A deep vein puncture path generation method and device
By acquiring continuous multi-frame ultrasound images to identify anatomical target features and calculate the shortest venous puncture path, the problem of inaccurate puncture caused by insufficient experience of medical staff is solved, the accuracy and safety of puncture are improved, and the promotion of ultrasound imaging is facilitated.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- GUANGDONG GENERAL HOSPITAL
- Filing Date
- 2023-03-02
- Publication Date
- 2026-05-29
Smart Images
Figure CN116704213B_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of data processing technology, and in particular to a method and apparatus for generating deep vein puncture paths. Background Technology
[0002] Deep vein puncture is one of the most common procedures in intensive care units, frequently used for medication administration, intravenous infusion, and other similar procedures. The challenge of performing deep vein puncture lies in finding the appropriate puncture path for needle insertion, and the quality of this insertion directly affects the success of the procedure. Currently, medical staff typically use blind puncture or ultrasound-guided needle insertion methods when performing deep vein punctures. Medical staff often rely on experience to select the needle insertion path, which demands a high level of skill and extensive experience from the staff. Furthermore, the path chosen based on experience is often inaccurate, affecting the accuracy and outcome of the puncture.
[0003] Interventional ultrasound technology, a branch of modern ultrasound medicine, is a new technology developed based on ultrasound imaging to further meet the needs of clinical diagnosis and treatment. Ultrasound-guided puncture, in particular, is widely used in various clinical fields. Compared with manual puncture, it offers advantages such as greater precision, simpler operation, higher success rate, and less collateral damage. However, currently, determining the puncture approach point, needle insertion position, and angle in ultrasound images still requires the operator's judgment, demanding extensive experience and presenting significant challenges for beginners. Distinguishing between arteries and veins is often difficult in clinical practice, requiring expertise in ultrasound imaging. This limits the accuracy of the judgment, as it is dependent on the operator's experience and skills, and carries a risk of misjudgment. Furthermore, this restricts the promotion and widespread adoption of ultrasound-guided deep vein puncture. Summary of the Invention
[0004] This application provides a method and apparatus for generating deep vein puncture paths, which identifies anatomical target classes in ultrasound images and outputs the shortest vein puncture path, providing medical staff with reliable object recognition and path guidance, avoiding the risk of misjudgment due to insufficient skills or experience among medical staff, and improving the accuracy of vein puncture.
[0005] Firstly, this application provides a method for generating a deep vein puncture path, including:
[0006] Acquire ultrasound images of the puncture site; wherein the ultrasound images are sequential ultrasound images of multiple consecutive frames;
[0007] The categories of anatomical targets are identified based on the color and shape features of the ultrasound images, and corresponding feature vectors are established; wherein each feature vector corresponds to an anatomical target.
[0008] The ultrasound image after establishing feature vectors is input into a preset path model, so that the path model establishes a venous puncture path according to each feature vector and performs a traversal query to output the shortest venous puncture path; wherein, the path model establishes a venous puncture path for each feature vector according to a preset objective function.
[0009] In this way, acquiring multiple consecutive frames of temporal ultrasound images and identifying anatomical targets based on the color and shape features of the ultrasound images eliminates the need for manual ultrasound image recognition. This reduces the professional knowledge requirements for ultrasound imaging operators, facilitating the widespread use of ultrasound imaging, and also reduces the risk of misjudgment due to insufficient operator experience and skills. Furthermore, after establishing a corresponding feature vector for each anatomical target, the ultrasound image is input into a preset path model. The model then establishes a venous puncture path based on the feature vector and iterates through the path to find the shortest venous puncture path. The path model determines the shortest venous puncture path based on the preset target through data processing and calculation, providing medical personnel with a reliable path guide during puncture procedures. This avoids hospital staff selecting incorrect or inappropriate puncture paths due to insufficient experience or skills, improving the accuracy and safety of venous puncture.
[0010] In one implementation, after acquiring the ultrasound image of the puncture site, the method further includes:
[0011] The pixel changes of the ultrasound image in various directions are obtained based on the maximum density projection method.
[0012] The pixel change information of the ultrasound image is fused to generate a first ultrasound image.
