Methods, apparatus, equipment and storage media for the extraction of aortic dissection membrane

Abstract

This application provides a method, apparatus, device, and storage medium for extracting aortic dissection membrane. The method includes: determining the corresponding dissection membrane boundary region based on the true lumen region and false lumen region within an aortic image; determining seed point pairs within the aortic image based on the pixel gradient within the dissection membrane boundary region; and clustering the pixels within the aortic image using the seed point pairs to obtain the corresponding aortic dissection membrane region. The embodiments of this application can achieve accurate extraction of the aortic dissection membrane region within an aortic image, ensuring automated and effective identification of the aortic dissection membrane region within an aortic image, enhancing the versatility and efficiency of aortic dissection membrane extraction, and reducing the treatment risk of aortic dissection.

Description

Technical Field

[0001] This application relates to the field of image processing technology, specifically to a method, apparatus, device, and storage medium for extracting aortic dissection membrane. Background Technology

[0002] Aortic dissection refers to the separation of blood from the aortic media by tearing through the intima, causing the blood to flow into the aortic wall and creating a true and false lumen separation along the long axis of the aorta. Aortic dissection is an acute, rapidly progressing, and serious cardiovascular disease. If it is not diagnosed and treated early, it can severely threaten the patient's life.

[0003] Currently, the localization and extraction of aortic dissection membrane usually relies on doctors manually annotating the images based on the patient's computed tomography angiography (CTA) images. This method requires a certain level of experience and skill from the doctor and cannot guarantee the accuracy of aortic dissection membrane localization, thus posing certain treatment risks. Summary of the Invention

[0004] This application provides a method, apparatus, device, and storage medium for extracting aortic dissection membrane, enabling accurate extraction of the aortic dissection membrane region within aortic imaging, enhancing the versatility and efficiency of aortic dissection membrane extraction, and reducing the treatment risks of aortic dissection.

[0005] In a first aspect, embodiments of this application provide a method for extracting aortic dissection membrane, the method comprising:

[0006] The corresponding dissecting membrane junction area is determined based on the true lumen area and false lumen area in the aortic image;

[0007] Seed point pairs in the aortic image are determined based on the pixel gradient in the interface region of the dissecting membrane.

[0008] Using the seed point pairs, the pixels in the aortic image are clustered to obtain the corresponding aortic dissection membrane region.

[0009] Secondly, embodiments of this application provide an extraction device for aortic dissection membrane, the device comprising:

[0010] The junction region determination module is used to determine the corresponding dissecting membrane junction region based on the true lumen region and the false lumen region in the aortic image;

[0011] The seed point pair determination module is used to determine the seed point pairs in the aortic image based on the pixel gradient in the interlaminar membrane junction area.

[0012] The aortic dissection membrane extraction module is used to cluster the pixels in the aortic image using the seed point pairs to obtain the corresponding aortic dissection membrane region.

[0013] Thirdly, embodiments of this application provide an electronic device, which includes:

[0014] A processor and a memory, the memory being used to store a computer program, and the processor being used to call and run the computer program stored in the memory to perform the method for extracting aortic dissection membrane provided in the first aspect of this application.

[0015] Fourthly, embodiments of this application provide a computer-readable storage medium for storing a computer program that causes a computer to perform the aortic dissection membrane extraction method as provided in the first aspect of this application.

[0016] Fifthly, embodiments of this application provide a computer program product, including a computer program / instructions, characterized in that, when the computer program / instructions are executed by a processor, they implement the method for extracting aortic dissection membrane as provided in the first aspect of this application.

[0017] This application provides a method, apparatus, device, and storage medium for extracting aortic dissection membrane. First, the corresponding dissection membrane boundary region is determined based on the true lumen region and false lumen region within the aortic image. Then, based on the pixel gradient within the dissection membrane boundary region, seed point pairs are determined within the aortic image. These seed point pairs are used to cluster the pixels within the aortic image to extract the corresponding aortic dissection membrane region. This achieves accurate extraction of the aortic dissection membrane region within the aortic image, ensuring automated and effective identification of the aortic dissection membrane region, enhancing the versatility and efficiency of aortic dissection membrane extraction, and reducing the treatment risks of aortic dissection. Attached Figure Description

[0018] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0019] Figure 1 This is a flowchart illustrating a method for extracting aortic dissection membrane according to an embodiment of this application;

[0020] Figure 2 This is a flowchart illustrating another method for extracting aortic dissection membrane according to an embodiment of this application;

[0021] Figure 3a This is a schematic diagram of a true cavity segmentation mask shown in an embodiment of this application;

[0022] Figure 3b This is a schematic diagram of a pseudo-cavity segmentation mask shown in an embodiment of this application;

[0023] Figure 3c This is a schematic diagram of the interlayer membrane interface mask shown in an embodiment of this application;

[0024] Figure 3d This is a schematic diagram of the interlayer membrane junction region shown in an embodiment of this application;

[0025] Figure 4 This is a schematic diagram illustrating the process of determining seed point pairs within an aortic image, as shown in an embodiment of this application.

