A method and system for assisted identification of aortic dissection

Abstract

The application provides an auxiliary identification method and system for aortic dissection, and the method comprises the following steps: acquiring chest cavity plain scan images of a user's aorta through a CT device, and identifying cross-sectional contour images of the aorta in each chest cavity plain scan image through an image edge recognition algorithm; identifying edge structure information of each cross-sectional contour image, and extracting image features of each cross-sectional contour image; identifying each abnormal image area and abnormal information thereof in each cross-sectional contour image, identifying the abnormal type and abnormal degree of each cross-sectional contour image based on the abnormal information of each abnormal image area, and obtaining auxiliary identification information of aortic dissection of the user through an aortic dissection analysis strategy, aortic dissection probability prediction and prediction reasons. The scheme can improve the auxiliary identification efficiency of aortic dissection in emergency.

Description

Technical Field

[0001] This invention relates to the field of image recognition and data analysis technology, and in particular to an auxiliary identification method and system for aortic dissection. Background Technology

[0002] Aortic dissection (AD) is a pathological change in which high-velocity, high-pressure blood flow enters the aortic media through a tear in the aortic intima, causing separation of the media and propagation along the long axis of the aorta, resulting in two distinct lumens: a true lumen and a false lumen. This disease has a rapid onset, rapid progression, and high mortality rate; therefore, the detection of aortic dissection is a current research focus.

[0003] Traditional methods for detecting aortic dissection involve obtaining a trunk scan of the user using contrast-enhanced CT imaging to help doctors identify whether the user has aortic dissection. However, due to the rapid onset and progression of aortic dissection, some patients present with atypical symptoms and are difficult to differentiate from other diseases, leading clinicians to fail to order contrast-enhanced CT examinations in the first instance. Some patients are not suitable for contrast-enhanced CT examinations due to various factors, and in the emergency room, they often only have plain CT images with much lower contrast than contrast-enhanced CT images, resulting in low efficiency in assisting the identification of aortic dissection in the emergency room. Summary of the Invention

[0004] The main objective of this invention is to provide an auxiliary identification method for aortic dissection, aiming to solve the problem that the existing technology, due to the shortage of equipment and personnel resources for CT enhanced imaging, cannot efficiently assist doctors in identifying the patient's aortic dissection condition in the emergency room, resulting in low auxiliary identification efficiency of aortic dissection in the emergency room.

[0005] To achieve the above objectives, the present invention provides an auxiliary identification method for aortic dissection, the method comprising:

[0006] Acquire a plain thoracic scan image of the user's aorta, and use an image edge recognition algorithm to identify the cross-sectional contour image of the aorta in each of the plain thoracic scan images;

[0007] Identify the edge structure information of each cross-sectional contour image and extract the image features of each cross-sectional contour image;

[0008] For each cross-sectional contour image, based on the image features in the cross-sectional contour image and the edge structure information of the cross-sectional contour image, the aortic anomaly recognition model is used to identify each abnormal image region in the cross-sectional contour image and the abnormal information of each abnormal image region. Based on the abnormal information of each abnormal image region, the abnormal type and the degree of abnormality of the cross-sectional contour image are identified.

[0009] Based on the anomaly type and degree of anomaly of each cross-sectional contour image, an aortic dissection analysis strategy is used to predict the user's aortic dissection probability and the reason for the prediction of the aortic dissection probability. The user's aortic dissection probability and the reason for the prediction of the aortic dissection probability are used as auxiliary identification information for the user's aortic dissection.

[0010] Optionally, the identification of edge structure information for each cross-sectional contour image includes:

[0011] Based on the aortic location corresponding to each cross-sectional contour image, the sample structure location range of each aortic structure contained in the cross-sectional contour image is identified, and for each cross-sectional contour image, based on each edge contour line in the cross-sectional contour image, the image content contained in each edge contour line is identified.

[0012] Based on the sample structure location range of each aortic structure and the content location range corresponding to the image content contained in each edge contour line, the aortic structure corresponding to each image content is identified;

[0013] Based on each image content, each adjacent edge contour line is identified, and for each pair of adjacent edge contour lines, the sub-edge spacing information between the adjacent edge contour lines is identified. The sub-edge spacing information between each pair of adjacent edge contour lines, as well as the aortic structure corresponding to the image content contained in each pair of adjacent edge contour lines, are used as the edge structure information of the cross-sectional contour image.

[0014] Optionally, the extraction of image features from each cross-sectional contour image includes:

[0015] For each cross-sectional contour image, an image feature extraction network is used to extract the initial image features of the cross-sectional contour image and identify the feature type of each initial image feature.

[0016] Add a type identifier corresponding to the feature type of each initial image feature to each initial image feature, and establish a correspondence between each initial image feature and the image content represented by each initial image feature;

[0017] The initial image features, which include type identifiers and corresponding relationships, are used as the image features of the cross-sectional contour image.

[0018] Optionally, the step of identifying each abnormal image region in the cross-sectional contour image and the edge structure information of the cross-sectional contour image, using an aortic anomaly recognition model, and identifying the abnormal information of each abnormal image region, includes:

[0019] Based on the feature type of each image feature in the cross-sectional contour image, the database is queried for the regular features corresponding to each image feature and the feature image corresponding to each regular feature. Based on the feature image corresponding to each regular feature and the image content of the image feature corresponding to each regular feature, a similarity recognition network is used to identify the similarity between each image feature and the regular feature corresponding to each image feature.

[0020] Image content corresponding to abnormal image features below the similarity threshold is selected as abnormal image regions. Each abnormal image region, the abnormal image features of each abnormal image region, the edge structure information between each abnormal image region, and the feature type of each abnormal image feature are input into the aortic anomaly recognition model to identify the image signs of each abnormal image region.

[0021] All image features of each abnormal image region are used as the abnormal information for each abnormal image region.

[0022] Optionally, identifying the anomaly type and degree of anomaly of the cross-sectional contour image based on the anomaly information of each of the abnormal image regions includes:

[0023] For each abnormal image region, based on each image feature of the abnormal image region, query the image feature database for the sub-abnormality type and the sub-abnormality degree of each image feature.

[0024] The sub-anomaly types of all image features are taken as the anomaly types of the cross-sectional contour image, and the sub-anomaly degrees of all image features are taken as the anomaly degrees of the cross-sectional contour image.

