An intelligent puncture positioning method and device for anterior talofibular ligament lesion based on ultrasound imaging

Abstract

The present invention discloses a method and device for intelligent puncture localization of lesions in the anterior talofibular ligament using ultrasound images. The method comprises: obtaining an ultrasound image dataset of the anterior talofibular ligament; the ultrasound image dataset of the anterior talofibular ligament includes N ultrasound image data of the anterior talofibular ligament, where N is a positive integer; preprocessing the ultrasound image data of the anterior talofibular ligament to obtain preprocessed ultrasound image data; segmenting the preprocessed ultrasound image data to obtain segmented ultrasound image data; performing data enhancement on the segmented ultrasound image data to obtain enhanced ultrasound image data; processing the enhanced ultrasound image data using a three-dimensional reconstruction model to obtain a three-dimensional reconstruction model; and processing the three-dimensional reconstruction model to obtain puncture path planning information. The present invention achieves accurate identification of the anterior talofibular ligament and improves the accuracy of puncture localization.

Description

Technical Field

[0001] The present invention relates to the technical field of medical image processing, and in particular to a method and device for intelligent puncture positioning of lesions of the anterior talofibular ligament using ultrasonic images. Background Art

[0002] The anterior talofibular ligament (ATFL) is one of the most important ligaments on the lateral side of the ankle, playing a key role in maintaining ankle stability. ATFL injuries are the most common type of ankle sprain, particularly in sports injuries. Accurate diagnosis and localization of ATFL injuries are crucial for developing effective treatment plans.

[0003] Although ultrasound imaging technology has the advantages of non-invasiveness, real-time, and dynamic observation in the diagnosis of musculoskeletal diseases, the following difficulties still exist in actual clinical application, especially in scanning the anterior talofibular ligament:

[0004] 1. Complex anatomical structure: The anterior talofibular ligament is located on the outside of the ankle joint. The surrounding anatomical structures are complex, including the fibula, talus, joint capsule, etc. These structures are prone to produce artifacts in ultrasound images, affecting the clear display of the ligament.

[0005] 2. Individual differences among patients: The anatomical structures of different patients vary. For example, obese patients have a thicker subcutaneous fat layer, which further increases the difficulty of obtaining ultrasound images.

[0006] 3. High technical requirements for operation: Ultrasonic scanning requires the operator to have a high level of technical skills, including accurate placement of the probe and precise control of the scanning angle, which places high demands on the operator's experience and skills.

[0007] For clinicians who are not specialized in ultrasound, accurate identification of the anterior talofibular ligament is difficult. The main reasons include:

[0008] 1. Insufficient professional knowledge: Non-ultrasound specialists lack systematic ultrasound imaging training and find it difficult to accurately identify complex anatomical structures.

[0009] 2. Difficulty in image interpretation: The dynamic and real-time nature of ultrasound images increases the difficulty of image interpretation. It is difficult for non-professional physicians to accurately judge the status of the ligament in a short period of time.

[0010] 3. Lack of experience accumulation: The interpretation of ultrasound images requires a lot of practical experience accumulation, and non-professional physicians have obvious shortcomings in this regard.

[0011] The learning curve for musculoskeletal ultrasound is relatively steep, mainly reflected in the following aspects:

[0012] 1. High technical requirements: Ultrasound scanning requires mastery of multiple techniques such as probe operation, image acquisition, and image interpretation.

[0013] 2. Complex knowledge system: It is necessary to master multidisciplinary knowledge such as anatomy, pathology, and ultrasound imaging.

[0014] 3. Limited practice opportunities: Ultrasound scanning requires a lot of practice opportunities, but the actual clinical operation opportunities are limited, which restricts the physician's learning and experience accumulation.

[0015] Therefore, it is of great significance to study an intelligent puncture positioning method for lesions of the anterior talofibular ligament based on ultrasound imaging. Summary of the Invention

[0016] The technical problem to be solved by the present invention is to provide a method and device for intelligent puncture positioning of lesions in the anterior talofibular ligament based on ultrasound imaging, use deep learning methods to achieve accurate identification of the anterior talofibular ligament, and plan the syringe puncture path through an intelligent navigation system to improve puncture accuracy and safety.

[0017] In order to solve the above technical problems, the first aspect of the embodiments of the present invention discloses a method for intelligent puncture location of lesions in the anterior talofibular ligament using ultrasound imaging, the method comprising:

[0018] S1, obtaining an ultrasound image dataset of the anterior talofibular ligament; the ultrasound image dataset of the anterior talofibular ligament includes N ultrasound image data of the anterior talofibular ligament, where N is a positive integer;

[0019] S2, preprocessing the ultrasound image data of the anterior talofibular ligament to obtain preprocessed ultrasound image data;

[0020] S3, segmenting the preprocessed ultrasonic image data to obtain segmented ultrasonic image data;

[0021] S4, performing data enhancement on the segmented ultrasound image data to obtain enhanced ultrasound image data;

[0022] S5, processing the enhanced ultrasound image data using the three-dimensional reconstruction model to obtain a three-dimensional reconstruction model;

[0023] S6, processing the three-dimensional reconstructed model to obtain puncture path planning information.

[0024] As an optional implementation manner, in the first aspect of the embodiment of the present invention, preprocessing the ultrasound image data of the anterior talofibular ligament to obtain preprocessed ultrasound image data includes:

[0025] S21, labeling the ultrasound image data of the anterior talofibular ligament to obtain label information data;

[0026] S22, performing denoising processing on the ultrasound image data of the anterior talofibular ligament to obtain denoised ultrasound image data;

[0027] S23, performing contrast enhancement processing on the denoised ultrasonic image data to obtain pre-processed anterior talofibular acoustic image data.

[0028] As an optional implementation, in the first aspect of the embodiment of the present invention, the denoising process is performed on the ultrasound image data of the anterior talofibular ligament to obtain the denoised ultrasound image data, including:

[0029] S221, processing the ultrasound image data of the anterior talofibular ligament using a first feature extraction model to obtain first feature information;

[0030] S222: downsample the first feature information to obtain second feature information;

[0031] S223, using a second feature extraction model, processing the second feature information to obtain third feature information;

[0032] S224, upsampling the third feature information to obtain fourth feature information;

[0033] S225, using a third feature extraction model to process the fourth feature information to obtain fifth feature information;

[0034] S226, using a convolution model to process the fifth feature information to obtain sixth feature information;

[0035] S227: Process the sixth characteristic information to obtain denoised ultrasound image data.

[0036] As an optional implementation manner, in the first aspect of the embodiment of the present invention, performing contrast enhancement processing on the denoised ultrasound image data to obtain preprocessed pre-talofibular acoustic image data includes:

[0037] S231, decomposing the denoised ultrasonic image data to obtain base layer data and detail layer data;

[0038] S232, performing illumination equalization processing on the base layer data to obtain enhanced base layer data;

[0039] S233, performing enhancement processing on the detail layer data to obtain enhanced detail layer data;

[0040] S234, synthesizing the enhanced base layer data and the enhanced detail layer data to obtain pre-processed anterior talofibular acoustic image data.

