Venipuncture correction system and use thereof

Abstract

A venipuncture correction system and its application, including a vein image acquisition module, a data preprocessing and analysis module, a puncture correction and feedback module, a system integration and control module, an interaction and display module, the vein image acquisition module is used for real-time acquisition of vein image data, the data preprocessing and analysis module is used for vein imaging data preprocessing operation and analysis and identification of venipuncture position, the puncture correction and feedback module is used for real-time puncture path correction and dynamic adjustment according to feedback, the system integration and control module is used for communication and operation management of the integrated system, the interaction and display module is used for display of puncture path correction information and control of system parameters, the application of the venipuncture correction system, a venipuncture position recognition algorithm based on multi-objective optimization is proposed to analyze and identify the venipuncture position, and a puncture path correction algorithm based on neural network is proposed to correct the venipuncture path in real time.

Description

Technical Field

[0001] This invention relates to the fields of multi-objective optimization, puncture correction, and neural network technology, specifically to a venous puncture correction system and its application. Background Technology

[0002] Multi-objective optimization technology is a technique that comprehensively optimizes multiple objectives and constraints simultaneously. It aims to solve the challenge of balancing different needs in multi-objective problems. In the field of venipuncture correction, this technology can effectively integrate multiple variables in the puncture process, such as vein location, depth, puncture angle, and needle precision. Through optimization algorithms, it achieves the optimal puncture path and operation method. Multi-objective optimization technology can not only improve the success rate of puncture but also reduce patient discomfort, reduce the risk of puncture failure, and ensure efficient and safe venipuncture operation.

[0003] Puncture correction technology is a technique that adjusts the puncture path and needle position through real-time monitoring and feedback. It aims to solve the problem of failure caused by operational errors during puncture. This technology usually combines sensors, image processing and feedback mechanisms to correct the puncture action in real time and ensure that the needle accurately enters the target vein. Puncture correction technology is particularly important for complex venous systems and individual patient differences. With puncture correction technology, the venous puncture system can correct deviations in real time during dynamic operation, which can greatly improve the accuracy and success rate of puncture operation.

[0004] Neural network technology is a machine learning technique that simulates the human nervous system. It aims to make accurate predictions and decisions by learning patterns and rules from large amounts of data. In the venipuncture correction system, neural networks can be used to analyze and predict complex issues such as vein location and needle insertion angle. By training on a large number of vein images and puncture data, neural networks can automatically identify and adjust errors in the puncture operation, thereby providing more accurate puncture paths and real-time feedback. The introduction of neural networks can improve the intelligence level of the puncture correction system, enabling the system to achieve autonomous optimization and precise control in complex and variable environments.

[0005] Existing venipuncture correction systems and their applications suffer from low adaptability and accuracy in complex environments. Due to the lack of multi-objective optimization algorithms, the system cannot comprehensively consider multiple factors such as vein depth, angle, and puncture path, resulting in inaccurate puncture paths and a high failure rate. Furthermore, existing venipuncture correction systems fail to fully utilize the intelligent learning capabilities of neural networks, lacking in-depth analysis and dynamic adjustment of vein images and puncture data. This makes the correction and feedback mechanisms during the puncture process relatively simple, unable to effectively cope with special cases and complex conditions, thus affecting the puncture success rate and user experience. Summary of the Invention

[0006] The purpose of this invention is to provide a venipuncture correction system and its application, in order to solve the problems mentioned in the background art regarding the low adaptability and accuracy of existing venipuncture correction systems and their applications in complex environments, the inability of the system to comprehensively consider multiple factors such as vein depth, angle, and puncture path, resulting in inaccurate puncture path and high puncture failure rate, and the lack of in-depth analysis and dynamic adjustment of vein images and puncture data, which makes the correction and feedback mechanism during the puncture process relatively simple and unable to effectively cope with special cases and complex conditions, thus affecting the puncture success rate and user experience.

[0007] To achieve the above objectives, the present invention provides the following technical solution: a vein puncture correction system and its application, comprising a vein image acquisition module, a data preprocessing and analysis module, a puncture correction and feedback module, a system integration and control module, and an interaction and display module, characterized in that: the vein image acquisition module is used to acquire real-time image data of veins, ensuring accurate acquisition of image data containing morphological and positional information of veins, ensuring basic data support for subsequent analysis and correction; the data preprocessing and analysis module includes a vein data preprocessing unit and a puncture position recognition unit, the vein data preprocessing unit is used to perform noise filtering, image enhancement, and feature extraction on the acquired vein image data, ensuring the generation of clear vein images to improve the accuracy of data analysis, and the puncture position recognition unit proposes a vein puncture position recognition algorithm based on multi-objective optimization, used to analyze vein image data, identify suitable puncture positions, and generate suggested puncture points. The system includes a puncture path detection module to ensure the safety and accuracy of the puncture procedure. The puncture correction and feedback module comprises an intelligent correction unit and a puncture success feedback unit. The intelligent correction unit uses a neural network-based puncture path correction algorithm to predict the location and angle of the vein and generate dynamic correction suggestions, ensuring the needle enters the vein along the optimal path, significantly improving the accuracy and success rate of the puncture procedure. The puncture success feedback unit detects whether the needle has successfully entered the target vein and performs dynamic correction adjustments by feeding back puncture success information to the system. The system integration and control module manages and coordinates the operation of each module, ensuring efficient collaboration between the acquisition, analysis, correction, and feedback stages, achieving overall system stability. The interaction and display module displays vein images, puncture paths, and correction suggestions to the user in real time, supports system parameter adjustments and operation record storage, ensuring efficient system operation and adjustment for the user.

