Ultrasound lesion standard section selection method and system
By matching the ultrasound image feature vector and standard section average image based on deep learning, the standard section type and position of ultrasound lesions are determined, which solves the empirical dependence problem of probe position and angle in ultrasound image acquisition, and improves the scientificity and data of clinical trials accuracy.
Patent Information
- Application Number
- PCT/CN2024/090787
- Authority / Receiving Office
- WO · WO
- Patent Type
- Applications
- Current Assignee / Owner
- Priority Date
- 2023-06-02
- Filing Date
- 2024-04-30
- Publication Date
- 2024-12-05
AI Technical Summary
During ultrasound image scanning, the probe position and scanning angle are limited by the operator's experience, making it challenging to achieve standardized ultrasound image acquisition. The prior art ignores information redundancy between different directions, which may lead to underestimation of information or Overestimate.
Using a deep learning-based method, the standard section with the highest similarity is determined by obtaining the feature vector of the ultrasonic image to be processed and the average image feature vector corresponding to each standard section type. , use image registration algorithm for adaptive adjustments to better utilize information redundancy.
Improve the comparability and accuracy of clinical trial data, reduce subjective factors in ultrasound scanning operations, enhance scientificity and reliability, and can more comprehensively describe the characteristics of each pixel in the ultrasound image, capturing different tissues or organs. difference between.
Smart Images

Figure CN2024090787_05122024_PF_FP_ABST
Abstract
Description
Method and system for selecting standard sections of ultrasound lesions
[0001] Related applications
[0002] This application claims priority to Chinese patent application number 202310654567.X, filed on June 2, 2023, entitled “A method and system for selecting standard sections of ultrasound lesions based on deep learning,” the entire text of which is hereby incorporated by reference. Technical Field
[0003] The present application relates to the field of ultrasound technology, and in particular to a method and system for selecting standard sections of ultrasound lesions. Background Art
[0004] Clinical trials are an important means of evaluating the safety and efficacy of new drugs, therapies, or medical devices. Ultrasound technology is widely used in these trials to detect and assess lesions. However, during ultrasound image acquisition, scanning parameters such as probe position and scanning angle are limited by the experience of the scanning operator, making standardized ultrasound image acquisition in clinical trials challenging.
[0005] In related technologies, real-time ultrasound images are usually matched with the feature vectors of several preset standard section ultrasound images to screen out the standard section ultrasound image with the highest similarity; this section selection method ignores the information redundancy between different directions, and may therefore cause the information in some directions to be underestimated or overestimated.
[0006] Summary of the Invention
[0007] According to various embodiments of the present application, the present application provides a method and system for selecting standard sections of ultrasound lesions that can be applied in clinical trials.
[0008] In a first aspect, the present application provides a method for selecting a standard section of an ultrasound lesion, comprising:
[0009] Acquiring a first ultrasound image to be processed, and extracting a first eigenvector of the first ultrasound image;
[0010] Matching the first feature vector with the feature vector of the average image corresponding to each standard section type, and determining the standard section type corresponding to the average image having the highest similarity to the first feature vector;
[0011] The position and angle corresponding to the standard section are determined according to the standard section type.
[0012] In one embodiment, the deep learning-based ultrasound lesion standard section selection method further includes:
[0013] The real-time ultrasound image after adaptively adjusting the position and angle is registered with the standard section corresponding to the determined standard section type.
[0014] In one embodiment, an image registration algorithm is used for registration, and the image registration algorithm adopts a phase correlation algorithm, a mutual information algorithm or a deformation field algorithm.
[0015] In one embodiment, the step of obtaining the feature vector of the average image corresponding to each standard section type includes:
[0016] acquiring a second ultrasound image set including the lesion;
[0017] performing feature extraction on all second ultrasound images in the second ultrasound image group to obtain a feature vector group;
[0018] Clustering the feature vector group based on a clustering algorithm to obtain cluster groups corresponding to different lesion morphologies and location types;
[0019] respectively calculating an average value of the feature vectors of all the second ultrasound images in each cluster of the cluster group to obtain an average image of each cluster and a feature vector of the average image;
[0020] The characteristic vectors of all average images corresponding to the cluster groups are classified to determine the standard section type corresponding to each cluster.
