Anesthesia puncture auxiliary positioning method and system based on image technology

Abstract

The invention provides an anesthesia puncture auxiliary positioning method and system based on an image technology, and relates to the technical field of image processing, and the method comprises the steps: obtaining an ultrasonic image of a patient; the ultrasonic image is preprocessed, and preprocessing comprises noise reduction, contrast enhancement and image segmentation; carrying out feature extraction on the preprocessed ultrasonic image through Hadamard transformation, and determining direction features; performing feature extraction on the preprocessed ultrasonic image through a convolutional neural network to determine spatial features; fusing the direction features and the spatial features to determine fusion features; and determining the target position of the needle head for anesthesia puncture according to the fusion characteristics. According to the invention, the operation efficiency in the anesthesia process is obviously improved.

Description

Technical Field

[0001] The present invention relates to the technical field of image processing, and particularly to an anesthesia puncture assistance positioning method and system based on image technology. Background Art

[0002] The anesthesia puncture assistance positioning method based on image technology is a technical method that uses medical image processing and artificial intelligence technology to help doctors accurately locate the puncture point of the anesthesia needle. This method analyzes medical image data such as ultrasonic images, combines computer vision and deep learning technology, extracts effective features from the images, and determines the target position of the needle, thereby improving the success rate, accuracy, and safety of anesthesia puncture.

[0003] Problems such as the differences in patient anatomical structures, image noise, and limited field of view increase the difficulty and failure rate of puncture, and may even cause serious complications. However, through feature extraction, image segmentation, and target positioning using ultrasonic images and deep learning technology, not only can the success rate of puncture be effectively improved, but also the risk of complications can be significantly reduced, thus meeting the requirements of modern medicine for safety, efficiency, and accuracy.

[0004] However, traditional anesthesia puncture methods usually rely on doctors' experience and touch, which easily lead to puncture failure. Speckle noise commonly exists in ultrasonic images, which can cover up key anatomical landmarks, interfere with the recognition and positioning of the target area, increase the difficulty of target positioning, and make it difficult to achieve high-precision target positioning. Summary of the Invention

[0005] In order to solve the technical problems that traditional anesthesia puncture methods usually rely on doctors' experience and touch, which easily lead to puncture failure, speckle noise commonly exists in ultrasonic images, which can cover up key anatomical landmarks, interfere with the recognition and positioning of the target area, increase the difficulty of target positioning, and make it difficult to achieve high-precision target positioning, the present invention provides an anesthesia puncture assistance positioning method and system based on image technology.

[0006] The technical solutions provided by the embodiments of the present invention are as follows:

[0007] First aspect:

[0008] An anesthesia puncture assistance positioning method based on image technology provided by an embodiment of the present invention includes:

[0009] S1: Obtain the ultrasonic image of the patient;

[0010] S2: Preprocess the ultrasonic image, where the preprocessing includes noise reduction, contrast enhancement, and image segmentation;

[0011] S3: Extract features from the preprocessed ultrasound image through Hadamard transform to determine the directional features;

[0012] S4: Extract features from the preprocessed ultrasound image through a convolutional neural network to determine the spatial features;

[0013] S5: Fuse the directional features and spatial features to determine the fused features;

[0014] S6: Determine the target position of the needle for anesthesia puncture based on the fused features.

[0015] Second aspect:

[0016] An anesthesia puncture assistance positioning system based on image technology provided by an embodiment of the present invention includes:

[0017] A processor;

[0018] A memory, on which computer-readable instructions are stored. When the computer-readable instructions are executed by the processor, the anesthesia puncture assistance positioning method based on image technology as in the first aspect is implemented.

[0019] Third aspect:

[0020] A computer-readable storage medium provided by an embodiment of the present invention, on which a computer program is stored. When the program is executed by the processor, the anesthesia puncture assistance positioning method based on image technology as in the first aspect is implemented.

[0021] The beneficial effects brought by the technical solutions provided by the embodiments of the present invention at least include:

[0022] In the embodiments of the present invention, by preprocessing the ultrasound image, the noise is effectively reduced and the contrast of the ultrasound image is enhanced, improving the visualization of the target anatomical structure. Combining Hadamard transform and convolutional neural network to extract image features realizes the comprehensive extraction of the directional information and local spatial dependence of the ultrasound image, improves the expression ability of the features, avoids subjective dependence, fuses the directional features and spatial features, effectively integrates the information in the sequence domain and spatial domain, generates high-dimensional fused features, improves the integrity and accuracy of the image information, overcomes the limitations of a single feature extraction method, and significantly improves the operation efficiency during anesthesia. Description of the Drawings

[0023] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following will briefly introduce the drawings required for the description of the embodiments. Obviously, the following drawings are only some embodiments of the present invention. For those of ordinary skill in the art, without creative efforts, other drawings can be obtained based on these drawings.

