Abdominal aorta imaging method and related apparatus

Abstract

The application provides an abdominal aorta imaging method, which comprises the following steps: obtaining an ultrasonic echo signal of a target object abdominal space, reconstructing a three-dimensional ultrasonic image of the abdominal aorta based on the ultrasonic echo signal, selecting a target cross-section image from the three-dimensional ultrasonic image, and calculating related information of the abdominal aorta based on the target cross-section image. The three-dimensional ultrasonic image directly shows the overall structure information of the abdominal aorta, the provided abdominal aorta information is more abundant, the calculated related information of the abdominal aorta is more accurate, and therefore the accuracy of a disease diagnosis result based on the three-dimensional ultrasonic image is also higher. In addition, the application embodiment further provides an ultrasonic detection device to ensure the application and implementation of the above method in practice.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of medical equipment, more particularly, to an abdominal aorta imaging method and related equipment. BACKGROUND

[0002] Abnormalities such as dilation may occur in the abdominal aorta, which is an aortic aneurysm, threatening people's life and health. At present, the main way to detect the abdominal aorta is ultrasonic scanning, that is, using an ultrasonic probe to perform continuous transverse and longitudinal scanning of each segment of the abdominal aorta and multi-section scanning of the lesion area, to obtain a two-dimensional image of the abdominal aorta, and then observing the state of the arterial wall and lumen according to the two-dimensional image to assess whether the abdominal aorta has an aneurysm or other lesions.

[0003] However, this kind of ultrasonic detection method can only obtain one section image of the abdominal aorta each time, and the diagnostic results based on the image data are not accurate enough. SUMMARY

[0004] To this end, the present application provides an abdominal aorta imaging method and related equipment to solve the technical problem of inaccurate ultrasonic diagnosis of the abdominal aorta.

[0005] In a first aspect, an embodiment of the present application provides an abdominal aorta imaging method, comprising:

[0006] controlling an ultrasonic detection probe to emit ultrasonic waves to the abdominal space of a target object, and controlling the ultrasonic detection probe to receive ultrasonic echo signals returned from the abdominal space of the target object;

[0007] performing three-dimensional image reconstruction based on the ultrasonic echo signals to obtain a three-dimensional ultrasonic image of the abdominal aorta;

[0008] selecting a target section image from the three-dimensional ultrasonic image, the target section image comprising an abdominal aorta transverse section image and / or an abdominal aorta longitudinal section image;

[0009] calculating related information of the abdominal aorta based on the target section image

[0010] In a second aspect, an embodiment of the present application provides an abdominal aorta imaging method, comprising:

[0011] controlling an ultrasonic detection probe to emit ultrasonic waves to the abdominal space of a target object, and controlling the ultrasonic detection probe to receive ultrasonic echo signals returned from the abdominal space of the target object;

[0012] performing three-dimensional image reconstruction based on the ultrasonic echo signals to obtain a three-dimensional ultrasonic image of the abdominal aorta;

[0013] displaying the three-dimensional ultrasonic image of the abdominal aorta.

[0014] In a third aspect, the embodiments of the present application provide an ultrasound detection device, comprising:

[0015] an ultrasound detection probe configured to emit ultrasound waves to an abdominal space of a target object;

[0016] a processor configured to control the ultrasound detection probe to emit ultrasound waves to the abdominal space of the target object, and control the ultrasound detection probe to receive ultrasound echo signals returned from the abdominal space of the target object; perform three-dimensional image reconstruction based on the ultrasound echo signals to obtain a three-dimensional ultrasound image of the abdominal aorta; select a target cross-sectional image from the three-dimensional ultrasound image, the target cross-sectional image comprising an abdominal aorta transverse cross-sectional image and / or an abdominal aorta longitudinal cross-sectional image; and calculate relevant information of the abdominal aorta based on the target cross-sectional image.

[0017] In a fourth aspect, the embodiments of the present application provide an ultrasound detection device, comprising:

[0018] an ultrasound detection probe configured to emit ultrasound waves to an abdominal space of a target object;

[0019] a processor configured to control the ultrasound detection probe to emit ultrasound waves to the abdominal space of the target object, and control the ultrasound detection probe to receive ultrasound echo signals returned from the abdominal space of the target object; perform three-dimensional image reconstruction based on the ultrasound echo signals to obtain a three-dimensional ultrasound image of the abdominal aorta;

[0020] a display configured to display the three-dimensional ultrasound image of the abdominal aorta.

[0021] As can be seen from the above technical solutions, the embodiments of the present application have at least the following advantages:

[0022] After obtaining the ultrasound echo signals of the abdominal space of the target object, a three-dimensional ultrasound image of the abdominal aorta is reconstructed based on the ultrasound echo signals, a target cross-sectional image is obtained from the three-dimensional ultrasound image, and relevant information of the abdominal aorta is calculated based on the target cross-sectional image. The three-dimensional ultrasound image can provide more abundant information of the abdominal aorta for medical personnel, and the calculated relevant information of the abdominal aorta is more accurate, thereby improving the accuracy of diagnosis of abdominal aorta diseases to a certain extent. BRIEF DESCRIPTION OF DRAWINGS

[0023] In order to more clearly illustrate the technical solutions of the embodiments of the present application or the prior art, the following will briefly introduce the drawings needed to be used in the embodiments or the prior art description. Obviously, the drawings in the following description are only some embodiments of the present application, and for those skilled in the art, other drawings can also be obtained without creative labor based on these drawings.

[0024] Figure 1 Flow chart for one embodiment of the abdominal aorta detection method;

[0025] Figure 2 Flow chart for another embodiment of the abdominal aorta detection method;

[0026] Figure 3 Schematic diagram of two implementations for determining the abdominal aorta region from a three-dimensional ultrasound image;

[0027] Figure 4 Schematic diagram of two ways of identifying the cross section of the abdominal aorta region;

[0028] Figure 5A and Figure 5B Schematic diagram of two different ways of marking the abdominal aorta region;

[0029] Figure 6 Schematic diagram of one implementation of feature matching to identify the abdominal aorta region based on an image library;

[0030] Figure 7 Flow chart for yet another embodiment of the abdominal aorta detection method;

[0031] Figure 8A Schematic diagram of one implementation of inner diameter measurement based on longitudinal section images of the abdominal aorta;

[0032] Figure 8B Schematic diagram of one implementation of inner diameter measurement based on transverse section images of the abdominal aorta;

[0033] Figure 9 Schematic diagram of one implementation of transverse section inner diameter variation curve;

[0034] Figure 10 Schematic diagram of one implementation of the abdominal aorta detection device;

[0035] Figure 11 Schematic diagram of another implementation of the abdominal aorta detection device. DETAILED DESCRIPTION

[0036] The technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all the embodiments of the present application. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative work fall within the scope of protection of the present application.

[0037] Currently, the main examination method of abdominal aorta is to use an ultrasound detection device to collect an ultrasound image of the abdominal aorta, and to diagnose the lesion of the abdominal aorta based on the ultrasound image, including but not limited to abdominal aortic aneurysm. Ultrasound examination has the advantages of safety, convenience, no radiation, low cost, etc., and is widely used in clinical examination and has become one of the main auxiliary means for the diagnosis of many diseases. However, the inventors found at least the following technical problems in the implementation of the prior art: medical staff usually need to continuously scan various sections of the abdominal aorta, which takes a long time, and the accuracy of the disease diagnosis result obtained from the ultrasound image needs to be further improved.

[0038] To solve at least one of the above technical problems, an embodiment of the present application provides an abdominal aorta imaging method, which can generate and display a three-dimensional ultrasound image of the abdominal aorta, and the abdominal aorta disease diagnosis result obtained from the three-dimensional ultrasound image is more accurate. The method can be applied to an ultrasound detection device, as shown in Figure 1 , one embodiment of the abdominal aorta imaging method specifically includes steps 101-103.

[0039] 101, control the ultrasound detection probe to emit ultrasound waves to the abdominal space of the target object, and control the ultrasound detection probe to receive ultrasound echo signals returned from the abdominal space of the target object.

[0040] Specifically, the ultrasound detection device has an ultrasound detection probe, and the processor can control the probe to emit ultrasound waves. In the application scenario of the embodiment of the present application, the medical staff can aim the ultrasound detection probe at the abdominal space of the target object, and the ultrasound detection device can receive the ultrasound echo signals returned from the abdominal space. The target object can be a human being or other types of animals with an abdominal cavity.

[0041] 102, three-dimensional image reconstruction is performed based on the ultrasound echo signals to obtain a three-dimensional ultrasound image of the abdominal aorta.

[0042] Specifically, the abdominal space of the target object has an abdominal aorta, and the ultrasound echo signal carries information related to the abdominal aorta. The ultrasound echo signal is reconstructed to obtain a three-dimensional ultrasound image of the abdominal aorta. One reconstruction method is that the ultrasound echo signal is a three-dimensional ultrasound echo signal and has corresponding spatial position information. For example, a detection probe with three-dimensional detection capability such as a volume probe or a surface array probe is used to obtain a three-dimensional ultrasound echo signal, the three-dimensional ultrasound echo signal carries spatial position information, and the three-dimensional ultrasound echo signal is directly used for three-dimensional image reconstruction to obtain a three-dimensional ultrasound image. Another reconstruction method is that the ultrasound echo signal is processed to obtain a plurality of two-dimensional ultrasound data of the abdominal aorta, and spatial position information corresponding to the two-dimensional ultrasound data is obtained; based on the plurality of two-dimensional ultrasound data and the spatial position information corresponding to the two-dimensional ultrasound data, a three-dimensional ultrasound image of the abdominal aorta is reconstructed. The second reconstruction method is described in detail below.

[0043] The two-dimensional ultrasound data is obtained by an ultrasound detection device performing ultrasound detection on the abdominal aorta region, and the ultrasound detection method includes but is not limited to two-dimensional B-mode, color flow, spectral Doppler, etc., which is not limited here. Specifically, the ultrasound detection device includes a detection probe such as a convex array probe, and a medical staff can use the detection probe to scan the abdominal aorta from top to bottom or from bottom to top. The starting point of the transverse section observation is usually located below the diaphragm, and the end point reaches the bifurcation level of the left and right common iliac arteries. When a convex array probe is used for scanning, continuous scanning from the starting point to the end point can be performed; or the medical staff can also use a detection probe such as a volume probe to scan after finding the position of the abdominal aorta segment of interest in advance.

[0044] When scanning, the detection probe can emit ultrasound waves to the target object and receive ultrasound echo signals returned from the target object, and the ultrasound detection device processes the ultrasound echo signals to obtain two-dimensional ultrasound images of the target object. The two-dimensional ultrasound data in this step can be two-dimensional ultrasound images or ultrasound echo signals. If the ultrasound echo signals are obtained, the two-dimensional ultrasound images can be obtained by processing the ultrasound echo signals, and then the subsequent three-dimensional reconstruction step is performed, or the ultrasound echo signals are directly used for the subsequent three-dimensional reconstruction step.

[0045] It should be noted that in order to realize the reconstruction of the three-dimensional ultrasound image, not only the two-dimensional ultrasound data is needed, but also the spatial position information corresponding to the two-dimensional ultrasound data is needed. The spatial position information represents the position information of the abdominal aorta scanned by the two-dimensional ultrasound data in the abdominal space. Exemplarily, the spatial position information can specifically include spatial coordinate information and orientation information, or other information capable of representing the spatial position.

[0046] One way of obtaining the spatial position information is that the detection probe can carry a spatial positioning device, which can perceive the motion trajectory of the detection probe in the three-dimensional space, and the detection probe can collect two-dimensional ultrasound data, and the spatial positioning device provides corresponding spatial position information for the two-dimensional ultrasound data.

[0047] After obtaining the multiple frames of two-dimensional ultrasound data and the spatial position information of the two-dimensional ultrasound data, the three-dimensional ultrasound image can be reconstructed. The two-dimensional ultrasound data can represent what the image data of a plane of the abdominal aorta is, and according to the spatial position information of the two-dimensional ultrasound data, the plane corresponding to the two-dimensional ultrasound data in the abdominal space can be determined, and the relative positions between the two-dimensional ultrasound data can also be determined, so that the three-dimensional ultrasound image can be reconstructed.

[0048] The reconstruction of the three-dimensional ultrasound image can specifically include various implementation manners, and embodiments of the present application take a volume data reconstruction manner as an example for description. Specifically, the volume data reconstruction manner includes two steps of volume data construction and voxel value mapping.

[0049] 1. Volume data construction: determining the abdominal aorta volume data of the three-dimensional ultrasound image to be reconstructed.

[0050] Specifically, the volume data can be considered as a hypothetical three-dimensional space structure, which is the three-dimensional abdominal aorta to be reconstructed in the application scenario of the embodiments of the present application. The determined volume data can specifically include parameters such as coordinate origin, dimension, physical interval between voxels, and the like, and can specifically adopt a bounding box technology or other methods, which are not limited here.

[0051] 2. Voxel value mapping: according to the mapping relationship between the pixels in the two-dimensional ultrasound data and the voxels of the abdominal aorta volume data, the pixel values of the multiple frames of two-dimensional ultrasound data are mapped to the voxel values of the abdominal aorta volume data, so as to obtain the three-dimensional ultrasound image of the abdominal aorta.

[0052] First, a two-dimensional ultrasound image is obtained from the two-dimensional ultrasound data, and it needs to be noted that the two-dimensional ultrasound image can not be output for display. According to the spatial position information of the multiple frames of two-dimensional ultrasound images, a mapping relationship between the pixels in the multiple frames of two-dimensional ultrasound images and the voxels of the abdominal aorta volume data is established. Specifically, the pixels included in the two-dimensional ultrasound image are extracted, and a mapping relationship between each pixel and the corresponding spatial position voxel of the abdominal aorta volume data is established based on the spatial position of each pixel. Then, according to the mapping relationship, the pixel values of the multiple frames of two-dimensional ultrasound images are mapped to the voxel values of the abdominal aorta volume data, and the pixel value mapping process can select a forward mapping, a reverse mapping or a function-based mapping and the like.

