Transcranial Doppler flexible ultrasonic probe detection data management system and method

Abstract

The invention discloses a transcranial Doppler flexible ultrasonic probe detection data management system and method, and relates to the technical field of ultrasonic probe design, and the method comprises the steps: obtaining and preprocessing head image data of a patient, and extracting structural features of a temporal window part from the preprocessed data; constructing a three-dimensional grid model containing a blood vessel structure according to the extracted structural features, and analyzing an angle relation between an ultrasonic probe and a target blood vessel in the model and signal quality performance of the ultrasonic probe; collecting head image data, blood flow measurement data and probe angle data of a plurality of patients to construct a piezoelectric ceramic angle prediction model, and predicting the optimal angle of each piezoelectric ceramic in the ultrasonic probe under different patient head structures and blood flow conditions; the surface of the ultrasonic probe is covered with a flexible material, a closed-loop control system is introduced into the probe, the voltage of piezoelectric ceramics is adjusted according to a model prediction result, and angle control over the probe is achieved.

Description

Technical Field

[0001] The present invention relates to the technical field of ultrasonic probe design, and particularly to a transcranial Doppler flexible ultrasonic probe detection data management system and method. Background Art

[0002] As a non-invasive cerebrovascular detection technology, transcranial Doppler (TCD) is widely used in clinical practice to evaluate the hemodynamic status of cerebral arteries. This technology measures the blood flow velocity and blood flow parameters of arteries at the base of the brain by penetrating the skull with ultrasonic waves, providing an important basis for the diagnosis of cerebrovascular diseases. However, during the traditional TCD detection process, the incident angle of the probe is crucial for accurately obtaining the cerebral artery blood flow signal. Due to the differences in individual head structures and temporal window positions, doctors usually need to manually adjust the angle of the probe to find the optimal ultrasonic incident path, which not only increases the complexity of the operation but also may lead to subjectivity and instability of the detection results.

[0003] In actual operation, doctors need to make multiple attempts and adjustments based on the patient's head shape and the specific position of the temporal window, relying on experience and feel, to find the appropriate incident angle of the probe. This process is not only time-consuming and laborious but also may affect the accuracy of the detection results due to differences in doctors' skill levels. In addition, due to the diverse head structures and temporal window morphologies of different patients, traditional fixed-angle or standardized probe designs often fail to meet the detection needs of all patients, limiting the wide application and accuracy improvement of TCD technology. Summary of the Invention

[0004] The purpose of the present invention is to provide a transcranial Doppler flexible ultrasonic probe detection data management system and method to solve the problems raised in the prior art.

[0005] To achieve the above purpose, the present invention provides the following technical solution: A transcranial Doppler flexible ultrasonic probe detection data management method, the method comprising the following steps:

[0006] Step S100: Obtain the head image data of the patient and perform preprocessing, and extract the structural features of the temporal window part from the preprocessed data;

[0007] Step S200: Construct a three-dimensional grid model containing vascular structures according to the extracted structural features of the temporal window part, and analyze the incident angle of the ultrasonic probe relative to the target blood vessel in the model and evaluate the signal quality performance of the ultrasonic probe;

[0008] Step S300: Collect the head image data, blood flow measurement data, and probe angle data of multiple patients to construct a piezoelectric ceramic angle prediction model for predicting the optimal angle of each piezoelectric ceramic in the ultrasonic probe under different patient head structures and blood flow conditions;

[0009] Step S400: Cover the surface of the ultrasonic probe with a flexible material, and introduce a closed-loop control system inside the probe to adjust the voltage of the piezoelectric ceramic according to the result predicted by the model, thereby controlling the angle of the probe.

[0010] Furthermore, the step S100 includes:

[0011] Step S101: Scan the head through a medical imaging device to obtain head image data including the temporal window part;

[0012] Step S102: Preprocess the head image data, including denoising and enhancing the contrast. The preprocessed head image data is represented as a three-dimensional array I(x, y, z), where I represents the gray value, and I(x, y, z) represents the gray value at the coordinate (x, y, z); extract the structural features of the temporal window part from the preprocessed image data. The structural features of the temporal window part include locating the position of the temporal window in the image using the Canny edge detection algorithm, and using the region growing method to extract the contour of the temporal window; using the Hessian matrix filter to enhance the vascular structure in the temporal window part of the image, and adopting the gradient tracking algorithm to track the centerline of the blood vessel by calculating the gradient of the blood vessel, where the gradient represents the direction of the blood vessel, and the curvature is calculated to represent the degree of bending of the blood vessel; separate the vascular structure from the image background through threshold segmentation. According to the contrast between the vascular structure and the background in the enhanced image, set a threshold, determine the area with pixel values greater than the threshold in the image as the vascular structure, and the area less than the threshold as the background. Through the segmented binary image, obtain the position information of the vascular pixels, and calculate the direction vector of the blood vessel at each pixel point through the eigenvector of the Hessian matrix; obtain the radius of the blood vessel by measuring the distance from the blood vessel edge to the centerline; finally, obtain the structural features of the temporal window part including the position and contour of the temporal window in the image and the position, direction vector, centerline and radius of the blood vessels in the temporal window part.

[0013] In the above technical solution, by obtaining the head image data of the patient and preprocessing it, the structural features of the temporal window part can be accurately extracted, including key information such as the position, direction, centerline and radius of the blood vessels. These information provide a solid foundation for subsequent construction of a three-dimensional mesh model, analysis of the incident angle of the ultrasonic probe, and evaluation of the signal quality, thereby improving the accuracy and efficiency of transcranial Doppler ultrasound examination.

[0014] Furthermore, the step S200 includes:

[0015] Step S201: Initialize a three-dimensional mesh model M containing vertices V, edges E, and faces F according to the temporal window contour extracted in step 102. Using the region growing algorithm, start from the initial point P0(x0, y0, z0) of the mesh and expand the mesh M according to the gray values in the image. Set a threshold T. For the neighborhood point P1(x1, y1, z1) of P0, if |I(x1, y1, z1) - I(x0, y0, z0)| < T, then add P1 to the mesh M, where I(x1, y1, z1) represents the gray value at the coordinate (x1, y1, z1) in the image, and I(x0, y0, z0) represents the gray value at the initial point P0. Repeat the above comparison process until no new points satisfy the growth condition. During the mesh expansion process, convert the centerline and radius of the blood vessel extracted in step S102 into point, line, and face data in the mesh. The centerline of the blood vessel is represented by the edge E in the mesh, and the radius of the blood vessel is used to determine the size and shape of the face F in the mesh. The centerline of the blood vessel is composed of the points {P1, P2, …, P n}, where P n represents the nth point. For each point P n , there is a corresponding radius r n . Add all the points to the mesh M and construct the corresponding face F according to the radius r n . Finally, calculate the Laplacian coordinates of each point in the mesh through the Laplacian smoothing algorithm to update the positions of the mesh points, and continuously iterate the smoothing process until the preset smoothing degree is reached to complete the optimization process of the mesh M;