[0024] Figure 1 It is a schematic flowchart of an anesthesia puncture assisted positioning method based on image technology provided by an embodiment of the present invention;

[0025] Figure 2 It is a schematic structural diagram of an anesthesia puncture assisted positioning system based on image technology provided by an embodiment of the present invention. Specific embodiments

[0026] Next, the technical solutions in the present invention will be described with reference to the accompanying drawings.

[0027] Referring to the attached Figure 1 figures, it shows a schematic flowchart of an anesthesia puncture assisted positioning method based on image technology provided by an embodiment of the present invention.

[0028] An embodiment of the present invention provides an anesthesia puncture assisted positioning method based on image technology. This method can be implemented by an anesthesia puncture assisted positioning device based on image technology, and this anesthesia puncture assisted positioning device based on image technology can be a terminal or a server. The processing flow of the anesthesia puncture assisted positioning method based on image technology can include the following steps:

[0029] S1: Obtain the ultrasonic image of the patient.

[0030] Among them, the ultrasonic image is image data obtained by an ultrasonic device using the propagation and reflection of sound waves in human tissues, and is commonly used in medical diagnosis to display the structural characteristics of soft tissues and organs.

[0031] Specifically, obtaining the real-time image of the patient through an ultrasonic device can clearly display the position and characteristics of human soft tissues and target anatomical structures under the conditions of non-invasive and non-radiative.

[0032] S2: Preprocess the ultrasonic image, where the preprocessing includes noise reduction, contrast enhancement, and image segmentation.

[0033] Among them, noise reduction refers to removing the speckle noise or other random interference signals in the ultrasonic image while retaining the key anatomical details in the image. Contrast enhancement refers to improving the contrast between the target area and the background by adjusting the brightness and darkness differences of the image, making the key features more obvious. Image segmentation refers to dividing the image into regions and backgrounds with specific attributes and extracting the region of interest.

[0034] It should be noted that through preprocessing, the noise in the ultrasonic image is effectively removed, improving the image quality. Contrast enhancement highlights the key anatomical structures, facilitating doctors' observation. Image segmentation is a technology in computer vision used to divide an image into several meaningful regions for independent processing or analysis of different regions.

[0035] In a possible implementation, S2 specifically includes:

[0036] S201: Denoise the ultrasound image through anisotropic diffusion filtering.

[0037] Among them, anisotropic diffusion filtering is an image noise reduction technology that controls the diffusion process to ensure that the smooth areas of the image are not disturbed and the details of the edge areas remain clear.

[0038] It should be noted that the anisotropic diffusion filter can effectively remove random noise in the image, retain important edges and details in the image, and avoid edge blurring.

[0039] S202: Performing contrast enhancement on the denoised ultrasound image through morphological grayscale transformation to obtain a high-contrast image.

[0040] Among them, morphological grayscale transformation is a nonlinear operation in image processing. It uses structural elements to corrode, dilate, open, close, etc. the image to improve the shape characteristics of the image. In grayscale image processing, morphological operations can enhance the contrast in the image and highlight the details, especially tiny structures and edges.

[0041] It should be noted that morphological grayscale transformation can effectively enhance the details and contrast in ultrasound images, highlight the boundaries and structures of target objects, and enhance the recognizability of images.

[0042] S203: Perform image segmentation on the high contrast image using an adaptive threshold binarization algorithm.

[0043] Among them, the adaptive threshold binarization algorithm is an image segmentation method that dynamically determines the threshold based on the grayscale information of the local area of ​​the image. It can use different thresholds in different parts of the image and effectively segment objects in images with uneven lighting or local contrast differences.

[0044] It should be noted that the adaptive threshold binarization algorithm can adjust the segmentation criteria according to the local characteristics of different regions to improve the robustness of segmentation, which is particularly suitable for medical image processing with complex backgrounds and more details.

[0045] In a possible implementation, S201 specifically includes:

[0046] S2011: Perform Gaussian convolution on the ultrasound image to obtain a multi-scale grayscale image:

[0047] I(x,y,t)=I0(x,y)*G(x,y,t)

[0048] Among them, I(x, y, t) represents the multi-scale grayscale image at time t, I0(x, y) represents the ultrasound image, and G(x, y, t) represents the Gaussian kernel function.