[0053] The forward mapping is a mapping mode from pixels in the two-dimensional ultrasound image to voxels in the abdominal aorta volume data. The specific process includes: traversing each pixel included in the two-dimensional ultrasound image, mapping the pixel to the corresponding voxel according to the transformation matrix of the spatial coordinate position of each pixel, and there may be multiple pixels mapped to the same voxel in the mapping process. For this case, the corresponding voxel can be valued according to a certain method, such as the nearest neighbor pixel method or the pixel mean method. Due to the sparsity of the two-dimensional ultrasound data sampling process, there may be some unvalued voxels after forward mapping. For such unvalued missing voxels, the values of the neighboring voxels can be used for difference calculation, and the calculation result is used as the value of the voxel to ensure the comprehensiveness of the voxel value.

[0054] The reverse mapping is a mapping mode from voxels in the abdominal aorta volume data to pixels in the two-dimensional ultrasound image. The specific process includes: traversing each voxel in the abdominal aorta volume data, finding a set of pixel groups corresponding to the current voxel through spatial position transformation. Then, the current voxel is valued by using the pixel set according to certain rules, such as the nearest neighbor voxel method based on a pixel value, or various interpolation algorithms using multiple pixel values (distance weighted interpolation, median filter interpolation, etc.). The specific valuation method can be determined according to the actual situation, which is not limited here.

[0055] The function-based mapping is to construct the mapping function relationship between the pixels in the two-dimensional ultrasound image and the voxels in the abdominal aorta volume data, and to map based on the mapping function relationship. The specific process includes: constructing the mapping function relationship between the pixels and the voxels according to the pixel points and their spatial position information in the two-dimensional ultrasound image, fitting the mapping function relationship, and calculating the voxel value of the voxel in the volume data according to the fitted mapping function relationship.

[0056] It can be understood that the selection of the mapping mode needs to be comprehensively considered based on factors such as the three-dimensional ultrasound image imaging effect expected to be achieved and the image processing time consumption, and the specific mapping mode used can be adjusted according to the actual situation, which is not limited here. After the voxel value mapping, the voxel value of the abdominal aorta volume data can be obtained, and the abdominal aorta volume data is given the voxel value, that is, the three-dimensional ultrasound image of the abdominal aorta is obtained.

[0057] 103. Displaying the three-dimensional ultrasound image of the abdominal aorta.

[0058] Specifically, after obtaining the voxel value of the abdominal aorta volume data, the abdominal aorta volume data can be rendered and displayed according to the voxel value. The display method can include surface rendering, volume rendering, etc. The volume rendering method can specifically include light projection algorithm, shear-warp algorithm, frequency domain volume rendering algorithm, and snowball algorithm, etc. This is not limited here.

[0059] Take the light ray projection algorithm as an example for illustration. Based on different display purposes, the maximum value, minimum value, average value (X-ray mode) and other attributes on the light ray projection path can be selected for display, or the effect of the light source can be increased according to the light illumination model. It can be understood that while displaying the three-dimensional ultrasound image of the abdominal aorta, various orthogonal cross-sectional displays included in the three-dimensional ultrasound image can also be provided, thereby providing intuitive and rich abdominal aorta structure information.

[0060] As can be seen from the above technical solutions, the abdominal aorta imaging method provided by the embodiments of the present application reconstructs the three-dimensional ultrasound image of the abdominal aorta, and can display the three-dimensional ultrasound image. The three-dimensional ultrasound image intuitively displays the overall structure information of the abdominal aorta, and the provided abdominal aorta information is more abundant, so that the accuracy of the disease diagnosis result based on the three-dimensional ultrasound image is also higher.

[0061] It should be noted that in another embodiment of the present application, after obtaining the three-dimensional ultrasound image of the abdominal aorta in step 102, a target cross-sectional image such as an abdominal aorta transverse cross-sectional image and / or an abdominal aorta longitudinal cross-sectional image can be selected from the three-dimensional ultrasound image; and the related information of the abdominal aorta is calculated based on the target cross-sectional image. In this embodiment, the step of displaying the three-dimensional ultrasound image in step 103 can be performed or omitted. The process of selecting the target cross-sectional image and calculating the related information of the abdominal aorta is described in detail below, and will not be described here.

[0062] The three-dimensional ultrasound image obtained in the above embodiments can be the overall structure of the abdominal space, which includes not only the abdominal aorta blood vessel but also other human organ or tissue structures. In order to facilitate the diagnosis of the abdominal aorta, the present application can also identify the three-dimensional ultrasound image after obtaining the three-dimensional ultrasound image to locate the abdominal aorta region in the three-dimensional ultrasound image, and intuitively display the position of the abdominal aorta in the three-dimensional ultrasound image to medical personnel, so that the medical personnel can observe or measure the abdominal aorta in the region for disease diagnosis operations.

[0063] See Figure 2 which shows another embodiment of the abdominal aorta imaging method, which is based on the embodiment shown in Figure 1 The embodiment shown in the embodiment further includes step 204. It should be noted that steps 201 to 203 in the embodiment are similar to the above Figure 1 The steps 101 to 103 in the corresponding embodiment are similar, and will not be described here. Only step S204 will be described below.

[0064] 204, based on the structural characteristics of the abdominal aorta, image recognition is performed on the three-dimensional ultrasound image to identify the abdominal aorta region in the three-dimensional ultrasound image.

[0065] Specifically, the abdominal aorta has its own structural characteristics, which can include anatomical features of the abdominal aorta, image features of the abdominal aorta, or other features. The abdominal aorta region is located from the three-dimensional ultrasound image according to the structural characteristics.

[0066] As shown in Figure 3 The present application provides two ways to determine the abdominal aorta region: 1. Identify the abdominal aorta region based on three-dimensional body data. 2. Identify the abdominal aorta region based on two-dimensional ultrasound images. The two methods are described below.

[0067] 1. Identify the abdominal aorta region based on three-dimensional ultrasound image three-dimensional body data.

[0068] Specifically, a deep learning image segmentation method based on three-dimensional body data can be used to identify the abdominal aorta region from a three-dimensional ultrasound image.

[0069] First, a pre-trained neural network model is needed, which is trained by a deep learning algorithm on three-dimensional ultrasound body data with abdominal aorta region labels. The neural network model can be specifically selected from 3D Unet, V-Net, DeepMedic, Thickened 2D Networks, etc. which perform well in three-dimensional data recognition problems. It can be understood that the architecture of the neural network model can be adjusted according to application requirements in actual implementation, and the specific place is not limited.

[0070] During training, an ultrasound body database can be constructed, which stores three-dimensional ultrasound body data. The three-dimensional ultrasound body data is labeled, and then the labeled three-dimensional ultrasound body data is used to train the neural network model. The deep learning algorithm used by the neural network model can optimize the neural network model itself, so that the trained neural network model has the ability to identify whether there is an abdominal aorta region in the input image and mark the range of the abdominal aorta region.

[0071] After training the neural network model, the three-dimensional ultrasound image can be input into the trained neural network model to obtain the recognition result output by the neural network based on the learned abdominal aorta features. The recognition result is used to represent the spatial position of the abdominal aorta region in the three-dimensional ultrasound image. Exemplarily, the recognition result can be specifically the boundary range of the abdominal aorta.

[0072] 2. Identify the abdominal aorta region based on two-dimensional ultrasound images in three-dimensional ultrasound images.

[0073] In actual implementation, the recognition of the abdominal aorta can also be completed based on multiple two-dimensional ultrasound images in the three-dimensional ultrasound image. Specifically, the three steps of selecting a section, region section recognition, and region section splicing can be included.

[0074] 2.1, selecting a section. Multiple two-dimensional abdominal aorta section images are selected from the three-dimensional ultrasound image, and the abdominal aorta section images include abdominal aorta transverse section images and / or abdominal aorta longitudinal section images.

[0075] Specifically, multiple abdominal aorta section images participating in the recognition process are selected from the three-dimensional ultrasound image, and the abdominal aorta section images are two-dimensional images. In order to ensure the recognition effect of the abdominal aorta region, all abdominal aorta transverse section images or all abdominal aorta longitudinal section images included in the three-dimensional ultrasound image can be selected for recognition. Alternatively, considering the recognition efficiency, part of the abdominal aorta section images in the three-dimensional ultrasound image can also be selected to participate in the recognition process, such as extracting abdominal aorta section images at intervals of one frame or several frames to reduce the amount of data participating in the recognition process, and then obtaining the recognition results of other frames of abdominal aorta section images in an interpolation manner, and then positioning the abdominal aorta region according to the overall recognition result. The specific selection method of the abdominal aorta section image can be determined according to the actual situation, which is not limited here.

[0076] 2.2, region section recognition. Based on the structural characteristics of the abdominal aorta, the abdominal aorta region is recognized in the abdominal aorta section image.

[0077] Specifically, as described above, the abdominal aorta has its own structural characteristics, and based on the structural characteristics, the abdominal aorta region in the abdominal aorta section image can be recognized. The recognized abdominal aorta region can be referred to as an abdominal aorta region section.

[0078] See Figure 4, as to the identification of the abdominal aorta region section, the application provides the following identification methods: deep learning-based identification and non-deep learning-based identification. The deep learning-based identification refers to identifying the abdominal aorta region based on a pre-trained neural network model. The non-deep learning-based identification refers to matching the image features of the pre-constructed image library with the abdominal aorta section image to obtain the abdominal aorta region. It should be noted that the difference between the two identification methods is that the former can use deep learning algorithm to automatically learn image features from image database and use the learned image features for identification, while the latter needs to use manually set image features for identification. More specifically, the deep learning-based identification can include two specific implementation methods: deep learning-based target detection method and deep learning-based image segmentation method. The non-deep learning-based identification can also include two specific implementation methods: non-deep learning-based target detection method and non-deep learning-based image segmentation method. The four specific implementation methods are described below.

[0079] (1) Deep learning-based target detection method.

[0080] Specifically, a pre-trained neural network model is obtained. When training the neural network model, the training set used includes multiple two-dimensional ultrasound images, and the two-dimensional ultrasound images have annotation information. That is, if the abdominal aorta region exists in the two-dimensional ultrasound image, the region of interest, i.e., the abdominal aorta region, is labeled using a regular shape box surrounding the region. The neural network model can optimize itself based on the image information in the box and the position information of the box, thereby enabling the neural network model to have the ability to identify whether the two-dimensional ultrasound image includes the abdominal aorta region, and in the case of yes, to have the ability to use a regular shape box to mark the abdominal aorta region. Exemplarily, the neural network model can be a Faster-RCNN, YOLO, SSD, RetinaNet, EfficientDet, FCOS, CenterNet, etc. detector. It can be understood that the form of the neural network model can be adjusted during actual implementation of the scheme, which is not limited here.

[0081] The abdominal aorta section image is input into the neural network model to obtain the identification result output by the neural network model based on the structural characteristics of the abdominal aorta. Since the annotation information in the training set is a regular shape box surrounding the abdominal aorta region, the identification result includes a regular shape box surrounding the abdominal aorta region. Exemplarily, as shown in FIG. 4, the abdominal aorta region in the abdominal aorta section image is marked by a regular shape box. Figure 5AAs shown, a regular rectangular frame is used to locate the abdominal aorta region in a two-dimensional ultrasound image. It should be noted that if the abdominal aorta region does not exist in the abdominal aorta section image, no regular rectangular frame will appear. Therefore, based on the recognition results, it is possible to determine whether the abdominal aorta region exists in the abdominal aorta section image and, if so, its approximate location and extent.

[0082] One specific implementation of this method is to use deep learning-based bounding-box detection and recognition. Specifically, by stacking convolutional layers and fully connected layers, the constructed image database is trained to learn features and regress parameters. For an input abdominal image to be identified, a neural network model can be used to directly regress the corresponding detection box of the region of interest and simultaneously obtain the classification of the tissue structure within the region of interest. Common neural network models are described above and will not be repeated here.

[0083] (2) Image segmentation method based on deep learning.

[0084] Specifically, a pre-trained neural network model is obtained, and when the neural network model is trained, the training set used includes multiple two-dimensional ultrasound images and the two-dimensional ultrasound images have annotation information. If there is an abdominal aorta region in the two-dimensional ultrasound image, the boundary line of the abdominal aorta region is used to mark the specific boundary range of the abdominal aorta region. Similarly, the neural network model can optimize itself based on the image information within the boundary line and the position information of the boundary line, so that the neural network model has the ability to identify whether the two-dimensional ultrasound image includes the abdominal aorta region, and the ability to use the boundary line to mark the abdominal aorta region if so. Exemplarily, the neural network model can be a network model such as FCN, Unet, SegNet, DeepLab, Mask RCNN, etc. It can be understood that the form of the neural network model can be adjusted during the actual implementation of the solution, and this is not limited here. After training based on the training set, the neural network model has the ability to mark the abdominal aorta region in the two-dimensional ultrasound image with a boundary line.

[0085] The cross-sectional image of the abdominal aorta is input into the neural network model to obtain the recognition result output by the neural network model based on the structural features of the abdominal aorta. Since the labeled information in the training set is the boundary line marking the abdominal aorta area, the corresponding recognition result includes the boundary line marking the abdominal aorta area. For example, Figure 5B As shown in Figure 2, the abdominal aorta region is located using boundary lines in a two-dimensional ultrasound image. It should be noted that if the abdominal aorta region does not exist in the abdominal aorta section image, no boundary lines appear. Therefore, based on the recognition results, it is possible to determine whether the abdominal aorta region exists in the abdominal aorta section image and, if so, the specific boundary range of the abdominal aorta region.

[0086] One specific implementation of the method is an end-to-end semantic segmentation network method based on deep learning. Specifically, the neural network model used by this method is similar in structure to the deep learning-based bounding box detection method, except that the fully connected layer is removed and an up-sampling or de-convolution layer is added to make the size of the input abdominal image to be recognized and the output abdominal image to be recognized the same, so as to directly obtain the abdominal aortic region of the input abdominal image to be recognized and its corresponding category. Common neural network models are as described above and will not be repeated here.