[0016] Step S202: Measure the reflection frequency of ultrasonic waves in the blood through a Doppler ultrasound device, and calculate the blood flow velocity in the blood vessel according to the Doppler effect. According to the formula:

[0017]

[0018] where v represents the blood flow velocity in the blood vessel, f s is the received reflection ultrasonic wave frequency, f0 is the emitted ultrasonic wave frequency, and c is the sound velocity;

[0019] Step S203: Through the Doppler ultrasound device, record the emission direction of the probe ultrasonic wave during manual operation and calculate the corresponding direction vector. By analyzing the angle between the direction vector of the blood vessel and the probe emission direction vector, obtain the incident angle θ of the probe relative to the target blood vessel:

[0020]

[0021] where represents the direction vector of the blood vessel, A x , A y , Az respectively represent the projected lengths of the blood vessel direction on the x-axis, y-axis, and z-axis, represents the probe emission direction vector, B x , B y , B z represent the projected lengths of the ultrasonic waves emitted by the probe on the x-axis, y-axis, and z-axis;

[0022] Step S204: Analyze the effect of the ultrasonic probe signal quality based on the blood flow velocity, pulsatility index, resistance index, and probe angle. According to the formula:

[0023]

[0024] where S represents the signal quality index, V p is the maximum systolic velocity, V d is the maximum diastolic velocity, Y is the pulsatility index, K is the resistance index, and the where V m represents the average velocity. When the signal quality index is the largest, it indicates that the ultrasonic probe signal quality is the best, and at this time, the probe angle is the best.

[0025] In the above technical solution, a three-dimensional grid model is constructed using the extracted temporal window structure features, and the incident angle of the ultrasonic probe relative to the target blood vessel and the signal quality performance are analyzed. This analysis process helps to optimize the use angle of the ultrasonic probe, so as to ensure that the ultrasonic waves can penetrate the skull more effectively and focus on the target blood vessel, improving the signal clarity and signal-to-noise ratio; at the same time, the evaluation of the signal quality also provides an important reference for the subsequent adjustment of the probe angle.

[0026] Further, the step S300 includes:

[0027] Step S301: Obtain the head image data of m patients, and obtain all the data of the temporal window parts of the m patients according to steps S102 to S204, including the temporal window contour, blood vessel structure data, blood flow velocity v, pulsatility index Y, resistance index K, the incident angle θ of the ultrasonic probe relative to the target blood vessel, and the effect S of the ultrasonic probe signal quality. Associate the basic information of the patients with the extracted data to form a data set and store it in the database;

[0028] Step S302: Obtain the relevant data sets of each patient from the database, divide them into a training set, a test set, and a validation set according to a ratio, and construct a convolutional neural network model, including an input layer, a hidden layer, and an output layer. The input layer is used to receive the feature vector of the temporal window structure data, the hidden layer is used for feature extraction and transformation, and the output layer is used to output the angles {θ1, θ2, θ3,..., θ n} of each piezoelectric ceramic, where θ nThe angle of the nth piezoelectric ceramic; training the model using the training set, adjusting the model parameters through the backpropagation algorithm to minimize the prediction error; during the training process, regularly validating the model using the validation set; testing the temporal window data of different patients using the test set to test the adaptive ability of the model, and finally using the trained model to automatically predict the angle of each piezoelectric ceramic according to the temporal window structure data of the patient.

[0029] In the above technical solution, by collecting the head image data, blood flow measurement data, and probe angle data of multiple patients, a piezoelectric ceramic angle prediction model is constructed. The model automatically predicts the optimal angle of each piezoelectric ceramic in the ultrasonic probe according to the head structure and blood flow conditions of different patients. This function improves the adaptability and intelligence level of ultrasonic examination, reduces the complexity and error of manual operation, and also improves the efficiency and accuracy of the examination.

[0030] Further, in step S400, a flexible material is covered on the surface of the transcranial Doppler probe, and a closed-loop control system is introduced inside the probe, including an angle sensor, a controller, and a drive circuit. The angle sensor is used to monitor the angle change of the piezoelectric ceramic element under the action of voltage in real time. The controller is used to receive the output signal of the angle sensor and calculate the required voltage adjustment amount according to the real-time feedback of the angle sensor and the predicted angle of each piezoelectric ceramic. The drive circuit applies the adjusted voltage to the piezoelectric ceramic element to control the angle conversion.

[0031] In the above technical solution, a flexible material is covered on the surface of the ultrasonic probe, and a closed-loop control system is introduced. By monitoring the angle change of the piezoelectric ceramic element in real time and calculating the required voltage adjustment amount according to the prediction model and real-time feedback, the angle of the probe is automatically adjusted to match the best examination conditions. This not only improves the flexibility and accuracy of ultrasonic examination, but also reduces the operation difficulty and human error, bringing a more comfortable and efficient examination experience to patients.

[0032] A transcranial Doppler flexible ultrasonic probe detection data management system, the system includes a data acquisition module, a data analysis module, an ultrasonic probe angle prediction module, and an ultrasonic probe angle adaptive adjustment module;

[0033] The data acquisition module is used to obtain the head image data of the patient and perform preprocessing, and extract the structural features of the temporal window part from the preprocessed data;

[0034] The data analysis module constructs a three-dimensional grid model containing blood vessel structures according to the extracted structural features of the temporal window part, and analyzes the incident angle of the ultrasonic probe relative to the target blood vessel in the model and evaluates the signal quality performance of the ultrasonic probe;

[0035] The ultrasonic probe angle prediction module is used to collect head image data, blood flow measurement data, and probe angle data of multiple patients to construct a piezoelectric ceramic angle prediction model, which is used to predict the optimal angle of each piezoelectric ceramic in the ultrasonic probe under different head structures and blood flow conditions of patients;

[0036] The ultrasonic probe angle adaptive adjustment module covers the surface of the transcranial Doppler probe with a flexible material and introduces a closed-loop control system inside the probe, including an angle sensor, a controller, and a drive circuit. The angle sensor is used to monitor the angle change of the piezoelectric ceramic element under the action of voltage in real time. The controller is used to receive the output signal of the angle sensor and calculate the required voltage adjustment amount according to the real-time feedback of the angle sensor and the predicted angle of each piezoelectric ceramic. The drive circuit applies the adjusted voltage to the piezoelectric ceramic element to control the angle conversion.