[0049] Among them, Gaussian convolution refers to using the Gaussian kernel function to perform a smoothing operation on the image to remove fine noise and generate grayscale images of different scales simultaneously.

[0050] S2012: Determine the anisotropic diffusion filtering of the multi-scale grayscale image:

[0051]

[0052] Among them, I t represents the change rate of the multi-scale grayscale image at time t, div represents, c(x, y, t) represents the diffusion coefficient function, represents the gradient of the multi-scale grayscale image, e represents the exponential function, f represents the coefficient controlling the diffusion behavior, and S represents the smoothing parameter.

[0053] Specifically, anisotropic diffusion filtering refers to smoothing the image through a diffusion equation, where the degree of diffusion is controlled by the pixel gradient and the diffusion coefficient function. It can smooth the region while protecting the edges. The diffusion coefficient function can define the diffusion intensity, which is related to the magnitude of the image gradient and restricts the diffusion of the edges to prevent edge blurring.

[0054] S2013: Denoise the ultrasound image by adjusting the diffusion coefficient function of the anisotropic diffusion filtering.

[0055] It should be noted that through the combination of Gaussian convolution and anisotropic diffusion filtering, the edge features and anatomical details of the ultrasound image are retained while denoising. Gaussian convolution can effectively smooth the image and generate multi-scale information. Anisotropic diffusion filtering adaptively adjusts the diffusion intensity according to the image gradient, eliminates speckle noise while protecting the edge information, and provides high-quality image data for subsequent target localization.

[0056] In a possible implementation manner, S202 specifically includes:

[0057] S2021: Create a disk-shaped structuring element through morphological grayscale transformation.

[0058] Among them, the disk-shaped structuring element is a template for morphological operations, with a disk shape, defining the range and mode of image transformation.

[0059] S2022: Perform top-hat transformation and bottom-hat transformation on the denoised ultrasound image according to the disk-shaped structuring element to determine the top-hat image and the bottom-hat image.

[0060] Among them, the Top-Hat Transform and the Bottom-Hat Transform are two common operations in morphological image processing. The Top-Hat Transform is used to extract bright regions smaller than the structuring element from the original image, while the Bottom-Hat Transform is used to extract dark regions larger than the structuring element.

[0061] S2023: Perform opening and closing operations on the top-hat image and the bottom-hat image respectively to obtain an opened image and a closed image.

[0062] Among them, the opening operation refers to removing small bright spots or noises in the image through dilation and erosion operations, and the closing operation refers to filling small dark spots or gaps in the image through erosion and dilation operations.

[0063] It should be noted that the Top-Hat Transform and the Bottom-Hat Transform can reveal small-scale features and local structures in the image, effectively separate the target region from the background, and improve the image contrast.

[0064] S2024: Combine the top-hat image, the bottom-hat image, the opened image, and the closed image to determine the bright region and the dark region:

[0065] I b = I t - I c

[0066] I d = I s - I a

[0067] Among them, I b represents the bright region, I t represents the closed image, I c represents the bottom-hat image, I d represents the dark region, I s represents the opened image, I a represents the bottom-hat image.

[0068] Among them, the bright region is the region with higher brightness in the image, usually representing the more prominent feature part in the image, and the dark region is the region with lower gray value in the image, usually representing the darker part in the image.

[0069] S2025: Enhance the contrast of the denoised ultrasonic image according to the bright region and the dark region to obtain a high-contrast image:

[0070] I n = I f - aI d + aI b

[0071] Among them, I nRepresents a high-contrast image, I f Represents the denoised ultrasound image, and a represents the proportional transformation constant.

[0072] Specifically, through morphological gray-scale transformation, combined with top-hat transformation and bottom-hat transformation, the bright and dark regions in the image can be effectively distinguished. The opening operation and closing operation further clean up the noise, improve the regional continuity and detail clarity of the image. Finally, the contrast is enhanced by combining the bright and dark regions, highlighting the key anatomical structures.

[0073] In a possible implementation manner, S203 specifically includes:

[0074] S2031: Calculate the binarization threshold of the high-contrast image according to the gray value of the pixel:

[0075]

[0076] δ(x,y) = I n (x,y) - m(x,y)

[0077] Wherein, T(x,y) represents the binarization threshold, m(x,y) represents the local average value, I n (x,y) represents the gray value of the high-contrast image at the pixel point (x,y), k represents the deviation controlling the threshold change, and δ(x,y) represents the local average deviation.