[0087] The above two recognition methods of the abdominal aortic region, in the process of positioning and recognizing the region of interest (abdominal aortic region), use machine learning methods to learn the features or rules that can distinguish the target region and the non-target region in the image database, and then position and recognize the region of interest in other images to be recognized according to the features or rules. Specifically, it can include the following two steps: first, build an image database, which usually contains multiple abdominal images and the corresponding abdominal aortic region labeling results. The labeling results can be set according to the actual task needs, which can be a ROI (region of interest) box containing the abdominal aorta, or a Mask (mask) for accurate segmentation of the abdominal aorta. If the actual task requires positioning multiple categories of abdominal aortas, the category of each ROI box or Mask also needs to be specified; the second step is positioning and recognition, that is, after building the image database, the features or rules that can distinguish the abdominal aortic region and the non-abdominal aortic region in the image database are learned based on the machine learning algorithm to realize the positioning and recognition of the abdominal aortic region in the abdominal image to be recognized. It can be seen that in the two recognition methods of the abdominal aortic region, both are based on a neural network model constructed by deep learning. The annotation information form used by neural network models with different structures is different. The annotation information of the target detection method is a regular shape box surrounding the abdominal aortic region, and the annotation information of the image segmentation method is the boundary line of the abdominal aortic region, but the annotation information can also be other shapes, as long as it can mark the approximate position range of the abdominal aortic region. Based on this, those skilled in the art can think of using other structures of neural network model to realize the recognition of the abdominal aortic region. Therefore, the recognition method based on neural network can be summarized as follows: obtaining a pre-trained neural network model, the neural network model is trained by a deep learning algorithm on multiple two-dimensional abdominal aortic cross-sectional image samples with annotation information, the annotation information is used to represent the abdominal aortic region in the abdominal aortic cross-sectional image sample; input the abdominal aortic cross-sectional image into the neural network model to obtain the recognition result output by the neural network model based on the structural characteristics of the abdominal aorta, the recognition result is used to represent the abdominal aortic region included in the abdominal aortic cross-sectional image.

[0088] (3) Non-deep learning-based target detection method.

[0089] An image library is pre-constructed, and the image library contains abdominal aorta two-dimensional images, and the abdominal aorta two-dimensional images are pre-labeled with a region of interest, i.e., an abdominal aorta region, using a regular shape box. Further, an image feature of the abdominal aorta region in the abdominal aorta two-dimensional image can be obtained. The image feature is used for comparison with an image feature of an abdominal aorta cross-section image to be identified in actual implementation. It should be noted that, in order to speed up the processing efficiency, the image feature can be pre-processed and stored in the image library, or in order to reduce the storage space, the image feature is not pre-stored but is processed in real time after the abdominal aorta cross-section image to be identified is obtained.

[0090] In actual implementation, after the abdominal aorta cross-section image is obtained, the target detection algorithm is used to detect the image region of interest from the abdominal aorta cross-section image based on the structural characteristics of the abdominal aorta, and the regular shape box is used to label the image region of interest, and the image feature of the image region of interest is extracted. For example, a group of candidate image regions of interest are framed in the abdominal aorta cross-section image by sliding window or selective search, and then the candidate frame regions are respectively subjected to feature extraction, and the PCA, LDA, HOG, Harr, LBP, SIFT, texture, or neural network image features can be extracted.

[0091] The extracted image feature is matched with the image feature of the abdominal aorta region pre-labeled by the regular shape box, and discriminators such as linear classifiers, support vector machines (SVM), nearest neighbors (KNN), random forests, or simple neural networks can be used for matching. According to the matching result, it can be determined whether the image region of interest of the abdominal aorta cross-section image contains the abdominal aorta region.

[0092] (4) Non-deep learning-based image segmentation method.

[0093] An image library is pre-constructed, and the image library is set in a manner similar to the above-mentioned image library-based target segmentation method, except that the abdominal aorta region is labeled by a boundary line instead of a regular shape box, and the specific boundary range of the abdominal aorta region is labeled by the boundary line. Other descriptions can be found in the above content, and will not be repeated here.

[0094] In actual implementation, after obtaining the abdominal aorta cross-section image, based on the structural characteristics of the abdominal aorta, an image segmentation algorithm is used to segment the image region of interest from the abdominal aorta cross-section image, and the contour of the image region of interest is marked. For example, the image is pre-segmented by threshold segmentation, snake, level set, GraphCut and other image processing methods, a group of candidate target structure boundary ranges are selected as the image region of interest in the pre-segmented image, and then the features of the region surrounded by the boundary range of the image region of interest are extracted, such as PCA, LDA, HOG, Harr, LBP, SIFT and other feature types, or neural network extracted feature types.

[0095] The extracted image features are matched with the image features of the abdominal aorta region pre-marked by the boundary line, such as linear classifier, support vector machine (Support Vector Machine, SVM) or simple neural network classifier. According to the matching result, it can be determined whether the abdominal aorta cross-section image region of interest contains the abdominal aorta region.

[0096] One specific implementation of the method is to first locate the abdominal aorta region, which can be a target region of interest (ROI) or a mask (Mask), then extract features from the located region, which can include PCA features, LDA features, Harr features, texture features, etc., or features extracted by a deep neural network, then use a discriminator to match and classify the extracted features and image features extracted from the image database to determine whether the abdominal aorta cross-section image region of interest contains the abdominal aorta region. The discriminator can be KNN, SVM, random forest, neural network, etc.

[0097] The above two recognition methods of the abdominal aorta region are both based on non-deep learning algorithms, the difference is that the annotation information in the image database is different, the annotation information of the target detection method is a regular shape box surrounding the abdominal aorta region, and the annotation information of the image segmentation method is the boundary line of the abdominal aorta region, but the annotation information can also be other shapes, as long as it can mark the approximate position range of the abdominal aorta region. Based on the above two implementation methods, those skilled in the art can think of other implementation methods that also use non-deep learning algorithms for feature matching to achieve the purpose of recognizing the abdominal aorta region. For reference Figure 6The recognition method based on a non-deep learning algorithm can be summarized as follows: based on the structural characteristics of the abdominal aorta, an image region of interest is selected from the abdominal aorta cross-sectional image, and image features of the image region of interest are extracted; an abdominal aorta two-dimensional image is obtained from a pre-constructed image library, and image features of an abdominal aorta region pre-labeled in the abdominal aorta two-dimensional image are obtained; and the extracted image features are matched with the image features of the pre-labeled abdominal aorta region to determine whether the image region of interest contains the abdominal aorta region.

[0098] The various recognition schemes provided above can all recognize the abdominal aorta region in the abdominal aorta cross-sectional image based on the structural characteristics of the abdominal aorta. In actual implementation, the specific recognition manner of the abdominal aorta region can be selected according to actual needs, and can be adjusted according to actual conditions, and the specific implementation is not limited here.

[0099] 2.3, region cross section splicing. Based on the abdominal aorta region in the multiple-frame two-dimensional abdominal aorta cross-sectional images, the abdominal aorta region of the three-dimensional ultrasound image is spliced.

[0100] It can be understood that the abdominal aorta region recognized from each frame of abdominal aorta cross-sectional image is a cross section. If the abdominal aorta cross-sectional image includes both abdominal aorta transverse cross-sectional images and abdominal aorta longitudinal cross-sectional images, the same type of abdominal aorta cross-sectional image cross sections are spliced and combined to obtain a three-dimensional abdominal aorta region.

[0101] The above detailed two determination manners of the abdominal aorta region, wherein step 1 describes a manner based on three-dimensional body data, and steps 2.1-2.3 describe a manner of obtaining a three-dimensional abdominal aorta region based on two-dimensional ultrasound images. The first implementation manner considers the correlation between adjacent images in the image sequence, and the second implementation manner utilizes the image information in the entire two-dimensional ultrasound image to make the image information involved in the processing more comprehensive. Both manners can improve the accuracy of the abdominal aorta region recognition result to some extent.

[0102] As can be seen from the embodiments shown in Figure 2 The technical scheme can display a three-dimensional abdominal aorta region, which not only enables medical staff to more intuitively determine the position of the abdominal aorta in the three-dimensional ultrasound image, but also more comprehensively displays the characteristics of the abdominal aorta, and the accuracy of disease diagnosis by medical staff is also higher.

[0103] It should be noted that in the above positioning process of the abdominal aorta, the target cross-sectional image can be obtained, and based on the target cross-sectional image, information related to whether the target cross-sectional image contains a dissection aneurysm can also be recognized. For specific descriptions, please refer to the following identification process of the dissection aneurysm, which will not be described here.

[0104] InFigure 1 On the basis of displaying the three-dimensional ultrasound image, further, some parameter information of the abdominal aorta region can be calculated and analyzed based on the three-dimensional ultrasound image. As shown in the method embodiment, the method embodiment includes steps 701-704. It should be noted that the execution sequence of the step 704 and the step 703 of displaying the three-dimensional ultrasound image is not limited to Figure 7 As shown, the method embodiment includes steps 701-704. It should be noted that the execution sequence of the step 704 and the step 703 of displaying the three-dimensional ultrasound image is not limited to Figure 7 As shown, the method embodiment includes steps 701-704. It should be noted that the execution sequence of the step 704 and the step 703 of displaying the three-dimensional ultrasound image is not limited to Figure 1 The second reconstruction mode of reconstructing the three-dimensional ultrasound image in the embodiment will not be described here. The following will only describe the added step 704.

[0105] 704. Select a target cross-sectional image from the three-dimensional ultrasound image, and calculate the related information of the abdominal aorta based on the target cross-sectional image.

[0106] Specifically, the cross-sectional image is selected from the reconstructed three-dimensional ultrasound image in a certain way. The selected cross-sectional image is referred to as a target cross-sectional image. The target cross-sectional image includes an abdominal aorta transverse cross-sectional image and / or an abdominal aorta longitudinal cross-sectional image.

[0107] In an embodiment, the target cross-sectional image can be manually selected by medical personnel or automatically selected by the device. The following will describe the two different selection methods.

[0108] (1) Manual selection by medical personnel. The ultrasound detection device can provide a function of rotating or moving the three-dimensional ultrasound image to display the three-dimensional ultrasound image under different spatial perspectives, thereby helping the medical personnel to select the cross-sectional image of interest. For example, the medical personnel can select one or more cross-sectional images of interest from the displayed three-dimensional ultrasound image or from the positioned abdominal aorta region, or the medical personnel can input a target perspective or select a specific perspective, etc. The ultrasound detection device determines the cross-sectional image of interest according to the information. In response to the operation of the user selecting the abdominal aorta cross-sectional image of interest in the three-dimensional ultrasound image under the target spatial perspective, the abdominal aorta cross-sectional image of interest selected by the medical personnel is determined as the target cross-sectional image.

[0109] (2) Automatic selection according to feature recognition of the three-dimensional ultrasound image.

[0110] Based on the structural features of the abdominal aorta, the three-dimensional ultrasound image is subjected to image recognition to identify the abdominal aorta region in the three-dimensional ultrasound image; the center line of the abdominal aorta region is determined, and the target cross-sectional image is selected based on the center line of the abdominal aorta region. The identification of the abdominal aorta region from the three-dimensional ultrasound image can be achieved by the implementation mode of the step 204 in the above Figure 2 Embodiment, which will not be described here.

[0111] The center line of the abdominal aorta region can represent the central position of the abdominal aorta region, can be directly measured, or can also be obtained by fitting. The implementation process of the fitting method includes: selecting at least one abdominal aorta transverse section image from the three-dimensional ultrasound image; determining the central position from the abdominal aorta region identified in the at least one abdominal aorta transverse section image; and fitting the center line of the abdominal aorta region based on the central position of the abdominal aorta region in the at least one abdominal aorta transverse section image.

[0112] Specifically, different center line fitting methods are used for the abdominal aorta region extracted by the image segmentation method or the target detection method.

[0113] If the abdominal aorta region is determined based on the image segmentation method, the region boundary has been marked by the boundary line, so a circle can be fitted on the abdominal aorta region boundary in the selected abdominal aorta transverse section image by using least square fitting, Hough transform, RANSAC, etc., and the center position of the circle is taken as the central position, and then the central positions are fitted into the center line of the abdominal aorta region by using machine learning methods such as least square fitting, ridge regression, local weighted regression, KNN, SVM, etc.

[0114] If the abdominal aorta region is determined based on the target detection method, the region boundary has been marked by the regular shape box, so the center point of the regular shape box can be taken as the central position, and the center line is obtained based on the central position; or further, the boundary range of the abdominal aorta region is refined by using image processing methods such as edge extraction, threshold segmentation, region-based segmentation, and then the central position is obtained from the boundary range and the center line is fitted.

[0115] After obtaining the center line of the abdominal aorta region, an abdominal aorta longitudinal section image passing through the center line can be selected as a target section image, or an abdominal aorta transverse section image orthogonal to the abdominal aorta longitudinal section image can be selected as a target section image, or both of them can be selected, which is not limited here. If the target section image is displayed to medical personnel, the image perspective is better, which provides more information for medical personnel and is more conducive to medical personnel to perform measurement and other operations.

[0116] The target section image can be further displayed to medical personnel for observation or further execution of other steps such as measurement. Or the target section image is not displayed, and various related information of the abdominal aorta is calculated based on the target section image, and the related information is further displayed to medical personnel. The abdominal aorta related information calculated based on the target section image includes but is not limited to pipe diameter, whether it contains an aneurysm, etc., which is not limited in the present application.

[0117] It should be noted that the target cross-section image obtained in some embodiments may not have the identification result of the abdominal aorta region, for example, if the target cross-section image is autonomously selected by the user. Then, the abdominal aorta region in the target cross-section image can be identified by the following two methods, and then the relevant information of the abdominal aorta is calculated based on the identification result.

