Break detection method, device, equipment, storage medium and program product

Abstract

The application relates to a crack detection method, device, equipment, storage medium and program product. The method comprises the following steps: determining a first mask image of intermediate tissue according to an acquired original image; the first mask image comprises the intermediate tissue and an initial crack area on the intermediate tissue; completing the initial crack area on the intermediate tissue to determine a second mask image; the second mask image comprises complete intermediate tissue; processing the second mask image according to the original image to determine at least one candidate crack area; and determining a target crack area according to the at least one candidate crack area. The method can improve the accuracy of the crack detection result.

Description

Technical Field

[0001] This application relates to the field of image technology, and in particular to a method, apparatus, device, storage medium, and program product for detecting breaches. Background Technology

[0002] Aortic dissection refers to the process by which blood from within the aortic lumen enters the aortic media through a tear in the aortic intima, causing the media to separate and extend along the long axis of the aorta, creating a true and false lumen separation within the aortic wall. Blood can flow between these two lumens. Detecting the tear in aortic dissection provides a valuable data foundation for further image analysis.

[0003] Currently, when detecting ruptures in aortic dissections, the image values ​​of the true and false lumens can be compared to distinguish them, and the connecting part between the true and false lumens can be used as the rupture point to achieve rupture detection.

[0004] However, due to the complexity of breaches in real-world scenarios, the aforementioned techniques suffer from inaccurate breach detection results. Summary of the Invention

[0005] Therefore, it is necessary to provide a method, apparatus, equipment, storage medium, and program product for detecting breaches that can improve the accuracy of breach detection results, in response to the above-mentioned technical problems.

[0006] Firstly, this application provides a method for detecting breaches, the method comprising:

[0007] A first mask image of the intermediate tissue is determined based on the acquired original image; the first mask image includes the intermediate tissue and the initial tear region on the intermediate tissue;

[0008] The initial tear area on the intermediate tissue is filled in to determine the second mask image; the second mask image includes the complete intermediate tissue.

[0009] The second mask image is processed based on the original image to determine at least one candidate breach region;

[0010] The target breach region is determined based on at least one candidate breach region.

[0011] In one embodiment, determining the target breach region based on at least one candidate breach region includes:

[0012] At least one candidate breach region is classified using a pre-defined classification network to determine the target breach region.

[0013] The above classification network is trained based on multiple sample mask images. Each sample mask image includes a sample breakage region and a standard category of the sample breakage region. The standard category is used to characterize whether the sample breakage region is a target breakage region.

[0014] In one embodiment, the above-described process of completing the initial tear area on the intermediate tissue to determine the second mask image includes:

[0015] Based on the position information of each point on the intermediate tissue in the first mask image, surface reconstruction processing is performed on the intermediate tissue and the initial tear area to determine the closed surface corresponding to the intermediate tissue.

[0016] The second mask image is determined based on the closed surface and the first mask image.

[0017] In one embodiment, the above-mentioned surface reconstruction processing of the intermediate tissue and the initial tear area based on the position information of each point on the intermediate tissue in the first mask image to determine the closed surface corresponding to the intermediate tissue includes:

[0018] Based on the position information of each point on the intermediate tissue in the first mask image, the point cloud sparse sampling process is performed on the intermediate tissue and the initial tear area to determine the point cloud data corresponding to the intermediate tissue.

[0019] The point cloud data is reconstructed using surface reconstruction to determine the closed surface corresponding to the intermediate tissue.

[0020] In one embodiment, the above-mentioned surface reconstruction processing of point cloud data to determine the closed surface corresponding to the intermediate tissue includes:

[0021] Calculate the normal vector direction of each point in the point cloud data, and adjust the normal vector direction of each point to the same direction;

[0022] The point cloud data after adjusting the normal vector direction is processed by surface reconstruction to determine the closed surface corresponding to the intermediate tissue.

[0023] In one embodiment, the above-described processing of the second mask image based on the original image to determine at least one candidate breach region includes:

[0024] Based on the original image, the second mask image is processed by image block cropping to determine at least one candidate breach region.

[0025] In one embodiment, the above-mentioned image block cropping process of the second mask image based on the original image to determine at least one candidate breach region includes:

[0026] Based on the original image, determine the image values ​​corresponding to each point on the complete intermediate tissue in the original image;

[0027] Based on the image values ​​of each point, determine at least one candidate point among them;

[0028] Image block cropping is performed at the location of at least one candidate point in the second mask image to determine at least one candidate breach region.

[0029] In one embodiment, determining the first mask image of the intermediate tissue based on the acquired original image includes:

[0030] A mask image of the tissue to be segmented is determined based on the acquired original image; wherein the tissue to be segmented includes intermediate tissue.

[0031] The first mask image of the intermediate tissue is determined based on the original image and the mask image of the tissue to be segmented.

[0032] In one embodiment, the intermediate tissue is an intima sheet or a blood vessel.

[0033] Secondly, this application also provides a breakage detection device, which includes:

[0034] The first mask determination module is used to determine a first mask image of the intermediate tissue based on the acquired original image; the first mask image includes the intermediate tissue and the initial tear region on the intermediate tissue;

[0035] The second mask determination module is used to complete the initial tear area on the intermediate tissue and determine the second mask image; the second mask image includes the complete intermediate tissue;

[0036] The processing module is used to process the second mask image based on the original image to determine at least one candidate breach region;

[0037] The detection module is used to determine the target breach area based on at least one candidate breach area.

[0038] Thirdly, this application also provides a computer device, which includes a memory and a processor. The memory stores a computer program, and the processor executes the computer program to perform the following steps:

[0039] A first mask image of the intermediate tissue is determined based on the acquired original image; the first mask image includes the intermediate tissue and the initial tear region on the intermediate tissue;

[0040] The initial tear area on the intermediate tissue is filled in to determine the second mask image; the second mask image includes the complete intermediate tissue.

[0041] The second mask image is processed based on the original image to determine at least one candidate breach region;

[0042] The target breach region is determined based on at least one candidate breach region.