[0025] Optionally, the prediction of the user's aortic dissection probability and the reasons for the prediction of the aortic dissection probability based on the anomaly type and the degree of anomaly of each of the cross-sectional contour images, using an aortic dissection analysis strategy, includes:

[0026] The aortic dissection analysis strategy is broken down into sub-analysis strategies for each abnormality type. For each cross-sectional contour image, based on the sub-abnormality degree of each sub-abnormality type in the cross-sectional contour image, the probability of sub-aortic dissection for each sub-abnormality type is analyzed through the sub-analysis strategy for each abnormality type.

[0027] Obtain the aortic dissection weight value for each abnormality type, and based on the aortic dissection weight value for each abnormality type, perform a weighted calculation on the sub-aortic dissection probability for each sub-abnormality type to obtain the first aortic dissection probability for each sub-abnormality type.

[0028] The sub-abnormality types of the cross-sectional contour image are sorted in descending order of the probability of the first aortic dissection for each sub-abnormality type. Then, the sub-abnormality degree of a preset number of sub-abnormality types is selected in the order of the sub-abnormality type sequence from front to back as the sub-prediction cause of the cross-sectional contour image.

[0029] The probability of first aortic dissection is obtained by summing all the probabilities of first aortic dissection in the cross-sectional contour image, and the probability of second aortic dissection in all cross-sectional contour images is averaged to obtain the probability of aortic dissection of the user.

[0030] The sub-predicted causes of all cross-sectional contour images are used as the predicted causes of the aortic dissection probability.

[0031] Furthermore, to achieve the above objectives, the present invention also provides an auxiliary identification system for aortic dissection, the auxiliary identification system for aortic dissection comprising:

[0032] The acquisition module is used to acquire a plain thoracic scan image of the user's aorta and identify the cross-sectional contour image of the aorta in each of the plain thoracic scan images using an image edge recognition algorithm.

[0033] The extraction module is used to identify the edge structure information of each cross-sectional contour image and extract the image features of each cross-sectional contour image.

[0034] The identification module is used to identify, for each cross-sectional contour image, each abnormal image region and the abnormal information of each abnormal image region in the cross-sectional contour image based on the image features in the cross-sectional contour image and the edge structure information of the cross-sectional contour image, through the aortic abnormality identification model, and to identify the abnormality type and the degree of abnormality of the cross-sectional contour image based on the abnormal information of each abnormal image region.

[0035] The analysis module is used to predict the user's aortic dissection probability and the reason for the prediction of the aortic dissection probability based on the anomaly type and the degree of anomaly of each cross-sectional contour image, using an aortic dissection analysis strategy, and to use the user's aortic dissection probability and the reason for the prediction of the aortic dissection probability as auxiliary identification information for the user's aortic dissection.

[0036] Optionally, the extraction module is specifically used for:

[0037] Based on the aortic location corresponding to each cross-sectional contour image, the sample structure location range of each aortic structure contained in the cross-sectional contour image is identified, and for each cross-sectional contour image, based on each edge contour line in the cross-sectional contour image, the image content contained in each edge contour line is identified.

[0038] Based on the sample structure location range of each aortic structure and the content location range corresponding to the image content contained in each edge contour line, the aortic structure corresponding to each image content is identified;

[0039] Based on each image content, each adjacent edge contour line is identified, and for each pair of adjacent edge contour lines, the sub-edge spacing information between the adjacent edge contour lines is identified. The sub-edge spacing information between each pair of adjacent edge contour lines, as well as the aortic structure corresponding to the image content contained in each pair of adjacent edge contour lines, are used as the edge structure information of the cross-sectional contour image.

[0040] Optionally, the extraction module is specifically used for:

[0041] For each cross-sectional contour image, an image feature extraction network is used to extract the initial image features of the cross-sectional contour image and identify the feature type of each initial image feature.

[0042] Add a type identifier corresponding to the feature type of each initial image feature to each initial image feature, and establish a correspondence between each initial image feature and the image content represented by each initial image feature;

[0043] The initial image features, which include type identifiers and corresponding relationships, are used as the image features of the cross-sectional contour image.

[0044] Optionally, the identification module is specifically used for:

[0045] Based on the feature type of each image feature in the cross-sectional contour image, the database is queried for the regular features corresponding to each image feature and the feature image corresponding to each regular feature. Based on the feature image corresponding to each regular feature and the image content of the image feature corresponding to each regular feature, a similarity recognition network is used to identify the similarity between each image feature and the regular feature corresponding to each image feature.

[0046] Image content corresponding to abnormal image features below the similarity threshold is selected as abnormal image regions. Each abnormal image region, the abnormal image features of each abnormal image region, the edge structure information between each abnormal image region, and the feature type of each abnormal image feature are input into the aortic anomaly recognition model to identify the image signs of each abnormal image region.

[0047] All image features of each abnormal image region are used as the abnormal information for each abnormal image region.

[0048] Optionally, the identification module is specifically used for:

[0049] For each abnormal image region, based on each image feature of the abnormal image region, query the image feature database for the sub-abnormality type and the sub-abnormality degree of each image feature.

[0050] The sub-anomaly types of all image features are taken as the anomaly types of the cross-sectional contour image, and the sub-anomaly degrees of all image features are taken as the anomaly degrees of the cross-sectional contour image.

[0051] Optionally, the analysis module is specifically used for:

[0052] The aortic dissection analysis strategy is broken down into sub-analysis strategies for each abnormality type. For each cross-sectional contour image, based on the sub-abnormality degree of each sub-abnormality type in the cross-sectional contour image, the probability of sub-aortic dissection for each sub-abnormality type is analyzed through the sub-analysis strategy for each abnormality type.

[0053] Obtain the aortic dissection weight value for each abnormality type, and based on the aortic dissection weight value for each abnormality type, perform a weighted calculation on the sub-aortic dissection probability for each sub-abnormality type to obtain the first aortic dissection probability for each sub-abnormality type.

[0054] The sub-abnormality types of the cross-sectional contour image are sorted in descending order of the probability of the first aortic dissection for each sub-abnormality type. Then, the sub-abnormality degree of a preset number of sub-abnormality types is selected in the order of the sub-abnormality type sequence from front to back as the sub-prediction cause of the cross-sectional contour image.

[0055] The probability of first aortic dissection is obtained by summing all the probabilities of first aortic dissection in the cross-sectional contour image, and the probability of second aortic dissection in all cross-sectional contour images is averaged to obtain the probability of aortic dissection of the user.

[0056] The sub-predicted causes of all cross-sectional contour images are used as the predicted causes of the aortic dissection probability.

[0057] Thirdly, this application provides a computer device. The computer device includes a memory and a processor, the memory storing a computer program, and the processor executing the computer program to implement the steps of the method described in any one of the first aspects.