[0013] In one implementation, the step of identifying the category of the anatomical target based on the color and shape features of the ultrasound image and establishing a corresponding feature vector specifically includes:
[0014] Obtain the RGB color feature values of each anatomical target in the first ultrasound image;
[0015] Shape feature values in the first ultrasound image are obtained according to the Hessian matrix; wherein each shape feature value corresponds to the shape feature orientation of the anatomical target;
[0016] Cluster the shape feature values according to the category of the anatomical target;
[0017] Clustering optimization is performed using the maximum inter-class variance method to output the category corresponding to each anatomical target and establish the corresponding feature vector.
[0018] In one implementation, before inputting the ultrasound image with the established feature vectors into the preset path model, the method further includes: specifically including:
[0019] The tolerance coefficient and risk coefficient of each of the aforementioned feature vectors are determined based on clinical requirements and anatomical relationships;
[0020] Establish the coordinates corresponding to each feature vector in the first ultrasound image; wherein, the coordinates are represented by the following coordinate formula:
[0021] ;
[0022] in, Let X be the coordinate position of the feature vector at the puncture site; The coordinate position is The red, green, and blue color characteristic values; The category of the anatomical target; The tolerance coefficient value of the anatomical target; The risk coefficient value of the dissected target.
[0023] In one implementation, the objective function is specifically:
[0024] ;
[0025] Where d is the Euclidean distance corresponding to each of the aforementioned feature vectors; For each of the described anatomical targets, a risk coefficient value is provided. For each of the described anatomical targets, the tolerance coefficient value is given.
[0026] In one implementation, after the path model establishes a vein puncture path based on each feature vector and performs a traversal query to output the shortest vein puncture path, it further includes obtaining the actual needle insertion point on the human body surface, specifically:
[0027] ;
[0028] in, The actual needle entry point on the human body surface; The width of the ultrasonic probe; The coordinate width value of the first ultrasound image; The needle insertion point is the point along the shortest venous puncture path.
[0029] Secondly, this application also provides a deep vein puncture path generation device, including a data acquisition module, an object recognition module, and a path generation module, specifically:
[0030] The data acquisition module is used to acquire ultrasound images of the puncture site; wherein, the ultrasound images are sequential ultrasound images of multiple consecutive frames;
[0031] The object recognition module is used to identify the category of the anatomical target based on the color and shape features of the ultrasound image, and to establish a corresponding feature vector; wherein each feature vector corresponds to an anatomical target;
[0032] The path generation module is used to input the ultrasound image after the feature vector is established into a preset path model, so that the path model establishes a venous puncture path according to each feature vector and performs a traversal query to output the shortest venous puncture path; wherein, the path model establishes a venous puncture path for each feature vector according to a preset objective function.
[0033] In this way, after the data acquisition module acquires multiple consecutive frames of time-series ultrasound images, the object recognition module identifies anatomical targets based on the color and shape features of the ultrasound images. This eliminates the need for manual ultrasound image recognition, reducing the professional knowledge requirements for ultrasound imaging operators and facilitating the widespread use of ultrasound imaging. It also reduces the risk of misjudgment due to insufficient operator experience and skills. Furthermore, after establishing a corresponding feature vector for each anatomical target, the ultrasound image is input into a preset path model. The path model then establishes a venous puncture path based on the feature vector and iterates through the path to output the shortest venous puncture path. The model determines the shortest venous puncture path based on the preset target through data processing and calculation, providing reliable path guidance for medical personnel during puncture procedures. This avoids hospital staff selecting incorrect or inappropriate puncture paths due to insufficient experience or skills, improving the accuracy and safety of venous punctures.
[0034] In one implementation, after the data acquisition module acquires an ultrasound image of the puncture site, it further includes:
[0035] The pixel changes of the ultrasound image in various directions are obtained based on the maximum density projection method.
[0036] The pixel change information of the ultrasound image is fused to generate a first ultrasound image.
[0037] In one implementation, the object recognition module is used to identify the category of the anatomical target based on the color and shape features of the ultrasound image, and to establish a corresponding feature vector, specifically including:
[0038] Obtain the RGB color feature values of each anatomical target in the first ultrasound image;
[0039] Shape feature values in the first ultrasound image are obtained according to the Hessian matrix; wherein each shape feature value corresponds to the shape feature orientation of the anatomical target;
[0040] Cluster the shape feature values according to the category of the anatomical target;
[0041] Clustering optimization is performed using the maximum inter-class variance method to output the category corresponding to each anatomical target and establish the corresponding feature vector.
