratory rate and amplitude; The respiratory rate and amplitude of various patients are dynamically collected, and the respiratory rate and amplitude are divided into four levels on average based on the respiratory rate and amplitude. The amplitude is set to four levels: 5mm, 10mm, 20mm, and 30mm; the respiratory rate is set to four levels: 10 breaths / min, 12 breaths / min, 15 breaths / min, and 20 breaths / min. Furthermore, the CBCT scanner parameters are set based on the set respiratory rate and amplitude, and the set CBCT scanning area is acquired based on the set CBCT scanner parameters. The acquired CBCT signal data under each respiratory mode are then summarized. Furthermore, referring to Figure 1 As shown, the CBCT signal data acquired under each respiratory mode are processed using image processing methods to obtain the processed CBCT signal data under each respiratory mode, including the following steps: S21, Through Filters and The filter filters the CBCT signal data acquired under each breathing mode to obtain filtered CBCT signal data. The filtering function is shown below: ; The filtering function is shown below: ; in, This represents CBCT signal data acquired under various respiratory modes. Indicates passage After filter processing, the filtered CBCT signal data is obtained. Indicates passage After filter processing, the filtered CBCT signal data is obtained. Indicates bandwidth. Represents a singular time function. Represents a time-domain function; Furthermore, on Filters and The filtered CBCT signal data is fitted to obtain the filtered CBCT signal data. S22. Calculate the noise of the filtered CBCT signal data, and determine whether the filtered CBCT signal data is qualified based on the calculation results; Based on the filtered CBCT signal data for each band, excluding the DC band, and calculating the total power of the entire peak band based on the band with the maximum value (max),... ; Furthermore, based on each band of the filtered CBCT signal data, the DC band, signal peak band, and harmonic band are removed (set to zero). Then, the median of the remaining portion is calculated, some values ​​are assigned, and finally, the sum of the entire power spectrum is calculated. ; Furthermore, the signal-to-noise ratio is set as a noise evaluation index standard; The signal-to-noise ratio formula is as follows: ; in, This indicates the signal-to-noise ratio (SNR). The higher the SNR, the stronger the noise suppression effect. Set a signal-to-noise ratio (SNR) threshold. If the SNR calculated from the filtered CBCT signal data is less than the set SNR threshold, the filtering result is unqualified. Filter again. If the SNR calculated from the filtered CBCT signal data is greater than or equal to the set SNR threshold, the filtering result is qualified. S23. Summarize the qualified filtered CBCT signal data to obtain the processed CBCT signal data for each respiratory mode. Furthermore, referring to Figure 1 As shown, based on the processed CBCT signal data under each respiratory mode, the CBCT signal data under each respiratory mode is reconstructed using a reconstruction algorithm to obtain the reconstructed CBCT 3D model under each respiratory mode, including the following steps: S31. Initialize the processed CBCT signal data for each respiratory mode to obtain the initialized CBCT signal data for each respiratory mode. S32. Set the projection plane and project the initialized CBCT signal data corresponding to each breathing mode onto the corresponding plane to obtain the projected CBCT signal data. At the same time, record the time frame corresponding to the projected CBCT signal data. S33. Measure and compare the projected CBCT signal data of adjacent time frames, and calculate the residuals of all projected data; The projected CBCT signal data contours were obtained by measuring the projected CBCT signal data from adjacent time frames. Projected data is calculated using a similarity comparison function. The degree of overlap of the residuals; ; in, This represents the residual of the CBCT signal data after projection of adjacent frames. These represent the CBCT signal data after projection of the i-th frame and the (i-1)-th frame, respectively. For two contours The area of ​​the overlapping portion. and These are the areas of the corresponding contours; S34. Weight the residuals of all projection data and iteratively update the projection image; S35. After the update is completed, when the obtained projection images converge, the iteration stops, and the reconstructed CBCT three-dimensional models under each breathing mode are obtained. The formula for updating the projected image is as follows: ; in, This indicates the projected image to be updated. Indicates the number of iterations. This represents the relaxation factor, used to control the convergence rate. This represents the parameter matrix of the reconstruction algorithm. For projection image, Indicates matrix transpose; Furthermore, referring to Figure 1 As shown, the reconstructed CBCT 3D models under various breathing modes are optimized using deep learning. The calculation of correction parameters for the CBCT 3D models under each breathing mode includes the following steps: S41. Acquire a 3D model of the CBCT scan area in a static state, and set multiple reference points in the 3D model of the CBCT in a static state; S42. Set up a three-dimensional coordinate system and determine the coordinates of all reference points in the CBCT three-dimensional model in the static state and the reconstructed CBCT three-dimensional model. S43. Compare the differences in the coordinates of the corresponding reference points in the CBCT 3D model under static conditions and the reconstructed CBCT 3D model, and construct a corrected model based on the comparison results; The coordinates of a reference point in a static CBCT 3D model are set as the standard coordinates, and the coordinates of the corresponding reference point in the reconstructed CBCT 3D model are set as the real-time coordinates. The difference between the standard coordinates and the real-time coordinates is calculated, and a set of data is obtained. ,in, These are the real-time