[0058] Fourthly, this application provides a computer-readable storage medium having a computer program stored thereon that, when executed by a processor, implements the steps of the method described in any one of the first aspects.

[0059] Fifthly, this application provides a computer program product. The computer program product includes a computer program that, when executed by a processor, implements the steps of the method described in any one of the first aspects.

[0060] This invention provides an auxiliary identification method and system for aortic dissection. The method includes: acquiring a chest CT scan image of the user's aorta, and identifying the cross-sectional contour image of the aorta in each chest CT scan image using an image edge recognition algorithm; identifying the edge structure information of each cross-sectional contour image and extracting each image feature of each cross-sectional contour image; and for each cross-sectional contour image, based on the image features and edge structure information of the cross-sectional contour image, identifying each abnormality in the cross-sectional contour image using an aortic abnormality identification model. The image region and the abnormal information of each abnormal image region are analyzed. Based on the abnormal information of each abnormal image region, the abnormal type and the degree of abnormality of the cross-sectional contour image are identified. Based on the abnormal type and the degree of abnormality of each cross-sectional contour image, the aortic dissection analysis strategy is used to predict the user's aortic dissection probability and the prediction cause of the aortic dissection probability. The user's aortic dissection probability and the prediction cause of the aortic dissection probability are used as the user's aortic dissection auxiliary identification information. This solution utilizes a standard, high-efficiency plain CT scanner to obtain chest images of the aorta positions within the user's thoracic cavity. Then, it extracts image features and performs edge detection on each cross-sectional contour image to identify its features and edge structure information. Next, an aortic anomaly detection model identifies abnormal image regions and their abnormal information within the cross-sectional contour images. This approach avoids the limitation of CT-enhanced imaging resources being unavailable to ordinary patients, while simultaneously improving the efficiency and accuracy of anomaly identification by combining traditional plain CT scanning with the aortic anomaly detection model. Furthermore, this solution uses the abnormal information from the abnormal image regions in each cross-sectional contour image to identify... The analysis of abnormality types and degrees, along with an aortic dissection analysis strategy, predicts the probability of aortic dissection for the user and the reasons for this prediction. This multi-image, comprehensive, and multi-angle analysis of the user's aortic dissection probability and predicted causes helps generate aortic dissection auxiliary identification information. This avoids the problem of misjudging aortic dissection due to omissions or identification errors when doctors manually review chest CT scans. It uses plain CT scans instead of CT-enhanced images, improving the efficiency of aortic dissection identification. Furthermore, by analyzing the abnormality types and degrees of cross-sectional contour images of multiple aortic chest CT scans, the aortic dissection analysis strategy predicts the user's aortic dissection probability and predicted causes, enhancing the comprehensiveness and accuracy of aortic dissection auxiliary identification.This comprehensively improves the efficiency of auxiliary identification of aortic dissection in emergency situations. Attached Figure Description

[0061] To more clearly illustrate the solutions in this application, the accompanying drawings used in the description of the embodiments of this application will be briefly introduced below. Obviously, the accompanying drawings described below are some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0062] Figure 1 This is a flowchart of the auxiliary identification method for aortic dissection provided in an embodiment of the present invention;

[0063] Figure 2 This is a schematic diagram of the structure of the aortic dissection auxiliary identification system provided in an embodiment of the present invention;

[0064] Figure 3 An internal structural diagram of a computer device provided in an embodiment of the present invention. Detailed Implementation

[0065] The aortic dissection auxiliary identification method provided in this invention is applied to an aortic dissection auxiliary identification system. Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application belongs. The terminology used herein is for the purpose of describing particular embodiments only and is not intended to limit this application. The terms "comprising" and "having," and any variations thereof, in the specification, claims, and accompanying drawings of this application are intended to cover non-exclusive inclusion. The terms "first," "second," etc., in the specification, claims, or accompanying drawings of this application are used to distinguish different objects, not to describe a particular order.

[0066] In this document, the term "embodiment" means that a particular feature, structure, or characteristic described in connection with an embodiment may be included in at least one embodiment of this application. The appearance of this phrase in various places throughout the specification does not necessarily refer to the same embodiment, nor is it a separate or alternative embodiment mutually exclusive with other embodiments. It will be explicitly and implicitly understood by those skilled in the art that the embodiments described herein can be combined with other embodiments.

[0067] To enable those skilled in the art to better understand the present application, the technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the accompanying drawings.

[0068] The aortic dissection auxiliary identification method provided in this application embodiment can be applied to the application environment of aortic dissection auxiliary identification. This method can be applied to a terminal, a server, or a system including both a terminal and a server, and is implemented through interaction between the terminal and the server. The terminal can be, but is not limited to, various personal computers. Specifically, the terminal uses a real-time and efficient plain scanning device to perform a plain scan of the aortic locations in the user's chest cavity to obtain chest cavity images. Then, image feature extraction and edge recognition are performed on each cross-sectional contour image to identify the image features and edge structure information of the cross-sectional contour image. Next, an aortic anomaly identification model is used to identify each abnormal image region and its abnormal information in the cross-sectional contour image. This avoids the limitation of CT enhanced image scanning resources being unavailable to ordinary patients, and improves the efficiency and accuracy of abnormal information identification by using traditional plain scanning methods and the aortic anomaly identification model to identify abnormal information in the user's cross-sectional contour image. Furthermore, this solution further identifies the cross-sectional contour image by using the abnormal information of the abnormal image regions in each cross-sectional contour image. The system analyzes the types and degrees of abnormalities in the aortic dissection images, and then uses an aortic dissection analysis strategy to predict the probability of aortic dissection in the user, as well as the reasons for this prediction. This comprehensive analysis of multiple images from all angles and perspectives helps generate auxiliary aortic dissection identification information for the user. This avoids the problem of misjudging aortic dissection due to omissions or identification errors when doctors manually review chest CT scans. It uses plain CT scans instead of enhanced CT images, improving the efficiency of aortic dissection identification. Furthermore, by analyzing the types and degrees of abnormalities in the cross-sectional contour images of multiple aortic chest CT scans, the aortic dissection analysis strategy predicts the probability and reasons for aortic dissection in the user, improving the comprehensiveness and accuracy of auxiliary aortic dissection identification. This comprehensively improves the efficiency of auxiliary aortic dissection identification in emergency situations.