[0041] As an optional implementation manner, in the first aspect of the embodiment of the present invention, performing data enhancement on the segmented ultrasound image data to obtain enhanced ultrasound image data includes:

[0042] S41, using a preset generator network to process the label information data and random noise to obtain a generated image;

[0043] S42, processing the segmented ultrasound image data and the generated image to obtain a cyclic correlation coefficient;

[0044] S43, using the cyclic correlation coefficient, performing adversarial training on the preset discriminator network and the preset generator network to obtain an optimized discriminator network and an optimized generator network;

[0045] S44, using the optimized generator network, processing the label information data and random noise to obtain an amplified image;

[0046] S45 , integrating the amplified image and the segmented ultrasonic image data to obtain enhanced ultrasonic image data.

[0047] As an optional implementation manner, in the first aspect of the embodiment of the present invention, the processing of the enhanced ultrasound image data using the three-dimensional reconstruction model to obtain the three-dimensional reconstruction model includes:

[0048] S51, using a three-dimensional reconstruction model, performing three-dimensional reconstruction on the enhanced ultrasound image data to obtain three-dimensional data information;

[0049] S52: Visualize the three-dimensional data information to obtain a three-dimensional reconstructed model.

[0050] As an optional implementation, in the first aspect of the embodiment of the present invention, processing the three-dimensional reconstructed model to obtain puncture path planning information includes:

[0051] S61, analyzing the three-dimensional reconstructed model to obtain starting point coordinate information and ending point coordinate information;

[0052] S62, setting the cost function and constraints of the path planning problem;

[0053] S63, using Logistic chaos mapping to generate the initial population of particle swarm;

[0054] S64, using the artificial bee colony algorithm to perform iterative search and update the individual position to move it towards a lower fitness value;

[0055] S65, using chaotic mapping to search the local solution space to make it jump out of the local optimal value;

[0056] S66, repeat S64 and S65 until the maximum number of iterations is reached or the fitness value change threshold is reached;

[0057] S67: Output the optimal fitness value and obtain puncture path planning information.

[0058] A second aspect of an embodiment of the present invention discloses an intelligent puncture and positioning device for lesions of the anterior talofibular ligament using ultrasound imaging, the device comprising:

[0059] a data acquisition module, configured to acquire an ultrasound image dataset of the anterior talofibular ligament; the ultrasound image dataset of the anterior talofibular ligament comprising N ultrasound image data of the anterior talofibular ligament, where N is a positive integer;

[0060] a preprocessing module, configured to preprocess the ultrasound image data of the anterior talofibular ligament to obtain preprocessed ultrasound image data;

[0061] an image segmentation module, configured to segment the preprocessed ultrasound image data to obtain segmented ultrasound image data;

[0062] a data enhancement module, configured to perform data enhancement on the segmented ultrasound image data to obtain enhanced ultrasound image data;

[0063] a three-dimensional reconstruction module, configured to process the enhanced ultrasound image data using a three-dimensional reconstruction model to obtain a three-dimensional reconstruction model;

[0064] The path planning module is used to process the three-dimensional reconstructed model to obtain puncture path planning information.

[0065] As an optional implementation, in the second aspect of the embodiment of the present invention, preprocessing the ultrasound image data of the anterior talofibular ligament to obtain preprocessed ultrasound image data includes:

[0066] S21, labeling the ultrasound image data of the anterior talofibular ligament to obtain label information data;

[0067] S22, performing denoising processing on the ultrasound image data of the anterior talofibular ligament to obtain denoised ultrasound image data;

[0068] S23, performing contrast enhancement processing on the denoised ultrasonic image data to obtain pre-processed anterior talofibular acoustic image data.

[0069] As an optional implementation, in the second aspect of the embodiment of the present invention, the denoising process is performed on the ultrasound image data of the anterior talofibular ligament to obtain the denoised ultrasound image data, including:

[0070] S221, processing the ultrasound image data of the anterior talofibular ligament using a first feature extraction model to obtain first feature information;

[0071] S222: downsample the first feature information to obtain second feature information;

[0072] S223, using a second feature extraction model, processing the second feature information to obtain third feature information;

[0073] S224, upsampling the third feature information to obtain fourth feature information;

[0074] S225, using a third feature extraction model to process the fourth feature information to obtain fifth feature information;

[0075] S226, using a convolution model to process the fifth feature information to obtain sixth feature information;

[0076] S227: Process the sixth characteristic information to obtain denoised ultrasound image data.

[0077] As an optional implementation, in the second aspect of the embodiment of the present invention, performing contrast enhancement processing on the denoised ultrasound image data to obtain preprocessed pre-talofibular acoustic image data includes:

[0078] S231, decomposing the denoised ultrasonic image data to obtain base layer data and detail layer data;

[0079] S232, performing illumination equalization processing on the base layer data to obtain enhanced base layer data;

[0080] S233, performing enhancement processing on the detail layer data to obtain enhanced detail layer data;

[0081] S234, synthesizing the enhanced base layer data and the enhanced detail layer data to obtain pre-processed anterior talofibular acoustic image data.

[0082] As an optional implementation manner, in the second aspect of the embodiment of the present invention, performing data enhancement on the segmented ultrasound image data to obtain enhanced ultrasound image data includes:

[0083] S41, using a preset generator network to process the label information data and random noise to obtain a generated image;

[0084] S42, processing the segmented ultrasound image data and the generated image to obtain a cyclic correlation coefficient;

[0085] S43, using the cyclic correlation coefficient, performing adversarial training on the preset discriminator network and the preset generator network to obtain an optimized discriminator network and an optimized generator network;

[0086] S44, using the optimized generator network, processing the label information data and random noise to obtain an amplified image;

[0087] S45 , integrating the amplified image and the segmented ultrasonic image data to obtain enhanced ultrasonic image data.

[0088] As an optional implementation, in the second aspect of the embodiment of the present invention, the processing of the enhanced ultrasound image data using the three-dimensional reconstruction model to obtain the three-dimensional reconstruction model includes:

[0089] S51, using a three-dimensional reconstruction model, performing three-dimensional reconstruction on the enhanced ultrasound image data to obtain three-dimensional data information;

[0090] S52: Visualize the three-dimensional data information to obtain a three-dimensional reconstructed model.

[0091] As an optional implementation, in the second aspect of the embodiment of the present invention, processing the three-dimensional reconstructed model to obtain puncture path planning information includes:

[0092] S61, analyzing the three-dimensional reconstructed model to obtain starting point coordinate information and ending point coordinate information;

[0093] S62, setting the cost function and constraints of the path planning problem;

[0094] S63, using Logistic chaos mapping to generate the initial population of particle swarm;

[0095] S64, using the artificial bee colony algorithm to perform iterative search and update the individual position to move it towards a lower fitness value;

[0096] S65, using chaotic mapping to search the local solution space to make it jump out of the local optimal value;

[0097] S66, repeat S64 and S65 until the maximum number of iterations is reached or the fitness value change threshold is reached;

[0098] S67: Output the optimal fitness value and obtain puncture path planning information.