[0021] In one embodiment, in step A3, the clustering algorithm is a K-means clustering algorithm, a density-based clustering algorithm, or a hierarchical clustering algorithm.
[0022] In one embodiment, the average value of the feature vectors of all the second ultrasound images in each cluster of the cluster group is calculated using a feature calculation function; wherein the feature calculation function includes a feature fusion function.
[0023] In one embodiment, determining the position and angle corresponding to the standard section according to the standard section type further includes:
[0024] Selecting the corresponding standard section template from the database according to the determined standard section type; wherein the database stores a standard section template corresponding to each standard section type, and the standard section template at least indicates a position and angle corresponding to the standard section;
[0025] The position and angle of the first ultrasound image are adaptively selected according to the standard section template.
[0026] In a second aspect, the present application further provides an ultrasound lesion standard section selection system for implementing the above-mentioned ultrasound lesion standard section selection method, comprising:
[0027] a first feature extraction module, configured to acquire a first ultrasound image to be processed and extract a first feature vector of the first ultrasound image;
[0028] a feature matching module, connected to the first feature extraction module, configured to match the first feature vector with a feature vector of an average image corresponding to each standard section type, and determine the standard section type corresponding to the average image having the highest similarity to the first feature vector;
[0029] The section selection module is connected to the feature matching module and is used to determine the position and angle corresponding to the standard section according to the standard section type.
[0030] In one embodiment, the ultrasound lesion standard section selection system further includes:
[0031] A section registration module is connected to the section selection module and is used to register the real-time ultrasound image after adaptively adjusting the position and angle with the standard section corresponding to the determined standard section type.
[0032] In one embodiment, the ultrasound lesion standard section selection system further includes:
[0033] An image acquisition module, configured to acquire a second ultrasound image group containing the lesion;
[0034] a second feature extraction module, connected to the image acquisition module, configured to extract features from all second ultrasound images in the second ultrasound image group to obtain a feature vector group;
[0035] a clustering module, connected to the second feature extraction module, configured to perform clustering processing on the feature vector group based on a clustering algorithm to obtain cluster groups corresponding to different lesion morphologies and location types;
[0036] an average processing module, connected to the clustering module, configured to respectively calculate the average of the feature vectors of all the second ultrasound images in each cluster of the cluster group, to obtain an average image of each cluster and a feature vector of the average image;
[0037] A classification module is connected to the average processing module and is used to classify the feature vectors of all average images corresponding to the cluster groups and determine the standard section type corresponding to each cluster.
[0038] The details of one or more embodiments of the present application are set forth in the following drawings and description to make other features, objects, and advantages of the present application more readily apparent. BRIEF DESCRIPTION OF THE DRAWINGS
[0039] In order to more clearly illustrate the technical solutions in the embodiments of the present application or related technologies, the following briefly introduces the drawings required for use in the embodiments or related technical descriptions. Obviously, the drawings described below are merely embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on the disclosed drawings without any creative work.
[0040] FIG1 is a flow chart of a method for selecting a standard section of an ultrasound lesion in some embodiments of the present application.
[0041] FIG2 is a flow chart of a method for selecting a standard section of an ultrasound lesion in some embodiments of the present application.
[0042] FIG3 is a flow chart of the steps for obtaining the feature vector of the average image in some embodiments of the present application.
[0043] FIG4 is a schematic diagram of a process for determining the position and angle corresponding to a standard section according to the standard section type in some embodiments of the present application.
[0044] FIG5 is a structural block diagram of an ultrasound lesion standard section selection system in some embodiments of the present application.
[0045] FIG6 is a structural block diagram of an ultrasound lesion standard section selection system in some embodiments of the present application.