[0118] The first method is that the medical staff manually marks the abdominal aorta region in the target cross-section. That is, in response to the operation of the medical staff marking the boundary of the abdominal aorta on the target cross-section image, the image region surrounded by the abdominal aorta boundary is determined as the abdominal aorta region. For example, the ultrasound detection device provides an edge drawing tool, and the medical staff can use the tool to mark the boundary of the region of interest, i.e., the abdominal aorta region. The second method is to identify the abdominal aorta region in the target cross-section image based on the structural characteristics of the abdominal aorta. The specific identification process can be referred to the implementation of step 204 in the above-mentioned embodiment, which will not be repeated here. Figure 2 The implementation of the embodiment of step 204 can be referred to for the implementation of the second method, which will not be repeated here.

[0119] After determining the target cross-section image, the relevant information of the abdominal aorta can also be calculated to assist the doctor in completing the diagnosis process. The relevant information of the abdominal aorta can include the pipe diameter of the abdominal aorta and / or the relevant information of the abdominal aortic aneurysm. The two types of information are described below.

[0120] 1. The pipe diameter of the abdominal aorta. The pipe diameter of the abdominal aorta can specifically include the long-axis direction inner diameter of the abdominal aorta and / or the short-axis direction inner diameter of the abdominal aorta.

[0121] Figure 8A For the longitudinal cross-section image of the abdominal aorta, the abdominal aorta region is identified, and the center line of the abdominal aorta region is determined. Figure 8A The long-axis direction inner diameter of the abdominal aorta is explained, and the calculation process of the inner diameter is described. The long-axis direction inner diameter of the abdominal aorta can also be referred to as the longitudinal cross-section inner diameter, which refers to the inner diameter of the abdominal aorta lumen obtained according to the longitudinal cross-section image of the abdominal aorta. It can be obtained by the following two methods.

[0122] If the abdominal aorta longitudinal cross-section image has a boundary line marking the abdominal aorta region, the center line of the abdominal aorta region is determined, and the long-axis direction inner diameter of the abdominal aorta is calculated based on the boundary line and the center line. The specific calculation process is, for example Figure 8AAs shown, a plurality of points are sampled on the boundary of the anterior wall or the posterior wall of the abdominal aorta vessel, and for each sampling point, a corresponding point on the other side wall boundary is found, such as generating a straight line passing through the sampling point and perpendicular to the center line of the abdominal aorta, finding the point where the straight line intersects the other side of the tube wall, and then determining the distance between the two points on the tube wall as the long axis direction inner diameter of the abdominal aorta at the sampling point. Each sampling point can determine a long axis direction inner diameter of the abdominal aorta in this way, and then the long axis direction inner diameters corresponding to all sampling points are statistically analyzed, such as maximum value, minimum value, mean value, variance, etc., to obtain the long axis direction inner diameter of the abdominal aorta.

[0123] If the longitudinal section image of the abdominal aorta has a regular shape box marking the abdominal aorta region, the center line of the abdominal aorta region is determined, and the long axis direction inner diameter of the abdominal aorta is calculated based on the regular shape box and the center line. The specific calculation process is, for example, to take one side boundary of the regular shape box surrounding the abdominal aorta as the boundary of one side of the abdominal aorta vessel, and select some points from the boundary as sampling points. Similar to the above method, the long axis direction inner diameter of the abdominal aorta is measured. It should be noted that the measurement result obtained from the boundary of the regular shape box may not be accurate enough, so based on the regular shape box, further image processing methods such as edge extraction, threshold segmentation, region-based segmentation are used to obtain more accurate boundaries of the abdominal aorta, and then the above boundary range-based method is used for sampling measurement and statistical analysis, etc. Steps to obtain more accurate long axis direction inner diameter of the abdominal aorta.

[0124] Figure 8B For the abdominal aorta transverse section image, in combination Figure 8B The short axis direction inner diameter of the abdominal aorta is explained and the calculation process of the inner diameter is described. The short axis direction inner diameter of the abdominal aorta can also be referred to as the transverse section inner diameter, which refers to the inner diameter of the abdominal aorta lumen obtained from the abdominal aorta transverse section image.

[0125] If the abdominal aorta transverse section image has a boundary line marking the abdominal aorta region, the short axis direction inner diameter of the abdominal aorta is calculated based on the boundary line. For example, the circumferential radius of the boundary range, the maximum inner diameter of the boundary range, or the length of the boundary range of the longitudinal meridian of interest can be calculated as the short axis direction inner diameter of the abdominal aorta.

[0126] If the abdominal aorta transverse section image has a regular shape box marking the abdominal aorta region, the length or width of the regular shape box is determined as the short axis direction inner diameter of the abdominal aorta. For example, the length or width of the regular shape box can be selected as the inner diameter, or further image processing methods such as edge extraction, threshold segmentation, region-based segmentation are used to refine the boundary range of the abdominal aorta region within the regular shape box, and then the above boundary range-based method is used for sampling measurement and statistical analysis, etc. Steps to obtain more accurate long axis direction inner diameter of the abdominal aorta.

[0127] The short axis direction inner diameter of each frame of the abdominal aorta transverse section image can be measured. The medical staff can slide the trackball of the ultrasound diagnosis equipment to select the abdominal aorta transverse section image of different frames, and then view the inner diameter of the abdominal aorta transverse section image of the current frame. Of course, the medical staff can also view the inner diameter of each frame of the abdominal aorta longitudinal section image in this way.

[0128] Further, in order to show the differences of the inner diameters measured by the frames of images, the long axis direction inner diameter change curve and / or the short axis direction change curve can be drawn.

[0129] Specifically, according to the multiple frames of the abdominal aorta transverse section images and the short axis direction inner diameters corresponding to the abdominal aorta transverse section images, the short axis inner diameter change curve is generated. The horizontal coordinate of the short axis inner diameter change curve is the position of the abdominal aorta transverse section image on the center line of the abdominal aorta, and the vertical coordinate is the short axis direction inner diameter measured by the abdominal aorta transverse section image.

[0130] According to the multiple frames of the abdominal aorta longitudinal section images and the long axis direction inner diameters corresponding to the abdominal aorta longitudinal section images, the long axis inner diameter change curve is generated. The horizontal coordinate of the long axis inner diameter change curve is the position of the abdominal aorta longitudinal section image on the center line of the abdominal aorta, and the vertical coordinate is the long axis direction inner diameter corresponding to the abdominal aorta longitudinal section image. Exemplarily, Figure 9 An example of the short axis inner diameter change curve is shown.

[0131] In order to intuitively prompt some inner diameter information to the medical staff, some information in the inner diameter change curve can be marked. Generally, if the inner diameter is greater than a certain value, it means that there is a greater possibility of abdominal aortic aneurysm at this position, so in an embodiment, the maximum value of the short axis direction inner diameter can be marked on the short axis inner diameter change curve, and the abdominal aorta transverse section image corresponding to the maximum value of the inner diameter is further displayed. Similarly, the maximum value of the long axis direction inner diameter can also be marked on the long axis inner diameter change curve, and the abdominal aorta transverse section image corresponding to the maximum value of the inner diameter is further displayed. The marking method can be highlighting, adding an indication, adding a color, and various other methods with prompting effect, which are not specifically limited in the present application.

[0132] 2. Related information of abdominal aortic aneurysm. Based on the target section image, the related information of the abdominal aortic aneurysm, such as the related information of the dissection aneurysm and / or the related information of the non-dissection aneurysm, can be calculated.

[0133] Specifically, abdominal aortic aneurysm refers to the aneurysmal dilatation of abdominal aorta, which is the most common aortic abnormality, and can be accompanied by thrombosis, intimal dissection or rupture. The rupture of abdominal aortic aneurysm can even lead to death, which seriously threatens people's life safety. Compared with the normal artery diameter, the abdominal aortic aneurysm is usually defined as more than 50% increase in diameter. For abdominal aorta, this means that the diameter is more than 3.0 cm to consider aneurysm. Abdominal aortic aneurysm can be divided into dissection aneurysm and non-dissection aneurysm.

[0134] The information related to the dissection aneurysm can be obtained by a neural network model trained by a deep learning algorithm on abdominal aortic cross-sectional images with labeled information related to the dissection aneurysm. The target cross-sectional image is input into the pre-trained neural network model to obtain the information related to the dissection aneurysm output by the neural network model.

[0135] Depending on the type of deep learning algorithm used, the determination method of abdominal aortic aneurysm can be divided into image classification method, object detection method and image segmentation method.

[0136] In the image classification method based on deep learning, the types of neural network models used can include AlexNet, VGG, ResNet, Inception, MobileNet, etc. The training set used is an abdominal aortic cross-sectional image with labels, and the labels are used to indicate whether the abdominal aortic cross-sectional image contains a dissection aneurysm. After training the neural network model using the training set, the target cross-sectional image extracted from the three-dimensional ultrasound image is input into the trained neural network model. The neural network model calculates the probability of the target cross-sectional image containing a dissection aneurysm and not containing a dissection aneurysm, and outputs the label corresponding to the maximum probability as the recognition result of the target cross-sectional image. Based on the output result, it can be determined whether the target cross-sectional image contains a dissection aneurysm, that is, the information related to the dissection aneurysm output by the neural network model is whether the target cross-sectional image contains a dissection aneurysm. It can be understood that, in order to ensure the accuracy of the judgment result, multiple target cross-sectional images can be input at a time to obtain a more accurate recognition result.

[0137] In the target detection method based on deep learning, the neural network model used can include Faster-RCNN, YOLO, SSD, RetinaNet, EfficientDet, FCOS, CenterNet, etc. The training set used is an abdominal aortic section image with annotations, and the annotation information is the region position of the dissection aneurysm in the abdominal aortic section image. Specifically, when there is a dissection aneurysm in the abdominal aortic section image, a regular shape box surrounding the dissection aneurysm is used to mark it, and the position of the regular shape box can be represented by its coordinate information. After training the neural network model using the training set, the target section image extracted from the three-dimensional ultrasound image is input into the trained neural network model, and the neural network model will output whether there is a dissection aneurysm in the image and the region position of the dissection aneurysm when there is one. Specifically, the dissection aneurysm is surrounded by a regular shape box, and the coordinate information of the regular shape box can represent the region position of the dissection aneurysm in the abdominal aortic section image.

[0138] In the image segmentation method based on deep learning, the neural network model used can include FCN, Unet, SegNet, DeepLab, Mask RCNN, etc. The training set used has annotated abdominal aortic section images, and the annotation information is the region position of the dissection aneurysm in the abdominal aortic section image. Specifically, when there is a dissection aneurysm in the abdominal aortic section image, the contour of the dissection aneurysm is marked. After training the neural network model using the training set, the target section image extracted from the three-dimensional ultrasound image is input into the trained neural network model, and the neural network model will output whether there is a dissection aneurysm in the image and the region position of the dissection aneurysm when there is one. Specifically, the region range of the dissection aneurysm is delineated using the boundary line.

[0139] It can be seen that both the target detection method based on deep learning and the image segmentation method based on deep learning can output the region position of the dissection aneurysm in the target section image.

[0140] To improve the recognition efficiency of abdominal aortic aneurysm, the recognition of dissection aneurysm can be performed during the recognition of abdominal aortic region. That is, the target section image is input into a pre-trained neural network model to obtain the dissection aneurysm-related information output by the neural network model. The neural network model is obtained by training the abdominal aortic section image, and the abdominal aortic section image has annotation information related to the dissection aneurysm. The training method includes deep learning algorithms such as image classification, target detection, or image segmentation based on deep learning, etc. It can also include other non-deep learning algorithms.

[0141] Specifically, the target cross-section image can be obtained in the identification process of the abdominal aorta region, and the identified target cross-section image is input into the neural network model for identifying the dissection aneurysm, so as to complete the identification of the abdominal aorta region and the dissection aneurysm at the same time. The identification process of the dissection aneurysm is completed in the identification process of the abdominal aorta region part of the three-dimensional ultrasound image, and there is no mutual conflict in the execution order.

[0142] In addition to the information related to the dissection aneurysm, information related to the non-dissection aneurysm can also be determined, and the related information can be whether the target cross-section image contains the non-dissection aneurysm, which can be determined based on the diameter measurement result of the abdominal aorta region. Specifically, according to the long axis direction inner diameter of the abdominal aorta and the short axis direction inner diameter of the abdominal aorta, it is determined whether the target cross-section image contains the non-dissection aneurysm, and the non-dissection aneurysm is prompted on the basis of being determined to be yes. The specific non-dissection aneurysm prompt mode can include:

[0143] (1) If the long axis direction inner diameter of the abdominal aorta exceeds the preset long axis inner diameter threshold or the short axis direction inner diameter of the abdominal aorta exceeds the preset short axis inner diameter threshold, the target cross-section image is prompted to contain the non-dissection aneurysm.

[0144] Specifically, the inner diameter threshold (long axis inner diameter threshold or short axis inner diameter threshold) corresponding to the non-dissection aneurysm can be set according to clinical experience, including but not limited to 3.0 centimeters. Specifically, the maximum long axis direction inner diameter or the maximum short axis direction inner diameter can be selected for judgment, and if the threshold is exceeded, it means that the abdominal aorta may have lesions and may have non-dissection aneurysm.

[0145] (2) If the average value of the long axis direction inner diameter of the abdominal aorta in multiple frames exceeds the preset long axis inner diameter average threshold, or if the average value of the short axis direction inner diameter of the abdominal aorta in multiple frames exceeds the preset short axis inner diameter average threshold, the target cross-section image is prompted to contain the non-dissection aneurysm.

[0146] Specifically, the average inner diameter of the abdominal aorta in the long axis direction or the short axis direction is calculated, and whether there is a non-dissection aneurysm is judged based on the relationship between the average inner diameter and the preset threshold. If the average inner diameter is greater than the preset threshold, it means that there may be a non-dissection aneurysm, which can be prompted.

[0147] (3) If the difference between the maximum value and the minimum value of the long axis direction inner diameter of the abdominal aorta accounts for the minimum value exceeds the preset long axis inner diameter ratio threshold, or if the difference between the maximum value and the minimum value of the short axis direction inner diameter of the abdominal aorta accounts for the minimum value exceeds the preset short axis inner diameter ratio threshold, the target cross-section image is prompted to contain the non-dissection aneurysm.