[0099] A third aspect of the present invention discloses another intelligent puncture and localization device for lesions of the anterior talofibular ligament using ultrasound imaging, the device comprising:

[0100] a memory storing executable program code;

[0101] a processor coupled to the memory;

[0102] The processor calls the executable program code stored in the memory to execute part or all of the steps in the method for intelligent puncture positioning of lesions in the anterior talofibular ligament using ultrasound imaging disclosed in the first aspect of the embodiment of the present invention.

[0103] The fourth aspect of the present invention discloses a computer-storable medium, which stores computer instructions. When the computer instructions are called, they are used to execute some or all of the steps in the intelligent puncture positioning method for lesions in the anterior talofibular ligament ultrasound imaging disclosed in the first aspect of the embodiment of the present invention.

[0104] Compared with the prior art, the embodiments of the present invention have the following beneficial effects:

[0105] This invention discloses a method and device for intelligent puncture localization of anterior talofibular ligament lesions using ultrasound imaging. This method enables precise identification of the anterior talofibular ligament and improves the accuracy of puncture localization. By performing three-dimensional reconstruction and path planning, it can assist an intelligent navigation system in planning the syringe puncture path, reducing the risk of puncture. BRIEF DESCRIPTION OF THE DRAWINGS

[0106] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0107] Figure 1 This is a flow chart of a method for intelligent puncture and localization of lesions of the anterior talofibular ligament using ultrasound imaging disclosed in an embodiment of the present invention;

[0108] Figure 2 This is an ultrasound image of the anterior talofibular ligament disclosed in an embodiment of the present invention;

[0109] Figure 3 is a schematic diagram of a method for denoising anterior talofibular ligament ultrasound images disclosed in an embodiment of the present invention;

[0110] Figure 4 1 is a schematic structural diagram of an intelligent puncture and positioning device for lesions of the anterior talofibular ligament using ultrasound imaging disclosed in an embodiment of the present invention;

[0111] Figure 5 It is a structural schematic diagram of another intelligent puncture and positioning device for lesions of the anterior talofibular ligament using ultrasound imaging disclosed in an embodiment of the present invention. DETAILED DESCRIPTION

[0112] In order to enable those skilled in the art to better understand the solutions of the present invention, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making any creative efforts shall fall within the scope of protection of the present invention.

[0113] The terms "first," "second," and so on, in the description and claims of the present invention and the accompanying drawings are used to distinguish between different objects, not to describe a specific order. Furthermore, the terms "including," "having," and any variations thereof, are intended to cover non-exclusive inclusions. For example, a process, method, apparatus, product, or device comprising a series of steps or elements is not limited to the listed steps or elements but may optionally include steps or elements not listed, or may optionally include other steps or elements inherent to the process, method, product, or device.

[0114] References herein to "embodiments" mean that a particular feature, structure, or characteristic described in connection with the embodiments may be included in at least one embodiment of the present invention. The appearance of this phrase in various places in the specification does not necessarily refer to the same embodiment, nor does it constitute a separate or alternative embodiment that is mutually exclusive of other embodiments. It is understood, both explicitly and implicitly, by those skilled in the art that the embodiments described herein may be combined with other embodiments.

[0115] The anterior talofibular ligament (ATFL) is the weakest and most easily ruptured ligament tissue on the outside of the ankle. If the ATFL is severely injured, such as a tear or complete rupture, the patient is usually unable to heal on their own. In this case, surgical treatment, such as ligament reconstruction and suturing, is required to restore the integrity and function of the ligament. Knowing the exact location of the ATFL helps to identify problems in a timely manner when the ankle is injured and take appropriate measures to prevent further damage. Currently, the diagnosis of ligament injuries mainly relies on manual reading of films, which is labor-intensive and requires a high level of professional qualifications and clinical experience from the doctor.

[0116] The present invention discloses a method and device for intelligent puncture localization of lesions of the anterior talofibular ligament using ultrasound images. The method comprises: obtaining an ultrasound image dataset of the anterior talofibular ligament; the ultrasound image dataset of the anterior talofibular ligament includes N ultrasound image data of the anterior talofibular ligament, where N is a positive integer; preprocessing the ultrasound image data of the anterior talofibular ligament to obtain preprocessed ultrasound image data; segmenting the preprocessed ultrasound image data to obtain segmented ultrasound image data; performing data enhancement on the segmented ultrasound image data to obtain enhanced ultrasound image data; processing the enhanced ultrasound image data using a three-dimensional reconstruction model to obtain a three-dimensional reconstruction model; processing the three-dimensional reconstruction model to obtain puncture path planning information. The present invention achieves accurate identification of the anterior talofibular ligament and improves the accuracy of puncture localization. Detailed descriptions are given below.

[0117] Example 1

[0118] See also Figure 1 , Figure 1 This is a flow chart of a method for intelligent puncture location of anterior talofibular ligament lesions using ultrasound imaging disclosed in an embodiment of the present invention. Figure 1 The method for intelligent puncture location of lesions in the anterior talofibular ligament ultrasound image described above is applied to the field of medical image processing technology, and the embodiments of the present invention do not limit this. Figure 1 As shown, the method for intelligent puncture location of lesions in the anterior talofibular ligament ultrasound image may include the following operations:

[0119] S1, obtaining an ultrasound image dataset of the anterior talofibular ligament; the ultrasound image dataset of the anterior talofibular ligament includes N ultrasound image data of the anterior talofibular ligament, where N is a positive integer;

[0120] Ultrasound imaging data of the anterior talofibular ligament were obtained using a high-resolution ultrasound probe or from the database of the medical institution.

[0121] S2, preprocessing the ultrasound image data of the anterior talofibular ligament to obtain preprocessed ultrasound image data;

[0122] S3, segmenting the preprocessed ultrasonic image data to obtain segmented ultrasonic image data;

[0123] Segmenting the pre-processed ultrasound image data using a threshold segmentation method to obtain segmented ultrasound image data, where the segmented ultrasound image data includes the anterior talofibular ligament and the lesion area;

[0124] S4, performing data enhancement on the segmented ultrasound image data to obtain enhanced ultrasound image data;

[0125] S5, processing the enhanced ultrasound image data using the three-dimensional reconstruction model to obtain a three-dimensional reconstruction model;

[0126] S6, processing the three-dimensional reconstructed model to obtain puncture path planning information.

[0127] Figure 2 This is an ultrasound image of the anterior talofibular ligament disclosed in an embodiment of the present invention. In the figure, LM represents the fibula (lateral malleolus), Ta represents the talus, and the arrow represents the anterior talofibular ligament.

[0128] Optionally, preprocessing the ultrasound image data of the anterior talofibular ligament to obtain preprocessed ultrasound image data includes:

[0129] S21, labeling the ultrasound image data of the anterior talofibular ligament to obtain label information data;

[0130] Experienced ultrasound imaging physicians annotated the collected ultrasound image data, including the boundary of the anterior talofibular ligament and the lesion area (such as tearing and swelling, which are further divided into 1st, 2nd and 3rd degrees according to the degree of tearing and swelling), with a total of 7 labels.

[0131] S22, performing denoising processing on the ultrasound image data of the anterior talofibular ligament to obtain denoised ultrasound image data;

[0132] S23, performing contrast enhancement processing on the denoised ultrasonic image data to obtain pre-processed anterior talofibular acoustic image data.