[0026] Figure 5 This is a schematic block diagram of an aortic dissection membrane extraction device shown in an embodiment of this application;

[0027] Figure 6 This is a schematic block diagram of an electronic device shown in an embodiment of this application. Detailed Implementation

[0028] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0029] It should be noted that the terms "first," "second," etc., in the specification, claims, and accompanying drawings of this invention are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of the invention described herein can be implemented in orders other than those illustrated or described herein. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion; for example, a process, method, system, product, or server that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or devices.

[0030] This application applies to various aortic images obtained by performing CT angiography (CTA) on the aortic vascular structure of any patient after contrast agent injection using computed tomography (CT). These aortic images can be aortic CTA images.

[0031] Considering the inability to ensure the accuracy of aortic dissection membrane when manually annotating it in aortic images, this application designs an automated extraction scheme for aortic dissection membrane. This scheme enables automated and effective identification of aortic dissection membrane in aortic images, improving the efficiency and stability of aortic dissection membrane extraction while ensuring accurate extraction of the aortic dissection membrane region in aortic images. It also features strong versatility in aortic dissection membrane extraction.

[0032] Figure 1 This is a flowchart illustrating a method for extracting aortic dissection membrane according to an embodiment of this application. (Refer to...) Figure 1 The method may specifically include the following steps:

[0033] S110, determine the corresponding dissecting membrane junction region based on the true lumen region and false lumen region in the aortic image.

[0034] In this application, the aortic image can be obtained by performing a CTA examination on the aortic vascular structure of a patient with aortic dissection using any CT scanning device after the corresponding contrast agent is injected into the aorta.

[0035] It should be understood that the aortic images in this application are generally three-dimensional images, which can be composed of multi-slice aortic images represented by CTA sequences.

[0036] Therefore, for each layer of the aortic image, the corresponding aortic dissection region is extracted from the aortic image, and the aortic dissection region of each layer can then form a three-dimensional aortic dissection membrane.

[0037] For any patient with aortic dissection, after blood enters the aortic media from the tear in the aortic intima, a pathological lumen is added at the tear site, in addition to the original normal blood flow in the aorta. At this time, the chamber where the original normal blood flow in the aorta is located is the true lumen in this application, while the newly added pathological lumen at the aortic tear site is the false lumen in this application.

[0038] Moreover, the true lumen and the false lumen usually form corresponding separation patterns on the aortic wall, and the membrane on the aortic wall used to separate the true lumen and the false lumen is regarded as the corresponding aortic dissection membrane.

[0039] Therefore, in order to accurately extract the aortic dissection membrane within the aortic image, this application first acquires the aortic image of each layer. Then, it performs corresponding chamber segmentation processing on the aortic image of each layer, thereby accurately segmenting the corresponding true lumen region and false lumen region from the aortic image of each layer.

[0040] Then, considering that the true lumen and false lumen within the aorta are typically separated by the aortic dissection membrane, it is understood that the true lumen and false lumen are located on opposite sides of the aortic dissection membrane. Therefore, for the true lumen region and false lumen region within each layer of the aortic image, by analyzing the distribution positions of the true lumen region and false lumen region within that layer of the aortic image and the matching edge points when the cavities are aligned, the gap region formed between the true lumen region and false lumen region can be preliminarily determined within that layer of the aortic image, serving as the dissection membrane boundary region in this application. Following the same method described above, the dissection membrane boundary region within each layer of the aortic image can be determined.

[0041] It should be understood that the aortic dissection membrane boundary region within each aortic image layer in this application can preliminarily represent the aortic dissection membrane within that layer, but it cannot guarantee the accuracy of aortic dissection membrane segmentation. This boundary region may contain not only pixels within the aortic dissection membrane but also some pixels from the true lumen boundary or false lumen boundary connected to the aortic dissection membrane. Therefore, this application requires further optimization of the aortic dissection membrane boundary region within each aortic image layer to accurately segment the corresponding aortic dissection membrane region.

[0042] S120, determine the seed point pairs in the aortic image based on the pixel gradient in the interlaminar region.

[0043] Considering that the aortic dissection membrane junction region contains the various boundary pixels where the aortic dissection membrane connects to the true lumen and false lumen respectively when separating the true lumen and false lumen, then for each boundary pixel, from a specific direction, one side of the boundary pixel mainly consists of pixels within the true lumen or false lumen, while the other side mainly consists of pixels within the aortic dissection membrane. This indicates that each boundary pixel exhibits significant pixel variation when viewed from a specific direction.

[0044] In general, the gradient value of any point in a specific direction can represent how fast that point changes in that specific direction.

[0045] In this application, since the aortic dissection membrane is a closed region, there may be a pixel with the most significant pixel change in any direction. Therefore, after determining the dissection membrane boundary region within each aortic layer image, any reference direction can be pre-selected within this boundary region. Then, the gradient of each pixel within the dissection membrane boundary region is calculated based on the pixel grayscale values ​​of each pixel within that region.