[0069] In one embodiment, such as Figure 1 As shown, an auxiliary identification method for aortic dissection is provided. Taking the application of this method to a terminal as an example, the method includes the following steps:

[0070] Step S101: Obtain a plain thoracic scan image of the user's aorta, and use an image edge recognition algorithm to identify the cross-sectional contour image of the aorta in each plain thoracic scan image.

[0071] In this embodiment, the terminal uses a CT scanner to perform plain scans at different locations within the user's chest cavity, obtaining different chest cavity images. Then, in response to the staff's image recognition information upload operation, the terminal obtains the aortic location corresponding to each chest cavity image. Next, the terminal uses an image edge recognition algorithm to identify the cross-sectional contour image of the aorta in each chest cavity image. This cross-sectional contour image of the aorta includes the aorta itself and the cross-sectional contour images of its branches. The specific recognition process will be described in detail later.

[0072] Step S102: Identify the edge structure information of each cross-sectional contour image and extract each image feature of each cross-sectional contour image.

[0073] In this embodiment, the terminal identifies the edge structure information of each cross-sectional contour image. Since the cross-sectional contour image includes multiple aortic structural parts, the edge structure information is the edge spacing information of each aortic structural part. This edge spacing information includes the furthest edge spacing and the nearest edge spacing between two aortic structures. Then, the terminal extracts image features from each cross-sectional contour image, where the image features include the image features corresponding to each aortic structure.

[0074] Step S103: For each cross-sectional contour image, based on the image features in the cross-sectional contour image and the edge structure information of the cross-sectional contour image, the aortic abnormality recognition model is used to identify each abnormal image region in the cross-sectional contour image and the abnormal information of each abnormal image region. Based on the abnormal information of each abnormal image region, the abnormality type and the degree of abnormality of the cross-sectional contour image are identified.

[0075] In this embodiment, for each cross-sectional contour image, the terminal, based on the image features and edge structure information of the cross-sectional contour image, uses an aortic anomaly recognition model to identify each abnormal image region and the abnormal information of each abnormal image region. An abnormal image region is the image content corresponding to an abnormal image feature; the method for identifying abnormal image features will be described in detail later. The abnormal information of each abnormal image region is the image sign information of that region, which includes, but is not limited to, calcification displacement, linear high density within the aortic lumen, aortic high density, uneven density within the aortic lumen, aortic widening and branch vessel widening, irregular aortic morphology, peripheral effusion, pericardial effusion, pleural effusion, and high-density shadows in the dorsal lung field. Based on the abnormal information of each abnormal image region, the terminal identifies the abnormal type and degree of abnormality in the cross-sectional contour image. The abnormal type refers to the sign type corresponding to each sign. For example, calcification displacement corresponds to calcification displacement type, linear high density within the aortic lumen corresponds to linear high density type, aortic high density corresponds to aortic high density type, heterogeneous density within the aortic lumen corresponds to heterogeneous density type, aortic widening and branch vessel widening corresponds to vessel widening type, and aortic irregularity corresponds to irregularity type, etc. The degree of abnormality is identified based on the degree of the observed signs; different abnormality types correspond to different degrees of abnormality. For example, the abnormality degree corresponding to the aortic high density type includes different degrees of abnormality corresponding to different high density ranges; the abnormality degree corresponding to the calcification displacement type includes the abnormality degree corresponding to a displacement ≥3mm inward of the outer edge of the calcified aortic concentric circle, and the abnormality degree corresponding to a displacement <3mm inward of the outer edge of the calcified aortic concentric circle, etc.

[0076] Step S104: Based on the anomaly type and the degree of anomaly of each cross-sectional contour image, the aortic dissection analysis strategy is used to predict the user's aortic dissection probability and the reason for the prediction of the aortic dissection probability. The user's aortic dissection probability and the reason for the prediction of the aortic dissection probability are used as the user's aortic dissection auxiliary identification information.

[0077] In this embodiment, based on the anomaly type and degree of each cross-sectional contour image, the terminal uses an aortic dissection analysis strategy to predict the user's aortic dissection probability and the reason for the prediction. This aortic dissection probability and the reason for the prediction are then used as auxiliary information for the user's aortic dissection identification. Different sub-analysis strategies correspond to different anomaly types in the aortic dissection analysis strategy. The aortic dissection probability predicted by the terminal is based on the probability value corresponding to the degree range to which each anomaly type belongs. The correspondence between each degree range and the aortic dissection probability value constitutes the sub-analysis strategy corresponding to different anomaly types. The specific identification process will be explained in detail later. The reason for predicting the probability of each aortic dissection is based on the degree of abnormality corresponding to the larger probability value of a preset number of abnormal types. Based on the above scheme, a conventional plain scan device is used to perform a plain scan of the aortic location in the user's chest cavity to obtain a chest cavity plain scan image. Then, image feature extraction and edge recognition are performed on each cross-sectional contour image to identify the image features and edge structure information of the cross-sectional contour image. Then, the aortic abnormality identification model is used to identify each abnormal image region in the cross-sectional contour image and the abnormal information of each abnormal image region. While avoiding the deficiency of CT enhanced image scanning resources that are not available for ordinary patients, this scheme uses traditional plain scan methods and the aortic abnormality identification model to identify the abnormal information of abnormal image regions in the user's cross-sectional contour image, thereby improving the identification efficiency and accuracy of abnormal information. Secondly, this scheme further uses each cross-section... This method analyzes abnormal information in the contour images, identifies the type and degree of abnormality in the cross-sectional contour images, and then uses an aortic dissection analysis strategy to predict the user's probability of aortic dissection and the reasons for this prediction. By comprehensively analyzing the user's aortic dissection probability and predicted causes through multiple images from all angles, it assists in generating aortic dissection auxiliary identification information. This avoids the problem of misjudging aortic dissection due to omissions or identification errors when doctors manually review chest CT scans. It uses plain CT scans instead of CT-enhanced images, improving the efficiency of aortic dissection identification. Furthermore, by analyzing the abnormal types and degrees of abnormalities in the cross-sectional contour images of multiple aortic chest CT scans, the aortic dissection analysis strategy predicts the user's probability of aortic dissection and the predicted causes, improving the comprehensiveness and accuracy of aortic dissection auxiliary identification. This comprehensively improves the efficiency of aortic dissection auxiliary identification in emergency situations.

