Artery abnormality detection device, electronic apparatus, and storage medium
Through the arterial abnormality detection device, image processing technology is used to detect arterial abnormalities, which solves the problem of lack of effective indications in the prior art, and realizes accurate detection of arterial abnormalities and provides detailed information.
Patent Information
- Application Number
- CN202411749157.4
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-11-29
- Publication Date
- 2025-03-25
AI Technical Summary
The prior art lacks effective indications when detecting arterial abnormalities, making it difficult for doctors to accurately judge the specific abnormalities of the arterial.
An arterial abnormality detection device is proposed, which can detect arterial abnormality through modules such as image acquisition, image segmentation, position determination, diameter calculation, ratio calculation, abnormality determination, abnormal position determination and abnormal length determination, and provide more indications for arterial abnormality.
Accurate detection of arterial abnormalities is achieved, and information on the location and length of arterial abnormalities is provided to help doctors better judge the condition.
Smart Images
Figure CN119693317A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the field of computer technology, and in particular to an arterial abnormality detection device, an aortic dissection prediction model training device, an electronic device and a computer-readable storage medium. Background Art
[0002] Arterial abnormalities may include aneurysms, aneurysmal dilatation, arterial stenosis, and aortic dissection.
[0003] Aneurysms are bulges or weak spots on the walls of arterial vessels. Usually, an arterial diameter that exceeds a certain ratio of the normal arterial diameter, such as more than 50%, can be diagnosed as an aneurysm or aneurysmal dilatation. Aneurysms can generally be caused by high blood pressure, atherosclerosis, rheumatic immune diseases, etc., and may also occur due to factors such as infection. High blood pressure can break through blood vessels and cause aneurysms, which can be life-threatening after rupture, especially cerebral aneurysms, which can cause stroke or even death once ruptured. Atherosclerosis is the most common cause of aneurysms. Due to the appearance of many atherosclerotic plaques in the arteries, these plaques erode the arterial walls, causing thickening of the walls, arterial stenosis or aneurysmal dilatation, and eventually developing into an aneurysm. If you suffer from rheumatic immune diseases, such as Behcet's disease, aneurysmal dilatation will occur, and eventually lead to an aneurysm, and the probability of this disease causing the tumor to rupture and bleed is very high.
[0004] Aneurysmal dilatation refers to an artery that exceeds the normal diameter, but does not reach the standard for diagnosis of an aneurysm, so this stage is called aneurysmal dilatation. Aneurysmal dilatation is mostly asymptomatic when it occurs. The most common cause is arteriosclerosis. In addition, it can also be seen in arterial dilatation secondary to damage to the arteries by some diseases, such as syphilitic arteritis and large arteritis. Arteriosclerosis or arteritis in patients can lead to structural damage to the arterial wall, causing the arterial wall to lose its inherent elasticity and become fragile. Under the continuous impact of high blood pressure, the diameter of the artery gradually stretches, expands, and bulges, eventually forming an aneurysm.
[0005] Arterial stenosis often occurs in the carotid artery, intracranial artery, and coronary artery. Some patients may also experience stenosis of the lower limb arteries. The main causes of arterial stenosis are atherosclerosis, arteritis, and aortic dissection.
[0006] Aortic dissection refers to an aortic disease in which the blood in the aorta enters the middle layer of the arterial wall through the torn aortic intima to form a dissection hematoma, which expands along the long axis of the blood vessel to form a true and false lumen of the artery. Aortic dissection is a fatal disease. It is caused by various causes, such as hypertension, aortic coarctation, atherosclerosis, aortic aneurysm, and large arteritis, which lead to increased pressure or structural changes in the aortic wall, a rupture in the vascular intima, and blood enters between the intima and the media through the rupture to form a false lumen. Symptoms mainly include severe tearing or knife-like pain in the chest and back, even extending to the abdomen. Blood pressure rises, and there is a large difference in blood pressure between the upper limbs and between the upper limbs and lower limbs. Hypotension occurs in severe cases. It may be accompanied by short-term collapse, coma, blindness, loss of sensation or mobility on one side of the limbs, chest tightness, and difficulty breathing.
[0007] Therefore, it is very necessary to detect arterial abnormalities in a timely manner. Currently, after obtaining the patient's aortic image, doctors usually judge the specific abnormality of the artery based on experience, and no more indications are given to doctors. Summary of the invention
[0008] In view of this, the embodiments of the present application propose, on the one hand, an arterial abnormality detection device, and on the other hand, propose a training device, electronic device and computer-readable storage medium for an aortic dissection prediction model, so as to realize the detection of arterial abnormalities and provide users with more indications of arterial abnormalities.
[0009] An arterial abnormality detection device proposed in an embodiment of the present application includes: an image acquisition module, which is used to acquire an image containing an artery in a region to be detected of a current patient; an image segmentation module, which is used to segment an arterial region from the image containing the artery; a position determination module, which is used to determine the position of the centerline of the artery and the cross-section at each position on the centerline from the arterial region; a diameter calculation module, which is used to calculate the maximum diameter of each cross-section; a ratio calculation module, which is used to calculate the ratio between the maximum diameter of each cross-section and the corresponding standard diameter; an abnormality determination module, which is used to compare the ratio between the maximum diameter of each cross-section calculated in sequence and the corresponding standard diameter with a set threshold value, and determine whether the artery has an abnormality according to the comparison result; an abnormality position determination module, which is used to determine the position of the arterial abnormality based on the cross-section when determining that the artery has an abnormality; and an abnormality length determination module, which is used to determine the length of the arterial abnormality based on the position corresponding to the cross-section on the centerline.
[0010] In one embodiment, the abnormality determination module compares the ratio between the maximum diameter of each cross-section calculated in sequence and the corresponding standard diameter with a set arterial dilation threshold, and determines that the artery has an arterial dilation-type abnormality when there is at least one ratio greater than the arterial dilation threshold; and / or compares the ratio between the maximum diameter of each cross-section calculated in sequence and the corresponding standard diameter with a set arterial stenosis threshold, and determines that the artery has an arterial stenosis abnormality when there is at least one ratio less than the arterial stenosis threshold.
[0011] In one embodiment, when the abnormality determination module determines that the artery has an abnormality of arterial dilatation, the abnormal position determination module arranges the maximum diameters of the cross sections at various positions on the center line into a sequence according to their order on the center line; performs change point detection on the sequence to determine the cross section where the starting point and the end point of the arterial widening are located; the abnormal length determination module determines the length of the widened part of the artery based on the corresponding positions of the cross section where the starting point and the end point of the arterial widening are located on the center line;
[0012] In one embodiment, the device further includes: a first abnormality type determination module, which is used to calculate the ratio of the length of the widened part of the artery to the maximum width to obtain a length-to-width ratio. When the length-to-width ratio is within a set spherical threshold range, the widened part is determined to be an aneurysm; otherwise, the widened part is determined to be an aneurysmal dilation; wherein the maximum width is the largest of the maximum diameters of the cross-section of the widened part.