[0046] Furthermore, for the dissecting membrane junction region within each layer of the aortic image, the junction pixel with the most significant pixel change can be identified based on the gradient magnitude of each pixel within this junction region. From the selected reference direction, one side of this junction pixel primarily consists of pixels within the true or false lumen, while the other side primarily consists of pixels within the aortic dissecting membrane. Therefore, according to this reference direction, one pixel can be selected from each side of this junction pixel to form the seed point pair in this application. Thus, one pixel in the seed point pair will be within the true or false lumen, while the other pixel will be within the aortic dissecting membrane.

[0047] Following the same method described above, corresponding seed point pairs can be selected from the dissecting membrane junction region within each layer of aortic image.

[0048] S130: Using seed point pairs, cluster the pixels in the aortic image to obtain the corresponding aortic dissection region.

[0049] Given that the true and false lumens of the aorta are primarily composed of blood, while the aortic dissection membrane is mainly composed of elastic fiber cells, pixels within the true and false lumens regions of an aortic image may exhibit similar pixel features. However, pixels within either the true or false lumens regions will differ from those within the aortic dissection membrane region.

[0050] For each seed point pair in the aortic image, since one pixel in the seed point pair is a pixel in the true lumen or false lumen, while the other pixel is a pixel in the aortic dissection membrane, it means that the two pixels in the seed point pair will have different pixel features.

[0051] Therefore, for each layer of aortic image, this application can employ a corresponding clustering algorithm, using two pixels from the seed point pair within that layer of aortic image as initial cluster centers. The pixel features of each pixel in that layer of aortic image are then compared with the pixel features of the two pixels from the seed point pair to determine the feature differences between each pixel and the two pixels in the seed point pair. Then, pixels with small feature differences within each aortic image are clustered together. Since the two pixels in the seed point pair have different pixel features, based on these two pixels, all pixels can be clustered into two groups: one cluster contains one pixel from the seed point pair, and the other cluster contains the other pixel from the seed point pair.

[0052] Furthermore, since one pixel in a seed point pair is a pixel within the true lumen or false lumen, while the other pixel is a pixel within the aortic dissection membrane, the two regions formed by the two clusters obtained after clustering the pixels in each aortic image layer can be either a chamber region composed of the true lumen and false lumen within that aortic image layer, or a region of the aortic dissection membrane within that aortic image layer.

[0053] Therefore, based on the distinguishing features between pixels within the true and false lumens and pixels within the aortic dissection membrane, the corresponding aortic dissection membrane region can be selected from these two regions by analyzing the pixel features within the two regions formed by the aforementioned two clusters. Following the same method, the aortic dissection membrane region within each layer of the aortic image can be determined. Then, by combining the aortic dissection membrane regions within each layer of the aortic image according to their order, the three-dimensional aortic dissection membrane region can be determined.

[0054] The technical solution provided in this application first determines the corresponding aortic dissection boundary region based on the true lumen region and the false lumen region within the aortic image. Then, based on the pixel gradient within the aortic dissection boundary region, seed point pairs are determined within the aortic image. These seed point pairs are then used to cluster the pixels within the aortic image to extract the corresponding aortic dissection region. This achieves accurate extraction of the aortic dissection region within the aortic image, ensuring automated and effective identification of the aortic dissection region, enhancing the versatility and efficiency of aortic dissection extraction, and reducing the treatment risk of aortic dissection.

[0055] As an optional implementation scheme in the embodiments of this application, in order to ensure the accuracy of aortic dissection membrane region extraction, this application can provide a detailed description of the pixel gradient determination method and the specific clustering method used in the aortic dissection membrane region extraction process, which involves the true lumen region, false lumen region and dissection membrane boundary region.

[0056] Figure 2 This is a flowchart illustrating another method for extracting aortic dissection membrane, as shown in an embodiment of this application.

[0057] like Figure 2 As shown, the method may specifically include the following steps:

[0058] S210 inputs the aortic image into a pre-built chamber segmentation model to predict the true lumen segmentation mask and the false lumen segmentation mask of the aortic image.

[0059] Specifically, the process begins by acquiring images of each layer of the aorta within the aortic imaging system and determining the grayscale value of each pixel within each layer.

[0060] For example, each layer of aortic image within an aortic imaging system can be represented as:

[0061]

[0062] Among them, a mn This represents the grayscale value of the pixel in the m-th row and n-th column of each aortic image layer.

[0063] Furthermore, to ensure the accuracy of chamber segmentation in each layer of aortic images within the aortic imaging, this application can pre-acquire a large number of aortic images as training samples, and label the corresponding true lumen region and false lumen region in each aortic image as corresponding sample labels. Then, using appropriate machine learning methods, and utilizing the aforementioned large number of training samples and corresponding sample labels, a chamber segmentation model capable of accurately segmenting the mask image represented by the true lumen region and the false lumen region is pre-trained.

[0064] In this application, each layer of aortic image within an aortic image can be input into a pre-constructed chamber segmentation model. The trained mesh parameters and network structure within the chamber segmentation model are used to perform corresponding chamber segmentation on each layer of aortic image, thereby predicting the true lumen segmentation mask corresponding to the true lumen region and the false lumen segmentation mask corresponding to the false lumen region within each layer of aortic image.