[0043] Fourthly, this application also provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, performs the following steps:

[0044] A first mask image of the intermediate tissue is determined based on the acquired original image; the first mask image includes the intermediate tissue and the initial tear region on the intermediate tissue;

[0045] The initial tear area on the intermediate tissue is filled in to determine the second mask image; the second mask image includes the complete intermediate tissue.

[0046] The second mask image is processed based on the original image to determine at least one candidate breach region;

[0047] The target breach region is determined based on at least one candidate breach region.

[0048] Fifthly, this application also provides a computer program product, comprising a computer program that, when executed by a processor, performs the following steps:

[0049] A first mask image of the intermediate tissue is determined based on the acquired original image; the first mask image includes the intermediate tissue and the initial tear region on the intermediate tissue;

[0050] The initial tear area on the intermediate tissue is filled in to determine the second mask image; the second mask image includes the complete intermediate tissue.

[0051] The second mask image is processed based on the original image to determine at least one candidate breach region;

[0052] The target breach region is determined based on at least one candidate breach region.

[0053] The aforementioned method, apparatus, device, storage medium, and program product for detecting breaches determine a first mask image of intermediate tissue from the original image, completes the initial breach area on the intermediate tissue to determine a second mask image of the complete intermediate tissue, then processes the second mask image using the original image to obtain at least one candidate breach area, and finally determines the target breach area based on the candidate breach areas. In this method, because the initial breach area of ​​the intermediate tissue can be completed first, and then candidate breach areas are determined by combining the original image with the mask image of the completed intermediate tissue, and finally the target breach area is determined based on the candidate breach areas, this avoids the problem of low accuracy in detecting breach areas when there are non-real breach areas in the initial breach area. By completing the intermediate tissue and further detecting the breach areas, this problem can be avoided, thereby improving the accuracy of the finally detected breach area. Attached Figure Description

[0054] Figure 1 Example diagram of existing aortic dissection;

[0055] Figure 2 This is an internal structural diagram of a computer device in one embodiment;

[0056] Figure 3 This is a flowchart illustrating a breach detection method in one embodiment;

[0057] Figure 4 This is a flowchart illustrating the breach detection method in another embodiment;

[0058] Figure 5 This is a flowchart illustrating the breach detection method in another embodiment;

[0059] Figure 6 Example images of intermediate tissue before and after sparse sampling in another embodiment;

[0060] Figure 7 Example diagrams showing the intermediate tissue normal vector direction before and after adjustment in another embodiment;

[0061] Figure 8 Example diagrams before and after intermediate tissue reconstruction in another embodiment;

[0062] Figure 9 This is a flowchart illustrating the breach detection method in another embodiment;

[0063] Figure 10 This is a flowchart illustrating the breach detection method in another embodiment;

[0064] Figure 11 This is a structural block diagram of a breach detection device in one embodiment. Detailed Implementation

[0065] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application.

[0066] Normally, after a tear in the aortic intima, blood flow can enter the aortic wall through the rupture, dividing the original single-lumen structure of the aorta into a true lumen (TL) and a false lumen (FL). See also Figure 1 As shown, the true and false lumens may be connected or disconnected. The connection between the true and false lumens is the rupture point (i.e., the selected area in the figure), and a further rupture may exist distally. When the true and false lumens are not connected, they are separated by the aortic intima. Currently, when detecting ruptures in aortic dissections, the image values ​​of the true and false lumens can be compared to distinguish them, and the connecting part between the true and false lumens is taken as the rupture point for detection. However, due to the complexity of rupture situations in real-world scenarios, the above-mentioned techniques suffer from inaccurate rupture detection results. Therefore, this application provides a rupture detection method, apparatus, device, storage medium, and program product to solve the above-mentioned technical problems.

[0067] The tear detection method provided in this application can be applied to computer equipment, which can be a terminal or a server. Taking a terminal as an example, its internal structure diagram can be as follows: Figure 2 As shown, the computer device includes a processor, memory, communication interface, display screen, and input devices connected via a system bus. The processor provides computing and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system and computer programs. The internal memory provides an environment for the operation of the operating system and computer programs stored in the non-volatile storage media. The communication interface is used for wired or wireless communication with external terminals; wireless communication can be achieved through Wi-Fi, mobile cellular networks, NFC (Near Field Communication), or other technologies. When executed by the processor, the computer program implements a breach detection method. The display screen can be an LCD screen or an e-ink screen. The input devices can be a touch layer covering the display screen, buttons, a trackball, or a touchpad on the computer device's casing, or an external keyboard, touchpad, or mouse.

[0068] Those skilled in the art will understand that Figure 2The structure shown is merely a block diagram of a portion of the structure related to the present application and does not constitute a limitation on the computer device to which the present application is applied. Specific computer devices may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements.

[0069] In one embodiment, such as Figure 3 As shown, a method for detecting breaches is provided, which is applied to... Figure 1 Taking a computer device as an example, the method may include the following steps:

[0070] S202, determine a first mask image of the intermediate tissue based on the acquired original image; the first mask image includes the intermediate tissue and the initial tear region on the intermediate tissue.

[0071] Here, the original image can include the intermediate tissue. Generally, the original image of the intermediate tissue is not acquired separately; instead, it is acquired by acquiring the original image of the tissue to be segmented, which includes the intermediate tissue. The tissue to be segmented can be, for example, the aorta, carotid artery, or iliac artery, or other tissues that have multiple compartments due to aortic dissection and contain intermediate tissue. As an optional embodiment, the intermediate tissue can be an intimal flap or a blood vessel.

[0072] The original image can be obtained by real-time reconstruction of the tissue to be segmented after scanning, or it can be obtained from images of the tissue to be segmented pre-stored in the cloud or on a server. Of course, other methods of obtaining the original image are also possible, and no specific limitation is made here. The original image can be a CTA (CT angiography) image, or other types of CT (Computed Tomography) images or MRI (Magnetic Resonance Imaging) images.