[0148] Specifically, the deformation of the abdominal aorta can also be used as a basis for judging whether the target cross-sectional image contains a non-dissecting aneurysm. The deformation can be determined by an inner diameter ratio, and a threshold value of the inner diameter ratio, such as 0.5, is preset. Regardless of the long axis direction inner diameter or the short axis direction inner diameter of the abdominal aorta, if the difference between the maximum value and the minimum value of the inner diameter exceeds the preset inner diameter ratio threshold value, it is considered that the abdominal aorta is deformed and the deformation is serious, and thus a non-dissecting aneurysm exists. It should be noted that the inner diameter ratio threshold value corresponding to the long axis direction inner diameter and the inner diameter ratio threshold value corresponding to the short axis direction inner diameter can be the same or different, and the present application does not make a specific limitation.

[0149] It should be noted that when automatically measuring the information of the abdominal aorta, if the target cross-sectional image is manually selected by medical personnel, the orientation information of the slice tool used when selecting the target cross-sectional image is first used to determine whether the target cross-sectional image belongs to a transverse cross-sectional type or a longitudinal cross-sectional type, and then the abdominal aorta region in the target cross-sectional image is positioned. The specific positioning method can be referred to the detailed description above. In addition to automatically measuring the inner diameter of the abdominal aorta, the ultrasound detection device can also provide a manual measurement tool for medical personnel to measure on the selected target cross-sectional image and display the measurement results in real time.

[0150] In the existing abdominal aorta detection process, medical personnel observe the abdominal aorta wall condition and lumen condition, such as the anteroposterior diameter and transverse diameter, according to two-dimensional ultrasound images, and also need to select the best two-dimensional ultrasound image in the abdominal aorta scanning process under multiple cross-sectional images of the lesion area that is beneficial to diagnosis, and manually measure the pipe diameter and other information. This operation process is relatively cumbersome and time-consuming, has low efficiency, and for medical personnel with limited level and experience, it may not be easy to accurately select the best cross-sectional image for evaluating the lesion area to make accurate manual measurement. However, the embodiments provided by the present application can automatically select the cross-sectional image, or can automatically position the abdominal aorta region and automatically measure the related important diagnostic indicators, thereby not only simplifying the operation process of medical personnel, but also improving the efficiency and accuracy of abdominal aorta ultrasound examination of medical personnel.

[0151] The abdominal aorta imaging method reconstructs a three-dimensional ultrasound image of the abdominal aorta through multiple two-dimensional ultrasound data and corresponding spatial position information, and displays the three-dimensional ultrasound image. The three-dimensional ultrasound image directly displays the overall structural information of the abdominal aorta, and the provided abdominal aorta information is more abundant, so that the accuracy of the disease diagnosis result based on the three-dimensional ultrasound image is also higher.

[0152] In order to ensure the application and implementation of the above method embodiments in practice, the present application also provides an ultrasound detection device, which specifically includes an ultrasound detection probe and a processor.

[0153] An ultrasonic detection probe configured to emit ultrasonic waves to an abdominal space of a target object;

[0154] A processor configured to control the ultrasonic detection probe to emit ultrasonic waves to an abdominal space of a target object, and control the ultrasonic detection probe to receive ultrasonic echo signals returned from the abdominal space of the target object; perform three-dimensional image reconstruction based on the ultrasonic echo signals to obtain a three-dimensional ultrasonic image of the abdominal aorta; select a target cross-sectional image from the three-dimensional ultrasonic image, the target cross-sectional image comprising an abdominal aorta transverse cross-sectional image and / or an abdominal aorta longitudinal cross-sectional image; and calculate relevant information of the abdominal aorta based on the target cross-sectional image.

[0155] In an implementation manner, the relevant information of the abdominal aorta calculated by the processor comprises at least one of the following: a long-axis direction inner diameter of the abdominal aorta, a short-axis direction inner diameter of the abdominal aorta, and relevant information of an abdominal aortic aneurysm; the relevant information of the abdominal aortic aneurysm comprises relevant information of a dissection aneurysm and / or relevant information of a non-dissection aneurysm.

[0156] In an implementation manner, the processor determines the relevant information of the dissection aneurysm based on the target cross-sectional image, specifically configured to: input the target cross-sectional image into a pre-trained neural network model to obtain dissection aneurysm related information output by the neural network model; the neural network model is obtained by training abdominal aorta cross-sectional images by a deep learning algorithm, and the abdominal aorta cross-sectional images have annotation information related to the dissection aneurysm.

[0157] In an implementation manner, the annotation information is whether the abdominal aorta cross-sectional image contains a dissection aneurysm, and the dissection aneurysm related information output by the neural network model is whether the target cross-sectional image contains a dissection aneurysm.

[0158] In an implementation manner, the annotation information is a region position of the dissection aneurysm in the abdominal aorta cross-sectional image, and the dissection aneurysm related information output by the neural network model is a region position of the dissection aneurysm in the target cross-sectional image.

[0159] In an implementation manner, the ultrasonic detection device further comprises a display; the processor determines the relevant information of the non-dissection aneurysm based on the target cross-sectional image, specifically configured to: generate prompt information according to the long-axis direction inner diameter of the abdominal aorta and / or the short-axis direction inner diameter of the abdominal aorta, the prompt information being used to prompt whether the target cross-sectional image contains a non-dissection aneurysm; and the display is configured to display the prompt information.

[0160] The application also provides an ultrasonic detection device, such as Figure 10As shown, the ultrasonic detection device can specifically include: an ultrasonic detection probe 1001, a processor 1002, and a display 1003.

[0161] The ultrasonic detection probe 1001 is configured to emit ultrasonic waves to the abdominal space of the target object.

[0162] The processor 1002 is configured to control the ultrasonic detection probe to emit ultrasonic waves to the abdominal space of the target object, and control the ultrasonic detection probe to receive ultrasonic echo signals returned from the abdominal space of the target object; perform three-dimensional image reconstruction based on the ultrasonic echo signals to obtain a three-dimensional ultrasonic image of the abdominal aorta.

[0163] It should be noted that the ultrasonic detection device can be integrated with an ultrasonic probe and an ultrasonic wave processing circuit, so as to process the ultrasonic waves by the ultrasonic detection device itself to obtain the two-dimensional ultrasonic data. Alternatively, the ultrasonic detection device can also be an ultrasonic imaging device and does not integrate an ultrasonic probe and an ultrasonic wave processing circuit, and after the two-dimensional ultrasonic data is processed by other ultrasonic detection devices, the two-dimensional ultrasonic data is sent to the ultrasonic imaging device, and the three-dimensional ultrasonic image is reconstructed by the ultrasonic imaging device.

[0164] The description of the two-dimensional ultrasonic data can be referred to Figure 1 Embodiments related content, which will not be described here.

[0165] In addition, in the case that the ultrasonic detection device is provided with an ultrasonic probe, it can also be provided with a spatial positioning device, such as a magnetic field spatial positioning device, to provide spatial position information of the two-dimensional ultrasonic data. For example, the magnetic field spatial positioning device can specifically include an electromagnetic field generator, a spatial position sensor (or receiver) and a microprocessor, which can sense the motion trajectory of the ultrasonic probe in the three-dimensional space, and provide the spatial coordinates and orientation information of each frame of two-dimensional ultrasonic image required for three-dimensional reconstruction. Alternatively, in the case that the ultrasonic detection device is not provided with an ultrasonic probe, the spatial position information of the two-dimensional ultrasonic data can be obtained by other devices that can measure the three-dimensional spatial position, and sent to the ultrasonic detection device together with the two-dimensional ultrasonic data.

[0166] The display 1003 is configured to display the three-dimensional ultrasonic image of the abdominal aorta.

[0167] In an implementation manner, the processor performs three-dimensional image reconstruction based on the ultrasonic echo signals to obtain the three-dimensional ultrasonic image of the abdominal aorta, and specifically is configured to:

[0168] The ultrasound echo signals are processed to obtain a plurality of two-dimensional ultrasound data of the abdominal aorta, and spatial position information corresponding to the two-dimensional ultrasound data is obtained, the spatial position information being used to represent position information of the abdominal aorta scanned by the two-dimensional ultrasound data in the abdominal space; and based on the plurality of two-dimensional ultrasound data and the spatial position information corresponding to the two-dimensional ultrasound data, a three-dimensional ultrasound image of the abdominal aorta is reconstructed. More specifically, abdominal aorta volume data of the three-dimensional ultrasound image to be reconstructed is determined; a mapping relationship between pixels in the plurality of two-dimensional ultrasound data and voxels of the abdominal aorta volume data is established according to the spatial position information of the plurality of two-dimensional ultrasound data; and pixel values of the plurality of two-dimensional ultrasound data are mapped to voxel values of the abdominal aorta volume data according to the mapping relationship, so as to obtain the three-dimensional ultrasound image of the abdominal aorta.

[0169] In an implementation manner, the processor is further configured to select a target cross-sectional image from the three-dimensional ultrasound image, the target cross-sectional image including an abdominal aorta transverse cross-sectional image and / or an abdominal aorta longitudinal cross-sectional image; and calculate related information of the abdominal aorta based on the target cross-sectional image.

[0170] In an implementation manner, the processor selects a target cross-sectional image from the three-dimensional ultrasound image, and specifically for:

[0171] The three-dimensional ultrasound image is rotated and / or moved to display the three-dimensional ultrasound image in different spatial perspectives; and in response to a user operation of selecting an abdominal aorta cross-sectional image of interest in the three-dimensional ultrasound image in a target spatial perspective, the abdominal aorta cross-sectional image of interest selected by the user is determined as the target cross-sectional image.

[0172] In an implementation manner, the processor selects a target cross-sectional image from the three-dimensional ultrasound image, and specifically for:

[0173] Based on structural characteristics of the abdominal aorta, image recognition is performed on the three-dimensional ultrasound image to identify an abdominal aorta region in the three-dimensional ultrasound image; a center line of the abdominal aorta region is determined; and based on the center line of the abdominal aorta region, a target cross-sectional image is selected.

[0174] In an implementation manner, the processor determines the center line of the abdominal aorta region, and specifically for:

[0175] At least one abdominal aorta transverse cross-sectional image is selected from the three-dimensional ultrasound image; a center position of the abdominal aorta region identified from the at least one abdominal aorta transverse cross-sectional image is determined; and based on the center position of the abdominal aorta region of the at least one abdominal aorta transverse cross-sectional image, a center line of the abdominal aorta region is fitted.

[0176] In one implementation, the processor selects a target section image based on a centerline of the abdominal aorta region, specifically for:

[0177] From the three-dimensional ultrasound image, a longitudinal section image of the abdominal aorta passing through the center line is selected as the target section image; or, from the three-dimensional ultrasound image, a transverse section image of the abdominal aorta orthogonal to the longitudinal section image of the abdominal aorta is selected as the target section image.

[0178] In one implementation, the processor is further configured to perform image recognition on the three-dimensional ultrasound image based on structural features of the abdominal aorta, so as to identify the abdominal aorta region in the three-dimensional ultrasound image.

[0179] In one implementation, the processor performs image recognition on the three-dimensional ultrasound image based on structural features of the abdominal aorta, specifically for:

[0180] A plurality of frames of two-dimensional abdominal aorta section images are selected from the three-dimensional ultrasound image, wherein the abdominal aorta section images include a transverse abdominal aorta section image and / or a longitudinal abdominal aorta section image; based on the structural characteristics of the abdominal aorta, an abdominal aorta region is identified in the abdominal aorta section image; and based on the abdominal aorta regions in the plurality of two-dimensional abdominal aorta section images, the abdominal aorta region of the three-dimensional ultrasound image is spliced.

[0181] In one implementation, the processor identifies the abdominal aorta region in the abdominal aorta section image based on the structural features of the abdominal aorta, specifically for:

[0182] A pre-trained neural network model is obtained, wherein the neural network model is trained by a deep learning algorithm on multiple frames of two-dimensional abdominal aorta section image samples with labeled information, and the labeled information is used to represent the abdominal aorta region in the abdominal aorta section image samples; the abdominal aorta section image is input into the neural network model to obtain a recognition result output by the neural network model based on the structural characteristics of the abdominal aorta, and the recognition result is used to represent the abdominal aorta region included in the abdominal aorta section image.

[0183] In one implementation, the annotation information is a regularly shaped frame enclosing the abdominal aorta region, and the recognition result includes a regularly shaped frame enclosing the abdominal aorta region; or, the annotation information is a boundary line marking the abdominal aorta region, and the recognition result includes a boundary line marking the abdominal aorta region.

[0184] In one implementation, the processor identifies the abdominal aorta region in the abdominal aorta section image based on the structural features of the abdominal aorta, specifically for:

[0185] select an image region of interest from the abdominal aorta cross-section image based on the structural feature of the abdominal aorta, and extract image features of the image region of interest; obtain an abdominal aorta two-dimensional image from a pre-constructed image library, and obtain image features of an abdominal aorta region pre-labeled in the abdominal aorta two-dimensional image; and match the extracted image features with the image features of the abdominal aorta region pre-labeled to determine whether the image region of interest contains the abdominal aorta region.

[0186] In an implementation manner, the processor selects the image region of interest from the abdominal aorta cross-section image, specifically for:

[0187] detects the image region of interest from the abdominal aorta cross-section image by using a target detection algorithm, and labels the image region of interest by using a regular shape box; or, segments the image region of interest from the abdominal aorta cross-section image by using an image segmentation algorithm, and labels a contour of the image region of interest.

[0188] In an implementation manner, the processor performs image recognition on the three-dimensional ultrasound image based on the structural feature of the abdominal aorta, specifically for:

[0189] obtains a pre-trained neural network model, the neural network model being trained by a deep learning algorithm on three-dimensional ultrasound volume data with an abdominal aorta region label; and inputs the three-dimensional ultrasound image into the neural network model to obtain a recognition result output by the neural network based on learned features of the abdominal aorta, the recognition result being used to represent a spatial position of the abdominal aorta region in the three-dimensional ultrasound image.