[0065] For example, such as Figure 3a As shown, the true cavity segmentation mask in this application can be represented as:

[0066]

[0067] Among them, b mn This represents the mask value indicating whether the pixel in the m-th row and n-th column of each aortic image layer is a pixel within the true lumen region. If a pixel is within the true lumen region, then the mask value for that pixel is 1; otherwise, it is 0.

[0068] like Figure 3b As shown, the pseudo-cavity segmentation mask in this application can be represented as:

[0069]

[0070] Among them, c mn This represents the mask value indicating whether the pixel in the m-th row and n-th column of each aortic image layer is a pixel within the false lumen region. To distinguish it from the true lumen segmentation mask, if a pixel is within the false lumen region, then the mask value of that pixel is 2; otherwise, it is 0.

[0071] S220, Based on the expansion results of the true cavity segmentation mask and the false cavity segmentation mask, determine the corresponding interlayer membrane boundary mask.

[0072] Considering that the aortic dissection membrane is used to separate the true lumen and the false lumen, primarily forming the gap between them, the pixels within the aortic dissection membrane that come into contact with the true lumen and the false lumen can be merged into the true lumen segmentation mask and the false lumen segmentation mask respectively by expanding the boundary points of the true lumen within the true lumen segmentation mask and the false lumen segmentation mask accordingly. Then, by finding the intersection of the mask values ​​of each pixel in the expanded true lumen segmentation mask and the expanded false lumen segmentation mask, the pixels with non-zero mask values ​​in both masks are identified, thus obtaining the corresponding dissection membrane boundary mask.

[0073] As an optional implementation scheme in this application, the interlayer membrane interface mask can be determined by the following steps:

[0074] The first step is to expand the true cavity segmentation mask and the false cavity segmentation mask respectively to obtain the expanded true cavity segmentation mask and the expanded false cavity segmentation mask.

[0075] This application can use corresponding expansion elements to expand the true lumen segmentation mask and the false lumen segmentation mask respectively, so that the boundaries of the true lumen segmentation mask and the false lumen segmentation mask expand outward, and the aortic dissection membrane formed between the true lumen and the false lumen is merged into the true lumen segmentation mask and the false lumen segmentation mask respectively, thereby obtaining the expanded true lumen segmentation mask and the expanded false lumen segmentation mask.

[0076] The second step is to determine the corresponding interlayer membrane boundary mask based on the overlap between the expanded true cavity segmentation mask and the expanded false cavity segmentation mask.

[0077] By analyzing the mask values ​​of each pixel within the expanded true lumen segmentation mask and the expanded false lumen segmentation mask, we can identify pixels with non-zero mask values ​​in both masks. These pixels belong to both the expanded true lumen and the expanded false lumen regions, representing overlapping pixels within both. Since there are no overlapping pixels between the true and false lumens, these overlapping pixels are likely pixels merged from the aortic dissection membrane after the expansion of the true and false lumen segmentation masks. Therefore, these overlapping pixels can be considered as pixels within the aortic dissection membrane to determine the corresponding dissection membrane boundary mask.

[0078] For example, suppose the expanded true cavity segmentation mask can be represented as a Mask Aorta-true-p The expanded pseudo-cavity segmentation mask can be represented as Mask Aorta-false-p Among them, Mask Aorta-true-p The mask value is 1 for pixels belonging to the true cavity region, and 0 otherwise. The Mask... Aorta-false-p The mask value is 2 for pixels belonging to the false cavity region, and 0 otherwise.

[0079] Therefore, by adding the expanded true cavity segmentation mask and the expanded false cavity segmentation mask, a new mask can be obtained, which can be represented as Mask. Aorta-p =Mask Aorta-true-p +Mask Aorta-false-p At this point, the mask... Aorta-p The mask value of each pixel can be one of four values: 0, 1, 2, or 3.

[0080] Among them, the mask Aorta-p A pixel with a mask value of 3 can represent a pixel within the aortic dissection membrane. Therefore, it can be determined from this mask. Aorta-p Pixels with a mask value of 3 are selected, and their mask values ​​are changed to 1. The mask values ​​of other pixels are changed to 0. This yields the interlayer membrane boundary mask in this application, which can be represented as Mask. Aorta-mo The interlayer membrane interface mask in this application can be as follows: Figure 3c As shown.

[0081] S230, Based on the dissecting membrane junction mask, determine the corresponding dissecting membrane junction area from the aortic image.

[0082] After determining the interaortic membrane boundary mask for each layer of the aortic image, the interaortic membrane boundary mask can be added to that layer of the aortic image to correct the grayscale value of each pixel in that layer of the aortic image, thereby determining the interaortic membrane boundary region in that layer of the aortic image.

[0083] For example, suppose an image of a certain aortic layer is represented as Image, and the interstitial membrane junction mask of that aortic layer is represented as Mask. Aorta-mo So, if Figure 3d As shown, the dissecting membrane junction region within this layer of the aortic image can be represented as...

[0084] Among them, f mn This represents the grayscale value of the pixel in the m-th row and n-th column of the aortic image. If a pixel belongs to the interface region of the dissecting membrane, then the grayscale value of that pixel is the original grayscale value; otherwise, it is 0.