[0037] The data acquisition module scans the head through a medical imaging device to obtain head image data including the temporal window part, and preprocesses the head image data, including denoising and enhancing contrast. The preprocessed head image data is represented as a three-dimensional array I(x, y, z), where I represents the gray value, and I(x, y, z) represents the gray value at the coordinates (x, y, z); the structural features of the temporal window part are extracted from the preprocessed image data. The structural features of the temporal window part include using the Canny edge detection algorithm to locate the position of the temporal window in the image and using the region growing method to extract the contour of the temporal window; using the Hessian matrix filter to enhance the blood vessel structure in the image of the temporal window part and using the gradient tracking algorithm to track the path of the blood vessels; separating the blood vessel structure from the image background to obtain blood vessel structure data including the position, direction vector, center line, and radius of the blood vessels.

[0038] The data analysis module includes a grid construction unit, a blood flow velocity analysis unit, an ultrasonic probe incident angle analysis unit, and an ultrasonic probe signal quality analysis unit; the grid construction unit initializes a three-dimensional grid model M including vertices V, edges E, and faces F according to the temporal window contour extracted in step 102; using the region growing algorithm, starting from the initial point P0(x0, y0, z0) of the grid, the grid M is expanded according to the gray value in the image; a threshold T is set. For the neighborhood point P1(x1, y1, z1) of P0, if |I(x1, y1, z1) - I(x0, y0, z0)| < T, then P1 is added to the grid M, where I(x1, y1, z1) represents the gray value at the coordinate (x1, y1, z1) in the image, and I(x0, y0, z0) represents the gray value at the initial point P0. Repeat the above comparison process until no new points meet the growth conditions; during the grid expansion process, the center line and radius of the blood vessel extracted in step S102 are converted into point, line, and surface data in the grid. The center line of the blood vessel is represented by the edge E in the grid, and the radius of the blood vessel is used to determine the size and shape of the face F in the grid. The center line of the blood vessel is composed of points {P1, P2, …, P n}, where P n represents the nth point. For each point P n , there is a corresponding radius r n . All points are added to the grid M, and the corresponding face F is constructed according to the radius r n ; finally, the Laplacian coordinates of each point in the grid are calculated by the Laplacian smoothing algorithm to update the positions of the grid points, and the smoothing process is continuously iterated until the preset smoothing degree is reached, completing the optimization process of the grid M;

[0039] The blood flow velocity analysis unit calculates the blood flow velocity in the blood vessel by measuring the reflection frequency of ultrasonic waves in the blood through a Doppler ultrasound device; the ultrasonic probe incident angle analysis unit obtains the incident angle of the probe relative to the target blood vessel by analyzing the included angle between the direction vector of the blood vessel and the emission direction vector of the probe; the ultrasonic probe signal quality analysis unit analyzes the effect of the ultrasonic probe signal quality according to the blood flow velocity, pulsatility index, resistance index, and probe angle.

[0040] The ultrasonic probe angle prediction module includes a data storage unit and a model construction unit; the data storage unit obtains the head image data of m patients, and obtains all the data of the temporal window parts of the m patients according to step S102 to step S204, including the temporal window contour, blood vessel structure data, blood flow velocity v, pulsatility index Y, resistance index K, the incident angle θ of the ultrasonic probe relative to the target blood vessel, and the effect S of the ultrasonic probe signal quality. The basic information of the patient is associated with the extracted data to form a data set and stored in the database;

[0041] The model construction unit obtains the relevant data sets of each patient from the database, divides them into a training set, a test set, and a validation set according to a certain proportion, and constructs a convolutional neural network model, including an input layer, a hidden layer, and an output layer. The input layer is used to receive the feature vectors of the temporal window structure data, the hidden layer is used for feature extraction and transformation, and the output layer is used to output the angles of each piezoelectric ceramic {θ1, θ2, θ3, …, θ n}, where θ n is the angle of the nth piezoelectric ceramic; the model is trained using the training set, and the model parameters are adjusted through the backpropagation algorithm to minimize the prediction error; during the training process, the validation set is used to validate the model regularly; the test set is used to test the temporal window data of different patients to test the adaptive ability of the model, and finally the trained model is used to automatically predict the angles of each piezoelectric ceramic according to the temporal window structure data of the patient.

[0042] The ultrasonic probe angle adaptive adjustment module covers a flexible material on the surface of the transcranial Doppler probe and introduces a closed-loop control system inside the probe, including an angle sensor, a controller, and a drive circuit. The angle sensor is used to monitor the angle change of the piezoelectric ceramic element under the action of voltage in real time. The controller is used to receive the output signal of the angle sensor and calculate the required voltage adjustment amount according to the real-time feedback of the angle sensor and the predicted angles of each piezoelectric ceramic; the drive circuit applies the adjusted voltage to the piezoelectric ceramic element to control the angle conversion.

[0043] Compared with the prior art, the beneficial effects of the present invention are:

[0044] By extracting the structural features of the temporal window part in the patient's head image data, constructing a three-dimensional vascular grid model, and combining with the analysis of the incident angle of the ultrasonic probe, the present invention ensures that the relative position relationship between the ultrasonic probe and the target blood vessel is more accurate; compared with the traditional method, this method can be customized and adjusted according to the specific anatomical structure of individual patients, improving the quality of ultrasonic signals;

[0045] By constructing a convolutional neural network model, the present invention can not only predict the optimal probe angle according to the head structure and blood flow conditions of different patients, but also realize adaptive adjustment under different operating environments and patient conditions; compared with the traditional fixed probe angle or single angle adjustment method, this method is more flexible and universal, and can meet a wider range of clinical needs;

[0046] In the present invention, by covering the surface of the ultrasonic probe with a flexible material and introducing a closed-loop control system, an angle sensor is combined to monitor the angle change of the probe. According to the real-time feedback, the voltage is automatically adjusted to control the angle change of the piezoelectric ceramic, ensuring that the probe always maintains the optimal angle throughout the operation process. This not only improves the flexibility and comfort of the ultrasonic probe, but also enhances the convenience of operation, effectively reducing the unstable factors during the operation process. Description of the Drawings

[0047] Figure 1 It is a schematic diagram of probe data analysis for a transcranial Doppler flexible ultrasonic probe detection data management method of the present invention;

[0048] Figure 2 It is a flowchart of a method for a transcranial Doppler flexible ultrasonic probe detection data management method of the present invention. Detailed Embodiment

[0049] Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the scope of protection of the present invention.

[0050] Embodiment: As Figure 1 - Figure 2 shown, the present invention provides a transcranial Doppler flexible ultrasonic probe detection data management method, and the method includes the following steps:

[0051] Step S100: Obtain the head image data of the patient and perform preprocessing, and extract the structural features of the temporal window part from the preprocessed data;

[0052] Step S200: Construct a three-dimensional grid model containing blood vessel structures according to the extracted structural features of the temporal window part, analyze the incident angle of the ultrasonic probe relative to the target blood vessel in the model, and evaluate the signal quality performance of the ultrasonic probe;

[0053] Step S300: Collect the head image data, blood flow measurement data, and probe angle data of multiple patients to construct a piezoelectric ceramic angle prediction model for predicting the optimal angle of each piezoelectric ceramic in the ultrasonic probe under different patient head structures and blood flow conditions;

[0054] Step S400: Cover the surface of the ultrasonic probe with a flexible material and introduce a closed-loop control system inside the probe for adjusting the voltage of the piezoelectric ceramic according to the results predicted by the model, thereby controlling the angle of the probe.