[0013] In one embodiment, the spherical threshold range is [0.5, 1.5].
[0014] In one embodiment, when the abnormality determination module determines that the artery has an arterial stenosis abnormality, the abnormal position determination module arranges the maximum diameters of the cross sections at various positions on the center line into a sequence according to their order on the center line; performs change point detection on the sequence to determine the cross section where the starting point and the end point of the aortic narrowing are located; and the abnormal length determination module determines the length of the arterial stenosis part based on the corresponding positions of the cross section where the starting point and the end point of the arterial narrowing are located on the center line.
[0015] In one embodiment, when the abnormality determination module determines that the artery has an abnormality of arterial dilatation, and the image of the artery in the area to be detected is an image of the chest containing the aorta, the device further includes: a second abnormality type determination module, which is used to fill the image area outside the artery area with a background color; arrange and cut the image containing the artery area filled with the background color along the center line to obtain at least one image block of equal size; and sequentially input the at least one image block into a pre-trained aortic dissection prediction model for prediction to obtain an aortic dissection prediction result output by the aortic dissection prediction model; when the aorta has an aortic dissection, the aortic dissection prediction result has an aortic dissection area mark; the aortic dissection prediction model The training samples of the type include: a positive data sample pair and a negative data sample pair; the positive data sample pair includes an image block with dissection as an input sample and an image block with an aortic dissection area mark as an output sample, and the negative data sample pair includes an image block without dissection as an input sample and an image block without an aortic dissection area mark as an output sample; the abnormal position determination module determines the various cross sections or the nearest cross sections to which the aortic dissection area belongs according to the aortic dissection area mark, and determines the cross section where the starting position of the aortic dissection is located and the cross section where the end position is located; the abnormal length determination module determines the length of the aortic dissection based on the corresponding positions of the cross section where the starting position of the aortic dissection is located and the cross section where the end position is located on the center line.
[0016] In one embodiment, the aortic dissection area is marked as an aortic dissection mask.
[0017] A training device for an aortic dissection prediction model proposed in an embodiment of the present application includes: a first module for acquiring historical images containing the aorta that meet the quantity requirements; a second module for segmenting the aortic region from each historical image containing the aorta; a third module for filling the image area outside the aortic region with a background color; a fourth module for arranging and cutting the image containing the aorta region filled with the background color along the center line to obtain at least one image block of equal size; a fifth module for using the image block with dissection as a positive data input sample and the image block without dissection as a negative data input sample and output sample; a sixth module for marking the aortic dissection region of the image block with dissection to obtain the corresponding positive data output sample; and a seventh module for training an artificial intelligence network using positive data sample pairs including positive data input and output samples and negative data sample pairs including negative data input and output samples to obtain an aortic dissection prediction model.
[0018] An electronic device proposed in an embodiment of the present application includes: at least one memory and at least one processor, wherein: the at least one memory is used to store a computer program; the at least one processor is used to call the computer program stored in the at least one memory to perform the following operations: obtain an image of the current patient's area to be detected that includes an artery; segment an arterial area from the image including the artery; determine the position of the centerline of the artery and the cross-section at each position on the centerline from the arterial area; calculate the maximum diameter of each cross-section; calculate the ratio between the maximum diameter of each cross-section and the corresponding standard diameter, and compare the ratio between the maximum diameter of each cross-section calculated in sequence and the corresponding standard diameter with a set threshold value, and determine whether the artery has an abnormality based on the comparison result; when it is determined that the artery has an abnormality, determine the position of the abnormality based on the cross-section, and determine the length of the abnormality based on the position corresponding to the cross-section on the centerline.
[0019] A computer-readable storage medium proposed in an embodiment of the present application stores a computer program thereon; the computer program can be executed by a processor and implement the following operations: obtaining an image of the current patient's area to be detected including an artery; segmenting the arterial area from the image including the artery; determining the position of the centerline of the artery and the cross-section at each position on the centerline from the arterial area; calculating the maximum diameter of each cross-section; calculating the ratio between the maximum diameter of each cross-section and the corresponding standard diameter, and comparing the ratio between the maximum diameter of each cross-section calculated in sequence and the corresponding standard diameter with a set threshold, and determining whether the artery has an abnormality based on the comparison result; when it is determined that the artery has an abnormality, determining the position of the abnormality based on the cross-section, and determining the length of the abnormality based on the position of the cross-section corresponding to the centerline.
[0020] It can be seen from the above scheme that in the embodiment of the present application, the artery region is segmented from the patient's image containing the artery by using image processing technology, and the position of the centerline of the artery and the cross section at each position on the centerline are determined from the artery region; then the ratio between the maximum diameter of each cross section and the corresponding standard diameter is calculated, and each ratio calculated in turn is compared with a set threshold value, and whether the artery is abnormal is determined based on the comparison result; when it is determined that the artery is abnormal, the position of the artery abnormality is determined based on the cross section, and the length of the artery abnormality is determined based on the corresponding position of the cross section on the centerline. Thereby, the detection of arterial abnormalities is realized, and more abnormal information including the abnormal position and length can be provided to doctors.
[0021] In addition, when determining the presence of abnormalities such as arterial dilatation, by calculating the aspect ratio of the abnormal area and comparing it with a spherical threshold range, it can be determined whether the abnormal type is an aneurysm, thereby realizing the detection of aneurysms and aneurysmal dilatation. When determining the presence of abnormalities such as aortic dilatation, a pre-trained aortic dissection prediction model is used to predict whether there is an arterial dissection in the aortic area, thereby realizing the detection of arterial dissection.
[0022] In addition, when it is determined that there is an abnormality such as arterial dilatation, an aortic dissection prediction model is pre-trained, and the patient's aortic image area is divided into at least one image block of equal size, and the image block is input into the aortic dissection prediction model, so that the prediction result of aortic dissection can be obtained, thereby realizing the detection of aortic dissection.
[0023] At the same time, the technical solution in the embodiment of the present application outputs results in a short time. The output display is intuitive and easy to understand, and the generated quantitative and qualitative results can assist doctors in judging the condition. The measurement results are stable, and the repeatability and ease of use are good. BRIEF DESCRIPTION OF THE DRAWINGS
[0024] The preferred embodiments of the present invention will be described in detail below with reference to the accompanying drawings so that the above and other features and advantages of the present invention will be more apparent to those skilled in the art. In the accompanying drawings:
[0025] Figure 1 This is an exemplary flow chart of the aortic abnormality detection method in an embodiment of the present application.
[0026] Figures 2 to 8 They are Figure 1 An exemplary flow chart of one embodiment of a method for detecting an aortic abnormality is shown.
[0027] Fig. 9 This is an exemplary flow chart of a method for training an aortic dissection prediction model in an embodiment of the present application.