[0078] Specifically, I n (x,y) represents the gray value of the high-contrast image at the pixel point (x,y). The gray value of the pixel reflects the light and dark contrast of the image. δ(x,y) represents the deviation between the current pixel value and the average value of its neighborhood, reflecting the difference degree between the local pixel and the background. m(x,y) represents the local average value reflecting the overall brightness level of the local area. k is the deviation coefficient controlling the threshold change, and its value range is [0,1]. The smaller the k value, the larger the threshold, which is suitable for enhancing the low-brightness area. The larger the k value, the smaller the threshold, which is suitable for enhancing the high-brightness area.

[0079] S2032: Perform image segmentation on the high-contrast image according to the binarization threshold:

[0080]

[0081] Wherein, d(x,y) represents the segmentation result of the high-contrast image, that is, the gray value of the preprocessed ultrasound image at the pixel point (x,y).

[0082] Wherein, for each pixel point, the threshold is dynamically adjusted according to the local brightness distribution of its neighborhood, so that a higher threshold is used in the area with high brightness and a lower threshold is used in the area with low brightness, thereby realizing segmentation adapting to different regional characteristics.

[0083] It should be noted that by using the binarization threshold calculated adaptively for the local area, it is possible to flexibly handle the situation of uneven illumination or complex details in the image. By dynamically adjusting the threshold through the local average and deviation, the target and background in the image can be accurately segmented, avoiding the limitations of the global threshold method in an uneven illumination environment.

[0084] S3: Perform feature extraction on the preprocessed ultrasonic image through Hadamard transform to determine the directional features.

[0085] Among them, the Hadamard transform is a matrix transform. The ultrasonic image is transformed through the Hadamard matrix to extract the directional features in the image.

[0086] It should be noted that by performing feature extraction on the preprocessed ultrasonic image through Hadamard transform, the local directional characteristics of the image can be efficiently decomposed, the directional information can be enhanced, and the target area can be accurately identified.

[0087] In a possible implementation manner, S3 specifically includes:

[0088] S301: Perform Hadamard transform on the preprocessed ultrasonic image to obtain the Hadamard transform coefficients:

[0089]

[0090] Among them, W(u, v) represents the Hadamard transform coefficient in the transform domain at (u, v), d(x, y) represents the gray value of the preprocessed ultrasonic image at the pixel point (x, y), x = 0, 1,..., N - 1, y = 0, 1,..., N - 1, N represents the size of the image, ψ(u, v, x, y) represents the transform kernel function, g k (x) represents the k-th bit of the binary representation of x, g k (u) represents the k-th bit of the binary representation of u, g k (y) represents the k-th bit of the binary representation of y, g k (v) represents the k-th bit of the binary representation of v, k = 0, 1,..., l - 1, l represents the total number of bits of the binary representation of the image size.

[0091] S302: Convert the Hadamard transform coefficients into matrix form to determine the transform coefficient matrix:

[0092]

[0093] Among them, W represents the overall transform coefficient matrix, H Ndenotes the Hadamard matrix, and D denotes the preprocessed ultrasonic image matrix. denotes the transpose of the Hadamard matrix.

[0094] S303: Recursively segment the preprocessed ultrasonic image according to the transformation coefficient matrix to obtain a segmented image.

[0095] Among them, recursive segmentation is a process of layer-by-layer segmentation of an image. Each segmentation is based on the result of the previous layer segmentation. Recursive segmentation extracts features at different levels by continuously dividing the image.

[0096] S304: Perform multi-scale transformation on the segmented image to determine the Hadamard transformation coefficients of the segmented image:

[0097]

[0098] Among them, denotes the Hadamard coefficient of the j-th segmented image, denotes the Hadamard matrix of the j-th segmented image, and P j denotes the j-th segmented image.

[0099] Among them, multi-scale transformation refers to the transformation and processing of an image at different scales (resolutions), which can extract features at different levels and details from the image and capture richer image information.

[0100] S305: According to the Hadamard transformation coefficients of the segmented image, use directional filters to extract the directional features of the segmented image:

[0101]

[0102] Among them, v j denotes the directional feature of the j-th segmented image, Var denotes the statistical variance, Fvt denotes the filter in the vertical direction, Fhr denotes the filter in the horizontal direction, and Fdg denotes the filter in the diagonal direction. denotes the Hadamard product.