[0078] Optionally, identifying the edge structure information of each cross-sectional contour image includes: identifying the sample structure location range of each aortic structure contained in the cross-sectional contour image based on the aortic location corresponding to each cross-sectional contour image; and for each cross-sectional contour image, identifying the image content contained in each edge contour line based on each edge contour line in the cross-sectional contour image; identifying the aortic structure corresponding to each image content based on the sample structure location range of each aortic structure and the content location range corresponding to the image content contained in each edge contour line; identifying each adjacent edge contour line based on each image content; and for each pair of adjacent edge contour lines, identifying the sub-edge spacing information between adjacent edge contour lines; and using the sub-edge spacing information between each pair of adjacent edge contour lines and the aortic structure corresponding to the image content contained in each pair of adjacent edge contour lines as the edge structure information of the cross-sectional contour image.

[0079] In this embodiment, for each cross-sectional contour image, the terminal identifies the image content contained in each edge contour line based on the edge contour lines in the cross-sectional contour image. Based on the sample structure location range of each aortic structure and the content location range corresponding to the image content contained in each edge contour line, the terminal identifies the aortic structure corresponding to each image content.

[0080] Based on the aortic location corresponding to each cross-sectional contour image, the terminal queries the database for the structural range of each aortic structure at each aortic location, as well as the adjacency relationships between each aortic structure. Then, the terminal locates the sample structural location range of each aortic structure in the cross-sectional contour images of each aortic location. This location method involves the terminal identifying the edge contour range of each cross-sectional contour image, then identifying the degree of overlap between each structural range and that edge contour range (i.e., the range of image content contained within the edge contour line) (i.e., the percentage of the area of ​​the overlapping range to the area of ​​the structural range). The terminal then selects the aortic structure corresponding to the structural range with the maximum degree of overlap as the aortic structure of that edge contour range.

[0081] Then, based on each image content, the terminal identifies each adjacent edge contour line, and for each pair of adjacent edge contour lines, identifies the sub-edge spacing information between the adjacent edge contour lines, and uses the sub-edge spacing information between each pair of adjacent edge contour lines, as well as the aortic structure corresponding to the image content contained in each pair of adjacent edge contour lines, as the edge structure information of the cross-sectional contour image.

[0082] Based on the above scheme, the aortic structure of each edge contour range is identified by the overlap of structural ranges. Then, based on the spacing information between each edge contour line, the spacing information of each sub-edge between adjacent edge contour lines is identified, which improves the accuracy of identifying the aortic structure corresponding to each image content and the edge structure information of each cross-sectional contour image.

[0083] Optionally, extract each image feature of each cross-sectional contour image, including: for each cross-sectional contour image, extract initial image features of the cross-sectional contour image through an image feature extraction network, and identify the feature type of each initial image feature; add a type identifier corresponding to the feature type of each initial image feature to each initial image feature, and establish a correspondence between each initial image feature and the image content represented by each initial image feature; use the initial image features containing the type identifier and the correspondence as each image feature of the cross-sectional contour image.

[0084] In this embodiment, for each cross-sectional contour image, the terminal extracts initial image features using an image feature extraction network and identifies the feature type of each initial image feature. This image feature extraction network is a Convolutional Neural Network (CNN) based on a self-attention mechanism, which adds a type identifier corresponding to the feature type of each initial image feature. The feature type refers to the structural type of the aortic structure to which the image feature belongs. The type identifier can be a text identifier, a symbol identifier, or a color identifier, etc.

[0085] The terminal establishes a correspondence between each initial image feature and the image content represented by each initial image feature. Finally, the terminal uses the initial image features, which include type identifiers and correspondences, as the image features of the cross-sectional contour image.

[0086] Based on the above scheme, the structural features of each aortic structure are identified by image feature extraction, thereby improving the comprehensiveness and accuracy of aortic structure identification.

[0087] Optionally, based on the image features in the cross-sectional contour image and the edge structure information of the cross-sectional contour image, the aortic anomaly recognition model identifies each abnormal image region in the cross-sectional contour image and the abnormal information of each abnormal image region. This includes: based on the feature type of each image feature in the cross-sectional contour image, querying the database for the regular features corresponding to each image feature and the feature image corresponding to each regular feature; based on the feature image corresponding to each regular feature and the image content of each regular feature, identifying the similarity between each image feature and the regular features corresponding to each image feature through a similarity recognition network; filtering the image content corresponding to the abnormal image features below the similarity threshold as abnormal image regions; and inputting each abnormal image region, the abnormal image features of each abnormal image region, the edge structure information between each abnormal image region, and the feature type of each abnormal image feature into the aortic anomaly recognition model to identify each image sign of each abnormal image region; and using all image signs of each abnormal image region as the abnormal information of each abnormal image region.

[0088] In this embodiment, the terminal, based on the feature type of each image feature in the cross-sectional contour image, queries the database for the corresponding conventional features and the corresponding feature images for each image feature. Based on the feature images and image content of each conventional feature, a similarity recognition network identifies the similarity between each image feature and its corresponding conventional features. The similarity between the image feature and the conventional features includes the cosine similarity between the two features (calculated using a cosine similarity algorithm) and the image similarity between the feature image and its image content identified by the image similarity recognition network. Then, the terminal calculates the average of the two similarities to obtain the similarity between each image feature and its corresponding conventional features.

[0089] Then, the terminal filters image content corresponding to abnormal image features below a preset similarity threshold as abnormal image regions. Each abnormal image region, its abnormal image features, the edge structure information between each abnormal image region, and the feature type of each abnormal image feature are input into the aortic anomaly recognition model to identify the image signs of each abnormal image region. This aortic anomaly recognition model is a classifier based on a reinforcement learning neural network. Finally, the terminal uses all the image signs of each abnormal image region as the anomaly information for that region.

[0090] Based on the above scheme, by using dual similarity recognition—feature similarity and image similarity—Padding can identify whether each aortic structure has abnormalities, thus improving the accuracy and comprehensiveness of the identification. Then, through an aortic anomaly recognition network, image features of each abnormal image region are identified, avoiding the error rate and anomalies of manual identification. This improves the accuracy of identifying image features of each abnormal image region.

[0091] Optionally, based on the abnormal information of each abnormal image region, the abnormal type and the degree of abnormality of the cross-sectional contour image are identified, including: for each abnormal image region, based on each image feature of the abnormal image region, querying the sub-abnormal type and the sub-abnormal degree of each image feature in the image feature database; taking the sub-abnormal types of all image features as the abnormal type of the cross-sectional contour image, and taking the sub-abnormal degrees of all image features as the degree of abnormality of the cross-sectional contour image.