[0028] Fig.10 This is an exemplary structural diagram of an arterial abnormality detection device in an embodiment of the present application.
[0029] Fig.11 This is an exemplary structural diagram of a training device for an aortic dissection prediction model in an embodiment of the present application.
[0030] Fig.12 This is a schematic diagram of the structure of an electronic device in an embodiment of the present application.
[0031] The reference numerals are as follows:
[0032]
[0033] DETAILED DESCRIPTION
[0034] In the embodiment of the present application, in order to obtain more indications of aortic abnormalities, it is considered to use computer technologies such as image processing and related algorithms to extract more indication information for users after obtaining the aortic image.
[0035] In order to make the purpose, technical scheme and effect of the present invention more clearly understood, the specific implementation mode of the present invention is now described with reference to the accompanying drawings. In the drawings, the same reference numerals represent components with the same structure or similar structures but the same functions.
[0036] In this document, “exemplary” and “illustrative” mean “serving as an example, instance or illustration”, and any illustration or implementation described in this document as “exemplary” or “illustrative” should not be construed as a more preferred or more advantageous technical solution.
[0037] In order to simplify the drawings, each figure only schematically shows the parts related to the present invention, which do not represent the actual structure of the product.
[0038] In this article, "a" or "an" not only means "only one", but also means "more than one". In this article, "first", "second", etc. are only used to distinguish each other, rather than to indicate their importance and order.
[0039] Figure 1 FIG. 1 is an exemplary flow chart of the method for detecting aortic abnormality in an embodiment of the present application. Figure 1 As shown, the method may include the following steps:
[0040] Step S11, obtaining an image of the area to be detected of the current patient including the artery.
[0041] In this embodiment, the image containing the artery can be a computed tomography (CT) image, such as a chest (including the scanning position of the chest) CT plain scan image, or a contrast-enhanced CT (CE-CT) image, or a CT arterial angiography (CTA) image, etc., or a digital subtraction angiography (DSA) image, etc., or a magnetic resonance angiography (MRA) image or a contrast-enhanced magnetic resonance angiography (CE-MRA) image, etc.
[0042] Step S12, segmenting the artery region from the image containing the artery.
[0043] In some applications, the segmented artery region may also be referred to as an artery mask.
[0044] Step S13, determining the position of the centerline of the artery and the cross section of each position on the centerline from the artery region.
[0045] In this step, each position on the center line may be each pixel position on the center line, or may be a position on the center line that meets a set requirement, such as positions spaced at the same distance.
[0046] Step S14, calculating the maximum diameter of each cross section.
[0047] Step S15, calculating the ratio between the maximum diameter of each cross section and the corresponding standard diameter in turn.
[0048] Step S16, comparing the ratio between the maximum diameter of each cross section calculated in sequence and the corresponding standard diameter with a set threshold, and determining whether the artery has an abnormality based on the comparison result.
[0049] In one embodiment, this step may include: step S16A, comparing the ratio between the maximum diameter of each cross-section calculated in sequence and the corresponding standard diameter with a set arterial dilation threshold, to determine whether there is at least one ratio greater than the arterial dilation threshold? When it is determined that there is at least one ratio greater than the arterial dilation threshold, it can be determined that the artery has an arterial dilation abnormality, and then step S17 is performed; and / or, step S16B, comparing the ratio between the maximum diameter of each cross-section calculated in sequence and the corresponding standard diameter with a set arterial stenosis threshold, to determine whether there is at least one ratio less than the arterial stenosis threshold? When it is determined that there is at least one ratio less than the arterial stenosis threshold, it can be determined that the artery has an arterial stenosis abnormality, and then step S17 is performed.
[0050] When step S16A and step S16B are included at the same time, step S16A may be executed first and then step S16B, or step S16B may be executed first and then step S16A, or step S16A and step S16B may be executed simultaneously.
[0051] In one example, the arterial dilation threshold may be 1.5. In one example, the arterial stenosis threshold may be 0.5.
[0052] Step S17, when it is determined that the artery has an abnormality, determining the location of the arterial abnormality based on the cross section.
[0053] Step S18, determining the length of the arterial anomaly based on the corresponding position of the cross section on the center line.
[0054] Afterwards, the information of the arterial abnormality can be provided to a user such as a doctor. The information of the arterial abnormality can include: the starting position and length of the arterial abnormality on the artery. Figures 2 to 8In each of the illustrated embodiments, the information on arterial abnormalities may further include the type of arterial abnormalities, such as aneurysm, aneurysmal dilatation, arterial stenosis, and aortic dissection.
[0055] Figure 2 for Figure 1 An exemplary flow chart of one embodiment of a method for detecting aortic abnormality is shown. Figure 2 In the illustrated embodiment, after step S15, step S16A is executed, and when it is determined in step S16A that the artery has an arterial dilatation abnormality, step S17 may specifically include:
[0056] Step S17A, arranging the maximum diameters of the cross sections at various positions on the center line into a sequence according to their order on the center line.
[0057] Step S17B, performing change point detection on the sequence to determine the cross section where the starting point and the end point of the arterial widening are located.
[0058] In the step, the change point detection refers to confirming a point on the sequence through statistical methods, such as the Bayes method, the maximum likelihood method, the local comparison method, etc., that the point set before the point conforms to one distribution, and the point set after the point conforms to another distribution, then the point is a change point of the sequence.
[0059] Among them, the starting point and the end point of arterial widening are respectively a change point on the sequence, the point set before the starting point of arterial widening conforms to the standard diameter distribution, the point set after the starting point conforms to the diameter distribution larger than the standard diameter, the point set before the end point of arterial widening conforms to the diameter distribution larger than the standard diameter, and the point set after the end point conforms to the standard diameter distribution.
[0060] Accordingly, step S18 includes step S18A, determining the length of the widened portion of the artery based on the corresponding positions of the cross section where the starting point of the widened artery is located and the cross section where the end point is located on the center line.
[0061] In addition, the method may further comprise the steps of:
[0062] Step S19, taking the largest of the maximum diameters of the cross section of the widened artery as the maximum width of the widened artery, and calculating the ratio of the length of the widened artery to the maximum width to obtain a length-to-width ratio.
[0063] Step S20, determining whether the aspect ratio is within a set quasi-spherical threshold range. If the aspect ratio is within a set quasi-spherical threshold range, executing step S21; otherwise, executing step S22.
[0064] In an example, the spherical threshold range may be [0.5, 1.5], that is, greater than or equal to 0.5 and less than or equal to 1.5.
[0065] Step S21, determining that the widened portion is an aneurysm.
[0066] Step S22, determining that the widened portion is an aneurysmal dilatation.
[0067] Figure 3 for Figure 1 FIG. 1 is an exemplary flow chart of another embodiment of a method for detecting aortic abnormality. Figure 3 In the illustrated embodiment, after step S15, step S16B is executed, and when it is determined in step S16B that the artery has arterial stenosis abnormality, step S17 may specifically include:
[0068] Step S17C: Arrange the maximum diameters of the cross sections at various positions on the center line into a sequence according to their order on the center line.