[0042] In one implementation, before inputting the ultrasound image with the established feature vectors into the preset path model, the method further includes:
[0043] The tolerance coefficient and risk coefficient of each of the aforementioned feature vectors are determined based on clinical requirements and anatomical relationships;
[0044] Establish the coordinates corresponding to each feature vector in the first ultrasound image; wherein, the coordinates are represented by the following coordinate formula:
[0045] ;
[0046] in, Let X be the coordinate position of the feature vector at the puncture site; The coordinate position is The red, green, and blue color characteristic values; The category of the anatomical target; The tolerance coefficient value of the anatomical target; The risk coefficient value of the dissected target.
[0047] In one implementation, the objective function is specifically:
[0048] ;
[0049] Where d is the Euclidean distance corresponding to each of the aforementioned feature vectors; For each of the described anatomical targets, a risk coefficient value is provided. For each of the described anatomical targets, the tolerance coefficient value is given.
[0050] In one implementation, after the path model establishes a vein puncture path based on each feature vector and performs a traversal query to output the shortest vein puncture path, it further includes obtaining the actual needle insertion point on the human body surface, specifically:
[0051] ;
[0052] in, The actual needle entry point on the human body surface; The width of the ultrasonic probe; The coordinate width value of the first ultrasound image; The needle insertion point is the point along the shortest venous puncture path.
[0053] Thirdly, this application also provides a terminal device, including a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor, wherein the processor executes the computer program to implement the deep vein puncture path generation method as described above.
[0054] Fourthly, this application also provides a computer-readable storage medium, the computer-readable storage medium including a stored computer program, wherein, when the computer program is executed, it controls the device where the computer-readable storage medium is located to perform the deep vein puncture path generation method as described above. Attached Figure Description
[0055] Figure 1 This is a flowchart illustrating a method for generating a deep vein puncture path according to an embodiment of the present invention;
[0056] Figure 2 This is a schematic flowchart of a maximum density projection method provided in an embodiment of the present invention;
[0057] Figure 3 This is a module structure diagram of a deep vein puncture path generation device provided in an embodiment of the present invention. Detailed Implementation
[0058] The specific embodiments of the present invention will be described in further detail below with reference to the accompanying drawings and examples. The following examples are for illustrative purposes only and are not intended to limit the scope of the invention.
[0059] The terms "first" and "second," etc., in the specification, claims, and accompanying drawings of this application are used to distinguish different objects, not to describe a specific order. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion. For example, a process, method, system, product, or apparatus that includes a series of steps or units is not limited to the listed steps or units, but may optionally include steps or units not listed, or may optionally include other steps or units inherent to such processes, methods, products, or apparatus.
[0060] In this document, the term "embodiment" means that a particular feature, structure, or characteristic described in connection with an embodiment may be included in at least one embodiment of this application. The appearance of this phrase in various places throughout the specification does not necessarily refer to the same embodiment, nor is it a separate or alternative embodiment mutually exclusive with other embodiments. It will be explicitly and implicitly understood by those skilled in the art that the embodiments described herein can be combined with other embodiments.
[0061] Example 1
[0062] See Figure 1 , Figure 1 This is a flowchart illustrating a method for generating a deep vein puncture path according to an embodiment of the present invention. The embodiment of the present invention provides a method for generating a deep vein puncture path, including steps 101 to 103, each step of which is detailed below:
[0063] Step 101: Acquire ultrasound images of the puncture site; wherein the ultrasound images are sequential ultrasound images of multiple consecutive frames;
[0064] Step 102: Identify the category of the anatomical target based on the color and shape features of the ultrasound image, and establish a corresponding feature vector; wherein each feature vector corresponds to an anatomical target;
[0065] Step 103: Input the ultrasound image after establishing the feature vector into a preset path model, so that the path model establishes a vein puncture path according to each feature vector, and performs a traversal query to output the shortest vein puncture path; wherein, the path model establishes a vein puncture path for each feature vector according to a preset objective function.