coordinates of the corresponding reference point. These are the standard coordinates of the corresponding reference point. To determine the difference between standard coordinates and real-time coordinates, we collect comparison results from all reference points and create a dataset for all reference points. Furthermore, a modified model is constructed using a polynomial fitting method; The corrected model formula is shown below: ; ; in, This represents the horizontal component offset value between the standard coordinates of the corresponding reference point and the real-time coordinates of the corresponding reference point. This represents the vertical component offset value from the standard coordinates of the corresponding reference point to the real-time coordinates of the corresponding reference point. This represents the horizontal component of the real-time coordinates of the corresponding reference point. This represents the vertical component of the real-time coordinates of the corresponding reference point. This represents the difference between standard coordinates and real-time coordinates; , Indicates two groups Secondary Lagrange basis functions; S44. Based on the constructed correction model, the reconstructed CBCT 3D model under each breathing mode is optimized by deep learning to determine the correction parameters of the CBCT 3D model under each breathing mode. S441. Summarize the population of CBCT 3D models reconstructed under each respiratory mode; Initialize the constructed population, define each particle as a set of reconstructed CBCT 3D models of the detection point, set the population size, and set the maximum number of iterations. Random position of particles Particle velocity and inertia factor ; S442. Set the constructed modified model as a fitness function, and calculate the fitness of each particle based on the fitness function; The formula for calculating the fitness of each particle is as follows: ; in, Indicates particle fitness; S443, updating the optimal position of a single particle; The formula for updating the velocity of a single particle is as follows: ; in, Represents particles In the Speed ​​during each iteration Represents particles In the Speed ​​during each iteration Represents particles In the Position in the next iteration. , Represents the acceleration constant. , Represents a random number within the interval [0,1]. Represents particles Individual extreme values, Represents the global extremum of all particles; The formula for updating the position of a single particle is as follows: ; in, Represents particles In the Position in the next iteration; For each particle being computed, its fitness at its current position is compared with the best position it has passed. The fitness of the current position is compared with that of the best position it has visited. Based on the fitness level, the current position is taken as the current optimal position. If the fitness of the current position is less than or equal to the best position it has visited, The fitness of the position does not change the current optimal position. ; S444, Update the group's optimal position; For each particle being computed, its fitness at its current position is compared to the best position passed by any particle in the population. The fitness of the current position is compared with that of the best position traversed by particles in the population. Based on the fitness level, the current position is taken as the current optimal position. If the fitness of the current position is less than or equal to the best position passed by a particle in its population, The fitness of the position does not change the current optimal position. ; S45. Update the inertia factor. Based on the updated inertia factor, update the position and velocity of all particles. The formula for updating the inertia factor is as follows: ; in, The inertia factor represents the initial inertia factor at the start of the iteration. This represents the inertia factor at the final iteration. Indicates the current iteration number. Indicates the maximum number of iterations; S446. Repeat steps S442-S445 until the maximum number of iterations is reached, and output the CBCT 3D model correction parameters corresponding to each breathing mode at the optimal position. Furthermore, referring to Figure 1 As shown, the real-time acquisition of the patient's respiratory mode and the corresponding CBCT 3D model under each respiratory mode, and the correction of the real-time acquired CBCT 3D model based on the correction parameters of the CBCT 3D model under each respiratory mode, includes the following steps: The system acquires the patient's breathing pattern and the corresponding CBCT 3D model in real time, and matches the patient's breathing pattern with the set gear. After the matching is completed, the CBCT scanner acquires the CBCT 3D model of the corresponding area of ​​the patient in real time. After the acquisition is completed, the CBCT three-dimensional model of the corresponding area of ​​the real-time acquired patient is corrected based on the correction parameters of the CBCT three-dimensional model in each breathing mode. In one specific embodiment, the medical rotating CBCT scanning imaging system further includes: a data acquisition module, a data processing module, a data reconstruction module, a model correction module, and a correction and adjustment module; The data acquisition module is used to preset the CBCT scanner and acquire CBCT signal data after acquisition under each breathing mode; The data processing module is used to process the acquired CBCT signal data; The data reconstruction module is used to reconstruct the CBCT signal data processed under each breathing mode using a reconstruction algorithm, so as to obtain the reconstructed CBCT three-dimensional model under each breathing mode. The model correction module is used to optimize the reconstructed CBCT 3D model under each breathing mode through deep learning, and to calculate the correction parameters of the CBCT 3D model under each breathing mode. The correction and adjustment module is used to correct the CBCT 3D model acquired in real time.