[0085] S240: Based on the centerline normal vector within the interlayer membrane interface region, determine the gradient of each pixel within the interlayer membrane interface region to obtain the target pixel with the largest gradient.

[0086] After identifying the dissecting membrane junction region within each layer of aortic imaging, in order to accurately analyze whether there are significant changes in each pixel within the dissecting membrane junction region, such as... Figure 4 As shown, this application can first obtain the center line Aorta-center in the interlayer membrane boundary region, and then obtain the corresponding center line normal vector Aorta-fa in the interlayer membrane boundary region with the center line Aorta-center as the reference.

[0087] Furthermore, for the dissecting membrane junction region within each aortic layer image, the vector direction of the centerline normal vector Aorta-fa can be used as the reference direction. A corresponding gradient algorithm is then employed to calculate the gradient value of each pixel within the dissecting membrane junction region along this reference direction. Finally, the target pixel with the largest gradient is selected from all pixels within the dissecting membrane junction region.

[0088] S250 determines the adjacent pixels of the target pixel in the aortic image with the centerline normal vector as the reference direction, and uses them as seed point pairs in the aortic image.

[0089] For a target pixel within the dissecting membrane junction region of each aortic layer in the aortic image, the position coordinates of the target pixel within that aortic layer can be determined, thereby locating the target pixel within that aortic layer.

[0090] Then, within each aortic image layer of the aortic imaging, the vector direction of the centerline normal vector Aorta-fa within that aortic image layer can be determined. Furthermore, as... Figure 4 As shown, the vector direction of the centerline normal vector Aorta-fa is used as the reference direction to determine the two adjacent pixels of the target pixel in the aortic image of this layer in the reference direction.

[0091] Since the target pixel represents a boundary point of the aortic dissection membrane within the dissection membrane junction region, its adjacent pixel in the reference direction will be a pixel within the true or false lumen, while the other adjacent pixel will be a pixel within the aortic dissection membrane. Therefore, this application can use the two adjacent pixels of the target pixel in the reference direction within each aortic image layer as seed point pairs within that aortic image layer.

[0092] Following the same method described above, seed point pairs within each layer of aortic images can be determined.

[0093] S260, taking the first and second seed points in the seed point pair as the initial cluster centers, and clustering the pixels in the aortic image according to the tissue density of each pixel in the aortic image, to obtain the first cluster region and the second cluster region after clustering.

[0094] After identifying the seed point pairs within each layer of the aortic image, for each seed point pair, one pixel in the pair can be designated as the first seed point, and the other pixel as the second seed point. Furthermore, the first and second seed points belong to different regions within the cavity region composed of the true lumen and the false lumen, and the aortic dissection membrane region, respectively.

[0095] Furthermore, considering that the hemochromatographic tissue density (i.e., HU value) of the chamber region composed of the true lumen and the false lumen is usually much higher than the tissue density (i.e., HU value) of the aortic dissection membrane region, this application can determine the tissue density HU value of each pixel in each aortic image layer, including the tissue density HU values ​​of the first seed point and the second seed point in the aortic image layer.

[0096] Then, this application can use first and second seed points in different regions belonging to the chamber region and the aortic dissection membrane region, respectively, as initial cluster centers. Then, by comparing the tissue density HU value of each pixel in the aortic image with the tissue density HU values ​​of the two cluster centers, the difference in tissue density HU value between each pixel and the two cluster centers is determined. Furthermore, pixels with smaller differences in tissue density HU values ​​are clustered together to obtain the first cluster corresponding to the first seed point and the second cluster corresponding to the second seed point.

[0097] For example, this application can use the k-means clustering algorithm to cluster each pixel in each layer of the aortic image. The specific clustering process can be as follows:

[0098] In each layer of the aortic image, the first and second seed points of the seed point pair are used as the initial cluster centers, and k in the k-means clustering algorithm can be set to 2. Then, among all pixels in the aortic image layer, iteratively search for a cluster with a small difference in tissue density HU value from the first seed point and another cluster with a small difference in tissue density HU value from the second seed point, minimizing the loss function corresponding to the clustering result, thereby dividing all pixels in the aortic image layer into two clusters.

[0099] In this application, the loss function in the k-means clustering algorithm can be:

[0100] Where, x i c represents the tissue density HU value of the i-th pixel in each layer of aortic image. i x represents i The cluster to which the represented pixel belongs. c i The cluster represents the center point of the cluster, and M is the number of pixels in each layer of the aortic image.

[0101] S270, based on the pixel density in the first cluster region and the second cluster region, determine one of the first cluster region and the second cluster region as the aortic dissection membrane region.

[0102] For each layer of aortic imagery, after determining the first and second cluster regions after clustering the pixels within that layer, since the first and second seed points belong to different regions within the chamber region composed of the true lumen and the false lumen, and the aortic dissection region, respectively, it can be concluded that one of the first and second cluster regions can be the chamber region composed of the true lumen and the false lumen, while the other region is the aortic dissection region. Furthermore, the hemochromatographic tissue density (HU) value within the chamber region composed of the true lumen and the false lumen is typically much higher than the tissue density (HU) value within the aortic dissection region.