[0066] In this embodiment of the invention, short-term temporal ultrasound images of the deep vein puncture area are acquired. Multiple consecutive frames of ultrasound images are arranged sequentially, and the pixel variations in each direction of the ultrasound images are obtained using the maximum density projection method. See also... Figure 2 , Figure 2 This is a flowchart illustrating a maximum intensity projection (MIP) method provided in an embodiment of the present invention. Specifically: A ray of light is emitted from back to front along the slice direction and projected onto a two-dimensional plane. The maximum pixel value along the path of the ray is the pixel value of the image on that two-dimensional plane. Rays of light are emitted sequentially from each direction of the ultrasound image and projected onto a two-dimensional plane. The pixels with the highest density in the image are retained and projected onto the two-dimensional plane. The pixel change information in each direction is fused to generate a first ultrasound image, namely a MIP reconstructed image (maximum intensity projection, MIP). Even small density changes can be fully displayed on the MIP reconstructed image, and the stenosis, dilation, and filling defects of blood vessels can also be displayed well.
[0067] In this embodiment of the invention, after generating the first ultrasound image, the RGB color feature values of each anatomical target in the first ultrasound image are obtained. Specifically, the first ultrasound image is input into a preset image feature extraction model, and the GBR color component feature values of the first ultrasound image are obtained by calling the corresponding function formula according to the model. Preferably, in this embodiment, color feature extraction is performed using Matlab. For ease of calculation, the color feature values can be directly obtained from the color ultrasound image containing blood vessels. Further, shape feature values in the first ultrasound image are obtained according to the Hessian matrix; wherein each shape feature value corresponds to the shape feature orientation of one anatomical target. Specifically, the concavity / convexity of each anatomical target is determined by the sign of the feature value obtained from the Hessian matrix, thereby determining the shape feature orientation of each anatomical target. In this embodiment of the invention, K-means clustering is used to cluster the shape feature values, and the categories of each anatomical object are initially classified. After optimizing and adjusting the clustering results according to the maximum inter-class method, the category corresponding to each anatomical target is output, and a corresponding feature vector is established. In this embodiment of the invention, the categories of anatomical targets include: veins, arteries, muscle soft tissue, and other tissues. After determining the category and corresponding feature vector of each anatomical target, the first ultrasound image is input into a preset path model. In this embodiment of the invention, before inputting the first ultrasound image into the preset path model, the tolerance coefficient and risk coefficient of each feature vector are determined based on clinical requirements and anatomical relationships. Corresponding coordinates are then established for each feature vector. Specifically, the coordinates can be expressed using the following formula:
[0068] ;
[0069] in, Let X be the coordinate position of the feature vector X at the puncture site; It is the coordinate position. The red, green, and blue color characteristic values; Category of the anatomical target; The tolerance coefficient value of the anatomical target; This represents the risk coefficient value for the anatomical target. Preferably, the tolerance coefficient value is less than 1. A higher tolerance coefficient (closer to 1) indicates a distance closer to 1, making it suitable for needle insertion. If the anatomical target is determined to be an artery, its risk coefficient value is set to the maximum, indicating that the puncture path at that point is blocked or has a maximum distance. Conversely, if it is not an artery, it is set to 1, indicating that the risk of puncture at that point is minimal.
[0070] In this embodiment of the invention, a linear model is used. ,in , where is the coordinate of each feature vector. These are the coefficients of the linear model. In this embodiment of the invention, the linear model establishes the venous puncture path corresponding to the feature vector according to a preset objective function; wherein, the objective function can be specifically expressed by the following formula:
[0071] ;
[0072] d is the Euclidean distance corresponding to the feature vector; The risk coefficient value for the dissection target; This represents the tolerance coefficient value of the anatomical target. The linear model establishes the puncture path for each feature vector based on this objective function, and iterates through the generated puncture paths according to a preset function, outputting the shortest venous puncture path. The preset function can be expressed by the following formula: .
[0073] As a preferred embodiment of the present invention, after determining the generated shortest venous puncture path, the method further includes obtaining the actual needle insertion point on the human body surface, which can be expressed by the following formula: ;
[0074] in, The actual needle entry point on the human body surface; The width of the ultrasonic probe; The coordinate width value of the first ultrasound image; The needle insertion point is the point along the shortest venous puncture path.
[0075] In this embodiment of the invention, a deep vein puncture path generation device is also provided, including a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor. When the processor executes the computer program, it implements the above-described deep vein puncture path generation method.
[0076] In this embodiment of the invention, a computer-readable storage medium is also provided, which includes a stored computer program, wherein the computer program controls the device where the computer-readable storage medium is located to execute the above-described deep vein puncture path generation method when it is running.
[0077] For example, the computer program may be divided into one or more modules, which are stored in the memory and executed by the processor to complete the present invention. The one or more modules may be a series of computer program instruction segments capable of performing specific functions, which describe the execution process of the computer program in the deep vein puncture path generation device.