[0046] FIG7 is a structural block diagram of an ultrasound lesion standard section selection system in some embodiments of the present application. DETAILED DESCRIPTION
[0047] The following will be combined with the accompanying drawings in the embodiments of this application to clearly and completely describe the technical solutions in the embodiments of this application. Obviously, the embodiments described are only part of the embodiments of this application, not all of the embodiments. Based on the embodiments of this application, all other embodiments obtained by ordinary technicians in this field without making any creative work are within the scope of protection of this application.
[0048] It should be noted that, unless there is any conflict, the embodiments and features in the embodiments of this application can be combined with each other.
[0049] The present application will be further described below with reference to the accompanying drawings and specific embodiments, but they are not intended to limit the present application.
[0050] Referring to FIG1 , some embodiments of the present application provide a method for selecting a standard section of an ultrasound lesion in a clinical trial, comprising:
[0051] Step S1, obtaining a first ultrasound image to be processed, and extracting a first eigenvector of the first ultrasound image;
[0052] Step S2, matching the first eigenvector with the eigenvector of the average image corresponding to each standard section type, and determining the standard section type corresponding to the average image having the highest similarity to the first eigenvector;
[0053] Step S3: determining the position and angle corresponding to the standard section according to the standard section type.
[0054] Specifically, in this embodiment, a first ultrasound image to be processed is first acquired, which is the image currently being acquired in real time. A first eigenvector of the first ultrasound image acquired in real time is extracted and matched with the eigenvectors of average images corresponding to several preset standard slice types. Their similarity is calculated, and the average image most similar to the first eigenvector is determined. The standard slice type corresponding to the average image is used as the standard slice type for the ultrasound image. The corresponding position and angle of the standard slice are then determined based on the standard slice type. In comparison, the method of using the eigenvector of the average image (i.e., the average value of the ultrasound image eigenvectors) can better utilize information redundancy between different directions, thereby more comprehensively describing the characteristics of each pixel in the ultrasound image. This method can better capture the differences between different tissues or organs, automatically select a standard slice based on the location and morphology of the specific lesion, and has wide applicability and practicality. It can also improve the comparability and accuracy of clinical trial data and reduce subjective factors in ultrasound scanning operations, thereby enhancing the scientific nature and reliability of clinical trials.
[0055] As an optional implementation, as shown in FIG2 , after step S3, the following steps are further included:
[0056] Step S4: registering the real-time ultrasound image after the adaptive adjustment of the position and angle with the standard section corresponding to the determined standard section type.
[0057] Specifically, in this embodiment, the standard section is registered with the ultrasound image to solve the problem of the difference between the standard section and the ultrasound image due to deformation and noise of the ultrasound image in the related art.
[0058] As an optional implementation, in step S4, an image registration algorithm is used for registration, and the image registration algorithm adopts a phase correlation algorithm, a mutual information algorithm or a deformation field algorithm.
[0059] Specifically, in this embodiment, in order to register the standard section with the ultrasound image, an image registration algorithm such as a phase correlation algorithm, a mutual information algorithm, or a deformation field algorithm may be used.
[0060] Furthermore, after step S4, the above step further includes: testing the accuracy of the selected standard sections. In order to verify the accuracy and reliability of the implementation scheme of this application, in actual application, some experienced doctors or ultrasound diagnostic experts evaluate and compare the standard sections, calculate their sensitivity and specificity and other indicators, and implement the testing of the selected standard sections.
[0061] As an optional embodiment, as shown in FIG3 , the step of obtaining the feature vector of the average image corresponding to each standard section type in step S2 includes:
[0062] Step A1, acquiring a second ultrasound image group containing the lesion;
[0063] A set of ultrasound images is acquired to form a second ultrasound image set. The ultrasound images in the second ultrasound image set are from patients in a clinical trial. Each second ultrasound image contains at least one lesion. Optionally, each ultrasound image should contain a clear boundary of the lesion and sufficient information for feature extraction.