[0103] Therefore, this application can select one of the first and second cluster regions as the aortic dissection membrane region by analyzing the size of the pixel tissue density HU value in the first and second cluster regions.

[0104] As an optional implementation of this application, the method for determining the aortic dissection membrane region from the first cluster region and the second cluster region can be as follows: for any cluster region in the first cluster region and the second cluster region, the average tissue density in the cluster region is determined according to the tissue density of each pixel in the cluster region; the cluster region with the smaller average tissue density in the first cluster region and the second cluster region is taken as the aortic dissection membrane region in the aortic image.

[0105] In other words, for any cluster region within the first and second cluster regions of each aortic image layer, the tissue density HU value of each pixel within that cluster region can be obtained. The average tissue density HU value of each pixel within that cluster region is then calculated to obtain the average tissue density HU value for that cluster region. Since the tissue density HU value of the blood angiography within the chamber region composed of the true lumen and the false lumen is usually much higher than that within the aortic dissection membrane region, the cluster region with the smaller average tissue density HU value can be selected from the first and second cluster regions as the aortic dissection membrane region within that aortic image layer. Following the same method, the aortic dissection membrane region within each aortic image layer can be determined. Then, by combining the aortic dissection membrane regions within each aortic image layer according to the order of the aortic images, the three-dimensional aortic dissection membrane region can be determined.

[0106] The technical solution provided in this application first determines the corresponding aortic dissection boundary region based on the true lumen region and the false lumen region within the aortic image. Then, based on the pixel gradient within the aortic dissection boundary region, seed point pairs are determined within the aortic image. These seed point pairs are then used to cluster the pixels within the aortic image to extract the corresponding aortic dissection region. This achieves accurate extraction of the aortic dissection region within the aortic image, ensuring automated and effective identification of the aortic dissection region, enhancing the versatility and efficiency of aortic dissection extraction, and reducing the treatment risk of aortic dissection.

[0107] Figure 5 This is a schematic block diagram of an aortic dissection membrane extraction device shown in an embodiment of this application.

[0108] like Figure 5 As shown, the device 500 may include:

[0109] The junction region determination module 510 is used to determine the corresponding dissecting membrane junction region based on the true lumen region and the false lumen region in the aortic image.

[0110] Seed point pair determination module 520 is used to determine seed point pairs in the aortic image based on the pixel gradient in the interlaminar membrane junction area;

[0111] The aortic dissection membrane extraction module 530 is used to cluster the pixels in the aortic image using the seed point pairs to obtain the corresponding aortic dissection membrane region.

[0112] In some implementations, the boundary region determination module 510 may include:

[0113] The true and false lumen segmentation unit is used to input the aortic image into a pre-constructed chamber segmentation model and predict the true lumen segmentation mask and false lumen segmentation mask of the aortic image;

[0114] The boundary mask determination unit is used to determine the corresponding interlayer membrane boundary mask based on the expansion results of the true cavity segmentation mask and the false cavity segmentation mask.

[0115] The junction region determination unit is used to determine the corresponding junction region of the dissecting membrane from the aortic image based on the junction mask of the dissecting membrane.

[0116] In some implementations, the boundary mask determining unit can be specifically used for:

[0117] The true cavity segmentation mask and the false cavity segmentation mask are respectively subjected to expansion processing to obtain an expanded true cavity segmentation mask and an expanded false cavity segmentation mask;

[0118] Based on the overlap between the expanded true cavity segmentation mask and the expanded false cavity segmentation mask, the corresponding interlayer membrane boundary mask is determined.

[0119] In some implementations, the seed point pair determination module 520 can be specifically used for:

[0120] Based on the centerline normal vector within the interlayer membrane interface region, the gradient of each pixel within the interlayer membrane interface region is determined to obtain the target pixel with the largest gradient.

[0121] Within the aortic image, using the centerline normal vector as the reference direction, the adjacent pixels of the target pixel are determined as seed point pairs within the aortic image.

[0122] In some implementations, the interlayer membrane extraction module 530 may include:

[0123] A clustering unit is used to cluster the pixels in the aortic image based on the tissue density of each pixel in the aortic image, with the first seed point and the second seed point in the seed point pair as the initial clustering centers, respectively, to obtain the first cluster region and the second cluster region after clustering.

[0124] The dissecting membrane extraction unit is used to determine, based on the pixel tissue density in the first cluster region and the second cluster region, one of the first cluster region and the second cluster region as the aortic dissecting membrane region.

[0125] In some implementations, the sandwich membrane extraction unit can be specifically used for:

[0126] For any one of the first cluster region and the second cluster region, the average tissue density within the cluster region is determined based on the tissue density of each pixel within that cluster region.

[0127] The cluster region with the lower average tissue density between the first cluster region and the second cluster region is taken as the aortic dissection region in the aortic image.

[0128] In this embodiment, the corresponding aortic dissection boundary region is first determined based on the true lumen region and the false lumen region within the aortic image. Then, seed point pairs within the aortic image are determined based on the pixel gradient within the aortic dissection boundary region. These seed point pairs are used to cluster the pixels within the aortic image to extract the corresponding aortic dissection region. This achieves accurate extraction of the aortic dissection region within the aortic image, ensuring automated and effective identification of the aortic dissection region, enhancing the versatility and efficiency of aortic dissection extraction, and reducing the treatment risk of aortic dissection.