[0078] The deep vein puncture path generation device can be a computing device such as a desktop computer, laptop, handheld computer, or cloud server. The deep vein puncture path generation device may include, but is not limited to, a processor, memory, and display. Those skilled in the art will understand that the above components are merely examples of deep vein puncture path generation devices and do not constitute a limitation on the deep vein puncture path generation device. It may include more or fewer components than described above, or a combination of certain components, or different components. For example, the deep vein puncture path generation device may also include input / output devices, network access devices, buses, etc.
[0079] The processor can be a Central Processing Unit (CPU), or other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor can be a microprocessor or any conventional processor. The processor is the control center of the deep vein puncture path generation device, connecting all parts of the device via various interfaces and lines.
[0080] The memory can be used to store the computer program and / or modules. The processor implements various functions of the deep vein puncture path generation device by running or executing the computer program and / or modules stored in the memory and calling the data stored in the memory. The memory may mainly include a program storage area and a data storage area. The program storage area may store the operating system, at least one application program required for a function (such as sound playback function, text conversion function, etc.), etc.; the data storage area may store data created according to the use of the mobile phone (such as audio data, text message data, etc.). In addition, the memory may include high-speed random access memory, and may also include non-volatile memory, such as hard disk, memory, plug-in hard disk, smart media card (SMC), secure digital card (SD) card, flash card, at least one disk storage device, flash memory device, or other volatile solid-state storage device.
[0081] The module integrated into the deep vein puncture path generation device, if implemented as a software functional unit and sold or used as an independent product, can be stored in a computer-readable storage medium. Based on this understanding, all or part of the processes in the above embodiments of the present invention can also be implemented by a computer program instructing related hardware. The computer program can be stored in a computer-readable storage medium, and when executed by a processor, it can implement the steps of the various method embodiments described above. The computer program includes computer program code, which can be in the form of source code, object code, executable files, or certain intermediate forms. The computer-readable medium can include: any entity or device capable of carrying the computer program code, recording media, USB flash drives, portable hard drives, magnetic disks, optical disks, computer memory, read-only memory (ROM), random access memory (RAM), electrical carrier signals, telecommunication signals, and software distribution media, etc. It should be noted that the content included in the computer-readable medium can be appropriately added or removed according to the requirements of legislation and patent practice in the jurisdiction. For example, in some jurisdictions, according to legislation and patent practice, computer-readable media do not include electrical carrier signals and telecommunication signals. This can be understood and implemented by those skilled in the art without any creative effort.
[0082] This invention provides a method for generating deep vein puncture paths. It acquires multiple consecutive frames of temporal ultrasound images and identifies anatomical targets based on the color and shape features of the ultrasound images, eliminating the need for manual image recognition. Firstly, after establishing a corresponding feature vector for each anatomical target, the ultrasound image is input into a preset path model. The model then constructs a vein puncture path based on the feature vector and iterates through the paths to output the shortest possible puncture path. The path model determines the shortest puncture path based on the preset target through data processing and calculation, providing reliable path guidance for medical personnel during puncture procedures. This avoids hospital staff selecting incorrect or inappropriate puncture paths due to insufficient experience or skill, improving the accuracy and safety of vein puncture. Secondly, it reduces the professional knowledge requirements for ultrasound imaging operators, facilitating the widespread use of ultrasound imaging. Furthermore, it reduces the risk of misjudgment due to insufficient operator experience and skill.
[0083] Example 2
[0084] See Figure 3 , Figure 3This is a modular structure diagram of a deep vein puncture path generation device provided in an embodiment of the present invention. The device includes a data acquisition module 301, an object recognition module 302, and a path generation module 303. Specifically:
[0085] The data acquisition module 301 is used to acquire ultrasound images of the puncture site; wherein, the ultrasound images are sequential ultrasound images of multiple consecutive frames;
[0086] The object recognition module 302 is used to identify the category of the anatomical target based on the color and shape features of the ultrasound image, and to establish a corresponding feature vector; wherein each feature vector corresponds to an anatomical target;
[0087] The path generation module 303 is used to input the ultrasound image after establishing feature vectors into a preset path model, so that the path model establishes a venous puncture path according to each feature vector and performs a traversal query to output the shortest venous puncture path; wherein, the path model establishes a venous puncture path for each feature vector according to a preset objective function.