[0069] Step S17D: performing change point detection on the sequence to determine the cross section where the starting point and the end point of the arterial narrowing are located.
[0070] Accordingly, step S18 includes step S18B, determining the length of the narrowed portion of the artery based on the corresponding positions of the cross section where the starting point of the narrowed artery is located and the cross section where the end point is located on the center line.
[0071] In another embodiment, Figure 2 and Figure 3 The technical solutions in can also be combined together. Figure 4 An example is shown in Figure 2 and Figure 3 An exemplary flow chart of a technical solution combining the technical solution in .
[0072] Figure 4 In order to execute first after step S15 Figure 2 The indicated branch is executed again Figure 3 Taking the branch shown as an example, after step S15, step S16A is first executed. When it is determined in step S16A that the artery has an abnormality of arterial dilatation, steps S17A, S17B, S18A, and S19 to S22 are sequentially executed. After executing step S21 or step S22, step S16B is executed. When it is determined in step S16B that the artery has an abnormality of arterial stenosis, steps S17C, S17D, and S18B are executed. Alternatively, when it is determined in step S16A that the artery does not have an abnormality of arterial dilatation, step S16B is executed.
[0073] In addition, for Figure 2 and Figure 3 In the case of combining the technical solutions in Figure 4 In other embodiments, after step S15, the Figure 3 The indicated branch is executed again Figure 2 The branch shown, that is, after step S15, firstly execute step S16B, and when it is determined in step S16B that the artery has an abnormality of arterial stenosis, execute steps S17C, S17D and S18B, and then execute S16A, and when it is determined in step S16A that the artery has an abnormality of arterial dilatation, execute steps S17A, S17B, S18A, and steps S19 to S22 in sequence. Alternatively, when it is determined in step S16B that the artery does not have an abnormality of arterial stenosis, execute step S16A.
[0074] In addition, for Figure 2 and Figure 3 In the case of combining the technical solutions in the above, it is also possible to execute in parallel after step S15. Figure 2 The branches and Figure 3 The branch shown, that is, after step S15, the branches shown by steps S16A, S17A, S17B, S18A, and steps S19 to S22, and the branches shown by steps S16B, S17C, S17D and S18B are executed in parallel.
[0075] Figure 5 for Figure 1 FIG. 1 is an exemplary flow chart of another embodiment of a method for detecting aortic abnormality. Figure 5 In the illustrated embodiment, after step S15, step S16A is executed, and when it is determined in step S16A that the artery has an abnormality of arterial dilatation, and the arterial image of the area to be detected is an image of the thoracic aorta, the method may further include before step S17:
[0076] Step S23, filling the image area outside the artery area with a background color (also called Padding processing in some applications).
[0077] Step S24, arranging and cutting the image containing the artery region filled with the background color along the center line to obtain at least one image block of equal size.
[0078] For example, in one example, the size of the image block may be 32*32*32. Of course, in other examples, the image block may be of other sizes.
[0079] Step S25, inputting the at least one image block into a pre-trained aortic dissection prediction model in sequence for prediction, and obtaining an aortic dissection prediction result output by the aortic dissection prediction model; when the aorta has an aortic dissection, the aortic dissection prediction result has an aortic dissection area mark.
[0080] Wherein, the aortic dissection prediction model can be obtained by training an artificial intelligence network using training samples, wherein the training samples may include: positive data sample pairs and negative data sample pairs; the positive data sample pairs include image blocks with dissection as input samples and image blocks with aortic dissection area labels as output samples, and the negative data sample pairs include image blocks without dissection as input samples and image blocks without aortic dissection area labels as output samples.
[0081] In one example, the aortic dissection area marker can be an aortic dissection mask.
[0082] Accordingly, step S17 may include: step S17E, determining each cross section or the nearest cross section to which the aortic dissection area belongs according to the aortic dissection area mark, and determining the cross section where the starting position of the aortic dissection is located and the cross section where the end position is located.
[0083] For example, when the aortic dissection region mark can be an aortic dissection mask, in this step S17E, the cross section or the nearest cross section to which each voxel on the aortic dissection mask belongs can be checked. Specifically, the cross section closest to the end point on the aortic sinus side can be used as the cross section where the dissection starting position is located; the cross section closest to the end point on the abdominal aorta side can be used as the cross section where the dissection end position is located.
[0084] Accordingly, step S18 may include: step S18C, determining the length of the aortic dissection based on corresponding positions of the cross section at the starting position and the cross section at the ending position of the aortic dissection on the center line.
[0085] In another embodiment, Figure 5 The technical solution shown can be Figure 2 The technical solutions shown are combined together. Figure 6 An example is shown in Figure 2 and Figure 5 An exemplary flow chart of a technical solution combining the technical solution in .
[0086] Figure 6 In order to execute first after step S16A Figure 5 The indicated branch is executed again Figure 2Taking the branch shown in FIG. 1 as an example, after step S16A, firstly, steps S23 to S25 are executed, and then step S26 is executed to determine whether the aorta has aortic dissection according to the aortic dissection prediction result. When it is determined that the aorta has aortic dissection according to the aortic dissection prediction result, steps S17E and S18C are executed. When it is determined that the aorta does not have aortic dissection according to the aortic dissection prediction result, steps S17A, S17B, S18A, and steps S19 to S22 are executed.
[0087] Of course, in this method, after determining that the aorta has aortic dissection according to the aortic dissection prediction result and executing step S17E and step S18C, further steps S17A, step S17B, step S18A, and steps S19 to S22 may be executed. Figure 2 In the process of the branch shown, if it is determined in step S20 that the aspect ratio is not within a set spherical threshold range, step S22 may not be executed, but the process ends directly. This is because when aortic dissection exists in the same segment, its aspect ratio is also not within a set spherical threshold range. That is, for the same segment on the aorta, in the arterial abnormality information reported to the user, aortic dissection and aortic aneurysm may exist at the same time, but aortic dissection and aortic aneurysmal dilatation may not exist at the same time.
[0088] In addition, for Figure 2 and Figure 5 In the case of combining the technical solutions in Figure 6 In other embodiments, after step S16A, the following steps may be performed first: Figure 2 The indicated branch is executed again Figure 5 The branch shown in FIG. 1 is that after step S16A, firstly, steps S17A, S17B, S18A, and steps S19 to S22 are executed, and then steps S23 to S25, S17E, and S18C are executed. At this time, for the same segment on the aorta, after executing Figure 2 When the branch shown in FIG. 1 is followed and the presence of aortic aneurysm is determined in step S22, and then steps S23 to S25 are executed to determine the presence of aortic dissection, the final test result reported to the user may only include aortic dissection but not aortic aneurysm, that is, the priority of the test result of aortic dissection is higher than the priority of the test result of aortic aneurysm. In other words, for the same segment of the aorta, in the arterial abnormality information reported to the user, aortic dissection and aortic aneurysm may exist at the same time, but aortic dissection and aortic aneurysm may not exist at the same time.