[0092] In this embodiment, for each abnormal image region, the terminal queries the image feature database to find the sub-abnormality type and sub-abnormality degree of each image feature based on the image features of the abnormal image region. The image feature database includes multiple sub-abnormality degrees corresponding to each abnormality type, as well as the correspondence between each abnormality type and image feature, and the correspondence between each sub-abnormality degree and abnormal feature. By querying these correspondences, the terminal identifies the sub-abnormality type and sub-abnormality degree of the image feature, improving the accuracy and efficiency of the identification.

[0093] Finally, the terminal treats the sub-anomaly types of all image features as the anomaly types of the cross-sectional contour image, and treats the sub-anomaly degrees of all image features as the anomaly degrees of the cross-sectional contour image.

[0094] Based on the above scheme, by identifying the sub-anomaly type and sub-anomaly degree of each image feature, the anomaly type and degree of the cross-sectional contour image can be determined, thereby improving the comprehensiveness of the identification.

[0095] Optionally, based on the anomaly type and degree of each cross-sectional contour image, an aortic dissection analysis strategy is used to predict the user's aortic dissection probability and the reasons for the prediction. This includes: breaking down the aortic dissection analysis strategy into sub-analysis strategies for each anomaly type; for each cross-sectional contour image, based on the degree of sub-anomaly of each sub-anomaly type, analyzing the sub-aortic dissection probability of each sub-anomaly type using the sub-analysis strategy for each anomaly type; obtaining the aortic dissection weight value for each anomaly type; and performing a weighted calculation of the sub-aortic dissection probability for each sub-anomaly type based on the aortic dissection weight value for each anomaly type. The probability of first aortic dissection for each sub-abnormality type is obtained; the probabilities of first aortic dissection for each sub-abnormality type are sorted in descending order to obtain a sequence of sub-abnormality types in the cross-sectional contour image; and the sub-abnormality degree of a preset number of sub-abnormality types is selected as the sub-predicted cause of the cross-sectional contour image in the order of the sub-abnormality type sequence; all first aortic dissection probabilities are summed to obtain the second aortic dissection probability of the cross-sectional contour image; and the second aortic dissection probabilities of all cross-sectional contour images are averaged to obtain the user's aortic dissection probability; and all sub-predicted causes of the cross-sectional contour image are used as the predicted cause of the aortic dissection probability.

[0096] In this embodiment, the terminal breaks down the aortic dissection analysis strategy into sub-analysis strategies for each abnormality type. For each cross-sectional contour image, based on the sub-abnormality degree of each sub-abnormality type, the terminal analyzes the sub-aortic dissection probability for each sub-abnormality type using the sub-analysis strategies for each abnormality type. As described above, each abnormality type's sub-analysis strategy includes the aortic dissection probability corresponding to each abnormality degree. Then, based on this correspondence, the terminal identifies the sub-aortic dissection probability corresponding to the sub-abnormality degree of each abnormality type.

[0097] The terminal acquires the aortic dissection weight value for each abnormality type and, based on this weight value, calculates the sub-aortic dissection probability for each sub-abnormality type, thus obtaining the first aortic dissection probability for each sub-abnormality type. The aortic dissection weight value represents the probability that each aortic abnormality type will cause aortic dissection. For example, the weight values ​​for calcification displacement, linear high-density areas, and high-density aortic regions are high, while other abnormality types have low weight values.

[0098] The terminal sorts the first aortic dissection probabilities of each sub-abnormality type in descending order to obtain the sub-abnormality type sequence of the cross-sectional contour image. Then, according to the sub-abnormality type sequence from front to back, it filters the sub-abnormality degree of a preset number of sub-abnormality types as the sub-prediction cause of the cross-sectional contour image.

[0099] Finally, the terminal sums all the probabilities of first aortic dissection to obtain the probability of second aortic dissection in the cross-sectional contour image, and then averages the probabilities of second aortic dissection in all cross-sectional contour images to obtain the user's aortic dissection probability. The terminal uses the sub-prediction causes of all cross-sectional contour images as the predicted causes of aortic dissection probability.

[0100] Based on the above scheme, probability allocation is performed by using aortic dissection weight values ​​to ensure attention is paid to the degree of sub-abnormality of the abnormal type corresponding to the main signs, thereby improving the accuracy of the judgment of aortic dissection probability.

[0101] It should be understood that although the steps in the flowcharts of the embodiments described above are shown sequentially according to the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless explicitly stated herein, there is no strict order restriction on the execution of these steps, and they can be executed in other orders. Moreover, at least some steps in the flowcharts of the embodiments described above may include multiple steps or multiple stages. These steps or stages are not necessarily completed at the same time, but can be executed at different times. The execution order of these steps or stages is not necessarily sequential, but can be performed alternately or in turn with other steps or at least some of the steps or stages of other steps.

[0102] Based on the same inventive concept, this application also provides an auxiliary identification system for aortic dissection to implement the aforementioned auxiliary identification method. The solution provided by this system is similar to the implementation described in the above method; therefore, the specific limitations of one or more auxiliary identification system embodiments for aortic dissection provided below can be found in the limitations of the auxiliary identification method for aortic dissection described above, and will not be repeated here.

[0103] Further reference Figure 2 As a response to the above Figure 1 The present application provides an embodiment of an auxiliary identification system 200 for aortic dissection, which includes an acquisition module 210, an identification module 220, an identification module 230, and an analysis module 240, wherein:

[0104] The acquisition module 210 is used to acquire a plain thoracic scan image of the user's aorta and identify the cross-sectional contour image of the aorta in each of the plain thoracic scan images using an image edge recognition algorithm.

[0105] The extraction module 220 is used to identify the edge structure information of each cross-sectional contour image and extract the image features of each cross-sectional contour image;

[0106] The identification module 230 is used to identify, for each cross-sectional contour image, each abnormal image region and the abnormal information of each abnormal image region in the cross-sectional contour image based on the image features in the cross-sectional contour image and the edge structure information of the cross-sectional contour image, through the aortic abnormality identification model, and to identify the abnormality type and the degree of abnormality of the cross-sectional contour image based on the abnormal information of each abnormal image region.

[0107] The analysis module 240 is used to predict the user's aortic dissection probability and the reason for the prediction of the aortic dissection probability based on the anomaly type and the degree of anomaly of each of the cross-sectional contour images, using an aortic dissection analysis strategy, and to use the user's aortic dissection probability and the reason for the prediction of the aortic dissection probability as the user's aortic dissection auxiliary identification information.

[0108] Optionally, the extraction module 220 is specifically used for:

[0109] Based on the aortic location corresponding to each cross-sectional contour image, the sample structure location range of each aortic structure contained in the cross-sectional contour image is identified, and for each cross-sectional contour image, based on each edge contour line in the cross-sectional contour image, the image content contained in each edge contour line is identified.