[0103] Specifically, by using filters in different directions, directional features such as edges and textures in the image can be identified. The statistical variance is a statistic used to measure the degree of data change, reflecting the degree of dispersion of data points relative to the average value. In image processing, variance is often used to describe the texture features of an image.

[0104] The filter in the vertical direction is specifically:

[0105]

[0106] Among them, u and v respectively represent the horizontal direction index and the vertical direction index of the sequence domain, and α represents the parameter of the filter coverage area percentage.

[0107] The filter in the horizontal direction is specifically:

[0108]

[0109] Among them, T represents transpose.

[0110] The filter in the diagonal direction is specifically:

[0111] F dg = 1 - (F vt + F hr ).

[0112] S4: Through the convolutional neural network, perform feature extraction on the preprocessed ultrasonic image to determine the spatial features.

[0113] Among them, the convolutional neural network (CNN) is a deep learning model suitable for image processing tasks. It extracts local features of the image through convolutional layers and performs dimensionality reduction and feature screening through pooling layers. The spatial features are the features representing anatomical structures or physical properties in the image, used to describe the position, shape, and local texture information of the target area.

[0114] In a possible implementation manner, S4 is specifically:

[0115] Perform multiple convolution operations on the preprocessed ultrasonic image for feature extraction to determine the spatial features:

[0116]

[0117] Among them, Y f represents the f-th spatial feature, down represents the downsampling operation, F represents the activation function, * represents the convolution operator, K f represents the f-th convolution kernel, X i represents the i-th input feature map generated by the previous layer, and b f represents the bias of the f-th spatial feature.

[0118] Among them, the convolution operation is the core operation in deep learning for feature extraction. It slides a convolution kernel over the input data to calculate the weighted sum of the local area, thereby extracting specific features.

[0119] Specifically, by performing multiple convolution operations on the ultrasound image, information from low-level features to high-level features can be extracted layer by layer. Downsampling reduces the redundant information in the feature map and improves the computational efficiency, while the activation function enhances the non-linear expression ability of the model, enabling the network to more accurately capture the features of complex anatomical structures and target regions.

[0120] S5: Fuse the directional feature and the spatial feature to determine the fused feature.

[0121] It should be noted that by fusing the directional feature and the spatial feature, the local directional information of the image is organically combined with the global spatial structure, which can more comprehensively describe the features of the target region, improve the data expression ability, and enhance the accuracy and robustness of target localization.

[0122] In a possible implementation manner, S5 specifically includes:

[0123] S501: Vectorize the directional feature and the spatial feature to obtain a directional feature vector and a spatial feature vector.

[0124] S502: Fuse the vectorized directional feature and spatial feature:

[0125] y = f LDH Ω + vec(Y f )Λ

[0126] Ω = [I (|P|d×|P|d) ∣[0] (|P|d×MNL)

[0127] Λ = [[0] (MNL×|P|d) ∣I (MNL×MNL)

[0128] where y represents the fused feature, Y f represents the f-th spatial feature, vec represents vectorization, vec(Yf) represents the spatial feature vector, f LDH represents the directional feature vector, Ω represents the mapping operation of the directional feature vector, Λ represents the mapping operation of the spatial feature vector, I represents the identity matrix, [0] represents the zero matrix, d represents the dimension of each directional feature, |P| represents the total number of segmented images, and MNL represents the total dimension of the convolutional features.

[0129] In a possible implementation manner, the directional feature vector is specifically:

[0130]

[0131] where f LDH represents the directional feature vector, T i ​​The mapping matrix representing the i-th directional feature, where i = 1, 2, ..., |P|, and |P| represents the total number of segmented images.

[0132] S6: Determine the target position of the needle for anesthetic puncture based on the fused features.

[0133] It should be noted that this step can more accurately locate the target area in complex anatomical structures, avoid puncture failure or misoperation, greatly improve the accuracy and safety of puncture positioning, and provide intelligent auxiliary support for anesthetic operations.

[0134] In a possible implementation manner, S6 is specifically:

[0135] Determine the target position of the needle for anesthetic puncture through the fully connected layer according to the fused features:

[0136] z = ReLU(Wy + b)

[0137] Where z represents the target detection result, that is, the target position of the needle, ReLU represents the ReLU activation function, W represents the weight matrix, y represents the fused features, and b represents the bias term.