[0089] In addition, for Figure 2 and Figure 5 In the case of combining the technical solutions in the above, it is also possible to perform the following steps in parallel after step S16A. Figure 2 The branches and Figure 5 That is, after step S16A, steps S17A, S17B, S18A, and steps S19 to S22, and steps S23 to S25, S17E, and S18C are executed synchronously. Figure 2 The branches and Figure 5 When there is a branch shown, if the test results include both aortic dissection and aortic aneurysm-like dilatation, the final test result reported to the user may only include aortic dissection but not aortic aneurysm-like dilatation, that is, the priority of the test result of aortic dissection is higher than the priority of the test result of aortic aneurysm-like dilatation. In other words, for the same segment of the aorta, in the arterial abnormality information reported to the user, aortic dissection and aortic aneurysm may exist at the same time, but aortic dissection and aortic aneurysm-like dilatation may not exist at the same time.
[0090] In yet another embodiment, Figure 5 The technical solution shown can be Figure 3 The technical solutions shown are combined together. Figure 7 An example is shown in Figure 3 and Figure 5 An exemplary flow chart of a technical solution combining the technical solution in .
[0091] Figure 7 In order to execute first after step S15 Figure 5 The indicated branch is executed again Figure 3 Taking the branch shown as an example, after step S15, step S16A is first executed. When it is determined in step S16A that the artery has an abnormality of arterial dilatation, steps S23 to S25, step S17E and step S18C are sequentially executed, and then step S16B is executed. When it is determined in step S16B that the artery has an abnormality of arterial stenosis, steps S17C, step S17D and step S18B are executed. Alternatively, when it is determined in step S16A that the artery does not have an abnormality of arterial dilatation, step S16B is executed.
[0092] In addition, for Figure 5 and Figure 3 In the case of combining the technical solutions in Figure 7 In other embodiments, after step S15, the Figure 3 The indicated branch is executed again Figure 5The branch shown, that is, after step S15, firstly execute step S16B, and when it is determined in step S16B that the artery has an abnormality of arterial stenosis, execute steps S17C, S17D and S18B, and then execute S16A, and when it is determined in step S16A that the artery has an abnormality of arterial dilatation, execute steps S23 to S25, S17E and S18C in sequence. Alternatively, when it is determined in step S16B that the artery does not have an abnormality of arterial stenosis, execute step S16A.
[0093] In addition, for Figure 5 and Figure 3 In the case of combining the technical solutions in the above, it is also possible to execute in parallel after step S15. Figure 5 The branches and Figure 3 The branch shown, that is, after step S15, the branch shown by steps S16A, steps S23 to S25, step S17E and step S18C, and the branch shown by steps S16B, step S17C, step S17D and step S18B are executed in parallel.
[0094] In yet another embodiment, Figure 5 The technical solution shown can be Figure 2 and Figure 3 The technical solutions shown are combined together. Figure 8 An example is shown in Figure 2 , Figure 3 and Figure 5 An exemplary flow chart of a technical solution combining the technical solution in .
[0095] Figure 8 In order to execute first after step S15 Figure 5 The indicated branch is executed again Figure 2 The branch shown is then executed Figure 3Taking the branch shown as an example, after step S15, step S16A is first executed. When it is determined in step S16A that the artery has an abnormality of arterial dilatation, steps S23 to S25 are first executed, and then step S26 is executed to determine whether the aorta has an aortic dissection according to the aortic dissection prediction result. When it is determined that the aorta has an aortic dissection according to the aortic dissection prediction result, steps S17E and S18C are executed. When it is determined that the aorta does not have an aortic dissection according to the aortic dissection prediction result, steps S17A, S17B, S18A, and steps S19 to S22 are executed, and finally step S16B is executed. When it is determined in step S16B that the artery has an abnormality of arterial stenosis, steps S17C, S17D, and S18B are executed. Alternatively, when it is determined in step S16A that the artery does not have an abnormality of arterial dilatation, step S16B is executed. Of course, in this method, after determining that the aorta has aortic dissection according to the aortic dissection prediction result and executing step S17E and step S18C, further executing Figure 2 At this time, for the same segment of the aorta, after confirming that the segment has aortic dissection, Figure 2 In the process of the branch shown, if it is determined in step S20 that the aspect ratio is not within a set spherical threshold range, step S22 may not be performed, but the process ends directly, because when aortic dissection exists in the same section, its aspect ratio is also not within a set spherical threshold range.
[0096] In addition, for Figure 2 , Figure 3 and Figure 5 In the case of combining the technical solutions in Figure 8 In other embodiments, when it is determined in step S16A that the artery has an arterial dilatation abnormality, the following steps may be performed first: Figure 2 The indicated branch is executed again Figure 5 The branch shown, that is, after step S16A, first execute step S17A, step S17B, step S18A, and step S19 to step S22, and then execute step S23 to step S25, step S17E and step S18C. Then, execute step S16B. If it is determined in step S16B that the artery has arterial stenosis abnormality, execute step S17C, step S17D and step S18B. Alternatively, if it is determined in step S16A that the artery does not have arterial dilatation abnormality, execute step S16B. At this time, for the same segment on the aorta, when executing Figure 2When the branch shown in the figure determines that aortic aneurysmal dilatation exists in step S22 and then executes steps S23 to S25 and determines that aortic dissection exists, the final test result reported to the user may only include aortic dissection but not aortic aneurysmal dilatation, that is, the priority of the test result of aortic dissection is higher than the priority of the test result of aortic aneurysmal dilatation.
[0097] Alternatively, when it is determined in step S16A that the aorta has an arterial dilatation abnormality, the following steps may be performed simultaneously: Figure 5 The branches and Figure 2 Then execute the branch corresponding to step S16B. At this time, for the same segment on the aorta, execute Figure 2 The branches and Figure 5 When the branch is shown, if the test results include both aortic dissection and aortic aneurysmal dilatation, the final test results reported to the user may only include aortic dissection but not aortic aneurysmal dilatation, that is, the priority of the test result of aortic dissection is higher than the priority of the test result of aortic aneurysmal dilatation.
[0098] In addition, for Figure 2 , Figure 3 and Figure 5 In the case of combining the technical solutions in the embodiment, after step S15, the branch corresponding to step S16A and the branch corresponding to step S16B can be executed in parallel. And, after step S16A, both the branch corresponding to step S16A and the branch corresponding to step S16B can be executed in parallel. Figure 2 The branches and Figure 5 The branches shown can also be Figure 8 Execute first as shown in Figure 5 The indicated branch is executed again Figure 2 The branch shown, or you can execute first Figure 2 The indicated branch is executed again Figure 5 Branches shown.