[0073] After obtaining the original image of the tissue to be segmented, a segmentation model or algorithm can be used to segment the tissue and intermediate tissue in the original image to obtain a first mask image of the intermediate tissue. The intermediate tissue included in this first mask image is generally not a complete intermediate tissue. Due to segmentation or the existence of actual breaks, there may be some initial break regions on the intermediate tissue. The number of these initial break regions can be one or more, and the initial break region can be a region composed of one or more points.

[0074] S204, complete the initial tear area on the intermediate tissue to determine the second mask image; the second mask image includes the complete intermediate tissue.

[0075] In this step, after obtaining the intermediate tissue including the initial breach area, morphological methods (such as dilation or morphological opening), mesh filling, or planar or curved surface reconstruction can be used to fill the initial breach area on the intermediate tissue, completing the intermediate tissue into a complete intermediate tissue without breaches.

[0076] Specifically, in this step, the initial tear area of ​​the intermediate tissue can be filled in on the first mask image to obtain the complete intermediate tissue, that is, to obtain the second mask image.

[0077] S206, Process the second mask image based on the original image to determine at least one candidate breach region.

[0078] In this step, the second mask image is obtained after obtaining the first mask image from the original image. Therefore, the positional information of the intermediate tissue in the second mask image and the original image corresponds. Generally, the original image contains more information about the intermediate tissue, while the mask image, being a binary image, contains less information about the intermediate tissue. Here, the image information at the corresponding position of the intermediate tissue can be obtained from the original image using its relative position. Then, this image information is combined with image processing of the intermediate tissue in the second mask image to delineate at least one candidate tear region from the intermediate tissue in the second mask image.

[0079] As an optional embodiment, the second mask image can be processed by image block cropping based on the original image to determine at least one candidate tear region. That is, based on the image information of the corresponding position of the intermediate tissue in the original image, the intermediate tissue in the second mask image is processed by image block cropping to obtain at least one image block, which can be used as a candidate tear region.

[0080] S208, determine the target breach region based on at least one candidate breach region.

[0081] In this step, after obtaining at least one candidate rupture region, each candidate rupture region can be detected to determine whether it is a real rupture region or a non-real rupture region caused by errors in segmenting intermediate tissue. Finally, based on the judgment result, the real rupture region can be selected from the candidate rupture regions and recorded as the target rupture region. This target rupture region can be one or more.

[0082] Furthermore, after obtaining the target rupture area, the target rupture area can be directly pushed to the doctor's end for display, or a bounding box image can be re-extracted from the original image with the target rupture area as the center and displayed to the doctor's end.

[0083] In the aforementioned tear detection method, a first mask image of the intermediate tissue is determined using the original image. The initial tear region on the intermediate tissue is then filled in to determine a second mask image of the complete intermediate tissue. The second mask image is then processed using the original image to obtain at least one candidate tear region. Finally, the target tear region is determined based on the candidate tear regions. This method avoids the problem of low accuracy in tear detection caused by the presence of non-real tear regions in the initial tear region. By filling in the intermediate tissue and further detecting the tear regions, this problem can be avoided, thus improving the accuracy of the final detected tear region.

[0084] The above embodiments mention that the target breach region can be determined by at least one candidate breach region. In order to improve the efficiency and accuracy of the determination of the target breach region, a classification network can be used for determination. The following embodiments will describe the process in detail.

[0085] In another embodiment, another method for detecting breaches is provided. Based on the above embodiments, S208 may include the following steps:

[0086] Step A: At least one candidate breach region is classified using a preset classification network to determine the target breach region.

[0087] The classification network can be a neural network, and there are no specific restrictions on the specific network architecture and type. This classification network can be a binary classification network, mainly used to distinguish whether a candidate breach region is a target breach region or a non-target breach region.

[0088] Before classifying candidate breach regions using a classification network, the network can be pre-trained. During training, this network is trained on multiple sample mask images. Each mask image includes a breach region and its standard category. The standard category characterizes whether a breach region is a target breach region. This standard category can be represented by labels; for example, a breach region with a standard category of 1 is considered a target breach region, while a breach region with a standard category of 0 is considered a non-target breach region. Target breach regions can be considered as the actual breach regions.

[0089] During the training of the initial classification network, the sample mask image is input into the initial classification network to obtain the predicted category of the sample tear region. Then, the loss between the predicted category and the corresponding standard category is calculated, and the loss is backpropagated to adjust the parameters of the initial classification network. Once the loss reaches a threshold or stabilizes, the parameters of the initial classification network are fixed to obtain the trained classification network.

[0090] After the classification network is trained, each candidate breach region obtained above can be input into the classification network to obtain the classification result corresponding to each candidate breach region. The classification result can indicate whether the breach region is the target breach region. Thus, the target breach region in the candidate breach region can be determined by the classification result.

[0091] In this embodiment, a classification network is used to classify candidate breach regions to obtain the target breach region, which improves the efficiency and accuracy of target breach region determination. Furthermore, the classification network is trained based on multiple sample mask images, each of which includes a sample breach region and its standard category. Training the classification network with multiple samples and their labels makes the network more accurate, further enhancing classification accuracy.

[0092] The above embodiments mentioned that the initial tear area on the intermediate tissue can be repaired. The following embodiments will explain in detail how to repair the tear area.

[0093] In another embodiment, another method for detecting breaches is provided, based on the above embodiments, such as... Figure 4 As shown, the above S204 may include the following steps:

[0094] S302, based on the position information of each point on the intermediate tissue in the first mask image, perform surface reconstruction processing on the intermediate tissue and the initial tear area to determine the closed surface corresponding to the intermediate tissue.

[0095] In this step, when obtaining the first mask image, the positional information of each point on the intermediate tissue can also be obtained. Then, using the positional information of each point, surface reconstruction is performed on both the initial and non-initial tear regions on the intermediate tissue using one or more methods such as surface reconstruction, interpolation, and least squares. Finally, the entire intermediate tissue is reconstructed into a complete closed surface. There are no tear regions on this closed surface.