[0138] It should be noted that by processing the fused features through the fully connected layer, the expression ability of high-dimensional features can be fully utilized, converting complex directional and spatial information into specific target positions of the needle. The fully connected layer has a strong feature combination ability, can accurately capture non-linear relationships, improve the accuracy and robustness of target position prediction, provide reliable positioning support for anesthetic puncture, and reduce the subjective error of human judgment at the same time.

[0139] The beneficial effects brought by the technical solution provided by the embodiments of the present invention at least include:

[0140] In the embodiments of the present invention, by preprocessing the ultrasonic image, effectively reducing noise and enhancing the contrast of the ultrasonic image, improving the visualization of the target anatomical structure, extracting image features by combining the Hadamard transform and the convolutional neural network, realizing the comprehensive extraction of the directional information and local spatial dependence of the ultrasonic image, improving the expression ability of the features, avoiding subjective dependence, fusing the directional features and spatial features, effectively integrating the information in the sequence domain and the spatial domain, generating high-dimensional fused features, enhancing the integrity and accuracy of the image information, overcoming the limitations of a single feature extraction method, and significantly improving the operation efficiency during anesthesia.

[0141] Refer to the attached Figure 2 , which shows the structural schematic diagram of an anesthetic puncture auxiliary positioning system based on image technology provided by the present invention.

[0142] The present invention also provides an anesthesia puncture assistance positioning system 20 based on image technology, which is applied to the above-mentioned anesthesia puncture assistance positioning method based on image technology, and includes:

[0143] A processor 201.

[0144] A memory 202, on which computer-readable instructions are stored. When the computer-readable instructions are executed by the processor 201, the anesthesia puncture assistance positioning method based on image technology as in the method embodiment is realized.

[0145] The anesthesia puncture assistance positioning system 20 provided by the present invention can execute the above-mentioned anesthesia puncture assistance positioning method based on image technology and achieve the same or similar technical effects. To avoid repetition, the present invention will not elaborate further.

[0146] The beneficial effects brought by the technical solution provided by the embodiment of the present invention at least include:

[0147] In the embodiment of the present invention, by preprocessing the ultrasonic image, the noise is effectively reduced and the contrast of the ultrasonic image is enhanced, improving the visualization of the target anatomical structure. Combining the Hadamard transform and the convolutional neural network to extract the image features realizes the comprehensive extraction of the direction information and local spatial dependence of the ultrasonic image, improves the expression ability of the features, avoids subjective dependence, fuses the direction features and the spatial features, effectively integrates the information in the sequence domain and the spatial domain, generates high-dimensional fusion features, improves the integrity and accuracy of the image information, overcomes the limitations of a single feature extraction method, and significantly improves the operation efficiency during anesthesia.

[0148] It should be understood that the processor in the embodiment of the present invention may be a central processing unit (CPU), and this processor may also be 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 may be a microprocessor or this processor may also be any conventional processor, etc.

[0149] It should also be understood that the memory in the embodiments of the present invention may be a volatile memory or a non-volatile memory, or may include both volatile and non-volatile memories. Among them, the non-volatile memory may be a read-only memory (ROM), a programmable ROM (PROM), an erasable PROM (EPROM), an electrically erasable PROM (EEPROM), or a flash memory. The volatile memory may be a random access memory (RAM), which is used as an external cache. By way of example but not limitation, many forms of random access memory (RAM) are available, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDR SDRAM), enhanced SDRAM (ESDRAM), synchlink DRAM (SLDRAM), and direct rambus RAM (DR RAM).

[0150] The above embodiments can be implemented in whole or in part by software, hardware (such as circuits), firmware, or any combination thereof. When implemented using software, the above embodiments can be implemented in whole or in part in the form of a computer program product. The computer program product includes one or more computer instructions or computer programs. When the computer instructions or computer programs are loaded or executed on a computer, the processes or functions according to the embodiments of the present invention are generated in whole or in part. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable devices. The computer instructions can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another. For example, the computer instructions can be transmitted from one website, computer, server, or data center to another website, computer, server, or data center via wired (such as infrared, wireless, microwave, etc.) means. The computer-readable storage medium can be any available medium that can be accessed by a computer or a data storage device such as a server or a data center that contains one or more collections of available media. The available medium can be a magnetic medium (such as a floppy disk, a hard disk, or a magnetic tape), an optical medium (such as a DVD), or a semiconductor medium. The semiconductor medium can be a solid-state drive.