[0055] The step S100 includes:

[0056] Step S101: Scan the head through a medical imaging device to obtain head image data containing the temporal window part;

[0057] Step S102: Preprocess the head image data, including denoising and enhancing contrast. The preprocessed head image data is represented as a three-dimensional array I(x, y, z), where I represents the gray value and I(x, y, z) represents the gray value at the coordinate (x, y, z); Extract the structural features of the temporal window part from the preprocessed image data. The structural features of the temporal window part include using the Canny edge detection algorithm to locate the position of the temporal window in the image, and using the region growing method to extract the contour of the temporal window; Use the Hessian matrix filter to enhance the vascular structure in the temporal window part of the image, and adopt the gradient tracking algorithm to track the centerline of the blood vessel by calculating the gradient of the blood vessel, where the gradient represents the direction of the blood vessel, and the curvature is calculated to represent the degree of bending of the blood vessel; Separate the vascular structure from the image background through threshold segmentation. According to the contrast between the vascular structure and the background in the enhanced image, set a threshold, and determine the area with pixel values greater than the threshold in the image as the vascular structure, and the area less than the threshold as the background. Through the segmented binary image, obtain the position information of the vascular pixels, and calculate the direction vector of the blood vessel at each pixel point through the eigenvector of the Hessian matrix; Obtain the radius of the blood vessel by measuring the distance from the blood vessel edge to the centerline; Finally, obtain the structural features of the temporal window part, including the position and contour of the temporal window in the image, and the position, direction vector, centerline and radius of the blood vessels in the temporal window part.

[0058] In the above technical solution, by obtaining the head image data of the patient and preprocessing it, the structural features of the temporal window part can be accurately extracted, including key information such as the position, direction, centerline and radius of the blood vessels. These information provide a solid foundation for subsequent construction of a three-dimensional grid model, analysis of the incident angle of the ultrasound probe, and evaluation of the signal quality, thus improving the accuracy and efficiency of transcranial Doppler ultrasound examination.

[0059] The step S200 includes:

[0060] Step S201: Initialize a three-dimensional mesh model M including vertices V, edges E, and faces F according to the temporal window contour extracted in step 102. Using the region growing algorithm, start from the initial point P0(x0, y0, z0) of the mesh and expand the mesh M according to the gray values in the image. Set a threshold T. For the neighborhood point P1(x1, y1, z1) of P0, if |I(x1, y1, z1) - I(x0, y0, z0)| < T, then add P1 to the mesh M, where I(x1, y1, z1) represents the gray value at the coordinate (x1, y1, z1) in the image, and I(x0, y0, z0) represents the gray value at the initial point P0. Repeat the above comparison process until no new points meet the growth condition. During the mesh expansion process, convert the centerline and radius of the blood vessel extracted in step S102 into point, line, and face data in the mesh. The centerline of the blood vessel is represented by the edge E in the mesh, and the radius of the blood vessel is used to determine the size and shape of the face F in the mesh. The centerline of the blood vessel is composed of the points {P1, P2, …, P n}, where P n represents the nth point. For each point P n , there is a corresponding radius r n . Add all the points to the mesh M and construct the corresponding face F according to the radius r n . Finally, calculate the Laplacian coordinates of each point in the mesh through the Laplacian smoothing algorithm to update the positions of the mesh points, and continuously iterate the smoothing process until the preset smoothing degree is reached to complete the optimization process of the mesh M;

[0061] Step S202: Measure the reflection frequency of ultrasonic waves in the blood through a Doppler ultrasound device, and calculate the blood flow velocity in the blood vessel according to the Doppler effect. According to the formula:

[0062]

[0063] where v represents the blood flow velocity in the blood vessel, f s is the received reflection ultrasonic wave frequency, f0 is the emitted ultrasonic wave frequency, and c is the sound velocity;

[0064] Step S203: Through the Doppler ultrasound device, record the emission direction of the probe ultrasonic wave during manual operation and calculate the corresponding direction vector. By analyzing the angle between the direction vector of the blood vessel and the emission direction vector of the probe, obtain the incident angle θ of the probe relative to the target blood vessel:

[0065]

[0066] where represents the direction vector of the blood vessel, A x , A y , Az respectively represent the projected lengths of the blood vessel direction on the x-axis, y-axis, and z-axis, represents the probe emission direction vector, B x , B y , B z represent the projected lengths of the ultrasonic waves emitted by the probe on the x-axis, y-axis, and z-axis;

[0067] Step S204: Analyze the effect of the ultrasonic probe signal quality based on the blood flow velocity, pulsatility index, resistance index, and probe angle. According to the formula:

[0068]

[0069] where S represents the signal quality index, V p is the maximum systolic velocity, V d is the maximum diastolic velocity, Y is the pulsatility index, K is the resistance index, and the where V m represents the average velocity. When the signal quality index is the largest, it indicates that the ultrasonic probe signal quality is the best, and the probe angle is the best at this time.

[0070] In the above technical solution, a three-dimensional grid model is constructed using the extracted temporal window structure features, and the incident angle of the ultrasonic probe relative to the target blood vessel and the signal quality performance are analyzed. This analysis process helps to optimize the use angle of the ultrasonic probe, so as to ensure that the ultrasonic waves can penetrate the skull more effectively and focus on the target blood vessel, improving the signal clarity and signal-to-noise ratio; at the same time, the evaluation of the signal quality also provides an important reference for the subsequent adjustment of the probe angle.

[0071] The step S300 includes:

[0072] Step S301: Obtain the head image data of m patients, and obtain all the data of the temporal window parts of the m patients according to steps S102 to S204, including the temporal window contour, blood vessel structure data, blood flow velocity v, pulsatility index Y, resistance index K, the incident angle θ of the ultrasonic probe relative to the target blood vessel, and the effect S of the ultrasonic probe signal quality. Associate the basic information of the patients with the extracted data to form a data set and store it in the database;

[0073] Step S302: Obtain the relevant data sets of each patient from the database, divide them into a training set, a test set, and a validation set according to a ratio, and construct a convolutional neural network model, including an input layer, a hidden layer, and an output layer. The input layer is used to receive the feature vector of the temporal window structure data, the hidden layer is used for feature extraction and transformation, and the output layer is used to output the angles {θ1, θ2, θ3,..., θ n} of each piezoelectric ceramic, where θ nThe angle of the nth piezoelectric ceramic; training the model using the training set, adjusting the model parameters through the backpropagation algorithm to minimize the prediction error; during the training process, validating the model using the validation set regularly; testing the temporal window data of different patients using the test set to test the adaptive ability of the model, and finally using the trained model to automatically predict the angle of each piezoelectric ceramic according to the temporal window structure data of the patient.