[0096] S304, determine the second mask image based on the closed curved surface and the first mask image.

[0097] In this step, the above-mentioned surface reconstruction process can be performed on the intermediate tissue on the first mask image. In this way, after obtaining the closed surface of the intermediate tissue, a mask image including the closed surface, i.e., the second mask image, can be obtained.

[0098] In this embodiment, the initial tear area is reconstructed using the position information of each point on the intermediate tissue in the first mask image. After obtaining the closed surface of the intermediate tissue, the second mask image is obtained. In this way, the intermediate tissue can be quickly and accurately reconstructed into a closed surface through surface reconstruction, thereby improving the efficiency and accuracy of subsequent determination of the tear area on the intermediate tissue.

[0099] Regarding the specific implementation of the surface reconstruction process mentioned in the above embodiments to obtain the closed surface of the intermediate tissue, the following embodiment provides a possible implementation method.

[0100] In another embodiment, another method for detecting breaches is provided, based on the above embodiments, such as... Figure 5 As shown, the above S302 may include the following steps:

[0101] S402, based on the position information of each point on the intermediate tissue in the first mask image, perform point cloud sparse sampling processing on the intermediate tissue and the initial tear area to determine the point cloud data corresponding to the intermediate tissue.

[0102] As mentioned in S302 above, when obtaining the first mask image, the intermediate tissue in the first mask image can be regarded as a collection of many voxel points. Therefore, the position information of each voxel point can also be obtained here. The voxel points on the intermediate tissue are relatively dense. Therefore, in order to reduce the computational load of subsequent surface reconstruction, the intermediate tissue and the initial tear area in the first mask image can be subjected to point cloud sparse sampling processing. Specifically, the voxel points of the intermediate tissue can be uniformly and sparsely sampled to obtain the sampled voxel points. The data composed of the sampled voxel points can be recorded as the point cloud data of the intermediate tissue.

[0103] For example, taking the intermediate tissue as an endometrial sheet as an example, see [link to relevant documentation]. Figure 6 As shown, the left image is the first mask image of the intermediate tissue, which is the original image before sparse sampling. The right image is the intermediate tissue after sparse sampling. It can be seen that the voxel points of the intermediate tissue after sparse sampling are more evenly distributed and fewer in number. This reduces the computational cost of reconstructing the intermediate tissue and thus improves the reconstruction efficiency.

[0104] S404 performs surface reconstruction processing on point cloud data to determine the closed surface corresponding to the intermediate tissue.

[0105] After obtaining point cloud data of the intermediate tissue through sparse sampling, the orientation of each point in the point cloud data may be inconsistent, which makes surface reconstruction difficult and affects the accuracy of the reconstruction. Therefore, to avoid this problem, as an optional embodiment, the normal vector direction of each point in the point cloud data can be calculated and adjusted to the same direction; the point cloud data after adjusting the normal vector direction is then processed for surface reconstruction to determine the closed surface corresponding to the intermediate tissue.

[0106] In other words, after obtaining point cloud data from sparsely sampled intermediate tissues, a surface centered on the current point can be fitted using the positions of each point and its neighboring points. Simultaneously, the curvature or inflection point of each point can be obtained. Then, the normal vector and its direction for each point can be calculated using methods such as least squares. Finally, the normal vector directions of each point can be corrected using the normal vectors of its neighboring points, ensuring that the normal vectors of all points point to the same surface, i.e., in the same direction, such as pointing to the front of a curved surface.

[0107] For example, taking the intermediate tissue as an endometrial sheet as an example, see [link to relevant documentation]. Figure 7 As shown, the left image shows the normal vectors of each point on the middle tissue pointing in different directions, while the right image shows the normal vectors of each point on the middle tissue pointing in the same direction and on the same surface after the direction is adjusted. It can be seen that the data after the direction is adjusted is more uniform, which can facilitate subsequent surface reconstruction processing.

[0108] After adjusting the normal vectors of all points in the point cloud data to the same surface, a reconstruction algorithm can be used to perform surface reconstruction processing on the point cloud data. This reconstructs the intermediate tissue and its initial breakage area into a complete closed surface, obtaining the closed surface corresponding to the intermediate tissue. Optionally, the reconstruction algorithm here can be the Poisson reconstruction algorithm. By performing surface reconstruction processing through point cloud sparsity processing and the Poisson reconstruction algorithm, a smooth filling surface corresponding to the intermediate tissue can be obtained, making the subsequent judgment of the breakage area of ​​the intermediate tissue more reasonable.

[0109] For example, taking the intermediate tissue as an endometrial sheet as an example, see [link to relevant documentation]. Figure 8 As shown, the left image is the reconstructed closed surface of the intermediate tissue, including the overlapping portion of the reconstructed intermediate tissue and the intermediate tissue segmented from the first mask image, representing a complete intermediate tissue; the middle image is the left end face of the reconstructed intermediate tissue, with the reconstructed intermediate tissue within the dashed box; the right image is the left end face of the intermediate tissue before reconstruction, with the intermediate tissue before reconstruction within the dashed box. It can be seen that the reconstructed intermediate tissue is more complete and is also a closed surface without any breaks. Furthermore, from... Figure 8As can be seen, the aorta can be divided into one or more connected domains by the intimal flap, which can be used for the segmentation of the true and false lumens of the aorta, and at the same time, it can achieve the pre-screening of the tear area of ​​the intimal flap.

[0110] In this embodiment, the initial tear area is first sparsely sampled using the positions of various points on the intermediate tissue in the first mask image to determine the point cloud data of the intermediate tissue. Then, surface reconstruction is performed on this data to obtain the corresponding closed surface. This sparse sampling results in a more uniform distribution and fewer voxel points in the intermediate tissue, reducing the computational load for subsequent reconstruction and thus improving reconstruction efficiency. Furthermore, by adjusting the normal vectors of all points in the point cloud data to the same direction before surface reconstruction, the closed surface of the intermediate tissue is obtained. This adjustment results in more uniform point cloud data, facilitating subsequent surface reconstruction.