[0110] Based on the sample structure location range of each aortic structure and the content location range corresponding to the image content contained in each edge contour line, the aortic structure corresponding to each image content is identified;

[0111] Based on each image content, each adjacent edge contour line is identified, and for each pair of adjacent edge contour lines, the sub-edge spacing information between the adjacent edge contour lines is identified. The sub-edge spacing information between each pair of adjacent edge contour lines, as well as the aortic structure corresponding to the image content contained in each pair of adjacent edge contour lines, are used as the edge structure information of the cross-sectional contour image.

[0112] Optionally, the extraction module 220 is specifically used for:

[0113] For each cross-sectional contour image, an image feature extraction network is used to extract the initial image features of the cross-sectional contour image and identify the feature type of each initial image feature.

[0114] Add a type identifier corresponding to the feature type of each initial image feature to each initial image feature, and establish a correspondence between each initial image feature and the image content represented by each initial image feature;

[0115] The initial image features, which include type identifiers and corresponding relationships, are used as the image features of the cross-sectional contour image.

[0116] Optionally, the identification module 230 is specifically used for:

[0117] Based on the feature type of each image feature in the cross-sectional contour image, the database is queried for the regular features corresponding to each image feature and the feature image corresponding to each regular feature. Based on the feature image corresponding to each regular feature and the image content of the image feature corresponding to each regular feature, a similarity recognition network is used to identify the similarity between each image feature and the regular feature corresponding to each image feature.

[0118] Image content corresponding to abnormal image features below the similarity threshold is selected as abnormal image regions. Each abnormal image region, the abnormal image features of each abnormal image region, the edge structure information between each abnormal image region, and the feature type of each abnormal image feature are input into the aortic anomaly recognition model to identify the image signs of each abnormal image region.

[0119] All image features of each abnormal image region are used as the abnormal information for each abnormal image region.

[0120] Optionally, the identification module 230 is specifically used for:

[0121] For each abnormal image region, based on each image feature of the abnormal image region, query the image feature database for the sub-abnormality type and the sub-abnormality degree of each image feature.

[0122] The sub-anomaly types of all image features are taken as the anomaly types of the cross-sectional contour image, and the sub-anomaly degrees of all image features are taken as the anomaly degrees of the cross-sectional contour image.

[0123] Optionally, the analysis module 240 is specifically used for:

[0124] The aortic dissection analysis strategy is broken down into sub-analysis strategies for each abnormality type. For each cross-sectional contour image, based on the sub-abnormality degree of each sub-abnormality type in the cross-sectional contour image, the probability of sub-aortic dissection for each sub-abnormality type is analyzed through the sub-analysis strategy for each abnormality type.

[0125] Obtain the aortic dissection weight value for each abnormality type, and based on the aortic dissection weight value for each abnormality type, perform a weighted calculation on the sub-aortic dissection probability for each sub-abnormality type to obtain the first aortic dissection probability for each sub-abnormality type.

[0126] The sub-abnormality types of the cross-sectional contour image are sorted in descending order of the probability of the first aortic dissection for each sub-abnormality type. Then, the sub-abnormality degree of a preset number of sub-abnormality types is selected in the order of the sub-abnormality type sequence from front to back as the sub-prediction cause of the cross-sectional contour image.

[0127] The probability of first aortic dissection is obtained by summing all the probabilities of first aortic dissection in the cross-sectional contour image, and the probability of second aortic dissection in all cross-sectional contour images is averaged to obtain the probability of aortic dissection of the user.

[0128] The sub-prediction causes of all cross-sectional contour images are used as the prediction causes of the aortic dissection probability. Each module in the aforementioned aortic dissection auxiliary identification system can be implemented entirely or partially through software, hardware, or a combination thereof. These modules can be embedded in or independent of the processor in a computer device, or stored in the memory of a computer device as software, so that the processor can call and execute the corresponding operations of each module.

[0129] In one embodiment, a computer device is provided, which may be a terminal, and its internal structure diagram may be as follows: Figure 3 As shown, the computer device includes a processor, memory, communication interface, display screen, and input system connected via a system bus. The processor provides computing and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system and computer programs. The internal memory provides an environment for the operation of the operating system and computer programs in the non-volatile storage media. The communication interface is used for wired or wireless communication with external terminals; wireless communication can be achieved through Wi-Fi, mobile cellular networks, NFC (Near Field Communication), or other technologies. When executed by the processor, the computer program implements an auxiliary identification method for aortic dissection. The display screen can be an LCD screen or an e-ink screen. The input system can be a touch layer covering the display screen, buttons, a trackball, or a touchpad on the computer device's casing, or an external keyboard, touchpad, or mouse.

[0130] Those skilled in the art will understand that Figure 3The structure shown is merely a block diagram of a portion of the structure related to the present application and does not constitute a limitation on the computer device to which the present application is applied. Specific computer devices may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements.

[0131] In one embodiment, a computer device is provided, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the steps of the method described in any one of the first aspects.

[0132] In one embodiment, a computer-readable storage medium is provided having a computer program stored thereon, which, when executed by a processor, implements the steps of the method described in any one of the first aspects.

[0133] In one embodiment, a computer program product is provided, including a computer program that, when executed by a processor, implements the steps of the method described in any one of the first aspects.

[0134] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, data stored, data displayed, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties.

[0135] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium, and when executed, it can include the processes of the embodiments of the above methods. Any references to memory, databases, or other media used in the embodiments provided in this application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetic random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM can take many forms, such as Static Random Access Memory (SRAM) or Dynamic Random Access Memory (DRAM). The databases involved in the embodiments provided in this application may include at least one type of relational database and non-relational database. Non-relational databases may include, but are not limited to, blockchain-based distributed databases. The processors involved in the embodiments provided in this application may be general-purpose processors, central processing units, graphics processing units, digital signal processors, programmable logic devices, quantum computing-based data processing logic devices, etc., and are not limited to these.

[0136] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0137] The embodiments described above are merely illustrative of several implementation methods of this application, and while the descriptions are specific and detailed, they should not be construed as limiting the scope of this patent application. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of this application, and these all fall within the protection scope of this application. Therefore, the protection scope of this application should be determined by the appended claims.