[0133] Optionally, performing denoising on the ultrasound image data of the anterior talofibular ligament to obtain denoised ultrasound image data includes:

[0134] S221, processing the ultrasound image data of the anterior talofibular ligament using a first feature extraction model to obtain first feature information;

[0135] S222: downsample the first feature information to obtain second feature information;

[0136] S223, using a second feature extraction model, processing the second feature information to obtain third feature information;

[0137] S224, upsampling the third feature information to obtain fourth feature information;

[0138] S225, using a third feature extraction model to process the fourth feature information to obtain fifth feature information;

[0139] S226, using a convolution model to process the fifth feature information to obtain sixth feature information;

[0140] S227: Process the sixth characteristic information to obtain denoised ultrasound image data.

[0141] Figure 3 This is a schematic diagram of the anterior talofibular ligament ultrasound image denoising method disclosed in an embodiment of the present invention. The first feature extraction model consists of a two-layer ConvNet network. The first layer includes three convolutions of different sizes and orientations: 3×3, 1×3, and 3×1. The second layer uses a 3×3 ConvNet + ReLU operation to further refine the fused features. Using convolution kernels of different sizes and orientations can capture more diverse features. The expansion rate of each ConvNet is 1.

[0142] y1=C 1×3 (x)

[0143] y2=C 3×1 (x)

[0144] y3=C 3×3 (x)

[0145] f1=Re(C 3×3 (Cat(y1,y2,y3)))

[0146] Where x is the ultrasound image data of the anterior talofibular ligament (noisy image), C 1×3 is a 1×3 convolution, C 3×1 is a 3×1 convolution, C 3×3 is a 3×3 convolution, y1 is the output of a 1×3 convolution, y2 is the output of a 3×1 convolution, y3 is the output of a 3×3 convolution, f1 is the first feature information, Re represents the ReLU activation function, and Cat is the fusion concatenation operation.

[0147] The second feature extraction model includes a first-scale model, a second-scale model, a third-scale model, and a fourth-scale model. The four models have the same structure and are composed of a multi-scale progressive cascade combination of four interacting residual units. The data of the four interacting residual units are Cat-concatenated and then subjected to 1×1 convolution to obtain the output.

[0148] The third feature extraction model is a dual-path global attention module. The upper branch extracts wide-area pixel information through two 7×7 Conv operations. A hidden layer is added in the middle to adjust the number of channels and improve network efficiency. The channel compression ratio r of the hidden layer is 8. A sigmoid function is used to capture the input-output weight relationship and weights it with the initial features by channel-by-channel multiplication. The lower branch first obtains aggregated features through downsampling. Then, a multi-layer perceptron unit composed of two 1×1 Conv operations is used to adjust the feature map through the hidden layer to capture the multidimensional hierarchical dependencies between image pixels. The compression ratio r is the same as that of the upper branch. The sigmoid function filters the output features to a value between (0, 1). Finally, the outputs of the upper and lower branches are added together to obtain the fifth feature information.

[0149] The convolution model is 3×3 convolution.

[0150] Processing the sixth characteristic information to obtain denoised ultrasonic image data includes:

[0151]

[0152] in, is the denoised ultrasound image data, x r It is the sixth characteristic information.

[0153] Optionally, performing contrast enhancement processing on the denoised ultrasound image data to obtain pre-processed pre-talofibular acoustic image data includes:

[0154] S231, decomposing the denoised ultrasonic image data to obtain base layer data and detail layer data;

[0155] Denoised ultrasound image data I(x,y) = base layer data I b (x,y)+detail layer data I d (x,y)

[0156] I(x,y)=I b (x,y)+I d (x,y)

[0157] The base layer data is obtained by minimizing the following objective function

[0158]

[0159] The objective function consists of two terms: the first is a difference term adapted to the texture component to preserve meaningful structures; and the second is a regularization term based on the total variation, which will constrain image details to the detail layer. represents the gradient operator, and λ is set to twice the global noise estimate σ, that is, λ = 2σ.

[0160] The global noise estimate σ is calculated as:

[0161]

[0162] Where * represents the convolution operator. W and H are the width and height of the image I(x,y) respectively. N is the convolution mask, defined as

[0163]

[0164] S232, performing illumination equalization processing on the base layer data to obtain enhanced base layer data;

[0165] Will I b (x,y) is converted to HSI color space to obtain intensity layer I i (x,y),I i (x,y)=L(x,y)·R(x,y), where L(x,y) is the illumination component and R(x,y) is the reflection component.

[0166] The enhanced illumination component is

[0167] L e (x,y)=L(x,y)·L sig (x,y)

[0168] L e (x,y) is the enhanced illumination component, r min and r max They represent the minimum and maximum values ​​of the reflected radiation component, α=0.5~7.5, β=0.002~0.014 respectively.

[0169] The enhanced reflection component is

[0170] R e (x,y)=cR(x,y) γ

[0171] Where c is a constant, c = 0.3 and γ = 0.6.

[0172] I′ i (x,y)=L e (x,y)·R e (x,y)

[0173] Among them, I′ i (x,y) is the enhanced strength layer, and I′ i (x,y) is converted to RGB space to obtain the enhanced base layer data.

[0174] S233, performing enhancement processing on the detail layer data to obtain enhanced detail layer data;

[0175] ω(x,y)=|I d (x,y)|*G(x,y)

[0176] Where * represents the convolution operator and G(x,y) represents the Gaussian filter, which is defined as

[0177]

[0178] In this paper, the Gaussian standard deviation δ=21, ω(x,y) is the enhanced detail layer data, and x,y are the coordinates of the pixel.

[0179] S234, synthesizing the enhanced base layer data and the enhanced detail layer data to obtain pre-processed anterior talofibular acoustic image data.

[0180] Pre-processed talofibular anterior acoustic image data = enhanced base layer data + enhanced detail layer data

[0181] Optionally, performing data enhancement on the segmented ultrasound image data to obtain enhanced ultrasound image data includes:

[0182] The Conditional Generative Adversarial Network (CGAN) guides the data generation process by inputting random noise and label information. It consists of a generator network (G) and a discriminator network (D). The generator G generates samples with label features, and the discriminator D distinguishes the generated samples from real samples.

[0183] S41, using a preset generator network to process the label information data and random noise to obtain a generated image;

[0184] S42, processing the segmented ultrasound image data and the generated image to obtain a cyclic correlation coefficient

[0185]

[0186] Where i and j represent the pixel positions, m and n are the image length and width, V1(i, j) is the segmented ultrasound image data, and V2(i, j) is the generated image.

[0187] S43, using the cyclic correlation coefficient, conduct adversarial training on the preset discriminator network and the preset generator network to obtain an optimized discriminator network and an optimized generator network; the training process is as follows.

[0188] (1) Random noise and label information data are input into the generator network G to obtain the generated image. The discriminator network D is fixed, and the cyclic correlation coefficients of the generated image and the label information data are calculated respectively. The maximum value of the above cyclic correlation coefficients is selected to improve the loss function and optimize the parameters of the generator network G.