[0074] In the above technical solution, by collecting the head image data, blood flow measurement data, and probe angle data of multiple patients, a piezoelectric ceramic angle prediction model is constructed. The model automatically predicts the optimal angle of each piezoelectric ceramic in the ultrasonic probe according to the head structure and blood flow conditions of different patients. This function improves the adaptability and intelligence level of ultrasonic examination, reduces the complexity and error of manual operation, and also improves the efficiency and accuracy of the examination.

[0075] In step S400, a flexible material is covered on the surface of the transcranial Doppler probe, and a closed-loop control system is introduced inside the probe, including an angle sensor, a controller, and a drive circuit. The angle sensor is used to monitor the angle change of the piezoelectric ceramic element under the action of voltage in real time. The controller is used to receive the output signal of the angle sensor and calculate the required voltage adjustment amount according to the real-time feedback of the angle sensor and the predicted angle of each piezoelectric ceramic. The drive circuit applies the adjusted voltage to the piezoelectric ceramic element to control the angle conversion.

[0076] In the above technical solution, a flexible material is covered on the surface of the ultrasonic probe, and a closed-loop control system is introduced. By monitoring the angle change of the piezoelectric ceramic element in real time and calculating the required voltage adjustment amount according to the prediction model and real-time feedback, the angle of the probe is automatically adjusted to match the optimal examination conditions. This not only improves the flexibility and accuracy of ultrasonic examination, but also reduces the operation difficulty and human error, bringing a more comfortable and efficient examination experience to patients.

[0077] A transcranial Doppler flexible ultrasonic probe detection data management system, the system includes a data acquisition module, a data analysis module, an ultrasonic probe angle prediction module, and an ultrasonic probe angle adaptive adjustment module;

[0078] The data acquisition module is used to obtain the head image data of the patient and perform preprocessing, and extract the structural features of the temporal window part from the preprocessed data;

[0079] The data analysis module constructs a three-dimensional grid model including blood vessel structures according to the extracted structural features of the temporal window part, analyzes the incident angle of the ultrasonic probe relative to the target blood vessel in the model, and evaluates the signal quality performance of the ultrasonic probe;

[0080] The ultrasonic probe angle prediction module is used to collect head image data, blood flow measurement data and probe angle data of multiple patients to construct a piezoelectric ceramic angle prediction model, which is used to predict the optimal angle of each piezoelectric ceramic in the ultrasonic probe under different head structures and blood flow conditions of patients;

[0081] The ultrasonic probe angle adaptive adjustment module covers the surface of the transcranial Doppler probe with a flexible material and introduces a closed-loop control system inside the probe, including an angle sensor, a controller and a drive circuit. The angle sensor is used to monitor the angle change of the piezoelectric ceramic element under the action of voltage in real time. The controller is used to receive the output signal of the angle sensor and calculate the required voltage adjustment amount according to the real-time feedback of the angle sensor and the predicted angle of each piezoelectric ceramic. The drive circuit applies the adjusted voltage to the piezoelectric ceramic element to control the angle conversion.

[0082] The data acquisition module scans the head through a medical imaging device to obtain head image data including the temporal window part, and preprocesses the head image data, including denoising and enhancing contrast. The preprocessed head image data is represented as a three-dimensional array I(x, y, z), where I represents the gray value and I(x, y, z) represents the gray value at the coordinates (x, y, z); the structural features of the temporal window part are extracted from the preprocessed image data. The structural features of the temporal window part include using the Canny edge detection algorithm to locate the position of the temporal window in the image and using the region growing method to extract the contour of the temporal window; using the Hessian matrix filter to enhance the blood vessel structure in the image of the temporal window part and using the gradient tracking algorithm to track the path of the blood vessels; separating the blood vessel structure from the image background to obtain blood vessel structure data including the position, direction vector of the blood vessels, as well as the center line and radius of the blood vessels.

[0083] The data analysis module includes a grid construction unit, a blood flow velocity analysis unit, an ultrasonic probe incident angle analysis unit, and an ultrasonic probe signal quality analysis unit; the grid construction unit initializes a three-dimensional grid model M including vertices V, edges E, and faces F according to the temporal window contour extracted in step 102; using the region growing algorithm, starting from the initial point P0(x0, y0, z0) of the grid, expand the grid M according to the gray value in the image; set a threshold T, for the neighborhood point P1(x1, y1, z1) of P0, if |I(x1, y1, z1) - I(x0, y0, z0)| < T, then add P1 to the grid M, where I(x1, y1, z1) represents the gray value at the coordinate (x1, y1, z1) in the image, and I(x0, y0, z0) represents the gray value at the initial point P0, repeat the above comparison process until no new points meet the growth condition; during the grid expansion process, convert the center line and radius of the blood vessel extracted in step S102 into point, line, and surface data in the grid. The center line of the blood vessel is represented by the edge E in the grid, and the radius of the blood vessel is used to determine the size and shape of the face F in the grid. The center line of the blood vessel is composed of points {P1, P2,..., P n}, where P n represents the nth point. For each point P n , there is a corresponding radius r n . Add all points to the grid M, and construct the corresponding face F according to the radius r n ; finally, calculate the Laplacian coordinates of each point in the grid through the Laplacian smoothing algorithm to update the positions of the grid points, and continuously iterate the smoothing process until the preset smoothing degree is reached to complete the optimization process of the grid M;

[0084] The blood flow velocity analysis unit calculates the blood flow velocity in the blood vessel by measuring the reflection frequency of ultrasonic waves in the blood through a Doppler ultrasonic device; the ultrasonic probe incident angle analysis unit obtains the incident angle of the probe relative to the target blood vessel by analyzing the included angle between the direction vector of the blood vessel and the transmitting direction vector of the probe; the ultrasonic probe signal quality analysis unit analyzes the effect of the ultrasonic probe signal quality according to the blood flow velocity, pulsatility index, resistance index, and probe angle.