[0099] Again, for Figure 2 , Figure 3 and Figure 5 In the case of combining the technical solutions in the embodiment, after step S15, the branch corresponding to step S16B may be executed first, and then the branch corresponding to step S16A may be executed. Moreover, after step S16A, both the branch corresponding to step S16B and the branch corresponding to step S16A may be executed in parallel. Figure 2 The branches and Figure 5 The branches shown can also be Figure 8 Execute first as shown in Figure 5 The indicated branch is executed again Figure 2 The branch shown, or you can execute first Figure 2 The indicated branch is executed again Figure 5 Branches shown.
[0100] That is to say, Figure 2 , Figure 3 and Figure 5 The technical solutions in the invention can be combined arbitrarily.
[0101] In addition, for Figure 5 The aortic dissection prediction model used in the technical solution shown, and a training method for the aortic dissection prediction model are also provided in the embodiments of the present application. Fig. 9 FIG. 2 shows an exemplary flow chart of a method for training an aortic dissection prediction model in an embodiment of the present application. Fig. 9 As shown, the method may include the following processing:
[0102] Step S91, obtaining historical images including the aorta that meet the quantity requirement.
[0103] Step S92: for each historical image containing the aorta, segment the aorta region from the historical image containing the aorta.
[0104] Step S93, filling the image area outside the aorta area with the background color.
[0105] Step S94, arranging and cutting the image containing the aorta region filled with the background color along the center line to obtain at least one image block of equal size.
[0106] Step S95, taking the image blocks with interlayer as positive data input samples, and taking the image blocks without interlayer as negative data input samples and output samples.
[0107] Step S96, marking the aortic dissection area of the image block with dissection to obtain the corresponding positive data output sample.
[0108] Step S97, using the positive data sample pairs including positive data input and output samples and the negative data sample pairs including negative data input and output samples to train an artificial intelligence network to obtain an aortic dissection prediction model.
[0109] The above is a detailed description of the arterial abnormality detection method and the aortic dissection prediction model training method in the embodiment of the present application. The arterial abnormality detection device and the aortic dissection prediction model training device in the embodiment of the present application are described in detail below. The arterial abnormality detection device and the aortic dissection prediction model training device in the embodiment of the present application can be used to implement the corresponding method in the embodiment of the present application. For details not disclosed in detail in the embodiment of the device of the present invention, please refer to the corresponding description in the embodiment of the method of the present invention, and will not be repeated here one by one.
[0110] Fig.10 FIG. 1 is an exemplary structural diagram of an arterial abnormality detection device in an embodiment of the present application. Fig.10 As shown, the device may include: an image acquisition module 1001, an image segmentation module 1002, a position determination module 1003, a diameter calculation module 1004, a ratio calculation module 1005, an abnormality determination module 1006, an abnormal position determination module 1007 and an abnormal length determination module 1008.
[0111] The image acquisition module 1001 is used to acquire an image of the area to be detected of the current patient including the artery.
[0112] The image segmentation module 1002 is used to segment an artery region from the image containing the artery.
[0113] The position determination module 1003 is used to determine the position of the centerline of the artery and the cross section of each position on the centerline from the artery region.
[0114] The diameter calculation module 1004 is used to calculate the maximum diameter of each cross section.
[0115] The ratio calculation module 1005 is used to calculate the ratio between the maximum diameter of each cross section and the corresponding standard diameter.
[0116] The abnormality determination module 1006 is used to compare the ratio between the maximum diameter of each cross section calculated in sequence and the corresponding standard diameter with a set threshold, and determine whether the artery has an abnormality according to the comparison result.
[0117] The abnormality position determination module 1007 is used to determine the position of the arterial abnormality based on the cross section when it is determined that the artery has an abnormality.
[0118] The abnormal length determination module 1008 is used to determine the length of the arterial abnormality based on the corresponding position of the cross section on the center line.
[0119] In one embodiment, the abnormality determination module 1006 can be used to compare the ratio between the maximum diameter of each cross-section calculated in sequence and the corresponding standard diameter with a set arterial dilation threshold, and when there is at least one ratio greater than the arterial dilation threshold, determine that the artery has an arterial dilation-type abnormality; and / or, compare the ratio between the maximum diameter of each cross-section calculated in sequence and the corresponding standard diameter with a set arterial stenosis threshold, and when there is at least one ratio less than the arterial stenosis threshold, determine that the artery has an arterial stenosis abnormality.
[0120] In another embodiment, when the abnormality determination module 1006 determines that the artery has an abnormality of arterial dilatation, the abnormal position determination module 1007 can arrange the maximum diameters of the cross sections at various positions on the center line into a sequence according to their order on the center line; perform change point detection on the sequence to determine the cross section where the starting point of the arterial widening is located and the cross section where the end point is located; the abnormal length determination module 1008 can determine the length of the widened part of the artery based on the corresponding positions of the cross section where the starting point of the arterial widening is located and the cross section where the end point is located on the center line. At the same time, the arterial abnormality detection device can further include: a first abnormality type determination module 1009, which is used to calculate the ratio of the length of the widened part of the artery to the maximum width to obtain an aspect ratio, and when the aspect ratio is within a set spherical threshold range, the widened part is determined to be an aneurysm; otherwise, the widened part is determined to be an aneurysmal dilatation; wherein the maximum width is the largest of the maximum diameters of the cross sections of the widened part.
[0121] In another embodiment, when the abnormality determination module 1006 determines that the artery has an abnormality of arterial dilatation, and the arterial image of the area to be detected is a chest aorta image, the arterial abnormality detection device may further include: a second abnormality type determination module 1010, which is used to fill the image area outside the arterial area with a background color; arrange and cut the image containing the arterial area filled with background color along the center line to obtain at least one image block of equal size; input the at least one image block into a pre-trained aortic dissection prediction model in sequence for prediction, and obtain the aortic dissection prediction result output by the aortic dissection prediction model; when the aorta has an aortic dissection, the aortic dissection prediction result has an aortic dissection area mark. The training samples of the aortic dissection prediction model include: a positive data sample pair and a negative data sample pair; the positive data sample pair includes an image block with a dissection as an input sample and an image block with an aortic dissection area mark as an output sample, and the negative data sample pair includes an image block without a dissection as an input sample and an image block without an aortic dissection area mark as an output sample. Accordingly, the abnormal position determination module 1007 can determine the various cross sections or the nearest cross sections to which the aortic dissection region belongs according to the aortic dissection region mark, and determine the cross section where the aortic dissection starts and where the end position is located. The abnormal length determination module 1008 can determine the length of the aortic dissection based on the corresponding positions of the cross sections where the aortic dissection starts and where the end position is located on the center line.