[0023] It should be noted that, The above content is merely an example and illustration of the concept of the present invention. Those skilled in the art can make various modifications or additions to the specific embodiments described, or use similar methods to replace them, as long as they do not deviate from the concept of the invention or exceed the scope defined by the present invention, and all such modifications and additions should fall within the protection scope of the present invention.

Claims

1. A medical rotational CBCT scanning imaging system, characterized in that, The method comprises the following steps: S1, presetting a patient breathing mode and a CBCT scanning area, and collecting the set CBCT scanning area through a CBCT scanner to obtain CBCT signal data collected under each breathing mode; S2, processing the CBCT signal data collected under each breathing mode through a signal processing mode to obtain CBCT signal data processed under each breathing mode; S3, reconstructing the CBCT signal data processed under each breathing mode through a reconstruction algorithm based on the CBCT signal data processed under each breathing mode to obtain a CBCT three-dimensional model reconstructed under each breathing mode; S4, optimizing the CBCT three-dimensional model reconstructed under each breathing mode through a deep learning mode, and calculating a correction parameter of the CBCT three-dimensional model under each breathing mode; S5, collecting a CBCT three-dimensional model under a corresponding breathing mode in real time based on the patient breathing mode, and correcting the CBCT three-dimensional model collected in real time based on the correction parameter of the CBCT three-dimensional model under each breathing mode.

2. A medical rotational CBCT scanning imaging system according to claim 1, wherein, The presetting of the patient breathing mode and the CBCT scanning area, and the collection of the set CBCT scanning area through the CBCT scanner to obtain the CBCT signal data collected under each breathing mode comprises the following steps: The setting of the breathing mode comprises a breathing frequency and an amplitude; The breathing frequency and the amplitude of various patients are dynamically collected, and the breathing frequency and the amplitude are averagely divided into four ranges based on the breathing frequency and the amplitude; The amplitude is set to four ranges of 5mm, 10mm, 20mm and 30mm, and the breathing frequency is set to four ranges of 10 times / min, 12 times / min, 15 times / min and 20 times / min; The CBCT scanner parameters are set based on the set breathing frequency and amplitude, the set CBCT scanning area is collected based on the set CBCT scanner parameters, and the CBCT signal data collected under each breathing mode is obtained by summarizing.

3. A medical rotational CBCT scanning imaging system according to claim 1, wherein, The processing of the CBCT signal data collected under each breathing mode through the image processing mode to obtain the CBCT signal data processed under each breathing mode comprises the following steps: S21, by filter and The filter filters the CBCT signal data collected under each breathing pattern to obtain filtered CBCT signal data. To The filter and The CBCT signal data after filter processing is fitted to obtain filtered CBCT signal data; S22, calculating the noise of the filtered CBCT signal data, and judging whether the filtered CBCT signal data is qualified based on the calculation result; S23, summarizing the qualified filtered CBCT signal data to obtain the CBCT signal data processed under each breathing mode.