[0008] Preferably, the vein image acquisition module, by combining infrared imaging technology, acquires real-time image information of vein location and morphology, ensuring that vein image data can be clearly acquired under different skin types, vein depths, and complex physiological conditions, providing reliable basic data for subsequent analysis and correction.

[0009] Preferably, the data preprocessing and analysis module includes a vein data preprocessing unit. The vein data preprocessing unit performs preprocessing operations such as filtering, denoising, contrast enhancement, and feature extraction on the vein imaging data to ensure that the generated vein image data is accurate and clear, eliminate errors caused by imaging equipment and human factors, and ensure the reliability of subsequent analysis results.

[0010] Preferably, the data preprocessing and analysis module includes a puncture location identification unit, which proposes a multi-objective optimization-based venipuncture location identification algorithm. By identifying and analyzing the preprocessed image, a suitable puncture location is located, ensuring the accessibility, safety, and ease of operation of the puncture location.

[0011] Preferably, the multi-objective optimization-based vein puncture location recognition algorithm is as follows: First, key features are extracted from the preprocessed vein image data, including the vein's location, diameter, and curvature. By extracting the vein's location information, the vein's diameter and curvature are calculated, where the diameter is defined as the number of pixels along the radial direction, and the curvature is defined as the degree of change in the vein's angle within a given length. The specific formula is expressed as follows:

[0012]

[0013] in, This represents the set of vein image data after preprocessing. Represented as independent vein image data, This represents the total number of vein image data. Represented as the second in the two-dimensional coordinates of a vein image point Data on the horizontal axis, Represented as the second in the two-dimensional coordinates of a vein image point Data on the vertical axis, Represented as a point in a vein image The data on the diameter of the vein at that time, Represented as a point in a vein image Local tangent angle data of the vein at that time. This is represented as overall vein diameter information. Represented as a point in a vein image The data on the diameter of the vein at that time, This represents the overall venous curvature information. This represents the total length of the vein. Represented as the index of the number of vein image data. The total number of vein image data points is represented by this value. Secondly, a multi-objective optimization model is constructed by establishing an objective function for the vein puncture process. The specific formula for the radial objective function is as follows:

[0014]

[0015] in, Represented as a radial objective function, it measures the suitability of the puncture point diameter and the ability to minimize radial deformation. This is represented as the vein diameter weighting adjustment factor. The cross-sectional area of ​​the vein, used to reflect the resistance to deformation, is expressed by the following formula for the axial objective function:

[0016]

[0017] in, This is expressed as an axial objective function, used to measure the uniformity of the curvature and diameter of the veins around the puncture point. Indicated as the number of venous puncture sites around the puncture site The diameter of each point This is represented as an index of the number of potential venipuncture sites. This represents the total number of potential venous puncture sites. Represented as curvature weighting factor, Indicated as the number of venous puncture sites around the puncture site The corresponding curvature of each point is determined. Then, a multi-objective optimization model is constructed that simultaneously considers the radial and axial requirements of the venous puncture location. The radial requirements include the diameter and deformation of the vein, while the axial requirements include the changes in the diameter and curvature of the vein. The specific formula of the multi-objective optimization model is expressed as follows:

[0018]

[0019] in, Let be the set of all non-dominated solutions. The set of solutions for all potential venous puncture sites This is represented as one of the solutions in the set of potential venous puncture sites. Represented as a superior solution The solution corresponding to a specific objective function. Indicated as belonging to, This indicates that it does not exist. This can be expressed as for all objective functions, Represented as AND solution Index of the number of relevant objective functions This means that there exists at least one objective function. Represented as AND solution Index of the number of relevant objective functions This can be expressed as seeking to minimize the objective function in a multi-objective optimization model. Represented as the first The radial diameter requirement for the first vein image point reflects the first... The diameter of the vein at each vein image point. Represented as the first The angle requirement of the vein image point reflects the first... The angle of the vein at each vein image point This is expressed as a weighting factor for the radial diameter requirement. This is expressed as a weighting factor for the radial angle requirement. This is expressed as a weighting factor for the required axial diameter. This is represented by a weighting factor for the axial curvature requirement. Represented as the first The axial diameter requirement of the first vein image point reflects the first The change in vein diameter at each vein image point Indicates the first The axial curvature requirement of the vein image point reflects the first The curvature of the vein at each venous image point is calculated. Finally, the optimal puncture location is selected from the set of non-dominated solutions using a constructed multi-objective optimization model to guide the actual venous puncture operation. The specific formula is expressed as follows:

[0020]

[0021] in, This represents the optimal puncture position selected from the set of non-dominated solutions calculated. This means finding the parameters that minimize the value of the function within the parentheses. The function, Let represent a candidate venipuncture site in the set of non-dominated solutions calculated. Indicates candidate venous puncture sites radial objective function at The value, Indicates the candidate venous puncture site The axial objective function value at that point, through... The optimal puncture location is determined by finding the puncture position in the solution set that minimizes the sum of the radial and axial objective function values. This ensures that the selected puncture location is optimized in both radial and axial aspects, thereby improving the safety and accuracy of venipuncture procedures.