[0151] It should be understood that the term "and / or" in this document is merely a description of the association relationship between associated objects, indicating that three relationships can exist. For example, A and / or B can represent: A exists alone, A and B exist simultaneously, and B exists alone. Here, A and B can be singular or plural. In addition, the character " / " in this document generally represents an "or" relationship between the associated objects before and after, but it may also represent an "and / or" relationship, which can be specifically understood with reference to the context before and after.

[0152] In the present invention, "at least one" means one or more, and "a plurality" means two or more. "At least one of the following" or its similar expressions refer to any combination of these items, including any combination of single items or plural items. For example, at least one of a, b, or c can represent: a, b, c, a - b, a - c, b - c, or a - b - c, where a, b, and c can be single or multiple.

[0153] It should be understood that in various embodiments of the present invention, the magnitudes of the sequence numbers of the above processes do not mean the order of execution. The order of execution of each process should be determined by its function and internal logic, and should not constitute any limitation to the implementation process of the embodiments of the present invention.

[0154] Those of ordinary skill in the art will realize that the units and algorithm steps of the examples described in connection with the embodiments disclosed herein can be implemented in electronic hardware, or in a combination of computer software and electronic hardware. Whether these functions are executed in hardware or software depends on the specific application and design constraints of the technical solution. A professional technician can use different methods to implement the described functions for each specific application, but such implementation should not be considered to exceed the scope of the present invention.

[0155] Those skilled in the art can clearly understand that for the convenience and brevity of description, the specific working processes of the devices, apparatuses, and units described above can refer to the corresponding processes in the foregoing method embodiments, and will not be elaborated herein.

[0156] In several embodiments provided by the present invention, it should be understood that the disclosed devices, apparatuses, and methods can be implemented in other ways. For example, the device embodiments described above are merely illustrative. For example, the division of units is only a logical function division. In actual implementation, there may be other division methods. For example, multiple units or components can be combined or integrated into another device, or some features can be ignored or not executed. Another point is that the displayed or discussed couplings or direct couplings or communication connections to each other can be through some interfaces. The indirect couplings or communication connections of devices or units can be electrical, mechanical, or other forms.

[0157] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place, or may be distributed to multiple network units. Some or all of the units can be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0158] In addition, the functional units in various embodiments of the present invention can be integrated in a processing unit, or each unit can exist physically alone, or two or more units can be integrated in one unit.

[0159] If a function is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or a part of this technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions for causing a computer device (which may be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the methods in various embodiments of the present invention. The aforementioned storage medium includes: various media such as USB flash drives, mobile hard disks, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical discs that can store program codes.

[0160] An embodiment of the present invention provides a computer-readable storage medium, on which a computer program is stored. When the program is executed by a processor, it implements the anesthesia puncture assistance positioning method based on image technology as in the method embodiment.

[0161] The computer-readable storage medium provided by the present invention can implement the steps and effects of the anesthesia puncture assistance positioning method based on image technology in the above method embodiment. To avoid repetition, the present invention will not elaborate further.

[0162] The beneficial effects brought by the technical solution provided by the embodiment of the present invention at least include:

[0163] In the embodiment of the present invention, by preprocessing the ultrasonic image, the noise is effectively reduced and the contrast of the ultrasonic image is enhanced, improving the visualization of the target anatomical structure. Combining the Hadamard transform and the convolutional neural network to extract image features realizes the comprehensive extraction of the direction information and local spatial dependence of the ultrasonic image, improves the expression ability of the features, avoids subjective dependence, fuses the direction features and spatial features, effectively integrates the information in the sequence domain and the spatial domain, generates high-dimensional fusion features, improves the integrity and accuracy of the image information, overcomes the limitations of a single feature extraction method, and significantly improves the operation efficiency during anesthesia.

[0164] The above is only the specific implementation manner of the present invention, but the protection scope of the present invention is not limited thereto. Any person skilled in the art within the technical scope disclosed by the present invention can easily think of changes or substitutions, which should all be covered within the protection scope of the present invention. Therefore, the protection scope of the present invention should be subject to the protection scope of the claims.

[0165] The following points need to be explained:

[0166] (1) The accompanying drawings of the embodiments of the present invention only relate to the structures involved in the embodiments of the present invention, and other structures can refer to the usual designs.

[0167] (2) For clarity, in the accompanying drawings used to describe the embodiments of the present invention, the thickness of layers or regions is enlarged or reduced, that is, these drawings are not drawn to actual scale. It can be understood that when an element such as a layer, film, region or substrate is referred to as being "on" or "under" another element, the element can be "directly" on or under the other element or there can be intervening elements.