[0064] Step A2, performing feature extraction on all second ultrasound images in the second ultrasound image group to obtain a feature vector group;
[0065] Among them, feature extraction is performed on each second ultrasound image to obtain a set of feature vectors to describe the morphology and position information of the ultrasound image. In one embodiment, the feature extraction step can be implemented based on deep learning, such as using a related convolutional neural network (CNN) model or an HIS (Hue, Intensity, Saturation) statistical grayscale distribution method. The convolutional neural network model can optionally use network models such as VGG (Visual Geometry Group), ResNet (Residual Network) or Inception. In another embodiment, the CNN model can also be customized according to specific needs to achieve personalized feature extraction. For each ultrasound image, it is input into the CNN model and the output of its last convolution layer is extracted as the feature vector of the image.
[0066] Furthermore, before feature extraction, a data preprocessing step is included: all ultrasound images are preprocessed, including image enhancement using anisotropic filtering and IIR (Infinite-Duration Impulse Response) filtering of previous and next frames, and normalization to improve clustering accuracy. Feature extraction is then performed on the preprocessed ultrasound images.
[0067] Step A3: clustering the feature vector group based on a clustering algorithm to obtain cluster groups corresponding to different lesion morphologies and location types;
[0068] The feature vectors of all images in the second ultrasound image group are clustered to obtain a set of clusters. In a preferred embodiment, the clustering operation can be implemented using a common clustering algorithm, such as K-Means, DBSCAN (Density-Based Spatial Clustering of Applications with Noise), or hierarchical clustering. The feature vectors are input into the clustering algorithm for clustering. The clustering algorithm automatically divides all feature vectors into different clusters, each cluster representing a lesion morphology and location type. Optionally, the number of clusters and parameters can be selected according to specific needs.
[0069] Step A4, respectively calculating the average value of the feature vectors of all second ultrasound images in each cluster of the cluster group to obtain an average image and a feature vector of the average image of each cluster;
[0070] For each cluster, the average image of all second ultrasound images is calculated, and a feature vector of this average image is extracted. For each cluster, the feature vectors of all ultrasound images are averaged to obtain the average feature vector for the cluster. This average feature vector is then input into the CNN model, and the output of the last convolutional layer is extracted as the feature vector of the average image for the cluster. In step A4, by calculating the average of the feature vectors in each cluster as the representative feature vector for that cluster, the complexity and subjectivity of feature extraction are reduced, and all information can be fully utilized.
[0071] It should be noted that, to improve model reusability, the CNN model used in step A4 can share the convolutional neural network model used in step A2. Furthermore, in step S1, the convolutional neural network model used in step A2 can also be used to perform feature extraction on the first ultrasound image to be processed. The first ultrasound image to be processed is input into the CNN model, and the output of the last convolutional layer is extracted as the feature vector of the image.
[0072] Step A5: classify the feature vectors of all average images corresponding to the cluster groups to determine the standard section type corresponding to each cluster.
[0073] Among them, the feature vectors of all average images are classified to determine the standard section type corresponding to each cluster. In one embodiment, the classification operation can be implemented using common classification algorithms, such as SVM (Support Vector Machine), random forest or deep neural network. When training the classifier, the feature vector of each average image is used as input and the corresponding standard section type is used as output. In practical applications, standard section selection: for the real-time second ultrasound image, after extracting its feature vector, it is compared with the representative feature vectors of all cluster clusters, and the standard section type corresponding to the representative feature vector of the most similar cluster cluster is selected.
[0074] Specifically, considering that the relevant method for calculating ultrasound image feature vectors is typically to obtain the feature vector by taking a weighted average of the grayscale value and directional information of each pixel, while this feature vector calculation method is simple and easy to implement, it ignores the information redundancy between different directions and may therefore cause information in some directions to be underestimated or overestimated. In this embodiment, the method of calculating the average value of the ultrasound image feature vector can better utilize the information redundancy between different directions, thereby more comprehensively describing the characteristics of each pixel in the ultrasound image; at the same time, this method can better capture the differences between different tissues or organs and improve the accuracy of classification and diagnosis.