[0085] The ultrasonic probe angle prediction module includes a data storage unit and a model construction unit; the data storage unit obtains the head image data of m patients, and obtains all the data of the temporal window parts of m patients according to step S102 to step S204, including the temporal window contour, blood vessel structure data, blood flow velocity v, pulsatility index Y, resistance index K, the incident angle θ of the ultrasonic probe relative to the target blood vessel, and the effect S of the ultrasonic probe signal quality. Associate the basic information of the patient with the extracted data to form a data set and store it in the database;

[0086] The model construction unit obtains the relevant data sets of each patient from the database, divides them into a training set, a test set and a validation set according to a ratio, and constructs a convolutional neural network model, including an input layer, a hidden layer and an output layer. The input layer is used to receive the feature vectors of the temporal window structure data, the hidden layer is used for feature extraction and transformation, and the output layer is used to output the angles of each piezoelectric ceramic {θ1, θ2, θ3, …, θ n}, where θ n is the angle of the nth piezoelectric ceramic; the model is trained using the training set, and the model parameters are adjusted through the backpropagation algorithm to minimize the prediction error; during the training process, the model is regularly validated using the validation set; the temporal window data of different patients are tested using the test set to test the adaptive ability of the model, and finally the trained model is used to automatically predict the angles of each piezoelectric ceramic according to the temporal window structure data of the patient.

[0087] The ultrasonic probe angle adaptive adjustment module covers the surface of the transcranial Doppler probe with a flexible material and introduces a closed-loop control system inside the probe, including an angle sensor, a controller and a drive circuit. The angle sensor is used to monitor the angle change of the piezoelectric ceramic element under the action of voltage in real time. The controller is used to receive the output signal of the angle sensor and calculate the required voltage adjustment amount according to the real-time feedback of the angle sensor and the predicted angles of each piezoelectric ceramic; the drive circuit applies the adjusted voltage to the piezoelectric ceramic element to control the angle conversion.

[0088] Embodiment of the present invention: The head of a patient is scanned by a CT device to obtain a three-dimensional grayscale image data with a size of 512*512*256, the grayscale value range is 0 to 255, the Gaussian filter is used to remove noise, the standard deviation is set to 1.5, and histogram equalization is used to increase the contrast of the image. The processed head image data is represented as a three-dimensional array I(x, y, z), where I represents the grayscale value, and I(x, y, z) represents the grayscale value at the coordinate (x, y, z);

[0089] Using the Canny edge detection algorithm, set the low threshold to 50, the high threshold to 150, the edge threshold to 0.1, locate the position of the temporal window in the image as (50, 60, 70)-(150, 160, 170), use the region generation method to extract the contour of the temporal window, and set the growth threshold to 5; use the Hessian matrix filter to enhance the vascular structure in the temporal window part of the image, set the filter scale to 2, adopt the gradient tracking algorithm to trace the path of the blood vessel, first calculate the gradient of the image, the gradient direction represents the direction of the blood vessel, calculate the curvature to represent the degree of curvature of the blood vessel, and gradually trace the center line of the blood vessel along the gradient direction, with the step size set to 1, separate the vascular structure from the image background through threshold segmentation, and set a threshold T s = 128 according to the contrast between the vascular structure and the background in the enhanced image, and determine the area in the image where the pixel value is greater than T s as the vascular structure, and the area less than T s as the background. Through the segmented binary image, obtain the position information of the vascular pixels, and calculate the direction vector of the blood vessel at each pixel point using the eigenvector of the Hessian matrix. The direction vector represents the tangent direction of the blood vessel at that point; obtain the radius of the blood vessel by measuring the distance from the blood vessel edge to the center line. For each vascular pixel point, calculate the shortest distance to the center line and take the average value as the radius of the blood vessel. Finally, the structural features obtained for the temporal window part include: the position of the temporal window in the image (50, 60, 70)-(150, 160, 170), the contour (the set of pixels obtained by region growing), and the position, direction vector, center line, and radius of the blood vessels in the temporal window part;

[0090] According to the extracted temporal window contour, initialize a three-dimensional mesh model M containing vertices V, edges E, and faces F, set the initial point P0(100, 110, 120) as the center point of the temporal window contour, and use the region growing algorithm to expand the mesh M. Set the growth threshold T = 10, and according to the gray value difference |I(x1, y1, z1)-I(x0, y0, z0)| of adjacent points, when the result is less than 10, add the point P1 to the mesh;

[0091] Obtain the ultrasonic emission frequency f0 = 4 MHz of the ultrasonic probe through the Doppler ultrasound device, and the received frequency f s = 4.02 MHz, and the sound speed c = 1540 m / s; calculate v = 0.385 m / s through the formula;

[0092] Analyze the angle between the direction vector of the blood vessel and the probe emission direction vector to obtain the incident angle θ of the probe relative to the target blood vessel. The blood vessel direction vector is (0.5, 0.3, 0.8), and the probe emission direction vector is = (0.6, 0.4, 0.6), representing the probe emission direction vector, and calculate the incident angle θ≈26°;

[0093] Obtain the maximum systolic velocity V p = 0.8 m / s, the maximum diastolic velocity V d = 0.2 m / s, the average velocity V m = 0.5 m / s, calculate the pulsatility index Y = 1.2, the resistance index K = 0.75, and the signal quality index S ≈ 0.96; During the examination, for the same patient, since the probe angle will continuously change to obtain the best signal quality, it is necessary to continuously repeat the calculation of three parameters: blood flow velocity, probe incident angle, and signal quality effect. Specifically, whenever the probe angle changes, the blood flow velocity at the current angle is recalculated through the real-time acquired ultrasound image data. At the same time, the incident angle is calculated based on the position of the probe and the direction vector of the blood vessel. Then, combining these parameters with the pulsatility index and resistance index of the blood flow, the signal quality effect at the current angle is evaluated through a formula; Then, all the signal quality data and blood flow velocity data of the same patient at different probe angles are integrated, which is an important input for subsequent construction of the piezoelectric ceramic angle prediction model;

[0094] Obtain the head image data of 100 patients, extract relevant data to construct a dataset including temporal window contour, blood vessel structure data, blood flow velocity, pulsatility index, resistance index, incident angle, and signal quality index. Associate the basic information of the patients (including patient ID, name, age, etc.) with the extracted data and store it as a CSV file; Construct a convolutional neural network model. The input layer receives the feature vector. The hidden layer contains two convolutional layers and two fully connected layers. The output layer outputs the angle of each piezoelectric ceramic. Use 80% of the data as the training set, 10% as the validation set, and 10% as the test set. Train the model through the training set, use the Adam optimizer, the learning rate is 0.001, train for 50 epochs, and finally use the trained model to automatically predict the angle of each piezoelectric ceramic according to the temporal window structure data of the patient, and output [θ1 = 30°, θ2 = 25°, θ3 = 35°, …];

[0095] Cover a layer of polydimethylsiloxane flexible material on the surface of the transcranial Doppler probe, and install an angle sensor to monitor the piezoelectric ceramic angle in real time; The controller uses an Arduino microcontroller, which is used to receive the output signal of the angle sensor, and calculates the required voltage adjustment amount according to the real-time feedback of the angle sensor and the predicted angle of each piezoelectric ceramic; Use an H-bridge circuit to adjust the voltage according to the controller output to control the piezoelectric ceramic angle.