[0190] In an implementation manner, the processor determines the related information of the abdominal aorta based on the target cross-section image, specifically for:

[0191] if the target cross-section image does not have the recognition result of the abdominal aorta region, in response to an operation of a user labeling an abdominal aorta boundary on the target cross-section image, determines an image region surrounded by the abdominal aorta boundary as the abdominal aorta region; and based on the abdominal aorta region of the target cross-section image, calculates the related information of the abdominal aorta.

[0192] In an implementation manner, the processor determines the related information of the abdominal aorta based on the target cross-section image, specifically for:

[0193] if the target cross-section image does not have the recognition result of the abdominal aorta region, identifies the abdominal aorta region in the target cross-section image based on the structural feature of the abdominal aorta; and based on the abdominal aorta region of the target cross-section image, calculates the related information of the abdominal aorta.

[0194] In an implementation, the information about the abdominal aorta includes at least one of: a long axis direction inner diameter of the abdominal aorta, a short axis direction inner diameter of the abdominal aorta, and information about an abdominal aortic aneurysm; the information about the abdominal aortic aneurysm includes information about a dissection aneurysm and / or information about a non-dissection aneurysm.

[0195] In an implementation, the processor calculates the long axis direction inner diameter of the abdominal aorta and / or the short axis direction inner diameter of the abdominal aorta based on the target cross-sectional image, in particular for:

[0196] If the abdominal aorta cross-sectional image has a boundary line marking an abdominal aorta region, the short axis direction inner diameter of the abdominal aorta is calculated based on the boundary line; and / or, if the abdominal aorta longitudinal cross-sectional image has a boundary line marking an abdominal aorta region, a center line of the abdominal aorta region is determined, and the long axis direction inner diameter of the abdominal aorta is calculated based on the boundary line and the center line.

[0197] In an implementation, the processor calculates the long axis direction inner diameter of the abdominal aorta and / or the short axis direction inner diameter of the abdominal aorta based on the target cross-sectional image, in particular for:

[0198] If the abdominal aorta cross-sectional image has a regular shape frame surrounding an abdominal aorta region, a length or a width of the regular shape frame is determined as the short axis direction inner diameter of the abdominal aorta; and / or, if the abdominal aorta longitudinal cross-sectional image has a regular shape frame surrounding an abdominal aorta region, a center line of the abdominal aorta region is determined, and the long axis direction inner diameter of the abdominal aorta is calculated based on the regular shape frame and the center line.

[0199] In an implementation, the processor is further configured to generate a short axis inner diameter variation curve according to a plurality of the abdominal aorta cross-sectional images and the short axis direction inner diameters corresponding to the abdominal aorta cross-sectional images; wherein an abscissa of the short axis inner diameter variation curve is a position of an abdominal aorta cross-sectional image on a center line of the abdominal aorta, and an ordinate is a short axis direction inner diameter corresponding to the abdominal aorta cross-sectional image; and / or, generate a long axis inner diameter variation curve according to a plurality of the abdominal aorta longitudinal cross-sectional images and the long axis direction inner diameters corresponding to the abdominal aorta longitudinal cross-sectional images; wherein an abscissa of the long axis inner diameter variation curve is a position of an abdominal aorta longitudinal cross-sectional image on the center line of the abdominal aorta, and an ordinate is a long axis direction inner diameter corresponding to the abdominal aorta longitudinal cross-sectional image.

[0200] In an implementation, the processor is further configured to mark a maximum value of the short axis direction inner diameter on the short axis inner diameter variation curve; and / or, mark a maximum value of the long axis direction inner diameter on the long axis inner diameter variation curve.

[0201] In an implementation manner, the display is further configured to display an abdominal aorta transverse section image corresponding to the maximum inner diameter in the short axis direction; and / or display an abdominal aorta longitudinal section image corresponding to the maximum inner diameter in the long axis direction.

[0202] In an implementation manner, the processor is configured to determine information related to a dissection aneurysm based on the target section image, and specifically configured to:

[0203] input the target section image into a pre-trained neural network model to obtain information related to the dissection aneurysm output by the neural network model; the neural network model is trained by a deep learning algorithm based on abdominal aorta section images having annotation information related to the dissection aneurysm.

[0204] In an implementation manner, the annotation information is whether the abdominal aorta section image contains the dissection aneurysm, and the information related to the dissection aneurysm output by the neural network model is whether the target section image contains the dissection aneurysm.

[0205] In an implementation manner, the annotation information is a region position of the dissection aneurysm in the abdominal aorta section image, and the information related to the dissection aneurysm output by the neural network model is a region position of the dissection aneurysm in the target section image.

[0206] In an implementation manner, the processor is configured to determine information related to a non-dissection aneurysm based on the target section image, and specifically configured to:

[0207] prompt whether the target section image contains the non-dissection aneurysm according to the inner diameter of the abdominal aorta in the long axis direction and / or the inner diameter of the abdominal aorta in the short axis direction.

[0208] In an implementation manner, the processor is configured to prompt whether the target section image contains the non-dissection aneurysm according to the inner diameter of the abdominal aorta in the long axis direction and / or the inner diameter of the abdominal aorta in the short axis direction, and specifically configured to:

[0209] if the inner diameter of the abdominal aorta in the long axis direction exceeds a preset long axis inner diameter threshold, prompt that the target section image contains the non-dissection aneurysm;

[0210] if the inner diameter of the abdominal aorta in the short axis direction exceeds a preset short axis inner diameter threshold, prompt that the target section image contains the non-dissection aneurysm;

[0211] if an average value of the inner diameter of the abdominal aorta in the long axis direction of multiple frames exceeds a preset long axis inner diameter average threshold, prompt that the target section image contains the non-dissection aneurysm;

[0212] If the average of the short axis direction inner diameters of the abdominal aorta in multiple frames exceeds a preset short axis direction inner diameter average threshold value, it is indicated that the target cross-section image contains a non-dissected aneurysm;

[0213] If the ratio of the difference between the maximum and minimum of the long axis direction inner diameter of the abdominal aorta to the minimum exceeds a preset long axis direction inner diameter ratio threshold value, it is indicated that the target cross-section image contains a non-dissected aneurysm;

[0214] If the ratio of the difference between the maximum and minimum of the short axis direction inner diameter of the abdominal aorta to the minimum exceeds a preset short axis direction inner diameter ratio threshold value, it is indicated that the target cross-section image contains a non-dissected aneurysm.

[0215] In an implementation manner, the processor is further configured to, after the step of identifying the abdominal aorta region in the abdominal aorta cross-section image based on the structural features of the abdominal aorta, input the target cross-section image into a pre-trained neural network model to obtain the information related to the dissected aneurysm output by the neural network model; the neural network model is obtained by training the abdominal aorta cross-section image by a deep learning algorithm, and the abdominal aorta cross-section image has annotation information related to the dissected aneurysm.

[0216] See Figure 11 The embodiment of the present application further provides a specific structure of an ultrasonic detection device, which comprises a probe 1101, a spatial positioning device 1102, a transmitting circuit 1103, a transmitting / receiving selection switch 1104, a receiving circuit 1105, a beam synthesis circuit 1106, a processor 1107, a display 1108 and a memory 1109.

[0217] The transmitting circuit 1103 can excite the probe 1101 to transmit ultrasonic waves to a target region, such as an abdominal aorta region; the receiving circuit 1105 can receive ultrasonic echoes returned from the target region through the probe 1101, so as to obtain ultrasonic echo signals / data; the ultrasonic echo signals / data are sent to the processor 1107 after being processed by the beam synthesis circuit 1106. The spatial positioning device 1102 can obtain the motion trajectory of the probe, so as to obtain the spatial position information of a two-dimensional ultrasonic image. The spatial position information is also sent to the processor 1107.

[0218] The processor 1107 processes the ultrasonic echo signals / data to obtain a two-dimensional ultrasonic image of the target region, and obtains a plurality of two-dimensional ultrasonic data of the abdominal aorta and spatial position information corresponding to the two-dimensional ultrasonic data, the spatial position information being used to represent the position information of the abdominal aorta scanned by the two-dimensional ultrasonic data in the abdominal space, and reconstructs a three-dimensional ultrasonic image of the abdominal aorta based on the plurality of two-dimensional ultrasonic data and the spatial position information corresponding to the two-dimensional ultrasonic data. In addition, the processor 1107 can also perform other steps related to the processor in each of the above method embodiments, which will not be described herein.

[0219] The three-dimensional ultrasonic image obtained by the processor 1107 can be stored in the memory 1109, and the three-dimensional ultrasonic image can be displayed on the display 1108.

[0220] In an embodiment, the display 1108 of the ultrasonic detection device can be a touch display screen, a liquid crystal display screen, etc., or can be a liquid crystal display, a television, etc. independent display device, or can be a display screen on a mobile phone, a tablet computer, etc. electronic device, etc.

[0221] In actual applications, the processor 1107 can be at least one of an Application Specific Integrated Circuit (ASIC), a Digital Signal Processor (DSP), a Digital Signal Processing Device (DSPD), a Programmable Logic Device (PLD), a Field Programmable Gate Array (FPGA), a Central Processing Unit (CPU), a controller, a microcontroller, and a microprocessor, so that the processor 1107 can execute the corresponding steps of the ultrasonic imaging method in each embodiment of the present application.

[0222] The memory 1109 can be a volatile memory (e.g., a Random Access Memory (RAM)) or a non-volatile memory (e.g., a Read Only Memory (ROM), a flash memory, a Hard Disk Drive (HDD), or a Solid-State Drive (SSD)), or a combination of the above, and provide the processor with instructions and data.

[0223] Various exemplary embodiments are described herein. However, those skilled in the art will recognize that changes and modifications can be made thereto without departing from the scope of the present disclosure. For example, the various steps and components involved in the operation steps can be implemented differently depending on the particular application or considerations associated with the operation of the system (e.g., one or more steps can be deleted, modified, or combined with other steps).

[0224] The terms "first", "second", and the like, herein do not denote any ordinality or "importance", but are used to distinguish one element from another. Furthermore, the terms "comprises", "comprising", "includes", "including" and the like, are inclusive only - they do not exclude the presence of other elements or steps. Furthermore, the terms "coupled" and "coupling" mean to be directly or indirectly connected, and are not necessarily limited to a physical or mechanical connection.

[0225] Additionally, as will be appreciated by those skilled in the art, the principles described herein can be reflected in a computer program product having a computer-readable medium having computer-readable program code embodied therein. Any tangible, non-transitory computer-readable storage media can be utilized, including magnetic storage devices (hard disks, floppy disks, etc.), optical storage devices (CD-ROMs, DVDs, Blu Rays, etc.), flash memory, and / or the like. These computer program instructions can be loaded onto a general purpose computer, a special purpose computer, or other programmable data processing apparatus to produce a machine, such that the instructions that execute on the computer or other programmable data processing apparatus create means for implementing the functions specified in the flowchart block or blocks. These computer program instructions can also be stored in a computer-readable memory that can direct a computer or other programmable data processing apparatus to function in a particular manner, such that the instructions stored in the computer-readable memory produce an article of manufacture including an implementation to

[0226] The foregoing detailed description has been presented for purposes of clarity and description. However, various modifications and changes can be made to the embodiments described without departing from the scope of the disclosure. Accordingly, the disclosure is intended to be illustrative, but not limiting, of the scope, and all modifications are intended to be included. Also, the advantages, other advantages, and solutions to problems have been described above with regard to various embodiments. However, the benefits, advantages, solutions to problems and any element(s) that can cause any of such should not be construed as critical, required or essential. The terms "comprises", "comprising", or any other variation thereof, used in this document, as well as any other similar terms, are intended to cover a non-exclusive inclusion, such that a process, method, article or apparatus that comprises a list of elements does not include only those elements but can include other elements not expressly listed or inherent to such process, method, article or apparatus. Also, as used in this document, the terms "coupled" and / or "coupling", together with any variations thereof, means physical connection, electrical connection, magnetic connection, optical connection, communicative connection, functional connection and / or any other connection.

[0227] The above embodiments only express several implementation manners, which are described in a more specific and detailed manner, but cannot be understood as the limitation of the patent scope of the present application. It should be noted that, for those skilled in the art, several modifications and improvements can be made without departing from the concept of the present application, which all belong to the protection scope of the present application. Therefore, the protection scope of the present application patent should be subject to the appended claims.

Claims

1. A method of imaging the abdominal aorta, characterized in that, The method comprises: controlling an ultrasonic detection probe to emit ultrasonic waves to an abdominal space of a target object, and controlling the ultrasonic detection probe to receive ultrasonic echo signals returned from the abdominal space of the target object; performing three-dimensional image reconstruction based on the ultrasonic echo signals to obtain a three-dimensional ultrasonic image of the abdominal aorta; selecting a target cross-sectional image from the three-dimensional ultrasonic image, the target cross-sectional image comprising an abdominal aorta transverse cross-sectional image and / or an abdominal aorta longitudinal cross-sectional image; calculating relevant information of the abdominal aorta based on the target cross-sectional image; The method further comprises: generating a short-axis inner diameter variation curve according to a plurality of the abdominal aorta transverse cross-sectional images and corresponding short-axis direction inner diameters of the abdominal aorta transverse cross-sectional images; wherein the horizontal coordinate of the short-axis inner diameter variation curve is the position of the abdominal aorta transverse cross-sectional image on the center line of the abdominal aorta, and the vertical coordinate is the corresponding short-axis direction inner diameter of the abdominal aorta transverse cross-sectional image; and / or generating a long-axis inner diameter variation curve according to a plurality of the abdominal aorta longitudinal cross-sectional images and corresponding long-axis direction inner diameters of the abdominal aorta longitudinal cross-sectional images; wherein the horizontal coordinate of the long-axis inner diameter variation curve is the position of the abdominal aorta longitudinal cross-sectional image on the center line of the abdominal aorta, and the vertical coordinate is the corresponding long-axis direction inner diameter of the abdominal aorta longitudinal cross-sectional image.