[0129] It should be understood that the device embodiments and method embodiments can correspond to each other, and similar descriptions can be referred to the method embodiments. To avoid repetition, further details will not be provided here. Specifically, Figure 5 The apparatus 500 shown can execute any of the method embodiments provided in this application, and the foregoing and other operations and / or functions of each module in the apparatus 500 are respectively for implementing the corresponding processes in the various methods of the embodiments of this application. For the sake of brevity, they will not be described in detail here.

[0130] The apparatus 500 of this application embodiment has been described above from the perspective of functional modules in conjunction with the accompanying drawings. It should be understood that this functional module can be implemented in hardware, in software instructions, or in a combination of hardware and software modules. Specifically, the steps of the method embodiments in this application can be completed by integrated logic circuits in the processor's hardware and / or by software instructions. The steps of the method disclosed in this application embodiment can be directly embodied as being executed by a hardware decoding processor, or by a combination of hardware and software modules in the decoding processor. Optionally, the software module can reside in a mature storage medium in the art, such as random access memory, flash memory, read-only memory, programmable read-only memory, electrically erasable programmable memory, registers, etc. This storage medium is located in memory, and the processor reads information from the memory and, in conjunction with its hardware, completes the steps in the above method embodiments.

[0131] Figure 6 This is a schematic block diagram of an electronic device 600 shown in an embodiment of this application.

[0132] like Figure 6 As shown, the electronic device 600 may include:

[0133] The system includes a memory 610 and a processor 620. The memory 610 stores computer programs and transfers the program code to the processor 620. In other words, the processor 620 can retrieve and run the computer program from the memory 610 to implement the methods described in the embodiments of this application.

[0134] For example, the processor 620 can be used to execute the above-described method embodiments according to instructions in the computer program.

[0135] In some embodiments of this application, the processor 620 may include, but is not limited to:

[0136] General-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc.

[0137] In some embodiments of this application, the memory 610 includes, but is not limited to:

[0138] Volatile memory and / or non-volatile memory. Non-volatile memory can be read-only memory (ROM), programmable read-only memory (PROM), erasable programmable read-only memory (EPROM), electrically erasable programmable read-only memory (EEPROM), or flash memory. Volatile memory can be random access memory (RAM), which is used as an external cache. By way of example, but not limitation, many forms of RAM are available, such as Static RAM (SRAM), Dynamic RAM (DRAM), Synchronous DRAM (SDRAM), Double Data Rate SDRAM (DDR SDRAM), Enhanced Synchronous DRAM (ESDRAM), Synchronous Link DRAM (SLDRAM), and Direct Rambus RAM (DR RAM).

[0139] In some embodiments of this application, the computer program may be divided into one or more modules, which are stored in the memory 610 and executed by the processor 620 to perform the method provided in this application. The one or more modules may be a series of computer program instruction segments capable of performing a specific function, which describe the execution process of the computer program in the electronic device.

[0140] like Figure 6 As shown, the electronic device may also include:

[0141] Transceiver 630, which can be connected to processor 620 or memory 610.

[0142] The processor 620 can control the transceiver 630 to communicate with other devices; specifically, it can send information or data to other devices or receive information or data sent by other devices. The transceiver 630 may include a transmitter and a receiver. The transceiver 630 may further include antennas, and the number of antennas may be one or more.

[0143] It should be understood that the various components in the electronic device are connected through a bus system, which includes a data bus, a power bus, a control bus, and a status signal bus.

[0144] This application also provides a computer storage medium storing a computer program thereon, which, when executed by a computer, enables the computer to perform the methods of the above-described method embodiments. Alternatively, this application also provides a computer program product containing instructions that, when executed by a computer, cause the computer to perform the methods of the above-described method embodiments.

[0145] When implemented using software, it can be implemented entirely or partially as a computer program product. This computer program product includes one or more computer instructions. When these computer program instructions are loaded and executed on a computer, all or part of the processes or functions described in the embodiments of this application are generated. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another. For example, the computer instructions can be transmitted from one website, computer, server, or data center to another via wired (e.g., coaxial cable, fiber optic, digital subscriber line (DSL)) or wireless (e.g., infrared, wireless, microwave, etc.) means. The computer-readable storage medium can be any available medium that a computer can access or a data storage device such as a server or data center that integrates one or more available media. The available medium can be a magnetic medium (e.g., floppy disk, hard disk, magnetic tape), an optical medium (e.g., digital video disc (DVD)), or a semiconductor medium (e.g., solid-state disk (SSD)).

[0146] Those skilled in the art will recognize that the modules and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0147] In the several embodiments provided in this application, it should be understood that the disclosed systems, apparatuses, and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of modules is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple modules or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between apparatuses or modules may be electrical, mechanical, or other forms.