4. A medical rotational CBCT scanning imaging system according to claim 3, wherein, The calculation of the noise of the filtered CBCT signal data, and the judgment of whether the filtered CBCT signal data is qualified based on the calculation result comprises the following steps: Based on each wave band of the filtered CBCT signal data, remove the direct current wave band, and calculate the sum power of the entire peak value wave band according to the maximum value max ; Based on the filtered CBCT signal data of each wave band, the direct current wave band is removed, the signal peak value section is removed, the harmonic section of the signal is removed, then the median of the remaining part is calculated, the removed part is assigned, and finally the sum of the entire power spectrum is calculated ; The signal-to-noise ratio is set as a noise evaluation index standard; The signal-to-noise ratio formula is as follows: ； wherein, represents the signal-to-noise ratio, the greater the set signal-to-noise ratio, the stronger the noise suppression effect; The signal-to-noise ratio threshold is set, when the signal-to-noise ratio calculated from the filtered CBCT signal data is less than the set signal-to-noise ratio threshold, it indicates that the filtering result is unqualified, and filtering is performed again, when the signal-to-noise ratio calculated from the filtered CBCT signal data is greater than or equal to the set signal-to-noise ratio threshold, it indicates that the filtering result is qualified.

5. The medical rotational CBCT scanning and imaging system of claim 1, wherein, The CBCT signal data processed under each breathing mode is reconstructed by a reconstruction algorithm to obtain a CBCT three-dimensional model after reconstruction under each breathing mode, including the following steps: S31, initializing the CBCT signal data processed under each breathing mode to obtain the corresponding initialized CBCT signal data under each breathing mode; S32, setting a projection plane and projecting the corresponding initialized CBCT signal data under each breathing mode to the corresponding plane to obtain the projected CBCT signal data, and recording the corresponding time frame of the projected CBCT signal data; S33, measuring and comparing the projected CBCT signal data of adjacent time frames to calculate the residual of all projection data; The measured post-projection CBCT signal data of the adjacent time frames are post-projection CBCT signal data profiles , and the degree of coincidence of the residual errors of the projection data is calculated by a similarity comparison function. ； wherein, represents a residual of the CBCT signal data after projection of adjacent frames, respectively represent the CBCT signal data after projection of the i-th frame and the i-1-th frame, are the areas of the two profiles of the overlapping portion, and are the areas of the corresponding profiles, respectively. S34, weighting the residual of all projection data and iteratively updating the projection image; S35, after updating, when the obtained projection image converges, the iteration stops, and the CBCT three-dimensional model after reconstruction under each breathing mode is obtained by summarizing; The formula for updating the projection image is as follows: ； wherein, denotes the projection image to be updated, denotes the number of iterations, denotes a relaxation factor for controlling the convergence speed, denotes a parameter matrix of the reconstruction algorithm, is a projection image, denotes the matrix transpose.

6. The medical rotational CBCT scanning and imaging system of claim 1, wherein, The CBCT three-dimensional model after reconstruction under each breathing mode is optimized by deep learning, and the correction parameters of the CBCT three-dimensional model under each breathing mode are calculated, including the following steps: S41, acquiring a CBCT three-dimensional model in a stationary state in a CBCT scanning area, and setting multiple reference points in the CBCT three-dimensional model in the stationary state; S42, setting a three-dimensional coordinate system to determine the coordinates of all reference points in the CBCT three-dimensional model in the stationary state and the CBCT three-dimensional model after reconstruction; S43, comparing the difference values of the corresponding reference point coordinates in the CBCT three-dimensional model in the stationary state and the CBCT three-dimensional model after reconstruction, and constructing a correction model based on the comparison result; S44, based on the constructed correction model, the CBCT three-dimensional model after reconstruction under each breathing mode is optimized by deep learning to determine the correction parameters of the CBCT three-dimensional model under each breathing mode.

7. A medical rotational CBCT scanning imaging system according to claim 6, wherein, The difference values of the corresponding reference point coordinates in the CBCT three-dimensional model in the stationary state and the CBCT three-dimensional model after reconstruction are compared, and a correction model is constructed based on the comparison result, including the following steps: Set the coordinate of the reference point in the CBCT three-dimensional model in the static state as the standard coordinate, the coordinate of the corresponding reference point in the CBCT three-dimensional model after reconstruction as the real-time coordinate, calculate the difference value of the standard coordinate and the real-time coordinate, and collect a group of data wherein, the real-time coordinate of the corresponding reference point, the standard coordinate of the corresponding reference point, the difference value of the standard coordinate and the real-time coordinate, collect the comparison results of all reference points, and establish the data set of all reference points; A correction model is constructed by polynomial fitting; The formula of the correction model is as follows: ； ； wherein, represents a horizontal component offset value of the corresponding reference point standard coordinate to the corresponding reference point real-time coordinate, represents a vertical component offset value of the corresponding reference point standard coordinate to the corresponding reference point real-time coordinate, represents a horizontal component of the corresponding reference point real-time coordinate, represents a vertical component of the corresponding reference point real-time coordinate, represents a difference value of the standard coordinate and the real-time coordinate; , represents two groups Lagrange base functions.