[0189] (2) Then fix the generator network G, and input the generated image, real image and label information data into the discriminator network D to judge whether the image is true or false, and improve the loss function based on the maximum value of the cyclic correlation coefficient between the image and the label image.

[0190] (3) Feed the results back to the generator network G and the discriminator network D, update the parameters, and repeat the above steps to balance the loss functions of the two and reduce the mismatch between the generated images and the labels.

[0191] S44, using the optimized generator network, processing the label information data and random noise to obtain an amplified image;

[0192] When amplifying an image, the average value of the cyclic correlation coefficients within each label in the original data set is used as a threshold. When the cyclic correlation coefficient between the generated image and the corresponding label image exceeds this threshold, the image is output to obtain the amplified image.

[0193] S45 , integrating the amplified image and the segmented ultrasonic image data to obtain enhanced ultrasonic image data.

[0194] Optionally, the process of processing the enhanced ultrasound image data using a three-dimensional reconstruction model to obtain a three-dimensional reconstruction model includes:

[0195] S51, using a three-dimensional reconstruction model, performing three-dimensional reconstruction on the enhanced ultrasound image data to obtain three-dimensional data information;

[0196] Optionally, the three-dimensional reconstruction model is a voxel-based three-dimensional reconstruction algorithm (VBM);

[0197] S52: Visualize the three-dimensional data information to obtain a three-dimensional reconstructed model.

[0198] Optionally, the three-dimensional data information is visualized by Smartbi visualization software or a self-designed graphical user interface.

[0199] Optionally, processing the three-dimensional reconstructed model to obtain puncture path planning information includes:

[0200] S61, analyzing the three-dimensional reconstructed model to obtain starting point coordinate information and ending point coordinate information;

[0201] S62, setting the cost function and constraints of the path planning problem;

[0202] F tr =ω4f fu +ω5f hi +ω6f da

[0203] Where ω4, ω5, and ω6 are the path length costs f fu , path height cost f hi , risk cost f da The weight factor, and ω4+ω5+ω6=1, F tr is the cost function.

[0204] S63, using Logistic chaos mapping to generate the initial population of particle swarm;

[0205] S64, using the artificial bee colony algorithm to perform iterative search and update the individual position to move it towards a lower fitness value;

[0206] S65, using chaotic mapping to search the local solution space to make it jump out of the local optimal value;

[0207] S66, repeat S64 and S65 until the maximum number of iterations is reached or the fitness value change threshold is reached;

[0208] S67: Output the optimal fitness value and obtain puncture path planning information.

[0209] This paper discloses a method and device for intelligently locating anterior talofibular ligament lesions using ultrasound imaging. This method enables precise identification of the anterior talofibular ligament and improves the accuracy of puncture location. By performing three-dimensional reconstruction and path planning, it can assist an intelligent navigation system in planning the syringe puncture path, reducing the risk of puncture.

[0210] Example 2

[0211] The steps of the intelligent puncture location method of the anterior talofibular ligament ultrasound image lesion in this embodiment include:

[0212] 1. Intelligent identification of the anterior talofibular ligament

[0213] 1.1 Data Collection and Preprocessing

[0214] Data acquisition: A high-resolution ultrasound probe was used to obtain ultrasound imaging data of the anterior talofibular ligament.

[0215] Data annotation: Experienced ultrasound imaging physicians annotate the collected ultrasound image data, including the boundaries of the anterior talofibular ligament and the lesion area (such as tearing, swelling, etc.).

[0216] Data preprocessing: including denoising, contrast enhancement, and normalization to improve image quality.

[0217] 1.2 Deep Learning Model Training

[0218] Model selection: Select a deep learning model suitable for medical image segmentation, such as U-Net, DeepLab, etc.

[0219] Training data: Use labeled ultrasound image data as the training set to train a deep learning model for accurate identification of the anterior talofibular ligament.

[0220] Data enhancement: Data enhancement technology is used to increase the diversity of training data and prevent model overfitting.

[0221] Loss function: Use the Dice coefficient loss function or combine it with the cross entropy loss function to optimize the model's segmentation accuracy for the lesion area.

[0222] 1.3 Model Validation and Optimization

[0223] Validation set test: Use an independent validation set to test model performance and calculate indicators such as the Dice coefficient and IoU to evaluate the model segmentation accuracy.

[0224] Hyperparameter tuning: Adjust model hyperparameters based on validation set results to achieve optimal performance.

[0225] Model integration: Model integration technology is used to further improve segmentation accuracy.

[0226] 2. Accurately guide the puncture path

[0227] 2.1 3D Reconstruction

[0228] Reconstructed model: Generate a 3D reconstructed model based on the anterior talofibular ligament and lesion area segmented by the deep learning model.

[0229] Visualization: Visualize the 3D reconstructed model to provide an intuitive view of the anatomical structure.

[0230] 2.2 Puncture path planning

[0231] Path calculation: Use the fast marching method or similar path planning algorithm to calculate the optimal puncture path.

[0232] Multi-objective optimization: During the path planning process, multiple objectives (such as path length, angle, safety, etc.) are considered and a multi-objective optimization algorithm is used for path optimization.

[0233] 2.3 Real-time Navigation

[0234] Real-time tracking: Track the syringe position through real-time ultrasound imaging and update the puncture path in real time.

[0235] Feedback mechanism: Establish a feedback mechanism to issue timely warnings and provide correction suggestions when the syringe deviates from the predetermined path.

[0236] User interface: Provides an intuitive user interface that displays real-time ultrasound images, puncture path, syringe position and other information.

[0237] 3. Device Design

[0238] 3.1 Hardware Composition

[0239] Ultrasound probe: High-resolution ultrasound probe with high frame rate and good spatial resolution.

[0240] Computer system: A high-performance computer system for running deep learning models and path planning algorithms.

[0241] Syringe positioning module: equipped with high-precision sensors to track the syringe position in real time.

[0242] User interface equipment: including monitor, keyboard, mouse, etc., providing an operating interface.

[0243] 3.2 Software Composition

[0244] Data acquisition software: controls the ultrasound probe to collect data.

[0245] Deep learning model: A trained deep learning model is used for accurate identification of the anterior talofibular ligament.

[0246] 3D reconstruction module: Generates a 3D reconstruction model based on the segmentation results.

[0247] Path planning module: calculates the optimal puncture path.

[0248] Real-time navigation module: tracks the syringe position in real time and provides navigation information.

[0249] User interface software: provides an operation interface that displays ultrasound images, puncture paths, syringe positions and other information.

[0250] As can be seen, the present invention can accurately identify the anterior talofibular ligament using deep learning technology, improving the accuracy of puncture positioning. Intelligent navigation: The intelligent navigation system plans the syringe puncture path, reducing the risk of puncture. Real-time feedback: The syringe position is tracked in real time, providing navigation information to ensure puncture accuracy. User-friendly: The intuitive user interface helps doctors perform precise punctures.