[0096] Finally, it should be noted that the above are only preferred embodiments of the present invention and are not used to limit the present invention. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art can still modify the technical solutions described in the foregoing embodiments or perform equivalent replacements for some of the technical features. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included within the protection scope of the present invention.

Claims

1. A method for managing detection data of a transcranial Doppler flexible ultrasound probe, characterized in that: The method includes the following steps: Step S100: Obtain the head image data of the patient and perform preprocessing, and extract the structural features of the temporal window part from the preprocessed data; Step S200: Construct a three-dimensional grid model including blood vessel structures according to the extracted structural features of the temporal window part, and analyze the incident angle of the ultrasound probe relative to the target blood vessel in the model and evaluate the signal quality performance of the ultrasound probe; Step S300: Collect the head image data, blood flow measurement data and probe angle data of multiple patients to construct a piezoelectric ceramic angle prediction model for predicting the optimal angle of each piezoelectric ceramic in the ultrasound probe under different head structures and blood flow conditions of patients; Step S400: Cover the surface of the ultrasound probe with a flexible material and introduce a closed-loop control system inside the probe to adjust the voltage of the piezoelectric ceramic according to the results predicted by the model, thereby controlling the angle of the probe.

2. The transcranial Doppler flexible ultrasonic probe detection data management method according to claim 1, wherein: The said step S100 includes: Step S101: Scan the head through a medical imaging device to obtain the head image data including the temporal window part; Step S102: Perform preprocessing on the head image data, including denoising and enhancing the contrast. The preprocessed head image data is represented as a three-dimensional array I(x, y, z), where I represents the gray value, and I(x, y, z) represents the gray value at the coordinate (x, y, z); Extract the structural features of the temporal window part from the preprocessed image data. The structural features of the temporal window part include using the Canny edge detection algorithm to locate the position of the temporal window in the image, and using the region growing method to extract the contour of the temporal window; Use the Hessian matrix filter to enhance the blood vessel structure in the temporal window part of the image, and use the gradient tracking algorithm to track the path of the blood vessel; Separate the blood vessel structure from the image background to obtain the blood vessel structure data including the position, direction vector of the blood vessel, and the center line and radius of the blood vessel.

3. A method for managing detection data of a transcranial Doppler flexible ultrasonic probe according to claim 1, characterized in that: The said step S200 includes: Step S201: Initialize a three-dimensional mesh model M including vertices V, edges E, and faces F according to the temporal window contour extracted in step 102. Using the region growing algorithm, start from the initial point P0(x0, y0, z0) of the mesh and expand the mesh M according to the gray value in the image. Set a threshold T. For the neighborhood point P1(x1, y1, z1) of P0, if |I(x1, y1, z1) - I(x0, y0, z0)| < T, then add P1 to the mesh M, where I(x1, y1, z1) represents the gray value at the coordinate (x1, y1, z1) in the image, and I(x0, y0, z0) represents the gray value at the initial point P0. Repeat the above comparison process until no new points meet the growth condition. During the mesh expansion process, convert the centerline and radius of the blood vessel extracted in step S102 into point, line, and face data in the mesh. The centerline of the blood vessel is represented by the edge E in the mesh, and the radius of the blood vessel is used to determine the size and shape of the face F in the mesh. The centerline of the blood vessel is composed of points {P1, P2, …, P n}, where P n represents the nth point. For each point P n , there is a corresponding radius r n . Add all points to the mesh M and construct the corresponding face F according to the radius r n . Finally, calculate the Laplacian coordinates of each point in the mesh through the Laplacian smoothing algorithm to update the positions of the mesh points, and continuously iterate the smoothing process until the preset smoothing degree is reached to complete the optimization process of the mesh M; Step S202: Measure the reflection frequency of ultrasonic waves in the blood by a Doppler ultrasound device to calculate the blood flow velocity in the blood vessel, according to the formula: where v represents the blood flow velocity in the blood vessel, F s is the received ultrasonic frequency, f0 is the transmitted ultrasonic frequency, and c is the speed of sound; Step S203: Analyze the included angle between the direction vector of the blood vessel and the emission direction vector of the probe to obtain the incident angle θ of the probe relative to the target blood vessel; Among them represents the direction vector of the blood vessel, A x , A y , A z respectively represent the projection lengths of the blood vessel direction on the x-axis, y-axis, and z-axis. represents the probe emission direction vector, B x , B y , B z represent the projection lengths of the ultrasonic wave emitted by the probe on the x-axis, y-axis, and z-axis; Step S204: Analyze the effect of the signal quality of the ultrasound probe according to the blood flow velocity, pulsatility index, resistance index and probe angle, according to the formula: where S represents the signal quality index, V p is the maximum systolic velocity, V d is the maximum diastolic velocity, Y is the pulsatility index, K is the resistance index, and the where V m represents the average velocity. When the signal quality index is maximum, it indicates that the signal quality of the ultrasonic probe is the best, and at this time the probe angle is optimal.

4. A method for managing detection data of a transcranial Doppler flexible ultrasonic probe according to claim 1, characterized in that: The said step S300 includes: Step S301: Obtain the head image data of m patients, and obtain all the data of the temporal window parts of m patients according to steps S102 to S204, including the temporal window contour, blood vessel structure data, blood flow velocity v, pulsatility index Y, resistance index K, the incident angle θ of the ultrasound probe relative to the target blood vessel, and the effect S of the signal quality of the ultrasound probe. Associate the basic information of the patient with the extracted data to form a data set and store it in the database; Step S302: Obtain the relevant data sets of each patient from the database, divide them into a training set, a test set, and a validation set according to a ratio, and construct a convolutional neural network model, including an input layer, a hidden layer, and an output layer. The input layer is used to receive the feature vectors of the temporal window structure data, the hidden layer is used for feature extraction and transformation, and the output layer is used to output the angles of each piezoelectric ceramic {θ1, θ2, θ3, …, θ n}, where θ n is the angle of the nth piezoelectric ceramic; use the training set to train the model, adjust the model parameters through the backpropagation algorithm to minimize the prediction error; during the training process, regularly use the validation set to validate the model; use the test set to test the temporal window data of different patients to test the adaptive ability of the model, and finally use the trained model to automatically predict the angles of each piezoelectric ceramic according to the temporal window structure data of the patient.

5. A method for managing detection data of a transcranial Doppler flexible ultrasonic probe according to claim 1, characterized in that: In step S400, a flexible material is covered on the surface of the transcranial Doppler probe, and a closed-loop control system is introduced inside the probe, including an angle sensor, a controller, and a drive circuit. The angle sensor is used to monitor the angle change of the piezoelectric ceramic element under the action of voltage in real time. The controller is used to receive the output signal of the angle sensor and calculate the required voltage adjustment amount according to the real-time feedback of the angle sensor and the predicted angle of each piezoelectric ceramic. The drive circuit applies the adjusted voltage to the piezoelectric ceramic element to control the angle conversion.