[0022] Preferably, the puncture correction and feedback module includes an intelligent correction unit. The intelligent correction unit proposes a puncture path correction algorithm based on a neural network. By acquiring data processed by the data preprocessing and analysis module, it analyzes the real-time position, angle, and trajectory of the venipuncture needle and adjusts the correction in real time according to the needle's deviation, ensuring that the needle can accurately reach the target vein position according to the planned path, thereby improving the accuracy and safety of venipuncture operation.

[0023] Preferably, the neural network-based puncture path correction algorithm is as follows: First, by constructing an improved U-Net neural network model, the vein image data corresponding to the identified vein puncture location after data preprocessing and analysis is analyzed, and the algorithm is further refined by shallow embedding... Convolutional kernels are used to enhance feature extraction capabilities, thereby improving the accuracy of image segmentation. This enables precise segmentation of vein images, providing a foundation for puncture path correction. The specific formula for the convolution kernel is expressed as follows:

[0024]

[0025] in, Represented as the convolution kernel function of a neural network. Represented as the x-axis coordinates of the vein image after rotation. Represented as the vertical coordinates of the vein image after rotation. The frequency of grayscale changes in a vein image determines the... The scale of the filter represents The sensitivity of the filter to the width of edges and textures. Represented as The standard deviation of the Gaussian factor in the filter represents Filter width, Represented as an exponential function, Represented as the imaginary unit, Represented as The phase parameter of the cosine function in the kernel represents The phase shift of the filter is then achieved through shallow embedding in the U-Net model. Convolutional kernels are used to enhance feature extraction capabilities. Convolutional kernels can capture texture and edge information in images. The specific formula for feature extraction is as follows:

[0026]

[0027] in, Represented as the extracted feature map, Let N and M represent the indices of the height and width of the convolution kernel, respectively, and let M represent the total number of heights and widths of the convolution kernels. Represented as a neural network weight matrix, Represented as pixel values ​​of the input vein image. This is represented as a bias term. Next, based on the feature map of the vein location information provided by the neural network, the feature map is input into subsequent layers of the neural network. The deep learning capability of the network is then used to predict the optimal puncture path. The specific formula is as follows:

[0028]

[0029] in, This represents the optimal puncture path predicted by the neural network. This is represented as the prediction function of a neural network. This is represented as the activation function of the output layer, used to transform the output into a probability distribution. This is represented as the weight matrix of a neural network convolutional layer. This is represented as the weight matrix of a pooling layer in a neural network. This is represented as the weight matrix of a fully connected layer in a neural network. This represents the bias term of the convolutional layer in a neural network. This is represented as the bias term of the pooling layer in a neural network. This is represented as the bias term of a fully connected layer in a neural network. Represented as a non-linear activation function, the input feature map is processed through convolutional layers and activation functions, then downsampled through pooling layers, and finally classified through fully connected layers to obtain the final prediction result. Secondly, a constructed neural network model monitors the puncture path in real time and performs dynamic corrections to address deviations in the venipuncture position. The specific formula is as follows:

[0030]

[0031] in, Represented as in time Puncture location The dynamic correction factor at the location, Represented as in time Puncture location The actual puncture path at the location, Represented as in time Puncture location The optimal puncture path predicted by the neural network model at that location. This represents the dynamic range of the puncture path. Secondly, real-time path adjustments are performed, as expressed by the following formula:

[0032]

[0033] in, This is represented as the adjusted new puncture path. This refers to adjusting the step size. By comparing the image position information before and after correction, the effect of puncture path correction is verified, thereby improving the accuracy and success rate of puncture. The specific formula is as follows:

[0034]

[0035] in, This is expressed as quantization error. Represented as the actual coordinates of the vein location. Represented as the coordinates of the ideal vein location. The smaller the value, the more effective the algorithm's correction.

[0036] Preferably, the puncture correction and feedback module includes an intelligent correction unit. The puncture success feedback unit monitors and judges the venipuncture status in real time and feeds back the puncture success information to the system. It performs dynamic correction and adjustment during the puncture process to ensure the accuracy and stability of venipuncture.

[0037] Preferably, the system integration and control module coordinates the data transmission and functional interaction of each module to ensure the smooth operation of vein image acquisition, data analysis, puncture correction and feedback mechanisms, while providing real-time status monitoring to ensure the system operates efficiently and stably.

[0038] Preferably, the interaction and display module presents real-time vein images, recommended puncture paths, correction feedback, and puncture status to the user in a visual manner, ensuring that the user can quickly understand the system information and make correct operational decisions, while also providing an interactive interface and system parameter adjustment functions.

[0039] Compared with the prior art, the beneficial effects of the present invention are:

[0040] 1. The puncture location recognition unit proposes a multi-objective optimization-based vein puncture location recognition algorithm. This algorithm extracts key features from preprocessed vein image data, including vein location, diameter, and curvature, to comprehensively evaluate the geometric characteristics of the vein and the feasibility of puncture. By calculating the diameter information of each vein point along the radial direction (pixels) and the curvature information reflecting the degree of vein angle change, the multi-objective optimization-based vein puncture location recognition algorithm can achieve a quantitative description of vein geometric characteristics, providing accurate input for subsequent optimization. Considering the actual needs of puncture operations, the algorithm establishes radial and axial objective functions to measure the suitability of the vein puncture point in the cross-section, including vein diameter and deformation, and the smoothness of the puncture path, including vein diameter variation and curvature uniformity. To achieve global optimization of the puncture location, the multi-objective optimization-based vein puncture location recognition algorithm constructs a multi-objective... The optimization model comprehensively considers radial and axial constraints and requirements. Through the calculation of non-dominated solution sets, it ensures that the optimization result can achieve a reasonable balance among multiple objectives. The algorithm can automatically select the optimal puncture position from a variety of potential puncture points and generate a specific puncture path. This process effectively reduces the subjectivity and uncertainty in human decision-making, while improving the safety and accuracy of puncture. The puncture position selected by the optimization model meets the requirements of sufficiently large vein diameter, smooth curvature, and small radial deformation, making it easier for the needle to accurately enter the target vein and reducing puncture failure. In addition, the algorithm has the ability to adapt to various complex conditions, further expanding the applicability of the vein puncture correction system. Overall, the vein puncture position recognition algorithm based on multi-objective optimization provides comprehensive and intelligent support for vein puncture operations, helps to improve the efficiency and quality of medical operations, and meets the high standards required for clinical applications.

[0041] 2. The intelligent correction unit proposes a puncture path correction algorithm based on neural networks. By constructing an improved U-Net neural network model, the algorithm embeds Gabor convolution kernels in the shallow layers, enhancing the ability to capture edge and texture information in vein images. Gabor convolution kernels can highlight the geometric features of veins, especially for complex textures and low-contrast regions, achieving higher image segmentation accuracy. Accurate segmentation provides a solid foundation for subsequent puncture path correction. During feature extraction, the neural network can efficiently extract key feature maps of vein location information and, combined with deep learning capabilities, predict the optimal puncture path, making puncture path planning more intelligent and precise. Furthermore, the algorithm monitors and adjusts the puncture path in real time through a dynamic correction mechanism, ensuring... This puncture correction system can adapt to complex situations involving vein displacement and patient position changes. The constructed neural network model generates dynamic correction factors based on changes in vein position information and updates the puncture path in real time, significantly reducing puncture failures. The dynamically corrected path effectively reduces puncture errors, ensuring the needle accurately enters the target vein. The correction mechanism also verifies the effect by comparing the puncture path errors before and after adjustment, ensuring the robustness and reliability of the system. In practical applications, the algorithm's dynamic path correction function is particularly suitable for patients with complex vein distribution and dynamic changes in vein position, improving the success rate and safety of puncture operations. It provides reliable technical support for medical professionals, reduces patient pain and puncture failure rates, and enhances the intelligence and practical value of a venous puncture correction system. Attached Figure Description

[0042] Figure 1 This is a schematic diagram of the structure of the present invention; Detailed Implementation

[0043] 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.

[0044] Please see Figure 1This invention provides a vein puncture correction system and its application, including a vein image acquisition module, a data preprocessing and analysis module, a puncture correction and feedback module, a system integration and control module, and an interaction and display module. The system is characterized by: the vein image acquisition module acquiring real-time image data of veins to ensure accurate acquisition of image data containing vein morphology and location information, providing fundamental data support for subsequent analysis and correction; the data preprocessing and analysis module including a vein data preprocessing unit and a puncture location identification unit; the vein data preprocessing unit performing noise filtering, image enhancement, and feature extraction on the acquired vein image data to ensure the generation of clear vein images and improve the accuracy of data analysis; and the puncture location identification unit proposing a multi-objective optimization-based vein puncture location identification algorithm to analyze vein image data, identify suitable puncture locations, and generate suggested puncture paths to ensure accurate puncture correction. The system ensures the safety and accuracy of puncture procedures. The puncture correction and feedback module includes an intelligent correction unit and a puncture success feedback unit. The intelligent correction unit proposes a neural network-based puncture path correction algorithm to predict the location and angle of vein puncture and generate dynamic correction suggestions, ensuring the needle enters the vein along the optimal path, significantly improving the accuracy and success rate of the puncture procedure. The puncture success feedback unit detects whether the needle has successfully entered the target vein and performs dynamic correction adjustments by feeding back puncture success information to the system. The system integration and control module manages and coordinates the operation of each module, ensuring efficient collaboration between the acquisition, analysis, correction, and feedback stages, achieving overall system stability. The interaction and display module displays vein images, puncture paths, and correction suggestions to the user in real time, supports system parameter adjustments and operation record storage, ensuring efficient system operation and adjustment for the user.

[0045] See Figure 1 Furthermore, the vein image acquisition module, by combining infrared imaging technology, acquires real-time image information on the location and morphology of veins, ensuring that vein image data can be clearly acquired under different skin types, vein depths, and complex physiological conditions, providing reliable basic data for subsequent analysis and correction.

[0046] See Figure 1 Furthermore, the data preprocessing and analysis module includes a vein data preprocessing unit. This unit performs preprocessing operations on the vein imaging data, such as filtering, denoising, contrast enhancement, and feature extraction, to ensure that the generated vein image data is accurate and clear, eliminate errors caused by imaging equipment and human factors, and guarantee the reliability of subsequent analysis results.

[0047] See Figure 1Furthermore, the data preprocessing and analysis module includes a puncture location identification unit, which proposes a multi-objective optimization-based venipuncture location identification algorithm. By identifying and analyzing the preprocessed image, it locates a suitable puncture location, ensuring the accessibility, safety, and ease of operation of the puncture location.