2. The abdominal aorta imaging method of claim 1, wherein the relevant information of the abdominal aorta comprises at least one of the following: a long-axis direction inner diameter of the abdominal aorta, a short-axis direction inner diameter of the abdominal aorta, and relevant information of an abdominal aortic aneurysm; the relevant information of the abdominal aortic aneurysm comprises relevant information of a dissection aneurysm and / or relevant information of a non-dissection aneurysm. The step of calculating the long-axis direction inner diameter of the abdominal aorta and / or the short-axis direction inner diameter of the abdominal aorta based on the target cross-sectional image comprises:

3. The method of imaging the abdominal aorta of claim 2, wherein, if the abdominal aorta transverse cross-sectional image has a boundary line marking the abdominal aorta region, calculating the short-axis direction inner diameter of the abdominal aorta based on the boundary line; and / or if the abdominal aorta longitudinal cross-sectional image has a boundary line marking the abdominal aorta region, determining a center line of the abdominal aorta region, and calculating the long-axis direction inner diameter of the abdominal aorta based on the boundary line and the center line. The step of calculating the long-axis direction inner diameter of the abdominal aorta and / or the short-axis direction inner diameter of the abdominal aorta based on the target cross-sectional image comprises:

4. The method of imaging the abdominal aorta of claim 2, wherein, if the abdominal aorta transverse cross-sectional image has a regular shape frame surrounding the abdominal aorta region, determining the length or width of the regular shape frame as the short-axis direction inner diameter of the abdominal aorta; and / or if the abdominal aorta longitudinal cross-sectional image has a regular shape frame surrounding the abdominal aorta region, determining a center line of the abdominal aorta region, and calculating the long-axis direction inner diameter of the abdominal aorta based on the regular shape frame and the center line. The method further comprises:

5. The method of imaging the abdominal aorta of claim 1, wherein, marking the maximum short-axis direction inner diameter on the short-axis inner diameter variation curve; and / or marking the maximum long-axis direction inner diameter on the long-axis inner diameter variation curve. The method further comprises:

6. The method of imaging the abdominal aorta of claim 1, wherein, displaying the abdominal aorta transverse cross-sectional image corresponding to the maximum short-axis direction inner diameter; and / or displaying the abdominal aorta longitudinal cross-sectional image corresponding to the maximum long-axis direction inner diameter. The method comprises: ​ 7. A method of imaging the abdominal aorta, characterized in that, ​ controlling an ultrasound detection probe to emit ultrasound waves to an abdominal space of a target object, and controlling the ultrasound detection probe to receive ultrasound echo signals returned from the abdominal space of the target object; performing three-dimensional image reconstruction based on the ultrasound echo signals to obtain a three-dimensional ultrasound image of the abdominal aorta; displaying the three-dimensional ultrasound image of the abdominal aorta; further comprising: selecting a target cross-sectional image from the three-dimensional ultrasound image, the target cross-sectional image comprising an abdominal aorta transverse cross-sectional image and / or an abdominal aorta longitudinal cross-sectional image; generating a short-axis inner diameter variation curve according to a plurality of the abdominal aorta transverse cross-sectional images and corresponding short-axis direction inner diameters of the abdominal aorta transverse cross-sectional images; wherein the horizontal coordinate of the short-axis inner diameter variation curve is the position of the abdominal aorta transverse cross-sectional image on the center line of the abdominal aorta, and the vertical coordinate is the corresponding short-axis direction inner diameter of the abdominal aorta transverse cross-sectional image; and / or generating a long-axis inner diameter variation curve according to a plurality of the abdominal aorta longitudinal cross-sectional images and corresponding long-axis direction inner diameters of the abdominal aorta longitudinal cross-sectional images; wherein the horizontal coordinate of the long-axis inner diameter variation curve is the position of the abdominal aorta longitudinal cross-sectional image on the center line of the abdominal aorta, and the vertical coordinate is the corresponding long-axis direction inner diameter of the abdominal aorta longitudinal cross-sectional image.

8. The method of imaging the abdominal aorta of claim 7, wherein, The step of performing three-dimensional image reconstruction based on the ultrasound echo signals to obtain a three-dimensional ultrasound image of the abdominal aorta comprises: processing the ultrasound echo signals to obtain a plurality of two-dimensional ultrasound data of the abdominal aorta, and obtaining spatial position information corresponding to the two-dimensional ultrasound data, the spatial position information being used to represent the position information of the abdominal aorta scanned by the two-dimensional ultrasound data in the abdominal space; reconstructing a three-dimensional ultrasound image of the abdominal aorta based on a plurality of the two-dimensional ultrasound data and the spatial position information corresponding to the two-dimensional ultrasound data.

9. The method of imaging the abdominal aorta of claim 7, wherein, Further comprising: calculating relevant information of the abdominal aorta based on the target cross-sectional image.

10. The method of imaging the abdominal aorta of claim 9, wherein, The step of selecting a target cross-sectional image from the three-dimensional ultrasound image comprises: rotating and / or moving the three-dimensional ultrasound image to display the three-dimensional ultrasound image at different spatial perspectives; in response to a user's operation of selecting an abdominal aorta cross-sectional image of interest in the three-dimensional ultrasound image at a target spatial perspective, determining the abdominal aorta cross-sectional image of interest selected by the user as the target cross-sectional image.

11. The method of imaging the abdominal aorta of claim 9, wherein, The step of selecting a target cross-sectional image from the three-dimensional ultrasound image comprises: performing image recognition on the three-dimensional ultrasound image based on the structural characteristics of the abdominal aorta to identify the abdominal aorta region in the three-dimensional ultrasound image; determining the center line of the abdominal aorta region; selecting the target cross-sectional image based on the center line of the abdominal aorta region.

12. The method of imaging the abdominal aorta of claim 11, wherein, The step of determining the center line of the abdominal aorta region comprises: selecting at least one abdominal aorta transverse cross-sectional image from the three-dimensional ultrasound image; determining the center position from the abdominal aorta region identified in the at least one abdominal aorta transverse cross-sectional image; fitting the center line of the abdominal aorta region based on the center position of the abdominal aorta region in the at least one abdominal aorta transverse cross-sectional image.

13. The method of imaging the abdominal aorta of claim 11, wherein, The step of selecting the target cross-sectional image based on the center line of the abdominal aorta region comprises: From the three-dimensional ultrasound image, an abdominal aorta longitudinal section image passing through the center line is selected as a target section image; or From the three-dimensional ultrasound image, an abdominal aorta transverse section image orthogonal to the abdominal aorta longitudinal section image is selected as a target section image.

14. The method of imaging the abdominal aorta of claim 7 wherein, Further comprising: Based on the structural characteristics of the abdominal aorta, image recognition is performed on the three-dimensional ultrasound image to identify the abdominal aorta region in the three-dimensional ultrasound image.

15. The method of imaging the abdominal aorta according to claim 11 or 14, wherein, The step of performing image recognition on the three-dimensional ultrasound image based on the structural characteristics of the abdominal aorta comprises: Selecting multiple two-dimensional abdominal aorta section images from the three-dimensional ultrasound image, the abdominal aorta section images including abdominal aorta transverse section images and / or abdominal aorta longitudinal section images; Based on the structural characteristics of the abdominal aorta, identifying abdominal aorta regions in the abdominal aorta section images; Based on the abdominal aorta regions in the multiple two-dimensional abdominal aorta section images, splicing to obtain an abdominal aorta region of the three-dimensional ultrasound image.

16. The method of imaging the abdominal aorta of claim 15, wherein, The step of identifying abdominal aorta regions in the abdominal aorta section images based on the structural characteristics of the abdominal aorta comprises: Obtaining a pre-trained neural network model, wherein the neural network model is trained by a deep learning algorithm on multiple two-dimensional abdominal aorta section image samples with annotation information, and the annotation information is used to represent the abdominal aorta region in the abdominal aorta section image sample; Inputting the abdominal aorta section image into the neural network model to obtain an identification result output by the neural network model based on the structural characteristics of the abdominal aorta, and the identification result is used to represent the abdominal aorta region included in the abdominal aorta section image.

17. The abdominal aorta imaging method of claim 16, wherein When the annotation information is a regular shape box surrounding the abdominal aorta region, the identification result includes a regular shape box surrounding the abdominal aorta region; or When the annotation information is a boundary line marking the abdominal aorta region, the identification result includes a boundary line marking the abdominal aorta region. The step of identifying abdominal aorta regions in the abdominal aorta section images based on the structural characteristics of the abdominal aorta comprises:

18. The method of imaging the abdominal aorta of claim 15, wherein, Based on the structural characteristics of the abdominal aorta, selecting an image region of interest from the abdominal aorta section image and extracting image features of the image region of interest; Obtaining an abdominal aorta two-dimensional image from a pre-constructed image library and obtaining image features of an abdominal aorta region pre-marked in the abdominal aorta two-dimensional image; Matching the extracted image features with the image features of the pre-marked abdominal aorta region to determine whether the image region of interest contains an abdominal aorta region. The step of selecting an image region of interest from the abdominal aorta section image comprises:

19. The method of imaging the abdominal aorta of claim 18, wherein, Using a target detection algorithm to detect an image region of interest from the abdominal aorta section image and using a regular shape box to mark the image region of interest; or Using an image segmentation algorithm to segment an image region of interest from the abdominal aorta section image and marking the contour of the image region of interest. ​ 20. The method of imaging the abdominal aorta of claim 11 or 14, wherein, The step of performing image recognition on the three-dimensional ultrasound image based on the structural features of the abdominal aorta comprises: obtaining a pre-trained neural network model, wherein the neural network model is trained by a deep learning algorithm on three-dimensional ultrasound volume data with abdominal aorta region labels; inputting the three-dimensional ultrasound image into the neural network model to obtain a recognition result output by the neural network based on learned abdominal aorta features, wherein the recognition result is used to represent the spatial position of the abdominal aorta region in the three-dimensional ultrasound image.

21. The method of imaging the abdominal aorta of claim 9, wherein, The step of determining the relevant information of the abdominal aorta based on the target cross-sectional image comprises: if the target cross-sectional image does not have a recognition result for the abdominal aorta region, determining an image region surrounded by the abdominal aorta boundary as the abdominal aorta region in response to a user marking the abdominal aorta boundary on the target cross-sectional image; calculating the relevant information of the abdominal aorta based on the abdominal aorta region of the target cross-sectional image.

22. The method of imaging the abdominal aorta of claim 9, wherein, The step of determining the relevant information of the abdominal aorta based on the target cross-sectional image comprises: if the target cross-sectional image does not have a recognition result for the abdominal aorta region, identifying the abdominal aorta region in the target cross-sectional image based on the structural features of the abdominal aorta; calculating the relevant information of the abdominal aorta based on the abdominal aorta region of the target cross-sectional image.

23. The abdominal aorta imaging method of claim 9, wherein: the relevant information of the abdominal aorta comprises at least one of the following: a long axis direction inner diameter of the abdominal aorta, a short axis direction inner diameter of the abdominal aorta, and relevant information of an abdominal aortic aneurysm; the relevant information of the abdominal aortic aneurysm comprises relevant information of a dissection aneurysm and / or relevant information of a non-dissection aneurysm.

24. The method of imaging the abdominal aorta of claim 23, wherein, The step of calculating the long axis direction inner diameter and / or the short axis direction inner diameter of the abdominal aorta based on the target cross-sectional image comprises: if the abdominal aorta transverse cross-sectional image has a boundary line marking the abdominal aorta region, calculating the short axis direction inner diameter of the abdominal aorta based on the boundary line; and / or, if the abdominal aorta longitudinal cross-sectional image has a boundary line marking the abdominal aorta region, determining a center line of the abdominal aorta region, and calculating the long axis direction inner diameter of the abdominal aorta based on the boundary line and the center line.

25. The method of imaging the abdominal aorta of claim 23, wherein, The step of calculating the long axis direction inner diameter and / or the short axis direction inner diameter of the abdominal aorta based on the target cross-sectional image comprises: if the abdominal aorta transverse cross-sectional image has a regular shape frame surrounding the abdominal aorta region, determining the length or width of the regular shape frame as the short axis direction inner diameter of the abdominal aorta; and / or, if the abdominal aorta longitudinal cross-sectional image has a regular shape frame surrounding the abdominal aorta region, determining a center line of the abdominal aorta region, and calculating the long axis direction inner diameter of the abdominal aorta based on the regular shape frame and the center line.

26. The method of imaging the abdominal aorta of claim 7 wherein, Further comprising: marking the maximum short axis direction inner diameter on the short axis inner diameter change curve; and / or, marking the maximum long axis direction inner diameter on the long axis inner diameter change curve.

27. The method of imaging the abdominal aorta of claim 26 wherein, Further comprising: displaying the abdominal aorta transverse cross-sectional image corresponding to the maximum short axis direction inner diameter; And / or, display the abdominal aorta longitudinal section image corresponding to the maximum value of the long axis direction inner diameter.

28. The method of imaging the abdominal aorta of claim 23, wherein, The step of determining the information related to the dissection aneurysm based on the target section image comprises: inputting the target section image into a pre-trained neural network model to obtain the information related to the dissection aneurysm output by the neural network model; The neural network model is trained by a deep learning algorithm on abdominal aorta section images with annotation information related to dissection aneurysm.

29. The abdominal aorta imaging method of claim 28, wherein The annotation information is whether the abdominal aorta section image contains a dissection aneurysm, and the information related to the dissection aneurysm output by the neural network model is whether the target section image contains a dissection aneurysm.

30. The abdominal aorta imaging method of claim 28, wherein The annotation information is the region position of the dissection aneurysm in the abdominal aorta section image, and the information related to the dissection aneurysm output by the neural network model is the region position of the dissection aneurysm in the target section image.

31. The method of imaging the abdominal aorta of claim 23 wherein, The step of determining the information related to the non-dissection aneurysm based on the target section image comprises: According to the long axis direction inner diameter of the abdominal aorta and / or the short axis direction inner diameter of the abdominal aorta, it is prompted whether the target section image contains a non-dissection aneurysm.