[0251] Please refer to Example 3 Figure 4 , Figure 4This is a schematic diagram of the structure of an intelligent puncture and positioning device for lesions of the anterior talofibular ligament using ultrasound imaging disclosed in an embodiment of the present invention. Figure 4 The intelligent puncture positioning device for lesions of the anterior talofibular ligament based on ultrasound images is applied in the field of medical image processing technology, and the embodiments of the present invention are not limited thereto. Figure 4 As shown, the intelligent puncture and positioning device for lesions of the anterior talofibular ligament using ultrasound imaging may include the following operations:

[0252] S301, a data acquisition module, configured to acquire an ultrasound image dataset of the anterior talofibular ligament; the ultrasound image dataset of the anterior talofibular ligament includes N ultrasound image data of the anterior talofibular ligament, where N is a positive integer;

[0253] S302, a preprocessing module, configured to preprocess the ultrasound image data of the anterior talofibular ligament to obtain preprocessed ultrasound image data;

[0254] S303, an image segmentation module, configured to segment the pre-processed ultrasound image data to obtain segmented ultrasound image data;

[0255] S304, a data enhancement module, configured to perform data enhancement on the segmented ultrasound image data to obtain enhanced ultrasound image data;

[0256] S305, a three-dimensional reconstruction module, configured to process the enhanced ultrasound image data using a three-dimensional reconstruction model to obtain a three-dimensional reconstruction model;

[0257] S305, a path planning module is used to process the three-dimensional reconstructed model to obtain puncture path planning information.

[0258] This paper discloses a method and device for intelligently locating anterior talofibular ligament lesions using ultrasound imaging. This method enables precise identification of the anterior talofibular ligament and improves the accuracy of puncture location. By performing three-dimensional reconstruction and path planning, it can assist an intelligent navigation system in planning the syringe puncture path, reducing the risk of puncture.

[0259] Example 4

[0260] See also Figure 5 , Figure 5 This is a schematic diagram of the structure of another intelligent puncture and positioning device for lesions of the anterior talofibular ligament using ultrasound imaging disclosed in an embodiment of the present invention. Figure 5 The intelligent puncture positioning device for lesions of the anterior talofibular ligament based on ultrasound images is applied in the field of medical image processing technology, and the embodiments of the present invention are not limited thereto. Figure 5 As shown, the intelligent puncture and positioning device for lesions of the anterior talofibular ligament using ultrasound imaging may include the following operations:

[0261] A memory 401 storing executable program code;

[0262] a processor 402 coupled to the memory 401;

[0263] The processor 402 calls the executable program code stored in the memory 401 to execute the steps of the method for intelligent puncture location of lesions of the anterior talofibular ligament using ultrasound imaging described in the first and second embodiments.

[0264] Example 5

[0265] An embodiment of the present invention discloses a computer-readable storage medium that stores a computer program for electronic data exchange, wherein the computer program enables a computer to execute the steps of the method for intelligent puncture localization of lesions in the anterior talofibular ligament ultrasound imaging described in Examples 1 and 2.

[0266] The device embodiments described above are merely illustrative. Modules described as separate components may or may not be physically separate, and components shown as modules may or may not be physical modules, i.e., they may be located in one place or distributed across multiple network modules. Some or all of the modules may be selected based on actual needs to achieve the objectives of the present embodiment. Persons of ordinary skill in the art will be able to understand and implement the present invention without inventive effort.

[0267] Through the detailed description of the above embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus the necessary general hardware platform, or of course, by means of hardware. Based on this understanding, the above technical solution, in essence, or the portion that contributes to the prior art, can be embodied in the form of a software product, which can be stored in a computer-readable storage medium, including a read-only memory (ROM), a random access memory (RAM), a programmable read-only memory (PROM), an erasable programmable read-only memory (EPROM), a one-time programmable read-only memory (OTPROM), an electronically erasable programmable read-only memory (EEPROM), a compact disc read-only memory (CD-ROM) or other optical disc storage, magnetic disk storage, magnetic tape storage, or any other computer-readable medium capable of carrying or storing data.

[0268] Finally, it should be noted that the intelligent puncture positioning method and device for lesions of the anterior talofibular ligament based on ultrasound imaging disclosed in the embodiment of the present invention only discloses a preferred embodiment of the present invention, which is only used to illustrate the technical solution of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, ordinary technicians in this field should understand that it is still possible to modify the technical solutions recorded in the aforementioned embodiments, or to make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the various embodiments of the present invention.