[0048] See Figure 1 Furthermore, the multi-objective optimization-based vein puncture location recognition algorithm is as follows: First, key features are extracted from the preprocessed vein image data, including the vein's location, diameter, and curvature. By extracting the vein's location information, the vein's diameter and curvature are calculated, where the diameter is defined as the number of pixels along the radial direction, and the curvature is defined as the degree of change in the vein's angle within a given length. The specific formula is expressed as follows:

[0049]

[0050] in, This represents the set of vein image data after preprocessing. Represented as independent vein image data, This represents the total number of vein image data. Represented as the second in the two-dimensional coordinates of a vein image point Data on the horizontal axis, Represented as the second in the two-dimensional coordinates of a vein image point Data on the vertical axis, Represented as a point in a vein image The data on the diameter of the vein at that time, Represented as a point in a vein image Local tangent angle data of the vein at that time. This is represented as overall vein diameter information. Represented as a point in a vein image The data on the diameter of the vein at that time, This represents the overall venous curvature information. This represents the total length of the vein. Represented as the index of the number of vein image data. The total number of vein image data points is represented by this value. Secondly, a multi-objective optimization model is constructed by establishing an objective function for the vein puncture process. The specific formula for the radial objective function is as follows:

[0051]

[0052] in, Represented as a radial objective function, it measures the suitability of the puncture point diameter and the ability to minimize radial deformation. This is represented as the vein diameter weighting adjustment factor. The cross-sectional area of ​​the vein, used to reflect the resistance to deformation, is expressed by the following formula for the axial objective function:

[0053]

[0054] in, This is expressed as an axial objective function, used to measure the uniformity of the curvature and diameter of the veins around the puncture point. Indicated as the number of venous puncture sites around the puncture site The diameter of each point Represented as an index of the number of potential venipuncture sites. This represents the total number of potential venous puncture sites. Represented as curvature weighting factor, Indicated as the number of venous puncture sites around the puncture site The corresponding curvature of each point is determined. Then, a multi-objective optimization model is constructed that simultaneously considers the radial and axial requirements of the venous puncture location. The radial requirements include the diameter and deformation of the vein, while the axial requirements include the changes in the diameter and curvature of the vein. The specific formula of the multi-objective optimization model is expressed as follows:

[0055]

[0056] in, Let be the set of all non-dominated solutions. The set of solutions for all potential venous puncture sites This is represented as one of the solutions in the set of potential venous puncture sites. Represented as a superior solution The solution corresponding to a specific objective function. Indicated as belonging to, This indicates that it does not exist. This can be expressed as for all objective functions. Represented as AND solution Index of the number of relevant objective functions This means that there exists at least one objective function. Represented as AND solution Index of the number of relevant objective functions This can be expressed as seeking to minimize the objective function in a multi-objective optimization model. Represented as the first The radial diameter requirement for the first vein image point reflects the first... The diameter of the vein at each vein image point. Represented as the first The angle requirement of the vein image point reflects the first... The angle of the vein at each vein image point This is expressed as a weighting factor for the radial diameter requirement. This is expressed as a weighting factor for the radial angle requirement. This is expressed as a weighting factor for the required axial diameter. This is represented by a weighting factor for the axial curvature requirement. Represented as the first The axial diameter requirement of the first vein image point reflects the first The change in vein diameter at each vein image point Indicates the first The axial curvature requirement of the vein image point reflects the first The curvature of the vein at each venous image point is calculated. Finally, the optimal puncture location is selected from the set of non-dominated solutions using a constructed multi-objective optimization model to guide the actual venous puncture operation. The specific formula is expressed as follows:

[0057]

[0058] in, This represents the optimal puncture position selected from the set of non-dominated solutions calculated. This means finding the parameters that minimize the value of the function within the parentheses. The function, Let represent a candidate venipuncture site in the set of non-dominated solutions calculated. Indicates candidate venous puncture sites radial objective function at The value, Indicates the candidate venous puncture site The axial objective function value at that point, through... The optimal puncture location is determined by finding the puncture position in the solution set that minimizes the sum of the radial and axial objective function values. This ensures that the selected puncture location is optimized in both radial and axial aspects, thereby improving the safety and accuracy of venipuncture procedures.

[0059] See Figure 1 Furthermore, the puncture correction and feedback module includes an intelligent correction unit, which proposes a puncture path correction algorithm based on a neural network. By acquiring data processed by the data preprocessing and analysis module, the real-time position, angle, and trajectory of the venipuncture needle are analyzed, and the correction is adjusted in real time according to the needle's deviation, ensuring that the needle can accurately reach the target vein position according to the planned path, thereby improving the accuracy and safety of venipuncture operation.

[0060] See Figure 1Furthermore, the puncture path correction algorithm based on neural networks is as follows: First, an improved U-Net neural network model is constructed to analyze the vein image data corresponding to the identified vein puncture location after data preprocessing and analysis. This is achieved by embedding the model in a shallow layer... Convolutional kernels are used to enhance feature extraction capabilities, thereby improving the accuracy of image segmentation. This enables precise segmentation of vein images, providing a foundation for puncture path correction. The specific formula for the convolution kernel is expressed as follows:

[0061]