[0111] Regarding the specific implementation method of determining candidate tear regions on the mask image for completing intermediate tissue mentioned in the above embodiments, the following embodiment provides a possible implementation method.

[0112] In another embodiment, another method for detecting breaches is provided, based on the above embodiments, such as... Figure 9 As shown, the above S206 may include the following steps:

[0113] S502, Based on the original image, determine the image values ​​corresponding to each point on the complete intermediate tissue in the original image.

[0114] The original image typically includes image values ​​(such as CT values, HU values, etc.), while the second mask image after filling the hole in the intermediate tissue is generally a binarized mask image, which generally does not include image values ​​such as CT values ​​and HU values.

[0115] As mentioned above, the positions of each point on the intermediate tissue in the second mask image correspond to the positions of each point on the original image. Therefore, the image values ​​of each point on the intermediate tissue can be obtained on the original image by mapping the positions of each point on the second mask image to the original image.

[0116] S504, Based on the image values ​​of each point, determine at least one candidate point among the points.

[0117] In this step, a threshold can be obtained from the image values ​​of each point on the original image. For example, the average HU value of blood vessels can be used as the threshold.

[0118] Next, the image values ​​of each point on the intermediate tissue can be compared with the threshold to obtain the comparison result for each point. If the image value of a point is less than or equal to the threshold, the point can be considered as intermediate tissue. If the image value of another point is greater than the threshold, the point can be considered as a suspected breach. In this case, the suspected breach can be used as a candidate point.

[0119] S506, perform image block cropping processing at the location of at least one candidate point in the second mask image to determine at least one candidate breach region.

[0120] After obtaining the candidate points of suspected ruptures, the positions of the candidate points can be mapped onto the second mask image of the complete intermediate tissue to obtain the positions of the corresponding points. Then, at the positions of the corresponding points on the second mask image, image block processing can be performed with a certain image block size to obtain the image block corresponding to the candidate point on the second mask image. This image block can be recorded as the candidate rupture area.

[0121] In this embodiment, by obtaining the image values ​​of each point on the complete intermediate tissue in the original image, and determining candidate points from each point accordingly, the image block of the candidate point sitting in the region can be extracted from the second mask image to obtain the candidate rupture region. In this way, the suspected rupture region on the complete intermediate tissue can be quickly obtained through image values, thus speeding up the detection of the rupture region.

[0122] The above embodiments mention that a mask image of intermediate tissue, including the initial rupture region, can be obtained from the original image. The following embodiments will describe in detail how to determine the mask image.

[0123] In another embodiment, another method for detecting breaches is provided, based on the above embodiments, such as... Figure 10 As shown, the above S202 may include the following steps:

[0124] S602, determine the mask image of the tissue to be segmented based on the acquired original image; wherein the tissue to be segmented includes intermediate tissue.

[0125] S604, determine the first mask image of the intermediate tissue based on the original image and the mask image of the tissue to be segmented.

[0126] As mentioned in S202 above, after obtaining the original image of the tissue to be segmented, a segmentation model or algorithm can be used to segment the tissue in the original image to obtain a segmented image of the tissue to be segmented. This segmented image can be a binary mask image that includes the tissue to be segmented and the background. Using the binary mask image of the tissue to be segmented, background information other than the tissue to be segmented can be excluded, improving the accuracy of identifying intermediate tissues in the subsequent steps. Furthermore, the segmentation model used to segment the tissue can be a neural network model, such as the V-net network; the specific model type is not specifically limited here. During the training process, the segmentation model can be trained using samples that include and do not include intermediate tissues, and the Dice loss method can be used for training.

[0127] Furthermore, since the intermediate tissue can be a part of the tissue to be segmented, a segmentation model or segmentation algorithm can be used to segment the intermediate tissue in the original image and the segmented image of the tissue to be segmented, so as to obtain the first mask image of the intermediate tissue.

[0128] Furthermore, intermediate tissues in the tissue to be segmented often appear in the lumen in a different form than the lumen (e.g., shadowed / darker areas), making them quite distinctive. Therefore, using the original image and the segmented image of the tissue to be segmented can achieve better segmentation results for intermediate tissues. Moreover, by incorporating the segmented image of the tissue to be segmented for intermediate tissue segmentation, the segmentation model or algorithm can focus solely on segmenting the intermediate tissue portion of the original image, resulting in higher efficiency and accuracy. Additionally, the segmentation model for intermediate tissue segmentation can also be a neural network model, such as the V-net network; the specific model type is not limited here. The Mask Dice loss method can be used to train the intermediate tissue segmentation model.

[0129] In this embodiment, the mask image of the intermediate tissue is determined by the mask image of the tissue to be segmented and the original image, which can improve the efficiency and accuracy of obtaining the intermediate tissue.

[0130] The following specific embodiment illustrates the technical solution of this application. Based on the above embodiment, the method may include the following steps:

[0131] S1, the acquired original image is segmented to obtain a mask image of the tissue to be segmented; wherein, the tissue to be segmented includes intermediate tissue; the intermediate tissue is an intima or blood vessel;

[0132] S2, based on the original image and the mask image of the tissue to be segmented, the intermediate tissue is segmented to obtain the first mask image of the intermediate tissue; the first mask image includes the intermediate tissue and the initial tear area on the intermediate tissue;

[0133] S3, based on the position information of each point on the intermediate tissue in the first mask image, perform point cloud sparse sampling processing on the intermediate tissue and the initial tear area to determine the point cloud data corresponding to the intermediate tissue;

[0134] S4, calculate the normal vector direction of each point in the point cloud data, and adjust the normal vector direction of each point to the same direction;

[0135] S5, perform surface reconstruction processing on the point cloud data after adjusting the direction of the normal vector to determine the closed surface corresponding to the intermediate tissue;

[0136] S6, determine the second mask image based on the closed curved surface and the first mask image;

[0137] S7, Based on the original image, determine the image values ​​corresponding to each point on the complete intermediate tissue in the original image;

[0138] S8, Based on the image values ​​of each point, determine at least one candidate point among the points;

[0139] S9, perform image block cropping at the location of at least one candidate point in the second mask image to determine at least one candidate breach region;

[0140] S10, a preset classification network is used to classify at least one candidate breach region to determine the target breach region. The classification network is trained based on multiple sample mask images. Each sample mask image includes a sample breach region and a standard category of the sample breach region. The standard category is used to characterize whether the sample breach region is the target breach region.