Claims

1. A method for intelligent puncture location of lesions in the anterior talofibular ligament using ultrasound imaging, characterized in that: The method comprises: S1, obtaining an ultrasound image dataset of the anterior talofibular ligament; the ultrasound image dataset of the anterior talofibular ligament includes N ultrasound image data of the anterior talofibular ligament, where N is a positive integer; S2, preprocessing the ultrasound image data of the anterior talofibular ligament to obtain preprocessed ultrasound image data, including: S21, labeling the ultrasound image data of the anterior talofibular ligament to obtain label information data; S22, performing denoising processing on the ultrasound image data of the anterior talofibular ligament to obtain denoised ultrasound image data, including: S221, processing the ultrasound image data of the anterior talofibular ligament using a first feature extraction model to obtain first feature information; The first feature extraction model consists of a two-layer Conv network. The first layer includes 3×3, 1×3, and 3×1 branches of convolution with different sizes and directions. The second layer uses 3×3Conv+ReLU operations to further refine the fused features. The use of convolution kernels of different sizes and directions can capture more diverse features. The expansion rate of Conv is 1. y1=C 1×3 (x) y2=C 3×1 (x) y3=C 3×3 (x) f1=Re(C 3×3 (Cat(y1,y2,y3))) Where x is the ultrasound image data of the anterior talofibular ligament (noisy image), C 1×3 is a 1×3 convolution, C 3×1 is a 3×1 convolution, C 3×3 is a 3×3 convolution, y1 is the output of a 1×3 convolution, y2 is the output of a 3×1 convolution, y3 is the output of a 3×3 convolution, f1 is the first feature information, Re represents the ReLU activation function, and Cat is the fusion concatenation operation; S222: downsample the first feature information to obtain second feature information; S223, using a second feature extraction model, processing the second feature information to obtain third feature information; The second feature extraction model includes a first-scale model, a second-scale model, a third-scale model, and a fourth-scale model. The four models have the same structure, which is composed of a multi-scale cascade combination of four interacting residual units. The data of the four interacting residual units are Cat-joined and then passed through a 1×1 convolution to obtain the output; S224, upsampling the third feature information to obtain fourth feature information; S225, using a third feature extraction model to process the fourth feature information to obtain fifth feature information; The third feature extraction model is a dual-path global attention module. The upper branch extracts wide-area pixel information through two 7×7 Conv operations. A hidden layer is set in the middle to adjust the number of channels and improve network operation efficiency. The channel compression ratio r of the hidden layer is 8. The Sigmoid function is used to capture the input-output weight relationship and weight the initial features by channel-by-channel multiplication. The lower branch first obtains aggregated features through a downsampling operation. Subsequently, a multi-layer perceptron unit composed of two 1×1 Conv is used to adjust the feature map through the hidden layer to capture the multi-dimensional hierarchical dependency between image pixels. The compression ratio r is the same as that of the upper branch. The Sigmoid function filters the output features to between (0, 1). The outputs of the upper and lower branches are added together to obtain the fifth feature information. S226, using a convolution model to process the fifth feature information to obtain sixth feature information; The convolution model is 3×3 convolution; S227, processing the sixth characteristic information to obtain denoised ultrasound image data, including: in, is the denoised ultrasound image data, x r is the sixth characteristic information; S23, performing contrast enhancement processing on the denoised ultrasound image data to obtain pre-processed pre-talofibular acoustic image data, including: S231, decomposing the denoised ultrasonic image data to obtain base layer data and detail layer data; Denoised ultrasound image data I(x,y) = base layer data I b (x,y)+detail layer data I d (x,y) I(x,y)=I b (x,y)+I d (x,y) The base layer data is obtained by minimizing the following objective function: The objective function consists of two terms: the first is a difference term adapted to the texture component, used to preserve meaningful structures; and the second is a regularization term based on total variation, which will limit the image details to the detail layer. represents the gradient operator, λ is set to 2 times the global noise estimate σ, that is, λ = 2σ; The global noise estimate σ is calculated as: Where * represents the convolution operator, W and H are the width and height of the image I(x,y), and N is the convolution template, which is defined as S232, performing illumination equalization processing on the base layer data to obtain enhanced base layer data; Will I b (x,y) is converted to HSI color space to obtain intensity layer I i (x,y),I i (x,y)=L(x,y)·R(x,y), where L(x,y) is the illumination component and R(x,y) is the reflection component. The enhanced illumination component is L e (x,y)=L(x,y)·L sig (x,y) L e (x,y) is the enhanced illumination component, r min and r max Respectively represent the minimum and maximum values ​​of the reflected radiation component, α = 0.5 ~ 7.5, β = 0.002 ~ 0.014; The enhanced reflection component is R e (x,y)=cR(x,y) γ Where c is a constant, c = 0.3 and γ = 0.6; I i ′(x,y)=L e (x,y)·R e (x,y) Among them, I i ′(x,y) is the enhanced strength layer, and I i ′(x,y) is converted to RGB space to obtain enhanced base layer data; S233, performing enhancement processing on the detail layer data to obtain enhanced detail layer data; ω(x,y)=|I d (x,y)|*G(x,y) Where * represents the convolution operator and G(x,y) represents the Gaussian filter, which is defined as Take the Gaussian standard deviation δ = 21, ω(x, y) is the enhanced detail layer data, x, y are the coordinates of the pixel; S234, synthesizing the enhanced base layer data and the enhanced detail layer data to obtain pre-processed anterior talofibular acoustic image data; S3, segmenting the preprocessed ultrasonic image data to obtain segmented ultrasonic image data; S4, performing data enhancement on the segmented ultrasound image data to obtain enhanced ultrasound image data, including: S41, using a preset generator network to process the label information data and random noise to obtain a generated image; S42, processing the segmented ultrasound image data and the generated image to obtain a cyclic correlation coefficient; Where i and j represent the pixel positions, m and n are the image length and width, V1(i, j) is the segmented ultrasound image data, and V2(i, j) is the generated image. S43, using the cyclic correlation coefficient, performing adversarial training on the preset discriminator network and the preset generator network to obtain an optimized discriminator network and an optimized generator network; S431: Input random noise and label information data into the generator network G to obtain a generated image. Fix the discriminator network D, calculate the cyclic cross-correlation coefficient of the generated image and the label information data, and select the maximum value of the cyclic cross-correlation coefficient to improve the loss function and optimize the parameters of the generator network G. S432: Then, the generator network G is fixed, and the generated image, the real image, and the label information data are input into the discriminator network D to judge whether the image is real or fake, and the loss function is improved based on the maximum cyclic correlation coefficient between the image and the label image; S433: Feed the results back to the generator network G and the discriminator network D, update the parameters, and repeat the above steps to balance the loss functions of the two networks, thereby reducing the mismatch between the generated images and the labels. S44, using the optimized generator network, processing the label information data and random noise to obtain an amplified image; When amplifying an image, the average value of the cyclic correlation coefficients within each label in the original data set is used as a threshold. When the cyclic correlation coefficient between the generated image and the corresponding label image exceeds this threshold, the image is output to obtain the amplified image. S45, integrating the amplified image and the segmented ultrasonic image data to obtain enhanced ultrasonic image data; S5, processing the enhanced ultrasound image data using the three-dimensional reconstruction model to obtain a three-dimensional reconstruction model; S6, processing the three-dimensional reconstructed model to obtain puncture path planning information, including: S61, analyzing the three-dimensional reconstructed model to obtain starting point coordinate information and ending point coordinate information; S62, setting the cost function and constraints of the path planning problem; S63, using Logistic chaos mapping to generate the initial population of particle swarm; S64, using the artificial bee colony algorithm to perform iterative search and update the individual position to move it towards a lower fitness value; S65, using chaotic mapping to search the local solution space to make it jump out of the local optimal value; S66, repeat S64 and S65 until the maximum number of iterations is reached or the fitness value change threshold is reached; S67: Output the optimal fitness value and obtain puncture path planning information.

2. The method for intelligent puncture and localization of lesions of the anterior talofibular ligament using ultrasound imaging according to claim 1, characterized in that: The method of processing the enhanced ultrasound image data using the three-dimensional reconstruction model to obtain the three-dimensional reconstruction model includes: S51, using a three-dimensional reconstruction model, performing three-dimensional reconstruction on the enhanced ultrasound image data to obtain three-dimensional data information; S52: Visualize the three-dimensional data information to obtain a three-dimensional reconstructed model.