[0075] Furthermore, calculating the average of ultrasound image feature vectors can reduce computational effort, thereby improving model training efficiency. This is because only the average of the feature vectors for each ultrasound image is required, rather than calculating the feature vector for each pixel. This significantly reduces computational time and memory requirements.
[0076] As an optional implementation, in step A3, the clustering algorithm is a K-means clustering algorithm, a density-based clustering algorithm, or a hierarchical clustering algorithm.
[0077] Specifically, a hierarchical clustering algorithm is used to cluster all ultrasound images. The feature vectors are input into the algorithm for clustering. The algorithm automatically divides all feature vectors into clusters, each representing a lesion morphology and location type, to obtain different lesion morphology and location types. The hierarchical clustering algorithm is a tree-based clustering method. Its main idea is to divide the data into different hierarchical structures and gradually merge data points with similar characteristics to ultimately form a tree structure.
[0078] As an optional implementation, in step A4, the average value of the feature vectors of all second ultrasound images in each cluster of the cluster group is calculated using a feature calculation function; wherein the feature calculation function includes a feature fusion function.
[0079] Specifically, in this embodiment, in step A4, the feature vector for each ultrasound image in the dataset is calculated using a feature calculation function in the Python-based scikit-learn library, such as the Feature Union function or the FMRI to Feature (Frequency Modulated Reconstruction Imaging to Feature Extraction) function. These functions can combine the feature vectors of multiple ultrasound images into a single feature vector. The mean of the feature vector for each ultrasound image is calculated using the mean function in NumPy (Numerical Python).
[0080] Furthermore, in step A5 above, a deep neural network is used for classification. The deep neural network includes convolutional layers, pooling layers, and fully connected layers. Unlike conventional networks such as VGG and RESNET, the first convolutional layer of this deep neural network uses a large 13*13 convolution kernel, and the output of the first convolutional layer is directly skipped and then added to the fully connected layer. This deep neural network architecture retains the features of the high and middle layers to the greatest extent, and is not easily lost as the layers deepen. Compared with networks with the same number of layers, its sensitivity and specificity are improved.
[0081] As an optional implementation, as shown in FIG4 , step S3 further includes:
[0082] Step S31, selecting a corresponding standard section template from a database according to the determined standard section type; wherein the database stores a standard section template corresponding to each standard section type, and the standard section template at least indicates a position and angle corresponding to the standard section;
[0083] Step S32: adaptively selecting the position and angle of the first ultrasound image according to the standard section template.
[0084] Specifically, in this embodiment, a standard section is selected based on the standard section type. For each standard section type, a standard section template is pre-determined and stored in a database. In actual applications, the standard section type corresponding to the real-time ultrasound image is determined through matching. Then, the standard section template corresponding to the standard section type is selected from the database, and a position and angle similar to the standard section template is adaptively selected in the real-time ultrasound image.
[0085] The present application also provides an ultrasound lesion standard section selection system for implementing the above-mentioned ultrasound lesion standard section selection method, as shown in FIG5 , comprising:
[0086] A first feature extraction module 1 is used to obtain a first ultrasound image to be processed and extract a first feature vector of the first ultrasound image;
[0087] Feature matching module 2, connected to the first feature extraction module 1, is used to match the first feature vector with the feature vector of the average image corresponding to each standard section type, and determine the standard section type corresponding to the average image with the highest similarity to the first feature vector;
[0088] The section selection module 3 is connected to the feature matching module 2 and is used to determine the position and angle corresponding to the standard section according to the standard section type.
[0089] As an optional embodiment, as shown in FIG6 , the ultrasound lesion standard section selection system further includes:
[0090] The section registration module 4 is connected to the section selection module 3 and is used to register the real-time ultrasound image after adaptively adjusting the position and angle with the standard section corresponding to the determined standard section type.