32. The method of imaging the abdominal aorta of claim 31, wherein, The step of prompting whether the target section image contains a non-dissection aneurysm according to the long axis direction inner diameter of the abdominal aorta and / or the short axis direction inner diameter of the abdominal aorta comprises: If the long axis direction inner diameter of the abdominal aorta exceeds a preset long axis inner diameter threshold, it is prompted that the target section image contains a non-dissection aneurysm; If the short axis direction inner diameter of the abdominal aorta exceeds a preset short axis inner diameter threshold, it is prompted that the target section image contains a non-dissection aneurysm; If the average value of the long axis direction inner diameter of the abdominal aorta exceeds a preset long axis inner diameter average threshold, it is prompted that the target section image contains a non-dissection aneurysm; If the average value of the short axis direction inner diameter of the abdominal aorta exceeds a preset short axis inner diameter average threshold, it is prompted that the target section image contains a non-dissection aneurysm; If the ratio of the difference between the maximum value and the minimum value of the long axis direction inner diameter of the abdominal aorta to the minimum value exceeds a preset long axis inner diameter ratio threshold, it is prompted that the target section image contains a non-dissection aneurysm; If the ratio of the difference between the maximum value and the minimum value of the short axis direction inner diameter of the abdominal aorta to the minimum value exceeds a preset short axis inner diameter ratio threshold, it is prompted that the target section image contains a non-dissection aneurysm.

33. The method of imaging the abdominal aorta of claim 15, wherein, After the step of identifying the abdominal aorta region in the abdominal aorta section image based on the structural characteristics of the abdominal aorta, the method further comprises: inputting the target section image into a pre-trained neural network model to obtain the information related to the dissection aneurysm output by the neural network model; The neural network model is trained by a deep learning algorithm on abdominal aorta section images with annotation information related to dissection aneurysm.

34. An ultrasonic testing apparatus characterized by, Comprise: An ultrasonic detection probe configured to emit ultrasonic waves to an abdominal space of a target object; a processor configured to control the ultrasonic detection probe to emit ultrasonic waves to an abdominal space of a target object, and control the ultrasonic detection probe to receive ultrasonic echo signals returned from the abdominal space of the target object; perform three-dimensional image reconstruction based on the ultrasonic echo signals to obtain a three-dimensional ultrasonic image of the abdominal aorta; select a target cross-sectional image from the three-dimensional ultrasonic image, the target cross-sectional image including an abdominal aorta transverse cross-sectional image and / or an abdominal aorta longitudinal cross-sectional image; and calculate relevant information of the abdominal aorta based on the target cross-sectional image; the processor is further configured to: generate a short-axis inner diameter variation curve according to a plurality of the abdominal aorta transverse cross-sectional images and corresponding short-axis direction inner diameters of the abdominal aorta transverse cross-sectional images; wherein the horizontal coordinate of the short-axis inner diameter variation curve is the position of the abdominal aorta transverse cross-sectional image on the center line of the abdominal aorta, and the vertical coordinate is the corresponding short-axis direction inner diameter of the abdominal aorta transverse cross-sectional image; and / or generate a long-axis inner diameter variation curve according to a plurality of the abdominal aorta longitudinal cross-sectional images and corresponding long-axis direction inner diameters of the abdominal aorta longitudinal cross-sectional images; wherein the horizontal coordinate of the long-axis inner diameter variation curve is the position of the abdominal aorta longitudinal cross-sectional image on the center line of the abdominal aorta, and the vertical coordinate is the corresponding long-axis direction inner diameter of the abdominal aorta longitudinal cross-sectional image.

35. The ultrasonic testing apparatus of claim 34, wherein, the relevant information of the abdominal aorta calculated by the processor includes at least one of the following: a long-axis direction inner diameter of the abdominal aorta, a short-axis direction inner diameter of the abdominal aorta, and relevant information of an abdominal aortic aneurysm; the relevant information of the abdominal aortic aneurysm includes relevant information of a dissection aneurysm and / or relevant information of a non-dissection aneurysm.

36. The ultrasonic testing apparatus of claim 35, wherein, the processor determines the relevant information of the dissection aneurysm based on the target cross-sectional image, and specifically for: inputting the target cross-sectional image into a pre-trained neural network model to obtain dissection aneurysm related information output by the neural network model; the neural network model is trained by a deep learning algorithm on abdominal aorta cross-sectional images having annotation information related to dissection aneurysms.

37. The ultrasonic detection device of claim 36, wherein: the annotation information is whether the abdominal aorta cross-sectional image contains a dissection aneurysm, and the dissection aneurysm related information output by the neural network model is whether the target cross-sectional image contains a dissection aneurysm.

38. The ultrasonic detection device of claim 36, wherein: the annotation information is the region position of the dissection aneurysm in the abdominal aorta cross-sectional image, and the dissection aneurysm related information output by the neural network model is the region position of the dissection aneurysm in the target cross-sectional image.

39. The ultrasonic testing apparatus of claim 35, wherein, further comprising a display; the processor determines the relevant information of the non-dissection aneurysm based on the target cross-sectional image, and specifically for: generating prompt information according to the long-axis direction inner diameter of the abdominal aorta and / or the short-axis direction inner diameter of the abdominal aorta, the prompt information being used to prompt whether the target cross-sectional image contains a non-dissection aneurysm; the display is configured to display the prompt information.

40. An ultrasonic testing apparatus, characterized by, ​ An ultrasonic detection probe configured to emit ultrasonic waves to an abdominal space of a target object; a processor configured to control the ultrasonic detection probe to emit ultrasonic waves to an abdominal space of a target object, and control the ultrasonic detection probe to receive ultrasonic echo signals returned from the abdominal space of the target object; perform three-dimensional image reconstruction based on the ultrasonic echo signals to obtain a three-dimensional ultrasonic image of the abdominal aorta; a display configured to display the three-dimensional ultrasonic image of the abdominal aorta; the processor is further configured to select a target cross-sectional image from the three-dimensional ultrasonic image, the target cross-sectional image comprising an abdominal aorta transverse cross-sectional image and / or an abdominal aorta longitudinal cross-sectional image; generate a short-axis inner diameter variation curve according to a plurality of the abdominal aorta transverse cross-sectional images and corresponding short-axis direction inner diameters of the abdominal aorta transverse cross-sectional images; wherein the horizontal coordinate of the short-axis inner diameter variation curve is the position of the abdominal aorta transverse cross-sectional image on the centerline of the abdominal aorta, and the vertical coordinate is the corresponding short-axis direction inner diameter of the abdominal aorta transverse cross-sectional image; and / or generate a long-axis inner diameter variation curve according to a plurality of the abdominal aorta longitudinal cross-sectional images and corresponding long-axis direction inner diameters of the abdominal aorta longitudinal cross-sectional images; wherein the horizontal coordinate of the long-axis inner diameter variation curve is the position of the abdominal aorta longitudinal cross-sectional image on the centerline of the abdominal aorta, and the vertical coordinate is the corresponding long-axis direction inner diameter of the abdominal aorta longitudinal cross-sectional image.

41. The ultrasonic testing apparatus of claim 40, wherein, the processor performs three-dimensional image reconstruction based on the ultrasonic echo signals to obtain a three-dimensional ultrasonic image of the abdominal aorta, specifically for: processing the ultrasonic echo signals to obtain a plurality of two-dimensional ultrasonic data of the abdominal aorta, and obtaining spatial position information corresponding to the two-dimensional ultrasonic data, the spatial position information being used to represent the position information of the abdominal aorta scanned by the two-dimensional ultrasonic data in the abdominal space; reconstructing a three-dimensional ultrasonic image of the abdominal aorta based on a plurality of the two-dimensional ultrasonic data and the spatial position information corresponding to the two-dimensional ultrasonic data.

42. The ultrasonic detection device of claim 40, wherein the processor is further configured to calculate relevant information of the abdominal aorta based on the target cross-sectional image.

43. The ultrasonic testing apparatus of claim 42, wherein, the processor selects a target cross-sectional image from the three-dimensional ultrasonic image, specifically for: rotating and / or moving the three-dimensional ultrasonic image to display the three-dimensional ultrasonic image under different spatial perspectives; in response to a user's operation of selecting an abdominal aorta cross-sectional image of interest in the three-dimensional ultrasonic image under a target spatial perspective, determining the abdominal aorta cross-sectional image of interest selected by the user as the target cross-sectional image.

44. The ultrasonic detection device of claim 40, wherein the processor is further configured to perform image recognition on the three-dimensional ultrasonic image based on the structural features of the abdominal aorta to identify the abdominal aorta region in the three-dimensional ultrasonic image.

45. The ultrasonic testing apparatus of claim 40, wherein, the processor identifies the abdominal aorta region in the abdominal aorta cross-sectional image based on the structural features of the abdominal aorta, specifically for: obtaining a pre-trained neural network model, wherein the neural network model is trained by a deep learning algorithm on a plurality of two-dimensional abdominal aorta cross-section image samples with labeled information, and the labeled information is used to represent an abdominal aorta region in the abdominal aorta cross-section image sample; inputting the abdominal aorta cross-section image into the neural network model to obtain an identification result output by the neural network model based on the structural characteristics of the abdominal aorta, and the identification result is used to represent the abdominal aorta region included in the abdominal aorta cross-section image.

46. The ultrasonic testing apparatus of claim 44, wherein, The processor performs image recognition on the three-dimensional ultrasound image based on the structural characteristics of the abdominal aorta, and is specifically configured to: obtain a pre-trained neural network model, wherein the neural network model is trained by a deep learning algorithm on a three-dimensional ultrasound volume data with an abdominal aorta region label; input the three-dimensional ultrasound image into the neural network model to obtain an identification result output by the neural network based on learned abdominal aorta characteristics, and the identification result is used to represent the spatial position of the abdominal aorta region in the three-dimensional ultrasound image.

47. The ultrasound detection device of claim 42, wherein the related information of the abdominal aorta includes at least one of the following: a long axis direction inner diameter of the abdominal aorta, a short axis direction inner diameter of the abdominal aorta, and related information of an abdominal aortic aneurysm; the related information of the abdominal aortic aneurysm includes related information of a dissection aneurysm and / or related information of a non-dissection aneurysm.

48. The ultrasonic testing apparatus of claim 47, wherein, The processor calculates the long axis direction inner diameter of the abdominal aorta and / or the short axis direction inner diameter of the abdominal aorta based on the target cross-section image, and is specifically configured to: if the abdominal aorta transverse cross-section image has a boundary line marking the abdominal aorta region, calculate the short axis direction inner diameter of the abdominal aorta based on the boundary line; and / or, if the abdominal aorta longitudinal cross-section image has a boundary line marking the abdominal aorta region, determine a center line of the abdominal aorta region, and calculate the long axis direction inner diameter of the abdominal aorta based on the boundary line and the center line.

49. The ultrasonic testing apparatus of claim 47, wherein, The processor calculates the long axis direction inner diameter of the abdominal aorta and / or the short axis direction inner diameter of the abdominal aorta based on the target cross-section image, and is specifically configured to: if the abdominal aorta transverse cross-section image has a regular shape frame surrounding the abdominal aorta region, determine the long or wide of the regular shape frame as the short axis direction inner diameter of the abdominal aorta; and / or, if the abdominal aorta longitudinal cross-section image has a regular shape frame surrounding the abdominal aorta region, determine a center line of the abdominal aorta region, and calculate the long axis direction inner diameter of the abdominal aorta based on the regular shape frame and the center line.

50. The ultrasound detection device of claim 40, wherein the processor is further configured to mark a maximum value of the short axis direction inner diameter on the short axis inner diameter change curve; and / or, mark a maximum value of the long axis direction inner diameter on the long axis inner diameter change curve.

51. The ultrasonic testing apparatus of claim 47, wherein, The processor determines the related information of the dissection aneurysm based on the target cross-section image, and is specifically configured to: input the target cross-section image into a pre-trained neural network model to obtain dissection aneurysm related information output by the neural network model; The neural network model is trained by a deep learning algorithm on abdominal aorta cross-section images, and the abdominal aorta cross-section images have annotation information related to dissection aneurysm.

52. The ultrasonic testing apparatus of claim 47, wherein, The processor determines information related to non-dissection aneurysm based on the target cross-section image, specifically for: According to the long axis direction inner diameter of the abdominal aorta and / or the short axis direction inner diameter of the abdominal aorta, it is prompted whether the target cross-section image contains non-dissection aneurysm.

53. The ultrasonic testing apparatus of claim 52, wherein, The processor prompts whether the target cross-section image contains non-dissection aneurysm according to the long axis direction inner diameter of the abdominal aorta and / or the short axis direction inner diameter of the abdominal aorta, specifically for: If the long axis direction inner diameter of the abdominal aorta exceeds the preset long axis inner diameter threshold, it is prompted that the target cross-section image contains non-dissection aneurysm; If the short axis direction inner diameter of the abdominal aorta exceeds the preset short axis inner diameter threshold, it is prompted that the target cross-section image contains non-dissection aneurysm; If the average value of the long axis direction inner diameter of the abdominal aorta exceeds the preset long axis inner diameter average threshold, it is prompted that the target cross-section image contains non-dissection aneurysm; If the average value of the short axis direction inner diameter of the abdominal aorta exceeds the preset short axis inner diameter average threshold, it is prompted that the target cross-section image contains non-dissection aneurysm; If the difference between the maximum value and the minimum value of the long axis direction inner diameter of the abdominal aorta accounts for the minimum value exceeds the preset long axis inner diameter ratio threshold, it is prompted that the target cross-section image contains non-dissection aneurysm; If the difference between the maximum value and the minimum value of the short axis direction inner diameter of the abdominal aorta accounts for the minimum value exceeds the preset short axis inner diameter ratio threshold, it is prompted that the target cross-section image contains non-dissection aneurysm.

54. The ultrasound testing device of claim 44, wherein: The processor is further configured to input the target cross-section image into a pre-trained neural network model to obtain dissection aneurysm related information output by the neural network model after identifying the abdominal aorta region in the abdominal aorta cross-section image based on the structural features of the abdominal aorta; the neural network model is trained by a deep learning algorithm on abdominal aorta cross-section images, and the abdominal aorta cross-section images have annotation information related to dissection aneurysm.

Images