Claims

1. An auxiliary identification method for aortic dissection, characterized in that, The method includes: Acquire a plain thoracic scan image of the user's aorta, and use an image edge recognition algorithm to identify the cross-sectional contour image of the aorta in each of the plain thoracic scan images; Identify the edge structure information of each cross-sectional contour image and extract the image features of each cross-sectional contour image; For each cross-sectional contour image, based on the image features in the cross-sectional contour image and the edge structure information of the cross-sectional contour image, the aortic anomaly recognition model is used to identify each abnormal image region in the cross-sectional contour image and the abnormal information of each abnormal image region. Based on the abnormal information of each abnormal image region, the abnormal type and the degree of abnormality of the cross-sectional contour image are identified. Based on the anomaly type and degree of anomaly of each cross-sectional contour image, an aortic dissection analysis strategy is used to predict the user's aortic dissection probability and the reason for the prediction of the aortic dissection probability. The user's aortic dissection probability and the reason for the prediction of the aortic dissection probability are used as auxiliary identification information for the user's aortic dissection.

2. The method according to claim 1, characterized in that, The identification of edge structure information for each cross-sectional contour image includes: Based on the aortic location corresponding to each cross-sectional contour image, the sample structure location range of each aortic structure contained in the cross-sectional contour image is identified, and for each cross-sectional contour image, based on each edge contour line in the cross-sectional contour image, the image content contained in each edge contour line is identified. Based on the sample structure location range of each aortic structure and the content location range corresponding to the image content contained in each edge contour line, the aortic structure corresponding to each image content is identified; Based on each image content, each adjacent edge contour line is identified, and for each pair of adjacent edge contour lines, the sub-edge spacing information between the adjacent edge contour lines is identified. The sub-edge spacing information between each pair of adjacent edge contour lines, as well as the aortic structure corresponding to the image content contained in each pair of adjacent edge contour lines, are used as the edge structure information of the cross-sectional contour image.

3. The method according to claim 2, characterized in that, The extraction of image features for each cross-sectional contour image includes: For each cross-sectional contour image, an image feature extraction network is used to extract the initial image features of the cross-sectional contour image and identify the feature type of each initial image feature. Add a type identifier corresponding to the feature type of each initial image feature to each initial image feature, and establish a correspondence between each initial image feature and the image content represented by each initial image feature; The initial image features, which include type identifiers and corresponding relationships, are used as the image features of the cross-sectional contour image.

4. The method according to claim 3, characterized in that, Based on the image features in the cross-sectional contour image and the edge structure information of the cross-sectional contour image, the aortic anomaly recognition model identifies each abnormal image region in the cross-sectional contour image and the abnormal information of each abnormal image region, including: Based on the feature type of each image feature in the cross-sectional contour image, the database is queried for the regular features corresponding to each image feature and the feature image corresponding to each regular feature. Based on the feature image corresponding to each regular feature and the image content of the image feature corresponding to each regular feature, a similarity recognition network is used to identify the similarity between each image feature and the regular feature corresponding to each image feature. Image content corresponding to abnormal image features below the similarity threshold is selected as abnormal image regions. Each abnormal image region, the abnormal image features of each abnormal image region, the edge structure information between each abnormal image region, and the feature type of each abnormal image feature are input into the aortic anomaly recognition model to identify the image signs of each abnormal image region. All image features of each abnormal image region are used as the abnormal information for each abnormal image region.

5. The method according to claim 4, characterized in that, The step of identifying the anomaly type and degree of anomaly in the cross-sectional contour image based on the anomaly information of each of the anomaly image regions includes: For each abnormal image region, based on each image feature of the abnormal image region, query the image feature database for the sub-abnormality type and the sub-abnormality degree of each image feature. The sub-anomaly types of all image features are taken as the anomaly types of the cross-sectional contour image, and the sub-anomaly degrees of all image features are taken as the anomaly degrees of the cross-sectional contour image.

6. The method according to claim 5, characterized in that, The method of predicting the user's aortic dissection probability and the reasons for predicting the aortic dissection probability based on the anomaly type and degree of anomaly in each of the cross-sectional contour images, using an aortic dissection analysis strategy, includes: The aortic dissection analysis strategy is broken down into sub-analysis strategies for each abnormality type. For each cross-sectional contour image, based on the sub-abnormality degree of each sub-abnormality type in the cross-sectional contour image, the probability of sub-aortic dissection for each sub-abnormality type is analyzed through the sub-analysis strategy for each abnormality type. Obtain the aortic dissection weight value for each abnormality type, and based on the aortic dissection weight value for each abnormality type, perform a weighted calculation on the sub-aortic dissection probability for each sub-abnormality type to obtain the first aortic dissection probability for each sub-abnormality type. The sub-abnormality types of the cross-sectional contour image are sorted in descending order of the probability of the first aortic dissection for each sub-abnormality type. Then, the sub-abnormality degree of a preset number of sub-abnormality types is selected in the order of the sub-abnormality type sequence from front to back as the sub-prediction cause of the cross-sectional contour image. The probability of first aortic dissection is obtained by summing all the probabilities of first aortic dissection in the cross-sectional contour image, and the probability of second aortic dissection in all cross-sectional contour images is averaged to obtain the probability of aortic dissection of the user. The sub-predicted causes of all cross-sectional contour images are used as the predicted causes of the aortic dissection probability.

7. An auxiliary identification system for aortic dissection, characterized in that, The system includes: The acquisition module is used to acquire a plain thoracic scan image of the user's aorta and identify the cross-sectional contour image of the aorta in each of the plain thoracic scan images using an image edge recognition algorithm. The extraction module is used to identify the edge structure information of each cross-sectional contour image and extract the image features of each cross-sectional contour image. The identification module is used to identify, for each cross-sectional contour image, each abnormal image region and the abnormal information of each abnormal image region in the cross-sectional contour image based on the image features in the cross-sectional contour image and the edge structure information of the cross-sectional contour image, through the aortic abnormality identification model, and to identify the abnormality type and the degree of abnormality of the cross-sectional contour image based on the abnormal information of each abnormal image region. The analysis module is used to predict the user's aortic dissection probability and the reason for the prediction of the aortic dissection probability based on the anomaly type and the degree of anomaly of each cross-sectional contour image, using an aortic dissection analysis strategy, and to use the user's aortic dissection probability and the reason for the prediction of the aortic dissection probability as auxiliary identification information for the user's aortic dissection.

8. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that, When the processor executes the computer program, it implements the steps of the method according to any one of claims 1 to 6.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the steps of the method according to any one of claims 1 to 6.

10. A computer program product, comprising a computer program, characterized in that, When the computer program is executed by a processor, it implements the steps of the method according to any one of claims 1 to 6.

Images