8. A medical rotational CBCT scanning imaging system according to claim 6, wherein, The CBCT three-dimensional model after reconstruction under each breathing mode is optimized by deep learning based on the constructed correction model to determine the correction parameters of the CBCT three-dimensional model under each breathing mode, including the following steps: S441, summarizing the CBCT three-dimensional model after reconstruction under each breathing mode to construct a population; Initialize the constructed population, set each particle to represent a set of reconstructed CBCT three-dimensional models of detection points, set the population size, the maximum number of iterations , random position of particles , particle velocity and inertia factor ; S442, setting the constructed correction model as a fitness function, and calculating the fitness of each particle based on the fitness function; The formula for calculating the fitness of each particle is as follows: ； wherein, represents the fitness of the particle; S443, updating the position of a single particle; The velocity update formula of a single particle is as follows: ； wherein representing the particles the velocity of the particles in the first iteration process, representing the particles the velocity of the particles in the first iteration process, representing the particles the position of the particles in the first iteration process, , representing the acceleration constant, , representing a random number in the interval [0,1], representing the individual extreme values of the particles , representing the global extreme values of the particles The position update formula of a single particle is as follows: ； wherein representing particles position in the first iteration process; For each particle of the calculation, the fitness of its current position is compared to the fitness of its passed best position If the fitness of the current position is greater than the fitness of its passed best position The current position is taken as the current best position If the fitness of the current position is less than or equal to the fitness of its passed best position The current best position is not changed ; S444, updating the best position of the population; For each particle of the calculation, the fitness of its current position is compared to the best position passed by the particle of its population , if the fitness of the current position is greater than the fitness of the best position passed by the particle of its population , the current position is taken as the current best position , if the fitness of the current position is less than or equal to the fitness of the best position passed by the particle of its population , the current best position is not changed ; S45, updating the inertia factor, and updating the position and velocity of all particles based on the updated inertia factor; The inertia factor update formula is as follows: ； wherein, represents the inertia factor at the beginning of the iteration, represents the inertia factor at the end of the iteration, represents the current iteration number, represents the maximum iteration number; S446, repeating steps S442-S445 until the maximum number of iterations is reached, and outputting the correction parameters of the CBCT three-dimensional model in each breathing pattern corresponding to the optimal position.

9. The medical rotational CBCT scanning and imaging system of claim 1, wherein, The real-time acquisition of the patient's breathing pattern and the CBCT three-dimensional model in the corresponding breathing pattern, and the correction of the real-time acquired CBCT three-dimensional model based on the correction parameters of the CBCT three-dimensional model in each breathing pattern include the following steps: Real-time acquisition of the patient's breathing pattern and the CBCT three-dimensional model in the corresponding breathing pattern, and matching the patient's breathing pattern with the set gear. After the matching is completed, the CBCT three-dimensional model of the corresponding region of the patient is acquired in real time by the CBCT scanner; After the acquisition is completed, the CBCT three-dimensional model of the corresponding region of the patient is corrected based on the correction parameters of the CBCT three-dimensional model in each breathing pattern.

10. A medical rotational CBCT scanning imaging system as claimed in claim 1, wherein, It comprises a data acquisition module, a data processing module, a data reconstruction module, a model correction module, and a correction adjustment module. The data acquisition module is used to preset the CBCT scanner and acquire the CBCT signal data after acquisition in each breathing pattern. The data processing module is used to process the acquired CBCT signal data. The data reconstruction module is used to reconstruct the processed CBCT signal data in each breathing pattern by a reconstruction algorithm to obtain the reconstructed CBCT three-dimensional model in each breathing pattern. The model correction module is used to optimize the reconstructed CBCT three-dimensional model in each breathing pattern by deep learning to calculate the correction parameters of the CBCT three-dimensional model in each breathing pattern. The correction adjustment module is used to correct the real-time acquired CBCT three-dimensional model.

Images