[0122] In another embodiment, when the abnormality determination module 1006 determines that the artery has an abnormality of arterial stenosis, the abnormality position determination module 1007 may arrange the maximum diameters of the cross sections at various positions on the center line into a sequence according to their order on the center line; perform change point detection on the sequence to determine the cross section where the starting point of aortic stenosis is located and the cross section where the end point is located. The abnormality length determination module 1008 may determine the length of the arterial stenosis part based on the corresponding positions of the cross section where the starting point of arterial stenosis is located and the cross section where the end point is located on the center line.
[0123] Fig.11 FIG. 1 is an exemplary structural diagram of a training device for an aortic dissection prediction model in an embodiment of the present application. The device can be used to train the aforementioned aortic dissection prediction model. Fig.11 As shown, the device may include: a first module 1101, a second module 1102, a third module 1103, a fourth module 1104, a fifth module 1105, a sixth module 1106 and a seventh module 1107.
[0124] The first module 1101 is used to obtain historical images containing the aorta that meet the quantity requirements.
[0125] The second module 1102 is used to segment the aorta region from each historical image containing the aorta.
[0126] The third module 1103 is used to fill the image area outside the aorta area with the background color.
[0127] The fourth module 1104 is used to arrange and cut the image containing the aorta region filled with the background color along the center line to obtain at least one image block of equal size.
[0128] The fifth module 1105 is used to take the image blocks with interlayer as positive data input samples, and the image blocks without interlayer as negative data input samples and output samples.
[0129] The sixth module 1106 is used to mark the aortic dissection area of the image block with dissection to obtain the corresponding positive data output sample.
[0130] The seventh module 1107 is used to train an artificial intelligence network using positive data sample pairs including positive data input and output samples and negative data sample pairs including negative data input and output samples to obtain an aortic dissection prediction model.
[0131] Fig.12 is a schematic diagram of the structure of an electronic device in an embodiment of the present application, such as Fig.12As shown, the device may include: at least one memory 1201 and at least one processor 1202. In addition, some other components may also be included, such as communication ports, etc. These components communicate via a bus 1203.
[0132] Among them, at least one memory 1201 is used to store a computer program. In one embodiment, the computer program can be understood to include Fig.10 The various modules or Fig.11 The modules of the training device for the aortic dissection prediction model shown in FIG. In addition, at least one memory 1201 can also store an operating system, etc. The operating system includes but is not limited to: Android operating system, Symbian operating system, Windows operating system, Linux operating system, etc.
[0133] At least one processor 1202 is used to call the computer program stored in at least one memory 1201 to execute the arterial abnormality detection method or the aortic dissection prediction model training method described in the embodiment of the present application. The processor 1202 can be a CPU, a processing unit / module, an ASIC, a logic module or a programmable gate array, etc. It can receive and send data through the communication port.
[0134] It should be noted that not all steps and modules in the above processes and structure diagrams are necessary, and some steps or modules can be ignored according to actual needs. The execution order of each step is not fixed and can be adjusted as needed. The division of each module is only for the convenience of describing the functional division adopted. In actual implementation, a module can be implemented by multiple modules, and the functions of multiple modules can also be implemented by the same module. These modules can be located in the same device or in different devices.
[0135] It is understood that the hardware modules in the above-mentioned embodiments can be implemented mechanically or electronically. For example, a hardware module may include a specially designed permanent circuit or logic device (such as a dedicated processor, such as an FPGA or ASIC) for completing a specific operation. The hardware module may also include a programmable logic device or circuit (such as a general-purpose processor or other programmable processor) temporarily configured by software for performing a specific operation. As for whether to implement the hardware module in a mechanical manner, or using a dedicated permanent circuit, or using a temporarily configured circuit (such as configured by software), it can be determined based on cost and time considerations.
[0136] In addition, a computer program product is also provided in the embodiment of the present application, including a computer program, which can be executed by a processor and implement the arterial abnormality detection method or the aortic dissection prediction model training method described in the embodiment of the present application. A computer-readable storage medium is also provided in the embodiment of the present application, on which a computer program is stored, and the computer program can be executed by a processor and implement the arterial abnormality detection method or the aortic dissection prediction model training method described in the embodiment of the present application. Specifically, a system or device equipped with a storage medium can be provided, on which a software program code that implements the functions of any of the above-mentioned embodiments is stored, and the computer (or CPU or MPU) of the system or device reads and executes the program code stored in the storage medium. In addition, the operating system operating on the computer can also be used to complete part or all of the actual operations based on the instructions of the program code. The program code read from the storage medium can also be written to a memory provided in an expansion board inserted into the computer or to a memory provided in an expansion unit connected to the computer, and then based on the instructions of the program code, a CPU installed on the expansion board or the expansion unit is made to perform part and all of the actual operations, thereby realizing the functions of any of the above-mentioned embodiments. Storage media implementations for providing program code include USB, floppy disk, hard disk, magneto-optical disk, optical disk (such as CD-ROM, CD-R, CD-RW, DVD-ROM, DVD-RAM, DVD-RW, DVD+RW), magnetic tape, non-volatile memory card and ROM, etc. Alternatively, the program code can be downloaded from a server computer via a communication network.
[0137] It can be seen from the above scheme that in the embodiment of the present application, the artery image is segmented from the patient's artery image by using image processing technology, and the position of the centerline of the artery and the cross section at each position on the centerline are determined from the artery image; then the ratio between the maximum diameter of each cross section and the corresponding standard diameter is calculated, and each ratio calculated in turn is compared with a set threshold value, and whether the artery is abnormal is determined based on the comparison result; when it is determined that the artery is abnormal, the position of the artery abnormality is determined based on the cross section, and the length of the artery abnormality is determined based on the corresponding position of the cross section on the centerline. Thus, the detection of arterial abnormalities is achieved, and more abnormal information including the abnormal position and length can be provided to doctors.
[0138] In addition, when determining the presence of arterial dilatation abnormalities, by calculating the aspect ratio of the abnormal area and comparing it with a spherical threshold range, it can be determined whether the abnormal type is an aneurysm, thereby achieving the detection of aneurysms and aneurysmal dilatation.
[0139] In addition, when it is determined that there is an abnormality such as arterial dilatation, an aortic dissection prediction model is pre-trained, and the patient's aortic image area is divided into at least one image block of equal size, and the image block is input into the aortic dissection prediction model, so that the prediction result of aortic dissection can be obtained, thereby realizing the detection of aortic dissection.
[0140] At the same time, the technical solution in the embodiment of the present application outputs results in a short time. The output display is intuitive and easy to understand, and the generated quantitative and qualitative results can assist doctors in judging the condition. The measurement results are stable, and the repeatability and ease of use are good.
[0141] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principle of the present invention should be included in the protection scope of the present invention.