[0168] (3) Without conflict, the embodiments of the present invention and the features in the embodiments can be combined with each other to obtain new embodiments.

[0169] The above is only the specific implementation manner of the present invention, but the protection scope of the present invention is not limited thereto. The protection scope of the present invention shall be subject to the protection scope of the claims.

Claims

1. An image-based anesthesia puncture auxiliary positioning method, characterized in that: include: S1: Acquire ultrasound images of the patient; S2: preprocessing the ultrasound image, wherein the preprocessing includes noise reduction, contrast enhancement and image segmentation; S3: extracting features from the preprocessed ultrasound image through Hadamard transform to determine directional features; S4: extracting features from the preprocessed ultrasound images through a convolutional neural network to determine spatial features; S5: fusing the directional feature and the spatial feature to determine a fusion feature; S6: Determine the target position of the needle for anesthesia puncture according to the fusion feature.

2. The anesthesia puncture auxiliary positioning method based on image technology according to claim 1 is characterized in that: The S2 specifically includes: S201: reducing noise on the ultrasound image by anisotropic diffusion filtering; S202: performing contrast enhancement on the denoised ultrasound image through morphological grayscale transformation to obtain a high-contrast image; S203: Perform image segmentation on the high contrast image using an adaptive threshold binarization algorithm.

3. The anesthesia puncture auxiliary positioning method based on image technology according to claim 2 is characterized in that: The S201 specifically includes: S2011: performing Gaussian convolution on the ultrasound image to obtain a multi-scale grayscale image; S2012: Determine anisotropic diffusion filtering of the multi-scale grayscale image; S2013: Denoising the ultrasound image by adjusting the diffusion coefficient function of the anisotropic diffusion filter.

4. The image-based anesthesia puncture auxiliary positioning method according to claim 2, characterized in that: The S202 specifically includes: S2021: creating a disk structure element through the morphological grayscale transformation; S2022: performing a top hat transformation and a bottom hat transformation on the denoised ultrasonic image according to the disk structure element to determine a top hat image and a bottom hat image; S2023: performing an opening operation and a closing operation on the top hat image and the bottom hat image respectively to obtain an opening operation image and a closing operation image; S2024: Determine a bright area and a dark area by combining the top hat image, the bottom hat image, the opening operation image, and the closing operation image; S2025: According to the bright area and the dark area, contrast enhancement is performed on the denoised ultrasound image to obtain the high-contrast image.

5. The anesthesia puncture auxiliary positioning method based on image technology according to claim 2 is characterized in that: The S203 specifically includes: S2031: Calculating a binarization threshold of the high contrast image according to the grayscale value of the pixel point; S2032: Perform image segmentation on the high contrast image according to the binarization threshold.

6. The anesthesia puncture auxiliary positioning method based on image technology according to claim 1 is characterized in that: The S3 specifically includes: S301: Performing Hadamard transform on the preprocessed ultrasound image to obtain Hadamard transform coefficients; S302: Convert the Hadamard transform coefficients into a matrix form to determine a transform coefficient matrix; S303: recursively segmenting the preprocessed ultrasound image according to the transformation coefficient matrix to obtain a segmented image; S304: performing a multi-scale transformation on the segmented image to determine a Hadamard transformation coefficient of the segmented image; S305: Extracting directional features of the segmented image using a directional filter according to the Hadamard transform coefficients of the segmented image.

7. The anesthesia puncture auxiliary positioning method based on image technology according to claim 1 is characterized in that: The S4 is specifically: The preprocessed ultrasound images are subjected to multiple convolution operations to extract features and determine spatial features.

8. The image-based anesthesia puncture auxiliary positioning method according to claim 1, characterized in that: The S5 specifically includes: S501: vectorizing the directional feature and the spatial feature to obtain a directional feature vector and a spatial feature vector; S502: Fusing the vectorized directional features and spatial features.

9. The image-based anesthesia puncture auxiliary positioning method according to claim 8, characterized in that: The S6 is specifically: Through the fully connected layer, the target position of the needle for anesthesia puncture is determined according to the fusion features.

10. An anesthesia puncture auxiliary positioning system based on image technology, characterized in that: include: processor; A memory having computer-readable instructions stored thereon, wherein when the computer-readable instructions are executed by the processor, the anesthesia puncture auxiliary positioning method based on image technology as described in any one of claims 1 to 9 is implemented.

Images