3. An intelligent puncture and localization device for lesions of the anterior talofibular ligament using ultrasound imaging, characterized in that: The device comprises: a data acquisition module, configured to acquire an ultrasound image dataset of the anterior talofibular ligament; the ultrasound image dataset of the anterior talofibular ligament comprising N ultrasound image data of the anterior talofibular ligament, where N is a positive integer; A preprocessing module, configured to preprocess the ultrasound image data of the anterior talofibular ligament to obtain preprocessed ultrasound image data, comprising: S21, labeling the ultrasound image data of the anterior talofibular ligament to obtain label information data; S22, performing denoising processing on the ultrasound image data of the anterior talofibular ligament to obtain denoised ultrasound image data, including: S221, processing the ultrasound image data of the anterior talofibular ligament using a first feature extraction model to obtain first feature information; The first feature extraction model consists of a two-layer Conv network. The first layer includes 3×3, 1×3, and 3×1 branches of convolution with different sizes and directions. The second layer uses 3×3Conv+ReLU operations to further refine the fused features. The use of convolution kernels of different sizes and directions can capture more diverse features. The expansion rate of Conv is 1. y1=C 1×3 (x) y2=C 3×1 (x) y3=C 3×3 (x) f1=Re(C 3×3 (Cat(y1,y2,y3))) Where x is the ultrasound image data of the anterior talofibular ligament (noisy image), C 1×3 is a 1×3 convolution, C 3×1 is a 3×1 convolution, C 3×3 is a 3×3 convolution, y1 is the output of a 1×3 convolution, y2 is the output of a 3×1 convolution, y3 is the output of a 3×3 convolution, f1 is the first feature information, Re represents the ReLU activation function, and Cat is the fusion concatenation operation; S222: downsample the first feature information to obtain second feature information; S223, using a second feature extraction model, processing the second feature information to obtain third feature information; The second feature extraction model includes a first-scale model, a second-scale model, a third-scale model, and a fourth-scale model. The four models have the same structure, which is composed of a multi-scale cascade combination of four interacting residual units. The data of the four interacting residual units are Cat-joined and then passed through a 1×1 convolution to obtain the output; S224, upsampling the third feature information to obtain fourth feature information; S225, using a third feature extraction model to process the fourth feature information to obtain fifth feature information; The third feature extraction model is a dual-path global attention module. The upper branch extracts wide-area pixel information through two 7×7 Conv operations. A hidden layer is set in the middle to adjust the number of channels and improve network operation efficiency. The channel compression ratio r of the hidden layer is 8. The Sigmoid function is used to capture the input-output weight relationship and weight the initial features by channel-by-channel multiplication. The lower branch first obtains aggregated features through a downsampling operation. Subsequently, a multi-layer perceptron unit composed of two 1×1 Conv is used to adjust the feature map through the hidden layer to capture the multi-dimensional hierarchical dependency between image pixels. The compression ratio r is the same as that of the upper branch. The Sigmoid function filters the output features to between (0, 1). The outputs of the upper and lower branches are added together to obtain the fifth feature information. S226, using a convolution model to process the fifth feature information to obtain sixth feature information; The convolution model is 3×3 convolution; S227, processing the sixth characteristic information to obtain denoised ultrasound image data, including: in, is the denoised ultrasound image data, x r is the sixth characteristic information; S23, performing contrast enhancement processing on the denoised ultrasound image data to obtain pre-processed pre-talofibular acoustic image data, including: S231, decomposing the denoised ultrasonic image data to obtain base layer data and detail layer data; Denoised ultrasound image data I(x,y) = base layer data I b (x,y)+detail layer data I d (x,y) I(x,y)=I b (x,y)+I d (x,y) The base layer data is obtained by minimizing the following objective function: The objective function consists of two terms: the first is a difference term adapted to the texture component, used to preserve meaningful structures; and the second is a regularization term based on total variation, which will limit the image details to the detail layer. represents the gradient operator, λ is set to 2 times the global noise estimate σ, that is, λ = 2σ; The global noise estimate σ is calculated as: Where * represents the convolution operator, W and H are the width and height of the image I(x,y), and N is the convolution template, which is defined as S232, performing illumination equalization processing on the base layer data to obtain enhanced base layer data; Will I b (x,y) is converted to HSI color space to obtain intensity layer I i (x,y),I i (x,y)=L(x,y)·R(x,y), where L(x,y) is the illumination component and R(x,y) is the reflection component. The enhanced illumination component is L e (x,y)=L(x,y)·L sig (x,y) L e (x,y) is the enhanced illumination component, r min and r max Respectively represent the minimum and maximum values ​​of the reflected radiation component, α = 0.5 ~ 7.5, β = 0.002 ~ 0.014; The enhanced reflection component is R e (x,y)=cR(x,y) γ Where c is a constant, c = 0.3 and γ = 0.6; I i ′(x,y)=L e (x,y)·R e (x,y) Among them, I i ′(x,y) is the enhanced strength layer, and I i ′(x,y) is converted to RGB space to obtain enhanced base layer data; S233, performing enhancement processing on the detail layer data to obtain enhanced detail layer data; ω(x,y)=|I d (x,y)|*G(x,y) Where * represents the convolution operator and G(x,y) represents the Gaussian filter, which is defined as Take the Gaussian standard deviation δ = 21, ω(x, y) is the enhanced detail layer data, x, y are the coordinates of the pixel; S234, synthesizing the enhanced base layer data and the enhanced detail layer data to obtain pre-processed anterior talofibular acoustic image data; an image segmentation module, configured to segment the preprocessed ultrasound image data to obtain segmented ultrasound image data; A data enhancement module is used to perform data enhancement on the segmented ultrasound image data to obtain enhanced ultrasound image data, including: S41, using a preset generator network to process the label information data and random noise to obtain a generated image; S42, processing the segmented ultrasound image data and the generated image to obtain a cyclic correlation coefficient; Where i and j represent the pixel positions, m and n are the image length and width, V1(i, j) is the segmented ultrasound image data, and V2(i, j) is the generated image. S43, using the cyclic correlation coefficient, performing adversarial training on the preset discriminator network and the preset generator network to obtain an optimized discriminator network and an optimized generator network; S431: Input random noise and label information data into the generator network G to obtain a generated image. Fix the discriminator network D, calculate the cyclic cross-correlation coefficient of the generated image and the label information data, and select the maximum value of the cyclic cross-correlation coefficient to improve the loss function and optimize the parameters of the generator network G. S432: Then, the generator network G is fixed, and the generated image, the real image, and the label information data are input into the discriminator network D to judge whether the image is real or fake, and the loss function is improved based on the maximum cyclic correlation coefficient between the image and the label image; S433: Feed the results back to the generator network G and the discriminator network D, update the parameters, and repeat the above steps to balance the loss functions of the two networks, thereby reducing the mismatch between the generated images and the labels. S44, using the optimized generator network, processing the label information data and random noise to obtain an amplified image; When amplifying an image, the average value of the cyclic correlation coefficients within each label in the original data set is used as a threshold. When the cyclic correlation coefficient between the generated image and the corresponding label image exceeds this threshold, the image is output to obtain the amplified image. S45, integrating the amplified image and the segmented ultrasonic image data to obtain enhanced ultrasonic image data; a three-dimensional reconstruction module, configured to process the enhanced ultrasound image data using a three-dimensional reconstruction model to obtain a three-dimensional reconstruction model; A path planning module is used to process the three-dimensional reconstructed model to obtain puncture path planning information, including: S61, analyzing the three-dimensional reconstructed model to obtain starting point coordinate information and ending point coordinate information; S62, setting the cost function and constraints of the path planning problem; S63, using Logistic chaos mapping to generate the initial population of particle swarm; S64, using the artificial bee colony algorithm to perform iterative search and update the individual position to move it towards a lower fitness value; S65, using chaotic mapping to search the local solution space to make it jump out of the local optimal value; S66, repeat S64 and S65 until the maximum number of iterations is reached or the fitness value change threshold is reached; S67: Output the optimal fitness value and obtain puncture path planning information.

4. An intelligent puncture and positioning device for lesions of the anterior talofibular ligament using ultrasound imaging, characterized in that: The device comprises: a memory storing executable program code; a processor coupled to the memory; The processor calls the executable program code stored in the memory to execute the intelligent puncture positioning method for lesions of the anterior talofibular ligament using ultrasound imaging as described in any one of claims 1-2.

5. A computer storable medium, characterized in that The computer storable medium stores computer instructions, and when the computer instructions are called, they are used to execute the intelligent puncture positioning method for lesions of the anterior talofibular ligament using ultrasound imaging as described in any one of claims 1-2.

Images