[0148] The modules described as separate components may or may not be physically separate. The components shown as modules may or may not be physical modules; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs. For example, the functional modules in the various embodiments of this application may be integrated into one processing module, or each module may exist physically separately, or two or more modules may be integrated into one module.

[0149] The above are merely specific embodiments of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the technical scope disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

Claims

1. A method for extracting an aortic dissection membrane, characterized by, The method comprises the steps of: determining a corresponding dissection membrane boundary region according to a true lumen region and a false lumen region in the aortic image; determining a seed point pair in the aortic image according to a pixel point gradient in the dissection membrane boundary region; performing clustering on pixel points in the aortic image by using the seed point pair to obtain a corresponding aortic dissection membrane region; wherein the step of determining the corresponding dissection membrane boundary region according to the true lumen region and the false lumen region in the aortic image comprises the steps of: inputting the aortic image into a pre-constructed lumen segmentation model to predict a true lumen segmentation mask and a false lumen segmentation mask of the aortic image; determining a corresponding dissection membrane boundary mask according to an inflation result of the true lumen segmentation mask and the false lumen segmentation mask; determining a corresponding dissection membrane boundary region in the aortic image according to the dissection membrane boundary mask; the step of determining the seed point pair in the aortic image according to the pixel point gradient in the dissection membrane boundary region comprises the steps of: determining a gradient of each pixel point in the dissection membrane boundary region according to a center line normal vector in the dissection membrane boundary region to obtain a target pixel point with the maximum gradient; determining a neighboring pixel point of the target pixel point as a seed point pair in the aortic image according to the center line normal vector as a reference direction in the aortic image; the step of performing clustering on the pixel points in the aortic image by using the seed point pair to obtain the corresponding aortic dissection membrane region comprises the steps of: respectively taking a first seed point and a second seed point in the seed point pair as initial clustering centers, and performing clustering on the pixel points in the aortic image according to tissue densities of the pixel points in the aortic image to obtain a first cluster region and a second cluster region after clustering; determining one of the first cluster region and the second cluster region as the aortic dissection membrane region according to the tissue densities of the pixel points in the first cluster region and the second cluster region.

2. The method of claim 1, wherein, the step of determining the corresponding dissection membrane boundary mask according to the inflation result of the true lumen segmentation mask and the false lumen segmentation mask comprises the steps of: respectively performing inflation processing on the true lumen segmentation mask and the false lumen segmentation mask to obtain an inflated true lumen segmentation mask and an inflated false lumen segmentation mask; determining the corresponding dissection membrane boundary mask according to a coincidence degree of the inflated true lumen segmentation mask and the inflated false lumen segmentation mask.

3. The method of claim 1, wherein, the step of determining one of the first cluster region and the second cluster region as the aortic dissection membrane region according to the tissue densities of the pixel points in the first cluster region and the second cluster region comprises the steps of: for any one of the first cluster region and the second cluster region, determining an average tissue density in the cluster region according to tissue densities of pixel points in the cluster region; taking a cluster region with a smaller average tissue density from the first cluster region and the second cluster region as the aortic dissection membrane region in the aortic image.

4. An apparatus for extracting an aortic dissection membrane, characterized by, The method comprises the steps of: a boundary region determination module configured to determine a corresponding dissection membrane boundary region according to a true lumen region and a false lumen region in the aortic image; A seed point pair determination module is configured to determine a seed point pair in the aortic image according to gradients of pixel points in the interface region of the dissection membrane. A dissection membrane extraction module is configured to cluster the pixel points in the aortic image by using the seed point pair to obtain a corresponding aortic dissection membrane region. The interface region determination module is specifically configured to: input the aortic image into a pre-constructed lumen segmentation model to predict a true lumen segmentation mask and a false lumen segmentation mask of the aortic image; determine a corresponding dissection membrane interface mask according to the bulging results of the true lumen segmentation mask and the false lumen segmentation mask; determine a corresponding dissection membrane interface region in the aortic image according to the dissection membrane interface mask; The seed point pair determination module is specifically configured to: determine the gradient of each pixel point in the dissection membrane interface region according to a center line normal vector in the dissection membrane interface region to obtain a target pixel point with the maximum gradient; determine a neighboring pixel point of the target pixel point as a seed point pair in the aortic image according to the center line normal vector as a reference direction in the aortic image. The dissection membrane extraction module is specifically configured to: cluster the pixel points in the aortic image according to the tissue density of each pixel point in the aortic image by taking the first seed point and the second seed point in the seed point pair as initial clustering centers to obtain a first cluster region and a second cluster region after clustering; determine one of the first cluster region and the second cluster region as an aortic dissection membrane region according to the pixel point tissue density in the first cluster region and the second cluster region.

5. An electronic device, comprising: The processor and the memory are configured to store a computer program and call and run the computer program stored in the memory to execute the aortic dissection membrane extraction method in any one of claims 1-3. The computer program is configured to enable a computer to execute the aortic dissection membrane extraction method in any one of claims 1-3.

6. A computer-readable storage medium, characterized in that, The computer program / instructions enable the processor to execute the aortic dissection membrane extraction method in any one of claims 1-3.

7. A computer program product comprising computer programs / instructions, characterized in that, ​

Images