[0141] It should be understood that although the steps in the flowcharts of the embodiments described above are shown sequentially according to the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless explicitly stated herein, there is no strict order restriction on the execution of these steps, and they can be executed in other orders. Moreover, at least some steps in the flowcharts of the embodiments described above may include multiple steps or multiple stages. These steps or stages are not necessarily completed at the same time, but can be executed at different times. The execution order of these steps or stages is not necessarily sequential, but can be performed alternately or in turn with other steps or at least some of the steps or stages of other steps.

[0142] Based on the same inventive concept, this application also provides a break detection device for implementing the break detection method described above. The solution provided by this device is similar to the solution described in the above method; therefore, the specific limitations in one or more break detection device embodiments provided below can be found in the limitations of the break detection method described above, and will not be repeated here.

[0143] In one embodiment, such as Figure 11 As shown, a tear detection device is provided, comprising: a first mask determination module, a second mask determination module, a processing module, and a detection module, wherein:

[0144] The first mask determination module is used to determine a first mask image of the intermediate tissue based on the acquired original image; the first mask image includes the intermediate tissue and the initial tear region on the intermediate tissue;

[0145] The second mask determination module is used to complete the initial tear area on the intermediate tissue and determine the second mask image; the second mask image includes the complete intermediate tissue;

[0146] The processing module is used to process the second mask image based on the original image to determine at least one candidate breach region;

[0147] The detection module is used to determine the target breach area based on at least one candidate breach area.

[0148] Optionally, the aforementioned intermediate tissue may be an intima sheet or a blood vessel.

[0149] In another embodiment, another breakage detection device is provided. Based on the above embodiments, the detection module may include:

[0150] The classification unit is used to classify at least one candidate breach region using a preset classification network to determine the target breach region. The classification network is trained based on multiple sample mask images. Each sample mask image includes a sample breach region and a standard category of the sample breach region. The standard category is used to characterize whether the sample breach region is the target breach region.

[0151] In another embodiment, another break detection device is provided. Based on the above embodiments, the second mask determination module may include:

[0152] The reconstruction unit is used to perform surface reconstruction processing on the intermediate tissue and the initial tear area based on the position information of each point on the intermediate tissue in the first mask image, and to determine the closed surface corresponding to the intermediate tissue.

[0153] An image determination unit is used to determine a second mask image based on the closed curved surface and the first mask image.

[0154] In another embodiment, another breach detection device is provided. Based on the above embodiments, the reconstruction unit may include:

[0155] The point cloud processing subunit is used to perform point cloud sparse sampling processing on the intermediate tissue and the initial tear area based on the position information of each point on the intermediate tissue in the first mask image, and to determine the point cloud data corresponding to the intermediate tissue.

[0156] The reconstruction sub-unit is used to perform surface reconstruction processing on point cloud data and determine the closed surface corresponding to the intermediate tissue.

[0157] Optionally, a sub-unit is reconstructed, specifically used to calculate the normal vector direction of each point in the point cloud data and adjust the normal vector direction of each point to the same direction; the point cloud data after adjusting the normal vector direction is processed by surface reconstruction to determine the closed surface corresponding to the intermediate tissue.

[0158] In another embodiment, another breach detection device is provided. Based on the above embodiments, the processing module is specifically used to perform image block cropping processing on the second mask image according to the original image to determine at least one candidate breach region.

[0159] Optionally, the above processing module may include

[0160] The image value determination unit is used to determine the image values ​​of each point on the complete intermediate tissue in the original image based on the original image.

[0161] The candidate point determination unit is used to determine at least one candidate point among the points based on the image values ​​of each point;

[0162] The cropping unit is used to perform image block cropping processing at the location of at least one candidate point in the second mask image to determine at least one candidate break region.

[0163] In another embodiment, another break detection device is provided. Based on the above embodiments, the first mask determination module may include:

[0164] The tissue mask determination unit is used to determine a mask image of the tissue to be segmented based on the acquired original image; wherein the tissue to be segmented includes intermediate tissue.

[0165] The first mask determination unit is used to determine the first mask image of the intermediate tissue based on the original image and the mask image of the tissue to be segmented.

[0166] Each module in the aforementioned breach detection device can be implemented entirely or partially through software, hardware, or a combination thereof. These modules can be embedded in the processor of a computer device in hardware form or independent of it, or stored in the memory of a computer device in software form, so that the processor can call and execute the operations corresponding to each module.

[0167] In one embodiment, a computer device is provided, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to perform the following steps:

[0168] A first mask image of intermediate tissue is determined based on the acquired original image; the first mask image includes intermediate tissue and an initial tear region on the intermediate tissue; the initial tear region on the intermediate tissue is completed to determine a second mask image; the second mask image includes complete intermediate tissue; the second mask image is processed based on the original image to determine at least one candidate tear region; the target tear region is determined based on at least one candidate tear region.

[0169] In one embodiment, the processor, when executing a computer program, also performs the following steps:

[0170] A pre-defined classification network is used to classify at least one candidate breach region to determine the target breach region. The classification network is trained based on multiple sample mask images. Each sample mask image includes a sample breach region and a standard category of the sample breach region. The standard category is used to characterize whether the sample breach region is the target breach region.