[0088] In this embodiment of the invention, after the data acquisition module 301 acquires the ultrasound image of the puncture site, it further includes:
[0089] Based on the maximum density projection method, the pixel changes of the ultrasound image in various directions are obtained; the pixel change information of the ultrasound image is fused to generate a first ultrasound image. In this embodiment of the invention, the object recognition module 302 identifies the category of the anatomical target based on the color and shape features of the ultrasound image and establishes a corresponding feature vector, specifically including:
[0090] Obtain the RGB color feature values of each anatomical target in the first ultrasound image;
[0091] Shape feature values in the first ultrasound image are obtained according to the Hessian matrix; wherein each shape feature value corresponds to the shape feature orientation of the anatomical target;
[0092] Cluster the shape feature values according to the category of the anatomical target;
[0093] Clustering optimization is performed using the maximum inter-class variance method to output the category corresponding to each anatomical target and establish the corresponding feature vector.
[0094] In this embodiment of the invention, before the path generation module 303 inputs the ultrasound image after establishing the feature vector into the preset path model, it further includes: specifically including:
[0095] The tolerance coefficient and risk coefficient of each feature vector are determined according to clinical requirements and anatomical relationships; the coordinates corresponding to each feature vector in the first ultrasound image are established; wherein, the coordinates are represented by the following coordinate formula:
[0096] ;
[0097] in, The coordinates of the feature vector X at the puncture site; The coordinate position is The red, green, and blue color characteristic values; The category of the anatomical target; The tolerance coefficient value of the anatomical target; The risk coefficient value of the dissected target.
[0098] In this embodiment of the invention, the objective function is specifically:
[0099] ;
[0100] Where d is the Euclidean distance corresponding to each of the aforementioned feature vectors; For each of the described anatomical targets, a risk factor value is provided. For each of the described anatomical targets, the tolerance coefficient value is given.
[0101] In this embodiment of the invention, after the path model establishes a vein puncture path based on each feature vector and performs a traversal query to output the shortest vein puncture path, it also includes obtaining the actual needle insertion point on the human body surface, specifically:
[0102] ;
[0103] in, The actual needle entry point on the human body surface; The width of the ultrasonic probe; The coordinate width value of the first ultrasound image; The needle insertion point is the point along the shortest venous puncture path.
[0104] Those skilled in the art will understand that, for the sake of convenience and brevity, the specific working process of the device described above can be referred to the corresponding process in the foregoing method embodiments, and will not be repeated here.
[0105] This invention provides a device for generating intravenous puncture paths. After a data acquisition module acquires multiple consecutive frames of time-series ultrasound images, an object recognition module identifies anatomical targets based on the color and shape features of the ultrasound images. This eliminates the need for manual identification of the ultrasound images, reducing the professional knowledge requirements for ultrasound imaging operators and facilitating the widespread use of ultrasound imaging. Furthermore, it reduces the risk of misjudgment due to insufficient operator experience and skills. Further, after establishing a corresponding feature vector for each anatomical target, the ultrasound image is input into a preset path model. The path model then establishes a intravenous puncture path based on the feature vector and iterates through the path to output the shortest intravenous puncture path. The model determines the shortest intravenous puncture path based on the preset target through data processing and calculation, providing reliable path guidance for medical personnel during puncture procedures. This avoids hospital staff selecting incorrect or inappropriate puncture paths due to insufficient experience or skills, improving the accuracy and safety of intravenous punctures.
[0106] The above description is only a preferred embodiment of the present invention. It should be noted that for those skilled in the art, several improvements and substitutions can be made without departing from the technical principles of the present invention, and these improvements and substitutions should also be considered within the scope of protection of the present invention.