[0091] Specifically, in practical applications, due to deformation and noise in ultrasound images, there may be certain differences between the standard section and the ultrasound image. In this embodiment, the standard section is registered with the ultrasound image to avoid differences between the standard section and the ultrasound image that may reduce the comparability and accuracy of the test data. Furthermore, to register the standard section with the ultrasound image, an image registration algorithm such as a phase correlation algorithm, a mutual information algorithm, or a deformation field algorithm can be used to implement the registration operation.
[0092] As an optional embodiment, as shown in FIG7 , the ultrasound lesion standard section selection system further includes:
[0093] An image acquisition module 5 is configured to acquire a second ultrasound image group including the lesion;
[0094] A second feature extraction module 6, connected to the image acquisition module 5, is used to extract features from all second ultrasound images in the second ultrasound image group to obtain a feature vector group;
[0095] A clustering module 7, connected to the second feature extraction module 6, is used to perform clustering processing on the feature vector group based on a clustering algorithm to obtain cluster groups corresponding to different lesion morphologies and location types;
[0096] an average processing module 8, connected to the clustering module 7, for respectively calculating the average value of the feature vectors of all second ultrasound images in each cluster of the cluster group, to obtain an average image of each cluster and a feature vector of the average image;
[0097] The classification module 9 is connected to the average processing module 8 and is used to classify the feature vectors of all average images corresponding to the cluster groups and determine the standard section type corresponding to each cluster.
[0098] The advantages or beneficial effects of adopting the above-mentioned technical solution are: this application determines the standard section type corresponding to the average image with the highest similarity to the first eigenvector by matching the feature vector of the acquired ultrasound image with the feature vector of the average image corresponding to each standard section type, and then determines the position and angle corresponding to the standard section according to the standard section type, which has wide applicability and practicality; at the same time, it can improve the comparability and accuracy of clinical trial data, and reduce the subjective factors of ultrasound scanning operations, thereby improving the scientificity and reliability of clinical trials.
[0099] It should be understood by those skilled in the art that the modules or steps of the present application described above can be implemented using a general-purpose computing device, they can be concentrated on a single computing device, or distributed across a network composed of multiple computing devices, optionally, they can be implemented using program code executable by the computing device, so that they can be stored in a storage device and executed by the computing device, and in some cases, the steps shown or described can be performed in a different order than herein, or they can be made into separate integrated circuit modules, or multiple modules or steps can be made into a single integrated circuit module for implementation. Thus, the present application is not limited to any specific combination of hardware and software.
[0100] It should be understood that, although the various steps in the flowcharts involved in the various embodiments described above are displayed in sequence according to the instructions of the arrows, these steps are not necessarily executed in sequence in the order indicated by the arrows. Unless otherwise specified herein, there is no strict order restriction on the execution of these steps, and these steps can be executed in other orders. Moreover, at least a portion of the steps in the flowcharts involved in the various embodiments described above can include multiple steps or multiple stages, and these steps or stages are not necessarily executed and completed at the same time, but can be executed at different times, and the execution order of these steps or stages is not necessarily to be carried out in sequence, but can be executed in turn or alternately with other steps or at least a portion of steps or stages in other steps.
[0101] The technical features of the above embodiments can be combined arbitrarily. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.
[0102] The above-described embodiments merely represent several implementation methods of the present application. While the descriptions are relatively specific and detailed, they should not be construed as limiting the scope of the present application. It should be noted that a person of ordinary skill in the art may make various modifications and improvements without departing from the spirit of the present application, and these modifications and improvements fall within the scope of protection of the present application. Therefore, the scope of protection of the present application shall be determined by the appended claims.
Claims
1. A method for selecting standard sections of ultrasonic lesions, characterized in that: include: Acquire a first ultrasonic image to be processed, and extract a first eigenvector of the first ultrasonic image; Matching the first feature vector with a feature vector of an average image corresponding to each standard section type, and determining the standard section type corresponding to the average image having the highest similarity to the first feature vector; The position and angle corresponding to the standard section are determined according to the standard section type.
2. The method for selecting standard sections of ultrasonic lesions according to claim 1, wherein: Also includes: The real-time ultrasound image after adaptively adjusting the position and angle is registered with the standard section corresponding to the determined standard section type.