[0171] In one embodiment, the processor, when executing a computer program, also performs the following steps:

[0172] Based on the position information of each point on the intermediate tissue in the first mask image, surface reconstruction processing is performed on the intermediate tissue and the initial tear area to determine the closed surface corresponding to the intermediate tissue; the second mask image is determined based on the closed surface and the first mask image.

[0173] In one embodiment, the processor, when executing a computer program, also performs the following steps:

[0174] Based on the position information of each point on the intermediate tissue in the first mask image, sparse sampling processing of point cloud is performed on the intermediate tissue and the initial tear area to determine the point cloud data corresponding to the intermediate tissue; surface reconstruction processing is performed on the point cloud data to determine the closed surface corresponding to the intermediate tissue.

[0175] In one embodiment, the processor, when executing a computer program, also performs the following steps:

[0176] Calculate the normal vector direction of each point in the point cloud data and adjust the normal vector direction of each point to the same direction; perform surface reconstruction processing on the point cloud data after adjusting the normal vector direction to determine the closed surface corresponding to the intermediate tissue.

[0177] In one embodiment, the processor, when executing a computer program, also performs the following steps:

[0178] Based on the original image, the second mask image is processed by image block cropping to determine at least one candidate breach region.

[0179] In one embodiment, the processor, when executing a computer program, also performs the following steps:

[0180] Based on the original image, determine the image values ​​corresponding to each point on the complete intermediate tissue in the original image; based on the image values ​​of each point, determine at least one candidate point among each point; perform image block cropping processing at the location of at least one candidate point in the second mask image to determine at least one candidate breach region.

[0181] In one embodiment, the processor, when executing a computer program, also performs the following steps:

[0182] A mask image of the tissue to be segmented is determined based on the acquired original image; wherein the tissue to be segmented includes intermediate tissue; a first mask image of the intermediate tissue is determined based on the original image and the mask image of the tissue to be segmented.

[0183] In one embodiment, the aforementioned intermediate tissue is an intima sheet or a blood vessel.

[0184] In one embodiment, a computer-readable storage medium is provided having a computer program stored thereon, the computer program performing the following steps when executed by a processor:

[0185] A first mask image of intermediate tissue is determined based on the acquired original image; the first mask image includes intermediate tissue and an initial tear region on the intermediate tissue; the initial tear region on the intermediate tissue is completed to determine a second mask image; the second mask image includes complete intermediate tissue; the second mask image is processed based on the original image to determine at least one candidate tear region; the target tear region is determined based on at least one candidate tear region.

[0186] In one embodiment, when the computer program is executed by a processor, it also performs the following steps:

[0187] A pre-defined classification network is used to classify at least one candidate breach region to determine the target breach region. The classification network is trained based on multiple sample mask images. Each sample mask image includes a sample breach region and a standard category of the sample breach region. The standard category is used to characterize whether the sample breach region is the target breach region.

[0188] In one embodiment, when the computer program is executed by a processor, it also performs the following steps:

[0189] Based on the position information of each point on the intermediate tissue in the first mask image, surface reconstruction processing is performed on the intermediate tissue and the initial tear area to determine the closed surface corresponding to the intermediate tissue; the second mask image is determined based on the closed surface and the first mask image.

[0190] In one embodiment, when the computer program is executed by a processor, it also performs the following steps:

[0191] Based on the position information of each point on the intermediate tissue in the first mask image, sparse sampling processing of point cloud is performed on the intermediate tissue and the initial tear area to determine the point cloud data corresponding to the intermediate tissue; surface reconstruction processing is performed on the point cloud data to determine the closed surface corresponding to the intermediate tissue.

[0192] In one embodiment, when the computer program is executed by a processor, it also performs the following steps:

[0193] Calculate the normal vector direction of each point in the point cloud data and adjust the normal vector direction of each point to the same direction; perform surface reconstruction processing on the point cloud data after adjusting the normal vector direction to determine the closed surface corresponding to the intermediate tissue.

[0194] In one embodiment, when the computer program is executed by a processor, it also performs the following steps:

[0195] Based on the original image, the second mask image is processed by image block cropping to determine at least one candidate breach region.

[0196] In one embodiment, when the computer program is executed by a processor, it also performs the following steps:

[0197] Based on the original image, determine the image values ​​corresponding to each point on the complete intermediate tissue in the original image; based on the image values ​​of each point, determine at least one candidate point among each point; perform image block cropping processing at the location of at least one candidate point in the second mask image to determine at least one candidate breach region.

[0198] In one embodiment, when the computer program is executed by a processor, it also performs the following steps:

[0199] A mask image of the tissue to be segmented is determined based on the acquired original image; wherein the tissue to be segmented includes intermediate tissue; a first mask image of the intermediate tissue is determined based on the original image and the mask image of the tissue to be segmented.

[0200] In one embodiment, the aforementioned intermediate tissue is an intima sheet or a blood vessel.

[0201] In one embodiment, a computer program product is provided, including a computer program that, when executed by a processor, performs the following steps:

[0202] A first mask image of intermediate tissue is determined based on the acquired original image; the first mask image includes intermediate tissue and an initial tear region on the intermediate tissue; the initial tear region on the intermediate tissue is completed to determine a second mask image; the second mask image includes complete intermediate tissue; the second mask image is processed based on the original image to determine at least one candidate tear region; the target tear region is determined based on at least one candidate tear region.

[0203] In one embodiment, when the computer program is executed by a processor, it also performs the following steps:

[0204] A pre-defined classification network is used to classify at least one candidate breach region to determine the target breach region. The classification network is trained based on multiple sample mask images. Each sample mask image includes a sample breach region and a standard category of the sample breach region. The standard category is used to characterize whether the sample breach region is the target breach region.

[0205] In one embodiment, when the computer program is executed by a processor, it also performs the following steps:

[0206] Based on the position information of each point on the intermediate tissue in the first mask image, surface reconstruction processing is performed on the intermediate tissue and the initial tear area to determine the closed surface corresponding to the intermediate tissue; the second mask image is determined based on the closed surface and the first mask image.