6. A transcranial Doppler flexible ultrasonic probe detection data management system, characterized in that: The system includes a data acquisition module, a data analysis module, an ultrasonic probe angle prediction module, and an ultrasonic probe angle adaptive adjustment module. The data acquisition module is used to obtain the head image data of the patient and perform preprocessing, and extract the structural features of the temporal window part from the preprocessed data. The data analysis module constructs a three-dimensional grid model containing blood vessel structures according to the extracted structural features of the temporal window part, analyzes the incident angle of the ultrasonic probe relative to the target blood vessel in the model, and evaluates the signal quality performance of the ultrasonic probe. The ultrasonic probe angle prediction module is used to collect the head image data, blood flow measurement data, and probe angle data of multiple patients to construct a piezoelectric ceramic angle prediction model, which is used to predict the optimal angle of each piezoelectric ceramic in the ultrasonic probe under different head structures and blood flow conditions of patients. In the ultrasonic probe angle adaptive adjustment module, a flexible material is covered on the surface of the transcranial Doppler probe, and a closed-loop control system is introduced inside the probe, including an angle sensor, a controller, and a drive circuit. The angle sensor is used to monitor the angle change of the piezoelectric ceramic element under the action of voltage in real time. The controller is used to receive the output signal of the angle sensor and calculate the required voltage adjustment amount according to the real-time feedback of the angle sensor and the predicted angle of each piezoelectric ceramic. The drive circuit applies the adjusted voltage to the piezoelectric ceramic element to control the angle conversion.

7. A transcranial Doppler flexible ultrasonic probe detection data management system according to claim 6, characterized in that: The data acquisition module scans the head through a medical imaging device to obtain the head image data containing the temporal window part, and performs preprocessing on the head image data, including denoising and enhancing the contrast. The preprocessed head image data is represented as a three-dimensional array I(x, y, z), where I represents the gray value, and I(x, y, z) represents the gray value at the coordinate (x, y, z). The structural features of the temporal window part are extracted from the preprocessed image data. The structural features of the temporal window part include using the Canny edge detection algorithm to locate the position of the temporal window in the image, and using the region growing method to extract the contour of the temporal window. The Hessian matrix filter is used to enhance the blood vessel structure in the temporal window part of the image, and the gradient tracking algorithm is used to track the path of the blood vessel. The blood vessel structure is separated from the image background, and the blood vessel structure data including the position, direction vector, center line, and radius of the blood vessel is obtained.

8. A transcranial Doppler flexible ultrasonic probe detection data management system according to claim 6, characterized in that: The data analysis module includes a grid construction unit, a blood flow velocity analysis unit, an ultrasonic probe incident angle analysis unit, and an ultrasonic probe signal quality analysis unit. The grid construction unit initializes a three-dimensional grid model M including vertices V, edges E, and faces F according to the temporal window contour extracted in step 102. Using the region growing algorithm, starting from the initial point P0(x0, y0, z0) of the grid, the grid M is expanded according to the gray value in the image. Set a threshold T. For the neighborhood point P1(x1, y1, z1) of P0, if |I(x1, y1, z1) - I(x0, y0, z0)| < T, then add P1 to the grid M, where I(x1, y1, z1) represents the gray value at the coordinate (x1, y1, z1) in the image, and I(x0, y0, z0) represents the gray value at the initial point P0. Repeat the above comparison process until no new points satisfy the growth condition; during the process of grid expansion, convert the centerline and radius of the blood vessel extracted in step S102 into point, line, and surface data in the grid. The centerline of the blood vessel is represented by the edge E in the grid, and the radius of the blood vessel is used to determine the size and shape of the surface F in the grid. The centerline of the blood vessel is composed of points {P1, P2, …, P n}, where P n represents the nth point. For each point P n , there is a corresponding radius r n . Add all points to the grid M and construct the corresponding surface F according to the radius r n . Finally, calculate the Laplacian coordinates of each point in the grid through the Laplacian smoothing algorithm to update the positions of the grid points, and continuously iterate the smoothing process until the preset smoothing degree is reached to complete the optimization process of the grid M; The blood flow velocity analysis unit calculates the blood flow velocity in the blood vessel by measuring the reflection frequency of ultrasonic waves in the blood using a Doppler ultrasound device. The ultrasonic probe incident angle analysis unit obtains the incident angle of the probe relative to the target blood vessel by analyzing the angle between the direction vector of the blood vessel and the emission direction vector of the probe. The ultrasonic probe signal quality analysis unit analyzes the effect of the ultrasonic probe signal quality based on the blood flow velocity, pulsatility index, resistance index, and probe angle.

9. The transcranial Doppler flexible ultrasonic probe detection data management system according to claim 6, characterized in that: The ultrasonic probe angle prediction module includes a data storage unit and a model construction unit. The data storage unit obtains the head image data of m patients, and obtains all the data of the temporal window parts of the m patients according to steps S102 to S204, including the temporal window contour, blood vessel structure data, blood flow velocity v, pulsatility index Y, resistance index K, the incident angle θ of the ultrasonic probe relative to the target blood vessel, and the effect S of the ultrasonic probe signal quality. The basic information of the patients is associated with the extracted data to form a data set and stored in the database. The model construction unit obtains the relevant data sets of each patient from the database, divides them into a training set, a test set and a validation set according to a ratio, and constructs a convolutional neural network model, including an input layer, a hidden layer and an output layer, where the input layer is used to receive the feature vectors of the temporal window structure data, the hidden layer is used for feature extraction and transformation, and the output layer is used to output the angles of each piezoelectric ceramic [θ1, θ2, θ3, …, θ n , where θ n is the angle of the nth piezoelectric ceramic; the model is trained using the training set, and the model parameters are adjusted through the backpropagation algorithm to minimize the prediction error; during the training process, the validation set is used to validate the model regularly; the temporal window data of different patients are tested using the test set to test the adaptive ability of the model, and finally the trained model is used to automatically predict the angles of each piezoelectric ceramic according to the temporal window structure data of the patient.

10. A transcranial Doppler flexible ultrasonic probe detection data management system according to claim 6, characterized in that: The ultrasonic probe angle adaptive adjustment module covers the surface of the transcranial Doppler probe with a flexible material and introduces a closed-loop control system inside the probe, including an angle sensor, a controller, and a drive circuit. The angle sensor is used to monitor the angle change of the piezoelectric ceramic element under the action of voltage in real time. The controller is used to receive the output signal of the angle sensor and calculate the required voltage adjustment amount according to the real-time feedback of the angle sensor and the predicted angle of each piezoelectric ceramic. The drive circuit applies the adjusted voltage to the piezoelectric ceramic element to control the angle conversion.

Images