Claims
1. A method for generating a deep vein puncture path, characterized in that, include: Acquire an ultrasound image of the puncture site; wherein the ultrasound image is a sequential ultrasound image of multiple consecutive frames; after acquiring the ultrasound image of the puncture site, the method further includes: acquiring the pixel changes of the ultrasound image in various directions based on the maximum density projection method; fusing the pixel change information of the ultrasound image to generate a first ultrasound image; The anatomical target category is identified based on the color and shape features of the ultrasound image, and a corresponding feature vector is established. Specifically, this includes: obtaining the RGB color features of each anatomical target in the first ultrasound image; obtaining shape feature values in the first ultrasound image based on the Hessian matrix; wherein each shape feature value corresponds to the shape feature orientation of an anatomical target; clustering the shape feature values according to the category of the anatomical target; optimizing the clustering using the maximum inter-class variance method, outputting the category corresponding to each anatomical target, and establishing a corresponding feature vector; wherein each feature vector corresponds to one anatomical target. The ultrasound image with established feature vectors is input into a preset path model, so that the path model establishes a venous puncture path based on each feature vector and performs a traversal query to output the shortest venous puncture path; wherein, before inputting the ultrasound image with established feature vectors into the preset path model, the method further includes: determining the tolerance coefficient and risk coefficient of each feature vector according to clinical requirements and anatomical relationships; establishing the coordinates corresponding to each feature vector in the first ultrasound image; wherein, the coordinates are represented by the following coordinate formula: ； in, Let X be the coordinate position of the feature vector at the puncture site; The coordinate position is The red, green, and blue color characteristic values; The category of the anatomical target; The tolerance coefficient value of the anatomical target; The risk coefficient value of the anatomical target; the path model establishes a venous puncture path for each feature vector according to a preset objective function, wherein the objective function is specifically: ； Where d is the Euclidean distance corresponding to each of the aforementioned feature vectors; For each of the described anatomical targets, a risk coefficient value is provided. For each of the described anatomical targets, the tolerance coefficient value is given.
2. The method for generating a deep vein puncture path as described in claim 1, characterized in that, The path model establishes a vein puncture path based on each feature vector, and after traversing and querying to output the shortest vein puncture path, it also includes obtaining the actual needle insertion point on the human body surface, specifically: ； in, The actual needle insertion point on the human body surface; The width of the ultrasonic probe; The coordinate width value of the first ultrasound image; The needle insertion point is the point along the shortest venous puncture path.
3. A deep vein puncture path generation device, characterized in that, It includes a data acquisition module, an object recognition module, and a path generation module, specifically: The data acquisition module is used to acquire ultrasound images of the puncture site; wherein, the ultrasound images are sequential ultrasound images of multiple consecutive frames; after acquiring the ultrasound images of the puncture site, the module further includes: acquiring the pixel changes of the ultrasound images in various directions based on the maximum density projection method; fusing the pixel change information of the ultrasound images to generate a first ultrasound image; The object recognition module is used to identify the category of anatomical targets based on the color and shape features of the ultrasound image, and to establish corresponding feature vectors. Specifically, it includes: obtaining the RGB color feature values of each anatomical target in the first ultrasound image; obtaining the shape feature values in the first ultrasound image based on the Hessian matrix; wherein each shape feature value corresponds to the shape feature orientation of an anatomical target; clustering the shape feature values according to the category of the anatomical target; optimizing the clustering using the maximum inter-class variance method, outputting the category corresponding to each anatomical target, and establishing a corresponding feature vector; wherein each feature vector corresponds to an anatomical target. The path generation module is used to input the ultrasound image after establishing feature vectors into a preset path model, so that the path model establishes a venous puncture path based on each feature vector and performs a traversal query to output the shortest venous puncture path; wherein, before inputting the ultrasound image after establishing feature vectors into the preset path model, the module further includes: determining the tolerance coefficient and risk coefficient of each feature vector according to clinical requirements and anatomical relationships; establishing the coordinates corresponding to each feature vector in the first ultrasound image; wherein, the coordinates are represented by the following coordinate formula: ； in, Let X be the coordinate position of the feature vector at the puncture site; The coordinate position is The red, green, and blue color characteristic values; The category of the anatomical target; The tolerance coefficient value of the anatomical target; The risk coefficient value of the anatomical target; the path model establishes a venous puncture path for each feature vector according to a preset objective function, wherein the objective function is specifically: ； Where d is the Euclidean distance corresponding to each of the aforementioned feature vectors; For each of the described anatomical targets, a risk coefficient value is provided. For each of the described anatomical targets, the tolerance coefficient value is given.
4. A terminal device, characterized in that, The device includes a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor, wherein the processor, when executing the computer program, implements the deep vein puncture path generation method as described in any one of claims 1 to 2.
5. A computer-readable storage medium, characterized in that, The computer-readable storage medium includes a stored computer program, wherein, when the computer program is executed, it controls the device containing the computer-readable storage medium to perform the deep vein puncture path generation method as described in any one of claims 1 to 2.