[0062] in, Represented as the convolution kernel function of a neural network. Represented as the x-axis coordinates of the vein image after rotation. Represented as the vertical coordinates of the vein image after rotation. The frequency of grayscale changes in a vein image determines the... The filter scale represents The sensitivity of the filter to the width of edges and textures. Represented as The standard deviation of the Gaussian factor in the filter represents Filter width, Represented as an exponential function, Represented as the imaginary unit, Represented as The phase parameter of the cosine function in the kernel represents The phase shift of the filter is then achieved through shallow embedding in the U-Net model. Convolutional kernels are used to enhance feature extraction capabilities. Convolutional kernels can capture texture and edge information in images. The specific formula for feature extraction is as follows:

[0063]

[0064] in, Represented as the extracted feature map, Let N and M represent the indices of the height and width of the convolution kernel, respectively, and let M represent the total number of heights and widths of the convolution kernels. Represented as a neural network weight matrix, Represented as pixel values ​​of the input vein image. This is represented as a bias term. Next, based on the feature map of the vein location information provided by the neural network, the feature map is input into subsequent layers of the neural network. The deep learning capability of the network is then used to predict the optimal puncture path. The specific formula is as follows:

[0065]

[0066] in, This represents the optimal puncture path predicted by the neural network. This is represented as the prediction function of a neural network. This is represented as the activation function of the output layer, used to transform the output into a probability distribution. This is represented as the weight matrix of a neural network convolutional layer. This is represented as the weight matrix of a pooling layer in a neural network. This is represented as the weight matrix of a fully connected layer in a neural network. This represents the bias term of the convolutional layer in a neural network. This is represented as the bias term of the pooling layer in a neural network. This is represented as the bias term of a fully connected layer in a neural network. Represented as a non-linear activation function, the input feature map is processed through convolutional layers and activation functions, then downsampled through pooling layers, and finally classified through fully connected layers to obtain the final prediction result. Secondly, a constructed neural network model monitors the puncture path in real time and performs dynamic corrections to address deviations in the venipuncture position. The specific formula is as follows:

[0067]

[0068] in, Represented as in time Puncture location The dynamic correction factor at the location, Represented as in time Puncture location The actual puncture path at the location, Represented as in time Puncture location The optimal puncture path predicted by the neural network model at that location. This represents the dynamic range of the puncture path. Secondly, real-time path adjustments are performed, as expressed by the following formula:

[0069]

[0070] in, This is represented as the adjusted new puncture path. This refers to adjusting the step size. By comparing the image position information before and after correction, the effect of puncture path correction is verified, thereby improving the accuracy and success rate of puncture. The specific formula is as follows:

[0071]

[0072] in, This is expressed as quantization error. Represented as the actual coordinates of the vein location. Represented as the coordinates of the ideal vein location. The smaller the value, the more effective the algorithm's correction.

[0073] See Figure 1 Furthermore, the puncture correction and feedback module includes an intelligent correction unit. The puncture success feedback unit monitors and judges the venipuncture status in real time and feeds back the puncture success information to the system. It performs dynamic correction and adjustment during the puncture process to ensure the accuracy and stability of venipuncture.

[0074] See Figure 1 Furthermore, the system integration and control module coordinates the data transmission and functional interaction of each module to ensure the smooth operation of vein image acquisition, data analysis, puncture correction, and feedback mechanisms. It also provides real-time status monitoring to ensure the system operates efficiently and stably.

[0075] See Figure 1 Furthermore, the interaction and display module presents real-time vein images, recommended puncture paths, correction feedback, and puncture status to users in a visual manner, ensuring that users can quickly understand system information and make correct operational decisions. It also provides an interactive interface and system parameter adjustment functions.

[0076] In practical use, the vein image acquisition module first acquires real-time image data of the veins, ensuring accurate acquisition of image data containing vein morphology and location information, thus providing basic data support for subsequent analysis and correction. Then, the data preprocessing and analysis module includes a vein data preprocessing unit and a puncture location recognition unit. The vein data preprocessing unit performs noise filtering, image enhancement, and feature extraction on the acquired vein image data, ensuring the generation of clear vein images to improve the accuracy of data analysis. The puncture location recognition unit proposes a multi-objective optimization-based vein puncture location recognition algorithm to analyze vein image data, identify suitable puncture locations, and generate suggested puncture paths, ensuring the safety and accuracy of the puncture operation. Secondly, the puncture correction and feedback module includes an intelligent correction unit and a puncture correction and feedback unit. The system comprises a feedback unit and an intelligent correction unit. The latter employs a neural network-based puncture path correction algorithm to predict the location and angle of vein puncture and generate dynamic correction suggestions, ensuring the needle enters the vein along the optimal path. This significantly improves the accuracy and success rate of puncture procedures. The puncture success feedback unit detects whether the needle has successfully entered the target vein and provides dynamic correction adjustments by feeding back puncture success information to the system. Finally, the system integration and control module manages and coordinates the operation of each module, ensuring efficient collaboration between data acquisition, analysis, correction, and feedback, thus achieving overall system stability. The interaction and display module displays real-time vein images, puncture paths, and correction suggestions to the user, supports system parameter adjustments and operation record storage, ensuring efficient system operation and adjustment.