Claims
1. An arterial abnormality detection device, characterized in that: include: An image acquisition module (1001) is used to acquire an image of the area to be detected of the current patient including the artery; An image segmentation module (1002), used for segmenting an artery region from the image containing the artery; A position determination module (1003) is used to determine the position of the center line of the artery and the cross section of each position on the center line from the artery region; A diameter calculation module (1004), used to calculate the maximum diameter of each cross section; A ratio calculation module (1005), used to calculate the ratio between the maximum diameter of each cross section and the corresponding standard diameter; An abnormality determination module (1006), used to compare the ratio between the maximum diameter of each cross section calculated in sequence and the corresponding standard diameter with a set threshold, and determine whether there is an abnormal section in the artery according to the comparison result; An abnormality location determination module (1007), configured to determine the location of the abnormal section based on the cross section when it is determined that the artery has an abnormality; and The abnormal length determination module (1008) is used to determine the length of the abnormal section based on the corresponding position of the cross section on the center line.
2. The arterial abnormality detection device according to claim 1, characterized in that: The abnormality determination module (1006) compares the ratio between the maximum diameter of each cross section calculated in sequence and the corresponding standard diameter with a set arterial dilation threshold, and determines that the artery has an arterial dilation abnormality when there is at least one ratio greater than the arterial dilation threshold; and / or, The ratio between the maximum diameter of each cross section calculated in sequence and the corresponding standard diameter is compared with a set arterial stenosis threshold value, and when there is at least one ratio smaller than the arterial stenosis threshold value, it is determined that the artery has arterial stenosis abnormality.
3. The arterial abnormality detection device according to claim 2, characterized in that: When the abnormality determination module (1006) determines that the artery has an abnormality of arterial dilatation, the abnormality position determination module (1007) arranges the maximum diameters of the cross sections at various positions on the center line into a sequence according to their order on the center line; performs change point detection on the sequence to determine the cross section where the starting point and the end point of the arterial widening are located; The abnormal length determination module (1008) determines the length of the widened portion of the artery based on the corresponding positions of the cross section where the starting point of the widened artery is located and the cross section where the end point is located on the center line; The device further comprises: a first abnormality type determination module (1009), which is used to calculate the ratio of the length of the widened part of the artery to the maximum width to obtain a length-to-width ratio. When the length-to-width ratio is within a set spherical threshold range, the widened part is determined to be an aneurysm; otherwise, the widened part is determined to be an aneurysmal dilation; wherein the maximum width is the largest of the maximum diameters of the cross section of the widened part.
4. The arterial abnormality detection device according to claim 3, characterized in that: The spherical threshold range is [0.5, 1.5].
5. The arterial abnormality detection device according to claim 2, characterized in that: When the abnormality determination module (1006) determines that the artery has an abnormality of arterial stenosis, the abnormality position determination module (1007) arranges the maximum diameters of the cross sections at various positions on the center line into a sequence according to their order on the center line; performs change point detection on the sequence to determine the cross section where the starting point and the end point of aortic stenosis are located; The abnormal length determination module (1008) determines the length of the arterial stenosis portion based on the corresponding positions of the cross section where the starting point of the arterial stenosis is located and the cross section where the end point of the arterial stenosis is located on the center line.
6. The arterial abnormality detection device according to any one of claims 1 to 5, characterized in that: When the abnormality determination module (1006) determines that the artery has an abnormality of arterial dilatation, and the image of the area to be detected including the artery is an image of the chest including the aorta, the device further comprises: The second abnormality type determination module (1010) is used to fill the image area outside the artery area with a background color; arrange and cut the image containing the artery area filled with the background color along the center line to obtain at least one image block of equal size; input the at least one image block in sequence into a pre-trained aortic dissection prediction model for prediction to obtain an aortic dissection prediction result output by the aortic dissection prediction model; when the aorta has an aortic dissection, the aortic dissection prediction result has an aortic dissection area mark; the training samples of the aortic dissection prediction model include: a positive data sample pair and a negative data sample pair; the positive data sample pair includes an image block with a dissection as an input sample and an image block with an aortic dissection area mark as an output sample, and the negative data sample pair includes an image block without a dissection as an input sample and an image block without an aortic dissection area mark as an output sample; The abnormal position determination module (1007) determines each cross section or the nearest cross section to which the aortic dissection region belongs according to the aortic dissection region mark, and determines the cross section where the aortic dissection start position is located and the cross section where the end position is located; The abnormal length determination module (1008) determines the length of the aortic dissection based on the corresponding positions of the cross section at the starting position and the cross section at the end position of the aortic dissection on the center line.
7. The arterial abnormality detection device according to claim 6, characterized in that: The aortic dissection area is marked as the aortic dissection mask.
8. A training device for an aortic dissection prediction model, characterized in that: include: A first module (1101) is used to obtain historical images including the aorta that meet the quantity requirements; A second module (1102) is used for segmenting the aorta region from each historical image containing the aorta; The third module (1103) is used to fill the image area outside the aorta area with a background color; A fourth module (1104) is used to arrange and cut the image containing the aorta region filled with background color along the center line to obtain at least one image block of equal size; A fifth module (1105) is used to take the image block with interlayer as a positive data input sample, and the image block without interlayer as a negative data input sample and output sample; The sixth module (1106) is used to mark the aortic dissection area of the image block with dissection to obtain the corresponding positive data output sample; and The seventh module (1107) is used to train an artificial intelligence network using positive data sample pairs including positive data input and output samples and negative data sample pairs including negative data input and output samples to obtain an aortic dissection prediction model.
9. An electronic device, characterized in that include: At least one memory (1101) and at least one processor (1102), wherein: The at least one memory (1101) is used to store a computer program; The at least one processor (1102) is used to call the computer program stored in the at least one memory (1101) to perform the following operations: Acquire an image of the area to be detected of the current patient including an artery; Segmenting an artery region from the image containing the artery; Determine the position of the centerline of the artery and the cross-section at each position on the centerline from the artery region; calculate the maximum diameter of each cross-section; Calculating the ratio between the maximum diameter of each cross section and the corresponding standard diameter, and comparing the ratio between the maximum diameter of each cross section and the corresponding standard diameter calculated in sequence with a set threshold, and determining whether the artery has an abnormality according to the comparison result; When it is determined that the artery has an abnormality, the location of the abnormality is determined based on the cross section, and the length of the abnormality is determined based on the corresponding position of the cross section on the center line.
10. A computer-readable storage medium having a computer program stored thereon; characterized in that: The computer program can be executed by a processor and implement the following operations: Acquire an image of the area to be detected of the current patient including an artery; Segmenting an artery region from the image containing the artery; Determine the position of the centerline of the artery and the cross-section at each position on the centerline from the artery region; calculate the maximum diameter of each cross-section; Calculating the ratio between the maximum diameter of each cross section and the corresponding standard diameter, and comparing the ratio between the maximum diameter of each cross section and the corresponding standard diameter calculated in sequence with a set threshold, and determining whether the artery has an abnormality according to the comparison result; When it is determined that the artery has an abnormality, the location of the abnormality is determined based on the cross section, and the length of the abnormality is determined based on the corresponding position of the cross section on the center line.