[0207] In one embodiment, when the computer program is executed by a processor, it also performs the following steps:

[0208] Based on the position information of each point on the intermediate tissue in the first mask image, sparse sampling processing of point cloud is performed on the intermediate tissue and the initial tear area to determine the point cloud data corresponding to the intermediate tissue; surface reconstruction processing is performed on the point cloud data to determine the closed surface corresponding to the intermediate tissue.

[0209] In one embodiment, when the computer program is executed by a processor, it also performs the following steps:

[0210] Calculate the normal vector direction of each point in the point cloud data and adjust the normal vector direction of each point to the same direction; perform surface reconstruction processing on the point cloud data after adjusting the normal vector direction to determine the closed surface corresponding to the intermediate tissue.

[0211] In one embodiment, when the computer program is executed by a processor, it also performs the following steps:

[0212] Based on the original image, the second mask image is processed by image block cropping to determine at least one candidate breach region.

[0213] In one embodiment, when the computer program is executed by a processor, it also performs the following steps:

[0214] Based on the original image, determine the image values ​​corresponding to each point on the complete intermediate tissue in the original image; based on the image values ​​of each point, determine at least one candidate point among each point; perform image block cropping processing at the location of at least one candidate point in the second mask image to determine at least one candidate breach region.

[0215] In one embodiment, when the computer program is executed by a processor, it also performs the following steps:

[0216] A mask image of the tissue to be segmented is determined based on the acquired original image; wherein the tissue to be segmented includes intermediate tissue; a first mask image of the intermediate tissue is determined based on the original image and the mask image of the tissue to be segmented.

[0217] In one embodiment, the aforementioned intermediate tissue is an intima sheet or a blood vessel.

[0218] It should be noted that all data involved in this application (including but not limited to data used for analysis, data stored, data displayed, etc.) are data that have been fully authorized by all parties.

[0219] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium, and when executed, it can include the processes of the embodiments of the above methods. Any references to memory, databases, or other media used in the embodiments provided in this application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetic random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM can take many forms, such as Static Random Access Memory (SRAM) or Dynamic Random Access Memory (DRAM). The databases involved in the embodiments provided in this application may include at least one type of relational database and non-relational database. Non-relational databases may include, but are not limited to, blockchain-based distributed databases. The processors involved in the embodiments provided in this application may be general-purpose processors, central processing units, graphics processing units, digital signal processors, programmable logic devices, quantum computing-based data processing logic devices, etc., and are not limited to these.

[0220] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0221] The embodiments described above are merely illustrative of several implementation methods of this application, and while the descriptions are specific and detailed, they should not be construed as limiting the scope of this patent application. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of this application, and these all fall within the protection scope of this application. Therefore, the protection scope of this application should be determined by the appended claims.

Claims

1. A method for detecting breaches, characterized in that, The method includes: A first mask image of the intermediate tissue is determined based on the acquired original image; the first mask image includes the intermediate tissue and an initial tear area on the intermediate tissue; the original image is a tissue image to be segmented including the intermediate tissue, which is an intima sheet or a blood vessel; A hole-filling operation is performed on the initial tear area on the intermediate tissue to complete the initial tear area and obtain a second mask image including the complete intermediate tissue. The second mask image is processed based on the original image to determine at least one candidate breach region; The target breach region is determined based on the at least one candidate breach region.

2. The method according to claim 1, characterized in that, Determining the target breach region based on the at least one candidate breach region includes: The at least one candidate breach region is classified using a preset classification network to determine the target breach region; The classification network is trained based on multiple sample mask images. Each sample mask image includes a sample breakage region and a standard category for the sample breakage region. The standard category is used to characterize whether the sample breakage region is a target breakage region.

3. The method according to claim 1, characterized in that, The step of completing the initial tear area on the intermediate tissue to determine the second mask image includes: Based on the position information of each point on the intermediate tissue in the first mask image, surface reconstruction processing is performed on the intermediate tissue and the initial tear area to determine the closed surface corresponding to the intermediate tissue. The second mask image is determined based on the sealed curved surface and the first mask image.

4. The method according to claim 3, characterized in that, The step of performing surface reconstruction processing on the intermediate tissue and the initial tear region based on the position information of each point on the intermediate tissue in the first mask image to determine the closed surface corresponding to the intermediate tissue includes: Based on the position information of each point on the intermediate tissue in the first mask image, point cloud sparse sampling processing is performed on the intermediate tissue and the initial tear area to determine the point cloud data corresponding to the intermediate tissue. The point cloud data is subjected to surface reconstruction processing to determine the closed surface corresponding to the intermediate tissue.

5. The method according to claim 4, characterized in that, The step of performing surface reconstruction processing on the point cloud data to determine the closed surface corresponding to the intermediate tissue includes: Calculate the normal vector direction of each point in the point cloud data, and adjust the normal vector direction of each point to the same direction; The point cloud data after adjusting the normal vector direction is subjected to surface reconstruction processing to determine the closed surface corresponding to the intermediate tissue.

6. The method according to any one of claims 1-5, characterized in that, The step of processing the second mask image based on the original image to determine at least one candidate breach region includes: Based on the original image, the second mask image is processed by image block cropping to determine at least one candidate breach region.

7. A breakage detection device, characterized in that, The device includes: The first mask determination module is used to determine a first mask image of the intermediate tissue based on the acquired original image; the first mask image includes the intermediate tissue and an initial tear area on the intermediate tissue; the original image is a tissue image to be segmented including the intermediate tissue, wherein the intermediate tissue is an intima sheet or a blood vessel; The second mask determination module is used to perform a hole-filling operation on the initial tear area on the intermediate tissue to fill the initial tear area and obtain a second mask image including the complete intermediate tissue. The processing module is used to process the second mask image based on the original image to determine at least one candidate breach region; The detection module is used to determine the target breach region based on the at least one candidate breach region.

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

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

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

Images