[0077] 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 make equivalent substitutions for some of the technical features. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A vein puncture correction system and its application, comprising a vein image acquisition module, a data preprocessing and analysis module, a puncture correction and feedback module, a system integration and control module, and an interaction and display module, characterized in that: The vein image acquisition module is used to acquire real-time image data of veins, ensuring accurate acquisition of image data containing vein morphology and location information, and ensuring basic data support for subsequent analysis and correction. The data preprocessing and analysis module includes a vein data preprocessing unit and a puncture location recognition unit. The vein data preprocessing unit is used to perform noise filtering, image enhancement, and feature extraction on the acquired vein image data to ensure the generation of clear vein images and improve the accuracy of data analysis. The puncture location recognition unit proposes a vein puncture location recognition algorithm based on multi-objective optimization to analyze vein image data, identify suitable puncture locations, and generate suggested puncture paths to ensure the safety and accuracy of the puncture operation. The puncture correction and feedback module includes an intelligent correction unit and a puncture success feedback form. The intelligent correction unit proposes a puncture path correction algorithm based on a neural network to predict the location and angle of vein puncture and generate dynamic correction suggestions, ensuring that the needle enters the vein along the optimal path, significantly improving the accuracy and success rate of the puncture operation. The puncture success feedback unit is used to detect whether the needle has successfully entered the target vein and to perform dynamic puncture correction adjustments by feeding back puncture success information to the system. The system integration and control module is used to manage and coordinate the operation of each module, ensuring efficient collaboration between the acquisition, analysis, correction, and feedback stages of the system, and achieving overall functional stability of the system. The interaction and display module is used to display vein images, puncture paths, and correction suggestion information to the user in real time, supports system parameter adjustment and operation record storage, and ensures that the user can operate and adjust the system efficiently.

2. The venipuncture correction system and its application according to claim 1, characterized in that: The vein image acquisition module, by combining infrared imaging technology, acquires real-time image information on the location and morphology of veins, ensuring that vein image data can be clearly acquired under different skin types, vein depths, and complex physiological conditions, providing reliable basic data for subsequent analysis and correction.

3. The venipuncture correction system and its application according to claim 1, characterized in that: The data preprocessing and analysis module includes a vein data preprocessing unit and a puncture location identification unit. The vein data preprocessing unit performs preprocessing operations such as filtering, denoising, contrast enhancement, and feature extraction on the vein imaging data to ensure the accuracy and clarity of the generated vein image data, eliminate errors caused by imaging equipment and human factors, and ensure the reliability of subsequent analysis results. The puncture location identification unit proposes a vein puncture location identification algorithm based on multi-objective optimization. By identifying and analyzing the preprocessed image, it locates a suitable puncture location, ensuring the accessibility, safety, and ease of operation of the puncture location.

4. The venipuncture correction system and its application according to claim 3, characterized in that, First, key features are extracted from the preprocessed vein image data, including the location, diameter, and curvature of the veins. By extracting the location information, the diameter and curvature of the veins are calculated. The diameter is defined as the number of pixels along the radial direction, and the curvature is defined as the degree of change in the vein's angle within a given length. The specific formulas are as follows: in, This represents the set of vein image data after preprocessing. Represented as independent vein image data, This represents the total number of vein image data. Represented as the second in the two-dimensional coordinates of a vein image point Data on the horizontal axis, Represented as the second in the two-dimensional coordinates of a vein image point Data on the vertical axis, Represented as a point in a vein image The data on the diameter of the vein at that time, Represented as a point in a vein image Local tangent angle data of the vein at that time. This is represented as overall vein diameter information. Represented as a point in a vein image The data on the diameter of the vein at that time, This represents the overall venous curvature information. This represents the total length of the vein. Represented as the index of the number of vein image data. The total number of vein image data points is represented by this value. Secondly, a multi-objective optimization model is constructed by establishing an objective function for the vein puncture process. The specific formula for the radial objective function is as follows: in, Represented as a radial objective function, it measures the suitability of the puncture point diameter and the ability to minimize radial deformation. This is represented as the vein diameter weighting adjustment factor. The cross-sectional area of ​​the vein, used to reflect the resistance to deformation, is expressed by the following formula for the axial objective function: in, This is expressed as an axial objective function, used to measure the uniformity of the curvature and diameter of the veins around the puncture point. Indicated as the number of venous puncture sites around the puncture site The diameter of each point Represented as an index of the number of potential venipuncture sites. This represents the total number of potential venous puncture sites. Represented as curvature weighting factor, Indicated as the number of venous puncture sites around the puncture site The corresponding curvature of each point is determined. Then, a multi-objective optimization model is constructed that simultaneously considers the radial and axial requirements of the venous puncture location. The radial requirements include the diameter and deformation of the vein, while the axial requirements include the changes in the diameter and curvature of the vein. The specific formula of the multi-objective optimization model is expressed as follows: in, Let be the set of all non-dominated solutions. The set of solutions for all potential venous puncture sites This is represented as one of the solutions in the set of potential venous puncture sites. Represented as a superior solution The solution corresponding to a specific objective function. Indicated as belonging to, This indicates that it does not exist. This can be expressed as for all objective functions. Represented as AND solution Index of the number of relevant objective functions This means that there exists at least one objective function. Represented as AND solution Index of the number of relevant objective functions This can be expressed as seeking to minimize the objective function in a multi-objective optimization model. Represented as the first The radial diameter requirement for the first vein image point reflects the first... The diameter of the vein at each vein image point. Represented as the first The angle requirement of the vein image point reflects the first... The angle of the vein at each vein image point This is expressed as a weighting factor for the radial diameter requirement. This is expressed as a weighting factor for the radial angle requirement. This is expressed as a weighting factor for the required axial diameter. This is represented by a weighting factor for the axial curvature requirement. Represented as the first The axial diameter requirement of the first vein image point reflects the first The change in vein diameter at each vein image point Indicates the first The axial curvature requirement of the vein image point reflects the first The curvature of the vein at each venous image point is calculated. Finally, the optimal puncture location is selected from the set of non-dominated solutions using a constructed multi-objective optimization model to guide the actual venous puncture operation. The specific formula is expressed as follows: in, This represents the optimal puncture position selected from the set of non-dominated solutions calculated. This means finding the parameters that minimize the value of the function within the parentheses. The function, Let represent a candidate venipuncture site in the set of non-dominated solutions calculated. Indicates candidate venous puncture sites radial objective function at The value, Indicates the candidate venous puncture site The axial objective function value at that point, through... The optimal puncture location is determined by finding the puncture position in the solution set that minimizes the sum of the radial and axial objective function values. This ensures that the selected puncture location is optimized in both radial and axial aspects, thereby improving the safety and accuracy of venipuncture procedures.