3. The method for selecting the standard section of ultrasonic lesions according to claim 2, wherein: An image registration algorithm is used for registration, and the image registration algorithm adopts a phase correlation algorithm, a mutual information algorithm or a deformation field algorithm.
4. The method for selecting standard sections of ultrasonic lesions according to claim 1, wherein: The step of acquiring the feature vector of the average image corresponding to each standard section type comprises: acquiring a second ultrasound image group including the lesion; Performing feature extraction on all second ultrasonic images in the second ultrasonic image group to obtain a feature vector group; Performing clustering processing on the feature vector group based on a clustering algorithm to obtain cluster groups corresponding to different lesion morphologies and location types; Respectively calculating the average values of the feature vectors of all the second ultrasound images in each cluster of the cluster group to obtain an average image of each cluster and a feature vector of the average image; The characteristic vectors of all average images corresponding to the cluster groups are classified to determine the standard section type corresponding to each cluster.
5. The method for selecting standard sections of ultrasonic lesions according to claim 4, wherein: The clustering algorithm is a K-means clustering algorithm, a density-based clustering algorithm, or a hierarchical clustering algorithm.
6. The method for selecting standard sections of ultrasonic lesions according to claim 4, wherein: The average value of the feature vectors of all the second ultrasound images in each cluster of the cluster group is calculated using a feature calculation function; wherein the feature calculation function includes a feature fusion function.
7. The method for selecting standard sections of ultrasonic lesions according to claim 1, wherein: Determining the position and angle corresponding to the standard section according to the standard section type also includes: According to the determined standard section type, the corresponding standard section template is selected from the database; wherein the database stores the standard section template corresponding to each standard section type, and the standard section template The plate at least indicates the position and angle corresponding to the standard section; The position and angle of the first ultrasound image are adaptively selected according to the standard section template.
8. An ultrasonic lesion standard section selection system, characterized in that: The method for selecting a standard ultrasonic lesion section according to any one of claims 1 to 7 comprises: A first feature extraction module, used to acquire a first ultrasonic image to be processed and extract a first feature vector of the first ultrasonic image; a feature matching module, connected to the first feature extraction module, for matching the first feature vector with a feature vector of an average image corresponding to each standard section type, and determining the standard section type corresponding to the average image having the highest similarity to the first feature vector; The section selection module is connected to the feature matching module and is used to determine the position and angle corresponding to the standard section according to the standard section type.
9. The ultrasonic lesion standard section selection system according to claim 8, wherein: Also includes: A section registration module is connected to the section selection module and is used to register the real-time ultrasound image after adaptively adjusting the position and angle with the standard section corresponding to the determined standard section type.
10. The ultrasonic lesion standard section selection system according to claim 8, wherein: Also includes: An image acquisition module, used for acquiring a second ultrasound image group containing the lesion; A second feature extraction module, connected to the image acquisition module, is used to extract features from all the second ultrasound images in the second ultrasound image group to obtain a feature vector group; A clustering module, connected to the second feature extraction module, is used to perform clustering processing on the feature vector group based on a clustering algorithm to obtain clustering cluster groups corresponding to different lesion morphologies and location types; an average processing module, connected to the clustering module, for respectively calculating the average values of the feature vectors of all the second ultrasound images in each cluster of the cluster group, to obtain an average image of each cluster and a feature vector of the average image; A classification module is connected to the average processing module and is used to classify the feature vectors of all average images corresponding to the cluster groups, and determine the standard section type corresponding to each cluster.
Citation Information
Patent Citations
Ultrasonic breast focus classification method and device and storage medium
CN113239951A
Liver ultrasonic standard section identification method and device, electronic equipment and storage medium
CN113658145A
Medical image registration method, medical image system and ultrasonic imaging system
CN116091516A
Ultrasonic focus standard section selection method and system based on deep learning
CN116681764A
Living body detection method and apparatus, electronic device, storage medium, and related system to which living body detection method is applied
US20210073516A1