5. The venipuncture correction system and its application according to claim 1, characterized in that: The puncture correction and feedback module includes an intelligent correction unit and a puncture success feedback unit. The intelligent correction unit proposes a puncture path correction algorithm based on a neural network. By acquiring data processed by the data preprocessing and analysis module, it analyzes the real-time position, angle, and trajectory of the venipuncture needle and adjusts the correction in real time according to the needle's deviation, ensuring that the needle can accurately reach the target vein position according to the planned path, thereby improving the accuracy and safety of venipuncture operation. The puncture success feedback unit monitors and judges the venipuncture status in real time and feeds back the puncture success information to the system. It performs dynamic correction and adjustment during the puncture process to ensure the accuracy and stability of venipuncture.

6. The venipuncture correction system and its application according to claim 5, characterized in that, First, an improved U-Net neural network model is constructed to analyze the vein image data corresponding to the identified vein puncture sites after data preprocessing and analysis. This is achieved by embedding the model in a shallow layer. Convolutional kernels are used to enhance feature extraction capabilities, thereby improving the accuracy of image segmentation. This enables precise segmentation of vein images, providing a foundation for puncture path correction. The specific formula for the convolution kernel is expressed as follows: in, Represented as the convolution kernel function of a neural network. Represented as the x-axis coordinates of the vein image after rotation. Represented as the vertical coordinates of the vein image after rotation. The frequency of grayscale changes in a vein image determines the... The filter scale represents The sensitivity of the filter to the width of edges and textures. Represented as The standard deviation of the Gaussian factor in the filter represents Filter width, Represented as an exponential function, Represented as the imaginary unit, Represented as The phase parameter of the cosine function in the kernel represents The phase shift of the filter is then achieved through shallow embedding in the U-Net model. Convolutional kernels are used to enhance feature extraction capabilities. Convolutional kernels can capture texture and edge information in images. The specific formula for feature extraction is as follows: in, Represented as the extracted feature map, Let N and M represent the indices of the height and width of the convolution kernel, respectively, and let M represent the total number of heights and widths of the convolution kernels. Represented as a neural network weight matrix, Represented as pixel values ​​of the input vein image. This is represented as a bias term. Next, based on the feature map of the vein location information provided by the neural network, the feature map is input into subsequent layers of the neural network. The deep learning capability of the network is then used to predict the optimal puncture path. The specific formula is as follows: in, This represents the optimal puncture path predicted by the neural network. This is represented as the prediction function of a neural network. This is represented as the activation function of the output layer, used to transform the output into a probability distribution. This is represented as the weight matrix of a neural network convolutional layer. This is represented as the weight matrix of a pooling layer in a neural network. This is represented as the weight matrix of a fully connected layer in a neural network. This represents the bias term of the convolutional layer in a neural network. This is represented as the bias term of the pooling layer in a neural network. This is represented as the bias term of a fully connected layer in a neural network. Represented as a non-linear activation function, the input feature map is processed through convolutional layers and activation functions, then downsampled through pooling layers, and finally classified through fully connected layers to obtain the final prediction result. Secondly, a constructed neural network model monitors the puncture path in real time and performs dynamic corrections to address deviations in the venipuncture position. The specific formula is as follows: in, Represented as in time Puncture location The dynamic correction factor at the location, Represented as in time Puncture location The actual puncture path at the location, Represented as in time Puncture location The optimal puncture path predicted by the neural network model at that location. This represents the dynamic range of the puncture path. Secondly, real-time path adjustments are performed, as expressed by the following formula: in, This is represented as the adjusted new puncture path. This refers to adjusting the step size. By comparing the image position information before and after correction, the effect of puncture path correction is verified, thereby improving the accuracy and success rate of puncture. The specific formula is as follows: in, This is expressed as quantization error. Represented as the actual coordinates of the vein location. Represented as the coordinates of the ideal vein location. The smaller the value, the more effective the algorithm's correction.

7. The venipuncture correction system and its application according to claim 1, characterized in that: The system integration and control module coordinates the data transmission and functional interaction of each module to ensure the smooth operation of vein image acquisition, data analysis, puncture correction, and feedback mechanisms. It also provides real-time status monitoring to ensure the system operates efficiently and stably.

8. The venipuncture correction system and its application according to claim 1, characterized in that: The interaction and display module presents real-time vein images, recommended puncture paths, correction feedback, and puncture status to users in a visual manner, ensuring that users can quickly understand system information and make correct operational decisions. It also provides an interactive interface and system parameter adjustment functions.

Images